thermoregulatory responses have been known to deteriorate with aging, which is likely associated with the decrease in peak oxygen consumption rate (V˙o2 peak). Tankersley et al. (32) suggested that 50–60% of the reduction in forearm skin blood flow (FBF) response to increased esophageal temperature (Tes) in older subjects was caused by a decrease in V˙o2 peak by performing a study that compared the FBF response with that in younger subjects whose V˙o2 peak was matched. Moreover, Havenith et al. (12) reconfirmed the results in a greater number of subjects, providing a regression equation to estimate the highest FBF during exercise at 65–70%V˙o2 peak fromV˙o2 peak and age. In addition to these cross-sectional studies, Thomas et al. (33) reported a 16-wk aerobic training for older men increased the FBF response in addition to theV˙o2 peak. Although these results suggest a close association between enhanced FBF response and increasedV˙o2 peak after exercise training in older subjects, the precise mechanisms remain unknown. Show Aerobic training increases blood volume (BV) in younger subjects, and the increased BV is considered to be one of the mechanisms involved in training-associated enhancement of FBF response by increasing the venous return to the heart (26, 31). Indeed, the maneuvers to increase the venous return to the heart, such as by intravenous saline infusion (22), head-out water immersion (21), and continuous negative pressure breathing (20), have been reported to enhance FBF response by stretching baroreceptors in addition to by increasing cardiac stroke volume during exercise in a hot environment. Hagberg et al. (9) suggested that in older subjects those who were aerobic trained showed greater BV than their sedentary counterparts, leading to higher cardiac stroke volume and cardiac output at a given relative intensity of exercise. Ho et al. (13) suggested that hypervolemia is the lone mechanism for the increased FBF response because splanchnic and renal vasoconstriction during exercise was not increased after aerobic training in older men, unlike in younger men. These results suggest that hypervolemia after aerobic training improves FBF response by increasing the venous return to the heart in older subjects. However, it has been controversial whether aerobic training increases BV in older subjects, and if it does, it is yet unknown how the increased BV enhances the FBF response. Some studies have reported that aerobic training increased BV in older subjects (4, 24), but others did not (29, 30, 35). In addition, few of these studies reported change in the FBF response after training. One study by Ho et al. (13) suggested that the enhanced FBF response was caused by increased cardiac output due to increased plasma volume (PV) after a 4-wk aerobic training, but they found no significant increase in PV because of too small number of subjects. In the present study, we examined the effect of 8- and 18-wk aerobic or resistance training on BV and FBF response in older subjects to elucidate the involvement of increased BV in the exercise training-induced enhancement of FBF response in older men. The reason for adding a resistance training trial was that the training enabled us to distinguish the mere effect of increasedV˙o2 peak on FBF response from other effects induced by aerobic training, such as more prolonged cardiovascular and/or heat loading. In addition, we measured the changes in sweat rate (SR) response to increased Tes after exercise training because, to our knowledge, there have been no studies on the effects of exercise training on this factor in older men. In addition, we assessed Tes thresholds for forearm skin vascular conductance (FVC) and SR responses and slopes of the response-Tes relationships to examine howV˙o2 peak and/or BV modify the responses to increased Tes. METHODSThe procedures in this study were approved by the Review Board on Human Experiments, Shinshu University School of Medicine. After the experimental protocols were fully explained, 23 older (58–72 yr) healthy men gave their written informed consent before participating in this study. The subjects were relatively active but did not participate in any regular exercise training program. All subjects were nonsmokers and had no overt history of cardiovascular or pulmonary diseases or any orthopedic limitations to the exercising test and training. During the experiment, no subject was taking medication that had a potential to impact cardiovascular and thermoregulatory function or BV and constituents. Subjects were randomly divided into the three training trials for 18 wk [control (C; n = 7), resistance training (RT;n = 8), and aerobic training (AT; n = 8); Table 1] to avoid differences in physical characteristics among the trials before training. In the C trial, subjects were not engaged in a specific training program except for walking of 9,465 ± 1,954 steps/day, 6.7 ± 0.2 days/wk. Subjects in the RT and AT trials trained under our supervision in addition to the walking of 10,353 ± 2,336 and 8,749 ± 490 steps/day, respectively. The training was performed between September and April to avoid any effect of heat acclimatization during the summer season. Averaged ambient temperature (Ta) in the city was 19°C in September, −1°C in January, and 4°C in April. Relative humidity (RH) was ∼70% throughout the period.
V˙o2 peak, ventilation threshold (VT), BV and constituents, and muscle strength for isometric knee extension were measured in all subjects before and after 8-wk and 18-wk training. FBF and SR were also measured during exercise in a hot environment. V˙o2 peak was measured while the subjects were in an upright position with the use of a cycle ergometer at Ta of 25.0 ± 0.1°C and RH of 46 ± 1% (means ± range). After baseline measurements at rest were taken for 3 min, the subjects started pedaling bicycles at 60 cycles/min without loading. Exercise intensity was increased by 30 W every 3 min until 120 W, and, above this intensity, it was increased by 15 W every 2 min until subjects could not maintain the rhythm. Oxygen consumption rate (V˙o2) was determined every 15 s from the oxygen and carbon dioxide fractions in expired gas and the expired ventilatory volume (Aeromonitor AE260, Minato, Tokyo, Japan). Heart rate (HR) was recorded every 1 min from the trace of an electrocardiogram (Life Scope 8, Nihon Kohden, Tokyo, Japan).V˙o2 peak was determined after the three largest consecutive values at the end of exercise were averaged. The criteria for determiningV˙o2 peak were that the respiratory exchange ratio was >1.1, V˙o2 leveled off despite increasing workload, and HR reached the age-predicted maximal value. VT was determined by the V-slope method and presented asV˙o2 at VT (2). Muscle strength for isometric extension was measured in each side of the knee with a dynamometer (Biodex 3, Biodex Medical System, Shirley, NY). After regular warming-up and familiarization protocols, the anatomic axis of the knee joint was aligned with the mechanical axis of the dynamometer arm to adjust the angle between the lower and upper legs to 105°. Then, three 3-s maximum voluntary contractions, intermitted by a 30-s recovery, were conducted. The peak torque averaged for three trials was adopted for the value for one side of the knee, and it is given as an averaged value of both sides of the knee in Table 1. On the day of the measurement, subjects reported to the laboratory at 7:00 AM normally hydrated but without having taken any food for 8 h before the experiment. PV was determined by the Evans blue dye-dilution method (7). The background absorbance due to turbidity was corrected by using a regression equation on the relationship between 620 and 740 nm, previously determined on 64 control plasma samples in 22 subjects according to the method reported elsewhere (5, 30). BV was calculated from PV and hematocrit (Hct) values after correction for plasma trapped among the red blood cells in the Hct tube (0.96) and an F-cell ratio (0.91) (8). The measurement error of BV was 2.1 ± 1.8% (means ± SD) (n = 4), which was obtained by measuring BV twice in the same subjects with a BV of 64.9–93.5 ml/kg after 2- to 3-wk intervals. The residues of blood samples drawn before the injection of the dye were used to determine the Hct (microcentrifuge method) and plasma albumin concentrations ([Alb]p; colorimetry). Total albumin content in plasma (Albtot) was determined as a product of PV and [Alb]p. The BV measurement was not performed in one of eight subjects in the AT trial who showed an allergic reaction to the patch test of the dye performed 24 h before the measurement on every subject. Subjects reported to the laboratory normally hydrated but having fasted for at least 2 h before the measurement, at the same time of day before and after the training regimens to avoid any effect of circadian rhythm. Clad in shorts and shoes, subjects emptied their bladders, entered the chamber controlled at 30.0 ± 0.1°C of Ta and 50 ± 1% of RH (means ± range), and sat in the contour chair of the cycle ergometer in a semirecumbent position for 45 min while all measurement devices were applied. After baseline measurements were taken at rest for 10 min, the subjects exercised in a semirecumbent position at 60% of their pretrainingV˙o2 peak for 20 min without fan cooling. Tes was monitored with a thermocouple in a polyethylene tubing (PE-90). The tip of the tube was advanced at a distance of one-fourth of the subject's standing height from the external nares. Mean skin temperature (T̄sk) was determined asT̄sk = 0.25 · Tfa + 0.43 · Tch + 0.32 · Tth(25), where Tfa, Tch, and Tth are skin surface temperature at the forearm, chest, and thigh measured with the thermocouples, respectively. SR was determined by capacitance hygrometry, calculated from the relative humidity and temperature of the air (THP-B3T, Shinei, Tokyo, Japan) flowing out of a 12.56-cm2 capsule at the rate of 1.5 l/min on the chest at 5 cm below the left clavicle. FBF was measured by venous occlusion plethysmography with a mercury-in-Silastic tube strain gauge placed around the upper side of the subject's left forearm positioned above the heart level, with the hand eliminated from the circulation by inflation of an occlusion cuff to a supra-arterial pressure (∼280 mmHg) (34). HR was recorded every 1 min as described inV˙o2 peak and VT. Systolic (SAP) and diastolic arterial blood pressures (DAP) were measured every 1 min from the right upper arm at the heart level by inflation of the cuff with a sonometric pickup of Korotkoff's sound (STPB-780, Colin, Komaki, Japan). Mean arterial blood pressure (MAP) was calculated as DAP + (SAP − DAP)/3. FVC was calculated as FBF/MAP (reported in units of ml · 100 ml−1 · min−1 · 100 mmHg−1). Tes,Tsk, and SR were recorded every 5 s, and FBF was measured twice every 1 min at rest and during exercise and presented every 1 min as an average. The Tes thresholds for increasing SR (THSR) and increasing FVC (THFVC) were determined on each subject as the Tes at 2–5 min after the start of exercise where SR or FVC increased above the baselines. The slopes of an increase in SR (SR/Tes) and FVC (FVC/Tes) at a given increase in Tes were determined on each subject from a linear regression equation on the measurements recorded at 5–20 min of exercise. Subjects in the RT and AT trials trained for 18 wk according to the protocol recommended by the American College of Sports Medicine (1). As warming-up and cooling-down protocols, subjects in the AT and RT trials performed a 5-min stretch exercise and a 5-min cycle ergometer exercise at 50%V˙o2 peak before and after the main exercise. Subjects in the RT trial performed an exercise protocol, consisting of a knee extension and flexion, chest press, pull-dip and arm curl with weight resistance machines (Athlete, Mizuno, Tokyo, Japan) at 60–80% of one repetition maximum (1 RM), two to three sets of eight repetitions per day, 3 days/wk. The exercise intensity was increased with the training days: two sets of each exercise at 60, 70, and 75% 1 RM in the 1st, 2nd, and 3rd wk, respectively, and three sets at 80% 1 RM after the 4th wk. In addition to the exercise, supportive upper back extension, pelvic rise, and crunch without weight loading were performed throughout the training period. Subjects in the AT trial performed a cycle ergometer exercise at 50–80% of V˙o2 peak for 60 min/day, consisting of four sets of 15-min exercise followed by a 5-min rest. The exercise intensity was increased with the training days: 50, 60, and 65% V˙o2 peak for the 1st, 2nd, and 3rd wk, respectively; 70%V˙o2 peak for the 4th to 8th wk, 75%V˙o2 peak for the 9th to 10th wk; and 80% V˙o2 peak after the 11th wk. HR was continuously monitored and recorded every 5 min during exercise. The exercise intensity was readjusted every 1 wk so that HR at 5 min of exercise was equivalent to the target exercise intensity. The environmental condition for the training room was controlled at Ta of ∼20°C and RH of ∼50% without any significant differences between the RT and AT trials. During exercise, the subjects were allowed access to water ad libitum, and the amount was monitored. Subjects were weighed before and after the training regimen each day to estimate sweat loss. Body weight loss after training per day was 4–6 ml/kg body wt for the RT trial and 8–10 ml/kg body wt for the AT trial. The effects of training on physical characteristics, BV, blood constituents, THSR, THFVC, SR/Tes, and FVC/Tes within each trial were tested by a 3 (C, RT, AT) × 3 (before, 8 wk, and 18 wk) ANOVA for repeated measures (Table 1 and see Table 3). The effects of training on cardiovascular and thermoregulatory responses in a hot environment within each trial were tested by three-way ANOVA for repeated measures (Table 2). Subsequent post hoc tests to determine significant differences in the various pairwise comparisons were performed by using Scheffé's test. The null hypothesis was rejected when there were values of P < 0.05. Regression analyses were performed by Brace's methods (3). Because two of eight subjects in the RT trial quit the last 10-wk training regimen, the comparison between 8- and 18-wk training was performed on only six subjects. Values are expressed as means ± SE for seven subjects in the C trial and for eight subjects in the RT and AT trials, except as noted.
RESULTSTable 1 shows the physical characteristics, BV, and PV before and after training. After 8-wk training,V˙o2 peak increased by 8.4 ± 2.9% (P < 0.01) in the RT trial and by 13.2 ± 2.4% in the AT trial (P < 0.0001) with respect to the pretraining values. After 18-wk training, it further increased by 9.7 ± 5.1% in the RT trial (P < 0.003) and by 20.0 ± 2.5% in the AT trial (P < 0.0001), whereas it remained unchanged in the C trial. There were no significant changes in body weight, maximal heart rate (HRmax), BV, PV, and [Alb]p after 8- and 18-wk training. Table 2 shows HR, MAP, Tes, and T̄skduring exercise in a hot environment before and after 8- and 18-wk training in three trials. Only the values at rest and at 5 and 20 min after the start of exercise are presented in the table to simplify. After 8- and 18-wk training, HR at rest decreased in the RT and AT trials but not in the C trial. The increase in HR during exercise was reduced in the AT trial but was enhanced in the RT and C trials. MAP at rest decreased significantly in all trials. The increase in MAP during exercise was reduced at 5 and 20 min in the AT trial, at 5 min in the RT trial, and at 20 min in the C trial, but it was enhanced at 20 min in the RT trial. Tes at rest decreased significantly in all trials. The increase in Tes during exercise was attenuated at 5 and 20 min in the RT and AT trials and at 5 min in the C trial.T̄sk at rest was not altered in any trials. The increase in T̄sk during exercise was reduced at 5 and 20 min in the AT trial and at 20 min in the C trial, but it increased at 5 min in the C trial. The SR and FVC responses to increased Tes during exercise in a hot environment are shown in Fig.1. THSR, THFVC, SR/Tes, and FVC/Tes, are summarized in Table3. THSR decreased by 0.22 and 0.28°C in the RT trial and by 0.15 and 0.17°C in the AT trial, after 8-wk and 18-wk training, respectively, but it did not change significantly in the C trial. Similarly, THFVC decreased by 0.27 and 0.32°C in the RT trial and by 0.15 and 0.29°C in the AT trial, after 8- and 18-wk training, respectively, but it did not change significantly in the C trial. There were no significant changes in SR/Tes and FVC/Tes before and after training in any trials. Fig. 1.Sweat rate (SR) and forearm skin vascular conductance (FVC) responses to increased esophageal temperature (Tes) during exercise in a warm environment with ambient temperature of 30°C and relative humidity of 50% at the intensity of 60% of pretraining peak oxygen consumption rate. A: control (C; circles). B: resistance training (RT; diamonds).C: aerobic training (AT; squares). Open symbols, before training; gray symbols, 8-wk training; solid symbols, 18-wk training. Bars indicate means ± SE. Regression analyses were performed on the measurements from 5 to 20 min after the start of exercise.
When the data from all the trials were pooled, the change inV˙o2 peak after training was weakly but significantly correlated with those in THSR(ΔTHSR; r = 0.30, P < 0.05) and THFVC (ΔTHFVC; r = 0.34, P < 0.03) but not with those in SR/Tes [Δ(SR/Tes); P > 0.05] or FVC/Tes [Δ(FVC/Tes);P > 0.4]. In contrast, the change in BV (ΔBV) after training was significantly correlated with Δ(SR/Tes) (r = 0.51, P < 0.0005) and Δ(FVC/Tes) (r = 0.45, P< 0.005) (Fig. 2, A andB) but not with ΔTHSR (P > 0.1) or ΔTHFVC (P > 0.3). As shown in Fig. 3 (A and B), ΔTHSR was significantly correlated with ΔTHFVC (r = 0.79, P < 0.0001), and Δ(SR/Tes) was significantly correlated with Δ(FVC/Tes) (r = 0.63, P< 0.0001). As shown in Fig. 4, the change in Albtot after 8- and 18-wk training was significantly correlated with that in PV (r = 0.68,P < 0.0001). Fig. 2.Relationship between the change in slope of an increase in SR at a given increase in Tes[Δ(SR/Tes)] and change in blood volume (ΔBV) (A) and relationship between the change in slope of an increase in FVC at a given increase in Tes[Δ(FVC/Tes)] and ΔBV (B) after training. Open symbols, changes between before training and 8-wk training; solid symbols, changes between 8- and 18-wk training. ΔBV was significantly correlated with Δ(SR/Tes) (r = 0.51,P < 0.0005; y = 0.23x− 0.12) and also with Δ(FVC/Tes) (r = 0.45, P < 0.005; y = 2.44x − 0.80). Fig. 3.Relationship between the change in Testhreshold for increasing FVC (ΔTHFVC) and Testhreshold for increasing SR (ΔTHSR) (A) and relationship between the change in Δ(SR/Tes) and Δ(FVC/Tes) (B) after training. Open symbols, changes between before training and 8-wk training; solid symbols, changes between 8-wk and 18-wk training. ΔTHSR was significantly correlated with ΔTHFVC (r = 0.79, P < 0.0001; y = 0.86x − 0.00). Δ(SR/Tes) was significantly correlated with Δ(FVC/Tes) (r = 0.63, P < 0.0001;y = 0.10x − 0.04). Fig. 4.Relationship between changes in plasma volume (ΔPV) and total albumin content in plasma (ΔAlbtot) after training. Open symbols, changes between before training and 8-wk training; solid symbols, changes between 8-wk and 18-wk training.The 2 parameters were significantly correlated (r = 0.68, P< 0.0001; y = 21.1x + 0.48). DISCUSSIONIn the present study, we verified the results previously reported in older men that BV did not increase after aerobic training (29,30, 35) and that THFVC and THSRdecreased with the increase inV˙o2 peak, whereas FVC/Tes and SR/Tes remained unchanged (33). Moreover, we confirmed the results not only after aerobic but also after resistance training. In addition, we clarified that the reductions in THFVC and THSR were more associated with increased V˙o2 peakthan with increased BV, whereas changes in FVC/Tes and SR/Tes were more associated with that in BV thanV˙o2 peak. As shown in Table 1, V˙o2 peakin the RT and AT trials increased after 8- or 18-wk training. The reductions in THFVC and THSR in the RT and AT trials were weakly but significantly correlated with the increase inV˙o2 peak. THFVC or THSR at a given absolute exercise intensity has been reported to decrease after aerobic training not only in younger (18, 25) but also in older subjects (33). Smolander et al. (28) demonstrated that THFVCincreased with relative exercise intensity in individual younger subjects. Thomas et al. (33) reported that 16-wk aerobic training decreased THFVC in subjects who increasedV˙o2 peak by >5%. Moreover, Ho et al. (13) suggested that, in older subjects, THFVC was not altered after a 4-wk training even when absolute exercise intensity was increased from 60 to 70% of pretraining V˙o2 peak, equivalent to 60% of posttrainingV˙o2 peak. These results suggest that the reduction in THFVC and/or THSRafter training was associated with reduced relative exercise intensity due to increased V˙o2 peak. BV in the AT trial did not increase, as shown in Table 1. It has been reported that the response of BV to aerobic training was lower in older subjects than in younger subjects (29, 30, 35). This may be caused by reduced fluid intake after thermal dehydration (17, 30) or water deprivation (23) in older men. Recently, Takamata et al. (30) studied the changes in body fluid response to dehydration before and after an exercise heat acclimatization regimen (80-min bicycle exercise at 40%V˙o2 peak per day for 6 days at 36°C of Ta and RH of 40%) and compared the results between older and younger men. They suggested that BV remained unchanged in older subjects, whereas it increased by ∼5% in younger men. They also suggested that recovery from body fluid loss during 2-h rehydration was twofold higher in younger men than that in older subjects and that the recovery was augmented after heat acclimatization in younger men but not in older men. They ascribed the results to the attenuated water intake and the reduced release of body fluid-retention hormones during rehydration in older men. Although in the present study, aerobic training was performed in a cooler environment and body fluid loss was less than in previous studies (30), the blunted body fluid conservation mechanisms in older men may be involved in no increase in BV for the AT trial. Another possible explanation for no increase in BV for the AT trial may be associated with no increase in Albtot for older men (Table 1 and Fig. 4). The exercise training-induced hypervolemia has been suggested to be dependent on an increase in Albtot, causing a fluid shift from the interstitial to intravascular fluid space according to the colloid osmotic pressure gradient between the spaces (10, 19, 27). In younger subjects, exercise training-induced hypervolemia has been reported to be typically accompanied by an increase in Albtot (10, 19,27). On the other hand, Zappe et al. (35) reported that, in older men, PV did not increase after 4 days of repeated exercise with a cycle ergometer because of attenuated increases in Albtot. They suggested that the failure to increase Albtot in older men after exercise was caused by the lower ability to synthesize (19) or translocate protein into the intravascular space than that reported in younger men (11). The interindividual variation in the increase in Albtot for the present study may be related to factors other than the active exercise training regimens, protein in diet (14), or heat acclimatization (27). As shown in Table 1, the increasedV˙o2 peak in the RT and AT trials was not accompanied by hypervolemia in older subjects. However, in younger subjects, it has been suggested that hypervolemia after aerobic training increased V˙o2 peak by increasing venous return to the heart and maximal cardiac stroke volume (26, 31). Frontera et al. (6) reported that, in older subjects, 12-wk strength training induced a 6% increase inV˙o2 peak and a 107% increase in 1 RM of the knee extensor, but they found no increase in BV. Recently, Jubrius et al. (15) studied the cellular energetic adaptation to 6-mo aerobic or resistance training in older subjects and reported that oxidative capacity increased by 31 and 57% after aerobic and resistance training, respectively. Because muscle strength for knee extensor in the AT trial increased by the same degree as that in the RT trial (Table 1), the increase inV˙o2 peak for the AT trial was caused by the increased oxidative capacity or oxygen extraction rate in the lower leg muscles. As shown in Fig. 2, ΔBV was positively correlated with Δ(FVC/Tes) and Δ(SR/Tes). To our knowledge, there have been no studies showing the effects of exercise training-induced hypervolemia on the slopes in older subjects. In younger subjects, the maneuvers to increase the venous return to the heart [saline infusion (22), head-out water immersion (21), or continuous negative pressure breathing (20)] increase FVC/Tes during exercise. These results suggest that increased BV enhances the FBF response by increasing cardiac output and/or by suppressing baroreflex-induced attenuation of skin vasodilation by increasing venous return to the heart in older subjects. Ho et al. (13) reported that a 4-wk aerobic training enhanced the FBF response during exercise of 60% of V˙o2 peak in a hot environment. They ascribed this to increased cardiac output by PV expansion, although they found no significant increase in PV before and after training as a result of the small number of subjects. Coupled with the results of the present study, it is suggested that the slopes were increased by hypervolemia, irrespective of the increase inV˙o2 peak in older men. The significant correlations between ΔTHFVC and ΔTHSR (Fig. 3A) and between Δ(FVC/Tes) and Δ(SR/Tes) (Fig.3B) suggested the close association of the active vasodilator and sudomotor systems (16). Mack et al. (16) demonstrated in young subjects that reduction of central venous pressure by lower body negative pressure decreased not only FVC/Tes but also SR/Tes during exercise, suggesting that the reductions were caused by suppression of the sudomotor and active vasodilator systems by unloading cardiopulmonary baroreceptors. Thus the sudomotor and active vasodilator systems are closely associated during dynamic exercise. We confirmed this in older men after exercise training. Summarizing these results, aerobic and/or resistance training in older men improved FVC and SR responses by the downward shift of THFVC and THSR rather than by their increased slopes of FVC/Tes and SR/Tes, which was associated more with the increasedV˙o2 peak than with BV regardless of trials. In contrast, the change in the slopes was associated more with the change in BV, which was not necessarily accompanied by increasedV˙o2 peak after training in older men. We thank the volunteer subjects for participating in this study. We also thank Drs. A. Takamata, Y. Yanagidaira, A. Sakai, and H. Endoh for helpful comments and discussion on this study. FOOTNOTESREFERENCES
Page 2glucose transport is the movement of glucose across the plasma membrane, whereas glucose uptake is the transport and phosphorylation of glucose by hexokinase that results in its clearance from the surrounding medium (13). Insulin binding to its receptor activates several intracellular signaling proteins, which ultimately leads to an increase in glucose transport and uptake (28, 29). It is generally believed that glucose transport is the rate-limiting step in glucose uptake and disposal. However, epinephrine acting via β-adrenergic receptors can shift the rate-limiting step from glucose transport to glucose uptake by inhibiting hexokinase and glucose phosphorylation (1, 4, 14, 18, 21). In a previous study from our laboratory (11), epinephrine was shown to attenuate the increase in muscle glucose uptake elicited by a moderate physiological insulin concentration. The results suggested that epinephrine attenuated glucose uptake by reducing the rate of glucose phosphorylation. However, epinephrine also blocked insulin-stimulated insulin receptor substrate-1 (IRS-1)-associated phosphatidylinositol 3-kinase (PI3-kinase) activity. Activation of PI3-kinase is an essential step for insulin-stimulated glucose transport (3, 5, 29). It is therefore possible that the reduction in insulin-stimulated glucose uptake by epinephrine we previously observed was due to an attenuation in glucose transport rather than or in addition to an inhibition of glucose phosphorylation. The purpose of the present study, therefore, was to investigate the effects of epinephrine on insulin-stimulated PI3-kinase activation and glucose transport in skeletal muscle. The findings of this study indicate that epinephrine can inhibit insulin-stimulated muscle glucose transport, but only when insulin is within a low to moderate physiological range. METHODSHarlan Sprague-Dawley rats (n = 40) between 110–120 g were randomly assigned to the following three groups: basal, insulin, insulin-epinephrine. All animals were obtained from and housed in the Animal Resource Center, University of Texas, Austin, TX. The temperature of the animal room was maintained at 21°C, and a 12:12-h light-dark cycle was set. All procedures were approved by the Animal Care and Use Committee of the University of Texas and conformed to the guidelines for the use of laboratory animals published by the US Department of Health and Human Services. Rats were anesthetized after an 8-h fast via an intraperitoneal injection of pentobarbital sodium (6.5 mg/100 g body wt), and the epitrochlearis (fast twitch) and soleus (slow twitch) muscles were then excised. The soleus was separated into strips weighing ∼15 mg and, with the epitrochlearis, was used to assess glucose transport and PI3-kinase activity after in vitro incubation. After isolation, epitrochlearis and soleus muscles were individually preincubated for 50 min in 1.5 ml of continuously gassed (95% O2-5% CO2) Krebs-Henseleit bicarbonate buffer containing 0.1% BSA, 32 mM mannitol, and 8 mM glucose. After preincubation, muscles were washed for 10 min in fresh buffer (1.5 ml) containing 0.1% BSA and 40 mM mannitol. Muscles were then incubated for 15 min in fresh buffer (1.5 ml) containing 0.1% BSA, 2 mM pyruvate, 38 mM mannitol, and 0.5 mg/ml ascorbic acid with either 0 insulin, 50 μU/ml insulin (Eli Lilly, Indianapolis, IN), or 50 μU/ml insulin plus 24 nM epinephrine (Sigma Chemical, St. Louis, MO). Next, the muscles were transferred to fresh buffer, and glucose transport was measured after a 15-min incubation in the presence of 0.1% BSA, 0.5 mg/ml ascorbic acid, 2 mM pyruvate, 6 mM 3-O-methyl-glucose (3-OMG), 280 μCi/mmol3H-labeled 3-OMG (Dupont NEN, Boston, MA), 32 mM mannitol, 10 μCi/mmol [14C]-mannitol (ICN Pharmaceuticals, Costa Mesa, CA), and the appropriate hormone concentration. For muscles in which IRS-1-associated PI3-kinase activity was to be measured, all radioactive isotopes were absent from the incubation medium. All incubations occurred at 29°C. After the last incubation period, muscles were blotted and freeze clamped with Wollenberg tongs cooled in liquid nitrogen. The ability of epinephrine (24 and 500 nM) to attenuate skeletal muscle glucose transport was also evaluated in the presence of 100 μU/ml insulin. Incubation procedures were identical to those described when 0 and 50 μU/ml insulin concentrations were evaluated with one exception. The initial 15-min incubation after the 10-min washout was eliminated. Thus glucose transport and PI3-kinase activity were assessed after 15 min of hormone exposure rather than after 30 min. The rationale for reducing the exposure time was based on recent findings by Song et al. (24). They demonstrated that PI3-kinase activity peaked at 6 min and then declined to basal values after exposure to a maximal insulin concentration in both the epitrochlearis and soleus muscles. In a previous study from our laboratory, we observed that PI3-kinase activity in the epitrochlearis and soleus muscles was maximal between 30 and 40 min of incubation with 50 μU/ml insulin (11). Thus, with an insulin concentration of 100 μU/ml, we reasoned that peak PI3-kinase activity would occur earlier than with insulin concentration of 50 μU/ml. Therefore, we reduced the exposure time to 15 min to ensure detection of an effect of epinephrine on insulin-stimulated PI3-kinase activity. Glucose transport was estimated by determining the incorporation rate of 3H-labeled 3-OMG into skeletal muscle. 3-OMG is a glucose analog that has transport rates similar to glucose but is not phosphorylated by hexokinase and provides a good estimate of the rate of glucose transport in skeletal muscle. Incubated muscles were weighed and dissolved in 1 M KOH for 15 min at 60°C. Dissolved samples were then neutralized with 1 M HCl, and 0.3 ml of the supernatant was added to 6 ml of Biosafe II scintillation fluid (Research Products International, Mount Prospect, IL). Samples were counted for3H and 14C in a LS-6000 liquid scintillation spectrophotometer (Beckman, Fullerton, CA). For PI3-kinase activity, muscle samples were homogenized, and aliquots of the homogenate were diluted to 2 μg/μl of protein in homogenization buffer. Samples were then immunoprecipitated by incubating for 2 h on ice with 2 μg of anti-IRS-1 antibody (Upstate Biotechnology, Lake Placid, NY). Protein A-Sepharose beads (Sigma Chemical) were added to the immunoprecipitation reaction, and incubation was continued for another 90 min at 4°C with rotation. The protein A- Sepharose beads-antibody complex was precipitated by centrifugation (5 min at 20,400 g). Immunoprecipitates were washed successively with 1) PBS containing 10% (octylphenoxy)polyethoxyethanol, 100 mM Na3VO4, and 1 M dithiothreitol; 2) 1 M Tris · HCl (pH 7.5), 2 M LiCl2, 100 mM Na3VO4, and 1 M dithiothreitol; and3) 1 M Tris · HCl (pH 7.5), 5 M NaCl, 10 mM EDTA, 100 mM Na3VO4, and 1 M dithiothreitol. Washing was done one time each with buffers 1 and 2 and twice with buffer 3. Packed beads were suspended, and 20 μl of phosphotidylinositol (Avanti Polar Lipids, Alabaster, AL) were dissolved in 4× HEPES and distilled H2O for a final concentration of 10 mg/ml. The kinase reaction was started by adding 10 μl of 10 mM ATP, 4× HEPES buffer, 0.4 M MgCl2, and 0.16 μCi/μl [γ-32P]ATP (Dupont). The reaction was incubated at room temperature with vigorous shaking and was terminated by adding 15 μl of 4 N HCl and 130 μl of MeOH-CHCl3 (1:1, vol/vol). After brief centrifugation, 20 μl of the organic solvent layer were spotted onto a thin-layer chromatography plate (Silica gel 60, Whatman, Hillsboro, OR) that had been activated with potassium oxalate. After phosphoinositides in running solvent (CHCl3-MeOH-H2O-NH4OH, 60:47:11:3.2, vol/vol/vol/vol) were separated, plates were dried and exposed. Spots were scraped from the plates and counted for32P in 6 ml of Biosafe II scintillation fluid in a scintillation counter. A one-way analysis of variance among experimental groups was performed on all variables. Fisher's protected least significance test was utilized to distinguish significant differences between groups. A level of P < 0.05 was set for significance for all test, and all values are expressed as means ± SE. RESULTSIn the isolated epitrochlearis and soleus muscles, an insulin concentration of 50 μU/ml increased glucose transport 53 and 138% above basal, respectively (Fig. 1, A andB). An insulin concentration of 100 μU/ml increased glucose transport 220% above basal in the epitrochlearis and 378% above basal in the soleus muscle (Fig. 2,A and B). Epinephrine (24 nM) inhibited glucose transport in the presence of 50 μU/ml insulin in both the epitrochlearis and soleus muscles. However, when muscles were incubated in 100 μU/ml insulin, neither 24 nor 500 nM epinephrine had a significant effect on glucose transport. Fig. 1.Insulin-stimulated (50 μU/ml) glucose transport in epitrochlearis (A) and soleus (B) muscles. Values are means ± SE. Ins, 50 μU/ml insulin; Ins/Epi, 50 μU/ml insulin and 24 nM epinephrine. * Significantly different from basal (P < 0.05). † Significantly different from insulin (P < 0.05). Fig. 2.Insulin-stimulated (100 μU/ml) glucose transport in epitrochlearis (A) and soleus (B) muscles. Values are means ± SE. Ins, 100 μU/ml insulin; Ins/Epi (24 nM), 100 μU/ml insulin and 24 nM epinephrine; Ins/Epi (500 nM), 100 μU/ml insulin and 500 nM epinephrine. * Significantly different from basal (P < 0.05). An insulin concentration of 50 μU/ml increased IRS-1-associated PI3-kinase activity in the epitrochlearis by 68% and in the soleus by 106% above basal (Fig. 3, A andB). An insulin concentration of 100 μU/ml significantly increased IRS-1-associated PI3-kinase activity in the epitrochlearis by 72% and in the soleus by 63% (Fig.4, A and B). Activation of IRS-1-associated PI3-kinase by 50 μU/ml insulin was inhibited by 24 nM epinephrine in both the epitrochlearis and soleus muscles. In the presence of 100 μU/ml insulin, neither 24 nor 500 nM epinephrine had an effect on insulin-stimulated IRS-1-associated PI3-kinase activity. Fig. 3.Insulin-stimulated (50 μU/ml) phosphatidylinositol 3-kinase (PI3-kinase) activity in epitrochlearis (A) and soleus (B) muscles. Values are means ± SE. Ins, 50 μU/ml insulin; Ins/Epi, 50 μU/ml insulin and 24 nM epinephrine. * Significantly different from basal (P < 0.05). † Significantly different from insulin (P < 0.05). Fig. 4.Insulin-stimulated (100 μU/ml) PI3-kinase activity in epitrochlearis (A) and soleus (B) muscles. Values are means ± SE. Ins, 100 μU/ml insulin; Ins/Epi (24nM), 100 μU/ml insulin and 24 nM epinephrine. * Significantly different from basal (P < 0.05). DISCUSSIONEpinephrine is effective in attenuating glucose uptake stimulated by a physiological concentration of insulin (11, 12, 14,21). It is generally accepted that epinephrine attenuates glucose uptake by inhibiting hexokinase and glucose phosphorylation (1, 4, 18, 21, 23). However, we recently found that, under certain in vitro conditions, this attenuation in insulin-stimulated glucose uptake can also coincide with a reduction in IRS-1-associated PI3-kinase activity (11, 12). Evidence suggests that the activation of PI3-kinase is necessary for the stimulation of glucose transport by insulin (3, 5, 20). Thus we hypothesized that inhibition of glucose uptake by epinephrine in the presence of a moderate physiological insulin concentration may be due to the inhibition of glucose transport rather than to a step distal to transport. In the present study, 50 and 100 μU/ml insulin increased IRS-1-associated PI3-kinase activity and glucose transport in the epitrochlearis and soleus muscles. Furthermore, 24 nM epinephrine was found to block activation of IRS-1-associated PI3-kinase by 50 μU/ml, and this coincided with a reduction in insulin-stimulated glucose transport. These results are in agreement with earlier studies in which insulin-stimulated glucose transport was found to require the activation of PI3-kinase (3, 5, 20). In studies by Cheatham et al. (3), Clarke et al. (5), and Lee et al. (20) in which the fungal metabolite wortmannin was used to block the activation of PI3-kinase, a corresponding inhibition in insulin-stimulated glucose transport was reported. The results of the present study, therefore, suggest that a physiological epinephrine concentration can reduce insulin-stimulated muscle glucose uptake by inhibiting glucose transport rather than by inhibiting hexokinase and glucose phosphorylation. However, this effect of epinephrine appears to be insulin-concentration specific because epinephrine had no attenuating effect on IRS-1-associated PI3-kinase activity or glucose transport activated by 100 μU/ml insulin. Even when epinephrine was increased from 24 to 500 nM, glucose transport was unabated in the presence of 100 μU/ml insulin. Therefore, the ability of epinephrine to attenuate insulin-stimulated glucose uptake by inhibiting transport appears to be limited to insulin concentrations in the low to moderate physiological range. Previous research strongly supports the concept that epinephrine has no effect on insulin-stimulated glucose transport. In the classic studies of Kipnis and colleagues (18, 19), it was demonstrated that epinephrine increased the intracellular free glucose concentration of skeletal muscle. Furthermore, epinephrine was found to inhibit maximal insulin-stimulated 2-deoxyglucose phosphorylation by 40%, whereas the intracellular concentration of 2-deoxyglucose increased proportionately. From these results, it was concluded that epinephrine inhibits insulin-stimulated glucose uptake in skeletal muscle by inhibiting hexokinase activity. The first study to directly assess the effect of epinephrine on insulin-stimulated glucose transport was conducted by Chiasson et al. (4). Using the rat hindlimb perfusion technique, they found that 100 nM epinephrine had no effect on 3-OMG transport activated by 1,000 μU/ml insulin but that it was effective in inhibiting the phosphorylation of 2-deoxyglucose. In studies that followed, Lee et al. (21) and Pittner et al. (23) found that even when epinephrine concentration was increased to 1,000 nM, it had no effect on glucose transport activated by ≥1,000 μU/ml insulin. In addition, Lee et al. (21) demonstrated that 1,000 nM epinephrine had no effect on glucose transport in the presence of 60 μU/ml insulin. The one commonality among these studies, aside from the study by Lee et al. (21), is that a pharmacological concentration of insulin was utilized. In the present study, 24 nM epinephrine had no effect on glucose transport activated by a high physiological concentration of insulin (100 μU/ml), which corroborates the findings of earlier studies. However, when a moderate physiological insulin concentration of 50 μU/ml was utilized, an inhibition of glucose transport was observed. We cannot at this time reconcile the difference in results between our findings and those of Lee et al. (21). Possibly, in the study by Lee et al., prior exposure of muscle to epinephrine caused downregulation of β-adrenergic receptors and reduced the efficacy of the epinephrine response. Regardless, our finding that epinephrine can inhibit insulin-stimulated glucose transport is not unique. Han and Bonen (9) reported that insulin-stimulated glucose transport in the three basic muscle fiber types of the rat could be inhibited by a physiological epinephrine concentration. Under certain circumstances, it is likely that a coordinated control of muscle glucose transport by insulin and epinephrine is required to achieve the most desirable physiological response. Recent studies have demonstrated that epinephrine or agents that increase cAMP can regulate the insulin signaling pathway by reducing tyrosine kinase activity, IRS-1 phosphorylation, and PI3-kinase activity (11, 25,26). Likewise, insulin-receptor activation can also lead to a reduction in β-adrenergic receptor function and adenylyl cyclase activity (8, 17). These results suggest that cross talk between these two signaling pathways may be a mode by which insulin and epinephrine regulate the opposing signaling pathway to control glucose transport. Therefore, it is possible that when the β-adrenergic signaling pathway has the greater level of receptor activation it will override the insulin signaling pathway, and when the insulin signaling pathway reaches a certain level of activation it will override the β-adrenergic signaling pathway. It is also possible that under elevated insulin conditions, glucose transport can be activated via an alternative pathway that is not influenced by β-adrenergic activation. There are several conditions when the modulating effects of epinephrine on insulin-stimulated glucose transport could be important, such as during exercise. Like insulin, contraction also increases skeletal muscle glucose transport (6, 12, 22). In combination, insulin and contraction produce a clear additive effect on glucose transport (6, 12, 22). During exercise when blood glucose starts to decline, increased sympathetic activity increases plasma epinephrine levels and lowers plasma insulin levels (27). Although plasma insulin levels decrease during exercise, there is, however, an increase in insulin exposure at the active tissue due to an increase in muscle blood flow. Thus a part of the increase in glucose transport during exercise can be attributed to the action of insulin. Blockage of the action of epinephrine with the use of β-adrenergic blockers can result in rapid, exercise-induced hypoglycemia (7,10). This hypoglycemic state induced by β-blockade is due to an increase in glucose utilization by skeletal muscle (2, 12, 15,16). In a previous study from our laboratory (12), 24 nM epinephrine reduced IRS-1-associated PI3-kinase activity and the additive effect of insulin and contraction on glucose uptake. The reduction in insulin/contraction-stimulated glucose uptake resulted in a glucose uptake rate similar to contraction alone. The present results suggest that epinephrine may play a significant role in regulating muscle glucose uptake during exercise by modulating insulin-stimulated glucose transport as well as by inhibition of hexokinase. Such a strategy would limit blood glucose utilization without excessive accumulation of intracellular free glucose; a response that could occur if inhibition of hexokinase were the only mechanism by which epinephrine limited glucose uptake. In summary, the present study demonstrates that epinephrine is capable of modulating glucose transport activated by insulin in the low to moderate physiological range. This could represent a significant means by which epinephrine regulates glucose utilization in the postabsorptive state and during exercise to prevent hypoglycemia, while also protecting the cell against excessive accumulation of intracellular free glucose. We are grateful for the excellent technical assistance provided by Zhenping Ding. FOOTNOTESREFERENCES
Page 3the ability to raise skin blood flow (SkBF) in response to environmental heat stress diminishes with advanced age (8, 9, 13, 17), particularly in those over the age of 65 yr. This attenuated SkBF response may contribute to the much higher incidence of heat-related illness and death in the elderly (5). Recent evidence has shown that the diminished ability to reflexively increase SkBF in the elderly is due to either an alteration in the active vasodilator mechanism or the skin vascular responsiveness to a given level of stimulation (9). A common tool used to evaluate changes in microvascular or endothelial function is the direct application of heat to an area of skin while evaluating the SkBF response. Locally heating an area of skin causes an increase in SkBF that is partially mediated by, or dependent on, nitric oxide (NO) (7, 12). Several investigators have shown that maximal SkBF to prolonged local heating is significantly reduced in older vs. young subjects (4,11, 15, 16, 22). It has been suggested that both functional and structural changes occur in the skin vasculature with aging that may account for the reduced SkBF to this stimulus (1, 11, 22). However, no studies have investigated the specific mechanisms that underlie changes in SkBF during local heating with aging, because until recently the mechanisms that mediate the SkBF response to local heating were unknown. Our laboratory recently reported that there are at least two independent mechanisms that contribute to the rise in SkBF during submaximal local heating: a fast-responding vasodilation primarily mediated by the axon reflexes and a more slowly responding vasodilator system that relies on local production of NO (12). There is evidence to suggest that NO-dependent mechanisms may be reduced with advanced age, and this may contribute to the attenuated SkBF responses. It has been reported that levels of the NO precursorl-arginine and the metabolites of NO, nitrite and nitrate, are reduced with aging (14). Furthermore, the SkBF response to iontophoresis of sodium nitroprusside, a NO donor, is reduced in older subjects (1). These data suggest that the ability to produce NO or the vascular responsiveness to NO may be attenuated with advanced age. However, depending on the local heating protocol, a number of different vasodilator mechanisms may be activated. For example, a very rapid rise in skin temperature, a prolonged period of heating, or a slight sensation of thermal pain during heating can cause the cutaneous vasodilation to be insensitive to NO synthase (NOS) inhibition (7, 10, 12). The specific mechanisms activated under these conditions are not known. Because our goal in the present study was to examine the role of NO in the attenuated SkBF response to local heating with age, we employed a local heating protocol that our laboratory previously determined would allow us to examine the contribution of NO and the axon reflexes to the SkBF response to local heating (12). We hypothesized that the sustained NO-dependent vasodilation, but not the initial rise in SkBF mediated by the axon reflex, would be diminished with advanced age during local heating. METHODSWe recruited 10 young (18–24 yr old; 5 men and 5 women) and 10 older subjects (69–84 yr old; 5 men and 5 women) to participate in this study. All young women were studied during menstruation (early follicular phase of the menstrual cycle) because the menstrual cycle is known to alter the SkBF response to local heating (2). All older women were postmenopausal and were not on hormone replacement therapy. All subjects were healthy, normotensive, nonsmokers, and not taking any medications. Institutional Review Board approval was obtained, and each subject gave informed consent before participation. Subjects wore a water-perfused suit to clamp whole body skin temperature between 32 and 33°C. The water-perfused suit did not cover the face or the forearm being studied, and it was used to minimize the influence of whole body skin temperature on reflex changes in SkBF at the local skin sites. Respiration and electrocardiogram were continuously measured throughout the study. Two microdialysis fibers (MD 2000, Bioanalytical Systems) with a membrane length of 10 mm and a 20-kDa cutoff were placed in the skin at least 5 cm apart in the ventral aspect of the nondominant forearm of each subject. The ventral aspect of the forearm was used to minimize the effects of chronic sun exposure on the results of the study and thereby isolate the effects of aging on the cutaneous vascular responses to local heating. Insertion of the microdialysis fibers was performed by first placing a 25-gauge needle just under the surface of the skin with entry and exit points ∼3 cm apart. The microdialysis fiber was then threaded through the needle so that the microdialysis membrane was 1 cm from the lumen of the needle. The needle was then partially withdrawn as the microdialysis membrane was pulled into place in the skin. Once the membrane was in place between the insertion and exit points, the needle was completely withdrawn. The microdialysis fibers were taped in place, and Ringer solution was perfused through the fibers at a rate of 2 μl/min. To obtain an index of SkBF, cutaneous red blood cell (RBC) flux (in mV) was measured over the two microdialysis sites by using a Moor laser-Doppler flowmetry (LDF) system (DRT-4). Skin temperature was controlled at the two microdialysis sites with Moor local heating units (SH02) each covering ∼100 mm2 of tissue. RBC flux was measured directly over the microdialysis membrane. To ensure that blood pressure was stable throughout the experimental protocols, blood pressure was measured at 10-min intervals by brachial auscultation in the dominant arm. After placement of the microdialysis fibers, RBC flux over the microdialysis sites was monitored to ensure that insertion trauma had resolved before the studies were started (between 60 and 130 min). The temperature of the local heating units at the microdialysis sites was kept constant at 33°C during the baseline period. One of the microdialysis sites was randomly assigned to receive 10 mMNG-nitro-l-arginine methyl ester (l-NAME; Calbiochem) dissolved in Ringer solution. We performed pilot studies in young and older subjects to determine that doses of l-NAME above 10 mM did not result in a greater inhibition of NO production by NOS. The microdialysis fibers were perfused at a rate of 2 μl/min for at least 30 min before local heating to inhibit NOS. Infusion of l-NAME in this site was continued throughout local heating. After l-NAME infusion and baseline measurements, the local heating protocols were performed as follows. Temperature of the local heating units was increased at a rate of 0.5°C every 5 s to a temperature of 42°C. This results in an increase in skin temperature to ∼40°C at the heating probe-skin surface interface (12). Subjects did not feel any sensations of pain during the rise in temperature at this rate or with prolonged heating at this temperature. The local heating units were held constant at 42°C throughout the entire protocol. After RBC flux in both sites had reached a stable plateau (∼30–40 min), 10 mM l-NAME was infused through the second microdialysis site (previously infused with Ringer solution only) at a rate of 2 μl/min. Local heating was continued in both sites until RBC flux at this site decreased to a new stable plateau for at least 10 min. The local heaters were then returned to 33°C, and 28 mM sodium nitroprusside (Nitropres, Ciba Phamaceuticals) was infused through the microdialysis fibers for 20–30 min to maximally vasodilate the skin at both sites. We performed pilot studies in young and older subjects to determine that 28 mM sodium nitroprusside was a sufficient dose to maximally vasodilate the skin at the microdialysis sites. Cutaneous vascular conductance (CVC) was calculated as LDF (mV)/mean arterial pressure (MAP; mmHg) to account for any differences in blood pressure between the groups. Data are expressed as percentage of maximal CVC obtained during nitroprusside infusion (%CVCmax). Data were digitized and stored on a computer at 100 Hz. Data were analyzed off-line with signal-processing software (Windaq, Dataq Instruments, Akron, OH). Baseline, plateau, and the reduction in CVC with NOS inhibition (post-NOS inhibition drop) were calculated by averaging values over a stable 10-min period. Initial peak and nadir CVC values were calculated by averaging the highest and lowest values, respectively, over a stable 30-s period. The phases of the SkBF response to local heating are presented in Fig.1. Fig. 1.A: representative tracing of cutaneous vascular conductance during the local heating protocol when nitric oxide synthase (NOS) was inhibited after 40 min of heating in a young and older subject. B: group data (means ± SE) during baseline, initial peak, nadir, plateau, and the reduction in cutaneous vascular conductance with NOS inhibition (post-NOS inhibition). *Significant difference from the young group, P < 0.05. All data are presented as means ± SE. Group characteristics, the contribution of NO to the initial peak and plateau phase of the SkBF response to local heating, and maximal CVC values to 28 mM sodium nitroprusside were compared by using t-tests. CVC during baseline, initial peak, nadir, plateau, and the drop in SkBF with NOS inhibition were analyzed by two-way ANOVA (age group × local heating phase) with repeated measures. When a significant interaction effect was observed, Tukey's post hoc analysis was used to identify significant differences in the pairwise comparisons. The level of significance was set at P < 0.05. RESULTSSubject characteristics are presented in Table1. No gender differences in CVC were observed in either age group. Thus the data from men and women were combined for each group. Baseline systolic blood pressure and MAP were significantly higher in the older subjects (both P < 0.05).
At the initiation of heating, skin temperature at the heating probe-skin surface interface rapidly increased to ∼40°C and remained stable throughout the heating protocol. No subject reported feeling any sensation of pain during local heating of either site. Figure 1A is a representative tracing of the CVC responses to the local heating protocol when NOS was inhibited after 40 min of heating in a young and older subject. Figure 1B displays the group data (means ± SE) during baseline, initial peak, nadir, plateau, and the decline in CVC with NOS inhibition (post-NOS inhibition drop). There was no significant main effect of age on the CVC response to local heating (P = 0.155). This was not unexpected because the general pattern of the SkBF response to local heating was similar in the two groups. There was a significant main effect for phase of local heating (P < 0.001). There was a significant interaction (age × local heating phase;P = 0.002) with significant differences by pairwise comparisons observed during the initial peak (young: 61 ± 2%CVCmax vs. older: 46 ± 4%CVCmax;P < 0.05) and plateau (young: 93 ± 2%CVCmax vs. older: 82 ± 5%CVCmax;P < 0.05) phases, displaying diminished responses in the older group. No significant differences in CVC were observed between the groups during baseline, nadir, or after NOS inhibition. Figure 2 displays the contribution of NO to the plateau phase of the SkBF response to local heating. Values were calculated by subtracting CVC values after NOS inhibition from CVC during the plateau phase in the site in which NOS was inhibited during the stable plateau phase. The contribution of NO to the plateau phase during local heating was significantly less in the older subjects (young: 75.0 ± 2.3%CVCmax vs. older: 61.1 ± 4.5%CVCmax; P < 0.05). Fig. 2.Calculated contribution of nitric oxide to the plateau phase of the cutaneous vascular conductance response to local heating when NOS was inhibited during heating. Values were calculated by subtracting cutaneous vascular conductance values after NOS inhibition from cutaneous vascular conductance during the plateau phase when NOS was inhibited during the stable plateau phase. The contribution of NO to the plateau phase during local heating was significantly less in the older compared with the young subjects (61.1 ± 4.5 vs. 75.0 ± 2.3% maximum cutaneous vascular conductance, respectively). Values are means ± SE. Figure 3A is a representative tracing of the CVC responses to the local heating protocol when NOS was inhibited before and throughout the local heating protocol in a young and older subject. Figure 3B displays the group data (means ± SE) during baseline, initial peak, nadir, and plateau phase when NOS was inhibited throughout the protocol. There was no significant main effect for age on the CVC response to local heating (P = 0.160). There was a significant main effect for local heating phase (P < 0.001), and for the interaction (age group × local heating phase; P = 0.04), with a significant difference by a pairwise comparison only observed during the initial peak (young: 52 ± 4%CVCmax vs. older: 38 ± 5%CVCmax;P = 0.004). No significant differences were observed between the groups during baseline, nadir, or between 30 and 40 min of local heating in this site. Fig. 3.A: representative tracing of the changes in cutaneous vascular conductance during the local heating protocol when NOS was inhibited before and throughout the local heating protocol in a young and older subject. B: group data (means ± SE) during baseline, initial peak, nadir, and the values at 40 min of the local heating protocol. *Significant difference from the young group,P < 0.05. To examine the contribution of NO to the initial peak in both groups of subjects, the difference in peak CVC values from the two sites (initial peak before NOS inhibition − initial peak with NOS inhibition; expressed as Δ%CVCmax) were compared. No difference between the groups was observed (young: 8.6 ± 3.6 Δ%CVCmax vs. older: 8.0 ± 3.5 Δ%CVCmax; P = 0.90). This demonstrates that the attenuated initial peak in the older subjects was not due to diminished NO-dependent vasodilation. Maximal CVC obtained by infusion of 28 mM sodium nitroprusside at the end of local heating (expressed as mV/100 mmHg) was significantly less in the older subjects (young: 192 ± 12 mV/100 mmHg vs. older: 156 ± 15 mV/100 mmHg;P = 0.03). DISCUSSIONThe goal of this study was to investigate the mechanisms that underlie the observed age-related attenuation of cutaneous vasodilator responses in humans. We performed two local heating trials in young and older subjects to examine the role of NO and the axon reflex in the SkBF response to a local heating protocol. We found that the initial rise in SkBF and the sustained vasodilation to local heating were significantly diminished in the older subjects. This finding suggests that healthy aging impacts the nerves that mediate the axon reflex or vascular responsiveness to the neurotransmitters released from these nerves. The smaller sustained rise in SkBF during prolonged local heating in the older subjects further suggests the ability to either produce or respond to NO is diminished with advanced age. Importantly, our findings cannot simply be explained by an age-related reduction in maximal SkBF, because the responses were evaluated as a percentage of each individual subject's maximal vasodilation obtained at each experimental site. The initial rise in SkBF during local heating appears to be predominantly mediated by an axon-reflex mechanism that remains robust when NOS is inhibited (12). In older subjects, we observed an attenuated initial peak response that was not due to diminished NO-dependent vasodilation. NO only contributes modestly to the initial peak response, and the magnitude of the reduction in the initial peak was similar in the two groups with NOS inhibition. Presently, we can only speculate on the neurotransmitters involved in the initial response to local heating and how it is reduced with advanced age. Recent evidence suggests that calcitonin gene-related peptide is the most likely neurotransmitter involved in the axon reflex (18). However, substance P cannot be ruled out because of the difficulty in measuring this peptide. Calcitonin gene-related peptide is one of the most abundant neuropeptides in the skin, and it is found alone or colocalized with substance P (21). Thus it appears likely that aging results in either a diminished release of these neurotransmitters or a diminished responsiveness of the vasculature to these neurotransmitters. This finding was in contrast to our hypothesis that only the NO-dependent portion of the local heating response would be diminished with advanced age. All of our older subjects were healthy, and none of them reported having decreased sensations to thermal or painful stimuli. Importantly, the rapid rise in SkBF in response to local heating is a vital step in protecting the skin from acute trauma. For example, the increase in blood flow will minimize the heat transferred to the tissues to protect the skin from damage. A diminished ability to rapidly increase SkBF in response to directly applied heat may make the elderly more susceptible to local tissue damage. Thus studies are needed to further explore the exact mechanisms involved in the initial rise of SkBF during local heating and to determine how aging impacts these mechanisms. The plateau phase of the SkBF response to local heating is primarily, but not entirely, mediated by NO (12). When NOS is inhibited during the sustained plateau in SkBF, SkBF decreases to a value that remains significantly elevated above baseline levels even when the cutaneous vasoconstrictor nerves are blocked (12). This suggests that an unknown vasodilator is present and mediates a portion of the plateau-phase dilation. Our finding of a diminished plateau phase (Fig. 1) and a significantly smaller contribution of NO to the plateau phase (Fig. 2) in the older subjects suggest that NO-dependent vasodilator mechanisms in the skin are diminished with advanced age. If plateau CVC values were similar in the two groups of subjects, then the decline in CVC with NOS inhibition would be similar and we would not be able to conclude that NO-mediated dilation was diminished with age. Alternatively, if post-NOS inhibition values were significantly less in the older subjects, then the lower plateau CVC values would be due to diminished vasodilation to the unknown vasodilator and not due to attenuated NO-mediated vasodilation. Taken together, we interpret our findings to suggest that the cutaneous vasculature either produces less NO or is less responsive to NO with age. The exact mechanism by which NO causes vasodilation in the skin during local heating is not known. It seems likely that the locally applied heat directly causes increased NO production by endothelial cells. However, there are other potential mechanisms by which NO may mediate cutaneous vasodilation. For example, NO is known to contribute ∼30% to the rise in SkBF during whole body heating, even though skin temperature at the measurement site does not increase (6, 19,20). Crandall and MacLean (3) were unable to measure a rise in NO concentration in the skin during whole body heating, suggesting that NO may play a “permissive role” in active vasodilation. That is, during whole body heating, NO may only need to be present to allow the activity of another vasodilator substance to achieve full expression. It is possible that NO is playing a similar role in mediating vasodilation to local heating, but this remains to be tested. Attenuated NO-dependent vasodilation in the elderly could be due to diminished levels of the NO precursor l-arginine. In support of this concept, Reckelhoff et al. (14) reported that levels of l-arginine and the metabolites of NO, nitrite and nitrate, are reduced with age. Alternatively, another possibility could be that the transduction of the NO signal in the smooth muscle of the skin could be diminished with age, causing reduced vascular responsiveness. In agreement with this concept, others have reported that the rise in SkBF to sodium nitroprusside iontophoresis is diminished with advanced age (1). Sodium nitroprusside is an endothelium-independent vasodilator acting through NO, so a reduced SkBF response to sodium nitroprusside iontophoresis in older subjects suggests the smooth muscle does not dilate to the same extent as in younger subjects for a given level of NO. We found that the rise in SkBF to the infusion of 28 mM sodium nitroprusside (sufficient to cause maximal vasodilation in both young and older subjects) was significantly lower in the older subjects. Although this finding is consistent with our concept of reduced vascular responsiveness to NO with age, we believe that this finding needs to be interpreted with caution. Site-to-site variability of the laser-Doppler technique makes a comparison of data expressed in arbitrary units of RBC flux between subjects, or even from one area of skin to another within a subject, tenuous at best. We used the infusion of 28 mM sodium nitroprusside as a tool to obtain the maximal vasodilator capacity of each site to allow comparisons between the groups of subjects. However, our finding of a reduced maximal SkBF with age by using these techniques is consistent with studies that maximally vasodilated the skin by prolonged local heating (4, 11, 15, 16,22). These studies found that the increase in SkBF during local heating sufficient to maximally vasodilate the skin is diminished in the older subjects (4, 11, 15, 16, 22). It was suggested that this represents a structural limitation in the skin of older individuals (1, 11, 22). However, maximally vasodilating the skin by local heating may activate a number of mechanisms that could mediate the vasodilation (7, 10, 12). Kellogg et al. (7) first reported that even a very brief sensation of mild pain (<5–10 s) during local heating rendered the vasodilation insensitive to NOS inhibition. We confirmed these findings and further found that a very rapid rise in local heating temperature altered the mechanisms of vasodilation and eliminated the bimodal increase in SkBF (12). Magerl and Treede (10) reported that a sustained skin temperature of >41°C may activate specific nociceptive neurons with or without the sensation of pain. Thus we chose to use a submaximal local heating protocol that we knew would allow us to specifically examine the role of NO and the axon reflex in our young and older subjects. Although we cannot rule out the possibility that differences in maximal cutaneous vasodilation may be due to structural limitations of the skin with age, our data may also suggest that there are functional changes in the nerves, neurotransmitters, or vascular responsiveness to locally applied heat with advanced age. In summary, the initial rise in SkBF and the maintained high SkBF during local heating of the skin are diminished with advanced age. These results suggest that axon reflex- and NO-dependent vasodilation are reduced in the skin of healthy older individuals. Future studies are needed to further investigate the exact mechanisms that underlie the increase in SkBF with local heating and to develop possible strategies to minimize the impact of age on these mechanisms. The considerable time and effort of the subjects are greatly appreciated. FOOTNOTESREFERENCES
Page 4recent studies in isolated muscle preparations have demonstrated an important role for Na+-K+-ATPase in muscular fatigue, via prevention of a rundown in transmembrane Na+ and K+ gradients and thus preservation of membrane excitability (for review, see Ref. 35). Attenuation of Na+-K+-ATPase activity via inhibition with ouabain accelerates muscle fatigability and retards subsequent recovery (6), whereas, conversely, stimulation of Na+-K+-ATPase activity delays muscle fatigability and accelerates subsequent recovery in muscles paralyzed in high-K+ solution (6, 7, 34). Reduced Na+-K+ gradients decrease rat soleus muscle M wave area and tetanic force, whereas subsequent muscle electrical stimulation of Na+-K+-ATPase or salbutamol-induced stimulation of Na+-K+-ATPase elicited a marked recovery (39, 40). These studies highlight the importance of Na+-K+-ATPase activity in skeletal muscle function in animal models. Muscle excitation elicits a dramatic and rapid increase above rest levels in Na+-K+-ATPase activity, measured as net Na+ extrusion, in isolated rat muscles (8,33). Na+-K+-ATPase activity in isolated rat soleus muscle may increase up to 22-fold above rest after only 10 s of 120-Hz stimulation, thus approaching the maximal theoretical Na+-K+-ATPase activity (8). There are no direct measures of Na+-K+-ATPase activation in contracting human skeletal muscle. However, this is likely to also be dramatic, as shown by a rapid decline in femoral venous plasma K+concentration ([K+]) after knee extensor exercise and by the rapid K+ clearance from blood after exercise (for review, see Ref. 42). Despite this increased Na+-K+-ATPase activation during muscle contractions, a direct, depressive effect of fatiguing exercise on the maximal Na+-K+-ATPase activity can be hypothesized. There is considerable structural homology of the catalytic subunits of the Ca2+-ATPase and Na+-K+-ATPase enzymes (20). It is now well known that fatiguing muscle contractions in humans induce an acute depression in the sarcoplasmic reticulum maximal Ca2+-ATPase activity or Ca2+-ATPase-mediated Ca2+ uptake rate in skeletal muscle (4, 16, 26). Hence it is conceivable that factors that adversely affect maximal Ca2+-ATPase activity may also impair Na+-K+-ATPase activity, and investigation into the possible effects of fatigue on Na+-K+-ATPase activity is of great interest. In human skeletal muscle, it is possible to measure the maximal in vitro Na+-K+-ATPase activity (12). However, this assay does not measure the increase in Na+-K+-ATPase activation but rather reflects the theoretical maximal Na+-K+-ATPase activity (12, 31). A recent study has indeed demonstrated depressed maximal in vitro Na+-K+-ATPase activity after repeated isometric contractions (11). However, repeated isometric contractions induce marked muscle ischemia, which might be causally linked with the impaired maximal in vitro Na+-K+-ATPase activity and the observed postcontractile depression (11). No studies thus far have investigated whether fatiguing, dynamic contractions depress maximal Na+-K+-ATPase activity in human skeletal muscle, and therefore this was the first aim of the present study. One characteristic of training is an enhanced resistance to fatigue during activity specific to the training regimen. Factors linked with Ca2+-ATPase inactivation with exercise are affected by training, with modified muscle sarcoplasmic reticulum Ca2+regulation during exercise by resistance and endurance training (16, 26) and increased endogenous antioxidant enzymes also evident with training (46). If the maximal Na+-K+-ATPase activity is depressed with fatigue, then chronic training could conceivably confer a protective effect on Na+-K+-ATPase activity, thereby attenuating this decline and contributing to enhanced muscular performance. Any possible protective effect of training on Na+-K+-ATPase activity may also be influenced by the total muscle Na+-K+-ATPase content, which is increased with training (9, 15, 29, 32). Thus an increased Na+-K+-ATPase content with training would offset any depressive effect of fatiguing exercise on maximal Na+-K+-ATPase activity. There have been no studies reporting both Na+-K+-ATPase content and maximal activity in trained muscle. Thus the second aim of this paper was to investigate whether chronic training protects against any possible decline in the maximal Na+-K+-ATPase activity in human muscle. No studies have examined the effects of training on maximal Na+-K+-ATPase activity per se in resting muscle in humans. The 165% increase in maximal Na+-K+-ATPase activity reported with endurance training in canine muscle (24) far exceeds the typical 14–29% increase in [3H]ouabain binding with training (31), probably because of poor and variable enzyme recovery in the isolated membrane fraction used for measurement (17). Thus the effects of chronic training on maximal Na+-K+-ATPase activity in skeletal muscle remain unknown and were also examined. Finally, the increased muscle Na+-K+-ATPase content with training has been suggested to be important in the improved plasma K+ regulation during exercise, but no correlative evidence was found (15, 29). Here we explore whether muscle Na+-K+-ATPase content and maximal Na+-K+-ATPase activity are correlated to an index of plasma K+ regulation during exercise, the rise in plasma K+ concentration per unit work output (18, 29). Three hypotheses were tested in this study: 1) that an acute bout of fatiguing dynamic exercise would depress maximal in vitro Na+-K+-ATPase activity in skeletal muscle in humans; 2) that this depression will be attenuated in chronically resistance-trained and endurance-trained athletes, and3) that muscle Na+-K+-ATPase content and maximal Na+-K+-ATPase activity will be correlated to the rise in plasma K+concentration per unit work output. METHODSEight healthy untrained male controls (UT), eight resistance-trained subjects (RT, 7 male, 1 female), and eight endurance-trained male subjects (ET) participated after being informed of risks associated with the study and giving written, informed consent. Our laboratory has recently reported the physical characteristics, training history, anthropometry, and muscle function test results in these subjects in a paper investigating muscle Ca2+ regulation at rest and fatigue (26). The UT subjects were recreationally active but were not well trained and did not participate in regular sporting activities. The ET and RT athletes had been training continuously for at least 2 yr. During this period the ET athletes had performed running and/or cycling endurance training for at least 5–6 h/wk and had a peak oxygen consumption (V˙o2 peak) exceeding 60 ml · min−1 · kg−1. The RT subjects trained with heavy weights, typically performing three sets, six to eight repetitions, for at least 1 h and at least three sessions/wk. All were able to perform a power-lifting-style squat exercise with free weights at least 1½ times their body mass. No significant differences existed between the three groups (means ± SD) for age (UT 26.4 ± 3.9, RT 26.8 ± 7.9, ET 26.4 ± 3.1 yr), body mass (UT 80.4 ± 6.8, RT 81.6 ± 3.3, ET 70.6 ± 9.9 kg), or height (UT 183.3 ± 5.7, RT 176.1 ± 4.7, ET 177.2 ± 7.1 cm). All protocols and procedures were approved by the Human Research Ethics Committee at Victoria University of Technology. Each subject completed two tests involving invasive procedures. The first comprised an incremental cycle ergometer exercise test with arterialized-venous blood sampling conducted before, during, and after exercise. The second was a muscle fatigue test, comprising repeated maximal quadriceps contractions on an isokinetic dynamometer. A vastus lateralis muscle biopsy was taken at rest and immediately after fatiguing contractions, with arterialized venous blood sampling before, during, and after exercise. Each subject refrained from vigorous exercise, alcohol, and caffeine consumption for 24 h before each test. An incremental exercise test (25 W/min, 60 rpm except ET where 80 rpm) was performed on an electrically braked cycle ergometer (Lode N.V. Groningen, Netherlands) to determineV˙o2 peak, as detailed elsewhere (26). A catheter (20-gauge, Jelco) was inserted into a superficial dorsal hand vein of the subject before the test, and all blood samples were arterialized by heating the hand in a hot (45°C) water bath for 10 min before samples were taken (29). The catheter was kept patent by periodic infusions of heparinized isotonic saline. Arterialized venous blood was sampled at rest, in the final 10 s of each minute during graded exercise, and at 1, 2, 5, 10, 20, and 30 min in recovery, and the samples were analyzed for Hb concentration, hematocrit, as well as [K+] and plasma hydrogen ([H+]) and lactate ([Lac−]) concentrations. The muscle fatigue test was performed on an isokinetic dynamometer (Cybex II Lumex, Ronkoukowany), as previously detailed and justified (26). A muscle biopsy was taken at rest and immediately postexercise, and blood samples were taken at rest, mid- and immediately postexercise, and at 1, 2, 5, 10, 20, and 30 min in recovery and were analyzed as described for the incremental test. Subjects were strapped to the Cybex dynamometer chair by belts across the hips, chest, and legs to stabilize the upper body and thigh. Subjects performed 50 maximal knee extensions at a velocity of 180°/s and at a rate of 0.5 Hz (duration 100 s). Peak torque was measured and fatigue index (percentage decline in peak torque) calculated as described previously (26). After injection of a local anesthetic into the skin and fascia (2% Xylocaine), two small incisions were made in the midportion of the vastus lateralis muscle of the right leg. The rest and fatigue biopsies were taken from separate incisions. Resting samples were analyzed for maximal in vitro Na+-K+-ATPase activity and Na+-K+-ATPase content, whereas fatigue samples were analyzed for maximal in vitro Na+-K+-ATPase activity. Muscle fiber-type composition, sarcoplasmic reticulum Ca2+ regulation, and metabolite contents are reported elsewhere (26). Approximately 20 mg of the frozen resting muscle was used to quantify the Na+-K+-ATPase content by using the [3H]ouabain binding method as previously described (21, 29, 37). Samples were cut into small pieces of 2–4 mg wet wt. In all experiments, freshly made vanadate solution was used. Samples were washed at 0°C for 20 min, with a change of medium after 10 min (2 × 10 min) in a buffer containing 10 mM Tris, 250 mM sucrose, 3 mM MgSO4, and 1 mM vanadate, pH 7.2–7.4. This procedure was used to thaw the samples and preincubate them with vanadate and to maintain low Na+ and K+ concentrations so as not to interfere with vanadate-facilitated [3H]ouabain binding. Incubations took place in a buffer containing 2 μCi/ml [3H]ouabain and ouabain added to a final concentration of 1 μM at 37°C for 2 h, with a change of medium after 1 h. After incubations, a washout at 0°C in unlabeled buffer for 2 h with a change of medium every 30 min (4 × 30 min) was performed to reduce the [3H]ouabain in the extracellular space and enhance the precision of the method. After washout, samples were blotted on dry filter paper, weighed, and soaked overnight in minivials containing 0.5 ml of 5% trichloroacetic acid. The next day, 2.5 ml of scintillator (Opti-fluor) was added before liquid scintillation counting of the [3H]ouabain activity was performed. The amount of [3H]ouabain taken up and retained by the samples was calculated on the basis of the sample wet weight and the specific activity of the incubation medium and samples. Skeletal muscle maximal in vitro Na+-K+-ATPase activity was determined by using the K+-stimulated 3-O-methylfluoroscein phosphatase (3-O-MFPase) assay in human muscle homogenates, as previously described in detail (12). The 3-O-MFPase assay was chosen because it is highly sensitive, capable of determining extremely low levels of Na+-K+-ATPase activity as found in human skeletal muscle (12). This assay has two to three times higher sensitivity, therefore requiring 50–100 times less tissue, than the K+-stimulatedp-nitro-phenyl-phosphatase (1, 36). Measurement of activity in whole muscle homogenates avoids the criticisms of very poor recovery inherent in techniques involving extensive enzymatic purification procedures (17). The assay was optimized for human skeletal muscle homogenates and specifically measures Na+-K+-ATPase activity, as evidenced by complete ouabain inhibition and K+stimulation of activity (12). The interassay (5.3%) and intra-assay (8.1%) variation were acceptably low (12). Briefly, muscle samples (30–40 mg) were immediately blotted on filter paper, weighed, then homogenized (5% wt/vol) at 0°C for 2 × 20 s, 20,000 rpm (Omni 1000, Omni International) in an homogenate buffer containing 250 mM sucrose, 2 mM EDTA, and 10 mM Tris (pH 7.40). Muscle homogenates were rapidly frozen and stored in liquid nitrogen for later determination of maximal in vitro K+-stimulated 3-O-MFPase activity. Before analysis, homogenates were freeze-thawed four times and then diluted one-fifth in cold homogenate buffer. The assay medium in which 3-O-MFPase activity was measured contained 5 mM MgCl2, 1.25 mM EDTA, 100 mM Tris, and an 80 nM 3-O-methyl fluorescein standard (pH 7.40). The freeze-thawed, diluted homogenate (30 μl) was incubated in 2.5 ml of assay medium at 37°C for 5 min before addition of 40 μl of 10 mM 3-O-MFP to initiate the reaction. After 60 s, 10 μl of 2.58 M KCl (final concentration 10 mM) was added to stimulate K+-dependent phosphatase activity, and the reaction was measured for a further 60 s. All assays were performed at 37°C, with continuous stirring, on a spectrofluorometer (Aminco Bowman AB2 SLM, Urbana, IL). Excitation wavelength was 475 nm, and emission wavelength was 515 nm, with 4-nm slit widths. The K+-stimulated 3-O-MFPase activity was calculated by subtracting the initial activity (comprising unspecific-ATPase activity and any spontaneous hydrolysis of 3-O-MFP) from the activity obtained after 10 mM KCl addition (12). Maximal in vitro 3-O-MFPase activity was expressed relative to muscle wet weight (nmol · min−1 · g−1wet wt) and to identify possible effects due to fluid shifts, also relative to muscle protein content (pmol · min−1 · mg−1protein). Protein content of the homogenate was determined spectrophotometrically by using bovine serum albumin as a standard. The relationship between [3H]ouabain binding site content and maximal in vitro 3-O-MFPase activity was examined in 22 of the 24 subjects, plus an additional six healthy untrained controls (age 40.7 ± 8.7 yr, body mass 60.8 ± 6.2 kg, height 163.9 ± 5.8 cm, means ± SD). The blood was mixed well, and air bubbles were removed from the syringe, which was capped tightly and placed on ice for subsequent duplicate analyses of plasma acid-base status and gas tensions (H+, Pco2, Po2, and [K+]) by use of an automated analyzer (865 Ciba Corning, Bayer). Hb concentration was determined in duplicate spectrophotometrically (Radiometer OSM2, Copenhagen, Denmark), whereas hematocrit was analyzed in triplicate after centrifugation (Hettich Zentrifugen D-7200, Tuttlingen, Germany). All analytical instruments were calibrated before and during the analyses with precision standards. An aliquot of whole blood was centrifuged at 4,000 rpm for 4 min, plasma was separated, a 200-μl aliquot of plasma was deproteinized in 600 μl of cold 3M perchloric acid, and the supernatant was later analyzed for plasma [Lac−] in triplicate by use of an enzymatic spectrophotometric technique. The percentage decline in plasma volume from rest (ΔPV), rise in plasma [K+] during exercise above rest (Δ[K+]) and the ratio of Δ[K+] per work done (Δ[K+]/work) were calculated as previously described (29, 30), with an example of the latter as follows. If, during incremental (25 W/min) exercise, Δ[K+] was 2.2 mM and the subject completed 1 min at 300 W, total work equals 117 kJ and the Δ[K+]/work is 18.8 nmol · l−1 · J−1. The decline in Na+-K+-ATPase activity with fatigue is calculated as rest minus fatigue in vitro 3-O-MFPase activity. Correlations involve pooled data from all subjects in the three groups (n = 22–24, as stated). Data are presented as means ± SE, except population data, for which means ± SD are shown. A two-way ANOVA (sample time, group) with repeated measures (time) was used to analyze most variables. A one-way ANOVA was used when only a single variable was compared between groups (e.g., [3H]ouabain binding). Post hoc analyses used the Newman-Keuls test. Correlations between variables were determined by least-square linear regression. Significance was accepted at P < 0.05. RESULTSAs a confirmation of training status, ET had a higher (P < 0.05) incremental exerciseV˙o2 peak than the other groups (UT 44.4 ± 1.8, RT 43.8 ± 3.6, ET 67.6 ± 1.5 ml · kg−1 · min−1, means ± SD), whereas RT had a higher (P < 0.05) quadriceps maximal peak torque during isokinetic contractions from 60–300°/s (26). Peak quadriceps muscle torque declined in all groups during the 50 contractions (P < 0.05), and the fatigue index was less (P < 0.05) in ET than in UT and RT (UT 47.4 ± 14.0, RT 43.4 ± 9.4, ET 29.9 ± 12.0%, means ± SD, Ref. 26). The resting muscle [3H]ouabain binding site content differed between groups, being 16.6 and 18.3% higher for ET than in UT and RT, respectively (P < 0.05, Fig.1). Fig. 1.Vastus lateralis muscle [3H]ouabain binding site content (Na+-K+-ATPase content) for untrained (UT), resistance-trained (RT), and endurance-trained (ET) subjects; n = 8 subjects/group (means ± SE). †ET > UT; ‡ET > RT (P < 0.05). A significant sample time main effect was shown for maximal in vitro 3-O-MFPase activity expressed per gram wet weight, with a decline of 13.8 ± 4.1% at fatigue (Fig.2A, P < 0.05). No significant group main effects or time-by-group interactions were seen, although resting 3-O-MFPase activity in ET tended to be higher (20.3%) than in UT. The decline in maximal in vitro 3-O-MFPase activity at fatigue did not differ significantly between the groups (absolute: UT −27 ± 8, RT −24 ± 13, ET −60 ± 17 nmol · min−1 · g wet wt−1; relative: UT −13 ± 4, RT −9 ± 6, ET −22 ± 6%). Fig. 2.Maximal in vitro K+-stimulated 3-O-methylfluoroscein phosphatase (3-O-MFPase) activity (Na+-K+-ATPase activity) in skeletal muscle obtained at rest and fatigue for UT, RT, and ET subjects.A: activity expressed per gram weight wet. B: activity expressed per milligram protein. Data are means ± SE;n = 7 subjects for UT and ET and n = 8 subjects for RT. * Main effect of exercise, fatigue < rest (P < 0.05). To determine whether the decline in activity with exercise may be due to muscle water accumulation, 3-O-MFPase activity was also expressed relative to muscle protein content, which did not differ between rest and fatigue samples or between groups (data not shown). A significant time main effect for maximal in vitro 3-O-MFPase activity expressed per milligram protein was again evident, with a −10.5 ± 3.4% reduction at fatigue (Fig. 2B,P < 0.05). No significant group main effects or time-by-group interactions were seen for maximal in vitro 3-O-MFPase activity per milligram of protein. There were no differences between groups in the absolute or percentage decline from resting values in maximal in vitro 3-O-MFPase activity (absolute: UT −141 ± 26, RT −115 ± 93, ET −159 ± 88 pmol · min−1 · mg protein−1; relative: UT −12 ± 2, RT −8 ± 9, ET −11 ± 6%). A significant time main effect was found for arterialized-venous plasma [K+] (P < 0.05), which increased midexercise (P < 0.05), peaked at fatigue, and returned to rest values by 2 min recovery (Fig.3A). No significant group main effect or time-by-group interactions were found for plasma [K+]. No between-group differences were found during the fatigue test for peak plasma [K+] (UT 4.81 ± 0.17, RT 4.57 ± 0.17, ET 4.60 ± 0.09 mmol/l), Δ[K+] (UT 0.92 ± 0.13, RT 0.60 ± 0.12, ET 0.85 ± 0.09 mmol/l), or Δ[K+]/work (UT 85.7 ± 13.0, RT 59.7 ± 11.6, ET 77.6 ± 8.5 nmol · l−1 · J−1). Fig. 3.Arterialized-venous plasma K+ concentration ([K+]) for UT (●), RT (□), and ET (▴) subjects during muscle fatigue test (A) and during recovery incremental test (B). InA, plasma K+ concentration ([K+]) is shown at rest (R), midexercise (M), and end exercise (E). InB, plasma [K+] is shown at rest, each minute during exercise, and during 30 min of recovery. In B, horizontal error bars indicate mean ± SE peak work rate. The peak incremental exercise plasma [K+] is replotted as the zero-recovery time point. * Main effect of exercise time (A) and/or work rate (B) [K+] > rest (P < 0.05). Data are means ± SE;n = 8 for UT and ET, n = 7 for RT. A significant time main effect was found for ΔPV (P< 0.05), whereby PV fell at fatigue until 2 min of recovery (P < 0.05) and returned to rest by 10 min recovery (Table 1). No significant group main effect or time-by-group interactions for ΔPV were found; thus electrolytes were not corrected for ΔPV. A significant time main effect was found for arterialized-venous plasma [Lac−] (P < 0.05), which peaked at 1–5 min recovery. A significant time-by-group interaction was found for plasma [Lac−], which was less in ET than in UT and RT from 2 until 10 min recovery (P < 0.05, Table 1). A significant time main effect was found for plasma [H+] (P < 0.05), which increased at fatigue, peaked at 5 min postexercise, and returned to rest levels by 20 min recovery. A significant time-by-group interaction was found for plasma [H+], which was lower at 5 min recovery in ET than UT and RT (P < 0.05, Table 1).
A significant time main effect was found for arterialized-venous plasma [K+] (P < 0.05), which increased above rest from 75 W until the peak incremental exercise work rate and had returned to rest by 5 min recovery (Fig. 3B). No significant group main effect or time-by-group interactions were found for plasma [K+]. No between-group differences were found for peak plasma [K+] (UT 6.14 ± 0.17, RT 6.07 ± 0.11, ET 6.43 ± 0.25 mmol/l) or Δ[K+] (UT 2.20 ± 0.16, RT 2.11 ± 0.15, ET 2.42 ± 0.19 mmol/l), but the Δ[K+]/work was 36% lower in ET (11.8 ± 0.4 nmol · l−1 · J−1) compared with UT (18.5 ± 2.3 nmol · l−1 · J−1,P < 0.05) and also tended to be lower than in RT (16.2 ± 2.2 nmol · l−1 · J−1,P = 0.09). A significant time main effect was found for ΔPV (P< 0.05), whereby PV fell below rest from 125 W until 1 min of recovery (P < 0.05) and returned to rest by 30 min recovery (Table 2). No significant group main effect or time-by-group interactions for ΔPV were found, and electrolytes were not corrected for ΔPV. The ΔPV at the peak incremental exercise work rate did not differ between groups (Table 2). Significant time main effects were found for arterialized-venous plasma [Lac−] and [H+] (P < 0.05), which rose above rest from 175 W, peaked at 1 and at 5 min recovery, respectively, and remained above rest at 30 min recovery (Table 2).
In resting muscle samples, a significant correlation was found between [3H]ouabain binding site content and the maximal in vitro 3-O-MFPase activity (Fig.4A, r = 0.61,n = 28, P < 0.05). The [3H]ouabain binding site content was inversely correlated with the fatigue index (Fig. 4B, r = −0.42,n = 24, P < 0.05) and correlated withV˙o2 peak (Fig.5A, r = 0.64,n = 23, P < 0.05). The maximal in vitro 3-O-MFPase activity was also correlated withV˙o2 peak (Fig. 5B,r = 0.46, n = 22, P < 0.05) but not with the fatigue index (r = 0.24, NS). Fig. 4.Relationship between [3H]ouabain binding site content (Na+-K+-ATPase content; pmol/g wet wt) and 3-O-MFPase activity (Na+-K+-ATPase activity; nmol · min−1 · g−1wet wt, n = 28) (A) and fatigue index during the muscle fatigue test (%, n = 24) (B) in UT (●), RT (□), ET (▴), and age-matched controls (■). Fig. 5.Relationship between peak O2 consumption (V˙o2 peak; l/min) and 3-O-MFPase activity (Na+-K+-ATPase activity, nmol min−1 · g−1 wet wt,n = 22) (A) and Na+-K+-ATPase content (pmol/g wet wt,n = 23) (B) for pooled UT (●), RT (□), and ET (▴) subjects. For the incremental test, significant inverse relationships were found between the incremental Δ[K+]/work and both the maximal in vitro 3-O-MFPase activity (r = −0.53,n = 22, P < 0.05, Fig.6A) and the Na+-K+-ATPase content (r = −0.49, n = 24, P < 0.05, Fig.6B). However, for the fatigue test, which involved only a small contracting muscle mass, no significant relationships were found between either maximal in vitro 3-O-MFPase activity or Na+-K+-ATPase content in resting muscle samples, against Δ[K+] or Δ[K+]/work (n = 22). Fig. 6.Relationships between the incremental exercise (Δ[K+]/work; nmol · l−1 · J−1) and maximal in vitro 3-O-MFPase activity (Na+-K+-ATPase activity; nmol · min−1 · g−1wet wt, n = 22) (A) and [3H]ouabain binding site content (Na+-K+-ATPase content; pmol/g wet wt,n = 24) (B) for UT (●), RT (□), and ET (▴) subjects. DISCUSSIONThe most important and novel finding of this study was that an acute bout of fatiguing dynamic exercise depressed the skeletal muscle maximal in vitro 3-O-MFPase activity, which is a measure of Na+-K+-ATPase activity. The similar decline at fatigue in 3-O-MFPase activity expressed relative to muscle wet weight or protein content argues strongly against the possibility of an artifactual effect due to a contraction-induced vascular fluid shift into muscle. Although we did not measure [3H]ouabain binding site content in the fatigued sample, it seems improbable that a loss of Na+-K+-ATPase pump units could occur in this time frame. Rather, an inactivation of these pump units is the more likely explanation. This is the first time such a depression has been demonstrated with dynamic exercise and implicates Na+-K+-ATPase as an additional potential site for muscle fatigue during intense muscle contractions. This finding appears somewhat paradoxical, because it is well established that Na+-K+-ATPase activity is increased during muscle contractions, both in isolated animal muscles (8, 33) and in humans (42, 45). This finding of depressed maximal in vitro Na+-K+-ATPase activity, consistent with a recent report after repeated isometric contractions (11), does not argue against an increase above rest in Na+-K+-ATPase activity in contracting muscle. Indeed, the rapid postexercise decline in plasma [K+] provides some evidence that Na+-K+-ATPase activity is raised well above resting levels. Rather, these findings indicate a reduction in the maximal attainable Na+-K+-ATPase activity with fatigue. In isolated rat soleus muscle stimulated at high frequency (e.g., 120 Hz), Na+-K+-ATPase activity, measured by intracellular Na+ extrusion, increased to maximal theoretical levels (8). In human muscles, however, in which excitation frequencies are much lower, it is likely that activation is less than maximal theoretical levels (see Ref.31). Thus a decline in the maximal in vitro Na+-K+-ATPase activity suggests a reduced safety factor in the attainable Na+-K+-ATPase activation, which may then be important in fatigue. Marked disturbances in muscle intracellular Na+ and both intracellular and extracellular K+ concentrations, reductions in muscle membrane potential, and excitability have been shown with fatigue in human muscles (Ref. 11 and references in Ref.42). We speculate that the depressed maximal Na+-K+-ATPase activity with fatigue might exacerbate these perturbations and thus accelerate muscular fatigue. It is important to note that ouabain-induced inhibition of Na+-K+-ATPase rat soleus muscle markedly enhanced the rate of fatigue (6). Whether the smaller fraction of Na+-K+-ATPase inhibited in this study has similar effects on fatigue has not yet been tested. Interestingly, this depression in Na+-K+-ATPase activity also appears to be reversible, at least after isometric contractions (11), further suggesting a link with fatigue rather than muscle damage. One possible criticism of this finding is that the 3-O-MFPase activity represents steps performing only part of the overall Na+-K+-ATPase cycle (2), and thus reduced phosphatase activity may not reflect reduced maximal Na+-K+-ATPase activity. It is important to appreciate, however, that methodological considerations govern our using this assay rather than utilizing the traditional direct measures of activity via rates of Piaccumulation. In human muscle, it is not possible to detect Na+-K+-ATPase activity by Piliberation because of the small sample yield of the biopsy technique, together with the overwhelmingly high total ATPase activity relative to Na+-K+-ATPase activity (3). Hence, the 3-O-MFPase activity assay was utilized here as the best available method (3, 12). Furthermore, through abolition of 3-O-MFPase activity by ouabain, we have demonstrated that this assay is specific for Na+-K+-ATPase (12). In addition, we report a significant correlation between 3-O-MFPase activity and ouabain binding site content in human muscle, as has been previously reported in rat (38) and human muscle (11). Assuming a molecular activity of 620 cycles/min, the 3-O-MFPase activity in UT muscle of 207 nmol · g wet wt−1 · min−1corresponds to an estimated [3H]ouabain binding site content value of 333 pmol/g, in excellent agreement with our measured value of 311 pmol/g. Although we cannot be certain that 3-O-MFPase activity inhibition will reflect inhibition of total Na+-K+-ATPase activity, this seems highly probable. It is not possible from this study to ascertain the exact mechanisms underlying this depression with fatigue in maximal in vitro 3-O-MFPase activity. However, given identical and controlled assay conditions for the rest and fatigue muscle samples, this most likely reflects a structural alteration in the Na+-K+-ATPase enzyme and/or altered characteristics of the membrane in which it is embedded. Fowles et al. (11) similarly reported a depression in the maximal in vitro 3-O-MFPase activity after repeated isometric contractions. Their depression was larger than in the present study, possibly reflecting the greater disturbances occurring with ischemia than with dynamic contractions. Possible underlying mechanisms include elevated intracellular Na+ and Ca2+ concentrations ([Na+] and [Ca2+], respectively) and reactive oxygen species. Intracellular [Na+] is increased twofold with exercise in human muscle (42) and causes reduced Na+-K+-ATPase activity in rat cerebellum slices, probably because of increased intracellular [Ca2+] (28). Intense muscle contractions induce Ca2+ entry via Na+ channels and Ca2+ accumulation (14), increased resting [Ca2+], and delayed posttetanic Ca2+transients (for references, see Ref. 26). Furthermore, increased [Ca2+] can decrease the Na+-K+-ATPase hydrolytic and transport activities (19, 43, 44, 48), even at nanomolar [Ca2+] (44). Reactive oxygen species are produced during intense muscle contractions (41) and inhibit Na+-K+-ATPase activity in a variety of tissues (25). Finally, these mechanisms may also be linked with increased [Na+] and [Ca2+] culminating in nitric oxide and peroxynitrate formation (47). Na+-K+-ATPase inactivation may also occur as a result of phosphorylation of the α subunit (5). It is unlikely that muscle metabolic perturbations can account for the present findings, because no significant correlations were found between muscle metabolites (26) and the maximal in vitro 3-O-MFPase activity (data not shown). It is unclear whether the depression in activity was due to a small depression in activity in all Na+-K+-ATPase enzymes in all muscle fibers or represented larger depressions in activity in Na+-K+-ATPase enzymes in selected fibers. However, we found no relationship between the decline in maximal in vitro 3-O-MFPase with fatigue and fiber composition (r = 0.19, NS), arguing against a fiber-specific effect. In the present and a previous paper (26), our laboratory has demonstrated for the first time an impairment with fatigue in both of the major cation active transport regulatory proteins, Na+-K+-ATPase and Ca2+-ATPase. With a similar structural homology (20), this leads to the intriguing question as to whether common mechanisms underlie these reductions. Surprisingly, however, we found no significant correlations between the percentage reduction with fatigue in the maximal in vitro 3-O-MFPase and Ca2+-ATPase activities (n = 22,r = −0.03). Thus different mechanisms appear to be involved in these processes. Further work is clearly required to determine the mechanisms involved in depression of Na+-K+-ATPase activity. The second major and unique finding from this study was that chronic training did not attenuate the decline in maximal in vitro Na+-K+-ATPase activity with fatigue. However, this observation further strengthens the validity of the depression in Na+-K+-ATPase activity, suggesting that this is an obligatory acute response to fatiguing muscular contractions. The ET athletes had higher V˙o2 peak and muscular fatigue resistance, suggesting that underlying enhanced muscle oxidative capacity typical of ET does not protect against the depression in Na+-K+-ATPase activity with fatigue. The lack of training status effect on the depression in maximal Na+-K+-ATPase activity does not invalidate our argument that this is intimately involved in fatigue. One possibility is that the upregulation in Na+-K+-ATPase content with training (9,15, 29) is an adaptive process to offset the functional consequences of a decline in maximal activity. Thus ET would demonstrate enhanced muscle performance despite an unchanged depression in maximal Na+-K+-ATPase activity. We acknowledge that one limitation in this study was the cross-sectional design, and thus our data cannot exclude a possible protective adaptation as may be ascertained with a longitudinal training program. We report for the first time skeletal muscle maximal Na+-K+-ATPase activity in athletes. The maximal in vitro 3-O-MFPase activity in resting muscle tended to be 20% higher in the ET group compared with UT, consistent with their 17% higher Na+-K+-ATPase content and with similar findings from longitudinal endurance training studies (9,15, 16, 27). The lack of significance may reflect a type II error due to the higher variability seen in the ET group. Surprisingly, no difference was found in either the 3-O-MFPase activity or [3H]ouabain binding site content between the RT and UT groups. This contrasts the 16% increase in [3H]ouabain binding content reported with resistance training (16), the 15% higher [3H]ouabain binding site with intensified resistance training (32), and the 45% higher [3H]ouabain binding site content in resistance-trained older men (23). The reason for this discrepancy is unclear but might reflect a lesser training level or shorter training duration in our subjects compared with previous studies (16, 23). We demonstrate an important functional role of Na+-K+-ATPase for muscle contractile performance in humans via relationships between the maximal in vitro 3-O-MFPase activity, [3H]ouabain binding site content, and two indexes of dynamic muscular performance: fatigability during repeated quadriceps contractions and the peak incremental exercise O2 uptake. Both the maximal in vitro 3-O-MFPase activity and the [3H]ouabain binding site content were significantly correlated withV˙o2 peak. A novel finding in human muscle was the significant inverse relationship between the [3H]ouabain binding site content and fatigability. This is consistent with studies in rat muscle in which Na+-K+-ATPase activation correlated with contractile performance (7, 33), although other studies in humans have failed to find a relationship between Na+-K+-ATPase content and muscle endurance during fatiguing isometric contractions (23) or repeated sprints (29). Others found depressions in each of muscle isometric force, M-wave area, and maximal in vitro 3-O-MFPase activity after isometric contractions, although these were not reported as being directly linked (11). Surprisingly, we did not find a positive relationship between either the maximal in vitro 3-O-MFPase activity or the fatigue-induced decline in 3-O-MFPase activity with the fatigue index. This may reflect variability in the 3-O-MFPase activity measures but is also consistent with multiple additional factors contributing to muscle fatigue, including impaired sarcoplasmic reticulum Ca2+ release and uptake (4, 26), K+ loss (42), metabolic perturbations such as increased intracellular Pi(10), and fatigue of the central nervous system (13). Finally, we tested whether the maximal in vitro 3-O-MFPase activity and the [3H]ouabain binding site content in muscle were linked with plasma K+ regulation during exercise. We have previously used Δ[K+]/work as a marker of adaptive training effects on plasma K+ regulation during exercise (18, 29). Here we show a lesser Δ[K+]/work during the incremental test in ET (and a tendency in RT) compared with UT subjects, in support of similar findings after sprint training (18, 29) and reduced hyperkalemia reported after endurance training (15, 22). The reduced Δ[K+]/work seen in the incremental cycling test was not evident during the muscle fatigue test, but this is not unexpected because of the smaller contracting muscle mass and consequent lower plasma [K+] during one-leg maximal exercise. An important functional role for muscle Na+-K+-ATPase in plasma K+regulation during exercise in humans was shown by the significant inverse relationship between the incremental exercise Δ[K+]/work and both the maximal in vitro 3-O-MFPase activity and the [3H]ouabain binding site content. The reduced Δ[K+]/work with chronically trained subjects may be due to reduced K+release from contracting muscles, as well as enhanced K+clearance by noncontracting muscle and/or other tissues, but the relative importance of these with training remains to be verified. Reduced Na+-K+-ATPase activity with fatigue in the endurance-trained subjects could be offset by their increased Na+-K+-ATPase content, thereby reducing the rise in plasma [K+] and helping explain their lower Δ[K+]/work. If an inhibition of Na+-K+-ATPase activity also occurred in noncontracting muscle, this would exacerbate the rise in plasma [K+]; however, this seems unlikely and was not tested in this study. In conclusion, acute exercise depressed maximal in vitro 3-O-MFPase activity in untrained, resistance-trained, and endurance-trained individuals. This finding, together with the related paper (26), points to a generalized downregulation of muscle cation regulatory active transport processes during intense contractions and suggests that these are intimately involved in fatigue. Important functional roles for muscle Na+-K+-ATPase were shown for both muscle performance and plasma K+ regulation during exercise. We thank our subjects for their generosity and hard work and Drs. Andrew Garnham, Peter Braun, and Judy Morton for performing the muscle biopsies. FOOTNOTESREFERENCES
Page 5anion channels, extensively expressed in the heart (17, 30), are mainly responsible for the transport of Cl−, the most abundant extracellular anion (14). The currents carried by Cl− are believed to play roles in the regulation of both the electrical activity of cardiac myocytes and cardiovascular functions (15, 17, 30). Previous studies have demonstrated that Cl− substitution has pronounced effects on electrical activities (12, 15, 29, 32) and contraction (13,25) in cardiac myocytes. In addition, Cl−substitution or putative anion channel blockers (ACBs) have frequently been reported to significantly modify the activity of other cardiac ion channels, such as the hyperpolarization-activated current (1,12), potassium channels (3, 39), and Na+ and Ca2+ channels (5, 24, 38). However, the mechanisms underlying the effects of ACBs and Cl− substitutes on other cardiac ion channels, which were hitherto attributed to nonspecific effects (5, 24, 38, 39) or unpredictable side effects (17), remain unclear. In cardiac ventricular myocytes, the most important physiological event for cardiac function is excitation-contraction coupling (ECC). During membrane depolarization, extracellular Ca2+ enters the cell via L-type Ca2+ channels and triggers a Ca2+-induced Ca2+ release from sarcoplasmic reticulum (SR) (10, 40). The resultant cytosolic Ca2+ transient (CaT) activates myofilament proteins and initiates contraction. Although different anion channel currents have been characterized in the heart (17, 30), the effects of these channels on ECC in cardiac myocytes remain largely unknown. Therefore, the purpose of the present study was to investigate the roles of anion channels in the regulation of cardiac ECC. MATERIALS AND METHODSVentricular myocytes were enzymatically isolated from the hearts of adult female Sprague-Dawley rats (200–250 g) by using essentially standard procedures, as reported previously (11). Briefly, the hearts were removed immediately after decapitation and retrogradely perfused with oxygenated Tyrode solution at 37°C for 5 min and then with Ca2+-free Tyrode solution for 5 min before the addition of 0.5 mg/ml collagenase (type I, Sigma Chemical, St. Louis, MO) and 1 mg/ml BSA to the same solution. After 35 min of digestion with the collagenase-containing solution, the hearts were perfused for 5 min with a Kraftbrühe (KB) (high K+) solution containing (in mM) 70 l-glutamic acid, 25 KCl, 20 taurine, 10 KH2PO4, 3 MgCl2, 0.5 EGTA, 10 glucose, and 10 HEPES-KOH (pH 7.35). Subsequently, the ventricular tissue was cut into small pieces in an oxygenated KB solution. After gentle stirring with a glass rod for 5 min, the myocyte-containing solution was filtered through a nylon mesh. The cells were maintained in KB solution at room temperature (23–25°C) for electrophysiological recordings. For measurements of cell shortening and CaT, after 30 min in KB solution, aliquots of cell suspension were sedimented by centrifugation at 100 g for 1 min and then resuspended in Ca2+-free Tyrode solution with 2% BSA. The Ca2+ concentration of the solution was gradually increased to normal (1.8 mM) within 30 min. Aliquots of cell suspension were transferred into a perfusion chamber on the stage of an inverted microscope (IMT-2, Olympus, Tokyo, Japan). Pipettes had tip resistances of 5–6 MΩ when filled with internal solution. Whole cell recordings were performed at room temperature by using a patch-clamp amplifier (Axopatch 200B, Axon Instruments, Foster, CA). Liquid junction potentials between external and pipette solutions were offset before the pipette touched the cell. In the anion substitution experiments, an agar-salt bridge was used as the reference electrode. To selectively activate the L-type Ca2+ current (ICa,L), the membrane potential was depolarized from −70 to −40 mV, where it was held for 100 ms to inactivate the Na+ and T-type Ca2+ channels. K+currents were eliminated by internal tetraethylammonium chloride (20 mM) and by omission of K+ from pipette and bath solutions. To monitor the effects of drugs, ICa,L was elicited by a depolarization from −40 to 0 mV for 300 ms every 5 s. The current-voltage relationship was assessed by measuring currents at voltage pulses (300 ms) from −40 to +50 mV, applied in 10-mV increments. The current signals were low-pass filtered at 2 kHz and digitized with an analog-to-digital converter (Digidata 1322) and pCLAMP 8.1 software (Axon Instruments, Foster, CA) at a sampling rate of 10 kHz. ICa,L was calculated as the difference between the peak inward current and the holding current level. Cytosolic calcium was measured by the fluorescent calcium indicator fura 2 by using a dual-fluorescence, calcium ion-sensing system (IonOptix, Milton, MA). The myocytes, suspended in Tyrode solution, were incubated with 5 μM fura 2-acetoxymethyl ester (AM) (St. Louis, MO) for 30 min at room temperature and then washed three times with fura 2-AM-free Tyrode solution. The loaded cells were kept at room temperature (23–25°C) for 1 h before measurements of intracellular Ca2+ concentration ([Ca2+]i) to allow the fura 2-AM in the cytosol to deesterify. Loading with a low concentration of fura 2-AM at a relatively low temperature of 23–25°C was done to minimize the effects of the compartmentalization of fura 2 (28). Fura 2-loaded myocytes were placed in a microperfusion chamber mounted on an inverted microscope (IX50, Olympus, Tokyo, Japan). A rod-shaped myocyte with clear striations and sharp edges was localized by microscopic observation, and contractions were elicited by field stimulation delivered at 0.25 Hz through two platinum electrodes mounted on either side of the chamber. Fura 2-loaded myocytes were alternately excited with a xenon lamp at wavelengths of 360 and 380 nm. The emission fluorescence was collected by the objective and reflected through a barrier filter to a photomultiplier tube, as previously described (26). Because each cell was used as its own control, [Ca2+]i was expressed as the 360-to-380 fura 2 ratio. Before myocyte [Ca2+]i was measured, the background fluorescence of the measuring area without a myocyte was set as zero. Cell length and contractile amplitude of myocytes were recorded with a video edge detector and a specialized data-acquisition software (SoftEdge Acquisition System and IonWizard, IonOptix, Milton, MA), as previously described (26). All whole cell patch-clamp recordings and measurements of cell shortening and CaT were performed at room temperature (23–25°C). The normal Tyrode solution contained (in mM) 143 NaCl, 5.4 KCl, 0.5 MgCl2, 1.8 CaCl2, 0.3 NaH2PO4, 5 glucose, and 5 HEPES-NaOH (pH 7.4, 310 mosmol/kgH2O). The nominally Ca2+-free Tyrode solution was prepared by omitting CaCl2 from the normal Tyrode solution. To obtain the desired Cl−substitution solutions, the 148.4 mM Cl− in normal Tyrode solution was replaced by equimolar Glt−, Asp−, acetate (Ac−), NO3−, SCN−, or Br−. In osmotic stress experiments, 90 mM mannitol was added to the normal Tyrode solution to obtain a hypertonic bath solution (400 mosmol/kgH2O). For the hypotonic experiments, a control isotonic solution (310 mosmol/kgH2O) was made by replacing the 35 mM NaCl in the normal Tyrode solution with 70 mM mannitol, and a hypotonic solution (240 mosmol/kgH2O) was obtained by omitting mannitol from the control isotonic solution. In whole cell patch-clamp recordings, the standard bath solution contained (in mM) 140 NaCl, 1.8 CaCl2, 0.5 MgCl2, 10 glucose, and 5 HEPES-NaOH (pH 7.4; 300 mosmol/kgH2O). The control pipette solution contained (in mM) 110 aspartic acid, 10 NaCl, 5 Mg-ATP, 5 EGTA, 20 tetraethylammonium chloride, 5.5 glucose, and 10 HEPES-CsOH (pH 7.3). In some whole cell patch-clamp experiments, the 140 mM Cl− in the standard bath solution was replaced by equimolar Glt−, Asp−, Br−, or SCN−, and Asp− in the pipette solution was replaced by equimolar Cl−. In the osmotic stress experiments, 100 mM mannitol was added to the standard bath solution to make a hypertonic solution (400 mosmol/kgH2O). In hypotonic experiments, a control isotonic bath solution (300 mosmol/kgH2O) was prepared by replacing the 40 mM NaCl in the standard bath solution with 80 mM mannitol, and a hypotonic solution (220 mosmol/kgH2O) was made by omitting mannitol from the control isotonic solution. The free Ca2+ concentration in bath solutions was calculated by using the CaBuf program (provided by Dr. G. Droogmans, Katholieke Universiteit, Leuven, Belgium) by taking the values of pH, temperature, ionic strength, association constants, and Mg2+ into consideration (36). The following chemicals, purchased from Sigma Chemical, were added to the bath solutions: 0.1–100 μM 5-nitro-2-(3-phenylpropylamino)benzoic acid (NPPB) and 0.1–500 μM niflumic acid (NFA). Stock solutions of fura 2-AM (0.5 mM), NPPB (100 mM), and NFA (500 mM) in DMSO were diluted to the desired final concentrations immediately before use. DMSO at a final concentration of ≤0.1% in the bath solution had no effect on contraction, CaT, orICa,L (data not shown). The data are presented as means ± SD. Statistical differences in the data were evaluated by Student's t-test and were considered significant at values of P < 0.05. RESULTSSimultaneous measurements of CaT and cell shortening showed that exposure of the myocytes to 50 μM NPPB decreased CaT and cell shortening by 92.9 ± 12.3 and 98.1 ± 4.6% (n = 13, P < 0.001, Fig.1A), respectively. Concomitant with the reduction in CaT, the diastolic (resting) length increased by 2.5 ± 1.6 μm (n = 13, P < 0.05, Fig. 1A, bottom). The reduced CaT and cell shortening recovered on withdrawal of NPPB. Similar effects were observed with NFA. Bath application of NFA (100 μM) decreased the amplitude of CaT and cell shortening by 36.8 ± 12.8 and 83.6 ± 7.7%, respectively (n = 12, P < 0.001, Fig. 1B). Both compounds suppressed the contraction in a dose-dependent manner with EC50 values of 12.3 μM (NPPB) and 44 μM (NFA) (Fig. 1C). Neither NPPB (50 μM,n = 4) nor NFA (100 μM, n = 4) had detectable effects on resting [Ca2+]i (data not shown). These results indicate that the putative ACBs, NPPB and NFA, have negative inotropic effects on rat cardiac ventricular myocytes as a result of decreased Ca2+ release from the SR and/or decreased Ca2+ influx. Fig. 1.Effects of 5-nitro-2-(3-phenylpropylamino)benzoic acid (NPPB) and niflumic acid (NFA) on electrically induced Ca2+transient (CaT) and contraction in fura 2-loaded cardiac myocytes. Changes in Ca2+ fluorescence ratio and cell shortening were recorded simultaneously. A: representative traces of CaT (top) and cell contraction (bottom) after electrical field stimulation at 0.25 Hz in the absence and presence of NPPB. B: representative traces of electrically induced CaT (top) and cell contraction (bottom) recorded before and after application of NFA. n = 13 (A) and 12 (B) experiments. C: concentration-response relations for the effects of NPPB (●) and NFA (■) on the electrically driven contraction of rat ventricular myocytes. Values are means ± SD; nos. in parentheses indicate the no. of cells observed. To further explore the mechanism of the effects of NPPB and NFA on cardiac ECC, we investigated the effects of the relatively impermeant Cl− substitutes Glt− and Asp− on the CaT and contraction of rat cardiac ventricular myocytes. Replacement of 148.4 mM extracellular Cl− by equimolar Glt− reversibly reduced the amplitudes of the electrically induced CaT and cell shortening by 69.9 ± 17.9 and 94.7 ± 6.2%, respectively (n = 7, P < 0.001, Fig. 2A). Asp−substitution decreased the amplitude of CaT and cell shortening by 34.8 ± 16.4 and 64.3 ± 22.4%, respectively (n = 9, P < 0.001, Fig.2B). These data showed that the inhibitory effects of Asp− substitution were weaker than those of Glt− substitution, although there was no significant alteration in the free [Ca2+]i after replacement of 148.4 mM extracellular Cl− with either equimolar Glt− (free Ca2+ = 1.77 mM) or Asp− (free Ca2+ = 1.75 mM). Neither Glt− (n = 3) nor Asp−(n = 3) induced noticeable changes in resting [Ca2+]i (data not shown). To further determine the role of extracellular anions on the ECC, we compared the effects of Glt− and Asp− with those of Ac−, a relatively small anion. In all 15 cells observed, Ac− substitution for Cl− provoked a biphasic response: an initially transient decrease in CaT and cell contraction followed by an increase (Fig. 2C). The second phase was also sensitive to NPPB (n = 3, Fig. 2C). The inhibitory effects of the putative ACBs and the relatively impermeant Cl− substitutes were comparable, suggesting that their actions may share a common pathway. Fig. 2.Effects of glutamate, aspartate, and acetate substitution for Cl− on electrically induced CaT and contraction of cardiac myocytes. Representative recordings are shown of electrical field stimulation-induced CaT (top) and cell contraction (bottom) before and after replacement of 148.4 mM extracellular Cl− with equimolar Glt−(A), Asp− (B), and Ac−(C). The CaT traces were obtained at times corresponding to that indicated by a–c (B) ora–d (C) in bottom traces. Replacement of extracellular Cl− in the bath solution by Glt− (A), Asp− (B), and Ac− (C) is indicated by the horizontal bars. The sarcolemmal L-type Ca2+ channels are crucial in ECC, because Ca2+ influx through the L-type Ca2+channels triggers Ca2+ release from the SR (40). Therefore, the effects of the ACBs, NPPB and NFA, and impermeant Cl− substitutes onICa,L were investigated. Bath application of NPPB (50 μM) and NFA (100 μM) reversibly decreased the peakICa,L by 60.1 ± 6.7% (n = 6, P < 0.01) and 45.4 ± 11.5% (n = 5, P < 0.01), respectively (Fig.3, A and B). NPPB and NFA reduced the amplitude of ICa,L without altering the reversal potential for ICa,L (Fig.3, C and D). Similarly, replacement of 140 mM extracellular Cl− by Glt− or Asp− suppressed the peak ICa,L by 78.2 ± 19.2% (n = 5, P < 0.01, Fig. 4A) and 36.5 ± 11% (n = 5, P < 0.01, Fig. 4B), respectively. The inhibitory effects of Glt−(n = 3) on ICa,L were not altered when the intrapipette Cl− was increased from the conventional level (30 mM) to 140 mM (data not shown). These data suggest that suppression of the L-type Ca2+ channel is responsible for the inhibition of cardiac ECC induced by the ACBs and the relatively impermeant Cl− substitutes. Fig. 3.Effects of NPPB and NFA on L-type Ca2+ current (ICa,L) in rat ventricular myocytes. Changes inICa,L were monitored by applying a depolarizing pulse (300 ms in duration) from −40 to 0 mV every 5 s.A and B: current traces in top panels were obtained at times indicated by a–c in bottompanels. To obtain current-voltage (I-V) curves of ICa,L, whole cellICa,L was elicited by 300-ms step depolarizations between −40 and +50 mV from a holding potential of −40 mV in increments of 10 mV. C and D:I-V curves of ICa,Lobtained in the absence and presence of NPPB (50 μM) and NFA (100 μM), respectively. Values are means ± SD (n = 5 cells). Fig. 4.Effects of glutamate and aspartate substitution for Cl− on ICa,L in rat ventricular myocytes. ICa,L was elicited by applying a depolarizing pulse (300 ms in duration) from −40 to 0 mV every 5 s. Right: replacement of 148.4 mM extracellular Cl− with equimolar Glt− (A) and Asp− (B) is indicated by the horizontal bars.Left: current traces were obtained at times indicated bya–c in the right panels. A andB each represent 5 experiments. To understand the role of the anion channels in the modulation ofICa,L, we examined the effects of the permeant inorganic anions Br−, NO3−, and SCN− on CaT and contraction in rat ventricular myocytes. In contrast to the effects of the relatively impermeant Cl− substitutes (Fig. 2), replacement of 148.4 mM extracellular Cl− with equimolar Br−increased cell shortening by 48.48 ± 17.5% (n = 7, P < 0.01, Fig.5A, bottom). NO3− and SCN− substitution for extracellular Cl− supported the cell contraction and CaT. NPPB (50 μM) abolished the electrically induced CaT and cell shortening after extracellular Cl− was substituted by Br−, NO3−, or SCN− (Fig. 5). In whole cell patch-clamp experiments, replacement of 140 mM extracellular Cl− with equimolar Br−significantly increased the peak ICa,L in six of seven cells observed (Fig.6A). SCN−substitution for extracellular Cl− supported theICa,L (Fig. 6B). It seemed that the maintenance of L-type Ca2+ channel activity depends, to a large degree, on anion channels. Fig. 5.Effects of replacement of extracellular Cl−by Br−, NO3−, and SCN− on electrically induced CaT and contraction of cardiac myocytes. Changes in Ca2+ fluorescence ratio and cell shortening were recorded simultaneously. Top: representative traces of electrically induced CaT before and after replacement of 148.4 mM extracellular Cl− by equimolar Br−(A), NO3− (B), and SCN− (C). Bottom: representative traces of the cell contraction. Application of NPPB (50 μM) is indicated by the horizontal bars. n = 7 (A), 5 (B), and 5 (C) experiments. Fig. 6.Effects of replacement of extracellular Cl− by Br− (A) and SCN− (B) on cardiac ICa,L. Left: representative traces of ICa,L obtained at times indicated bya and b in the right panels.Right: time courses of changes in the amplitude of theICa,L peak. Replacement of 148.4 mM extracellular Cl− with equimolar Br−(A) and SCN− (B) is indicated by the horizontal bars. n = 7 (A) and 6 (B) experiments. When the myocytes were perfused with hypotonic solution (240 mosmol/kgH2O), the amplitude of cell shortening increased by 28.4 ± 7.6% (n = 5, P < 0.01). The enhanced contraction in hypotonic conditions was also inhibited by NPPB (Fig. 7A). Hypotonic solution (220 mosmol/kgH2O) induced a significant increase in four of eight cells and counteracted the rundown of theICa,L in the other four cells (Fig.7B). On the contrary, exposure of the myocytes to hypertonic solutions (400 mosmol/kgH2O) decreased the cell shortening and ICa,L by 66.9 ± 17.6% (n = 4, P < 0.01, Fig. 7C) and 53.6 ± 7.2% (n = 6, P < 0.01, Fig. 7D), respectively. These results are in good agreement with the previous observation that hyperosmolality reduces the amplitude of ICa,L (27). Moreover, cell diastolic length in hypertonic solution decreased by 4.3 ± 2.2 μm due to cell shrinkage (n = 4,P < 0.05, Fig. 7C). Fig. 7.Effects of osmotic stress on cardiac contraction andICa,L. A: representative trace of cell shortening. Perfusion with hypotonic solution and bath application of NPPB (50 μM) are indicated by the horizontal bars. B: representative traces of ICa,L inleft panel were obtained at times indicated bya–c in the right panel.C: representative recording of cell shortening in isotonic and hypertonic bath solutions. D: representative traces ofICa,L in the left panel were obtained at times indicated by a–c in the rightpanel. Application of hypertonic bath solution is indicated by the horizontal bar. ICa,L was elicited by applying a depolarizing pulse (300 ms in duration) from −40 to 0 mV every 5 s. DISCUSSIONThe major findings of the present study are that the putative ACBs, NPPB and NFA, and the relatively impermeant Cl−substitutes, Glt− and Asp−, suppress the CaT and contraction of isolated rat cardiac ventricular myocytes. The effects of NPPB and NFA on cardiac contraction were similar to those obtained with other ACBs, the disulfonic stilbene derivatives (25). It is well known that the sarcolemmal L-type Ca2+ channel plays a crucial role in ECC, because Ca2+ influx through the L-type Ca2+ channels triggers Ca2+ release from the SR (40). Although little is known concerning the effects of NPPB and NFA on ECC of cardiac myocytes, a handful of studies have revealed that NPPB and NFA inhibit the contraction of vascular smooth muscles, probably mediated by Ca2+ channels (6-8). The present study suggests that NPPB- and NFA-induced depression of the cardiac ICa,L is mainly responsible for the decrease of the CaT and contraction. ACBs are frequently reported to have profound effects on some other ion channels (1, 5, 24, 39), including the L-type Ca2+ channels, in both vascular smooth muscles (8) and cardiac ventricular myocytes (5, 38). These unpredictable effects of ACBs were attributed to nonspecific effects, unrelated to Cl− channel block (5, 8, 19,24, 38, 39). However, these interpretations are challenged by the idea that the effect of NFA on vascular smooth muscles is related to a Cl− channel, which may lead to the opening of a calcium channel (6, 7, 22). In cardiac myocytes, Minocherhomjee et al. (25) have found that the disulfonic stilbene ACBs inhibit the uptake of both 36Cl and45Ca. This evidence suggests that Cl− and Ca2+ movements may be interrelated in some way. In the present experiments, we found that Glt− mimicked the inhibitory effects of NPPB and NFA on cardiac ECC and that Asp− induced partial inhibition of the CaT and contraction, whereas Ac− substitution for Cl−produces a transient inhibition followed by a NPPB-sensitive rebound response. The effects of Glt−, Asp−, and Ac− could not be attributed to their chelation of Ca2+, because there was no significant alteration in the free Ca2+ after replacement of most of the extracellular Cl− by these substitutes. The parallelism between the effects of the ACBs and the relatively impermeant Cl−substitutes suggests that they may share a common mechanism. It is likely that the effects of the ACBs and the relatively impermeant Cl− substitutes on ICa,L may involve either Cl− per se or anion channels. It is arguable that intracellular Cl− per se may play a role in cardiac ECC, because application of ACBs or relatively impermeant Cl− substitutes may result in intracellular Cl− depletion. However, the inhibitory effects of replacement of extracellular Cl− by Glt−remained unchanged while intracellular Cl− concentration was clamped at 140 mM. Moreover, permeant Cl− substitutes supported the contraction, CaT, and ICa,L of the cells (Figs. 5 and 6) but did not alter the inhibitory effects of NPPB (Fig. 5). These data suggest that the reduction in peakICa,L is an intracellular Cl−-independent effect. It appears that activation of the L-type Ca2+ channel needs the support of anion channel activity. The driving force for Cl− movement through the sarcolemmal membrane via anion channels depends on the equilibrium potential of Cl−. In the present electrophysiological experiments with the gradient of 144.6 mM extracellular Cl−concentration/30 mM intracellular Cl− concentration, the equilibrium potential of Cl− estimated by the Nernst equation was ≈−41 mV. If some Cl− moved across the membrane during the ICa,L measurements (from the holding potential of −40 to +50 mV), the current carried by Cl− would be expected to produce an outward-going current opposite in direction to the ICa,L due to Cl− influx. If so, inhibiting anion channels would be expected to increase the amplitude of the inward current. However, blockage of anion channels with ACBs inhibitedICa,L instead of enhancing it, an effect that was mimicked by the relatively impermeant Cl− substitutes, whereas permeant Cl− substitutes supported the L-type Ca2+ channels. These observations suggest that regulation of the L-type Ca2+ channel by an anion channel is not merely a question of an anion channel-mediated chloride conductance. The regulation may involve a channel-channel interaction mechanism, which has been well established among epithelial channels (4, 21,33). Three major types of Cl− channels have been detected in adult mammalian heart (30). Among these, the cystic fibrosis transmembrane regulator (CFTR), a Cl− channel activated by protein kinase A, is the least likely to be involved in modulating the L-type Ca2+ channel because CFTR is not detectable in adult rat ventricle (9). The Ca2+-activated Cl− channel (ICl,Ca) may play a role in the modulation ofICa,L; however, buffering by high intracellular EGTA prevents the activation of cardiac ICl,Ca(43). In the present patch-clamp experiments,ICl,Ca was expected to be eliminated by chelating internal free calcium with high EGTA (5 mM). Swelling-activated Cl− channel (ICl,swell), expressed ubiquitously in mammalian cells, is known to be blocked by NPPB, NFA, stilbene disulfonates, and diphenylamine-2-carboxylate (20, 31, 34, 35,42). The present study found that the cardiac contraction andICa,L were inhibited not only by NPPB and NFA but also by hypertonic stress, whereas hypotonic challenge increased the contraction and supported the ICa,L. These results suggest that the anion channel involved in the regulation ofICa,L is a ICl,swell. Because blockage of ICl,swell by NPPB and NFA under isotonic conditions produces pronounced effects on cardiac contraction and ICa,L, it seems that basalICl,swell activity is important for maintaining the activation of the cardiac L-type Ca2+ channel. Anion channels are ubiquitously present in cells. If there are channel-channel interactions between anion channels and other channels (such as cation channels), blocking anion channels with their inhibitors would be expected to alter the activities of other related channels. In epithelia, CFTR, an anion channel, is found to modulate other channels via channel-channel interaction (4, 21,33). For this reason, inhibition of CFTR with ACBs must have a profound influence on such ionic channels as those regulated by CFTR. Therefore, this channel-channel interaction may be responsible, at least partially, for the “nonspecific effects” of ACBs observed in epithelial cells. In heart, it is frequently reported that the potency of inhibition of cardiac cation channels by ACBs that differ in structure seems to be similar to that of reduction of anion channels (1, 5, 24, 38). However, the possibility of direct interaction between cardiac anion channels and cation channels is often ignored. Although some studies mentioned the possible relation between the cation channels and some anion channel before they attributed the effects of ACBs on ICa,L to nonspecificity (24, 38), the limitation of those studies is that some other important anion channels in heart (such asICl,swell) were not taken into consideration. Moreover, the nonspecific effects of ACBs observed in heterologous expression experiments should be carefully explained before excluding the possibility that the heterologously expressed ionic channels may be influenced by the intrinsic anion channels of host cells. The present study suggests that the mechanism of influence ofICa,L by ACBs and Cl− substitution may be mediated by direct channel-channel interaction between anion channels and L-type Ca2+ channels. Although the present study found that either ACBs, or impermeant Cl− substitutes, or hyperosmolality significantly inhibit the ICa,L, which may be mainly responsible for the decrease in Ca2+-induced Ca2+ release, some other factors may also be involved in the changes of cardiac ECC, because ACBs and Cl− substitution also influence other cardiac ion channels (5, 39) apart fromICa,L. Our primary study found that cardiac Na+ current was also significantly inhibited by either NPPB or relatively impermeant Cl− substitutes, Glt− and Asp−, although the resting membrane potential was not changed significantly under those conditions (data not shown). The mechanism of the ACB- and Cl−substitution-induced changes in Na+ current and the role of these changes in cardiac ECC are under investigation. Several studies reported that ACBs increased K+ currents in expressed cell systems or noncardiac cells. For example, NFA at low concentration (10 μM) increased a slow voltage-activated K+ current expressed in oocytes, but inhibited the slow voltage-activated K+ current at higher concentration (100 μM) (2). In smooth muscle cells from rabbit portal vein, NFA increased the noradrenaline-evoked Ca2+-activated K+ current but had no effect on spontaneous Ca2+-activated K+ current (16). In contrast, in rat heart, the transient outward K+ current (Ito), a major repolarizing ionic current in ventricular myocytes, was inhibited by either chloride channel blockers or relatively impermeant Cl− substitutes (23). In agreement with the previous observation, our experiments also verified that aspartate substitution for external Cl− strongly inhibited Ito(n = 3, data not shown). Moreover, Kv4.3, an important K+ channel responsible for Ito in rat cardiac cells, was significantly decreased by NFA, flufenamic acid, and disulfonic stilbenes (39). In addition, a volume-sensitive basolateral K+ current was blocked by NPPB in HT-29/B6 cells (18). It is well known that inhibition of Ito leads to a significant prolongation of the action potential duration (37). Accordingly, the cell contractility is expected to increase in this condition. Thus it seems unlikely that the strongly inhibitory effects of Cl−channel blockers and relatively impermeant Cl− substitutes on CaTs and cardiac contractility are due to increased K+currents. The effects of ACBs and Cl− substitution on cardiac K+ currents should be subjected to further investigation. Taken together, our data demonstrate that anion channels play an important role in cardiac ECC and suggest that the cardiacICl,swell might be involved in the regulation of cardiac ECC mainly by modulating the L-type Ca2+ channel. The present study provides insight into the mechanisms of the control of both electrical and contractile activities in cardiac ventricular myocytes. The L-type Ca2+ channel is crucial to cardiac ECC under physiological conditions. Swelling-activated Cl−current may increase during myocardial ischemia-reperfusion due to cell swelling (41). Therefore, investigation of the interaction between the ICl,swell and the L-type Ca2+ channel may have profound physiological and clinical significance. The authors thank Dr. Iain C. Bruce for reviewing the manuscript, Dr. G. R. Li for helpful discussion, and Dr. G. Droogmans for the kind gift of the CaBuf program. FOOTNOTESREFERENCES
Page 6insulin-like growth factor (IGF) I plays an important role in tissue anabolism by causing cell hypertrophy and hyperplasia in various cell types, including skeletal muscle myoblasts and tendon fibroblasts (1, 2,11, 41). In addition, exercise training per se affects tissue anabolism (29, 30), and may provide general health benefits (6, 33), but exactly how exercise training and tissue anabolism interact and what mechanisms are responsible remain unclear. Although inconsistency exists (22), several cross-sectional studies have found significant positive correlations between fitness level and circulating IGF-I levels (9, 35). Such a correlation indicates that fitness and exercise in healthy, adolescent subjects is associated with increased activity of the IGF system, favoring an anabolic state. In contrast, prospective studies on exercise training and IGF-I have reported decreased circulating IGF-I levels (9, 10, 24, 38), mimicking responses typically found during energy-deficient and catabolic states (10, 24, 38, 39). However, these studies often had a rather short duration (3 days to 5 wk), and one paper hypothesized that the IGF-I adaptation to chronic exercise in fact consisted of a two-phase response, where an initial catabolic-type response was followed by a more chronic anabolic state after prolonged (>5–6 wk) training (10). That prolonged training does result in an increased activity of the IGF system is supported by findings from animal studies demonstrating that longer periods of training (4–9 wk) result in increased IGF-I gene expression in skeletal muscular tissue (44) and in increased circulating IGF-I levels (43). However, it is unclear to what extent the hypothesized two-phase response of IGF-I adaptations also holds true for humans during prolonged physical training. The IGF-I action is believed in part to be mediated by circulating free IGF-I. During exercise and training, IGF binding protein (IGFBP)-3, through its proteolysis, represents a potent regulator of IGF-I bioactivity that can result in elevated concentrations of free IGF-I (37). Furthermore, increased IGFBP-3 proteolysis has been reported in one study that used an acute bout of exercise (37), whereas in a recent study on acute exercise no increase in IGFBP-3 proteolysis could be demonstrated (7). Independent of these diverse observations on responses to acute exercise, most previous studies have not addressed any potential role for IGFBP-3 proteolysis during a more chronic exposure to exercise and training. Nor has the influence of training on other IGFBPs and IGF-II been fully elucidated. Accordingly, the aim of the present investigation was to study the regulatory changes in the circulating IGF system, including binding proteins and IGFBP-3 proteolysis, in response to prolonged physical training in young men. It was investigated whether the adaptation consisted of a two-phase response with an early adaptive effect (after 4 wk of training) and a more chronic adaptation (after 11 wk) as proposed by Eliakim et al. in 1998 (10). This was studied in both well-trained and untrained subjects to determine any potential influence of training status on responses. We hypothesized that highly standardized training would result in more extensive changes in the circulating IGF system in untrained subjects because of lower fitness level. METHODSSeven untrained and twelve well-trained healthy men, previously screened by the military medical board, were chosen from a larger group of conscripts who had volunteered for military duty at the Royal Danish Life Guard training camp. The Royal Danish Life Guard is considered one of the elite troop regiments within the Army, consisting of highly motivated soldiers. The training status of the subjects was determined on the basis of a fitness questionnaire that addressed the weekly amount of physical training during the last 12 mo before the study. The questionnaire also included a subjective assessment of current physical fitness level compared with that of other men at the same age, an assessment that has previously been validated by Washburn et al. (42). Each subject's individual training status was verified by a direct maximal oxygen consumption (V˙o2 max) determination. Subject characteristics are provided in Table 1. None of the subjects performed any training for 1 wk before the study, none was on any medication, and all were nonsmokers. Informed written consent was obtained from each subject before he was included in the study, which conformed to the Declaration of Helsinki and was approved by the Ethical Committee for medical research in Copenhagen (KF-01-118/99). Participation in the study was blinded toward the military that was responsible for the physical training.
Fasting serum samples were collected before and after 4 and 11 wk of training. The 11 wk of standardized training consisted of 2–4 h daily training, categorized into four major groups (closed-order drill and marching, general conditioning, military-specific training, and open-order combat training), and each training session was conducted by a training officer. On average, the weekly training included 7 h of marching, 4 h of general conditioning (mostly running), and 9 h of military-specific training. Overall, the physical training primarily consisted of aerobic, lower extremity weight-bearing activities, but for 5–10% of the training time, intensity was aimed to be above the anaerobic threshold. The weekly amount of vigorous physical training was quantified by review of the recruit training schedules and by training reports from the training officers. V˙o2 max was determined during the week before training, and it was evaluated during week 11with an incremental V˙o2 max test on a motor-driven treadmill. TheV˙o2 max protocol used was designed to have the subjects reach the V˙o2 maxpoint within 3–7 min. After a 15-min warm-up, the subjects ran at an individually determined constant speed (13–16 km/h) throughout the test. After the first 2 min of the test, the treadmill was elevated 2° every 90-s until the subjects were unable to keep up with the pace. In the present study, all subjects completed the 2° elevation step (3.5 min) with an exercise range of 4.5–6.5 min, ensuring that V˙o2 max was reached. A criterion to establish that V˙o2 max was reached was that the respiratory quotient exceeded 1.10. Pulmonary oxygen uptake was measured with the Innovision Amis 2001 mixing chamber, on-line equipment (Odense, Denmark). All blood sampling was conducted after an overnight fast between 0630 and 0730 (baseline, week 4, and week 11). In addition, to avoid the effects of the previous training session, the subjects were not exposed to any strenuous exercise on the day before blood sampling, and each subject was ensured a minimum of 14 h of rest. The blood samples were drawn from an antecubital vein of the nondominant arm and were immediately iced and centrifuged at 5,000 rpm for 15 min (at 4°C). The serum was stored at −80°C until analysis. Serum total IGF-I and IGF-II were determined after acid-ethanol extraction by using noncompetitive time-resolved monoclonal immunofluorometric assays as previously described (16). Serum free IGF-I and IGF-II were determined by using ultrafiltration by centrifugation (19). Serum IGFBP-1 was determined by an enzyme-linked immunoassay (Medix Biochemica, Kauniainen, Finland). Serum IGFBP-3 was measured by immunoradiometric assay (IRMA) (Diagnostic System Laboratories, Webster, TX). The within-assay coefficient of variation (CV) for total IGF-I, IGF-II, IGFBP-3, and IGFBP-1 averaged <5%. The within-assay CV for free IGF-I and free IGF-II averaged 18 and 12%, respectively. All samples were determined within the same assay. Western ligand blotting (WLB), SDS-PAGE, and ligand blot analysis were performed in serum according to the method of Hossenlopp et al. (23) as previously described (12). Two microliters of serum were subjected to SDS-PAGE (10% polyacrylamide) under nonreducing conditions. Specificity of the IGFBP-2, IGFBP-3, and IGFBP-4 bands on WLBs was supported by competitive coincubation with unlabeled recombinant human IGF-I purchased from Bachem (Budendorf, Switzerland). The 125I-labeled IGFBP-3 degradation assay was performed as previously described (7, 28). 125I-IGFBP-3 (30,000 counts/min) (Diagnostic System Laboratories) was incubated for 18 h at 37°C. Two microliters of serum from the subjects were subjected to SDS-PAGE as described above. On each gel, serum samples from a healthy nonpregnant subject and term-pregnant woman were used as internal controls. Gels were fixed in a solution of 7% acetic acid, dried, and autoradiographed. The degree of proteolysis was calculated as a ratio of the absorbency of fragmented 125I-IGFBP-3 to the sum of all 125I-IGFBP-3-related optical densities in that lane and was expressed as a percent. The between-assay CVs for the WLBs and the 125I-IGFBP-3 degradation assay were below 10%. Autoradiograms of WLBs and the IGFBP-3 protease assay were quantified by densitometry with a Shimadzu CS-9001 PC dual-wavelength flying spot scanner (Shimadzu Europe, Duisburg, Germany). The relative density of the bands was measured as arbitrary absorbance units. All data are presented as means ± SD. One-way ANOVA for repeated measures, approached by general linear modeling, was used to test for time effect during the training period (baseline, week 4,week 11). If the ANOVA test revealed significant changes, post hoc analyses by Tukey's multiple comparison were used to compare specific pairs of means. To examine whether differences existed between untrained and well-trained subjects, the independent-samplest-test was used (effect of training status) (SPSS Standard Version 11.0). Pearson's correlation was used to address the relationship between changes in free IGF-I and changes in IGFBP-1 and IGFBP-2. An alpha level of <0.05 (2 tailed) was accepted as significant. RESULTSThe baseline characteristics of the subjects are presented in Table 1. Difference in training amounts,V˙o2 max, body weight, and body mass index were significant between well-trained and untrained subjects, whereas no significant differences were found between groups with regard to baseline IGF variables (Table2, Figs.1 and2).
Fig. 1.Serum total insulin-like growth factor I (IGF-I;A), free IGF-I (B), insulin-like growth factor binding protein (IGFBP)-1 (immunoradiometric assay; C), and IGFBP-2 [Western ligand blotting (WLB); D] in response to 11 wk of training (baseline, week 4, week 11) in 2 groups of young men with different initial training status. Values are means ± SD. Filled bars, well-trained individuals (n = 12); open bars, untrained individuals (n = 7). † Significantly different from baseline value, P < 0.05. Fig. 2.IGFBP-3 (immunoradiometric assay; A), IGFBP-3 (WLB; B) and IGFBP-3 proteolysis (C) in response to 11 wk of training (baseline, week 4, week 11) in 2 groups of young men with different initial training status. Values are means ± SD. Filled bars, well-trained individuals (n = 12); open bars, untrained individuals (n = 7). † Significantly different from baseline value, P < 0.05. In response to the training period, untrained individuals significantly improved their V˙o2 max by 16 ± 4%, whereas V˙o2 max was unchanged in the well-trained subjects, which resulted in no significant group difference being evident at the end of week 11 (data not presented). The training intervention resulted in significant decreases in both total and free IGF-I in the untrained group from baseline toweek 4 [by 14 ± 6% (total IGF-I) and 27 ± 19% (free IGF-I)] and from baseline to week 11 [by 15 ± 9% (total IGF-I) and 23 ± 18% (free IGF-I)] (Fig. 1). In the well-trained group, total IGF-I decreased by 9 ± 11% (P < 0.05) and free IGF-I decreased by 20 ± 24% (P < 0.05) from baseline to week 4 but had returned to baseline levels at week 11. IGFBP-2 increased significantly in the untrained group from baseline to weeks 4 and 11, whereas no significant changes from baseline were observed in well-trained group (Fig. 1). Overall (untrained + well trained), changes in free IGF-I were found to be inversely correlated with changes in IGFBP-1 and IGFBP-2 (−0.51 <r < −0.70, P < 0.05). IGFBP-3 (IRMA) and IGFBP-3 (WLB) transiently decreased in the untrained group from baseline to week 4 (P < 0.05) and returned to baseline levels at week 11, whereas no significant changes occurred in the well-trained group (Fig. 2). Compared with baseline, IGFBP-3 proteolysis increased in the untrained group (by 33 ± 26%; P < 0.05) after 4 wk and remained above baseline (by 36 ± 24%; P< 0.05) at 11 wk, whereas IGFBP-3 proteolysis in the well-trained group remained unchanged from baseline to week 11(Fig. 2). Training caused IGFBP-4 to decrease over the 11-wk period in both the well-trained and untrained groups (Table 2). DISCUSSIONIn the present study, total IGF-I decreased in both untrained and well-trained subjects after 4 wk of training, which is in contrast to results from a training study on collegiate swimmers that reported an increase in total and free IGF-I levels after 2 mo of training (26). A decrease in total IGF-I is, however, in accordance with other previously published endurance-type training studies (3, 9, 10, 24) and has recently been confirmed in a study on US Army Rangers performing 8 wk of physical training combined with sleep deprivation and relative reduced food intake (15). Findings from these relatively few studies investigating the response of IGF-I to chronic training would be in accordance with the view that both intensity and duration determine the resultant IGF-I levels. One of the central findings in the present study is that the immonureactive IGFBP-3 levels decreased in untrained subjects from baseline to week 4 and returned to baseline in week 11, whereas no change was observed for well-trained individuals. Other training studies that used different training protocols found either unchanged (3, 10, 31) or increased IGFBP-3 levels (26, 27). To our knowledge, only one study has previously demonstrated that IGFBP-3 levels decreased with 5 wk of training; however, interestingly in that study, a decrease was also observed in the control group that did not perform any training (9). The transient IGFBP-3 decrease observed during prolonged training in the untrained group, but not in the well-trained group, indicates that either initial training status or the relative physiological stress applied during training affected the response. From the present data, it may be suggested that a threshold exists with regard to the relative physiological stress needed during training to result in proteolysis of IGFBP-3. Thus it may be speculated that the standardized training (baseline to week 11) was a sufficient stimulus to cause changes in IGFBP-3 in the untrained subjects, whereas prolonged periods of training with higher intensities would be needed to exceed the threshold in well-trained subjects. Unpublished results from our group indicate that if the intensity is sufficiently high, changes in IGFBP-3 and IGFBP-3 proteolysis do occur in both well-trained and untrained subjects; however, further studies are warranted to fully elucidate this hypothesis. Interestingly, a recent study conducted on soldiers during a 4-day sustained operation was able to confirm the present findings of a decrease in IGFBP-3 in response to intense physical loading (34). The reduction in IGFBP-3 from baseline to week 4 in the untrained subjects was indicative of IGFBP-3 proteolysis. This was confirmed by the IGFBP-3 proteolysis assay, and it supports the notion that IGFBP-3 proteolysis can be increased in response to intense exercise and training. To our knowledge, there are no other studies investigating IGFBP-3 proteolysis during chronic training, whereas two studies have investigated the proteolytic response to acute exercise. Dall et al. (7) found albumin-adjusted IGFBP-3 proteolysis in elite rowers to be unaffected by 4× 5-min submaximal rowing exercise (55–85% V˙o2 max) followed by an 6- to 7-min all-out test (7). In contrast, Schwarz et al. (37) found a 10-min high-intensity cycle exercise in nonathletes to increase IGFBP-3 proteolytic activity by 44%. Schwarz et al. did not adjust for shifts in plasma volume, determined by a hematocrit (increase from 44 to 50%), which theoretically would account for at least 25% of the increased concentration in plasma-soluble variables (7). However, the increase in IGFBP-3 proteolytic activity with exercise in the study by Schwarz et al. cannot solely be accounted for by changes in hemoconcentration, and a marked difference in initial training status of the individuals could partly explain the differences in results between the two studies (7, 37). Serum free IGF-I is markedly affected by changes in levels of IGFBP-1 and -2, which both correlate inversely with free IGF-I (17) (an inverse correlation that was also found in the present study). In addition, free IGF-I may be affected by changes in IGFBP-3 proteolysis, which is believed to represent a compensatory mechanism serving to increase free IGF-I by lowering IGFBP-3 ligand affinity and thereby modifying IGF-I bioavailability by releasing IGF-I from its 150-kDa complex, consisting of IGF-I, intact IGFBP-3, and an acid-labile subunit (36). It has been proposed that exercise-induced IGFBP-3 proteolysis would contribute significantly to the anabolic effects of exercise (36, 37); however, whereas there is little doubt that IGFBP-1 and -2 are potent regulators of free IGF-I levels (18), the physiological significance of IGFBP-3 proteolysis remains to be clarified. In the present study, the IGFBP-3 proteolysis observed in untrained subjects in week 4 and week 11 was not accompanied by any increase in free IGF-I. In contrast, free IGF-I was significantly decreased at all times (Fig. 1). Conditions other than physical training have been shown to be accompanied by a marked increase in IGFBP-3 proteolysis, for example, pregnancy. However, in pregnancy, although a high degree of proteolysis has occurred, still the concomitant increase in levels of free IGF-I is only about twofold (21). IGFBP-3 proteolysis is also observed in patients with chronic renal failure. However, in this condition, serum free IGF-I is markedly reduced, a finding that has been explained by the upregulated levels of IGFBP-1 and -2 (17). Thus it can be argued that changes in IGFBP-1 and -2 have relatively greater impact on levels of free IGF-I than IGFBP-3 proteolysis. This view is supported by the present study, where the upregulated levels of IGFBP-1 and -2 may explain the reduction in free IGF-I (7, 13, 19, 36), despite an increased IGFBP-3 proteolysis. Total and free IGF-II remained unchanged in both untrained and well-trained subjects during the 11-wk training period, which is in contrast to findings of increased total IGF-II in a 3-mo training study on young subjects with cystic fibrosis (20). This increase in total IGF-II in patients with cystic fibrosis is likely to be due to low baseline levels. In healthy elderly marathon runners, measurements of total and free IGF-II were not different from sedentary age-matched controls, indicating that there is no effect of prolonged training on IGF-II in elderly subjects (8). IGF-II may play an important role in stimulation of bone growth and development (25), but the physiological response of IGF-II to prolonged training is very sparsely reported and the biological importance has yet to be determined (37). After 4 wk of training, IGFBP-4 had decreased in both well-trained and untrained subjects and remained lower after 11 wk also. IGFBP-4 has consistently been shown to inhibit IGF-I action but has also been shown to be able to undergo proteolytic cleavage, which has been speculated to increase the action of IGF-I (4, 5). Previous studies have shown that IGFBP-4 is degraded only in the presence of exogenous IGFs, but Fowlkes et al. (14) have shown that, in vitro, IGFBP-4 can be degraded in the absence of IGFs, an effect that was almost entirely inhibited by the addition of IGFBP-3 to the medium. Fowlkes et al. therefore speculated that IGFBP-3 has the potential to regulate IGFBP-4 proteolysis in vivo. The interaction between IGFs and IGFBP-3 and -4 is undoubtedly very complex, and even though the probable role is to regulate the IGF interaction at the cell surface, the in vivo physiological role has not yet been elucidated. In a recent paper by van Doorn et al. (40), the clinical relevance of IGFBP-4 in various pathological conditions was examined and the conclusion was that IGFBP-4 might be of minor clinical importance in patients with hyperthyroidism, hypothyroidism, growth hormone deficiency, acromegaly, and chronic renal failure. However, during intense physical training that seems to reflect a systemic catabolic state, proteolysis of IGFBP-4 could have an important compensatory local effect. Therefore, the reduction in IGFBP-4, found in the present study, could have the physiological role of counteracting the concomitant decrease in free and total IGF-I that was seen after 4 wk of training and thereby reduce the catabolic-like early effect of training and maybe even increase a local anabolic effect. It could be argued that the training regimen used in this study would lead to a general state of tissue catabolism. However, interestingly, it was shown in the same subjects that both degradation and synthesis of collagen type I were increased around the Achilles tendon after 4 wk of training followed by a net synthesis after 11 wk of training (29). This evidently does not allow for any quantitative statements regarding whole body anabolism or catabolism, but at least it indicates that enzymatic degradation of connective tissue primarily is dominant in the early training phase, indicative of a more catabolic state than later during the training period. Such a phenomenon may have been more pronounced in untrained vs. trained subjects, and it could support the view of a two-phased response in the IGF system with regard to prolonged training (7). Finally, it cannot be excluded that local tissue IGF-I levels were increased despite reduced circulating levels, that paracrine or autocrine growth promotion exists, or that increased numbers of cell-surface IGF receptors (or altered sensitivity) could be responsible for more differentiated local anabolic responses (13, 32, 36). In conclusion, the results from the present study indicate that 11 wk of physical training affect serum levels of total and free IGF-I, IGFBP-2, and IGFBP-3 and affect IGFBP-3 proteolysis differently in untrained and well-trained subjects, with more extensive changes seen in the untrained subjects. Furthermore, the training-induced reduction in IGFBP-3 and concomitant increase in IGFBP-3 proteolysis in previously untrained individuals suggest the existence of a threshold for relative physiological stress to be exceeded in order for IGFBP-3 proteolysis to occur. We thank Annie Høj, Karen Mathiassen, and Kirsten Nyborg for excellent technical assistance. FOOTNOTESREFERENCES
Page 7recently a paper emphasized that to calculate mechanical efficiency during knee-extension exercise there is a need for estimating internal power (IP) generated to overcome inertial and gravitational forces related to the movement of the lower limb when external power (EP) during such exercise was delivered to an ergometer at different contraction rates (11). This is an important aspect that for various forms of locomotion has been debated for many years, and various biomechanical models have been developed for estimating IP (1, 8,18, 23, 27, 33, 35-37). The methods and criteria for calculation of IP are manyfold, resulting in highly diverse values of IP, and the taxonomy used regarding the terms external and internal work or power is inconsistent. It is generaly agreed that work done on a load external to the body (e.g., a cycle ergometer) corresponds to EP, and IP during such activity has been suggested to include all potential energy (Epot) and kinetic energy (Ekin) changes due to movement of all body segments (33, 37). However, during various forms of locomotion such as walking and running, the energy changes of the whole body center of gravity are considered as “external work” and only the kinetic energy changes of the body segments relative to the whole body center of gravity are considered as “internal work” (8, 23,36). These conflicting conventions are probably due to the difference in nature of these activities: during running and walking horizontally on a treadmill, no power in strictly mechanical sense is performed external to the body, when calculated as a mean across one or more strides. However, because the muscles do perform power at each push-off that results in a lift of the body center of gravity, this power output has been considered as EP. Of note is that such EP is also performed during exercise such as cycling and knee extension, but it has not been included in the EP that during such activities is measured exclusively as work done against a load external to the body. The division of the work into internal and external as calculated during horizontal running and walking has been criticized as being based on an “imaginary” resultant force acting on a fictitious point (39), and the legitimacy or sense of computing total work as the sum of internal and external work has been questioned (32, 39). Therefore, there is a need to reevaluate whether the computation of total power (TP) from IP and EP is justified from the physiological responses such as heart rate, muscle blood flow, oxygen uptake (V˙o2), and efficiency. With the use of a “novel model” for calculating IP, mechanical efficiency during knee extension was reported to be lower at 100 than at 60 contractions/min (cpm) (11). However, such finding is fully dependent on the model used for calculating IP, which may vary widely as emphasized above. Therefore, the aim of the present study was to 1) use an established kinematic model for estimation of IP during knee extension performed at several contraction rates,2) validate this kinematic model and compare it with the “novel model,” and 3) analyze the relationship between TP and physiological responses for evaluating the legitimacy of IP estimates. METHODSEight men volunteered to participate in this study. Subjects provided informed, written consent for the investigation, which was approved by the local Ethics Committee. All subjects were physically active, but none was specifically trained. Their age was 24 ± 0.8 (SE) yr, height was 1.82 ± 0.03 m, and weight was 82 ± 4 kg. Single-leg, dynamic knee-extension exercise (2) was performed in sitting position (backrest 30° inclined) with the subjects secured to the seat with straps around torso, hip, and thighs. The heel was connected to an aluminum brace (weight 0.7 kg) that via a rod was connected to the pedal arm (length 0.22 m) of a friction-braked modified cycle ergometer (Monark, Varberg, Sweden) for the determination of EP. Special precision weights were manufactured that allowed to adjust the resistance with 1-g accuracy, resulting in a precision of 0.1 W or better for all contraction rates, and care was taken that these weights were always hanging free and unsupported to result in the corresponding friction. With the subjects sitting on a horizontal surface and the lower leg vertical, the knee angle was 107 ± 1° (180° = fully extended knee). During each contraction, the knee was extended to 150 ± 2° (knee angle range of movement was 43 ± 2°). The subjects were instructed to contract only the knee-extensor muscles during the extension phase and to relax during flexion. All subjects participated in training sessions before the actual experiments to become fully familiarized with the exercise. Furthermore, they had their maximal knee-extension performance measured at 60 cpm, and the values obtained ranged from 60 to 80 W. On the day of experimental series for this study, the subjects performed single-leg dynamic knee-extension at 20 and 40 W of EP at contraction rates of 45, 60, 75, 90, and 105 cpm in randomized order. Each bout was performed until steady state was attained before all simultaneous recordings were made (see below: IP,V˙o2, Heart rate, and Leg blood flow), and each bout lasted 5–7 min in all. IP corresponding to the changes in Epot and Ekin of the exercising leg during the movements was quantified from a biomechanical model (37) based on anthropometric (19) and kinematic measurements. Kinematic data were obtained from two-dimensional (2D) video recordings (50 Hz) in the sagittal plane of the leg. Round reflective markers were attached to the skin or brace of the exercising leg over the top point of trochanter major, epicondylus lateralis on femur, malleolus lateralis on tibia, and caput ossa metatarsalia 5, and also on three points of the aluminum brace, allowing the identification of its center of mass. Five knee extensions were digitized by using Peak Motus 2000 (Peak Performance Technologies Englewood, CO). The coordinates were filtered by fourth-order Butterworth low-pass filter. A 2D position file containing the x and ycoordinates for each marker was used as input for the software program ERGILA (developed in Matlab from The MathWorks, South Natick, MA), implementing the model as suggested by Winter (37). For the present study a two-segment model was used consisting of 1) thigh and 2) lower leg including foot and brace. In short, the model assumes instantaneous energy transfer between segments as well as between Epotand Ekin: total energy (Etot) = Epot + Ekin. IP is calculated as the sum of the positive Etot changes per second as mean from five knee extensions. Only the summed positive energy changes are included, with the summed negative energy changes being disregarded. TP was calculated by summing EP as measured on the modified cycle ergometer and IP as calculated from the kinematic model. Muscle mechanical efficiency was calculated as (EP + IP) divided by the corresponding rate of energy expenditure. Energy expenditure was estimated from V˙o2 for each subject and exercise bout after subtraction of restingV˙o2, which was assumed to be 0.00333 l/s, and applying the oxygen energy equivalent (OE) in joules per liter of oxygen determined from the respiratory exchange ratios (RER) according to Coyle et al. (10). Pulmonary V˙o2 was measured by using a breath-by-breath gas analysis system (CPX/D Metabolic Cart, Medical Graphics). The gas analyzers were routinely calibrated against certified calibration gases of known concentrations, and the ventilation sensor was calibrated with a 3-liter syringe. After 3 min of knee-extension exercise, recordings were sampled over 15 s, and the mean of eight such successive values was taken to represent theV˙o2 for each exercise bout. RER was also calculated from these data. Heart rate (HR) was recorded continuously by using the Finapres (Ohmeda 2300). Leg blood flow (LBF) was recorded simultaneously withV˙o2 during steady state by using the ultrasound Doppler technique. LBF was measured as femoral artery blood flow, and the procedure used to obtain these measurements has previously been validated and shown to produce accurate absolute values at rest and during exercise (28). In our hands, 105 cpm was the highest contraction rate at which proper LBF measurements could be attained. The equipment used was an ultrasound Doppler (model CFM 800, Vingmed Sound, Horten, Norway) equipped with an annular phased array transducer (APAT, Vingmed Sound) probe (11.5-mm diameter), operating at an imaging frequency of 7.5 MHz and variable Doppler frequencies of 4.0–6.0 MHz (high-pulsed repetition mode 4–36 kHz). The site for vessel diameter determination and blood velocity measurements in the common femoral artery was always distal to the inguinal ligament but above the bifurcation into the superficial and profound femoral branch. The blood velocity measurements were performed with the probe in as low insonation angle as physically possible, always below 60° (15). The mean vessel diameter was calculated in relation to the duration of the blood pressure curve according to the following formula: diameter = (systolic diameter/3) + 2 (diastolic diameter/3) (28). The diameter measurements were obtained under perpendicular insonation. LBF was calculated by multiplying the cross-sectional area of the femoral artery by the angle-corrected, time- and space-averaged, and amplitude (signal intensity)-weighted mean blood velocity (in m/s) (28). For further details, see Osada and Rådegran (26). Validation of the kinematic model for calculation of IP was based on independent estimations of IP from the metabolic data only (IPmet) EP+IPmet=(V˙O2 exercise−V˙O2 rest)·OE·DEEquation 1 where EP is 20 and 40 W, respectively, “V˙o2 exercise” is measured directly (in l/s), “V˙o2 rest” is set to ∼0.00333 l/s, OE (in J/l) is determined from RER (10) (corresponding to ∼20 kJ/l as a mean), and DE is delta efficiency for each particular knee-extension rate. DE was calculated as the difference in EP between the 20- and 40-W bouts (= 20 W) divided by the corresponding difference in the rate of energy expenditure determined from V˙o2, by again accounting for the RER. The basic assumption for calculating IPmet is that DE reflects a generic, movement-specific DE that is independent of EP (i.e., represents net efficiency at 20 W as well as at 40 W when performed with the same contraction rate).ANOVA was used followed by Fisher's post hoc least significant diffference test. Linear regression and curve fitting (third-degree polynomial) were made in EXEL 2000 (Microsoft, WA) and GraphPad PRISM (version 2.0, San Diego, CA), respectively. A Bland-Altman analysis was used to evaluate IP calculated from the kinematic model against IPmet (5). Data are presented as means ± SE or (range), and P < 0.05 was considered statistically significant. RESULTSThe major movement was the knee angular displacement as expected, but also the thigh showed some angular displacement relative to horizontal, although the energy changes of the lower leg were by far larger than those of the thigh (Fig. 1). The changes in Epot occurred in synchrony with the movement of the lower leg, whereas the changes in Ekin occurred at twice that frequency. Of note is the observation that the amplitude (peak − nadir) for the changes in Ekin increased in a curvilinear mode with contraction frequency, whereas for the changes in Epot the amplitude remained rather constant (Fig.1). The timewise changes of these energies were partially out of phase; i.e., peak Epot corresponded to nadir Ekin, and this attenuated the summed positive Etot because of the assumption of instantaneous energy transfer in the model. IP increased significantly with increasing contraction rate (45, 60, 75, 90, and 105 cpm): 5 ± 0.2, 7 ± 0.2, 13 ± 0.5, 21 ± 0.7, and 36 ± 1.4 W at 20 W of EP, and 5 ± 0.4, 8 ± 0.4, 13 ± 0.4, 21 ± 0.6, and 32 ± 1.7 at 40 W, respectively (Fig.2). There was no difference in IP between the two different levels of EP, in accordance with all the angular displacements showing highly similar time histories during 20 and 40 W (Fig. 1). A third-order polynomial fit (y =ax + bx3), including the coordinate (0, 0), resulted in the following equation: IP = 0.0299x + 0.00002617x3(R2 = 0.996), where IP is the mean value for 20 and 40 W, and x is contraction rate. Fig. 1.Sample plot from a representative subject of the changes in knee angle and thigh angle relative to horizontal in combination with energy changes estimated from a kinematic model during knee-extension exercise. Results are from exercise bouts at 20 W (thin lines) and 40 W (thick lines) of external power (EP) performed at the 5 different contraction rates (2-s periods shown). Estimated contributions of the sum of the positive total energy changes from thigh (ΔEtot thigh) and lower leg (ΔEtotlower leg) are presented separately, and also the potential (ΔEpot) and kinetic (ΔEkin) energy changes are shown separately. Finally, the instantaneous sum of all energy changes (ΔEtot) is shown. deg, Degrees; cpm, contractions/min. Fig. 2.Internal power (IP) vs. contraction rate. ●, Data from the present study shown as means ± SE. Symbol size exceeds ±1 SE for the low contraction rates. Solid line is drawn according to the following equation: IP = 0.0299x + 0.00002617x3(R2 = 0.996), where x is contraction rate. □, Data from Ferguson et al. (11). Dashed line is the best fit to those data including (0,0). When the subject's individual anthropometric differences are taken into account, IP tended to increase with body mass at the highest contraction rates (Fig. 3). However,R2 values were ≤0.5 and nonsignificant. Thus the effect of body mass (within 65–100 kg) on IP could be disregarded by taking into account the interindividual variation relative to the difference in IP between contraction rates, a finding that is in concert with that of Ferguson et al. (11). Therefore, independent of the subject's body mass, the IP during knee extension may be calculated from the formula for the polynomial fit above. Fig. 3.IP vs. body mass presented as individual data for the various contraction frequencies. ○, 45 cpm; ●, 60 cpm; ■, 75 cpm; ▴, 90 cpm; ⧫, 105 cpm. IP is for each subject taken as the mean value of 20 and 40 W. LBF, pulmonary V˙o2, and HR increased when constant EP was performed with increasing contraction rate and were consistently higher at 40 W compared with 20 W of EP (Fig.4). DE calculated from these data ranged from 16 to 23% between contraction rates but showed no significant differences, with the overall mean for DE being 20 ± 1%. The RER values used in these calculations for estimation of OE ranged from 0.85 to 1.01 for the different exercise bouts. The subsequent calculation of IPmet on the basis of Eq. 1 (as mean of 20 and 40 W) resulted in the following values: 5 ± 3.3, 7 ± 1.5, 10 ± 2.3, 19 ± 4.7, and 28 ± 6.6 W at the contraction rates of 45, 60, 75, 90, and 105 cpm, respectively. The above IP values calculated from the kinematic model were validated against these IPmet values by using the Bland-Altman analysis. Inclusion of all individual data sets (n = 74, because of 6 missing data sets) in one analysis showed that the estimated bias, calculated as the overall mean difference between IPmet and IP based on kinematics, was 2 W and that 93% of the differences between IPmet and IP based on kinematics were within ±2 SD of all these differences (Fig. 5). When the same analysis was used for each of the 10 exercise bouts (n = 8 or 7 in each bout), the overall difference ranged from 0 to 8 W, and corresponding differences within ± 2 SD of all differences ranged from 86 to 100%. Also, Bland-Altman analysis performed for each of the eight subjects (n = 10 exercise bouts for each subject) showed that 88–100% of the differences between IPmet and IP were within ± 2 SD of all these differences. The slopes of the regression lines for each of these plots ranged from −0.94 to 0.71 with a mean of 0.04. Thus assessment of the slope for the Bland-Altman analyses did not indicate a systematic relationship between the differences between IPmet and IP (on the basis of kinematics) vs. the average of these two variables. Fig. 4.Values (means ± SE) of leg blood flow (A), pulmonary oxygen uptake (V˙o2) (B), and heart rate (C) during exercise performed at 5 different contraction rates. Fig. 5.Bland-Altman plot of the difference between IP calculated from metabolic variables (IPmet) and IP calculated from the kinematic model (IPkinematic) vs. the average of these 2 values. The overall mean value of differences was 2 W and is depicted as the solid horizontal line together with the 2 dashed lines at difference ± 2 SD. When the physiological variables LBF,V˙o2, and HR were plotted vs. TP (= IP + EP), where IP was calculated from the kinematic data, they fell on the same lines independent of contraction rate (Fig.6). This was supported by the three linear regression lines calculated for the respective mean values (allR2 ≥ 0.92 and significant). We also calculated regression lines and R2 values for each subject for each of the three variables. TheR2 values ranged from 0.52 to 0.99, except for one subject's LBF data, where a large variability was seen and theR2 value was only 0.05. The correlations were reanalyzed excluding this subject, but this did not significantly affect our results and we decided to retain the subject in all data sets. The largest variability was seen for the LBF data at the highest contraction frequencies of 105 cpm, and these data tend to be on the high side of the regression line but did not deviate significantly. Fig. 6.Leg blood flow (A), pulmonaryV˙o2 (B), and heart rate (C) vs. total power output estimated as EP + IP, where IP is calculated according to a kinematic model. Values are means ± SE. Muscle mechanical efficiency at the five contraction rates was 22 ± 1.6, 23 ± 1.5, 24 ± 1.8, 23 ± 1.4, and 20 ± 1.3% at 20 W of EP, and 20 ± 1.1, 23 ± 1.0, 22 ± 0.9, 21 ± 1.3, and 18 ± 1.3% at 40 W of EP, respectively. The overall muscle mechanical efficiency was 22 ± 0.5% with no significant differences between working conditions. DISCUSSIONThe use of an established kinematic model for the calculation of IP during knee-extension exercise showed a curvilinear relationship between IP and contraction rate. This was expected because the Ekin of a movement increases with the velocity squared, which in turn corresponds to the contraction rate squared (21), and, therefore, the timewise Ekinchanges during the movements should increase with a third-order polynomial in relation to contraction rate (13,24). Because Epot per contraction is independent of the movement velocity it follows from a corresponding argument that the timewise Epot changes should increase proportionally with contraction rate, leading to the third-order polynomial fit applied for the present data: IP =ax + bx3, where x is the contraction rate. Surprisingly, some movement of the thigh occurred, which added a small but significant amount to the total IP during knee-extension exercise. This may be due to soft tissue compression of the thigh, e.g., during lifting of the lower leg against a resistance such that the posterior thigh will be compressed against the horizontal surface. In all, IP amounted to between 12 and 180% of the EP depending on contraction rate and magnitude of EP. The model used in the present study assumes instantaneous exchange of energy between all energy components (potential as well as kinetic translational and rotational) within each body segment and also between body segments (37). Only the summed positive energy changes are the basis for the power calculations here, and thus the minimum IP is reperesented, because energy changes in opposite directions will attenuate the sum. Storage of elastic energy during an eccentric work phase and release during a subsequent concentric work phase may further reduce the muscular work performed to overcome the TP but could be ignored during knee-extension exercise, because the knee extensors are almost completely relaxed in the eccentric phase and therefore unlikely to store elastic energy (2). The present kinematic model is based on theoretical considerations and has proven empirically to be valid in a number of studies with widely different movement patterns (20, 25, 33). In the present study, the validity of this model was further evaluated from metabolic data. The estimated bias of IP based on the kinematic model was only 2 W compared with IPmet, and the lack of agreement seemed not to depend on the magnitude of IP. The novel use of DE for estimating IP from metabolic variables is a method fundamentally different from kinematic model calculations, whether these are based on video recordings as in this study or on goniometer recordings as in a previous study (11). Thus the metabolic based calculation of IPmet may be used as “a gold standard” for validation of the kinematic estimation of IP. Previous studies have introduced “metabolic counterparts of internal power but not combined this with independent estimates of the mechanical power (13, 22). DE was taken as a movement specific efficiency being independent of EP and was calculated from aerobic metabolism only. This is justified from the experimental conditions in which healthy young men exercised at EP up to 40 W, which previously was shown to elicit only minor lactate changes (31). Furthermore, in the present study EP only up to 70% of maximum working capacity was performed and all the RER values were <1, which was in support of a prevailing aerobic metabolism. Importantly, for the conversion of V˙o2 to total energy turnover rate in watts, the RER value was used for estimation of the oxygen energy equivalent (10). The present estimates of IPmet from metabolic data were highly similar to IP estimates from kinematic data during knee extension, which is in support of the kinematic model assumptions used in this study. However, our data on IP were not in concert with those in a previous study because our data were generally lower than those reported by Ferguson et al. (11). This may in part be due to differences in the setup because the knee angular displacement in each contraction in the present study was ∼45°, but it was reported to be 80° in the study by Ferguson et al., although Ferguson et al. ignored movement of the thigh, which was shown to be significant in our setup and would have predicted our values to become higher (11). Surprisingly, Ferguson et al. reported IP to be lower at higher workloads than at lower workloads during high contraction rates, but not during low contraction rates, without giving any explanation for this finding (11). Of note is that in our calculation of IP no difference was found when the same contraction rate was performed with different external loads. This was expected because in these two conditions the same movement of the leg is performed in terms of timewise displacement of the lower leg (Fig. 1). Thus velocities and accelerations are similar, which in turn implies that the same muscular work has to be performed to move the leg. However, our finding of nonlinearity between IP and contraction rate is the most essential difference compared with the data by Ferguson et al. This relationship is of crucial importance in the comparison of muscle mechanical efficiency at different contractions rates. If IP increases proportional with contraction rate, as implicated by the data of Ferguson et al. (Fig. 1), then this may be the reason for their finding of lower muscle mechanical efficiency at high compared with low contraction rates. Importantly, the present study design with several different contraction rates allowed us to describe the dependency of IP on contraction rate. In the present study, no differences in muscle mechanical efficiency were found between contraction rates, neither when estimated as DE nor when based on the kinematic model. These two fully independent calculations of efficiency resulted in remarkably similar values, and in particular DE, or apparent efficiency, has previously been suggested as the most valid measure of muscular efficiency (34). Thus our findings are considered to be convincing evidence for the independency of muscle mechanical efficiency relative to contraction rate during the present contraction velocities. Studies on single fibers have demonstrated efficiency to depend on contraction velocity despite a wide spread in most data (16, 38). Delicate data on single human muscle fibers show fiber type-specific efficiencies relative to contraction velocity (6). The curves drawn to fit the data showed for type I fibers peak efficiency to occur at a velocity of ∼0.2 fiber lengths/s and for type IIA fibers at ∼0.3 fiber length/s at 20°C. On the basis of these single-fiber data, it is often implied that efficiency changes with contraction rate in voluntary movements also; however, scrutinizing the literature reveals contradictory findings. In the early studies of Asmussen (4), DE during bicycling seemed to be greater at higher contraction rates (68 vs. 102 cpm). This was confirmed in a number of later studies (7, 9, 22, 30), whereas another study was in support of a decrease in DE with increasing pedal rate (40–100 rpm) (14), and still another study presented data fitting an U-shaped curve (13). Finally, some studies do discuss efficiency to be independent of contractions velocity within a wide range of velocities (6, 29). Our knowledge on muscle mechanical efficiency during different voluntary contraction velocities is still limited, and changes in fiber-type recruitment occurring with changes in contraction intensity or velocity during human voluntary contractions may mask possible relationships between efficiency and contraction rate. This is because the fiber types show maximum efficiency at different contractions velocities, as pointed out above and discussed previously (40). Therefore, functionally constant efficiencies may exist during a wide range of voluntary activity. Of note is that the interpretation of data during varying contraction rates may not be comparable when differences in angular displacements or velocities are not accounted for. The contradictions regarding IP and mechanical efficiency during knee extension in the present study and the earlier study by Ferguson and co-workers (11) emphasize that the model assumptions for estimating IP are crucial. Ferguson and co-workers (11) stated that their model allowed IP to be determined accurately on the basis of test-retest reliability analysis, but they did not validate their model. Direct measurement of IP is not possible, and mechanistic discussions on the topic have come to a “dead end” (32). Nonetheless, the inclusion of IP is justified if, from human voluntary exercise, we can gain information on the muscular level. This was essentially the aim when the knee-extension model was introduced (2), and therefore it is equally essential to adopt a model for IP calculation that will not distort data on TP in a physiologically nonsensical way. This brings us back to the legitimacy or justification for estimating IP, which is that it contributes to a generic understanding of physiological responses to muscle contractile work. For this reason, we studied not only HR and V˙o2 but also blood flow of the exercising muscle group during the performance of various magnitudes of EP in combination with different magnitudes of IP due to different contraction rates. We measured LBF, which previously has been shown to reflect the increases in blood flow of the contracting knee-extensor muscles (3). Interestingly, LBF increased not only with increasing EP but also with increasing IP (Fig.4), and when plotted against TP (= EP + IP), LBF fell on the same line independent of the relative contribution of EP and IP (Fig.6), when IP was calculated based on kinematic model assumptions (37). This is at variance with the most recent paper by Ferguson at al. (12) in which they applied their previously published model for estimation of IP and showed LBF to be higher when the same TP (54 W) was performed at 100 cpm compared with 60 cpm. This may be exclusively due to their model for IP calculations, but also the experimental setup may play a role. Thus knee extension performed with an angular range of only 20° showed LBF to be similar when the same EP (<15 W) was performed at 60 and 80 cpm (17). This finding implies LBF to be relatively higher at 60 cpm compared with 80 cpm had the same TP been performed, which is in contradiction to Ferguson et al. (12). Interestingly, in our study, LBF was proportional to TP independent of contraction rate, and our experimental setup comprised both angular exertions and EP values between those applied in the studies of Ferguson et al. (12) and Hoelting et al. (17), respectively. In combination, these three studies suggest a U-shaped curve relationship between LBF and constant TP performed with increasing shortening velocities. Comprehensive studies covering a wide range of muscle contraction ranges and velocities as well as EPs are necessary to reveal the muscle blood flow dependency on contraction intensity and contraction rate to bridge the controversies and increase our generic understanding of physiological responses to muscular contractions, including validated models for estimation of TP. In conclusion, a novel metabolic validation of an established kinematic model for calculating IP supports the model assumptions for knee-extension exercise. The present polynomial fit (IP = 0.0299x + 0.00002617x3, wherex is contraction rate) accurately estimates IP for subjects having a body mass within 65–100 kg when exercising with knee angle displacements of ∼45° from a vertical lower leg position. When this model was applied, the physiological responses LBF,V˙o2, and HR were closely related to TP, supporting the legitimacy of IP estimates. Muscular mechanical efficiency (20 ± 1%) as well as DE (22 ± 0.5%) remained remarkably constant across contraction rates. A range of knee angular displacements and velocities as well as EPs have to be addressed in future studies for establishing a comprehensive data set for IP estimates to be used in the variety of knee-extension studies in which the knowledge of TP is pertinent. This study was financially supported by a grant from the Danish Research Council for Sports (to G. Sjøgaard). FOOTNOTESREFERENCES
Page 8results from several studies indicate that endothelial function in conduit arteries declines with age in humans and animals (3, 4, 7, 8, 10, 11, 16,23). The endothelial dysfunction induced by aging is characterized by blunted vasodilator responses of conduit arteries to select endothelium-dependent agonists (3, 4, 7, 8, 10, 11, 16,23). The mechanism(s) for the detrimental effects of age on endothelium-dependent dilation is not fully understood; however, age-associated decrements in the ability of endothelial cells to produce and/or release nitric oxide (NO) may contribute to the dysfunction (3, 11, 16). This speculation is supported by experimental evidence indicating that vasodilation in response to ACh and bradykinin is impaired in aorta from senescent subjects, whereas dilation to sodium nitroprusside (SNP) is not compromised (3, 5, 16). An age-associated decline in the expression of endothelial NO synthase (eNOS), decreasing local production of NO, is one mechanism that may contribute to impaired endothelium-mediated dilation in conduit arteries of senescent animals. Indeed, age-related reductions eNOS mRNA expression have been reported in aorta of senescent rats (3,6). Alternatively, decreased expression of Cu/Zn-dependent superoxide dismutase (SOD-1) may contribute to impaired endothelium-mediated dilation by compromising the ability to scavenge superoxide anion (O2−·), decreasing the biological half-life of NO. Although endothelial dysfunction is well documented in aorta of senescent rats, the effect of age on endothelium-dependent dilation in arteries that perfuse skeletal muscle is not known, and the effect of age on eNOS and SOD-1 expression has not been studied in feed arteries. Consequently, the purpose of this study was to test the hypothesis that aging decreases endothelium-dependent vasodilator responses in skeletal muscle feed arteries. In addition, we tested the hypothesis that attenuated vasodilator responses are associated with decreased eNOS and SOD-1 expression. Feed arteries from the soleus muscle (SFA) and gastrocnemius muscle (GFA) were studied because of their importance in regulating skeletal muscle blood flow during exercise (25). In addition, previous studies indicate the ability to increase blood flow to the soleus and gastrocnemius muscles during exercise is impaired in senescent rats (13). METHODSBefore initiation of this study, approval was received from the Institutional Animal Care and Use Committee at the University of Missouri. Male Fischer 344 rats (age 4 and 24 mo; n = 9/age group) were purchased from a commercial dealer (Harlan Sprague Dawley, Indianapolis, IN) and housed in the College of Veterinary Medicine's Animal Care Facility. The facility was maintained at 24°C with a 12:12-h light-dark cycle. Animals were provided food and water ad libitum, and the rats were examined daily by the investigators and by veterinarians affiliated with the University of Missouri's College of Veterinary Medicine. Rats were anesthetized with an intraperitoneal injection of pentobarbital sodium (50 mg/kg body wt ip). The soleus and gastrocnemius muscles were removed and placed in MOPS-buffered physiological saline solution (PSS) containing (in mM) 145.0 NaCl, 4.7 KCl, 2.0 CaCl2, 1.17 MgSO4, 1.2 NaH2PO4, 5.0 glucose, 2.0 pyruvate, 0.02 EDTA, and 25.0 MOPS, at pH 7.4. Feed arteries from the right leg were dissected and transferred to a Lucite chamber containing MOP-PSS (4°C) for cannulation. Feed arteries from the left leg were placed in ribonuclease-free microcentrifuge tubes and frozen at −80°C for analysis of eNOS and SOD-1 protein and mRNA expression. One end of each feed artery was cannulated with a glass micropipette and secured with 11-0 surgical silk. The opposite end of the feed artery was cannulated with a resistance-matched pipette and secured with 11-0 silk. The micropipettes were attached to separate reservoirs filled with MOPS-PSS supplemented with albumin (1 g/100 ml). The height of each reservoir was initially adjusted to set intraluminal pressure in each feed artery to 60 cmH2O (1 mmHg = 1.36 cmH2O) for 20 min. After 20 min, intraluminal pressure was raised to 90 cmH2O, and the feed arteries were allowed to equilibrate for an additional 40 min at 37°C. At the end of the equilibration period, arteries that did not develop at least 25% spontaneous tone were constricted with phenylephrine. All experimental protocols were subsequently conducted at an intraluminal pressure of 90 cmH2O to approximate in vivo intraluminal pressure (25). Endothelium-dependent dilation was assessed in feed arteries by adding increasing doses of ACh to the bath solution in cumulative doses over the range of 10−9 to 10−4 M in whole-log increments. Endothelium-independent dilation was assessed in feed arteries by adding increasing doses of adenosine (Ado) or SNP to the bath solution in cumulative doses over the range of 10−9 to 10−4 M in whole-log increments. SNP was utilized to determine whether aging alters the ability of vascular smooth muscle to respond to NO. At the end of each experiment, the PSS bath solution was replaced with Ca2+-free PSS. Feed arteries were incubated for 30 min to determine passive diameter at an intraluminal pressure of 90 cmH2O. All reagents used in dose-response experiments were obtained from Sigma Chemical (St. Louis, MO). Reagents were prepared on the day of the experiment. Relative differences in eNOS and SOD-1 mRNA expression in single feed arteries were assessed as described previously (26, 28). Briefly, single feed arteries were homogenized in 50 μl LiCl lysis buffer by vortexing the sample vigorously for 60 s. The sample was subsequently spun briefly in a microcentrifuge, and the process was repeated four to five times until the artery was completely digested. Poly(A)+ RNA was isolated from the crude lysate with paramagnetic oligo(dT) beads (Dynal), and first-strand cDNA synthesis was performed in a 20-μl volume (Superscript Preamplification System, GIBCO-BRL Life Technologies). Nine microliters of cDNA were used in a PCR using previously published primers and cycling conditions for eNOS or SOD-1 (26, 28). The PCR consisted of 25 cycles to ensure that the reaction was within the linear range for the eNOS and SOD-1 primer sets. All data were standardized by coamplifying eNOS or SOD-1 with glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and calculating an eNOS-to-GAPDH or SOD-1-to-GAPDH ratio for each feed artery (26). The sequence for the GAPDH primers has been published previously (28). Differences in eNOS and SOD-1 protein expression were assessed in single feed arteries (matched for diameter and length) by using immunoblot analysis as described previously in detail (15,28). eNOS protein expression was evaluated with a monoclonal antibody (1:1,600; catalog no. N30020, Transduction Laboratories). SOD-1 protein content was assessed with a polyclonal antibody (1:1,600; catalog no. SOD-100, Stressgen). Immunoblots were evaluated by densitometry by using NIH Image software (National Institutes of Health, Bethesda, MD), and data were expressed as densitometric units. GAPDH was used as an internal standard to control for small differences in protein loading. GAPDH protein content was assessed with a monoclonal antibody (1:10,000, catalog no. MAB374, Chemicon). All values are means ± SE. Between-group differences in body mass, passive diameter, as well as eNOS and SOD-1 mRNA and protein expression were assessed by using Student's t-tests for unpaired observations. Concentration-response curves were expressed as relative diameter and analyzed by two-way ANOVA with repeated measures on one factor (dose). Relative diameter was calculated asDdose/DP, whereDdose is measured diameter for a given dose (ACh, Ado, or SNP) and DP is maximal passive diameter. When a significant F value was obtained, post hoc analyses were performed with Duncan's multiple-range test. When more than one artery was studied from a single animal, concentration-response data were averaged; therefore, one animal counted as one observation. Statistical significance was set at theP ≤ 0.05 probability level. RESULTSCharacteristics of the young and old rats are shown in Table1. The body weight of old rats was significantly greater than young rats. Maximal passive diameter was greater in GFA than in SFA; however, there was not a significant main effect of age in SFA or GFA.
Before dose-response curves were initiated, percent tone in SFA and GFA was similar in young and old rats. Repeated-measures ANOVA indicated that ACh elicited a concentration-dependent dilation of feed arteries (Fig. 1) and a significant age × dose interaction (P = 0.004). Post hoc analyses revealed that endothelium-dependent dilation to high doses of ACh was significantly less in SFA from old rats than in SFA from young rats (Fig. 1A). In contrast, ACh-induced dilation was not impaired in GFA from old rats (Fig. 1B). Fig. 1.Dose-response relationship of soleus feed arteries (SFA;A) and gastrocnemius feed arteries (GFA; B) to ACh. B, baseline diameter before the first dose of ACh. Values are means ± SE; n = 9 rats per group. Beginning diameters were 88.4 ± 10.7 μm (young SFA), 109.0 ± 11.6 μm (old SFA), 159.5 ± 14.2 μm (young GFA), and 157.9 ± 10.1 μm (old GFA). Repeated-measures ANOVA indicated that dilation to ACh was reduced by aging in SFA but not in GFA. * Significantly different from young feed arteries, P < 0.05. Ado elicited a modest, but significant, dose-dependent dilation of SFA from young (18%) and old (12%) rats (Fig.2). Similarly, Ado elicited a dose-dependent dilation of GFA. There was not a significant main effect of age on Ado-induced dilation in SFA (P = 0.83) or GFA (P = 0.52) (Fig. 2, A and B). Fig. 2.Dose-response relationship of SFA (A) and GFA (B) to adenosine (Ado). B, baseline diameter before the first dose of Ado. Values are means ± SE; n = 9 rats per group. Beginning diameters were 96.3 ± 10.3 μm (young SFA), 120.6 ± 9.0 μm (old SFA), 173.8 ± 10.5 μm (young GFA), and 163.6 ± 7.9 μm (old GFA). Ado elicited a significant dose-dependent dilation of SFA and GFA; however, there was not a significant main effect of age in SFA or GFA. SNP elicited a concentration-dependent dilation of SFA and GFA (Fig.3). There was not a significant main effect of age on SNP-induced dilation in SFA (P = 0.15) or GFA (P = 0.42) (Fig. 3, A andB). Fig. 3.Dose-response relationship of SFA (A) and GFA (B) to sodium nitroprusside (SNP). B, baseline diameter before first dose of SNP. Values are means ± SE;n = 9 rats per group. Beginning diameters were 92.5 ± 12.1 μm (young SFA), 120.5 ± 9.6 μm (old SFA), 188.8 ± 11.1 μm (young GFA), and 176.3 ± 9.7 μm (old GFA). SNP elicited a dose-dependent dilation of SFA and GFA; however, there was not a significant main effect of age in SFA. The effect of aging on eNOS protein expression in feed arteries from the soleus and gastrocnemius muscle is shown in Fig.4. Immunoblot analysis revealed that eNOS protein expression declined with age in SFA (−71%) but not in GFA (Fig. 4, A and B). Semi-quantitative PCR revealed that eNOS mRNA expression was not altered by age in SFA or GFA (Fig.5). Fig. 4.Comparison of endothelial nitric oxide synthase (eNOS) protein expression in feed arteries from young (4 mo; lanes 1–4) and old (24 mo; lanes 5–8) rats. A: eNOS protein content in SFA. Inset: sample immunoblots for eNOS and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) in SFA. B: eNOS protein content in GFA. Inset: sample immunoblots for eNOS and GAPDH in GFA. Values are means ± SE; n = 9 rats per group. * Significantly different from young feed arteries, P < 0.05. Fig. 5.Comparison of eNOS mRNA expression (e-NOS-to-GAPDH ratio) in feed arteries from young (4 mo) and old (24 mo) rats.A: eNOS mRNA content in SFA. B: eNOS mRNA content in GFA. Values are means ± SE; n = 9 rats per group. Statistical analysis did not reveal a significant effect of age on eNOS mRNA expression in SFA or GFA. Immunoblot analysis revealed that SOD-1 protein expression declined with age in SFA (−54%) but not in GFA (Fig.6). SOD-1 mRNA expression was not altered by age in SFA or GFA (Fig. 7). Fig. 6.Comparison of superoxide dismutase-1 (SOD-1) protein expression in feed arteries from young (4 mo; lanes 1–4) and old (24 mo; lanes 5–8) rats. A: SOD-1 protein content in SFA. Inset: sample immunoblots for SOD-1 in SFA. B: SOD-1 protein content in GFA.Inset: sample immunoblots for SOD-1 in SFA. Values are means ± SE; n = 9 rats per group. * Significantly different from young feed arteries,P < 0.05. Fig. 7.Comparison of SOD-1 mRNA expression (SOD-1-to-GAPDH ratio) in feed arteries from young (4 mo) and old (24 mo) rats. A: SOD-1 mRNA content in SFA. B: SOD-1 mRNA content in GFA. Values are means ± SE; n = 9 rats per group. Statistical analysis did not reveal a significant effect of age on SOD-1 mRNA expression in SFA or GFA. DISCUSSIONThe purpose of this study was to test the hypothesis that aging decreases endothelium-dependent vasodilation in skeletal muscle feed arteries and that the attenuated vasodilator responses are associated with decreased expression of eNOS and/or SOD-1. The primary findings of this study were as follows. 1) Endothelium-dependent dilation to ACh was blunted in SFA but not in GFA. 2) Dilation to Ado was not compromised by age in SFA or GFA. 3) Endothelium-independent dilation to SNP was not significantly different in feed arteries from young and old rats. 4) eNOS and SOD-1 protein expression decreased with age in SFA but not in GFA. 5) eNOS and SOD-1 mRNA expression were not altered by age in SFA or GFA. These results suggest that age induces muscle-specific impairment of endothelium-dependent dilation in SFA. In the present study, vasodilator responses to ACh were blunted in SFA from old rats (Fig. 1). These data are in accord with previous studies indicating that ACh-induced dilation is impaired in aorta from senescent rats (3, 5, 16) and demonstrate that age-associated decrements in endothelial function can occur in skeletal muscle feed arteries. Given that ACh-induced dilation is mediated largely by endothelium-derived NO (14), blunted vasodilator responses to this agonist suggests that the ability of SFA to produce NO is impaired by aging. In the present study, Ado elicited a modest, but significant, vasodilator response in SFA (Fig. 2). Importantly, the dilator response to Ado was similar in young and old rats. In rats, Ado is a dilator that has been reported to mediate its effects primarily by acting directly on vascular smooth muscle (1). The finding that dilation to Ado was not different in old rats, whereas dilation to the endothelium-dependent agonist (ACh) was impaired, is consistent with our interpretation that aging selectively impairs NO-mediated vasodilator mechanisms. Importantly, dilation to SNP (a NO donor) was not impaired in SFA from senescent rats. Indeed, the trend was toward enhanced vasodilator responses to exogenous NO in SFA (P = 0.15). Enhanced smooth muscle responses to NO have been reported previously in aortic rings from senescent rats (16) and may represent smooth muscle adaptation to partially compensate for attenuated release of endothelium-derived NO in SFA from senescent rats. In this study, eNOS protein expression decreased with age in SFA (Fig.4). These data suggest that decreased eNOS protein expression may contribute to blunted ACh-induced dilation by impairing the ability to produce endothelium-derived NO. Further studies are needed, however, to determine whether age-related reductions in eNOS expression are associated with decrements in NO production in these arteries. The mechanism(s) by which aging decreases eNOS protein expression is not known. One possibility is that age-associated reductions in resting blood flow to the soleus muscle may lead to decreased shear stress on the endothelial lining of SFA. It is well documented that shear stress regulates eNOS expression in vascular endothelial cells (19-22, 24, 27). In addition, previous studies have demonstrated that chronic reductions in soleus muscle blood flow induced by hindlimb unweighting are associated with reductions in eNOS expression and ACh-induced dilation in SFA (15, 18). Although previous studies to determine the effect of aging on skeletal muscle blood flow did not reveal age-related decrements in resting blood flow to the rat hindlimb (9, 13), the effect of aging on intraluminal shear stress within SFA is not known. Decreased eNOS protein expression may also be mediated by age-associated reductions in physical activity. During exercise, blood flow increases to provide muscle with an adequate supply of oxygen and nutrients. It is well documented that flow/shear stress is an important signal regulating eNOS expression in blood vessels (19,27). Consequently, age-associated reductions in physical activity and associated increases in blood flow may remove an important signal for the maintenance of eNOS expression. In addition, decreased physical activity may reduce the frequency and intensity of soleus muscle contraction. Awolesi and associates (2) reported previously that eNOS expression can be regulated by cyclic strain in endothelial cells; therefore, it is conceivable that decreased muscular activity associated with advancing age may contribute to decreased eNOS expression in SFA. Interestingly, eNOS mRNA expression was similar in SFA from young and old rats (Fig. 5). These data are contrary to our hypothesis and suggest the possibility that age-associated reductions in eNOS protein expression may involve a decreased rate of protein synthesis and/or increased rate of protein degradation rather than changes in eNOS gene expression. Alternatively, decreased eNOS mRNA expression may have occurred at an earlier age, initiating the changes in eNOS protein expression observed in the 24-mo-old rats. An additional finding of this study was that aging was associated with decreased SOD-1 protein expression in SFA (Fig. 6). Age-related reductions in SOD-1 protein expression may contribute to impaired ACh-induced dilation of SFA by impairing the ability to scavenge O2−·, increasing free radical-mediated degradation of NO. The mechanism for decreased SOD-1 protein expression may involve age-associated reductions in resting blood flow and associated shear stress. Similar to eNOS, shear stress is a primary signal regulating SOD-1 expression in cultured endothelial cells (12) and in intact arteries (27). In addition, decreased soleus muscle blood flow, induced by hindlimb unweighting, is associated with reductions in SOD-1 expression and ACh-induced dilation in SFA (28). Interestingly, age-associated reductions in ACh-induced dilation did not occur in GFA (Fig. 1). Nor were there age-associated reductions in eNOS or SOD-1 protein expression. The mechanism by which aging induces muscle-specific impairment of endothelium-dependent dilation is not known but may involve differences in muscle fiber type and/or fiber recruitment patterns. The soleus muscle is a slow oxidative skeletal muscle that is recruited at rest and has relatively high resting blood flow (17). By comparison, the gastrocnemius muscle is composed of a mixture of slow oxidative, fast oxidative glycolytic, and fast glycolytic fibers (17). Consequently, this muscle is less active under resting conditions and has relatively lower blood flow under resting conditions (17). It is conceivable, therefore, that age-associated reductions in physical activity selectively reduce blood flow/shear stress or the frequency and intensity of contraction in the soleus muscle. In summary, the results of this study indicate that aging induces impairment of endothelium-dependent dilation in SFA but not in GFA. The age-induced change in endothelial function in SFA is characterized by attenuated vasodilator responses to ACh but not to Ado. In addition, the ability of vascular smooth muscle to respond to SNP was not impaired by aging. Collectively, these findings suggest endothelium-dependent vasodilator mechanisms are impaired in SFA from senescent rats. Age-associated reductions in eNOS protein expression in SFA may contribute to blunted endothelium-dependent dilation by decreasing NO production. Decreased SOD-1 protein expression may contribute to attenuated NO-mediated dilation by decreasing the ability to scavenge O2−·, reducing the half-life of NO. The authors gratefully acknowledge the expert technical assistance of Pam Thorne and Tammy Strawn. FOOTNOTESREFERENCES
Page 9although exercise is considered a physiological stimulus for cell release by the bone marrow (5), surprisingly few data are available on circulating hematopoietic precursors in athletes. Erythrocyte production was studied relative to athlete's anemia (25) and to assess the effects of intermittent hypoxic exposure on exercise performance (1). Conversely, little is known of the effects of exercise on myeloid precursors. Over 20 years ago, it was reported that colony-forming cells in peripheral blood increased after a short and intense exercise bout in normal subjects (2), but a detailed characterization of hematopoietic precursors in well-trained subjects was never obtained. The rationale to study myeloid precursors in athletes is that intense and prolonged exercise increases white blood cell (WBC) and neutrophil [polymorphonuclear neutrophil (PMN)] counts (5, 16), partly through mobilization of marginated PMNs (5) associated with increased endogenous plasma glucocorticoids (15). PMN activation also occurs during exercise, as indicated by increased plasma PMN elastase (8). In addition, endurance exercise causes release of hormones, such as cortisol and growth hormone (24), and mediators, such as tumor necrosis factor (TNF)-α, interleukin (IL)-6, and granulocyte colony-stimulating factor (G-CSF) (17), all known to promote the growth and release of hematopoietic progenitor cells (HPCs) (11). Whether exercise may cause release of other hematopoietic growth factors, such as fms-like tyrosine kinase-3 (flt3)-ligand, known to potently induce the growth and differentiation of HPCs in vivo and in vitro (13), has not been assessed. As for training-induced adaptations, exercise-induced neutrophilia was shown to become progressively blunted with training (22,23), but no study ever tested whether circulating HPC counts may differ between trained and sedentary subjects. Circulating immature cells are likely involved in angiogenesis (19) and repair processes (21), both mechanisms being possibly associated with strenuous exercise and progressive training. Given the large use of exercise-based rehabilitation programs in several diseases, knowledge of the physiological effects of training on HPCs might be of potential clinical use. In this study, we measured circulating CD34-positive (CD34+) cells and their subpopulations in healthy amateur runners. CD34+ cells are early HPCs with undifferentiated morphology, whose maturation steps are indicated by progressive acquisition of CD38, human leukocyte antigen (HLA)-DR, and CD33 markers. Eventually, these cells lose the CD34 marker as they enter the differentiation pathway, thereby developing lineage-specific features (26). Baseline HPC counts in runners, likely to reflect chronic, training-associated changes, were compared with data from sedentary controls. The response of HPCs to exercise was assessed by studying the same runners at the end of the 1999 Palermo International marathon (M) or half-marathon (HM), respectively, and the following morning. This experimental design was chosen to analyze whether amount or duration of endurance exercise could modulate inflammatory and stress mediators, as well as circulating HPC counts. METHODSSixteen male amateur runners participating in the 5th Palermo International M (n = 8) or HM (n = 8) were studied. The entire group had a mean age of 41.3 ± 13.4 yr and a racing experience of 11 ± 9 yr (range 1–35 yr). Mean body weight and height were 70 ± 7 kg and 173 ± 5 cm, respectively. On average, the HM and M groups trained 89 ± 32 and 99 ± 33 km/wk, respectively (not significant). M and HM runners differed significantly for age (M: 50.4 ± 9.5 yr, HM: 33.1 ± 11.5 yr; P < 0.005 by unpaired t-test). Nine sedentary healthy men (age 39.4 ± 10.2 yr, body weight 82 ± 9 kg, height 178 ± 6 cm) were studied at baseline as controls. All subjects were nonsmokers, clinically healthy, and with no history of recent infection or other disease. The protocol was approved by the local Ethics Committee, and all subjects gave written informed consent to the study. The 5th Palermo International M and HM were held on December 8, 1999 at sea level. The race began at 9:00 AM. Weather conditions were good (mean hourly data from 9:00 AM to 1:00 PM by the City of Palermo Weather Bureau: barometric pressure: 1,007 mbar; temperature: 12.1 ± 1.7°C; wind: 0.73 ± 0.24 m/s; solar irradiation: 315.6 ± 60.9 W/m2). Runners were allowed free intake of water during the race. Mean race time was 79 ± 7 min (range 69–91 min) in the HM group, and 207 ± 38 min (range 159–254 min) in the M group. Runners were studied under baseline conditions 9 ± 2 days before the race, shortly (13 ± 7 min) after completion of the race, and the next morning. On the day of the race, blood samples were obtained at the finish line and kept at 4°C during transportation to the V. Cervello Hospital in Palermo, where they were immediately processed. All other samples in runners and controls were collected at the hospital in the morning in fasting conditions, kept at 4°C, and immediately processed. Blood was drawn from the subject's antecubital vein into sterile tubes containing EDTA (Vacutainer, Becton Dickinson, San Jose, CA) for complete blood cell counts (ADVIA counter, Bayer) and analysis of HPC by cytofluorimetry. For surface marker analysis, peripheral blood samples were analyzed for the expression of the CD34 (human progenitor cell antigen-2 FITC), CD38 (Leu-17 phycoerythrin), CD33 (LeuM9 phycoerythrin), and HLA-DR (HLA-DR peridinin chlorophyll protein) antigens (Becton Dickinson) by using three-color staining. Ig isotype-negative controls were used. The samples were immunofluorescence labeled, and flow cytometry was performed on a FACScan Excalibur with CellQuest software (Becton Dickinson). Analysis was performed by using large contiguous gates on lymphocyte and monocyte regions (7). The percentage of CD34+cells was calculated by adding the percentage of CD34+cells on the lymphocyte and monocyte gates and subtracting the percentage of cells stained with the control reagents. The number of total cells acquired was 50,000 to ensure adequate sensitivity of the analysis. Total CD34+ cells were determined in triplicate in each subject, because three samples (for CD34/CD38, CD34/HLA-DR, and CD34/CD33 antigens) were analyzed; the mean value was used for analysis. Variability for total CD34+ cell counts was estimated by calculating the coefficient of variation in each subject. Circulating reticulocytes were also determined by the microscopic brilliant cresyl blue method. In each experimental condition, aliquots of plasma and serum were collected and stored at −80°C. Muscle enzymes [lactic dehydrogenase (LDH) and creatine kinase (CK)] and serum iron were measured by enzymatic assays (Olympus 640 kits and equipment, Olympus Diagnostica, Hamburg, Germany). Total plasma elastase was measured by a homogeneous enzyme immunoassay specific for human PMN elastase (detection threshold: 4 μg/l; Ecoline kit, Merck, Darmstadt, Germany). Serum cortisol was measured by radioimmunoassay (sensitivity 0.36 mg/dl; Immunotech, Marseille, France). Plasma levels of TNF-α, IL-6, G-CSF, flt3-ligand, and erythropoietin (EPO) were measured by immunoassay (R&D Systems Europe, Abingdon, UK). High-sensitivity kits were used for TNF-α (sensitivity: 0.18 pg/ml) and G-CSF (sensitivity: 0.8 pg/ml). The lower detection limit of the IL-6 assay was 0.7 pg/ml, and this value was used in statistical analysis for samples with undetectable IL-6 levels. Lower detection limits for the other assays were 7 pg/ml for flt3-ligand and 0.6 mIU/ml for EPO. All data are reported as means ± SD. Runners and controls at baseline were compared by unpaired t-test (normally distributed variables) and Mann-Whitney test (CD34+ cell counts). Data obtained at different time points in runners were analyzed by repeated-measures ANOVA or paired t-test with Bonferroni correction for post hoc comparisons. Relationships between variables were analyzed by simple linear regression. The statistical analysis package used was Statview 4.5 (Abacus Concept, Berkeley, CA). Significance was set at P < 0.05. RESULTSTable 1 reports main data on blood cell counts in runners and controls. Serum iron concentration was in the normal range in both groups, but circulating reticulocytes in runners were about one-half the value of controls (P < 0.0005). Total and differential WBC counts did not differ between groups.
Circulating CD34+ cells (Fig.1, Table 2) were three to four times higher in runners compared with controls, without significant differences between M and HM groups for total CD34+ cells or any subpopulation. Two HM runners showed very high total CD34+ cell counts, accounting for the trend toward higher values in the HM compared with M group (individual total CD34+ cell counts at baseline are reported in the x-axis of Fig.2A). CD34+/CD38− cells were very low in runners and controls, whereas total CD34+ and all other subpopulations were increased in runners. The coefficient of variation for total CD34+ cell counts was 9.5 ± 7.0% in runners and 13.2 ± 10.6% in controls (difference not significant). Fig. 1.Total circulating CD34+ cells and their subpopulations in controls and runners at baseline. Values are means ± SD. * Significant difference between groups by Mann-Whitney test, P < 0.005.
Fig. 2.Change in total CD34+ cell counts at the end of race vs. baseline total CD34+ counts (A) and training volume (B). ○, Half-marathon (HM) runners; ■, marathon (M) runners. A: the regression line refers to pooled HM and M points (r = −0.86,P < 0.001; regression on HM and M points only:r = −0.89, P < 0.005 andr = −0.76, P < 0.05, respectively).B: the regression line refers to M points (r= −0.88, P < 0.005), as it was not significant in the HM group. In runners, red blood cell counts were inversely correlated to training volume, estimated as kilometers per week (r = −0.51,P < 0.05). Neither circulating CD34+ cell or reticulocyte counts at baseline correlated with age or training volume. When the analysis of CD34+ cell counts was repeated after excluding two HM runners showing high baseline CD34+counts (see above), results of regression analyses did not change. Total WBC and PMN counts increased after the race, more after M than after HM (Table 1). Reticulocyte counts doubled after both M and HM (Table 1). All values had returned to baseline value by the morning after the race. Total circulating CD34+ cells or subpopulations at the end of the M or HM were unchanged compared with baseline (Table 2). The coefficients of variation for total CD34+ counts at the end of M and HM were 12.8 ± 9.1 and 11.7 ± 11.2%, respectively (not different from baseline or control values). In individual runners, the change in total CD34+ count at the end of the race was smallest in subjects with the highest baseline counts (r = −0.862, P < 0.0001 for pooled HM and M data; Fig. 2A). Absolute CD34+counts after the race did not correlate with training volume, but their change at the end of the race was smallest in M runners with the highest training volume (r = −0.88, P< 0.005 in the M group; Fig. 2B). Muscle enzymes at baseline were higher in runners (LDH: 352 ± 90 U/l; CK: 290 ± 173 U/l) than in controls (LDH: 288 ± 36 U/l; CK: 110 ± 69 U/l; P = 0.05 and <0.01, respectively, vs. runners). LDH increased immediately postrace (M: 523 ± 133 U/l; HM: 438 ± 84 U/l; P < 0.001 vs. baseline), whereas CK increased markedly on the day after the race (M: 1,279 ± 1,049 U/l; HM: 944 ± 856 U/l; P< 0.005 vs. baseline), without significant differences between M and HM runners. Figure 3 summarizes the plasma levels of total neutrophil elastase, TNF-α, IL-6, G-CSF, cortisol, and flt3-ligand in all groups and experimental time points. At baseline, runners and controls were similar, but flt3-ligand was lower in HM than M runners, possibly due to an age effect (flt3-ligand vs. age: controls, r = 0.76; runners, r = 0.63;P < 0.05 for both). TNF-α, PMN elastase, cortisol, and flt3-ligand increased after the race, irrespective of M and HM distance. Conversely, IL-6 and G-CSF increased more after M (40-fold and 3-fold, respectively) than after HM (10-fold and by 47%, respectively). EPO was unchanged after the race. Fig. 3.Inflammatory markers, cytokines, and hormone and growth factors in controls (open bars) and in runners (shaded bars, HM; solid bars, M) at baseline, end of race, and morning postrace. TNF-α, tumor necrosis factor-α; IL-6, interleukin-6; flt3-ligand, fms-like tyrosine kinase-3-ligand; G-CSF, granulocyte colony-stimulating factor. Values are means ± SD. * Significant changes compared with baseline,P < 0.01. Horizontal bars, significant differences between HM and M groups. In runners, plasma cortisol was positively associated with G-CSF (r = 0.76, P < 0.0001) and flt3-ligand (r = 0.71, P < 0.001), whereas its association with IL-6 was less consistent (r = 0.76,P < 0.05 in post-M samples only). Weaker associations were found between cortisol and markers of inflammation (TNF-α:r = 0.56, elastase: r = 0.72,P < 0.001 for both). Plasma cortisol correlated with WBC (r = 0.82, P < 0.0001), neutrophil (r = 0.83, P < 0.0001), and reticulocyte counts (r = 0.61, P < 0.0001) but not with CD34+ cell counts. Both TNF-α and PMN elastase correlated with flt3-ligand (r = 0.76 and 0.66, respectively; P < 0.0001) and reticulocyte counts (r = 0.47 and 0.66, respectively; P < 0.001) but not with CD34+ cell counts. In both HM and M groups, total CD34+ cell counts were lower on the morning postrace (Table 2) compared with end of the race (pooled data: from 18.0 ± 11.4 to 9.7 ± 5.1 cells/μl;P < 0.001). A similar trend was observed in CD34+/HLA-DR− and CD34+/CD33− subpopulations (P< 0.05 for both) only in M runners. We asked whether changes in HPC counts during recovery after the race (i.e., the change in cell counts between the morning postrace minus end of race) correlated with exercise-associated release of cytokines and growth factors. This hypothesis was not confirmed for plasma cortisol, IL-6, and G-CSF, as their levels at end of race did not correlate with total CD34+ cells or any subpopulation on the morning postrace. Conversely, the higher the level of flt3-ligand or TNF-α at the end of the race, the lower CD34+/HLA-DR+ and CD34+/CD33+ cell counts the morning postrace (Fig. 4). Fig. 4.flt3-Ligand (top) and TNF-α (bottom) plasma levels at end of race correlated inversely with CD34+/human leukocyte antigen-DR+ (left) and CD34+/CD33+ (right) cell counts on the morning postrace. Δ, Change. Symbols are as described in Fig. 2legend. Regression lines refer to pooled HM and M points. DISCUSSIONThe main result of this study is that chronic hematopoietic adaptations occur in well-trained athletes, as indicated by higher baseline circulating HPCs in runners than in sedentary controls. The second result is that HPCs did not change immediately after the race but were decreased on the following morning. Plasma cytokines and growth factors increased after both M and HM races and correlated with some changes in HPC counts observed the morning postrace, supporting the hypothesis that daily bouts of exercise, with release of hormones and other mediators, may act as chronic intermittent stimuli modulating HPCs in peripheral blood. A marathon is an established model of exercise-induced inflammation (6, 8, 16, 17, 23, 24). Exercise-induced neutrophilia partly results from mobilization of the marginated pool (5), secondary to increased cardiac output and plasma cortisol (15). Running a marathon increases plasma PMN elastase (8) and induces a complex pattern of pro- and anti-inflammatory mediators (17, 24), including cytokines and growth factors known to affect hematopoiesis, supporting the rationale of our study. Circulating HPCs were unaffected acutely by the M or HM. However, in both HM and M groups, subjects with the highest CD34+counts at rest showed the smallest change in CD34+ cells at the end of the race (Fig. 2A). Furthermore, the change in CD34+ counts at the end of the race in M runners correlated inversely with training volume, strongly supporting a training effect. Sedentary subjects undergoing a week of daily exercise showed progressively blunted exercise-induced neutrophilia and band cell release (22, 23). Our observations on more immature blood cells support exercise-dependent modulation of the early HPC pool. However, no significant relationship could be demonstrated between training volume and CD34+ cell counts at baseline. Exercise-induced release of stress and inflammatory mediators could be the link between exercise and modulation of hematopoiesis. Changes in circulating HPC counts appeared delayed compared with exercise-induced release of mediators. Total CD34+ cells decreased the morning postrace, together with CD34+/HLA-DR−and CD34+/CD33− subpopulation counts in M runners. Plasma cortisol increased similarly after M and HM and correlated with plasma levels of growth factors known to act on the bone marrow. Our study is the first to report that plasma flt3-ligand, a known activator of HPCs (13), doubled after exercise, irrespective of running distance. TNF-α, G-CSF, and IL-6 are all known to affect hematopoiesis (11, 18, 27). In humans, G-CSF is used to mobilize and apheretically collect HPCs for transplantation and acts synergistically with IL-1, IL-3, and IL-6 on the bone marrow (11, 27). G-CSF-induced mobilization of HPCs was shown to be mediated by release of elastase by PMNs in the bone marrow (12). A similar mechanism may operate in runners through exercise-induced PMN activation and elastase release. IL-6 also modulates hematopoietic cell growth and differentiation (11, 18) and is released by exercising muscles (6). However, because G-CSF and IL-6 increased less after HM than after M, their role on HPCs might be relatively minor compared with that of flt3-ligand. Indeed, the increase in flt3-ligand during exercise correlated inversely with decreased CD34+/HLA-DR+ and CD34+/CD33+ counts on the morning postrace. Conversely, we found no correlation between increased plasma IL-6 or G-CSF after the race and changes in CD34+ cells on the morning postrace. However, because several significant associations were found between plasma cortisol and inflammatory markers, it is likely that both stress and inflammation may come into play to modulate HPC subpopulations in runners. Only circulating CD34+/CD38− cells were low in both runners and controls, suggesting no effect of exercise on the very early CD34+ cell pool. Data from M and HM runners were similar for many variables. This is not surprising, because data after a real competition reflect the best possible performance in each runner, accounting for similar levels of markers of stress (cortisol) and muscle damage (enzymes) in both groups. Conversely, peripheral blood neutrophilia, IL-6, and G-CSF were higher after the M race (16), supporting the fact that they were influenced by the duration of exercise. Because M runners were slightly older than HM runners or controls, our conclusions may be affected at least partly by an age effect. We consider this hypothesis unlikely. CD34+ cells tend to decrease with age, but M runners showed higher, not lower, counts compared with controls. In addition, CD34+ cell counts did not differ significantly between M and HM runners, despite the significant age difference between groups. An effect of age, instead, was apparent for flt3-ligand at baseline in both runners and controls. The slight increase in flt3-ligand levels could reflect decreasing bone marrow sensitivity to this factor with age. We are aware that circulating HPC subpopulations reflect hematopoiesis only indirectly, and the precise mechanism of increased HPCs in runners or the effects of M or HM races on maturation of HPC cannot be inferred by our data. Ethical limitations prevent obtaining bone marrow samples in healthy subjects. On the other hand, our runners were not elite athletes, and the results of this study represent physiological responses in normal, healthy individuals. The data prompt the question on the physiological role of circulating HPCs in runners. It is possible that increased “turnover” of polymorphonuclear cells associated with exercise and the consequent inflammatory state may stimulate HPC production and mobilization; in this case, the decrease in HPC counts documented after the race might reflect maturation and replenishment of the peripheral PMN compartment after prolonged and intense exercise. Alternatively, we speculate on a possible HPC involvement in angiogenesis and repair processes associated with exercise and training. Circulating HPCs were involved in neovascularization after myocardial damage in humans (20) and include cells with an angioblastic potential (19). Exercise is a potent cause of angiogenesis in skeletal muscle (9), and circulating HPCs might contribute to this process. Further studies with measurements of specific markers of early endothelial or muscle cell differentiation, however, are necessary to support this hypothesis. As for the possible HPC involvement in tissue repair, it could apply to several tissues, including the lung, as bone marrow stem cells can differentiate into alveolar epithelium (10). In induced sputum of runners, we found very high PMN counts after a marathon and a pattern in PMN adhesion molecules, suggesting ongoing repair postrace (4). Should our findings be confirmed, they may be relevant in exercise-based cardiac and/or respiratory rehabilitation programs, as well as in other diseases. A study in well-trained subjects, however, cannot provide information on the exercise threshold necessary to trigger hematopoietic adaptations. This point deserves further study, together with the molecular and functional characterization of HPCs before and after training. Opposite to the behavior of HPCs, runners appeared to “spare” reticulocytes at rest and released them transiently after exercise. Our data confirm the low-reticulocyte counts at rest in runners studied during a training period (14). The increase in reticulocytes after the race correlated with plasma cortisol but was not associated with changes in EPO, as already reported by Bodary and coworkers (3). Besides stress, hemolysis secondary to the mechanical impact of running (25) could also modulate the reticulocyte response to exercise. In conclusion, HPC counts were increased in peripheral blood of runners at baseline and did not change immediately after a M or HM race. Circulating CD34+ cells decreased on the morning postrace, suggesting modulation of hematopoiesis during recovery after intense and prolonged exercise. Exercise-induced release of inflammatory and stress mediators, together with increased levels of hematopoietically active growth factors such as flt3-ligand, IL-6, and G-CSF, may be the pathophysiological link between exercise and chronically increased circulating HPCs in runners. The potential application of these findings to competitive training and exercise-based rehabilitation programs deserves further study. We thank all athletes for their cooperation, Dr. Maria Grazia Alaimo for providing data on weather conditions on the day of the race, Dr. Silvia Lenzi for help in data collection, and Drs. P. Marozzi and G. Alercia for cortisol measurements. We are indebted to Dr. Giuseppe Galfano and Prof. Enrico Cillari, Azienda Sanitaria Ospedale V. Cervello, Palermo, for sample processing by the laboratories of the hospital. FOOTNOTESREFERENCES
Page 10participation by women in both recreational and competitive sports has increased dramatically over the last two decades. In addition, the US Surgeon General's Report on Physical Activity and Health recommends that women of all ages, not just athletes, include a minimum of 30 min of moderate-intensity exercise on most days of the week (25). However, dietary energy insufficiency associated with high-intensity exercise training and competition can increase a woman's risk of experiencing an abnormal menstrual cycle (3, 13). Abnormal menstrual cycles, with chronically low ovarian hormones, may increase the risk for osteopenia, osteoporosis, and fractures (7). Oral contraceptives (OCs) are used for birth control in normally menstruating young women, and, although controversial, OCs have been used to prevent bone loss in amenorrheic athletes (8, 16). However, there is concern among athletes that these exogenous ovarian hormones affect exercise performance. Peak oxygen consumption (V˙o2 peak) is considered the “standard” for assessing aerobic exercise capacity (23), and V˙o2 peak in women could vary owing to ovarian hormone influences on stroke volume, pulmonary minute ventilation, oxygen-carrying capacity, blood flow, and muscle oxygen utilization. Although the cyclic endogenous ovarian hormone fluctuations across the normal menstrual cycle do not appear to affect V˙o2 peak (1, 6, 12), low-dose administration of exogenous estrogen and progesterone may have a greater influence on exercise capacity. Only a few studies have examined the effects of exogenous steroids on exercise performance by use of longitudinal study designs. Although short-term OC use (21 days) did not affect V˙o2 peak(2), 6 mo of monophasic OC use was associated with a significant decrease in V˙o2 peak in endurance-trained women (18). To our knowledge, no longitudinal studies have examined peak exercise capacity in moderately trained women before and after triphasic OC use. With monophasic OCs, the estrogen and progestin components remain constant throughout the pill cycle. In contrast, in triphasic OCs the amounts of estrogen and/or progestin vary across the pill cycle and more closely mimic the ovarian hormone variation that occurs during the normal menstrual cycle. Triphasic OCs contain lower per-cycle progestin levels to provide better cycle control and reduce the incidence of androgenic side effects such as alterations in carbohydrate and lipid metabolism (4) and therefore may not have the same influence on exercise capacity as monophasic OCs. The purpose of this investigation was to examine the effects of menstrual cycle phase (endogenous ovarian hormones) and triphasic OC use (exogenous ovarian hormone analogs) on peak exercise capacity, as measured byV˙o2 peak. MATERIALS AND METHODSEight subjects were recruited from the University of California, Berkeley campus, to participate in a series of experiments to examine the effects of ovarian hormones on cardiorespiratory function and substrate utilization during peak and prolonged submaximal exercise. Results from the submaximal exercise trials on normally menstruating women (22) are reported separately. Retrospective blood analyses revealed that two of the subjects failed to meet the ovarian hormone concentration criteria for the follicular and luteal phases of the menstrual cycle and thus were excluded from data analysis. The final subject pool, for peak exercise analysis, consisted of six healthy, nonsmoking, female subjects (25.5 ± 1.5 yr). Subjects habitually exercised 2–6 h/wk (3.5 ± 0.6 h/wk) but were not competitive athletes. The women were nulliparous; had been diet, weight, and exercise stable; and had not taken OCs for at least 6 mo. All subjects reported consistently normal menstrual cycles (22–32 days) and were injury and disease free as determined by health history questionnaire and physical examination. Informed, written consent was provided, and the University of California Committee for the Protection of Human Subjects approved the study protocol (no. 2001-8-132). Physical work capacity and V˙o2 peak were tested, in a randomized order, during the early follicular (FP, 4–8 days after the start of menses) and midluteal (LP, 17–25 after the start of menses and 6–9 days after ovulation) phases of the menstrual cycle before OC use. Ovulation was determined by using urine ovulation predictor kits (First Response, Carter Products, New York, NY). FP and LP were confirmed by plasma estradiol and progesterone concentrations from blood samples taken at rest before the peak exercise test or from blood sampled at rest on the same day of the next menstrual cycle. Progesterone levels above 3 ng/ml were used for verification of the luteal phase (21). Peak exercise testing was completed within one to two sequential menstrual cycles. After completion of the menstrual cycle phase testing, each subject began taking the same triphasic OC (one pill per day) for four complete cycles (28 days per cycle). For days 1–7 each pill contained 0.035 mg ethinyl estradiol and 0.18 mg norgestimate, fordays 8–14 each pill contained 0.035 mg ethinyl estradiol and 0.215 mg norgestimate, for days 15–21each pill contained 0.035 mg ethinyl estradiol and 0.25 mg norgestimate, and for days 22–28 the pills were absent of synthetic hormones. Physical work capacity andV˙o2 peak were reassessed during the week of the inactive pills (IP) and during the second week of active pill ingestion (HP). Subjects were instructed to refrain from exercise, caffeine, and medications 24 h before testing, to eat a light meal 3 h before arriving at the laboratory, and to maintain constant diet and exercise regimens throughout the entire experimental period. Three-day dietary records were collected and analyzed before and after the 4 mo of OC using the Nutritionist III program (N-Squared Computing, Salem, OR). Before each peak exercise test, subjects were weighed and body composition was determined (six-site skinfolds with a Harpenden skinfold caliper) (9). A continuously graded exercise test was conducted on an electronically braked cycle ergometer (Monark Ergometric 839E, Vansbro, Sweden). The workload began at 75 W and was increased by 25 W every 3 min until volitional exhaustion. The test was considered maximal if respiratory exchange ratio values exceeded 1.1. Respiratory gases were continuously collected and analyzed via an open-circuit indirect calorimetry system (Ametek S-3A1 O2 and Ametek CD-3A CO2 analyzers, Pittsburgh, PA), and respiratory parameters were recorded every minute by a real-time, on-line personal computer-based system. Heart rate was continuously monitored by a Quinton Q750 electrocardiograph (Bothell, WA). Blood was sampled at rest and immediately transferred to collection tubes containing EDTA for hormone determination. Plasma estradiol and progesterone concentrations were determined by 125I radioimmunoassay (Coat-A-Count kits; Diagnostic Products, Los Angeles, CA). All samples for each subject were analyzed together. The intra-assay coefficients of variation were 1–5%. Repeated-measures ANOVA and Fisher's protected least significant difference post hoc tests were used to determine phase differences in body weight, body composition, diet composition, estradiol and progesterone concentrations and peak power output, heart rate, pulmonary minute ventilation, oxygen consumption rate, carbon dioxide production, and respiratory exchange ratio by use of Statview 5.0.1 (SAS Institute, Cary, NC). Results are expressed as means ± SE throughout the text. The significance level was set at α < 0.05. RESULTSSubject characteristics for the six women that met the ovarian hormone criteria on the day of maximal exercise testing, for all phases, are presented in Table 1. Subject numbers and characteristics vary between the series of reports (22) from our laboratory because not all of the eight subjects met the phase criteria for every experimental protocol. There were no significant differences in body weight or body composition between FP and LP before OCs or between IP and HP with OCs, except a slightly higher fat mass in LP vs. FP. However, there was a small, but significant (P < 0.05), increase in body weight (3%) and fat mass (9%) after 4 mo of OC use.
There were no significant changes in total energy intake (1,920 ± 191 kcal before OC and 1,819 ± 260 kcal after OC), percentage of the energy intake as carbohydrate (58 ± 2.2% before OC and 54 ± 2.9% after OC), percentage of the energy intake as fat (27 ± 2.9% before OC and 31 ± 3.1% after OC), and percentage of the energy intake as protein (15 ± 1.6% before OC and 14 ± 1.3% after OC) with OC use. Day of cycle, days past ovulation, and the ovarian hormone profiles for each phase are shown in Table 2. Criteria for FP (progesterone < 1 ng/ml) and LP (progesterone > 3 ng/ml) were met in six subjects before OC use. LP was associated with significantly higher (P < 0.05) estradiol and progesterone concentrations than all other phases. Both ovarian hormones were low after OC use, validating the suppression of endogenous hormone production by synthetic ovarian steroids.
At peak effort, there were no significant differences in any of the cardiorespiratory variables between FP and LP before OC use or between IP and HP with OC (Table 3). However, after 4 mo of OC use, there were significant decreases (P < 0.05) in time to peak exercise (14%) and in the peak power output attained (8%). There were also significant (P < 0.05) reductions inV˙o2 peak measured in both liters per minute (11%) and milliliters per kilogram per minute (13%) and in peak carbon dioxide production (15%). There were no significant changes in peak heart rate, pulmonary minute ventilation, and respiratory exchange ratio. All six subjects experienced a decline inV˙o2 peak(ml · kg−1 · min−1) after 4 mo of OC use (Fig. 1).
Fig. 1.Peak oxygen consumption (V˙o2 peak) before and after 4 mo of oral contraceptives (OC) for each of the 6 subjects. BOC, before OC (mean = 42.5 ± 3.3 ml · kg−1 · min−1); AOC, after OC (mean = 36.9 ± 2.6 ml · kg−1 · min−1). Some of the same subjects as a previous report (22) were included in the BOC data. DISCUSSIONThis study confirms that, in the absence of OCs, menstrual cycle phase does not affect peak exercise capacity, with no significant changes in body weight, body composition, or cardiorespiratory factors, including V˙o2 peak, between the follicular and luteal phases. However, 4 mo of a low-dose triphasic OC resulted in a significant increase in body weight and fat mass and a significant 11% decrease in V˙o2 peaknot normalized to body mass. There was no change inV˙o2 peak between the inactive and high-dose phase with OCs, suggesting a persistent synthetic ovarian hormone effect despite a 1-wk cessation of ovarian steroid intake between cycles. That OCs, but not luteal phase menstrual cycle variations in ovarian hormones, affected V˙o2 peak suggests that steroid levels may be involved in suppression of peak exercise capacity. OCs mimic the estrogen profile during pregnancy, with high levels of ethinyl estradiol (>300 pg/ml), levels that are much higher than observed during the normal menstrual cycle (24). As well, the type of contraceptive pill may have an effect onV˙o2 peak. Our finding of an 11–13% decrease in V˙o2 peak in moderately trained women after 4 mo of triphasic OCs is greater than the 7% decrease in V˙o2 peak found in endurance-trained women with 6 mo of monophasic OCs (18). Moreover, C. M. Lebrun (unpublished observations) has observed a similar, small, but statistically significant decrease inV˙o2 peak with triphasic OC use in athletic women. And, finally, the duration of OC use may play a role. Longer than 1 mo of OC use appears to be necessary to induce physiological changes because a study examining 1–3 wk of monophasic OCs found no significant effect onV˙o2 peak (2). Although the number of subjects was small in our investigation, every subject experienced a drop in V˙o2 peakwith OC use, indicating a significant physiological effect. ThatV˙o2 peak was depressed during both IP (ethinyl estradiol levels ≈ 8 pg/ml) and HP (ethinyl estradiol levels > 300 pg/ml) phases of OC use is taken to indicate persistence of OC effects (24). In agreement with our findings are those of Lynch et al. (14), who looked at the effects of long-term OC use on intermittent exercise performance in untrained women. Factors that could reduce V˙o2 peakinclude decreases in stroke volume, oxygen-carrying capacity (hemoglobin levels), muscle blood flow, or oxygen extraction or changes in the pattern of substrate utilization. However, most of these do not appear to be candidates for an OC-induced negative effect onV˙o2 peak. A decrease in stroke volume is unlikely because estrogen replacement therapy has been shown to increase stroke volume (10) and OC use has been shown to increase the activity of the renin-angiotensin-aldosterone system at rest (19). Decreased hemoglobin concentration is also unlikely because most studies have found no difference in resting blood hemoglobin and ferritin concentrations (11, 17) and an increase in serum iron levels (17) with OC use, presumably owing to a decrease in menstrual blood loss (11). Although we did not directly assess sympathetic nervous system activity (SNA) in this study, decreased SNA and plasma catecholamine concentrations could explain the lower peak oxygen consumption observed with high ovarian hormone concentrations. Consistently high estrogen and progesterone concentrations, such as occur during pregnancy and with exogenous ovarian hormones, may blunt SNA and catecholamine levels as a protective mechanism to maintain blood flow to the uterus and prevent maternal hypoglycemia and uterine contractions (15). Both the sympathetic nervous and endocrine systems play roles in maintaining normal blood glucose concentrations. Because catecholamines do not begin to rise in the circulation until the level of effort becomes strenuous (e.g., >65%V˙o2 peak), catecholamines are directly involved in glycogen mobilization during strenuous exercise. In contrast, hormones such as human chorionic somatotropin, growth hormone, cortisol, and thyroid hormone play roles in maintaining glucose homeostasis during pregnancy (15), and their importance is more likely exhibited during submaximal prolonged exercise. However, catecholamines are directly involved in hepatic and muscle glycogen mobilization during strenuous exercise. The fetus relies almost exclusively on maternal glucose for growth and development (15), and suppression of epinephrine and norepinephrine release could be a means of preventing maternal liver glycogen depletion and low blood glucose concentrations. Pregnancy is associated with suppressed catecholamine levels during strenuous exercise (15), and exogenous estradiol administration has been shown to decrease SNA at rest (26), decrease catecholamine levels and glucose production and utilization during exercise (20), and increase the levels of the potent vasodilator nitric oxide (5). During exercise, increased SNA and the resultant vasoconstriction in nonactive tissue is essential for increasing blood flow to the working muscle. As exercise intensity increases, some vasoconstriction in the active muscle is also required to maintain mean arterial pressure. Blunting of SNA with high ovarian hormone concentrations, therefore, could limit peak exercise performance. Although oral contraceptives decrease peak exercise capacity in moderately trained young women, effects of these synthetic steroid hormones on prolonged endurance exercise performance in competitive athletes are less obvious and warrant further investigation. The decrement in V˙o2 peak induced by OC use may subside over time or become insignificant owing to training-induced adaptations in highly trained female athletes. In conclusion, these results suggest that 1) endogenous hormones have little effect on exercise performance as measured byV˙o2 peak, but 2) low-dose triphasic OCs (exogenous ovarian hormones) appear to decrease peak exercise performance in moderately physically active young women. The authors thank the subjects for their dedication to all aspects of the study. We also thank Joe Vivo, Zinta Zarins, and Christina Chueng for contributions to the data collection and blood analysis. We also thank Rosemary Agostini for commenting on the manuscript. FOOTNOTESREFERENCES
Page 11na+-k+-atpase, or na+-k+pump, catalyzes coupling of the chemical hydrolysis of ATP to the vectoral transport of Na+ out of and K+ into the cell. In skeletal muscle, basal activity of the Na+-K+ pump depends primarily on the distribution of Na+ and K+ on either side of the plasma membrane and utilizes only ∼2–8% of maximum pumping capacity in vivo (23). During contractile work, transport of the Na+ and K+ by the Na+-K+ pump rapidly restores ionic gradients after an excitatory electrical potential or “action potential” (35). The capacity of the Na+-K+pump can be challenged by heavy contractile demands and by changes in the local environment (i.e., ionic and metabolite buildup, substrate depletion), and, therefore, appropriate regulation of Na+-K+-ATPase activity is essential for maintaining transport capacity and muscle excitability. Despite a number of articles and reviews published recently on the role of the Na+-K+ pump in muscle excitability and fatigue (1, 12, 13, 20, 21, 35, 37), the factors controlling both the acute and long-term regulation of the pump with exercise are still largely undetermined. Acute regulation of the Na+-K+ pump can occur by influencing the activity of the pumps and by modulating the number of pumps at the cell surface (7). With contraction, the passive Na+ influx and K+ efflux can stimulate up to a 20-fold increase in Na+-K+ pump activity (12, 35). Substrates, cytoskeletal components, catecholamines, and hormones can also provide additional short-term activation (6, 16, 40). In addition, an increase in the number of pumps at the sarcolemma is believed to occur acutely by translocating a specific pool of Na+-K+ pump subunits from intracellular sites to the muscle membrane (27), potentially increasing the number of functional Na+-K+ pumps. Collectively, these cellular processes increase the transport capacity of Na+-K+-ATPase during exercise and, therefore, help to maintain ion gradients, excitability, and contractility. Despite the extensive research into mechanisms that may increase pump transport capacity during activity, it is unclear whether intrinsic Na+-K+-ATPase activity can be altered by exercise, as has been observed for the other ATPases in the cell. As an example, Williams et al. (41) observed that maximal activity of the actomyosin ATPase in frog muscle was reduced by 20% with a 5-min stimulation protocol involving repeated maximal contractions. Several investigators have also observed decreased Ca2+-ATPase activity after fatiguing exercise in rats (4, 9, 41). Although the exact mechanisms involved in the inactivation of these ATPases remain unclear, structural damage induced by the generation of free radicals is strongly suspected (33). The evidence from cardiac muscle Na+-K+-ATPase indicates that damage by circulating free radicals can occur (32). Collectively, the results suggest that catabolic processes associated with exercise may alter the structure of the skeletal Na+-K+-ATPase, reduce membrane excitability, impair conduction of action potentials, and contribute to fatigue. No published study presently exists, however, that examines the intrinsic activity of skeletal muscle Na+-K+-ATPase in response to exercise. The goal of this study was to determine whether the Na+-K+-ATPase activity is altered in muscles of different fiber composition after prolonged endurance running in rats. Our hypotheses were that the Na+-K+ pump is intrinsically modified during exercise, which results in a reduced Na+-K+-ATPase activity, and that the reduction in Na+-K+-ATPase activity is not specific to the fiber-type composition of the muscle. These hypotheses were tested in muscle homogenates using a 3-O-methylfluorescein K+-stimulated phosphatase assay (3-O-MFPase) as an indicator of Na+-K+-ATPase activity. METHODSUntrained, female Sprague-Dawley rats (age 12.1 ± 0.7 wk; weight 275 ± 21 g; means ± SE) were utilized for the study. Rats were housed in a room where the light cycle was controlled (12:12-h light-dark cycle), and rat chow and water were provided ad libitum. Care and treatment of the animals was in accordance with procedures outline by the Canadian Council on Animal Care. All procedures were approved by the University of Waterloo Office for Ethics in Research. To investigate the effect of a single session of aerobic exercise on Na+-K+ pump function, rats were randomly assigned to one of three groups (n = 10 per group). In one group (Run), rats were run on a treadmill at 21 m/min and 8% grade (∼65% peak aerobic power) until fatigue or to a maximum of 2 h. A second group of rats (Run+) were run on the treadmill as for the Run protocol, and then they were kept on the treadmill for an additional 45 min of low-intensity exercise (i.e., continued fast walking at 10 m/min). By reducing the speed of the treadmill, we were able to increase exercise duration. A third group of rats (Con) served as control for the anesthetic and surgical procedures. This design, including the exercise protocol, is comparable to that previously used to investigate Ca2+-ATPase activity after running and recovery (18). Directly after exercise (or at rest in Con), rats were anaesthetized with pentobarbital sodium (6 mg/100 g body wt), and a muscle sample was obtained from soleus (Sol), extensor digitorum longus (EDL), red vastus lateralis (RV), and white vastus lateralis (WV) and plunged into liquid nitrogen for later analysis of muscle metabolites, glycogen content, and Na+-K+ pump characteristics. Time for anesthetization and surgery after exercise averaged ∼5 min. Surgery was not initiated until animals showed no reflex response to pinching the foot. In an additional experiment, performed after the first experiment, we have investigated the effect of resting or passive recovery after Run on the changes in Na+-K+pump function. For this purpose, rats with similar characteristics (weight 258 ± 18 g; n = 32) were randomized into four groups (n = 8 per group), namely a control (Con) and groups that received 10 (Rec 10), 25 (Rec 25), and 45 (Rec 45) min of passive recovery after Run. Muscle glycogen and metabolites, including ATP, phosphocreatine, creatine, and lactate, were analyzed fluorometrically after extraction from freeze-dried tissue, according to procedures previously published (22). In addition, we have also measured the contents of the adenine nucleotides (ATP, ADP, AMP) and inosine monophosphate by using ion-pair reversed-phase high-performance liquid chromatography (28) as modified by our group (22). All samples were analyzed in duplicate. On a given analytical day, an equal number of tissue samples from each muscle and group were measured. Activity of Na+-K+-ATPase was assessed by using the K+-stimulated 3-O-MFPase modified from the procedures of Huang and Askari (26) and Horgan and Kuypers (25) but using higher substrate concentration (19,26). We have confirmed in a separate set of experiments (results not shown) that maximal activity was achieved at ∼160 μM substrate concentration in rat tissue. Additionally, the use of 1.25 mM EGTA and 5 mM NaN3 was also employed to optimize enzyme activity in rat muscle samples (2). Briefly, tissue from frozen muscle samples was homogenized (5% wt/vol) at 0°C for 2 × 20 s at 25,000 rpm (Polytron) in a buffer containing (in mM) 250 sucrose, 2 EDTA, 1.25 EGTA, 5 NaN3, and 10 Tris (pH 7.40). Homogenates were freeze thawed four times and diluted 1:5 in cold homogenate buffer. Approximately 30 μg of protein (∼30 μl homogenate) were incubated for 4 min in medium containing (in mM) 5 MgCl2, 1.25 EDTA, 1.25 EGTA, 5 NaN3, and 100 Tris (pH 7.40). The K+-stimulated activity of the Na+-K+-ATPase was determined by the increase in activity after the addition of 10 mM KCl at a substrate concentration of 160 μM 3-O-MFPase. The activity of 3-O-MFPase was determined by the difference in slope before and after the addition of KCl. We have shown that the change in slope with the addition of KCl is completely eliminated with ouabain (H. Green, unpublished observation). Na+-K+-ATPase activity, which was based on the average of three trials, is expressed in nanomoles per milligram of protein per hour. Protein content of the homogenate was determined by the method of Lowry as modified by Schacterle and Pollock (36). The intra-assay coefficient of variation, defined as the standard deviation divided by the mean, was 10.9% when averaged over all muscles. Statistical analysis was performed on Statistica for Windows ver. 4.5 software (Statsoft, Tulsa, OK). Descriptive statistics included means ± SE. Two-way ANOVA with repeated measures was used to analyze difference in Na+-K+ pump activity between the three conditions (Con, Run, Run+) and within muscle groups. Similarly, a two-way ANOVA was used to examine the effects of passive recovery (Con, Rec 10, Rec 25, and Rec 45) and muscle (WV, RV). Identical procedures were employed for analyses of the metabolite data. Paired analysis was used to assess the activity response to exercise between muscles. Post hoc analysis of mean values was performed by using Tukey's test. The probability level of statistical significance was accepted at P < 0.05. RESULTSThe average running duration at 21 m/min for the exercise was 102.4 ± 5.2 min for both Run and Run+ groups. The Run+ group exercised for an additional 45 min at 10 m/min. The glycogen depletion pattern indicated that each muscle was used during the prolonged, low-intensity protocol (Fig. 1). Muscle glycogen was reduced (P < 0.05) between 33 and 66% in the Run group in the muscles examined. No differences were observed between muscles in the Run group for glycogen content. No recovery in glycogen content was observed in the Run+ group. Fig. 1.Glycogen content in different locomotor muscles of rats. Values are means ± SE of n = 10 rats per group except extensor digitorum longus (EDL) where n = 7. Con, nonexercise control; Run, running at 21 m/min and 8% grade for 2 h; Run+, Run protocol plus 45 min of walk recovery at 10 m/min; Sol, soleus; RV, red vastus; WV, white vastus. Glycogen depletion ranged between 33 and 66% from Sol to WV muscle. There was a main effect for group whereby recovery from Run was different from Con (P < 0.05). Exercise had little effect in altering adenine nucleotides (Table1). The only difference observed was an increased ADP content in Run relative to Con (P < 0.05) in Sol. Exercise also had minimal effects on high-energy phosphates and metabolites, namely ATP, phosphocreatine, creatine, and lactate (Table 1). Increases in creatine in Run+ relative to Con (P < 0.05) were observed but only for EDL (P < 0.05). When averaged over all conditions, differences were observed for adenine nucleotides and for the high-energy phosphates and metabolites between muscles, typical to what our laboratory has previously published (14).
Na+-K+-ATPase activity, measured by 3-O-MFPase (Fig. 2), decreased (P < 0.05) by ∼12% from Con to Run+ when averaged over all muscles (168 ± 7 vs. 149 ± 7 nmol · mg protein−1 · h−1, respectively). The decrease was not specific to muscle. No differences were observed between Con and Run and between Run and Run+ for any muscle. No differences were found in nonspecific or background activity between Con, Run, and Run+ for any muscle. Fig. 2.Na+-K+-ATPase activity, measured by 3-O-methylfluorescein phosphatase activity (3-O-MFPase) in different locomotor muscles of rats. Values are means ± SE of n = 10 rats per group except EDL where n = 7. There was a main effect for group whereby Run was different from Con (P < 0.05). In an additional experiment designed to investigate the effect of the type of recovery on 3-O-MFPase activity, rats were allowed to recover for varying periods of time without exercise after Run. Using RV and WV as representative tissue, we could find no main effects between Con and Rec 10, Rec 25, or Rec 45 on 3-O-MFPase activity (Table 2). As with the previous experiment, no differences were found in nonspecific activity between Con and Rec groups. In this experiment, the average run time for each group was similar to the initial experiment.
DISCUSSIONAs hypothesized, we have found that the exercise protocol that we employed induced a reduction in Na+-K+-ATPase activity when measured in muscle homogenates in vitro. Although there is a strong indication for Na+-K+-ATPase activity to decrease in all muscles except the EDL at Run (P = 0.12), an additional 45 min at reduced exercise intensity were needed to obtain a significant reduction. In addition, we have also found that with up to 45 min of resting recovery, Na+-K+-ATPase activity is not different from nonexercised control animals. Our results for Run are similar to what has been found for the Ca2+-ATPase activity in response to a similar bout of exercise (18). However, unlike Ca2+ activity, where an overshoot was observed to occur during the additional exercise (18), our results for Na+-K+-ATPase activity indicate that further exercise depresses activity. The effect of the additional exercise on Na+-K+-ATPase activity did not depend on the muscle examined. The basis of our hypothesis, namely that a reduction in Na+-K+-ATPase activity would occur with prolonged exercise, was based on previous reports of declines in SR Ca2+-ATPase that occurred in response to similar types of exercise (4, 8, 41). However, not all studies report a reduction in Ca2+-ATPase activity with exercise (11,14, 18). Although the reasons for the discrepancies remain unclear, differences in exercise protocols, muscles examined, assay procedures, and species appear important (20). Our study suggests that at least some of these factors may be important with Na+-K+-ATPase activity given the additional exercise that was needed to induce changes. The diminishing Na+-K+-ATPase activity that we have observed with exercise is likely to represent structural modifications to the enzyme since our measurements were performed under optimal conditions in vitro. There are a number of cellular mechanisms that may explain the intrinsic changes to the enzyme and the specific effect on Na+-K+-ATPase activity. The most notable possibilities for acute inactivation may result from free radical damage (33), Ca2+-activated proteolysis (3), and heat denaturation (17), all of which can increase with exercise (38). Free radical damage has been demonstrated to reduce Na+-K+-ATPase activity in cardiac tissue (32) in a time- and concentration-dependent manner after response to a reactive compound (31). Evidence from studies on mouse diaphragm indicates that excessive intercellular Ca2+ can inhibit Na+-pump activity (39). Prolonged exposure to heat stress has been demonstrated to alter a number of metabolic processes in skeletal muscles, including Ca2+-ATPase activity (17). Further study is required to determine which of these potential mechanisms may predominate, as well as identifying the site on the enzyme that is altered. One possible mechanism that was investigated as a cause of the diminished activity in this experiment was that of substrate depletion. Skeletal muscle Na+-K+-ATPase appears to preferentially utilize aerobic glycolysis for metabolism (29). Furthermore, James et al. (30) have reported that activity of the pump is dependent on glycogen as a substrate, and, therefore, glycogen depletion may induce fatigue. In the absence of regional differences, it is unlikely that the range of 33–65% in glycogen depletion that was observed had a significant effect on Na+-K+-ATPase activity in this experiment. However, recent evidence suggests that Na+-K+-ATPase activity is regulated through ankyrin-spectrin links to the cytoskeleton (40), so it is possible that even subtle changes in glycogen may change the structural balance necessary for optimal enzymatic function. The analytical procedure for measuring 3-O-MFPase activity involves four freeze-thaw cycles to permeabilize membranes for the optimal activation by K+, so it is unlikely that altered membrane effects in vivo persist to alter activity beyond the “optimal” conditions in vitro. Moreover, by permeabilizing the membrane, we have also allowed measurement of the 3-O-MFPase activity throughout the cell and not just Na+-K+-ATPase regionalized to the plasma membranes. This is the first study to identify that an acute reduction in skeletal muscle Na+-K+-ATPase can occur with exercise in rats. These effects were observed after the Run protocol. Based on the results of the Run+ protocol, it would appear that the additional volume of exercise is a critical factor in inducing reductions in enzyme activity, at least under the conditions employed. One possible explanation for this might be that the mechanisms previously identified, in isolation or in combination, have a cumulative effect when exercise is continued, albeit at reduced intensity, in fatiguing muscles. The processing of free radicals to produce hydrogen peroxide, for example, may be what reduces skeletal muscle Na+-K+-ATPase in the same manner as is observed for cardiac Na+-K+-ATPase (31). Similarly, the effects of elevated Ca2+ may have a delayed time course because the causes of damage with this process can be attributed to the Ca2+-activated mobilization of calpain (5). Increased free Ca2+ has been presumed to activate a number of degradative cascades (5), and this effect may be related to the total amount of work done during an exercise bout (10). Similarly, heat denaturation due to elevated muscle temperature may also be cumulative. Because no changes in Na+-K+-ATPase activity were observed when the Run protocol was followed by passive recovery, continued contractile activity would appear essential in promoting the inhibitory effects on the Na+-K+ pump, regardless of the mechanism. This conclusion is based on the examination of two muscles, namely the WV and RV. It is possible that a different response could occur in Sol and EDL. In the experiment, we have also examined whether a muscle-specific response in Na+-K+-ATPase activity occurred with exercise. As indicated by a planned comparison, the response of EDL muscle was significantly different from that of the other muscles. EDL muscle exhibits considerably faster fatigue than Sol (15), and this is attributed, at least in terms of muscle excitability, to a greater number of Na+-channels relative to Na+-K+ pumps (24). This increased ion “leak” relative to ion “pump” capacity can result in a run down of ion gradients and result in fatigue (35). This fact may explain why EDL Na+-K+-ATPase activity was unaffected in the Run protocol compared with other muscles, possibly because of early fatigue and a lack of involvement in the treadmill running. Glycogen content measurements showed that EDL was similarly depleted to other muscles, so EDL was used either early or late in the Run protocol and/or was continued once the active recovery phase began. It is clear that additional exercise was needed for depressions in Na+K+-ATPase activity to occur in EDL in the Run+ protocol. A final noteworthy issue when results of this experiment are interpreted is whether Na+-K+-ATPase activity was measured in homogenates by using the 3-O-MFPase assay. K+-dependent hydrolysis of the 3-O-methylfluorescein phosphate chromogenic substrate substitutes for the aspartylphosphate intermediate of ATPase (25) to represent the terminal step in ATP hydrolysis (26). These phosphatase assays relate to actual ATPase activity (25, 26) but yield results that are less than the activity assessed by direct methods, possibly because of a reduced affinity for the artificial substrates. Although a direct measurement of Na+-K+-ATPase activity using a measurement of inorganic phosphate accumulation or a regenerating assay involving NAD/NADH changes would have been more desirable, such measurements on whole muscle homogenates are problematic given the low relative Na+-K+-ATPase activity in the cell relative to other ATPases. The observations from this experiment are consistent with previous findings (34) for cardiac muscle Na+-K+-ATPase and confirm a preliminary report that activity is reduced after repeated submaximal leg extension exercise to fatigue in skeletal muscle. These findings reintroduce the possibility that Na+-K+-ATPase may contribute to acute neuromuscular fatigue in skeletal muscle and highlight the role of the Na+-K+-ATPase as a key control site for regulating the electrical signal reaching the contractile apparatus. Confirmation of these findings and elucidation of the mechanisms for the reduced activity with exercise are avenues for future investigation. In the present study, we have shown that prolonged exercise to fatigue results in a trend toward a reduction in Na+-K+-ATPase activity as measured by the 3-O-MFPase reaction in all muscles examined except the EDL. Additional exercise of lower intensity was noted to associate with further reductions and to demonstrate significant change. Based on the findings of experiments using inactive recovery, it appears that the additional changes in 3-O-MFPase were due to exercise per se. It is unclear from the current work whether different forms of exercise could have resulted in more pronounced reductions in 3-O-MFPase and whether the changes are specific to muscles of different fiber-type composition. Yet to be determined, as well, is the physiological significance of the alterations in the 3-O-MFPase that were observed. Although it is tempting to speculate that a predisposition to failure in membrane excitation may result, this hypothesis needs to be investigated. In addition, the mechanisms associated with the reduction in 3-O-MFPase activity remain unknown. Future experiments should address the site on the enzyme (i.e, nucleotide binding, K+-Na+binding) that is subject to structural alterations and the intracellular stimulus responsible for inducing the change. This study was supported by the National Sciences and Engineering Research Council of Canada. FOOTNOTESREFERENCES
Page 12physicians have knownfor a long time that atelectasis plays a pivotal role in the development of hypoxemia during mechanical ventilation, both in general anesthesia for routine surgery and in various instances of acute lung injury. Recent practice has now moved toward the use of lower tidal volumes (Vt) in patients with acute lung injury and acute respiratory distress syndrome (ARDS) after two important studies demonstrated lower mortality outcomes by using this strategy (1, 1a). However, because of the progressive alveolar derecruitment that is often encountered with the use of lower Vt(27), there has been a renewed interest in the use of recruitment maneuvers (RMs) as a means of reopening these regions of atelectasis. Many studies have demonstrated improvement in oxygenation, respiratory system compliance, and even lung volume after sustained or deep inflations (DIs) in many different scenarios (3-5, 17,18, 26, 28), but some of these same studies were unable to show any lasting improvement when RMs were delivered during conventional positive pressure ventilation (4, 18, 25). These apparently conflicting results may be explained by the different experimental situations involved, as the effect of a RM is likely to depend on the size and duration of the inflation, the mode of ventilation in which the inflation is delivered, and the level of positive end-expiratory pressure (PEEP) employed. Furthermore, if the effects of a RM are transient, then its apparent benefits will depend on when they are measured. Different researchers have measured the effects of a RM at different time points after its administration (3, 4, 17, 26, 28), yet there has been little attempt to characterize the rate at which the benefits of a RM are lost in the period immediately after its delivery. The goal of our study, therefore, was to characterize the effects of a DI on lung mechanics during conventional mechanical ventilation in a mouse model of lung injury. We followed a measure of respiratory elastance (H) as an index of ongoing derecruitment after DI, both at baseline and after saline lavage at various levels of PEEP. We hypothesized that elastance would be decreased transiently by DI and that the rate and magnitude of subsequent recovery to its resting level would be increased in lung injury. A secondary hypothesis was that PEEP would retard the rate of recovery in elastance after DI. We also speculated that this influence of PEEP on the recovery of elastance would be more significant after lavage because an injured lung will be more prone to atelectasis. METHODSWe studied thirteen 8- to 9-wk-old BALB/c female mice (Jackson Laboratories, Bar Harbor, ME) weighing 19.8 ± 0.8 g. Each mouse was anesthetized with pentobarbital sodium via intraperitoneal injection at a dose of 90 mg/kg and then underwent tracheostomy with a secured 18-gauge metal cannula, and then connected to aflexiVent (SCIREC, Montreal, Canada) computer-controlled small animal ventilator. The mice were ventilated in a quasi-sinusoidal fashion at a rate of 200 breaths/min. Cylinder piston displacement was set at 0.25 ml, which resulted in a Vt of 0.20 ml (∼10 ml/kg) when gas compression was accounted for. PEEP was controlled by submerging the expiratory limb from the ventilator into a water trap. The animals were allowed 5 min to adjust to the ventilator at a PEEP of 3 cmH2O and were then paralyzed with an intraperitoneal injection of pancuronium bromide (0.5 ml/kg). To ensure adequate anesthesia, heart rate was monitored via continuous pulse plethysmography that was measured transcutaneously across the femoral crease. Halfway through the protocol, an additional dose of pentobarbital sodium (30 mg/kg) was administered for maintenance of deep anesthesia. After the initial stabilization period, the level of PEEP was set at 1 cmH2O, and two DIs were delivered at constant flow with a pressure limit of 25 cmH2O. Each DI lasted 2 s. The mouse was then returned to quasi-sinusoidal ventilation at 200 breaths/min with a Vt of 0.20 ml. Respiratory system input impedance (Zrs) was measured via a forced-oscillation technique (described in Data Analysis below) 4 s after the two DIs, then subsequently every 15 s for 5 min, and then every 30 s for an additional 2 min. The entire protocol was timed by a computer and repeated at a PEEP of 3 and 6 cmH2O. The order of PEEP was not randomized to minimize the possible damaging effects of the higher PEEP levels on measurements at lower PEEP. After completion of the protocol at each of the three different levels of PEEP, the mouse was disconnected from the ventilator, and 0.5 ml of PBS was instilled via the tracheal cannula and slowly suctioned back for a return of ∼0.3 ml in every animal. The mouse was then allowed 5 min to restabilize on the ventilator, and the same protocol was repeated at a PEEP level of 3, 1, and 6 cmH2O. Again, PEEP levels were not randomized, and this time a PEEP of 3 cmH2O was used first because the animals tended to do poorly at 1 cmH2O. Zrs was determined by measuring piston volume displacement and pressure in the ventilator cylinder while delivering 2-s oscillatory volume perturbations to the airway opening. These perturbations were composed of 13 superimposed sine waves of varying amplitude and frequency, ranging from 1 to 20.5 Hz. The frequencies were set at mutually primed values to reduce harmonic distortion that can occur in nonlinear systems (10). Before beginning the protocol, we obtained dynamic calibration signals necessary to correct for the physical characteristics of the flexiVent in subsequent measurements of Zrs (13, 32). Zrs itself was determined via Fourier transformation of the signals of ventilator piston volume and cylinder pressure as described previously (9, 13). Zrs was interpreted by being fit with the model Zrs=Raw+i2πfIaw G−iH(2πf)αEquation 1 whereα=2π arctanHGEquation 2 The parameters Raw and Iaw largely characterize the resistive and inertive properties, respectively, of the airways, whereasG characterizes the dissipative properties of the lung tissues (10). The symbol f represents frequency, and i is the square root of −1. In the present study, we focus our attention on H, which is essentially the conventional elastance of the respiratory system. (Indeed, it is precisely equal to H at an oscillation frequency of 1/2π Hz.) We thus obtained a set of H values vs. time for 7 min after the DIs in each mouse at each level of PEEP. Each data set began immediately after the two DIs at the lowest value for H and then increased toward a plateau as time progressed. Under baseline conditions, the data for H over time could be satisfactorily fitted by a first-order exponential function, which gave the time constant (τ). After saline lavage, the curves for H over time could not be well fit to an exponential function, and an effective τ was determined as the time needed for H to achieve 67% of its total excursion over the 7-min measurement period. Values for τ were also calculated from the combined data at each level of PEEP, both before and after lavage. All statistical analyses were performed by using SAS statistical analysis software (ver. 8.1, SAS Institute, Cary, NC). Data sets were subjected to natural logarithmic transformation before comparing differences between pre- and postlavage conditions and different levels of PEEP. Repeated-measures ANOVA was used to examine the within-subject effects of PEEP and lavage on the values for τ, baseline and plateau values for H, and a total recovery in H after the DI. Differences were considered significant at a P value of <0.05. This was followed by Tukey's procedure to compare means at each level of PEEP, before and after lavage, to determine between-condition differences.RESULTSAfter lavage, three mice died and three others began to spontaneously breathe or twitch, most often when placed on PEEP of 1 cmH2O. This made their data uninterpretable and presumably occurred because the amount of lung collapse was greatest at a PEEP of 1 cmH2O. This left interpretable data for n= 13 mice at every level of PEEP before lavage and n = 12, 7, and 10 at PEEPs of 3, 1, and 6 cmH2O, respectively, after lavage. Grouped values for H vs. time after two DIs for all interpretable data sets, both before and after lavage, are shown in Figs. 1 and2. Values of τ for each averaged data set are listed in Figs. 1 and 2and show that elastance approached its plateau value substantially more quickly after saline lavage compared with control conditions. Fig. 1.Mean elastance (H; ±SE) over time at various levels of positive end-expiratory pressure (PEEP) before saline lavage. Values for the time constant (τ) for the compiled data at each level of PEEP are shown above the respective curves. DI, deep inflation. Fig. 2.Mean H (±SE) over time at various levels of PEEP (cmH2O) after saline lavage. Values for τ for the compiled data at each level of PEEP are shown above the respective curves. To avoid bias during statistical analysis, data from a mouse were discarded unless the mouse had a complete set of interpretable data at each level of PEEP, both before and after lavage. Furthermore, analysis of the remaining seven mice showed one animal with a value for τ at a PEEP of 3 cmH2O that was well beyond 2 standard deviations from the mean. We suspect this may have been due to spontaneous movement that we had not been previously aware of, so data from this animal were omitted, as well, leaving a final n value of 6 mice. This significantly reduced variation in τ at a PEEP of 3 cmH2O after lavage but did not change the overall results or conclusions from analysis using the initial n = 7 mice. Saline lavage exhibited a significant effect on the values for τ overall (P = 0.0001), but no similar effect was demonstrated by the level of applied PEEP (P = 0.16). When data from individual animals were compared, no significant differences in τ were observed between any two levels of PEEP before or after lavage (Fig. 3). Baseline level of H was significantly increased by saline lavage (P = 0.0003) and significantly reduced by higher PEEP (P < 0.0001). When mean baseline levels ofH from each level of PEEP were compared, the only nonsignificant difference was between a PEEP of 1 and 3 cmH2O before lavage (P = 0.31; Fig.4). The plateau level of H was also significantly increased by lavage and reduced by higher PEEP (P < 0.0001). When separate levels of PEEP were compared, the only nonsignificant difference was again found between a PEEP of 1 and 3 cmH2O before lavage (P = 0.26; Fig. 5). Total recovery inH (difference between plateau and baseline) was significantly increased by lavage and reduced by higher PEEP (P < 0.0001). However, when separate levels of PEEP were compared, the total recovery in H was only significantly reduced by a PEEP of 6 cmH2O, both at baseline (PEEP 6 vs. 1 cmH2O, P < 0.05) and after lavage (PEEP 6 vs.1 and 3 cmH2O,P < 0.0001) (Fig. 6). Fig. 3.Mean τ (±SE) for the rise in H after 2 DIs at each level of PEEP (cmH2O), both before and after saline lavage. Significant difference in values taken before and after lavage: * P < 0.05, ** P < 0.01. Fig. 4.Mean H (±SE) immediately after two DIs at each level of PEEP (cmH2O) before and after saline lavage. NS, not significant. Significant difference: ** P < 0.01, *** P < 0.001. Fig. 5.Mean plateau level of H (±SE) after recovery after two DIs at each level of PEEP (cmH2O) before and after saline lavage. Significant difference: * P < 0.05, ** P < 0.01, *** P < 0.001. Fig. 6.Mean total recovery in H (±SE) after two DIs at each level of PEEP (cmH2O) before and after saline lavage. Significant difference: * P < 0.05, ** P < 0.01, *** P < 0.001. DISCUSSIONRecent clinical trials showing a mortality benefit from low Vt ventilation (1a, 2) have refocused investigation toward the prevention and treatment of atelectasis, a recognized consequence of this ventilation strategy (27). Studies have examined the peak pressure and duration of a DI required to completely overcome the atelectasis encountered in normal subjects during general anesthesia (29, 30). Still other studies utilizing computed tomography of the chest have demonstrated lasting improvements after RMs, again in healthy human subjects during general anesthesia (31). However, recruitment of collapsed regions has not been as well characterized in the injured lung. In particular, the time course of the benefits after a DI has received little attention but clearly has great bearing on the efficacy of any given RM regime. Our results illustrate that the effect of a DI on His transient, even in a normal lung, and can become fleeting in lung injury. The transience of the improvement in H derived from a DI in this study suggests that for any significant effect to be achieved through RMs, DI would need to be delivered frequently to maintain a significant fraction of the initial degree of open lung that existed immediately after its delivery. We can explore this issue quantitatively as follows. H is a measure of stiffness of the respiratory system; so at a given level of PEEP, we assume that it is inversely proportional to the amount of open lung. BecauseH is a function of time, we can describe the time course of the amount of open lung (O) relative to that immediately after the DI at time = 0 as O(t)=H(0)H(t)Equation 3 where t is time. If we give a DI every Tseconds, then O(t) is a periodic function with each period beginning at a value of 1 and reachingH(0)/H(T) at timeT. The initial value of 1 is a relative value describing the maximum amount of open lung achieved after a DI at a given level of PEEP. The mean value of O(t), when DIs are given every T seconds, is equal to its mean value over a single period of T seconds, which isŌ(T)=1T ∫0T H(0)H(t) dtEquation 4 Fig. 7.Mean open lung (Ō) plotted as a fraction against the required time interval between DIs for both control and saline-lavaged mice at each level of PEEP (cmH2O). Of course, the above considerations only apply to the brief DIs that we delivered to our mice. There is evidence that the majority of recruitment can occur as early as 1–2 s into an inhalation during volume-controlled ventilation of an injured lung (24). However, it is also possible that a more sustained DI may have produced a greater effect on H. Traditionally in human subjects, DIs are pressure limited at a range between 30 and 40 cmH2O and, if sustained, are held for 15–40 s (1a, 29–31). Nevertheless, nonsustained DIs have been shown before to provide benefit (26), and the nonsustained DIs in our study did reduce H, albeit transiently, even after a lavage. PEEP also has the potential to modulate the effect of a DI by preventing at least some parts of the lung from ever falling below their critical closing pressures (21, 28). The role of the RM is then to reach the higher critical opening pressures for segments not opened during normal tidal inflations (6, 23). This means that the amount of recruitable lung decreases with increasing PEEP (5, 17, 27). This was clearly the case in the present study after lung injury because we found that increasing PEEP markedly reduced baseline H (Fig. 4), its plateau level (Fig. 5), and particularly its total recovery (Fig. 6). This effect of PEEP achieved greater statistical significance after saline lavage (Figs. 4-6), implying that PEEP has more measurable effects when the lung is injured and prone to collapse. We initially hypothesized that increasing PEEP would slow down the closure process in general, thereby lengthening the benefits of a DI. Interestingly, this was not the case, either at baseline or after lavage (Fig. 3). This finding contrasts with that of a much earlier study in lambs and rabbits (35), although a more recent investigation using computer tomography scanning in injured pig lungs (24) found no significant differences between values for τ at different levels of PEEP. However, we are not able to conclude that closure in the lung is a pressure-independent process just because τ does not depend on PEEP. Because total recovery inH that follows a DI is smaller at higher levels of PEEP (Fig. 6), it could be that this phenomenon, when combined with lengthened closure times, results in a value for τ that is relatively unchanged by increasing PEEP. If closure in the lung is as strongly time dependent as our results indicate, it is also likely that the benefit derived from a DI will depend on the volume history of the lung before the point when the DI is given. This suggests that the effects of a DI may depend to some extent on the mode of ventilation being utilized. Indeed, studies have demonstrated reductions in lung compliance and increased lung volumes after RMs delivered during high-frequency ventilation (3, 4,18) but not during conventional positive pressure ventilation (4, 18). In addition, sustained DIs have been shown to result in greater recruited lung volumes when using low Vtor low levels of PEEP but not with high Vt or high PEEP (17). Nevertheless, despite using conventional positive pressure ventilation with a moderate Vt in our study, we were still able to demonstrate an improvement in H after a DI both at baseline and after lavage. Our positive observations were perhaps due in part to the fact that we made frequent measurements of mechanics with a very sensitive technique. Had we measured Hless frequently, beginning some minutes after DI, we might have missed observing any effect. For example, Pelosi et al. (26) were able to demonstrate immediate improvement in end-expiratory lung volume and elastance in ARDS patients during a frequent-sigh period of mechanical ventilation. However, their earliest follow-up measurements were made 30 min after sigh interruption, at which time these effect were lost. In another study involving saline-lavaged dogs, improvement in the arterial partial pressure of oxygen seen after a sustained DI was maintained at 15 min, but only in those animals ventilated with the lowest Vt and PEEP (17). This raises the point of how to define optimal PEEP. Conventionally, PEEP is chosen as the pressure at or just above the lower inflection point (LIP) of the inspiratory limb of the quasi-static pressure-volume curve (1a, 26, 27). LIP is generally thought to correspond to the major recruitment events occurring during inflation. However, several studies have shown evidence of recruitment beyond LIP (21, 23,27). The concept of ongoing recruitment throughout lung inflation has also been supported by mathematical models (12), human studies utilizing computer tomography of the chest (8), and animal models (33). Thus controversy already exists over how recruitment is best served by PEEP. The results of the present study raise another issue that complicates the question of optimal PEEP even further. That is, if recruitment depends on time, then the position of LIP during a quasi-static stepwise inflation of the lung may have little relevance for the recruitment that takes place during a normal, mechanically ventilated breath. This means that optimizing PEEP may need to be based on more than a static view of the lung and require a consideration of the dynamics of opening and closing of air spaces. Our study thus makes the case for recruitment and derecruitment in the lung being dynamic phenomena. This clearly has implications for optimizing mechanical ventilation. However, there clearly are anatomic differences that limit extrapolation of these results directly to humans. For example, there is obviously a significant size difference between humans and mice. In humans, gravity has a significant effect on regional distribution of critical opening pressures and recruitment from PEEP (6). In the mouse, regional differences throughout the lung due to gravity are presumably negligible. It has also been suggested that smaller animals may have a more rapid rate of loss in compliance due to the smaller surface area of their alveoli and subsequent rapid change in the proportional air-space configuration (35). Indeed, previous human studies suggest that the benefits of sustained inflations can persist as long as 40 min in normal subjects under general anesthesia (31) but not without some degree of decay over time. Another possible limitation is that we were unable to randomize the order in which the different levels of PEEP were delivered in our study. This was because our mice frequently did not tolerate being placed on a PEEP of 1 cmH2O immediately after lavage. The likelihood that order played a significant role was, in part, diminished by starting each protocol with two DIs and essentially beginning with a maximally inflated lung at each level of PEEP. Furthermore, in some mice, PEEP was returned from 6 to 3 cmH2O at the end of the experiment (data not shown), and the resulting H vs. time data were similar to those obtained at the initial PEEP of 3 cmH2O. Another possible limitation to our study is not having measured oxygenation. However, previous studies have demonstrated a significant correlation between RM-induced improvements in compliance and oxygenation (26, 34). It should also be noted that, in our study, we experimented with only a single type of lung injury, whereas it has been shown that RMs have different effects in different types of lung injury (7, 17, 18,26). Extrapulmonary forms of ARDS involve greater degrees of edema and alveolar collapse and are thus more amenable to recruitment than direct lung injury, which is characterized more by air-space consolidation (7, 26). We suspect that our model more resembles extrapulmonary ARDS since saline lavage is likely to predominantly produce surfactant depletion and atelectasis (19,22). Other models of lung injury that more closely mimic pneumonia and consolidation may behave differently. Increases in end-expiratory lung volumes seen with higher PEEP in the latter scenario could potentially be less due to alveolar recruitment and more due to overdistension of the less consolidated and more compliant regions. This could potentially result in higher, instead of lower,H with increasing PEEP (7, 34). The lower values for H observed with increasing PEEP in our study argue more for improved alveolar recruitment than alveolar overdistension. If this protocol were to be repeated in a different model of lung injury, such as one more consistent with pneumonia, the findings could be significantly different. Furthermore, measuring lung volumes in both types of lung injury at various levels of PEEP could help to better clarify the question of recruitment vs. overdistension. Another key point about our discussion thus far is that we have assumed that the transients we observed in H after the DIs were entirely due to progressive derecruitment of the lung. However, the mechanisms responsible for the long-term transients in Hseen during mechanical ventilation remain controversial. For example, one possible explanation for the observed recovery in H over time under baseline conditions could be stress adaptation within the tissues (15). The difference in shape between curves derived under baseline conditions and those after lavage suggest possible differences in the mechanisms at play, with derecruitment likely playing a greater role in the injured lung. Because mechanical stretch is known to stimulate secretion of surfactant by type II alveolar cells (37) and may lead to more effective dispersion of existing surfactant, an alternative mechanism is the gradual loss of surfactant from the air-liquid interface. Williams et al. (35) ascribed slow changes in H entirely to this latter mechanism, whereas Horie and Hildebrandt (14) proposed that a combination of surfactant dynamics together with airway closure at low lung volumes could be responsible. With the recent interest in optimal ventilation of critically ill patients, the attention of the scientific community seems to have focused on recruitment/derecruitment as an important process in the lung (12, 16, 18, 21, 23, 27, 28), particularly to describe the greatly increased hysteresis of the quasi-static pressure-volume curve in lung injury (18, 28). Even so, Wilson et al. (36) have just proposed a model that ascribes the knee in the pressure-volume curve from edematous lungs to the nonlinear coupled behavior of the parenchyma and alveolar fluid, which does not invoke recruitment/derecruitment at all. The truth may well involve contributions from all these mechanisms. Nevertheless, after lung injury, the rate and magnitude of increase in H are so dramatic after a DI that it is hard to imagine that derecruitment is not the principal process involved. Certainly, lung volume is significantly reduced by lavage at any given inflation pressure, as our laboratory has recently shown (20) by using body-box plethysmography in saline-lavaged mice. In summary, we have shown that DI in mice produces a transient decrease in H that returns rapidly toward its pre-DI value, with a rate of return that is greatly accelerated in lavage-induced lung injury. When we equate increases in H with decreases in the amount of open lung, we predict that DIs would have to be delivered several times per minute to keep an injured lung significantly more open than would be the case if no RMs were given at all. We conclude that higher levels of PEEP can significantly reduce the baseline, recovery, and final plateau of H after a DI, especially in an injured lung. However, added PEEP does not significantly reduce the rate of recovery in H after a DI. This means that PEEP does not reduce the frequency with which DIs need to be delivered in this model of lung injury. Statistical analysis was performed by Dr. Janice Yanushka Bunn, Research Assistant Professor, Medical Biostatistics, University of Vermont. FOOTNOTESREFERENCES
Page 13magnetic resonance imaging (MRI) can be applied on local skeletal muscle regions by analyzing variables pixel by pixel or within multiple local regions, including active as well as less active muscles (31). Therefore, simultaneous evaluation of water exchange in active and inactive muscles during dynamic exercise is valuable to elucidate the signal shift in muscle that is related to exercise and is beneficial to study how different MRI techniques that focus on water displacement relate. The physics of water exchange in the skeletal muscle is highly complex and multifactorial. The magnetic resonance signal has been shown to be multiexponential, which indicates a multicompartmental origin.1H-MRI using transverse relaxativity focuses on the intrinsic property of water and its exchange between the different compartments as well as the binding capacity of the water molecule to subcellular structures (11). Most present studies have used a monoexponential transverse relaxativity analysis, although a multiexponential behavior has been described (24). The monoexponential proton transverse relaxation time (T2) of skeletal muscle is known to increase by exercise (6, 20). Increased intracellular water content (4, 18) and mechanisms related to aerobic capacity, e.g., net intramuscular accumulation of osmoles (22), are assumed to be the most important factors related to prolonged T2 with exercise. To what degree other factors such as H+, phosphocreatine, or water related to the microvasculature (5) would contribute to the altered T2 still has not been determined. The increased water content in exercising muscles is known to affect both extra- and intracellular volumes (6, 27, 28). However, it is presumed that extracellular water affects the T2 more than total tissue and intracellular water in resting skeletal muscle (19). With diffusion-weighted MRI and calculation of the mean apparent diffusion capacity (ADC), it is possible to estimate water motion related to small and random movements in the tissue. These are probably related to both extra- and intracellular compartments, although the size of the extracellular volume seems to be the most important component. In vitro experiments have shown that ADC decreased when cells swell (1). A decreased ADC has also been related to cell swelling and decreased extracellular volume in the brain (2, 9,31). Increased ADC, found in exercising skeletal muscles, is accordingly presumed to reflect increased water motion, predominantly in the extracellular compartment, but effects of cytoplasmic motions are unclear. Furthermore, a temperature-related increase in ADC by ∼2%/°C (15) needs to be considered as being due to thermal storage in active muscle. Neurohumoral activity during exercise is known to affect resting skeletal muscle with exposure of increased sympathetic nerve activity (21) and vasoconstriction (3). Although not extensively studied, there are indications that muscle volume could decrease in nonexercising muscles. Reduced cross-sectional area (CSA) of nonexercising muscles has been documented in conjunction with dynamic exercise associated with biking (23) and plantar flexion (16). However, evidence of a water shift in nonexercising muscles during exercise has not been detected with isotope techniques (27). It has not been determined whether the decreased volume of nonexercising muscles is a result of reduced tissue water content or mostly of reduced vascular filling as a consequence of vasoconstriction. The study was designed to simultaneously evaluate exercising and nonexercising skeletal muscles during graded dynamic exercise and, moreover, to study whether MRI was able to measure altered extravascular volumes and how a measure of water diffusion and transverse relaxativity relates to muscle bulk volume. MRI was used to measure calf CSA, regional muscle T2, and water diffusion measured as the mean ADC. The primary aim of this study was to define whether dynamic exercise with a small muscle mass could induce a water shift in inactive muscles. The second aim was to evaluate how the responses of muscle transverse relaxativity, diffusion, and volume are related to graded dynamic exercise and during recovery in both active as well as inactive muscles. METHODSSix healthy male students were included with a mean age of 25 yr (range 23–26 yr). The Ethics Committee of the Karolinska Institutet approved the study. A specially designed foot ergometer in a nonmagnetic material constructed for plantar flexion exercise was used (16). Subjects were familiarized with the exercise setup within 1 wk before exercise in the magnetic resonance imager by performing unilateral graded plantar flexion exercise with the right foot and keeping the contralateral left leg resting. The highest workload sustained for 9 min was tried out. Two subjects exercised at the highest workload of 12 kg, and four subjects exercised at 22 kg. Exercise in the magnetic resonance imager was thereafter performed on 2 different days. Exercise started with a warm-up period of 3 min at 4 kg and continued with an additional 9 min at the predetermined load. Subjects were scheduled to exercise randomly with low (4 kg) or high workload (12 or 22 kg) on the first day and the other workload on the next day. Two exercise setups with the same workload were performed with a 50-min interval between bouts (Fig. 1). T2 acquisition was repeatedly scanned during the 45 min after the first stop, and diffusion weight imaging (ADC) was scanned during the 15 min after the second stop. Fig. 1.Schematic presentation of the exercise setup with its recurrent imaging and exercise periods. ADC, apparent diffusion capacity; T2, transverse relaxation time; post-ex, postexercise. Before imaging, all subjects had been resting in a supine position for at least 30 min. Single-slice imaging was performed with a 1.5-T magnet at 63 MHz by using a birdcage quadrature head coil with a uniform field (Signa, General Electric, Milwaukee, WI). Transaxial imaging was performed at the largest CSA of the calf. Padding was placed under the knees in the birdcage to prevent changes in calves' position during scanning. However, the padding did not ensure an identical angle of the knees between imaging days. The diagonal of the measured slice could therefore be different on the 2 imaging days and probably affected the measured CSA at rest. Accordingly, the percent change from rest was used when comparing exercise levels achieved on different days. T2 images were obtained by a multiple spin-echo sequence [repetition time (TR)/echo time (TE): 1,500/15, 30, 45, 60 ms; slice thickness: 10 mm; field of view: 40 × 20 mm, matrix 256 × 128 mm, which gave the in-plane resolution of 1.6 × 3.1 mm2], with a 2-min interval between points. CSA was measured with manual planimetry on the image with a TE of 15 ms performed by one examiner without knowing the subject's name and date of the study. Signal intensity values were obtained within a region of interest (ROI) of ∼2.2 cm2 on the lateral portion of the gastrocnemius, excluding visible vessels and fascia. T2 was calculated with the least square method by fitting four echoes to monoexponential decay by using the equation (regression coefficient · 10−1) · 103 as T2. Diffusion-weighted spin echo echo-planar imaging using Stejskal-Tanner diffusion was applied with tetrahedral gradients (29) with special concern to achieve a high temporal resolution. Acquisition parameters were b value = 600 s/mm2, TR/TE = 4,000/63 ms, field of view = 40 × 40, matrix = 128 × 128, and slice thickness = 10 mm, with an interval between points of 0.33 min. ROI was chosen to be large, placed within the lower half of the calf, including gastrocnemius, soleus, and peroneus longus, because the signal-to-noise ratio was presumed to be low. The outer and central parts (vessels, fibula, and tibia) of the calf border were excluded. The diffusion coefficient was also calculated according to a monoexponential model. When the immediate postexercise value was presented, a mean of the three initial values within the first minute was used. The statistical significance was evaluated by t-test for independent and dependent samples, and one-, two-, and three-way ANOVA and Pearson's product-moment correlation with regression line and 95% confidence interval were graphically displayed. Values from the whole acquisition were used in the ANOVA analysis. Significant statistical level was considered at P < 0.05, and values were expressed as means ± SD. All analyses used Statistica 5.5 software (StatSoft, Tulsa, OK). RESULTSImmediate postexercise values are presented in Table1. CSA, T2, and ADC at rest did not differ significantly between the 2 days. However, CSA differed numerically by ∼3% in the active skeletal muscle between low and high workloads, probably because of different knee padding in the coil. With exercise, a progressively decreased CSA was found in the inactive skeletal muscle because CSA increased in the active leg, which caused a significant interaction between the legs. T2 did not change in either leg at low workload, but an interaction between the legs was found at high workload (P < 0.05) because T2 shortened by 3.1 ± 2.4% in the inactive leg and was prolonged by 15.5 ± 11.5% in the active leg. Motion artifacts at high workload probably affected some ADC values in one subject; therefore, all values in the inactive leg and the 12 last values in the active leg were excluded. ADC did not change significantly in the inactive leg at either workload but increased significant in the active leg by 7.1 ± 3.0% at low workload and by 12.5 ± 6.9% at high workload.
Both CSA and ADC, but not T2, increased at low workload. A strikingly highly positive correlation was found between CSA and T2 in the exercising calf (r = 0.94, P < 0.001; Fig. 2A), and a nearly significant correlation was found in the nonexercising calf (r = 0.54, P = 0.07). Moreover, ADC correlated highly with CSA (r = 0.86, P< 0.001); however, this was mostly related to the correlation within the active leg (r = 0.66, P < 0.05), but a decreased or unchanged CSA, as seen in the inactive leg, was related to a decreased diffusion in half of the subjects (Fig.2B). Furthermore, a correlation of ADC with T2 was found (r = 0.70, P < 0.001). Fig. 2.A: correlation plot of cross-sectional area (dCSA) and transverse relaxation time (T2) from resting values in the exercising calf at low and high exercise intensity. The regression line with a confidence interval of 95% is plotted. B: correlation plot of changed CSA (dCSA) from resting situation and changed ADC (dADC) by using extrapolated ADC values in the nonexercising leg (N-leg) after low workload as a resting reference. Values are from N-leg at high workload (●) with missing data from one subject. The regression line with a confidence interval of 95% is plotted. Both CSA and T2 had a significant three-way interaction with ANOVA regarding leg, load, and time. ADC had fewer interaction terms but presented significant changes between loads and time (Table 1). After low workload, there were only minor, nonsignificant changes of T2 and CSA in the inactive leg, probably because of a rather large variation within the group. Significant differences between loads were present during the initial 8 min for CSA and during the first 2 min for T2 (P < 0.05). The immediate postexercise ADCs in the inactive leg were not significantly changed from resting values (Table1). The recovery of CSA in the inactive leg after high workload was defined as an exponential growth (Fig. 3). The half-time (t1/2) could be calculated, when considering a monoexponential function, and equaled 24.2 ± 12.6 min. Likewise, the decay of CSA and transit time was considered monoexponential in the active leg after high workload with at1/2 of 21.9 ± 12.6 min (range 4.3–36.5 min) and 22.3 ± 12.3 min (range 9.8–40.8 min), respectively, with closely correlated recovery times (r= 0.80, P = 0.05). The slow recovery of ADC in both legs, especially in the exercising leg, probably explained the higher values in the active leg at rest before the second exercise bout. When calculating the 50% recovery of ADC in the active leg at high workload, the extrapolated late ADC value obtained 45 min after the first exercise bout but before the ADC exercise bout was used. This gave a t1/2 of 65 ± 36 min (range 27–116 min), a significantly longer recovery time than that of both CSA and T2 (P < 0.05), which was still significantly prolonged if the suspect initial temperature-related values were excluded. Fig. 3.Recovery of T2 and CSA in N-Leg after contralateral high workload. Values are means ± SD. Half-time of the mean is indicated by arrows. T2 = 10.7 min (left); CSA = 24.8 min (right). *Significant difference (P< 0.05). The effect from the prior exercise bout preceding the T2 acquisition was probably not fully normalized at the time when resting ADC values were obtained. ADCs in the inactive leg at low workload were therefore chosen to be the reference resting value for both legs when the percent change was described. However, there could be a slight overestimation of resting ADC values becauset1/2 was 172 ± 152 min in the inactive leg at low workload, but with low ADCs the error would be small. If ADC values in active leg during the initial 4-min postexercise recovery were excluded, the level was still elevated by ∼6% at low workload and by ∼10% at high workload (Fig. 4), with a mean difference between loads of 74% (range 45–99%) during the recovery period (Fig.5B). Because of the large variation of ADC values, no significances were found between workloads in the immediate postexercise value. However, a significant interaction was found between workloads (Table 1), which indicated a workload-dependent elevation of ADCs in the recovery period that was indicative of a response to graded exercise. Fig. 4.Plot of measured ADC (mean values) and least squares line fit during 15-min recovery in the exercising calf after low (○; dotted line) and high workload (□; solid line). The suspect effect of temperature at high workload is presumed to relate to the elevated level between the upper extrapolated ADC level (fine dotted line) and the least square line fit during the initial 5-min period, as indicated by the arrow. Fig. 5.Values are percent change from resting values and exclude the suspect initial 4-min temperature-dependent ADC period and use the extrapolated 45-min ADC value from the prior exercise. Mean values are plotted with distance-weighted least squares line fit. A: 50% time course recovery is marked by arrows: T2 (22.4 min, first filled arrow); CSA (25.7 min, open arrow); and ADC (49.5 min, second filled arrow). Significant interaction terms found by using two-way ANOVA are found between T2-CSA (***P < 0.001) and T2-ADC (*P < 0.05), but not CSA-ADC [not significant (ns)]. B: mean ADC values after low- (open symbols, dotted line) and high-intensity (filled symbols, solid line) exercise. E-leg, exercising leg. DISCUSSIONThis study presents a novel correlation of skeletal muscle T2 with ADC and how they are related to muscle bulk volume during recovery from exercise in both active and inactive muscles. Furthermore, evidence of extravascular water absorption from inactive muscles at contralateral high-intensity dynamic plantar flexion exercise was found. Absorption of extravascular water in inactive muscles was indicated by decreased muscle CSA and shortened monoexponential T2 compared with resting values with known, nonsignificantly changed blood flow to the nonexercising calf during high-intensity contralateral plantar flexion exercise (16). The nearly significant correlation between T2 and CSA in the inactive leg and a significant, although weak, negative correlation with T2 in the active leg (r = −0.66, P < 0.05) indicated absorption of water in parallel with contralateral exercise intensity and neurohumoral activation. Despite unchanged mean ADC in inactive muscles between workloads, there was a significant correlation between changes in ADC and CSA that indicated small or even decreased water motion with unchanged or decreased CSA (Fig. 2B). Arterial osmolality was presumably relatively low during this exercise with a small muscle group at high intensity with a bulk flow of ∼1,500 ml/min at the high workload (17). Despite a relatively high-intensity plantar flexion exercise, arterial hematocrit was left unchanged or only slightly increased by ∼5–8%, and arterial lacate was elevated by <1 mmol (our own preliminary data). Therefore, increased oncotic pressure was not considered a dominant cause of dehydration of inactive muscles (13). Although with a large enough active muscle mass and high-intensity exercise, a decreased arterial plasma volume with increased colloid osmotic pressure and concentrations of ions and lactate may contribute to dehydration with an osmotic effect (26). However, these variables were not controlled during this study, and their relative importance has not been determined. Therefore, water absorption from the extravascular space in inactive muscles was considered to be primarily caused by increased sympathoadrenal vasoconstriction but with an additive effect by increased oncotic pressure of an undetermined magnitude. Absorption probably involved both intra- and extracellular water to the same extent since supposedly mainly solute-free water was withdrawn. One reason for capillary absorption of water in inactive muscles could be to counteract the regionally decreased plasma volume in vessel beds that supply exercising muscles with a gain of tissue water. The need for fluid compensation in the circulatory system during heavy exercise is a known demand because the loss of fluid into exercising muscles is larger than the decrease in plasma volume. Therefore, fluid absorption from inactive tissues seems to be needed (13); a mechanism that includes inactive muscles is supported by this study. Furthermore, there are some interesting results related to exercising muscles and especially with reference to the diffusion of water that could give some additional insight to water exchange in skeletal muscle as a result of exercise. It is known that water flux can be considered to be driven by a concentration gradient described by Fick's law (particle flux = −diffusion coefficient × concentration gradient) and not necessarily by an osmotic gradient, but the Fick's law equation follows a result of random motion that is measured with diffusion-sensitive MRI and is calculated as ADC. It is not only the size of the interstitial space that may affect ADC, although it may dominate, but also the altered flux in the compartment. Therefore, lymphatic flow could also, if considered random, affect the random motion of water and measured ADC since its flow is presumed to be hundreds to a thousand times slower than resting perfusion in skeletal muscles (7). To what degree lymphatic flow could affect tissue-water motion and calculated ADC is not known but cannot so far be neglected. Moreover, increased temperature is known to elevate random water motion; therefore, temperature-dependent elevated ADC values are expected in exercising muscles. The initial postexercise effect of hyperemia can be neglected because of nonrandom high velocities in vessels and would not give significantly false elevated values with our pulse sequence. However, tissue motion due to pulsations could be a problem. We did not measure local muscle temperature per second, but the presumed increased muscle temperature of ∼1°C (Fig. 4) is within the range of previously reported findings (8, 12, 25). Furthermore, González-Alonzo et al. (8) showed that during 6 min of recovery, ∼60% of stored muscle temperature was released with an exponential decline. Therefore, increased water motion by temperature is likewise expected in our study, and maybe throughout recovery, and is overlaid on the effect from the extravascular compartments, predominantly the interstitium. The temperature-related effect on ADC after high-intensity exercise could after 5 min of recovery probably only explain an additive effect by <0.5%. Moreover, we did not have any signs of transferred heating of the nonexercising calf because ADC was not significantly elevated in the inactive leg. It is reasonable to suggest that with prolonged elevated temperatures in exercising muscles, transferred heating would be expected in nonexercising muscles as previously described after heavy knee extension (10). The obtained ADC values in the exercising calf need to be interpreted cautiously, especially because the values showed a large signal variation with extrapolated resting values only. However, the immediate postexercise ADC value in exercising muscles was during high-intensity exercise increased in accordance with a previous study (14), but a graded response could also be documented in this study. A workload-dependent elevation of ∼7 and 12.5% during low- and high-intensity exercise, respectively, was found, and thereafter, there was a slow decline during 15 min of recovery (Figs. 4and 5B). The increased ADC, like T2, correlated in active muscles to muscle volume, presumably primarily to oxidative metabolic rate (e.g., extravascular accumulation of osmoles), and, to a presumably lesser extent, to hydrostatic forces (31); there was no attempt in this study to discriminate their relative importance. Furthermore, the delayed recovery of ADC relative to T2 could affirm the presumed slower restitution of fluids within the interstitial space than it could the contribution of water and its binding property within the intracellular compartment (26,31), a presumption also supported by the numerically shortert1/2 of T2 than of CSA in the inactive calf (Fig. 3). Both Sjögaard (27) and Ward et al. (31) recognized a slower capillary to extravascular exchange than intracellular to the interstitium, although with a different outcome of the size of the interstitium during exercise. It is, nevertheless, obvious that, in our study, increased ADC was detected at the first midacquisition time, 15 s after cessation of high-intensity exercise, and was affected by temperature as well as extravasular volume. None of the evaluated parameters in the exercising calf was fully normalized after high workload, which indicated that extravascular edema was still present after 45 min of recovery. The restoration rate of displaced water could to some degree depend on the positioning of the leg with a slightly bent knee and the foot slightly below knee level and above central veins. The evaluated parameters were, unfortunately, neither obtained within the same ROI nor evaluated within the same demarcated individual muscle. Different activation patterns of the calf muscles among subjects could therefore conceal significant changes and correlations. However, it was previously shown that both gastrocnemius and soleus muscles are activated by using this exercise setup (17). Measured ADC including both muscle groups would therefore reflect an average diffusion of the calf during this plantar flexion exercise. CSA of the calf would likewise be an adequate measure of exercise-induced volume change since muscles other than gastrocnemius and soleus are not activated to any substantial degree. However, results from a previous study (17) showed that tibialis anterior in the nonexercising calf was activated during high-intensity contralateral exercise; it was probably activated unintentionally to stabilize the pelvis during exercise. Regional CSA of tibialis anterior increased by ∼15%, and the true volume reduction of inactive muscles was therefore underestimated most likely by ∼0.3% when calf CSA was measured (unpublished data from Ref. 16), and this is probably also applicable to this present study. It is concluded that a combined MRI approach was feasible and capable of measuring water fluxes related to skeletal muscle exercise, and partly distinguished different compartments, however, were overlaid because of a multicompartmental origin of the measured parameters. During high-intensity dynamic exercise with a small muscle group, there was a small, although clearly reduced, volume of the inactive calf. This extravascular water absorption was probably induced primarily by sympathoadrenal vasoconstriction in inactive muscles, although minor contribution by osmosis cannot be excluded. Water diffusion had, just as T2, a graded response to exercise and correlated highly with muscle volume, which is indicative of extravascular water accumulation linked to perfusion and is likely a combined effect dominated by metabolic activity as driving forces and to a lesser extent by hydrostatic pressure. Furthermore, our findings support a faster restitution of intracellular water content and its binding property within the cell than that of extracellular water. The authors thank Dan Greitz, Magnus Karlsson, and YordsÖsterman for contributions to the technical execution of the study and Russell S. Richardson for providing preliminary data on hemodynamics during plantar flexion exercise. FOOTNOTESREFERENCES
Page 14administration of aerosolized bronchoconstrictor agents, such as histamine and methacholine (MCh), is widely used to document nonallergic bronchial hyperresponsiveness. Despite the large amount of work devoted to the study of the airway constrictor response evoked by inhalation of these and other agents, the time course of recovery of airway resistance after induced bronchoconstriction has received much less attention (11, 14). Spontaneous recovery from intense bronchoconstriction induced by MCh inhalation has been shown to occur slowly; up to 3 h may be required for its completion (11). Besides the natural degradation of the agonist and its clearance by the bronchial circulation (47), additional factors may be of importance in determining the time course of recovery. For instance, bronchoconstriction causes reflex increases in ventilation (24), which, in turn, may reflexly cause airway smooth muscle relaxation (13). Studies in the anesthetized rabbit have shown that increases in the amplitude and rate of mechanical ventilation decrease airway narrowing provoked by intravenous MCh administration, a phenomenon likely due to effects of stretching of the airway smooth muscle on force generation (41). It is well established that the adaptive responses evoked by exercise also include prominent increases in cardiorespiratory activity (48) and bronchodilation (45), which are to be ascribed partly to the reflex and mechanical effects exerted by the exercise hyperpnea (45). However, additional exercise-specific bronchodilator mechanisms have also been implicated. Submaximal exercise is associated with increased sympathetic activity (43); thus bronchodilation could be related to the increases in plasma catecholamine concentrations that occur during exercise (4), as well as to the release of inhibitory nonadrenergic noncholinergic neurotransmitters that are colocalized with norepinephrine (NE) in sympathetic motor nerves (52). Noticeably, epinephrine (E) infusions producing plasma concentrations similar to those achieved during intense exercise have been shown to induce bronchodilation and protection against induced bronchoconstriction, both in patients with bronchial asthma (25, 38) and in normal subjects (38). Animal experiments indicated that signals conveyed by thin myelinated (group III) and unmyelinated (group IV) muscle afferents, which have been shown to participate in the reflex cardiorespiratory adjustments induced by exercise (21), also influence airway smooth muscle tone. In fact, an inhibitory bronchomotor reflex has been demonstrated in the anesthetized animal; a decrease in tracheal smooth muscle tension is produced when chemical stimulants of group III and IV muscle afferents are injected into the arterial supply of hindlimb muscles (22) or when isometric contractions of hindlimb muscles are evoked by electrical stimulation of the appropriate ventral roots (23, 28). Bronchodilator responses of similar intensity may also originate from receptors located in the carotid body in response to pressor stimuli, such as those that are commonly observed during isometric exercise (39). Previous human studies (17, 18) have shown that sustained static muscular exercise of forelimb muscles (handgrip) elicits pressor and ventilatory responses. Rhythmic (intermittent) handgrip has been shown to induce cardiovascular responses similar to those of sustained static handgrip (2), but the respiratory adjustments evoked by this form of exercise have been less extensively investigated (20). So far, no attempts have been made at evaluating the influence of handgrip on airway tone, neither in normal subjects nor in patients with airway diseases. We observed in preliminary trials that, in normal (n = 4) and asthmatic (n = 5) subjects with normal baseline airway tone, 3-min static handgrip bouts at 30% of the maximum voluntary contraction (MVC) (sHG30) caused in most of them changes in airway tone ranging from 0 to −45% of baseline. Because, in some instances, handgrip-induced bronchodilation may be difficult to appreciate, especially when control airway caliber is within the normal range, we preferred to evaluate asthmatic patients in whom bronchoconstriction of similar degree (approximately +100% of baseline) had been induced experimentally. This approach also allowed us to minimize the confounding effects of differences in baseline airway smooth muscle tone that may be present in patients with natural asthma. Therefore, the present experiments were undertaken to investigate the effects of static and rhythmic handgrip on the time course of recovery of airway resistance after bronchoconstriction induced by MCh inhalation in asthmatic patients. METHODSSeventeen asthmatic outpatients (10 men, and 7 women, median age 27.0 yr, range 20–36 yr) participated in the study. All patients were nonsmokers and had allergic seasonal or perennial asthma, i.e., the presence of at least one positive immediate skin reaction to a battery of common inhaled allergens. They were in a stable clinical and functional state and were not on maintenance treatment with oral bronchodilators or corticosteroids. None of them had suffered from recent (within 4 wk) respiratory infections. They did not participate in competitive sports or take part in training programs that could alter their catecholamine response to physical exertion (19). All had normal arterial blood pressure (<140/90) and no history of cardiovascular diseases. They had participated in previous asthma studies and were fully familiarized with the laboratory equipment and provocation procedures. Patients' baseline lung function data are summarized in Table 1. The study protocol adhered to the recommendations of the Declaration of Helsinki for Human Experimentation and was approved by the local ethics committee; informed consent was obtained from each participant.
We measured airway resistance with the interruption method (Rint). Airway resistance is the ratio of the pressure gradient between the alveoli and the airway opening to the airflow. Pressure and flow were measured by means of a portable device (MicroRint, Micro Medical, Rochester, UK) consisting of an interrupting valve, a screen pneumotachograph, and a pressure transducer connected to a palmtop dedicated computer. This method for assessing airway resistance is based on the assumption that an imperceptible, brief airway occlusion during tidal breathing results in a pressure vs. time function that can be measured at the airway opening and used to estimate the alveolar pressure at the moment of the occlusion. The screen pneumotachograph, located between the pressure sensor and interrupter shutter, measures flow during tidal breathing. The ratio of the pressure difference between the estimated alveolar pressure at the moment of airway occlusion and the open-circuit pressure before occlusion to the flow measured immediately before airway occlusion yields the resistance of the conducting airways (Ref. 1, see also for further references). A brief interruption of expiratory flow was obtained by means of a rotating elliptical shutter. This design of shutter maximizes completeness of occlusion with a minimum of friction and loss of rapidity of closure. Manufacturer's specifications of the shutter state that closure occurs within 5–6 ms, with an occlusion period of 100 ms. Interruption was at peak expiratory flow, the preprogrammed manufacturer's default setting. This should correspond, approximately, to midexpiratory tidal volume (Vt), thus minimizing the influence of breath-to-breath variations in lung volume and, hence, in airway resistance (33). Interruptions occurred after a random number of breathing cycles (one to four), to prevent patients from anticipating shutter closure. The screen of the palmtop computer connected to the interrupter head displays the flow pattern during tidal breathing between interruptions. After a set of interruptions has been completed, this display can be used to validate mouth pressure and time functions (10) and to show simple summary statistics. Measurements were undertaken with the patient breathing through a disposable flanged mouthpiece and barrier filter (MicroGard, Sensormedics, Yorba Linda, CA) positioned between the patient's mouth and measuring device. Filter resistance was preliminarily checked in five pieces randomly selected from the available stock by means of a rotameter and a pressure transducer. Mean (±SD) filter resistance value was 0.035 ± 0.006 kPa · l−1 · s. This value was similar to that incorporated in the software of the device. Rint measurements are likely to be influenced by face and neck positioning, and, therefore, an arm supporting the measuring device was used to ensure standardized positioning of the patient's head and mouthpiece on a horizontal plane, with the neck slightly extended (5). Rint measurements were performed with a technician supporting the patient's cheeks and pharynx to reduce cheek and upper airway compliance (5). Breathing pattern was recorded by means of the respiratory inductive plethysmograph (RIP; Respitrace, Non-invasive Monitoring System, Ardsley, NY). The device was calibrated according to the procedure devised by Sackner et al. (36); validity of calibration was evaluated both by analyzing RIP waveforms during an isovolume maneuver and by comparing changes of Vt amplitude and end-expiratory lung volume measured by spirometry with RIP values (36). The sum of rib cage and abdominal RIP signals, which closely reflects the Vt measured at the airway opening, was fed to an eight-channel chart recorder (HP 7758A, Hewlett Packard, Palo Alto, CA). On paper recordings, we measured, on a breath-by-breath basis, the Vt, the duration of inspiratory (Ti) and expiratory times, and the total duration of the respiratory cycle (Tt). The respiratory rate (f = 60/Tt), the respiratory drive (Vt/Ti), and the inspiratory minute ventilation (V˙i = Vt × f) were subsequently calculated. Augmented breaths, i.e., the breaths approximating or exceeding twice the control Vt observed during handgrip and control trials (CTs) were counted. The fractional end-tidal CO2 (FetCO2) was also continuously monitored (Normocap CD 102, Datex, Helsinki, Finland). Static and rhythmic handgrip were performed by using the dominant hand, which grasped an isometric dynamometer connected to a strain-gauge transducer. The transducer signal was continuously displayed on an oscilloscope positioned in front of the patients to help them to maintain the desired level of contraction. Mean arterial blood pressure (MAP) and heart rate (HR) were continuously monitored noninvasively by using finger photoplethysmography (Finapres BP Monitor 2300, Ohmeda, Englewood, CO) from the second phalanx of the middle finger of the nonexercising hand, as previously reported (17, 18). The MAP and HR signals were also displayed on the chart recorder. MCh inhalations were carried out with a technique similar to that employed in previous studies (16). In brief, patients inhaled doubling MCh concentrations ranging from 0.125 to 16.00 mg/ml through a DeVilbiss no. 646 nebulizer (DeVilbiss, Somerset, PA) driven by a constant airflow (8 l/min). Each challenge was preceded by inhalation of 0.9% saline as a control solution. Both saline and MCh were inhaled during tidal breathing for 2 min (but seeProtocol); a 5-min interval was allowed between each inhalation period. The bronchial response was assessed by measuring 8–10 Rint values 90 s after inhalation of either saline or each MCh concentration. The highest and lowest recorded Rint values were always discarded, as well as those whose pressure-time curves were not consistent with those reported in the literature, e.g., showing a negative gradient after shutter closure, indicating artifacts due to air leakage (10). The remaining six to eight Rint values had a coefficient of variation <10%, and the mean value was used for data analysis (6). MCh inhalation was discontinued when Rint values increased by 100 ± 5% of the corresponding postsaline value. Samples of venous blood for catecholamine measurements were drawn from the cannulated large forearm vein of the nonexercising forearm. In all instances, an initial 3-ml sample of blood was discarded; 10-ml samples of blood were drawn into vacutainers containing 100 μl of a solution of glutathione (60 mg/ml) and EGTA (90 mg/ml). Blood samples were put immediately in an ice-water bath, centrifuged at 4°C, and stored at −70°C until assayed. Plasma E and NE concentrations were measured in duplicate and in the same assay for each patient by a radioenzymatic method previously described (30), using a CAT-A-Kit (Amersham, Buckinghamshire, UK) and according to the basic principles of Passon and Peuler (34). The sensitivity of the method was 20 and 15 pg/ml for NE and E, respectively; the between-assay variability was 10 and 7.5% for NE and E, respectively. Patients attended the laboratory on three separate occasions separated by an interval of ∼48 h. They were requested to abstain from all medications for the whole duration of the study and from caffeine-containing food for at least 12 h before each study day. After positioning of the RIP belts, patients were comfortably seated on a dentist's chair and were requested to relax fully with their head positioned in a slightly extended position. To this purpose, a soft cushion was placed between the patient's neck and the head rest. Patients were then connected to the airway resistance measuring device and were allowed to familiarize themselves with the equipment for 4–6 min. Subsequently, baseline measurements of Rint and cardiorespiratory variables were obtained during 3 min of tidal, regular breathing, as judged by inspection of the RIP signals. Care was taken to ensure that patients' baseline Rint values did not significantly vary between each study day. If Rint values turned out to be outside the accepted variability range (±15% compared with the first study day), the experimental session was rescheduled. If baseline Rint values proved to be within the accepted variability range, a 18- or 20-gauge Teflon catheter-over-needle intravenous line was inserted into a large forearm vein after local anesthesia, with injection of 1% lidocaine to the skin. This intravenous catheter was attached to a three-way stopcock apparatus, with one port available for blood sampling and the other connected to a 0.9% saline solution to keep the vein patent. After insertion of the venous line, each patient was allowed to relax for at least 30 min. A blood sample for baseline assessment of baseline catecholamine concentration was subsequently collected. Patients were then administered doubling MCh concentrations, preceded by inhalation of 0.9% saline as a control solution, while cardiorespiratory variables were continuously monitored. Changes in Rint were reassessed 90 s after inhalation of saline and each MCh concentration. To obtain, in each patient, a Rint increase as close as possible to +100% of the corresponding postsaline value, the last MCh administration could be tapered by reducing the inhalation time. Once the desired degree of bronchoconstriction had been achieved, an additional blood sample for measuring catecholamine plasma level was obtained. Then patients were randomly requested to rest for 3 min (CTs) or to perform either a 3-min sHG30 or a 3-min rhythmic (twelve 2-s contractions repeated every 3 s) handgrip at 60% of their MVC (rHG60). MVC values were determined as the mean value (coefficient of variation <5%) of the peak force developed during three consecutive MVCs lasting at least 3 s and performed at 5-min intervals. These handgrip paradigms producing the same total work output (9) were selected because, in previous studies (17, 18) and/or in preliminary trials, they were found to induce relatively intense cardiorespiratory responses and to be tolerated by the subjects without pain or discomfort. Changes in Rint values and in cardiovascular and respiratory variables were reassessed, at 1-min intervals, for 30 min after the completion of each 3-min period of rest or exercise. Additional blood samples for catecholamine assays were obtained 1 and 12 min after rest or exercise. To analyze changes in breathing pattern variables observed during MCh inhalation, CTs, sHG30, and rHG60 bouts, all recorded breaths were considered, except for those showing manifest alterations, as well as those immediately preceding or following sighs and short apneas. Changes in respiratory variables and Rint values induced by inhalation of each MCh concentration were expressed as a percentage of the corresponding postsaline (control) value; comparisons were made by the nonparametric analysis of variance for repeated measures followed by Dunn's test for multiple comparisons. Changes in respiratory variables observed after CTs, sHG30, and rHG60 bouts were expressed as a percentage of the corresponding post-MCh values and similarly compared. In most patients, the increased between-breaths variability during exercise did not allow us to record a sufficient number of reproducible Rint values and to provide reliable estimates of changes in airway resistance during exercise. Nonparametric analysis of variance and Dunn's posttests were also used to compare the magnitude of the early Rint changes observed 1 and 3 min after the completion of each motor task with those correspondingly recorded after CTs. To assess differences in the extent of the overall bronchodilator response and respiratory changes attained after CTs, sHG30, and rHG60 runs, values recorded at all scheduled time intervals during each 30-min observation period were averaged and compared by using the same statistical procedure described above. Changes in FetCO2 values and the differences in the number of augmented breaths observed in the three experimental conditions were analyzed by means of the one-way analysis of variance and Tukey's test for multiple comparisons. The same statistical procedure was used to compare changes in HR and MAP values recorded at baseline, at the predetermined level of MCh-induced bronchoconstriction, at 1-min intervals during both static and rhythmic handgrip trials, and during the first 2 min of the recovery period. Comparisons of plasma NE and E concentrations obtained at each scheduled interval were similarly performed. All reported values are means ± SE; in all instances, P < 0.05 was taken as significant. RESULTSAfter MCh inhalation, mean percent increases in postsaline Rint attained before CTs and sHG30 and rHG60 runs were 102.0 ± 1.8, 103.7 ± 1.7, and 103.9 ± 2.1%, respectively. No significant differences were found among these values. The time course of mean percent changes in Rint observed after the completion of CTs as well as of sHG30 and rHG60 bouts is depicted in Fig.1. Within 1 min after sHG30 and rHG60 runs, mean Rint values significantly decreased to similar extents (60.3 ± 4.0% after sHG30, and 65.3 ± 2.3% after rHG60) of the respective post-MCh values (P always <0.01). Three minutes after the completion of sHG30 and rHG60 runs, mean Rint values consistently displayed moderate increases up to levels corresponding to 74.1.1 ± 4.3 and 77.9 ± 2.9% of the respective post-MCh value (Fig. 1). In all experimental conditions, Rint values recorded from the 3rd min to the completion of the observation period proved to decrease in an approximately linear fashion, although those recorded after both types of handgrip consistently displayed levels lower than those recorded after CTs (Fig. 1). Fig. 1.Time course (30 min) of changes in airway resistance measured with the interrupter technique (Rint) after control trials (CTs) consisting of 3 min of rest (solid circles), static handgrip at 30% (sHG30; open circles), and rhythmic handgrip at 60% (rHG60; shaded circles) of maximum voluntary contraction. Changes in Rint are expressed as mean (±SE) percent variation of postsaline (control) values. Increases in baseline Rint observed at maximum methacholine (MCh)-induced bronchoconstriction (BC) are indicated at point BC on the time axis. Compared with post-MCh values, the overall (30-min periods) mean percent reductions in Rint after CTs and sHG30 and rHG60 runs were −13.6 ± 1.1, −36.1 ± 1.7, and −32.7 ± 0.7%, respectively (Fig. 2). Analysis of variance revealed that both sHG30 and rHG60 bouts caused more marked mean percent Rint decreases than CTs (P always <0.01), and that the reduction in Rint observed after sHG30 did not significantly differ from that attained after rHG60 (Fig. 2). Fig. 2.Overall mean (±SE) percent reductions (n= 17) in Rint after CTs, sHG30, and rHG60. * P < 0.01 compared with CTs. On each study day, mean baseline V˙i values (Fig.3A) were similar and corresponded to 9.25 ± 0.70, 9.95 ± 0.75, and 9.50 ± 0.74 l/min, respectively, on CTs and on sHG30 and rHG60 study days. In all experimental conditions (Fig. 3A), when Rint had augmented by ∼100% after MCh inhalation, mean V˙isignificantly rose from the corresponding control value (Palways <0.01). The increase in V˙i could be achieved, in some occasions, by prevailing increases in Vt or, more frequently, by increases in both Vt and f. In all instances, Vt/Ti also increased significantly (P < 0.01). Mean control FetCO2 diminished from 5.9 ± 0.1 to 4.6 ± 0.1% (P < 0.01). During both sHG30 and rHG60 bouts, V˙i displayed further, slight increases, mainly due to additional rises in Vt; these handgrip-induced increases in V˙i did not reach the level of statistical significance and did not cause additional changes in FetCO2. After exercise, overall mean (30-min periods) V˙i and Vt/Ti values remained significantly higher (P always <0.05) than the corresponding postsaline values. No differences were observed between changes in the pattern of breathing recorded after sHG30 and rHG60 runs. The number of augmented breaths was 2.0 ± 0.6 (range 0–8), 2.8 ± 0.7 (range 0–8), and 2.42 ± 0.6 (range 0–6) during CTs, sHG30, and rHG60, respectively; no significant differences were found among these values. Fig. 3.Mean (±SE) changes (n = 17) in control (C) minute ventilation (A), heart rate (B), and mean arterial pressure (C) observed at maximum BC during CTs (solid circles), 3-min bouts of sHG30 (open circles) and rHG60 (shaded circles) and during the early and late stages of the 30-min observation period. For clarity, data points were slightly shifted along the time axis. § P < 0.05 and * P < 0.01 compared with the corresponding BC values. Values of cardiorespiratory variables observed during 3-min handgrip bouts were significantly higher (P < 0.01) than those observed in CTs. On each of the 3 study days, baseline HR (70.73 ± 5.90, 69.48 ± 6.14, and 70.73 ± 5.60 beats/min) and MAP (80.31 ± 4.89, 77.60 ± 5.22, and 79.67 ± 5.22 mmHg) values did not significantly differ and showed only minimal, nonsignificant increases after MCh-induced bronchoconstriction (Fig. 3,B and C). As expected, cardiovascular variables did not change during CTs, and their values were significantly different (P always <0.01) from those observed during handgrip. In fact, during the 3-min exercise runs, both cardiovascular variables displayed progressive, significant increases up to peak values that were consistently attained by the end of the 3rd min of exercise (Fig. 3, B and C). There was no significant difference between mean peak HR and MAP values recorded during sHG30 and rHG60 runs. With the cessation of exercise, HR and MAP consistently resumed their control values within 2 min (Fig. 3,B and C) and did not display significant changes during the remaining 28 min of the observation period. As shown in Fig. 4, mean venous blood E and NE concentrations observed at baseline in CTs, as well as on sHG30 and rHG60 study days, were similar. MCh-induced bronchoconstriction did not induce significant changes in these variables. In contrast, 1 min after the completion of both sHG30 and rHG60 bouts, venous blood NE concentrations attained similar values that were higher than those recorded in control conditions and at maximum bronchoconstriction (P always <0.05). In all instances, NE concentrations measured 12 min after completion of each motor task were similar to those of control conditions; changes in venous blood E concentrations were small and inconsistent. No variations in plasma catecholamine concentrations were recorded during CTs. Fig. 4.Mean values (±SE) of plasma epinephrine (A) and norepinephrine (B) concentrations (n = 17) observed at baseline, at maximum BC, and at 1 and 12 min after 3 min of rest (solid bars), rHG60 (shaded bars), and sHG30 (open bars). * P < 0.05 compared with the corresponding baseline values. DISCUSSIONThis study in asthmatic patients analyzes the time course of recovery of Rint values after bronchoconstriction induced by MCh inhalation occurring either spontaneously or after 3-min bouts of static and rhythmic handgrip. The results of CTs indicate that the magnitude of the bronchoconstrictor response decreases with time in an approximately linear fashion, whereas the results of handgrip runs show that the recovery from induced bronchoconstriction follows a biphasic pattern. In the early postexercise phase (within 1 min), Rint displays a clear reduction, followed by a partial rebound toward higher values at 3 min (Fig. 1). Subsequently, i.e., during the remainder of the 30-min observation period, postexercise Rint progressively decreases at about the same rate as that observed after CTs, although consistently displaying lower levels. In addition, the magnitude of the overall bronchodilator response was similar after both handgrip paradigms and significantly greater than that observed after CTs (Fig. 2). We have monitored changes in airway resistance by means of the interrupter technique. The physiological basis and clinical utility of the interrupter technique have been reevaluated (e.g., Refs.12, 27), and the theoretical analysis by Bates et al. (1) has validated the technique. This technique has recently been proved to provide reliable estimates of airway resistance after MCh-induced bronchoconstriction (35), as well as to accurately assess the response to bronchodilators in the presence of preexisting bronchoconstriction (29). The interrupter technique is noninvasive, requires only passive cooperation, provides repeated estimates of airway resistance within a relatively short time interval, and, perhaps more importantly, allows the simultaneous assessment of other relevant variables, such as those derived from prolonged and undisturbed recordings of the breathing pattern. The use of variables derived from forced expirations for assessing changes in bronchial tone has been discarded, because these maneuvers require full inspirations that have been shown to alter airway smooth muscle tone (8, 40,42). After CTs, recovery from induced bronchoconstriction within the selected time frame was partial (Fig. 1). This finding is in keeping with previous observations in asthmatic patients showing limited recovery 30 min after MCh-induced bronchoconstriction (10,14). Spontaneous recovery from induced bronchoconstriction reflects not only natural degradation of the agonist and its clearance from the bronchial circulation (47) but also the reflex effects evoked by airway smooth muscle contraction and the mechanical action exerted by increased ventilation (7, 37, 40, 42). Responses specifically evoked by exercise likely contributed to define the magnitude and time course of bronchodilation after handgrip. In agreement with previous findings (24), the results show that, in asthmatic patients, MCh inhalation is accompanied by obvious increases in ventilation involving, as a rule, both the timing and drive components of the breathing pattern. These response are known to be reflex in origin (13) and can be ascribed to both direct (31) and indirect (13) airway receptor stimulation. The increases in Vt and V˙ibrought about by bronchoconstriction may have resulted in activation of other receptors, located both within and outside the respiratory tract. Signals arising from receptors located in the diaphragm that have been implicated in sympathetic activation during high-resistance breathing (43) may have reflexly influenced the time course of recovery from induced bronchoconstriction. In addition, feedback from slowly adapting pulmonary “stretch” receptors during the prolonged MCh-induced period of hyperventilation may have inhibited (32) the vagal bronchomotor neurons (vagal withdrawal). However, recent animal experiments indicate that vagotomy does not prevent suppression of MCh-induced airway narrowing caused by increases in the volume and frequency of mechanical ventilation (41) and suggest that maintenance of airway patency in vivo may depend on the static interactions between lung volume and the degree of smooth muscle activation, as well as on the stress exerted by the lung parenchyma on the airways during breathing (7, 37, 49). These mechanical events have been implicated in the reversal of MCh-induced bronchoconstriction by large lung inflation to near total lung capacity in both normal subjects (42) and patients with bronchial asthma (40). Inspiratory excursions approximating or exceeding twice the control Vt were occasionally recorded in our experiments and occurred during both CTs and HG runs. The effects of these “augmented breaths” were homogeneously distributed over all of the experimental conditions considered in the present study; thus they unlikely account for the bronchodilator responses observed immediately after handgrip cessation. Nevertheless, we can hypothesize that increases in ventilation of similar degrees can differentially affect the time course of Rint changes, depending on the level of airway smooth muscle tone on which they exert their mechanical action. Because handgrip consistently decreased Rint, the stretching effects of increased ventilation on dilated airways could have contributed in maintaining their caliber at a lower level throughout the whole 30-min observation period. The results confirm (17, 18, 20) that static and rhythmic handgrip paradigms producing the same total work output (9) evoked prominent rises in MAP and HR (Fig. 3). The cardiovascular responses elicited by isometric exercise are thought to be mediated by both an increase in efferent sympathetic nerve activity and a decrease in efferent vagal cardiac nerve activity (17, 18, see also for further references). Because handgrip caused no significant increase in circulating E but markedly increased HR, it is most likely that vagal withdrawal was the main mechanism involved not only in the mediation of tachycardia but also in the genesis of airway smooth muscle relaxation. Handgrip-induced cardiovascular responses were accompanied by slight, nonsignificant, additional increases in Vt andV˙i compared with those already attained after MCh inhalation. This finding is not surprising, because it seems obvious that the intense ventilatory response evoked by prior MCh inhalation did not warrant the development of the full range of ventilatory adjustments normally elicited by static (17, 18) and rhythmic (20) handgrip (occlusion phenomena). Animal studies have demonstrated that static contractions of the hindlimb muscles induced by electrical nerve stimulation (23,28), as well as injections of algesic substances into the arterial supply of the muscle (22), are accompanied by a reflex reduction in airway smooth muscle tone. It seems conceivable that the same mechanisms are implicated in the bronchodilator response evoked by both forms of handgrip. Whether supraspinal mechanisms, the so-called “central command” (48), have a role in the regulation of airway smooth muscle tone during voluntary exercise remains to be investigated. The reflexogenic drive arising from the working muscles, along with the central command, has also been implicated in sympathetic activation (43), which may occur also in response to prolonged, low-intensity (25% MVC) rhythmic handgrip causing no apparent stimulation of chemosensitive muscle endings (2). In agreement with previous findings (38), we found no significant changes in venous catecholamine concentrations in response to bronchoconstriction (Fig. 4). The finding of selective increases in NE plasma levels in response to isometric exercise is in keeping with previous observations obtained in normal subjects who performed static handgrip runs at the same contraction intensity as that used in the present study (e.g., Refs. 9, 30). Previous studies in normal subjects (46) have reported slight but significant increases in both E and NE plasma levels after handgrip bouts of the same intensity as that used in the present experiments but sustained until exhaustion. Handgrip runs were sustained without difficulty for 3 min by our patients and were not accompanied by muscle pain or discomfort (18). Whether the increases in circulating NE observed in the present experiments have a role in the genesis of the bronchodilation observed after sHG30 and rHG60 bouts remains unclear. In the light of earlier results showing no change in airway patency in response to NE infusions in asthmatic patients (3, 26), it seems unlikely that the mild, albeit significant, handgrip-induced rise in plasma NE concentration contributed to the postexercise decrease in Rint observed in the present experiments. However, the possibility exists that the action of NE is more evident when the airway smooth muscle is precontracted by MCh inhalation. Because handgrip-induced increases in circulating NE in this and previous experiments (30) have been shown to be short lasting, the effects exerted by this hormone, if any, were most likely confined to the decrease in Rint observed shortly after handgrip cessation. In light of available literature, inhibitory nonadrenergic noncholinergic neurotransmitters, such as nitric oxide and neuropeptide Y, which are coreleased with NE by sympathetic motor nerves (52), may contribute to airway smooth muscle relaxation. Due to its long-lasting activity (51), neuropeptide Y may also have a role in maintaining a decreased airway tone for the whole 30-min observation period (Fig. 1). Although studies in animal models (50) and in isolated human airway smooth muscle (5) have demonstrated airways dilation with severe hypoxemia, the results of studies performed in normal subjects seem to deny this possibility (15). Thus a significant effect of hypoxemia on the time course of recovery from induced bronchoconstriction appears unlikely and, furthermore, possibly counterbalanced by an opposing effect (44) exerted by the concomitant reduction in FetCO2provoked by hyperventilation in the present experiments. The carotid baroreflex has an influence on airway smooth muscle extending over the normal range of arterial pressures (39): increases in carotid body pressure decreases tracheal tension, whereas decreasing sinus pressure has the opposite effect (39). Airway smooth muscle relaxation is likely to represent a normal component of the adaptive response to isometric exercise. In fact, preliminary trials have documented a reduction in baseline airway resistance in most of the normal subjects and asthmatic patients with normal baseline airway caliber who performed 3-min static handgrip runs at the same contraction intensity as that employed in the present study. Additionally, previous studies have demonstrated airway dilation by brief, graded exercise in humans with normal or experimentally constricted airway smooth muscle (see Ref. 13 for references). In conclusion, we propose that the bronchodilator response induced by handgrip results mainly from the combined action of reflex withdrawal of cholinergic input to airway smooth muscle evoked by the simultaneous activation of skeletal muscle afferents and carotid sinus receptors (52), possibly with some contribution by mediators released in response to sympathetic activation. This work was supported by grants from the Ministero dell'Istruzione, dell'Università e della Ricerca of Italy. FOOTNOTESREFERENCES
Page 15many terrestrial animals live in environments that are topographically diverse and must often move up, down, and across slopes of varying degree. Grade-related changes in gravitational potential energy require that an animal expend more energy to move uphill at the same speed as on the level and dissipate energy to maintain this speed downhill. In accord with this, many physiological studies have shown that energetic costs, as measured by levels of oxygen consumption, increase during incline locomotion in diverse animal species (1, 9-11, 25, 30,38-41) and decrease during decline locomotion (1, 6,25, 30, 38, 41). Such shifts in locomotor energetics are largely due to changes in the recruitment and actions of various limb muscles. Whereas limb muscle recruitment typically increases on an incline (7, 8, 13, 20, 29, 31, 33), the reverse is true for decline locomotion relative to on the level (e.g., Ref.18). In fact, Smith and Carlson-Kuhta (36) noted that several major hip extensor muscles in cats remain completely inactive during decline walking on grades as low as 10%. In addition to these changes in muscle recruitment, alterations in limb posture and kinematics often accompany a change in surface grade (7, 17, 20, 37). For example, recent work on incline walking in cats demonstrated that, as grade increased from 25 to 100%, joint extension at the hip, knee, and ankle increased concomitant with electromyographic (EMG) intensity in extensors acting at those joints (7). In comparison, Smith et al. (37) showed that knee joint extension was reduced, whereas yield-related flexion at the ankle was increased on declined slopes of varying degree (also in cats). Such kinematic results suggest that limb muscles likely undergo relatively more active shortening on an incline but more stretching while active on a decline, which implies that shifts in muscle strain as well as recruitment are important for mediating changes in net muscle work and gravitational potential energy. Although quantification of joint kinematics provides insight into the length-change patterns of muscle-tendon units spanning those joints, kinematic data cannot directly resolve the specific muscle and/or tendon strains that underlie a given joint excursion in vivo. Sonomicrometry permits direct measurements of limb muscle fascicle strain during dynamic behaviors, but few studies have employed this technique to quantify alterations in muscle strain regimes associated with shifts in locomotor surface grade. Recent work by Roberts et al. (32) and Gabaldon et al. (12) has shown that lateral gastrocnemius fascicles in running turkeys actively shorten on an incline, remain nearly isometric on the level, and lengthen while active on a decline. These data support the notion that shifts in distal limb muscle strain are important for mediating mechanical work output or absorption in response to changes in surface grade, but the role of more proximal limb muscles in this context remains unexplored. To what extent might more proximal limb muscles also augment mechanical work output or dissipate energy in response to changes in surface grade? Hip and knee extensors generally have longer fibers than more pinnate ankle extensors, which suggests the capacity to actively shorten or stretch over relatively large distances, thus facilitating substantial contributions to energy production or absorption. However, aside from the conceptual understanding that incline and decline locomotion tend to bias muscles toward concentric vs. eccentric contractions, respectively, we have little appreciation for how length-change patterns in proximal limb muscles actually change with grade. In this study we use sonomicrometry and electromyography to measure patterns of muscle strain and activation in the biceps femoris (a biarticular muscle that acts in hip extension and knee flexion) and vastus lateralis (a major uniarticular knee extensor) of rats during uphill and downhill locomotion by having the rats move on a treadmill over a range of speeds and gaits. Our specific goals are threefold. First, we wish to test whether proximal muscles acting at the hip and knee exhibit different strain patterns depending on the surface grade. More specifically, we sought to determine whether these muscles shorten more on an incline than on the level and shorten less, or perhaps stretch, on a decline. Second, we address whether deactivation of the biceps femoris during stance, as observed in cats (36,37), is characteristic of decline locomotion in rats as well. If both rats and cats exhibit this pattern of deactivation, this may represent a relatively widespread neuromotor response to downhill grades among mammalian quadrupeds. Finally, by examining a range of speeds, we can explore the interaction between gait and grade on limb muscle actions during locomotion. For example, are differences observed among grades during walking also observed during running, or are the effects of grade more prominent at a particular gait or range of speeds? MATERIALS AND METHODSFemale Sprague-Dawley rats weighing 225–305 g (mean = 257 g) were obtained from Charles River Laboratories. Rats were housed in pairs in cages and maintained on a diet of IsoPro 3000. The room in which rats were held was kept at 21°C, and a 12:12-h light-dark cycle was established. Individuals were initially trained to walk, trot, and gallop on a small level treadmill with a 60 × 20-cm working section. Once individuals could maintain speed for 1 min at each gait, rats were then trained to move over a range of speeds with the treadmill either inclined or declined 15° (27% grade). Many animals were resistant to declined treadmill locomotion at high speeds, and thus the sample size of rats that performed trotting and galloping gaits downhill is relatively low. Animals ranged between 8 and 15 mo of age at the time of experiments, and all experimental procedures were approved by the University Committee for the Use and Care of Animals at Harvard University. To record patterns of muscle electrical activity and length change, fine-wire bipolar electrodes and piezoelectric sonomicrometry crystals were implanted unilaterally into the cranial or anterior aspect of the biceps femoris and central region of the vastus lateralis. The biceps is the largest muscle in the rat's hindlimb (3). Although it is biarticular, fascicles within the most anterior region of the biceps act primarily in hip extension and have little, if any, effect in flexing the knee. The vastus is the largest muscle of the quadriceps complex (3) and acts as a major extensor of the knee. Thus the two muscles of interest act as major extensors of the proximal hindlimb. In preparation for electrode and crystal implantation, rats were anesthetized with an intraperitoneal injection of pentobarbital sodium (35 mg/kg body mass). After anesthetization, the hindlimb and skull of the rat were shaved and scrubbed with a Povidone-iodine solution (EZ Prep, Becton Dickinson) for disinfection. A small skin incision was made over and parallel to the femur to expose the hindlimb muscles for implantation. A second incision was made through the skin over the skull, and a subcutaneous passage was created between the two incisions. To minimize wire exposure, all electrode and crystal wires were pulled through the incision at the skull, subcutaneously, to the limb incision. On the dorsal surface of the skull, a 10 × 15-mm area was cleared of all tissue, and a small hole was drilled into its center by using a dental drill. An epoxy block was secured onto the dorsum of the skull by using a small stainless-steel machine screw and dental cement. Before surgery, electrode and crystal wires were soldered into female miniature connectors (Microconnectors, GF-6), which in turn were glued to the sides of the epoxy block. Once the block was fixed on the skull, skin from the scalp was sutured around its base and sealed with silicone adhesive (Dow Corning). At the limb incision, two small pockets were created within the muscle tissue of the biceps and vastus by using the tips of fine watchmaker forceps. In each muscle, pockets were aligned along the trajectory of the muscle fibers ∼10 mm apart. Sonomicrometry crystals (1-mm + 38-gauge lead wires, Sonometrics) were placed within the pockets, which were then closed with 6-0 silk suture. Offset twist-hook bipolar silver-wire electrodes (22) with tips bared of insulation were implanted immediately adjacent to and between crystal pairs in each muscle by using a 21-gauge hypodermic needle. Because of the potential for motor unit compartmentalization within these muscles, there was the possibility that EMG recordings did not reflect the pattern of activation throughout the muscle as a whole. Nevertheless, because of the proximity of the electrode implant to the crystal implants, it is likely that fascicles in which length changes were measured are also the fascicles in which electrical activity was measured. Crystal and electrode wires were sutured with 6-0 silk through several superficial fibers onto the surface of the muscle to help prevent any dislodging during experiments. Once implantations were complete, the skin incision was sutured closed by using 4-0 silk. On completion of surgery, rats were allowed 24–48 h to recover before locomotor trials and recordings were started. The miniature female connectors fastened to the epoxy block on the rat's skull were connected via male counterparts (Microconnectors, GM-6) to lightweight shielded cables (Cooner Wire) that were attached to the EMG and sonomicrometry amplifiers. EMG signals were amplified (×1,000) and filtered (100- to 3,000-Hz band pass; 60-Hz notch) by using Grass P-511 preamplifiers. Sonomicrometry signals were amplified by using a Triton sonomicrometer (model 120-1001). Raw voltage data were digitized at 5,000 Hz with a 12-bit analog-to-digital converter (Digidata 1200B; Axon Instruments) and recorded onto computer with Digidata software (Axoscope). Table 1shows the individuals for which successful EMG and strain data were collected for the biceps and vastus.
Locomotor trials were performed on the same treadmill used for training, and all trials were recorded from a lateral perspective by using a digital high-speed video system (Redlake Motionscope; 125 frames/s). Once placed on the treadmill, rats were exercised at a given speed on the level until at least three to six strides were obtained at a steady speed; most trials recorded many more strides than this. Animals were then allowed to rest for 1–2 min before repeating another trial at a similar speed. Trials were then replicated with the treadmill inclined and declined at 15°, with rest intervals of 1–2 min between trials. The order of grade introduced was random. This procedure was repeated over a broad range of speeds until trials spanned gaits from walking to galloping. Because precise treadmill speeds could not be determined until after trials were completed, comparable speeds at each grade could only be approximated. As a result, data were organized by gait rather than speed for each individual. All trials were placed into one of five categories: slow walking (17–32 cm/s), fast walking (33–59 cm/s), trotting (52–83 cm/s), galloping first limb down (56–122 cm/s), and galloping second limb down (60–113 cm/s). The latter two categories distinguish between trials in which the experimental hindlimb was used as the first or second hindlimb down during a stride (i.e., “trailing” vs. “leading” limb). Table2 shows the breakdown of average speeds for each gait and grade combination from each individual.
An attempt was made to obtain 1–3 locomotor trials for each gait/grade combination. However, as Table 2 reflects, there were some combinations that were quite rare: trotting downhill, first-limb gallop uphill, second-limb gallop downhill. For each trial, visual frame-by-frame examination of video files was used to determine the timing of the experimental limb's foot-down and foot-up times for each stride. These times allowed determination of three statistical variables that were used to assess basic step-cycle parameters from each stride: stance duration, swing duration, and total cycle duration (stance + swing). With the use of these data, duty cycle was also calculated (stance duration/total cycle duration). Axoscope files were used to quantify fundamental parameters of the EMG bursts and fascicle strain regimes. Four statistical variables were used to assess EMG activity in the biceps and vastus: EMG onset and offset time (calculated relative to the timing of stance), duration, and intensity. Onset and offset times were determined visually by using magnified traces of the EMG signals. EMG intensity was calculated from the raw EMG data by averaging the spike amplitude of each rectified EMG signal. Intensities were scaled for every individual relative to the maximum value observed during level locomotion for each muscle (always observed during galloping). Thus an EMG intensity of 0.4 would reflect an average spike amplitude that was 40% of the maximum value observed during high-speed level galloping in that individual. Fascicle strains were estimated based on the changes in distance between sonomicrometry crystal pairs implanted into the muscles of interest. The signal-to-noise ratio of the sonomicrometry measurements averaged ∼40 across all animals and was estimated by comparing the baseline width of the signal at rest to the average amplitude of the signal during walking. More complete descriptions of the sonomicrometry technique are found in Olson and Marsh (28) and Biewener et al. (4), and more detail on our specific methods can be found in Gillis and Biewener (14). Briefly, any change in distance between crystals was assumed to represent a proportional length change over the entire fascicle. Fascicle strains were calculated as fractional length changes relative to a resting length. Rest length was determined while the animal was anesthetized. The animal was laid on its side (opposite the experimental limb), and the limb was then manually positioned so that the femur and tibia were oriented as during quiet stance. In this position, the distance between crystals in each muscle was measured, and these distances were used as resting lengths. Although these rest lengths likely represented a value that was slightly distinct from what would have been measured in an awake standing animal, making such measurements was impractical as animals tended to be quite mobile when awake and hooked up to the recording cables. Moreover, because rest lengths measured in vivo are often not linked to the force-length properties of the muscle under study, their main role is to serve as a standard metric against which fascicle length changes can be scaled, which our anesthetized measurements permit. Biceps strain patterns were characterized by a single statistical variable: stance-related shortening. Two variables were used to characterize vastus strain patterns: the initial lengthening strain observed during stance and the subsequent shortening strain (also during stance). Only stance-related strains were analyzed because these were the periods during which EMG activity was present and the muscles were likely generating active force. Basic statistics (i.e., mean and standard error) were calculated for each of the variables mentioned above to quantitatively assess aspects of the step cycle and patterns of strain and EMG activity in the biceps and vastus on incline and decline grades. Our recent work (14) on these proximal muscles during level locomotion allows for a direct quantitative assessment of the influence of surface grade on muscle recruitment and strain. To determine the effects of surface grade on these variables, repeated-measures ANOVAs were computed by using the mean values from each individual. Because slow walking was the only gait for which trials were obtained from all individuals on level, incline, and decline grades, this was the only gait for which repeated-measures ANOVAs were performed incorporating all three grades. Most other gaits had enough data from two of the grades to facilitate quantitative comparison. In these cases,t-tests were used to compare between grades, and they allowed for determination of whether trends observed in slow walking were also present at faster speeds and gaits. t-Tests were only used if data from at least the same four individuals were available at each grade. Thus comparisons were made between level and decline fast walking, level and incline trotting, and level and incline galloping (second limb). We used regression to analyze the relationships between speed and EMG intensity and relative duration. Analysis of covariance (ANCOVA) was then used, when appropriate, to compare these relationships among grades. A P value of <0.05 was used for all statistical tests to determine significance. RESULTSStep-cycle duration decreased in a curvilinear fashion with locomotor speed on level, incline, and decline surfaces (Fig.1A). At all grades, these decreases are mirrored by similar reductions in stance-phase duration (Fig. 1B). Swing-phase duration decreases minimally but significantly (P < 0.001; Fig. 1C). As a result, the proportion of the step cycle occupied by the stance phase (i.e., the duty cycle) exhibits a shallow curvilinear decrease with increasing speed regardless of grade (Fig. 1D). Fig. 1.Graphs of step-cycle duration (A), stance-phase duration (B), swing-phase duration (C), and duty cycle (D) as a function of speed for incline, level, and decline locomotion (black, gray, and white symbols, respectively). Different symbols represent different individuals (n = 7 rats). Note the comparable decrease in step-cycle and stance-phase duration, whereas swing-phase durations remain relatively constant or decrease only slightly. Plots of step-cycle, stance-phase, and swing-phase durations as a function of speed generally show substantial overlap among grades (Fig.1). However, more detailed analysis indicates that surface grade can influence the temporal nature of the step cycle, at least during relatively slow locomotor speeds (Fig.2A). For example, comparison of level, uphill, and downhill walking at comparable walking speeds reveals that the duration of the step cycle differs significantly among surface grades (P = 0.02;n = 7). Specifically, step-cycle durations are significantly shorter on a decline than on an incline (P = 0.03), and decline durations are also significantly shorter than on the level (P = 0.03); however, step-cycle durations are not significantly different between level and incline walking (P = 0.11). These effects on step-cycle duration are related to grade-dependent effects on both stance duration (P = 0.04; n = 7) and swing duration (P = 0.01; n = 7), which are always longer, on average, on an incline than on a decline (Fig. 2,B and C). Despite these differences in the absolute timing of the step cycle, duty cycle is not significantly affected by grade during slow walking (P > 0.05;n = 7) and averages 0.71 on both incline and decline surfaces (Fig. 2D). Because temporal aspects of the step cycle depend strongly on the speed of locomotion, quantitative comparisons of faster gaits are compromised by larger differences in average speeds across individuals. Fig. 2.Effects of surface grade on step-cycle duration (A), stance-phase duration (B), swing-phase duration (C), and duty cycle (D) during slow walking. Black bars, incline locomotion; gray bars, level locomotion; white bars, decline locomotion. Data are from the same 7 rats as Fig. 1at each grade. Speeds averaged 25 cm/s on the incline, 26 cm/s on the level, and 26 cm/s on the decline. Repeated-measures ANOVAs demonstrate a significant effect of grade on step-cycle, stance-phase, and swing-phase duration (P < 0.05). In all cases, durations are longer on an incline than on a decline. No significant effect of grade on duty cycle is detected (P > 0.05). Error bars denote standard errors. EMG activity in both the biceps and vastus typically begins near the time the foot makes ground contact during each stride and ends in the second half of the stance phase, regardless of gait or grade (Fig.3). Biceps activity, on average, begins slightly before the stance phase starts (Fig. 3). The major burst of vastus activity can begin slightly before or after the foot makes ground contact, but on average it starts after the onset of biceps activity (Fig. 3). Although a small burst of electrical activity is often present in the vastus late in the swing phase, this burst does not appear to be affected much by speed or grade. Absolute EMG burst durations in both muscles decrease with speed in a manner similar to the speed-dependent decrease in stance-phase duration (Fig. 4,A and C). As a result, on a given surface grade, EMG duration remains a nearly constant fraction of stance-phase duration, regardless of speed or gait (Fig. 4, B and D), Fig. 3.Schematic diagrams indicating the average timing of electromyographic (EMG) activity in the biceps and vastus during the step cycle for various gaits and grades. Within each gait, the stance and swing phases are distinguished by dark and light gray shading, respectively. Biceps data are shown in the top half of each diagram (above solid line), and vastus data are shown in the bottom half (below solid line). Average periods of EMG activity from level, incline, and decline locomotion are shown using gray, black and white bars, respectively. Numbers within each bar reflect the average speed of trials (in cm/s), followed by the number of rats (in parentheses) used to derive the EMG timing data. Data are only shown if at least 3 rats exhibited measurable EMG signals for that gait and grade. Note that EMG data from downhill trials for the biceps are often lacking due to little or no EMG activity in this muscle during downhill locomotion and not because of faulty recordings. Similarly, few animals exhibited downhill trotting or used the experimental limb as the trailing limb during uphill galloping. Error bars denote standard errors. Fig. 4.Effects of locomotor speed and surface grade on absolute EMG burst durations and durations as a fraction of stance-phase duration for the biceps (A and B; n = 4 rats) and vastus (C and D; n = 5 rats). Different symbols represent different rats (black = incline, gray = level). Best-fit linear-regression lines are shown inB and D for all incline and level locomotion trials. Data for decline locomotion are not included because EMG activity was rarely present across the entire range of speeds for an individual. For all regression lines shown, the slope is not significantly different from zero (see graphs for statistics). Thus, although absolute EMG burst duration decreases with speed, bursts occupy a consistent fraction of the stance-phase duration at all speeds. Analysis of covariance (ANCOVA) results show that this fraction is greater in the biceps on an incline than on the level (P = 0.006); however, the fraction of the stance phase occupied by vastus EMG activity is similar across these 2 grades (P = 0.34). In the biceps, EMG duration as a fraction of stance is greater, for any given speed, on an incline (mean = 0.85) than on the level (mean = 0.76) (P = 0.006, ANCOVA; Fig.4B). This difference in duration is due more to a shift in the relative timing of EMG offset than of EMG onset (Fig. 3). During decline locomotion, little or no EMG activity is present in the biceps except at high locomotor speeds (Figs.5A and6). Biceps EMG intensity increases significantly with speed on both level and incline surfaces. However, for decline locomotion, the slope of the line relating speed and EMG intensity is not significantly different from zero (Fig.5A). For any speed, biceps EMG intensity tends to be greater on an incline than on the level (P = 0.0001, ANCOVA) and is much reduced, on average, on a decline (Fig. 5A). Fig. 5.Effects of speed and grade on the relative intensity of EMG signals from the biceps (A; n = 4 rats) and vastus (B; n = 5 rats). Relative intensity values were obtained for each individual by dividing the absolute intensity of every EMG signal by the maximum intensity observed during level locomotion for that individual. Different symbols represent different individuals (black = incline, gray = level, white = decline). Best-fit linear-regression lines are shown for each grade (dashed line corresponds to decline data). Except for decline biceps data, all regression lines have a positive slope significantly different from zero that demonstrate increasing intensities with increasing speed (see graphs for statistics). ANCOVA results demonstrate that EMG intensities are significantly greater, on average, on an incline than on the level in the biceps (P = 0.0001) and vastus (P < 0.0001) and that vastus intensities are significantly greater on the level than on a decline (P < 0.0001). A lack of homogeneity among slopes precluded the use of ANCOVA for comparing biceps data from decline locomotion with other grades. Fig. 6.In vivo recordings of biceps strain and activation from the same rat during walking (top) and galloping (bottom) (all recordings from second limb down) for level, incline, and decline surfaces (left, middle, andright, respectively). Dark gray shading represents the stance phase, and light gray shading represents the swing phase. Note that the biceps largely shortens during stance, regardless of gait or grade, and that for both gaits stance-related shortening strains are greatest on an incline. In addition, traces from decline trials demonstrate minimal EMG activity, even at relatively fast speeds. All panels have the same scales for strain and EMG amplitude. Points of maximum strain, rather than the timing of stance onset, were used to define the intervals shown in each panel. Because maximum length was typically reached late in swing, the leftmost portion of each panel typically encompasses the very end of the swing phase (the small light gray area). L, muscle length; Lo, optimal muscle length. In the vastus, EMG duration as a fraction of stance is comparable across grades (Fig. 4D) and slightly lower, on average, than in the biceps (mean = 0.70, 0.67, and 0.70 for incline, level, and decline locomotion, respectively, when averaged across all speeds and individuals). In animals for which successful EMG recordings were obtained for both the vastus and biceps muscles, vastus activity typically began and ended slightly after biceps activity, regardless of grade (Fig. 3). Vastus EMG intensity increases significantly with speed on all three grades (Fig. 5B). At a given speed, vastus intensity is greater on an incline than on the level (P< 0.0001, ANCOVA) and is greater on the level than on a decline (P < 0.001, ANCOVA). In both muscles, fascicles undergo consistent and cyclic patterns of length change during each stride. In the biceps, fascicles generally shorten during stance and lengthen during swing, but both gait and grade influence the magnitude and/or trajectory of this length-change pattern (Fig. 6). Biceps fascicles generally begin to shorten just before the foot makes ground contact. After a brief bout of rapid shortening, fascicles often undergo a short period in which they remain relatively isometric, creating a small “shoulder” in the biceps strain trace early in stance (Fig. 6). This shoulder is typically prominent on level and decline grades but is diminished on an incline. After this shoulder, biceps fascicles shorten more considerably throughout the rest of stance. The biceps then begins to lengthen at the stance-swing transition and is stretched passively through all of the swing phase (Fig. 6). Total biceps shortening strains during stance increase with speed from a walk to a trot, and then decrease on transition to a gallop, regardless of grade (Fig. 7). In addition, for a given gait, shortening strains are generally highest on an incline, intermediate on the level, and lowest on a decline (Fig.7). This effect of grade is highly significant during slow walking (P < 0.001; n = 6). Specific comparisons show that strains on an incline (mean = 0.21) are significantly higher than strains on the level (mean = 0.16;P = 0.007) and on a decline (mean = 0.14;P = 0.003). During fast walking, strains are significantly lower on a decline (mean = 0.15) than on the level (mean = 0.18, P < 0.01; n = 6), and during trotting strains are significantly greater on an incline (mean = 0.24) than on the level (mean = 0.20,P < 0.01; n = 6). By averaging across all gaits, it was found that biceps shortening strains are 35% greater than those on the level (standard deviation among average values for each gait = 7%). Fig. 7.Histograms of average biceps shortening during stance at different gaits and grades for one individual. Gray, black, and white bars represent level, incline, and decline trials, respectively. Although variation exists among individuals with respect to shortening strain values, the grade-related changes exhibited here are generally representative of other animals. The only exception is the difference between incline and level galloping (lead limb), which is substantially greater in the animal depicted here than in the other rats (mean = 20% difference). Regardless of gait, however, shortening strains are typically highest on an incline, intermediate on the level, and lowest on a decline. Error bars represent standard errors. The strain regime of the vastus is markedly different from that observed in the biceps (Fig. 8). Vastus fascicles are typically stretched over the first half of the stance phase (Figs. 8 and 9). The amount of stretching incurred during this period varies among individuals but generally ranges between 8 and 15% of resting length, regardless of grade. After this stretch, fascicles exhibit one of three patterns through mid to late stance, depending on gait and grade: continued lengthening (generally observed on a decline), a brief period of nearly isometric behavior (most typical during walking), and a variable amount of shortening (generally greatest on an incline and at higher speeds). This “poststretch” or “late-stance” shortening strain is most prominent in the second limb down during galloping (Figs. 8 and 9), and it is significantly greater on an incline than on the level (P = 0.01; n = 5), averaging 0.14 and 0.10, respectively. During the transition from stance to swing, the vastus again begins to stretch as the knee is flexed and the foot is lifted off the ground (Figs. 8 and 9). Vastus fascicles continue to lengthen over the first third of the swing phase as the knee continues to flex and then shorten rapidly and over a large distance (15–20% of rest length) as the knee extends until the foot makes ground contact (Figs. 8 and 9). Fig. 8.In vivo recordings of vastus strain and activation from the same rat during walking (top) and galloping (bottom) (all recordings from second limb down) for level, incline, and decline surfaces (left, middle, andright, respectively). Dark gray shading represents the stance phase, and light gray shading represents the swing phase. Note that during stance the vastus always undergoes an initial period of stretch, regardless of gait or grade. After this relatively rapid stretch, fascicles exhibit a period of shortening, continued stretching, or relatively isometric behavior, depending on gait and grade. Late in stance, the muscle is again stretched rapidly as the limb transitions into swing. During the initial portion of the swing phase, fascicles continue to stretch before shortening rapidly and over a long distance until the foot again makes ground contact. The vastus typically exhibits EMG activity during the first two-thirds of stance, when the predominant strain is stretching. However, in the second limb down during galloping (bottom), the vastus typically undergoes more substantial shortening strains in the latter half of stance, and EMG activity is often present during much of this shortening. All panels have the same scales for strain and EMG amplitude. Fig. 9.Schematic vastus strain profiles for different gaits and grades. Vastus strain trajectories typically comprise 4 phases during each stride: 1) an intial stretch in the first half of stance; 2) a subsequent period in which fascicles exhibit variable degrees of shortening, lengthening, or remain nearly isometric; 3) another stretch during the transition from stance to swing; 4) a final period of substantial shortening over the second half of swing. Panels show the average timing and fascicle length excursions during each of these phases for different gaits and grades (n = 5 rats). Error bars are not shown in the figure because they obscure the presentation. Solid black lines reflect incline trials, gray lines represent level trials, and dashed lines represent decline trials. Within each gait, the stance and swing phases are distinguished by dark and light gray shading, respectively. Average speeds (in cm/s) and sample sizes (in parentheses) are shown atright. Animals rarely trotted downhill or used the experimental limb as the first limb down when galloping uphill; hence, strain trajectories for these gaits and grades are not shown. Note that the phase after the initial stretch in stance is what varies most across gaits and grades. Fascicles generally exhibit more shortening during this period as speed increases and as the grade inclines. In contrast, decline trials show little to no shortening over the same period. During decline locomotion, almost no late-stance shortening is ever observed in vastus fascicles. For example, during walking, vastus fascicles are stretched throughout the entire stance phase and exhibit only a shift in the relative velocity of stretch rather than a period of isometric or shortening behavior (Figs. 8 and 9). Even in the second limb down during decline galloping, fascicles often remain nearly isometric after their initial stretch, and, if they do shorten, strains are often no more than 2–3% (Figs. 8 and 9). DISCUSSIONWe undertook this study to determine the response of two proximal hindlimb muscles to changes in surface grade. The vastus lateralis is an uniarticular extensor of the knee; the biceps femoris is a biarticular muscle with an extensor moment at the hip and a flexor moment at the knee. Our results demonstrate that, at all speeds and gaits, both recruitment intensity and fascicle strain change markedly with grade in these muscles. Muscle activation levels are greater, for a given speed, on an incline than on the level and lower on a decline. In fact, the biceps femoris remains inactive during decline locomotion, except at relatively high speeds. Biceps fascicles shorten during stance on all grades, but for any gait they shorten more on an incline than on the level and more on the level than on a decline. The vastus lateralis is generally stretched 8–15% over the first half of stance, regardless of grade. After this initial stretch, vastus fascicles undergo a variable amount of shortening during the second half of stance on level and incline surfaces. This shortening is higher, on average, on an incline than on the level and increases with speed, reaching a maximum in the second limb down during galloping. In contrast, on a decline, vastus fascicles are generally stretched throughout stance, except during galloping, where a brief period of nearly isometric behavior follows the initial stretch. Taken together, these results suggest that proximal hindlimb muscles are likely as important for mediating levels of mechanical work output and absorption during grade locomotion as the distal ankle extensors that have been studied previously (e.g., Refs. 12, 15,31). Moderate slopes can influence temporal aspects of the step cycle during locomotion. For example, rats in this study used longer stride periods (i.e., lower stride frequencies) when walking up a 27% grade than when walking on the level at comparable speeds. Grade-related increases in stride duration have also been observed in rats moving up a 30% grade over a range of speeds (33), in cats walking up a 36% grade [see Fig. 1 in Pierotti et al. (29)], and in horses trotting up a 6–10% grade (16, 35). Although fewer studies have examined the effects of a decline grade on stride parameters, both rats (this study) and cats (37) appear to adopt relatively higher stride frequencies (i.e., shorter stride periods) and shorter stride lengths during decline walking than on the level. In fact, the study by Smith et al. (37) suggests that the steeper the slope, the more stride period and stride length decrease during decline walking. However, data regarding such grade-dependent temporal shifts are conflicting, even within the same species. Inclined grades had no effect on stride frequency in various studies of horses exercising at different speeds (9, 18, 31) or in cats walking freely (7) or running on a treadmill (29). Despite such disparity among compiled grade data for various animals to date, it seems fair to draw several tentative conclusions with respect to the effects of grade on temporal aspects of the step cycle. First, when an effect is present, it is manifested by an increase in stride duration on an incline (e.g., Refs. 16, 33,35) and a decrease in duration on a decline (e.g., Ref.37). Second, the magnitude of this effect may change with speed or gait; for example, several studies demonstrate an effect of grade during slow walking but not at faster locomotor speeds (e.g., Ref. 29). Third, the nature of the effect of grade may differ among species. Without more controlled studies using similar methods and large sample sizes, the impact of moderate surface grades on the temporal nature of the step cycle during quadruped locomotion will remain somewhat obscure. Although we did not set out to examine the interactions among grade, gait, and speed explicitly, examination of Table 2 provides some insight into these issues. First, animals were typically resistant to moving at high speeds on a decline (e.g., only 4 of 7 animals exhibited high-speed locomotion on a decline, whereas all 7 moved at high speeds on incline and level surfaces). We doubt such results are indicative of any natural disinclination toward downhill galloping in the wild but instead likely reflect the unnatural circumstance of running downhill fast on an enclosed treadmill. Second, animals rarely used trotting on a decline (e.g., only 1 of 7 animals exhibited downhill trotting). Most rats that were willing to gallop downhill were unwilling to exhibit a trotting gait on a decline at the speeds typically used for trotting on level and incline surfaces. Instead, these animals extended use of their fast walking gait at these speeds. This could reflect either a natural disinclination toward decline trotting or a response to using an enclosed treadmill. The fact that rats were willing to gallop on a decline but were nevertheless unwilling to trot at the speeds examined suggests that the reasons underlying the lack of one gait vs. the other may be different. Additional experiments are required to tease apart these issues. It is the activation and contraction of limb muscles that provide the forces and energy required to support and propel terrestrial animals during locomotion. Although there have been relatively few studies of the effects of surface grade on muscle activity patterns during animal locomotion, a number of results from this study are consistent with those observed among various vertebrate species moving up and down inclines. In general, extensor muscles become activated shortly before or at the time the foot makes ground contact, and activity persists over 60–80% of the stance phase. Although the effects of surface grade on the timing of muscle activity are relatively minor, the intensity of this activity is altered substantially. Relative to level locomotion, incline locomotion at similar speeds generally elicits an increase in EMG intensity in hindlimb extensors (7, 8, 19, 31,32), except those composed predominantly of slow-twitch fibers (29, 33). In contrast, decline locomotion generally elicits a reduction in EMG intensity in various limb extensors (37). Such alterations are typically interpreted as reflecting shifts in the level of activation or volume of muscle recruited to mediate the mechanical output required on different grades. A rather unexpected result of decline slope walking in cats was reported by Smith and colleagues in the mid 1990s (36,37). Major muscles that act to extend the hip (e.g., anterior biceps femoris and anterior semimembranosus) remained inactive during the stance phase on downhill grades as shallow as 10%. Whereas previous experiments had revealed an “immutable” synergy among limb extensors at the hip, knee, and ankle during various locomotor tasks (5), this synergy was broken during decline locomotion. Our data for downhill locomotion confirm the absence of EMG activity in the anterior region of the biceps femoris in rats as well, even at speeds faster than slow walking. In fact, even during slow galloping, EMG activity was occasionally absent on a downslope. Thus, among mammalian quadrupeds, inactivation of major muscles that act in extension at the hip joint may be a widespread neuromotor response to decline grades, thereby allowing hip extension to be largely passive (i.e., gravitational potential energy of the body can used to extend the hip when a rear foot is in contact with the ground), whereas active hip flexors may actually absorb mechanical energy during stance (37). It is well known that certain limb muscles exhibit biochemical and histological responses to exercise on different grades. More specifically, antigravity muscles typically experience substantial physical damage in untrained animals after bouts of downhill locomotion but less damage on the level (2). Such changes are assumed to be linked to grade-dependent differences in the mechanical actions of the underlying muscles. Limb muscles are presumed to undergo mainly concentric (i.e., shortening) contractions during uphill locomotion that result in relatively minimal damage. In contrast, downhill locomotion is known to bias limb muscles toward “eccentric” contractions, in which muscle fibers are stretched while actively generating force. Active muscle stretching dissipates energy as animals move downhill but also may result in rapid- and high-force development and injury, particularly to muscles composed mainly of slow fibers. For example, the vastus intermedius in rats, which is largely composed of slow oxidative fibers, incurs more damage after downhill locomotion than the vastus lateralis and medialis, which both have smaller proportions of slow fibers (2). Short-term training can mitigate the extent of muscle damage (34), perhaps by leading to an increase in the number of sarcomeres in series within myofibrils (23, 24) or by degeneration and/or regeneration of fibers susceptible to injury (2, 34). Despite recognizing major differences in the response of certain limb muscles to locomotion on different grades, the actual strain profiles experienced by such muscles have remained largely unknown. Strain data from the vastus lateralis of rats demonstrate a large degree of stretching over the first half of stance at all speeds (14) and grades (present study). EMG activity is generally coincident with this stretching, which suggests that eccentric contractions characterize vastus lateralis behavior regardless of grade. If one assumes that vastus intermedius strains are grossly similar to those of the vastus lateralis, this suggests that any differential morphological response of the muscle to training on different grades is not simply due to the presence of eccentric contractions on a decline and absence of such contractions on an incline. Instead, grade-dependent differences in vastus strain appear to be more subtle. Most in vivo studies of eccentric muscle contraction in rats exercise animals on declined slopes comparable to those used in this study (i.e., ∼15°) at 14–16 m/min or 23–27 cm/s (2, 23,24, 34), which is categorized herein as slow walking. At this grade and speed, vastus fascicles are typically stretched throughout all of stance and can experience continuous lengthening strains of nearly 25%, which has been shown to elicit substantial damage in situ in different mammalian muscle models (21, 27). Moreover, given the relatively higher limb-cycle frequencies used on a decline grade at this speed, stretching rates early in stance are between 5 and 50% greater than on an incline. Hence, differences in initial stretch velocity and/or the lack of any discrete fiber shortening are what differentiate the vastus strain regime between decline and incline grades at these speeds. The absence of fiber shortening and rapid rate of early stretch are likely the mechanical factors that induce the biochemical and morphological changes observed in vastus intermedius fibers after downhill locomotion (e.g., Refs. 2,23, 24, 34). Incline locomotion requires that the limb musculature as a whole produces larger amounts of mechanical work than on the level. Muscles produce mechanical work by actively shortening, and the greater the force produced and distance shortened, the higher the work output. Previous work on in vivo limb muscle force production and strain have typically focused on muscles acting at the ankle joint. Results from this work suggest that ankle extensors exert higher forces (15) and/or undergo greater shortening strains (32) in response to an inclined grade. Although we lack data on force production and thus cannot directly measure work output of these muscles on any grade, our strain and EMG data suggest that more proximal muscles also augment work output on inclined grades. During walking, trotting, and galloping, total biceps shortening and EMG intensity average 30–35% more on an incline than on the level. If one ignores force-velocity issues and makes a variety of other simplifying assumptions (e.g., EMG intensity is proportional to the volume of muscle recruited, all biceps shortening is active), these increases in shortening distance and EMG activity would suggest that the biceps likely increases its mechanical work output substantially in response to a 15° incline. Knee extensors might also augment work output during incline locomotion as the shortening observed late in stance on an incline is typically greater than that observed on the level in all gaits. For example, in the second limb down during galloping, vastus fascicles undergo 40% more late-stance shortening on an incline than on the level and show comparable increases in EMG intensity as well. Given the simplifying assumptions mentioned above, these changes likely lead to a major increase in work output by this muscle during late stance on uphill grades. Thus modulation of both strain and activation likely leads to substantial shifts in mechanical work production and absorption in major hip and knee extensor muscles of the rat during locomotion on different grades. Previous work on ankle extensors in different vertebrate species suggests that more distal pinnate muscles, with relatively long tendons, shift their function in a similar manner in response to changes in surface grade. Hence, despite substantial differences in architecture, major muscles throughout the hindlimb are likely important for amplifying and/or absorbing mechanical energy during locomotion on slopes of different degree. We thank Pedro Ramirez for help with animal care and training. Mike Williamson, Ty Hedrick, and Craig McGowan were of great help with various aspects of these experiments, and Ryan Monti provided useful comments on the manuscript. FOOTNOTESREFERENCES
Page 16the mechanomyogram(MMG) is a recording of the pressure wave produced by lateral expansion of a number of muscle fibers (22). In vitro studies using simple evoked contractions in frog muscles provide evidence that the vibratory signal from the muscle is an expression of the mechanical behavior of the muscle mass, although the main source of the MMG signal is motor unit activity (2, 12). A similar vibratory signal can be recorded easily from the surface of the skin in humans as a muscle contracts (surface MMG) (22), and it is expected that this technique will prove useful for investigating mechanical characteristics of muscle in the fields of physiology, clinical medicine, and rehabilitation. Surface MMG should prove useful if it is possible to ascertain the contributions of isolated motor units to the surface recording in human muscle. As has been extensively studied in the surface electromyogram (EMG), however, surface MMG is the sum of the signals emitted from a number of activated motor units, mediated and modulated by the architecture of the muscle-tendon complex, fat, and skin. To elucidate mechanical characteristics of the muscle from the recorded surface MMG, it is critical to determine how contractile properties and activation characteristics of individual motor units contribute to the surface MMG. Studies conducted on the gastrocnemius muscle of rats (5-7) and cats (23, 24, 28) have examined the contributions of isolated motor units to the surface MMG signals under well-controlled conditions. Bichler and colleagues (5-7), for example, investigated isolated motor units within the same muscles in vivo, and the contractile properties of the motor units were identified. The amplitude of MMG during repetitive stimulation of isolated motor units was shown to be associated with the amplitude of force oscillation (6, 7). Similar results were also obtained by Orizio et al. (23) in studies examining the medial gastrocnemius muscle of cats in situ. Therefore, evidence exists in mammals that isolated motor units contribute significantly to the surface MMG signal. In contrast to the research in mammals, there is little evidence supporting a close relation between surface MMG of the whole muscle and contractile properties and activation characteristics of the involved motor units in humans. In maximal twitch contractions evoked by direct whole muscle stimulation, the rising time of MMG (from the onset to the highest peak) was shown to be very similar to the contraction time of the single-twitch force, and it was found to be longer in the soleus muscle than in the vastus lateralis muscle (15). In studies using voluntary contractions, MMG from the soleus muscle contained an increased percentage of low frequencies compared with the biceps brachii muscles (18). Additionally, distinct responses to the increasing force level are observed in the amplitude of surface MMG between the soleus and medial gastrocnemius muscle (35). It has been further speculated that different contractile properties of motor units, in particular the speed of contraction and relaxation, may affect the development of fusion in relation to the contractile properties and discharge rate of motor units, thus affecting the amplitude of surface MMG (35). In these studies, the response of surface MMG from different muscles was interpreted with the presumption of the following fiber-type compositions: soleus contains the greatest abundance of slow fibers, followed by vastus lateralis (15), biceps brachii (18), and gastrocnemius (35). However, the contractile properties of the involved motor units were not examined. This kind of design is incomplete because it cannot discriminate the possible substantial effects due to slight differences in the setup of the sensor or the architecture of the muscle-tendon complex, fat, and skin (22). We hypothesized that the characteristics of surface MMG from the whole muscle in humans are largely dependent on the contractile properties of the activated motor units, especially on those features influencing the development of fusion (how fusion is developed in relation to contractile properties and stimulation frequency). This is based on the findings that fusion is developed at lower stimulation frequencies in slow-twitch fibers compared with fast-twitch fibers (10,31). Thus far, there is no direct evidence to support this hypothesis because previous studies do not identify motor unit activity in vivo (2, 12), were conducted on isolated motor units in other mammals than humans (5-7, 23, 24, 28), or were limited to the comparison between different muscles in humans (15, 18, 35). To test this hypothesis, we compared, in humans, the responses in surface MMG with the controlled intramuscular microstimulation of isolated motor units that belong to the same muscle with a wide range of contractile properties. With the intramuscular microstimulation technique, individual motor units can be isolated and the contractile properties of the stimulated motor units can be identified with regard to contraction time, half relaxation time, and development of fusion (13), thus allowing an accurate assessment of the contribution of individual motor units to the surface MMG in humans. METHODSEight isolated motor units were studied from the medial gastrocnemius muscle of four healthy male subjects. Their age, height, and body mass were 25.8 ± 0.4 (SE) yr, 177.2 ± 2.5 cm, and 68.3 ± 1.9 kg, respectively. The subjects had no medical history or physical signs of neuromuscular disorder. After subjects were fully informed about the nature of the experiment and possible risks involved, written informed consent was obtained from each subject. The protocol was in accordance with the criterion of the ethics committee review board of Kyoto University. Each subject was seated on an insulated, straight-back chair with wide belts crossing the chest and abdomen that tightly immobilized the body. An additional strap was used to secure the thigh to the chair. Force was measured with a strain-gauge transducer (model TB-654T, Nihon Kohden, Tokyo, Japan) positioned between a metal baseplate and a foot lever plate (Fig. 1). The bottom end of the foot lever plate had a semicircular attachment that surrounded and secured the heel. The heel was also secured with a strap at the bottom end of the foot lever plate. The strain-gauge transducer was aligned between the two plates near the distal part of the foot. The exact position of the entire device was carefully adjusted so that the knee was fully extended with the ankle joint angle at 100°. Fig. 1.Schematic diagram of the experimental setup. A strain-gauge transducer for the force measurement, a microphone for surface mechanomyogram (MMG), electrodes for surface electromyogram (EMG), and wire electrodes for the intramuscular microstimulation from an electrostimulator through an isolator are shown. Force, MMG, and EMG signals were amplified and stored on a personal computer via an analog-to-digital (A/D) converter. AC, alternating current; DC, direct current. The strain-gauge transducer measured the isometric plantar flexion force elicited by the electrical stimulations. The force signal was amplified through a direct-current (DC) amplifier (bandwidth: DC to 128 Hz; model SA-100, TEAC, Tokyo, Japan), and the system operated linearly between 0 and 9.8 N. Surface EMG was recorded with bipolar silver-silver chloride electrodes (9-mm diameter, 35-mm interelectrode distance) filled with conducting jelly that were applied over the belly of the medial gastrocnemius muscle. Electrode placement was preceded by abrasion of the skin surface to reduce the source impedance to 3 kΩ. The signals were band-pass filtered (5–1,000 Hz) and differentially amplified (model MEG-6100, Nihon Kohden; gain: ×1,000, input impedance: >100 MΩ, common mode rejection ratio: >80 dB) before recording. The method for measuring MMG signals was similar to methods previously employed in our laboratory (35, 36). The MMG was detected by an electret condenser microphone that was designed for research purposes (Daia Medical, Tokyo, Japan). The basic technique for detecting sound by this particular microphone is not different from other popular electret condenser microphones that are commercially available. Transverse mechanical activity of a muscle is transmitted to a thin diaphragm as it oscillates through an air column. The microphone is relatively small (10-mm diameter and 5-g mass), and it has a flat-frequency bandwidth between 3 and 2,000 Hz. The microphone was attached with adhesive tape over the belly of the medial gastrocnemius muscle between the electrodes for EMG measurement. The MMG signals were band-pass filtered (1–500 Hz) and amplified (model MEG-6100, Nihon Koden). The MMG, EMG, and force signals were displayed on a 20-MHz digital oscilloscope (model 5020A, Kikusui, Yokohama, Japan) and stored on a personal computer at a sampling rate of 2,048 Hz via an analog-to-digital converter (13-bit; TransEra 410, i2net, Tokyo, Japan). Two sets of fine-wire bipolar electrodes (stainless steel, 100-μm diameter, 5-μm uninsulated area, ∼200-μm interelectrode distance) were inserted into the medial gastrocnemius muscle ∼1.5–2.0 cm under the skin surface located directly beneath the MMG microphone sensor. Single rectangular electrical pulse waves of 0.5-ms duration were delivered from an electrostimulator (model SEN-7203, Nihon Koden) through a stimulator isolation unit (model SS 102J, Nihon Koden). To verify that only single motor units were stimulated, the position of the intramuscular fine-wire electrodes and the intensity of the stimulus (1–10 V) were carefully adjusted until reproducible all-or-nothing responses were acquired in both signals of EMG and force, simultaneously. The evoked EMG responses were monitored continuously to ensure that the same motor unit was stimulated throughout. Any trials with irregular EMG responses were rejected on-line. First, each motor unit was identified by measuring its mechanical properties during single-twitch contractions. The EMG, MMG, and force signals were averaged from 10 single-twitch contractions for each motor unit. Figure 2 shows examples of the averaged EMG, MMG, and force signals from two motor units that exhibited distinctively different mechanical properties. Peak twitch force, contraction time, and half relaxation time were calculated from the averaged force signals by following the methods of our laboratory's previous studies (19, 35). Contraction time was the time interval between the onset of force and the peak force. The onset of force was defined as the point at which the value exceeded three standard deviations of the baseline noise for three consecutive sampled points. Half relaxation time was defined as the time taken for the force to decline to one-half of the peak force value in the relaxation phase. Half relaxation time was employed for twitch force because it is a standard method to characterize the relaxation phase of motor units. As a measure of twitch contraction duration, contraction time and half relaxation time were summed and termed as “twitch duration” according to its usage in a previous study (8). Fig. 2.Examples of the surface EMG (A), MMG (B), and force (C) signals averaged for 10 single twitches of a motor unit. For comparison, data from a motor unit with the shortest twitch duration (163.2 ms, MU 1; left) and with the longest twitch duration (220.6 ms, MU 8; right) are shown. Onset of EMG corresponds to 0 ms on x-axis. MMG duration was measured in the following way from the averaged MMG signals. The MMG duration was defined as the interval between the onset and end of MMG signals produced by single-twitch contractions. In contrast to the analysis of twitch force, the end of MMG signals was employed because the MMG signals do not necessarily decrease monotonically after the peak. In this study, the end of MMG signals was defined as the point at which the rectified MMG declined to three standard deviations of the baseline noise for three consecutive sampled points. Similar to the methods used for force signals, the onset of MMG signals was also defined as the point at which the value exceeded three standard deviations of the baseline noise for three consecutive sampled points. Second, each individual motor unit was stimulated at 5, 10, 15, and 20 Hz. To exclude the trials that involved stimulation of surrounding motor units, evoked mass action potentials (M wave) were examined to determine whether they remained constant. Each stimulus lasted ∼6 s with a 10-min rest period between frequencies to avoid fatigue. Mean force was calculated from the middle 2-s segment of the signal. For further analysis of the force signals, both the DC component and linear trend of the force signals were eliminated by digital filtering to exclude transient responses (filtered force). From these signals, peak-to-peak amplitude of the MMG signals (MMG amplitude) and force signals (force fluctuations) were measured. For further analysis, the obtained MMG amplitude and force fluctuations were then normalized to the respective maximal values in each individual motor unit. Also, frequency analysis (4,096 points, Hamming window, fast Fourier transform) of the filtered force and corresponding MMG signals was performed over the same 2-s window to obtain the mean power frequency as in our laboratory's previous studies (20, 35). The mean power frequency was defined as the ratio between spectral moments of orders one and zero (20). The mean power frequency of MMG and force was analyzed to confirm whether their responses matched the stimulation frequency. Descriptive statistics include mean and SE. Statistical analyses were made by using linear correlation coefficients (Bravais-Pearson'sr). All values are expressed as means ± SE throughout the text, figures, and table. A probability level of P < 0.05 was considered to be statistically significant. RESULTSThe investigated motor units possessed a wide range of contractile properties with regard to twitch force, contraction time, half relaxation time, and twitch duration (Table1). The MMG duration also varied over a wide range. The differences in the contractile properties of the motor units and the corresponding surface MMG signals were obvious when the signals of the two motor units with the shortest and longest twitch duration were compared (Fig. 2). As can readily be seen, the MMG signal lasted for the full duration of the force twitch. The waveforms were not similar to those measured by accelerometers in which the MMG signal was rarely observed during relaxation phase (3, 4, 25). But the waveforms of the present study were similar to those usually obtained by electret condenser microphones (29) or piezoelectric microphones (5). The relations between MMG duration and contraction time, half relaxation time, and twitch duration of all the motor units are reported in Fig.3. MMG duration was strongly correlated with half relaxation time and twitch duration, but it was not related to contraction time.
Fig. 3.Relationships between MMG duration and contraction time (A), half relaxation time (B), and twitch duration (C) of all the motor units (n = 8). NS, not significant. In repetitive contractions, the characteristics of MMG and force signals varied, whereas the mean power frequency of MMG and force oscillations approximately matched the stimulation frequency, in all the motor units. The r values were 0.953 (P< 0.001) between mean power frequency of MMG and stimulation frequency and 0.980 (P < 0.001) between mean power frequency of force and stimulation frequency. Figure 4 shows two examples of surface MMG and force signals during repetitive stimulations of a motor unit with different stimulation frequencies. DC components of the force signals are removed as described in methods, and the motor units with the shortest and longest twitch duration correspond to the two motor units in Fig. 2. In both of the motor units, systematic reductions in the MMG amplitude and force fluctuations were observed concomitantly as the stimulation frequency increased. When the MMG amplitude and force fluctuations were normalized with respect to the maximal values for each individual motor unit, the MMG amplitude appeared to decline linearly in relation to the decrease in force fluctuations as the stimulation frequency increased (Fig.5A). The change in the MMG amplitude was most pronounced when the stimulation frequency increased from 5 to 10 Hz, where a highly significant correlation was exhibited between the changes in the MMG amplitude and force fluctuations (Fig.5B). These results indicate that the changes in the amplitude of surface MMG signals are significantly related to how fusion is developed in relation to the contractile properties of the activated motor units and stimulation frequency. Fig. 4.Examples of the surface MMG signals and filtered force (DC component and linear trend were eliminated) during repetitive stimulation of a motor unit at 5, 10, 15, and 20 Hz. MU 1 (top) and MU 8 (bottom) correspond to those in Fig. 2. Fig. 5.A: relationship between the amplitude of MMG and force fluctuations with different stimulation frequencies. Data are normalized to the maximal values of each measurement. Values are means ± SE. B: correlation for relative changes between the amplitude of MMG and force fluctuations in individual motor units (n = 8) by the increase in the stimulation frequency from 5 to 10 Hz. In addition, in the two motor units shown in Fig. 4, reductions in the MMG amplitude seem to be more pronounced in the motor unit with the slower contractile properties. In general (10, 34) and also in the present study (data not shown), contractions of slower motor units produce smaller force compared with faster motor units. Practically, mean force is most often related to the MMG signals in the studies utilizing voluntary contractions (26, 30, 35). Therefore, we further tried to identify and present the effects of contractile properties of the activated motor units on the relation between the characteristics of surface MMG and the magnitude of the mean force. Although the available number of motor units was small, four motor units with longer twitch duration (207.2 ± 6.9 ms) formed the “slower motor units” group and the four remaining motor units with shorter twitch duration (166.9 ± 2.0 ms) formed the “faster motor units” group. These groups were formed to ascertain whether any effect of the contractile properties of motor units appear. The MMG amplitude and force fluctuations of each group were investigated in relation to the mean force during repetitive stimulations. It should be noted that the mean force was calculated from the force signals that were not filtered. In Fig. 6, Aand B, greater declines in the group of slower motor units were apparent for the MMG amplitude and force fluctuations as mean force increased. Furthermore, we calculated the individual rate of decline in both MMG amplitude and force fluctuation against the increase in the mean force for each individual motor unit. In this calculation, regression analyses were applied to those values obtained at four different stimulation frequencies. Figure 7 shows that the rates of change in MMG amplitude and force fluctuations are significantly related to half relaxation time and twitch duration, with higher r values in those relations to half relaxation time. These results suggest that declines in the amplitude of surface MMG signals with force are more pronounced when the contractile properties of the activated motor unit are slower. Fig. 6.Relationships between the amplitude of MMG and mean force (A) and between force fluctuations and mean force (B) during repetitive stimulation with different stimulation frequencies. Values are means ± SE. □, Data from the 4 motor units with shorter twitch durations (Faster MUs); ●, data from the 4 motor units with longer twitch durations (Slower MUs). Fig. 7.Rate of changes with force in MMG amplitude (A andB) and in force fluctuations (C and D) compared with half relaxation time and twitch duration. Rate of change with force was calculated by the regression analysis of the relations between MMG amplitude (%) and mean force (N) for individual motor units or the relationships between force fluctuations (%) and mean force (N) for individual motor units. Data are from all the individual motor units (n = 8). DISCUSSIONAmong numerous attempts to clarify the significance of surface MMG, this is the first study that provides evidence in humans of a direct relation between the characteristics of surface MMG signals from a single whole muscle and the contractile properties of motor units. Major findings of this study were 1) that the duration of surface MMG was strongly correlated with the twitch duration of a motor unit, 2) that the decline in the amplitude of surface MMG was closely related to the decline in force fluctuations generated by increased frequency of contraction,3) and that fluctuations in surface MMG and force signals were related to the twitch duration and half relaxation time of the activated motor units. These results support the hypothesis that the characteristics of surface MMG are largely dependent on the contractile properties of the activated motor units, especially on those features influencing the development of fusion. In vitro studies (2, 12) confirm that the main sources of the MMG signals are the pressure waves generated by the gross lateral movement of the muscle fibers that occur during contraction and relaxation of the fibers. It is not clear, however, how these pressure waves can be reflected on MMG signals that are recorded at the skin surface in humans, although studies have tried to relate the characteristics of surface MMG signals to the contractile properties of muscle fibers by comparing different muscles (15, 18). In single-twitch contractions, pressure waves are generated from a simple movement of muscle fibers by a set of contraction and relaxation. Contraction time and relaxation time during single-twitch contractions are mostly related to myosin ATPase activity and reuptake of calcium, respectively, both of which characterize fiber types, and the latter can be acutely impaired by fatigue (11). If surface MMG signals reflect the pressure waves from motor unit activity, then a close correlation should exist between the duration of surface MMG and the required time for the muscle fibers to complete a set of contraction and relaxation. In the present study, the duration for a set of contraction and relaxation of a motor unit was evaluated by the sum of contraction time and half relaxation time (twitch duration). In the motor unit pool of the gastrocnemius muscle in humans, a strong correlation between duration of surface MMG and twitch duration was found (Fig. 3). Within the twitch duration, the duration of MMG was also correlated with half relaxation time but not with contraction time. This may simply be due to the wider range of half relaxation times compared with contraction times. The present results clearly indicate that the duration of surface MMG during twitch contractions depends on the contractile properties of activated motor units and suggest that the time characteristics of the pressure waves in twitch contractions is well reflected on surface MMG signals. The contributions of individual motor units to the surface MMG were examined in the medial gastrocnemius muscle of rats by stimulating isolated motor units (5-7). The close connection between contractile characteristics of motor units and surface MMG signals in rats has been demonstrated by the strong correlation between the MMG amplitude and the twitch tension of a motor unit (5). Together with the present findings on the strong correlation between the time characteristics of MMG and twitch tension in humans, it is likely that surface MMG during twitch contractions is largely dependent on the contractile characteristics of motor units. In tetanic contractions, lateral movement of muscle fibers may depend on the combination of the contractile properties of the muscle fibers and the frequency of the stimulation. Fluctuations of force during tetanic contractions are generated by the oscillation in tension during a series of contraction and relaxation. As a rule, the relative contribution of the relaxation phase to the fluctuations will be smaller as the stimulation frequency increases and fusion is developed. It is also known that the slower the twitch of the muscle, the lower the frequency at which fusion of the evoked mechanical events takes place (10, 31). As a corollary, force fluctuations at a given stimulation frequency would be smaller in motor units that have slower contractile properties. In the present study, this is demonstrated in the gastrocnemius muscle in humans as a greater rate of decline in force fluctuations in the slower motor units (Fig.6B). It is further supported by a significant correlation between the changes in force fluctuations and twitch duration as well as in the half relaxation time of the motor units (Fig. 7, C andD). Provided that both the lateral movement of muscle fibers and force fluctuations originate from a common source (2, 12,22-24), features influencing the development of fusion in the motor units can affect lateral movement of the muscle fibers and the characteristics of surface MMG as a consequence. In this study, the amplitude of MMG decreased in close relation to the force fluctuations produced by the different frequencies of stimuli (Fig. 5A), and the magnitude of these reductions was significantly correlated across the sampled motor units (Fig. 5B). Moreover, the rate of decline in the amplitude of MMG was significantly correlated with twitch duration as well as the half relaxation time of motor units (Fig. 7, A and B). Collectively, these results indicate that the changes in the amplitude of surface MMG during tetanic contractions of the gastrocnemius muscle in humans are largely dependent on the contractile characteristics and the development of fusion in the active motor unit. During unfused tetani with a variable degree of fusion in rats, progressively decreasing fluctuations in tension have been observed with increasing stimulation frequency, and the MMG amplitude and fluctuations in tension have been linearly related (7). This is consistent with the apparently linear decrease in MMG amplitude and force fluctuations with increasing stimulation frequency in humans (Fig. 5). Although the findings of the present study are limited because of the small number of motor units that were studied, previous literature on a rat muscle favors the present results in a human muscle and lends support to the hypothesis that the characteristics of surface MMG are largely dependent on the contractile properties of motor units. On the basis of the present findings and previous literature on MMG during electrically evoked contractions, various characteristics of surface MMG during voluntary contractions may reasonably be explained. Contrary to the tetanic contractions of individual motor units in the present study, voluntary contractions involve a greater number of activated motor units, and the discharge rate of these motor units varies and is generally not synchronized. It is of additional note that the amplitude of surface MMG is proportional to the amplitude of the motor unit action potential (35). The amplitude of MMG is known to increase with force in most muscles during voluntary contractions from low-to-moderate force levels, and the same holds true for the amplitude of EMG (1, 26, 30, 35). In this range of force, newly recruited motor units augment the gross lateral movement of the muscle fibers while the discharge rate of motor units is not yet high enough to attenuate the amplitude of MMG. At higher force levels, declines in the amplitude of MMG with increasing force have been observed, whereas the amplitude of EMG increases continuously because of the additional recruitment of motor units and the increase in their discharge rate (1, 26, 35). The decline in the amplitude of MMG may be attributed to the development of fusion where the high discharge rate of motor units may have greatly reduced the dimensional changes possible in muscle fibers. It is also expected that a muscle with a greater percentage of slow motor units is more likely to have a reduced amplitude of MMG (35). During voluntary contractions at a submaximal force level, the deficit in force of motor units during a fatiguing protocol is compensated for by increases in the discharge rate of already recruited motor units and/or by the additional recruitment of motor units (20). These changes in the activation patterns of motor units are reflected in the increased amplitude of EMG with time, whereas the amplitude of MMG declines in time with fatigue (27, 30). It is known that the relaxation time of muscle is prolonged by fatigue (8). The increased discharge rate of motor units with prolonged relaxation time under fatigue would facilitate the development of fusion and thus would decrease the amplitude of MMG. Collectively, the EMG-MMG relation during voluntary contractions can exhibit a linear relation under the condition where the positive effects on the amplitude of MMG and EMG by the recruitment of motor units are not substantially counteracted by the negative effects on the amplitude of MMG caused by the development of fusion. The present findings indicate that the influence of the development of fusion on MMG would explain this relation between EMG and MMG. It seems, therefore, that the relation between surface MMG and the force exerted during voluntary contractions of various muscles is dependent on the discharge rate and contractile properties of the activated motor units and on relative proportion of slow motor units in the muscle under investigation. This interpretation may also be applied to characterizing physiological tremor by MMG (17), where unfused tetani may play a role (16). Lastly, it is relevant to point out some technical considerations about the employed methods and limitations of the obtained results. In previous studies seeking a relation between surface MMG signals and the mechanical characteristics of a single motor unit, the spike-triggered averaging technique has been applied during voluntary contractions (29, 35). This technique is advantageous in that it does not require sophisticated stimulation techniques and voluntary contractions are possible, but it cannot necessarily measure true mechanical activity of a single motor unit because the averaged signals can easily be distorted by partial fusion even when it is discharging at a very low rate (14, 21). The present intramuscular microstimulation technique circumvents this problem (14) because it allows the control of discharge rate of a single motor unit (13, 33). Analysis or comparison of the recorded signals is further complicated by the types of transducers and by the layers of tissue between the muscle and the transducer (9, 22). With the utilization of an electret condenser microphone in the present study, MMG signals during twitch contraction lasted for the entire duration of the force twitch as in the previous studies that utilized electret condenser microphones (present study; Ref. 29) or piezoelectric microphones (5). In contrast, MMG signals detected by accelerometers have not exceeded the contraction phase of the single twitch with its main oscillations (3, 4, 25). Also, the mean frequency of MMG signals matches the stimulation frequency in the studies with electret condenser microphones (32,35), whereas the studies with accelerometers tend to have higher mean frequencies of MMG signals than the stimulation frequency (37, 38). From the basic mechanical properties of motor units and the consequent force signals in a single twitch, it is obvious that acceleration measured during the displacement of the muscle is much smaller during the relaxation phase compared with the contraction phase. As a consequence, the second-order derivative of the displacement during the contraction phase will be particularly emphasized in the signals recorded with accelerometers. For these reasons, the waveforms obtained by accelerometers (3, 4,25) could be different from the ones obtained by electret condenser microphones (present study; Ref. 35) or piezoelectric microphones (5). The skin and fat tissues beneath the transducer can act as low-pass filters (3, 22), and the occurrence of this is greatly different among subjects and individual muscles. It is true that the number of motor units is limited in the present study because of the major drawback of the electrical stimulation in humans, but many of the possible problems in the nature of the signals may be resolved by employing the intramuscular microstimulation technique on a single motor unit belonging to the same muscle and by utilizing the electret condenser microphones. Despite a limited number of motor units sampled, the mean frequency of MMG matched the stimulation frequency, and the observed range of contractile measures was substantially wide and closely consistent with the literature for the human gastrocnemius muscles (13, 33). This indicates that the effects of the medium are small or consistent and that the sampled motor units are well mixed in contractile properties. Still, the findings should be limited to the present setup for the human gastrocnemius muscles. Further studies on a larger number of motor units within a single muscle from a variety of muscles may extend the present findings. In conclusion, evidence from the present study supports a close relation between the surface MMG and the contractile properties of individual motor units within a single muscle in humans. It is suggested that the major characteristics of surface MMG are largely dependent on the contractile properties, especially on features in the development of fusion, of the activated motor units. The authors are grateful to Drs. Motoki Kouzaki, Tetsuo Fukunaga, and Hiroaki Kanehisa (University of Tokyo, Tokyo, Japan) for helpful suggestions and to Fumi Toyooka (University of Tokyo, Tokyo, Japan) for assistance with the figures. We thank Kevin G. Keenan (University of Colorado, Boulder, CO) for help proofreading and improving the manuscript. FOOTNOTESREFERENCES
Page 17the surface electromyographic (EMG) signal presents smaller bandwidth with respect to the intramuscular EMG, because the tissues separating the muscle fibers and the recording electrodes act as low-pass filters (15, 28). This determines low-spatial selectivity, which hinders the separation of the contributions of different motor units (MUs). For this reason, past research efforts in the surface EMG field were mainly devoted to the development of processing techniques in time and frequency domain, which gave indications about the global EMG activity, without aiming at an analysis at the single MU level. Although, recently, techniques for noninvasive assessment of single MU properties have been proposed and refined (4, 6, 8), their applicability is still limited by the need of complex detection systems, to low-contraction levels, and controlled laboratory conditions for signal recording. Thus the extraction of global information from the surface EMG signal remains of paramount importance in many applications, such as sport and occupational medicine, rehabilitation, and basic and applied physiology. Spectral analysis and amplitude estimation are usually performed from single differential signals (3) to obtain indications about the physiological processes occurring during sustained voluntary contractions. In addition to the efforts devoted to the technical issue of estimating amplitude and spectral variables in a reliable way (12, 32, 41), many studies were focused, in the past, on the clarification of the relationships between these global variables and the underlying physical processes, to extract information of physiological interest from the global analysis of the surface EMG signal. It has been clearly established that the rate of change of spectral variables and conduction velocity (CV) during a sustained contraction is indicative of muscle fatigue (31) and may be correlated with MU type (39, 43). It has also been shown, both theoretically (28, 41) and experimentally (2), that, during fatiguing contractions, CV and mean (MNF) or median spectral frequency (MDF) of the surface EMG signal are highly correlated; MNF and MDF reflect mainly the CV changes of the active MUs. However, the comparison of the percent rate of CV and MNF and MDF decrease showed that, although being the main contribution to spectral compression, CV is not the only determinant of changes of the characteristic spectral frequencies. A higher decrease of MNF and MDF with respect to CV was indeed observed in a number of past studies (refer, for example, to Ref. 31 and, recently, to Ref.30). To explain these results, changes in MU CV distribution spread, increase of the depolarization zone length, or increase in MU short-term synchronization has been suggested (31), although with no direct evidence. Among these factors, recent results indicated a major role of MU synchronization (25, 45). However, the ability of surface EMG spectral analysis as an indicator of MU short-term synchronization is still under debate. On the basis of the above considerations, nonlinear analysis of surface EMG signal has been exploited (35, 42). Recurrence quantification analysis (RQA), described by Eckmann et al. (5), is a technique for the detection of state changes in drifting dynamic systems that does not necessitate any a priori constraint on data size, stationarity, and statistical distribution (18). RQA has been used recently in a number of experimental studies (17, 23, 42), showing its potential in detecting changes in surface EMG due to fatigue. Webber et al. (42) tested the sensitivity of different indexes extracted from RQA. These authors indicated that subtle changes in surface EMG can be detected by the percentage of determinism (%Det), which reflects the amount of rule-obeying structure in the signal dynamic, and the percentage of recurrence (%Rec), which reflects the current state of the system. Furthermore, under particular experimental conditions, it was speculated (16, 18) that %Det should be more sensitive than spectral analysis to MU short-term synchronization, because it reveals embedded determinisms in an apparently stochastic signal. However, none of these studies has investigated in depth details which EMG signal parameters are indeed reflected by EMG variables extracted by RQA. For this reason, it is difficult to interpret experimental results, in particular when they are compared with those obtained by spectral analysis. In this work, we focused on two phenomena, the MU CV change and the degree of short-term synchronization, which clearly have an impact on surface EMG spectral variables. The theoretical derivation of the effects on RQA of mean MU CV and MU short-term synchronization and its relationship with classic spectral EMG analysis are difficult because of the complexity of the surface EMG generation and detection system. A modeling approach is, on the other hand, feasible and may be useful for the interpretation of experimental results. The aims of this work were thus 1) to investigate the effect of MU short-term synchronization and mean CV on EMG variables extracted from linear and nonlinear analysis, 2) to assess whether some variables of RQA and spectral analysis provide indications on the same neuromuscular system parameters or whether they have different sensitivity to variations of such parameters, and 3) to analyze advantages and limitations of the linear and nonlinear approach for assessment of changes in the surface EMG signals, as a consequence, for example, of fatigue, pathological conditions, or training. The study is performed on the basis of a simulation model, which allowed separate investigation of different muscle parameters. The modeling results are used to interpret experimental findings obtained from the biceps brachii muscle during voluntary isometric contractions. METHODSClassic spectral variables computed from the surface EMG signal are MNF and MDF. The estimates of these variables depend on the additive noise and the estimators adopted (12); on anatomical, physical, and detection-system parameters (9); and on the number, sizes, type, and firing rates of the active MUs (10). In the present study, these variables have been computed from synthetic and experimental surface EMG signals with well-known algorithms (for a recent review, see Ref. 12). In particular, the periodogram-based spectrum estimation was obtained from adjacent nonoverlapping signal epochs of 1 s. For the mathematical details about RQA, the reader can refer to a recent review by Filligoi and Felici (18), here summarized in the . The procedure is based on embedding EMG data in an N-dimensional Euclidean space. Recurrence maps are built from signal epochs, and quantification indexes are extracted from these maps (refer to the ). Among these indexes, %Rec and %Det will be investigated in this study. A model previously developed for the simulation of the surface EMG signal (13) was used in this study. The model simulates synthetic MU action potentials (MUAPs) generated by finite-length fibers and detected by surface electrodes with physical dimensions. The volume conductor is an anisotropic medium representing the muscle and two-layered isotropic media representing fat and skin tissues (15) (Fig. 1C). The transmembrane current density was described as indicated by Rosenfalck (37). The fixed-model parameters (such as the conductivities of the subcutaneous layers) are the same as reported in a previous study (10). Fig. 1.A: the independently generated firing patterns of all of the 65 active motor units (MUs). B: the same patterns to which synchronization has been added (10% firings, and MUs synchronized). C: the volume conductor and an example of location and territories of the active MUs. D: representative single differential simulated signal (10% firings and MUs synchronized). The simulated signals were detected by a single differential system with electrodes 5 mm long and 1 mm thick. The interelectrode distance was 20 mm. The detection probe was located in the middle between the centers of the innervation zone and the tendon region of a number of MUs with a mean semilength (in both directions) of 65 mm. The detection volume of the system was defined as suggested by Farina et al. (10) on the basis of a threshold in the energy of the surface MUAPs. The contributions to the surface EMG signal of the MUs out of the detection volume were neglected. The number of MU fibers was uniformly distributed between 50 and 450, and the MU fiber density was 20 fibers/mm2 (19). The MU territory was circular, and the fiber density in the muscle was 200 fibers/mm2. Thus the fibers of different MUs could be intermingled. The surface-recorded MUAPs were obtained as the sum of the action potentials generated by the muscle fibers belonging to each specific MU. Sixty-five MUs were simulated in each trial. This value was selected so that the total number of simulated fibers was similar to that of a real muscle in the detection volume (10). The simulated signals corresponded to a full recruitment of the MUs in the detection volume. To obtain the distribution of firing rates, it was assumed that the generated force was that corresponding to the recruitment threshold of the highest threshold MU in the simulated set. The MU recruitment threshold was computed with the exponential rule suggested by Fuglevand et al. (19) and recently applied by Farina et al. (10). Recruitment of MUs was assumed to take place at up to 80% of the maximal force, simulating a muscle that recruits MUs almost until the maximum contraction level, as is the case of the biceps brachii (26). The firing rates were inversely related to the recruitment threshold (34), with a minimum of 8 pulses/s at the recruitment and a maximum of 35 pulses/s. The standard deviation of the interpulse interval was fixed for all MUs to 15% of the mean interpulse interval (Gaussian distribution). The 65 MUs had a Gaussian distribution of CV with a standard deviation of 0.3 m/s. The mean value of the distribution was varied between 3 and 5 m/s to investigate its effect on the EMG variables. The distribution of CV was truncated at 2 and 7 m/s, which are the extremes of the experimentally observed values. The smallest MUs had the lowest CV (1), and the recruitment order followed the size principle (21), with the small and low CV MUs recruited at the beginning of the contraction (and thus having higher firing rate). The firing patterns of the active MUs were first generated independently (no synchronization) on the basis of the calculated mean firing rates and interpulse interval variability (Fig. 1A). After this step, some firing instants were moved to create the desired level of firing synchronization (Fig. 1B). In particular, the synchronization level was defined by two parameters, indicating the percentage of firings in each train synchronized with other firings (%F) and the number of firings synchronized together for each synchronization event [expressed as a percentage of the total number of MUs (%M)]. Given %F and %M, for each train, the preset number of MUAPs was synchronized with others by moving their firing instants. The MUAPs to be synchronized were moved to have their firing instants at a distance described by a Gaussian random variable with 0 mean and 2-ms standard deviation. In summary, after the generation of the independent firing patterns, the steps followed to induce MU synchronization were as follows: 1) compute the number Nm of firings of different MUs synchronized in each synchronization event as %M of the total number of MUs; 2) consider the first MUAP train generated and compute the number Nf of its firings, which should be synchronized with others as %F of the total number of firings of the train; 3) select randomly the Nf firings among those of the train; 4) for each of the Nf firings (reference firings), select randomly the (Nm − 1) MUs, of which one firing should be synchronized with the reference firing; 5) select, for each of the (Nm − 1) MUs, the firing closest to the reference firing; 6) move the selected firings in a position obtained as the realization of a Gaussian random variable with mean value of the reference firing position and standard deviation of 2 ms; and 7) repeat the operations described in steps 2–6 for all of the selected firings and for all of the MUAP trains. Once a firing has been shifted, it cannot be moved in subsequent operations. This procedure has been proposed by Yao et al. (45) and was validated by those authors by comparison of synchronization indexes proposed in the past in simulated and experimental signals. Figure 1 shows the representative firing patterns generated with and without MU synchronization and an example of the generated signal. Figure 2 shows examples of simulated signals with three levels of MU synchronization. The cross-histograms of two randomly selected firing patterns in the set are also reported. Fig. 2.Simulated signals with different degrees of synchronization among MU firing patterns. The percentage of synchronization is the percentage of potentials (with respect to the total no. in the train) of each MU synchronized with potentials of other MUs. The no. of potentials involved in each synchronization event is a percentage of the total no. of MUs (this percentage is equal to the percentage of synchronization). The signals are normalized with respect to the maximum value of the signal with highest degree of synchronization. The cross-histograms between 2 firing patterns (randomly selected) are also reported. See text for details. In each simulation, all of the 65 MUs were active, with firing patterns computed as described above on the basis of their recruitment threshold, size, CV, and degree of synchronization. In the simulations, the two parameters describing MU synchronization (%F and %M) were always set equal to reduce the number of parameters describing central changes. The values chosen were in the range between 0% (no synchronization) and 25% (very high synchronization) at steps of 5%. For each of these synchronization levels, the distribution of CV had a mean from 3 to 5 m/s at steps of 0.5 m/s. To take into account the variability of the results, depending on the locations of the MUs in the detection volume, the MUs were placed randomly in the muscle. For each physiological condition, 50 synthetic signals were generated. For each of the 50 signals in the same set, the positions of the MUs in the muscle were randomly selected (with uniform distribution in the detection volume), whereas the firing patterns (with synchronization), CV distribution, and number of fibers for each MU were fixed for the entire set. Thus, in each physiological condition, 50 anatomic conditions were tested, describing a basic source of variability between subjects. The simulated signals were all 5-s long, with a sampling frequency of 1,024 Hz and noise free. In total, 1,500 synthetic signals (corresponding to 5 CV values, 6 synchronization levels, and 50 signals for each condition) were generated. Ten healthy male volunteers (age, mean ± SD: 27.4 ± 5.2 yr; body mass, 72.3 ± 4.7 kg; stature, 174.3 ± 6.1 cm) participated in the study. No subject had known symptoms of neuromuscular disorders. The study was approved by the local ethics committee, and written, informed consent was obtained from all participants before inclusion. The torque of the elbow flexor muscles was measured with a modular brace, which incorporates two independent torque meters (model TR11, CCT Transducers, Torino, Italy), on each side of the brace. The force signal was digitized by a 12-bit analog-to-digital converter and sampled at 1,024 Hz. A linear array of electrodes (silver bar electrodes, 5 mm long, 1 mm diameter, 10 mm interelectrode distance) was applied between the two tendon regions of the long head of the biceps brachii muscle. A reference electrode was applied at the wrist. Before electrode placement, the skin was slightly abraded with abrasive paste and cleaned with water. The optimal orientation of the array was determined by visual inspection of the EMG signals in a few trials at the beginning of the experimental session. The array was held in place by elastic straps, which did not obstruct blood flow. The EMG signals were amplified and band-pass filtered (3 dB bandwidth = 10–500 Hz), sampled at 1,024 Hz, and converted in digital form by a 12-bit analog-to-digital converter. Subjects were seated comfortably and performed isometric contractions with the arm placed in the isometric torque brace with the shoulder angle at 90° flexion and the elbow angle at 120°. Measurements were performed on the right (always dominant) biceps. The hand was maintained halfway between pronation and supination. Maximal voluntary elbow flexion contraction was measured during three successive trials with 2 min of recovery in between. The greatest value was assumed as the maximal voluntary contraction (MVC) for the elbow flexion. After the determination of the MVC, 5-min rest was given to the subjects. After the MVC assessments, the subjects performed an 80% MVC isometric contraction of the elbow flexors lasting 10 s. Subjects held the contraction, matching a torque target by visual feedback. Each subject repeated the experimental session on 2 different days separated by more than 1 day. From the 15 single differential signals detected by the 16 electrodes of the array spaced by 10 mm, we computed the 14 single differential signals detected by electrodes spaced 20 mm. This was achieved by summation of consecutive single differential signals, as described in Ref. 11. The location of the main innervation zone was assessed by visual inspection of the multichannel EMG signals, and only the channels distal with respect to the innervation zone were considered for further analysis. From these signals, we selected that corresponding to the minimal variability of spectral EMG variables with respect to adjacent locations, as suggested in Ref. 30. From this channel, EMG variables from linear and nonlinear analyses were computed. Thus the signals used for subsequent analysis were detected by a single differential system, with electrodes spaced 20 mm, placed between the innervation zone and the distal tendon region, and providing the minimal variability of spectral variables to shifts of the recording system in that muscle zone. %Det, %Rec, and spectral variables were estimated from the selected signals from adjacent nonoverlapping epochs of 1 s. For each EMG variable, the first-order regression lines were computed from the values obtained during the 10-s-long contractions. The slope of EMG variables was defined as the slope of the regression line and the initial value as the intercept of the regression line. Normalized slopes, defined as the slopes divided by the initial values, were expressed in percentages per second. The experimental data were analyzed by using two-way repeated-measures ANOVA, followed by post hoc Student-Newman-Keuls pairwise comparisons, when required. Threshold for statistical significance was set to P = 0.05. Data are presented as means ± SD. RESULTSFigure 3 shows simulated signals with different mean values of CV distribution and degrees of synchronization level. The recurrence plots and the power spectra are also shown. The difference between the two recurrence maps in the two physiological conditions is evident (see also ). In particular, a higher number of rule-obeying structures in the signal is clearly shown by a higher regularity of the recurrence map. Fig. 3.From top to bottom: raw simulated signals, power spectra with mean (MNF) and median spectral frequency (MDF) values, and recurrence maps and their percentage of determinism (%Det) and recurrence (%Rec). The electromyographic signals are generated with 5 m/s conduction velocity (CV) and 0% MU synchronization level (A) and 3 m/s CV and 25% MU synchronization level (B). The signals in the top row as well as the power spectra in the middle row are in arbitrary units. The recurrence quantification analysis parameters are D = 15, R = 60, λ = 3, andL = 20 (see for definitions), providing a plot of 982 × 982 pixels. Figure 4 shows %Det and MNF as functions of mean CV and synchronization, normalized with respect to the highest mean CV simulated and to the lowest synchronization level. MDF led to results similar to those for MNF. Assuming, for example, decreasing mean CV (31) and increasing synchronization (25) during fatigue, these plots can be interpreted as indicative of the changes in linear and nonlinear variables during fatiguing contractions with normalization with respect to the initial values. Note that, for both CV and synchronization level changes, %Det shows greater relative changes with respect to MNF. The percent increase of %Rec for CV decreasing from 5 to 3 m/s was 130%. The percent increase of %Rec for the degree of synchronization increasing from 0 to 25% was 73%. The average nonnormalized values of %Det for 0% synchronization level and 3 m/s CV were 28.09 ± 12.31 and 22.93 ± 8.71%, respectively. For %Rec, they were 3.10 ± 0.88 and 3.52 ± 1.25%, respectively. The average values for the two conditions were 97.72 ± 16.27 and 101.34 ± 10.90 Hz for MNF and 94.77 ± 16.12 and 97.89 ± 11.46 Hz for MDF, respectively. Fig. 4.%Det and MNF as functions of the mean of CV distribution (A) and the degree of MU synchronization (B). The data are normalized with respect to (wrt) the mean values obtained for the maximum CV value (A) and the minimum degree of synchronization (B). The second-order polynomials interpolating the data are also shown. The mean CV values inA are normalized wrt the maximum value (5 m/s) to compare the relative rates of change of CV, %Det, and MNF. In each case, mean and SD are computed from the 50 anatomic conditions and by varying the nonfixed parameter over all of the values simulated. Figure 5 shows the scatterplot between %Det and MNF normalized with respect to the mean values computed over all of the simulation set (n = 1,500). A high correlation was found between the two variables in the simulated conditions (linear regression analysis, R = −0.95,P < 0.001). MDF showed very similar results, with a correlation with %Det resulting in R = −0.93 (P < 0.001). Changes in MNF or MDF corresponded to opposite changes in %Det, as already observed in Fig. 4. However, the normalized changes in %Det are almost twice those of the normalized MNF and MDF. Fig. 5.Scatterplot of %Det as function of MNF for all of the simulated signals (n = 1,500). The values are normalized wrt mean values (on all of the set of signals) to compare the relative changes of the 2 variables. %Rec showed a decrease with CV increasing and an increase with increasing degree of synchronization, as %Det. However, the correlation with spectral variables was poorer than for %Det (R = −0.74 for MNF and R = −0.72 for MDF), although statistically significant (P < 0.001). All subjects were able to maintain the selected force level constant for the time interval of the contraction within ±5% MVC. For all of the subjects, a statistically significant (linear regression analysis, P < 0.001) decrease in MNF and increase (P < 0.001) in %Det and %Rec was observed during the fatiguing task. As for the simulated signals, MNF and MDF showed very similar results; thus only those from MNF analysis are reported in the following paragraph. The mean MNF, %Det, and %Rec slopes (absolute values) were 2.03 ± 0.55 Hz/s, 2.32 ± 1.04%/s, and 0.27 ± 0.18%/s, respectively. The mean MNF, %Det, and %Rec initial values were 73.15 ± 8.34 Hz, 35.15 ± 9.31%, and 3.63 ± 1.04%, respectively. The normalized slopes (absolute values) were 2.66 ± 0.66%/s, 7.50 ± 5.05%/s, and 8.41 ± 7.32%/s, for MNF, %Det, and %Rec, respectively. A two-way (three EMG variables, MNF, %Det, and %Rec, and 2 days) ANOVA of the normalized slope was significant for the EMG variable (F < 9.38, P < 0.01) and not for the day. The post hoc Student-Newman-Keuls test disclosed pairwise differences (P < 0.01) between MNF normalized slopes and the normalized slopes of both %Det and %Rec. The initial values and normalized slopes of MNF and %Det were significantly correlated. For the initial values, the correlation coefficient was R = −0.78 (linear regression analysis,P < 0.001), whereas for the normalized slopes it wasR = −0.75 (P < 0.001). The correlation between MNF and %Rec was significant (P < 0.05), although, as for the simulations, much lower than that between MNF and %Det (R = −0.55 for the initial values and R = −0.46 for the normalized slopes). Figure6 reports the scatterplots between MNF and %Det initial values and normalized slopes. Fig. 6.Scatterplot of MNF initial values (A) and normalized slopes (B) as function of %Det initial values and normalized slopes for all of the subjects investigated and the 2 days (n = 20). DISCUSSIONAlthough RQA is being used in many studies on muscle assessment, surprisingly, it was not clearly known what information it provides. Mean CV and short-term synchronization were the two phenomena investigated in this work, because they have been proven to affect surface EMG spectral features during fatigue assessment. In this study, we were interested in analyzing the information that is extracted by RQA from the surface EMG signals and to compare this with that obtained by classic spectral analysis. We followed a simulation approach proposed in a previous study for addressing the issue of MU recruitment strategies (10). The main advantage of this approach is to take into account a number of anatomic conditions, thus avoiding results biased by a particular location of the MUs in the muscle. The simulation of the MUAPs was performed with a number of simplifications, which have been discussed already in previous works (9, 10, 13) and which are related to the choices of the simulation parameters (such as the conductivities of the subcutaneous layers), most of them not exactly known in practice. Different choices for these parameters probably would not affect the general conclusions drawn in this study, which are mostly related to the statistical properties of the MU firing patterns and to the relationships between CV and surface MUAPs. Another limitation of the physical model used for the generation of the surface MUAPs is the description of the subcutaneous layers and of the muscle as infinite parallel layers (13, 15). Other models may be better in this respect, as, for example, the approach followed by Kleine et al. (25), who assumed a cylindrical volume conductor, which probably simulates the arm better (40). Again, we do not think our main conclusions would be different by assuming a different volume conductor shape. Moreover, the agreement with the experimental results indicated that the modeling was accurate enough to clearly interpret the experimental findings. The assumed correlation among MU size, firing rate, and recruitment threshold is a consequence of the size principle (21), which has been validated in a number of previous studies. The association among CV, MU size, and firing rate is based on the results reported by Andreassen and Arendt-Nielsen (1), who indicated a relationship between CV and twitch MU properties and thus proposed CV as an additional size principle parameter. The correlation between firing rate and CV was also directly shown by Farina (6) using intramuscular and surface EMG techniques for assessing control and conduction properties of single MUs (7). Some previous studies aimed at the simulation of MU short-term synchronization. Hermens et al. (22) shifted the firings of independently generated firing patterns to create synchronization of all of the firings of MU pairs. The complete synchronization of MU firings was also assumed by Weytjens and van Steenberghe (44) for their analytic derivation of the power spectrum of the surface EMG signal. This simplification implies that the two synchronized MUs have the same mean firing rate. Yao et al. (45) proposed a method to introduce synchronization in a pool of MUs, again based on the shift of MU firings. This approach was validated by the analysis of classic indexes of synchronization (36) applied to the simulated signals and was adopted in the present study, although with a different model of MUAP generation and different simulation modalities. Recently, Kleine et al. (25) proposed a different approach based on the model suggested by Matthews (29), which described the afterhyperpolarization after a firing as an exponential function on which membrane noise is superimposed. The noise was then divided into a common and an individual part to simulate different degrees of synchronization. However, when comparing the synchronization level introduced by their method with that introduced by Yao et al. (45), these authors concluded that similar results were obtained (25); thus the choice of one of the two models is probably not critical in the context of this study. The effect of muscle fiber CV on spectral features of the EMG signal is well known from past studies (2, 31). Theoretically, neglecting the effects of the firing statistics on spectral features, the relative change of CV and MNF or MDF are equal if 1) no MUAP shape change occurs (only scaling), 2) the MU pool is stable (10), and 3) all of the MUs have the same CV. In our simulations,condition 3 was not met, to better approximate experimental signals. However, it was shown (Fig. 4) that the relative changes of MNF (MDF provided the same results) and CV are similar also in this condition. Direct interpretation of the plots of Fig. 4 in terms of changes occurring in CV during fatigue is, however, rather critical, because in real cases it is not known how the distribution of CV changes independently on the mean value, that is, how the relative weight of large and small MUs changes with fatigue. The dependence of %Det and %Rec on CV was not previously known, because the only results in this respect came from experimental fatigue studies in which a number of phenomena occurred and could not be separated from each other. The finding that the two variables extracted from the recurrence maps are changing with CV, as the spectral variables, clarifies that CV is a parameter strongly affecting nonlinear surface EMG analysis. The result can be explained by observing that the number of recurrent structures identified depends on the duration of the segments in the EMG signal, which repeat in time with almost the same shape, i.e., on the duration of the MUAPs. CV is directly related to the duration of the surface MUAPs if the length of the depolarized zone is constant during time, as assumed in the simulations. Thus, the higher CV is, the smaller is the duration of the surface MUAPs. From this point of view, %Det and %Rec detect changes in the frequency content of the signal, a conclusion important for the interpretation of experimental findings. From our results, it clearly appears that there is no possibility of separating the changes due to CV and to the degree of synchronization by using %Det or %Rec. The effect of the degree of synchronization on the spectral frequencies of the surface EMG signal was investigated in some previous works. Weytjens and van Steenberghe (44) extended the derivations of Lago and Jones (27) on the effect of MU firing statistics on the surface EMG power spectrum to the case of dependent firings and showed that cross-terms arise that increase the energy of the signal mostly at low frequencies. Hermens et al. (22) validated this analytic observation by a simulation model, and similar results were shown, more recently, by Yao et al. (45) and, with the inclusion of the electrode location variability, by Kleine et al. (25). Our results are in line with these previous studies and confirm the role of synchronization in determining MNF and MDF patterns during fatigue. %Det and %Rec also depended on MU synchronization, exhibiting an increase for increasing degree of synchronization. This study is the first that clearly shows that these EMG variables increase with an increasing level of MU short-term synchronization; the result is in agreement with the expectation that a higher number of deterministic structures in the signal should be detected by RQA. In previous studies, this dependence was suggested, but, in experimental investigations on muscle fatigue, it was not possible to separate changes of MU synchronization and CV. Felici et al. (16), for example, reported RQA analysis on a group of weight lifters, assuming that a higher degree of synchronization should occur with fatigue in this group with respect to a control group of sedentary persons. They reported a higher rate of change of %Det in the weight lifters with respect to the control group during sustained isometric contractions, but it was not possible to attribute this difference exclusively to the different change in MU synchronization degree, because CV was also changing. One of the main results of this study is the observed high correlation between %Det and MNF (or MDF) in all of the simulated conditions. The regression analysis reported in Fig. 5 shows a correlation coefficient between the two variables close to one. This result indicates that, as spectral variables, %Det also reflects physiological phenomena of interest in muscle assessment. Indeed the %Det changes found in previous studies could not be explained on a physiological basis before the present study. As it always happens in simulation studies, the crucial point is the validation and the extension of the results to experimental paradigms. We designed a simple fatigue protocol, and we assumed that, during the muscle contractions performed by the subjects, phenomena similar to those simulated occurred. The contraction level was selected to exclude recruitment of MUs during fatigue, because it corresponded to almost the maximal level at which MUs are recruited in the biceps brachii muscle. The simulation results were largely in agreement with the experimental findings. We found, indeed, a highly significant correlation between MNF and %Det in the experimental signals also, in agreement with the simulation predictions. Webber et al. (42) already showed a correlation between the rates of %Det and MNF change on 12 subjects who performed fatiguing contractions. We also observed high correlation between initial values. Thus, although many factors of variability among subjects were not included in the simulations for limiting the computational effort (differences of subcutaneous layer thicknesses, variability in the orientation of the detection system with respect to the fibers, length of the fibers, etc.) and within the limitations of the model used, as mentioned above (see Simulation of the MUAPs), the model results were still a valid tool for interpreting the experimental signals. For both the simulations and the experimental signals, %Rec resulted in poorer correlation with spectral variables with respect to %Det. Thus interpretation of %Rec changes from the surface EMG signals may be less direct than that of %Det changes. The observed larger changes in %Det and %Rec in response to fatigue than those in MNF were predicted by the simulation analysis. Webber et al. (42), in their experimental study, indicated an even higher difference between the normalized rate of change in MNF and %Det (changes in %Det were approximately five times larger than those in MNF). The discrepancy of the latter result with those of our study is probably due to the different experimental setup used (for example, interelectrode distances were different). The present results show that the two linear and nonlinear tools tested provide similar indications about CV and MU synchronization changes, two mechanisms that affect surface EMG signal features during fatigue. The claimed higher sensitivity of %Det to MU synchronization than to other fatigue-related changes was not confirmed, and thus RQA cannot be used to separate MU synchronization changes from other factors. During fatigue, in nonpathological conditions, %Det is highly correlated to spectral variables. This is in agreement with previous experimental studies in which RQA and spectral analysis always provided the same indications of muscle fatigue (16, 17, 42), although a correlation between variables obtained with the two tools was reported for experimental data only by Webber et al. (42). The results suggest that no new information is extracted by RQA with respect to that provided by classic spectral analysis. However, another important finding of the present study is that the sensitivity of RQA to muscle changes is higher than that of spectral variables. This was already suggested by Webber et al. and has been demonstrated in the present study on a large set of simulation conditions (including a number of anatomic configurations). From the above considerations, RQA seems promising for detection of muscle changes due to fatigue or other factors and could allow a better separation among groups of subjects and/or muscles. The main conclusions of this work are as follows. 1) Spectral variables, %Det, and %Rec are influenced by CV and the degree of synchronization; thus none of them can be used to separate the two effects during fatiguing contractions. 2) Spectral variables and %Det are highly correlated, and thus the information they provide on muscle properties is similar. The correlation between %Rec and spectral variables is much poorer than that for %Det, although %Rec, like %Det, is similarly sensitive to changes in muscle properties. These results are confirmed by both simulation and experimental findings and are particularly important for the correct interpretation of linear and nonlinear variables when comparing muscles or subjects. 3) Relative variations of %Det (and %Rec) in response to changes in muscle properties are significantly larger than relative variations of spectral variables. Thus nonlinear analysis of the surface EMG signal is a technique more sensitive than spectral analysis for the assessment of muscle fatigue and can be used for more detailed noninvasive muscle assessment. The application of RQA to surface EMG still needs, however, research efforts in determining its sensitivity to other factors of variability between subjects and muscles, such as the thickness of the subcutaneous layers, the orientation of the detection system, the interelectrode distance, the electrode location, the recruitment of MUs, the additive noise, and so on. Whereas the influence of all of these parameters on spectral variables has been analyzed in many past publications (see, for examples, Refs. 9-12, 14,20, 24, 33, 38,47), there is a complete lack of this information for RQA. The authors are sincerely grateful to Prof. Marco Marchetti of the Dipartimento di Fisiologia Umana, Università La Sapienza, Roma, and to Prof. Roberto Merletti of the Dipartimento di Elettronica, Politecnico di Torino, Torino, for useful discussions and comments on the manuscript. Although RQA has been described in detail in some past works, it is necessary in the context of this study to recall the basic principles of this technique. Given the surface EMG signal samples s(m) (m = 0, …, M), the following steps have been followed to compute the recurrence maps and the relative indexes. From the surface EMG, vectors representing the states of the system are extracted. To do this, data are lagged by an integer number λ of samples (“time delay”), which defines the time distance between two uncorrelated samples. Selection of λ is performed on individual time series; in the present study (simulated and experimental data), λ values ranged between 3 and 6. The space dimension (phase space), referred to as embedding dimension, has been selected in the present study as 15 (according to Ref.18). The following set of D-dimensional vectorsv(n) is extracted from the original myoelectric time series v(1)={s(1)s(λ+1)…s[(D−1)λ]}v(2)={s(2)s(λ+2)…s[(D−1)λ+1]}……………v(N)={s(N)s(N+λ)…s[N+(D−1)λ]}Equation 1 where N = M − (D − 1)λ.The states of the dynamic system under consideration are represented by the vectors v(n) defined in Eq.1. To define the recurrences between the states, their closeness should be quantified. To do so, the most commonly used metric is the Euclidean distance (d) d(i,j)={<[v(i)−v(j)]2>}½Equation 2 where < > indicates the summation of the elements of the argument vector.To make the results of the analysis independent of the energy of the observed signal, the values adopted in RQA are normalized with respect to the average distance dav between vectors dav=∑i=1N ∑i≠jN d(i,j)N(N−1)/2Equation 3 where the denominator represents the number of possible distances d(i,j).The collection of the normalized distances provides a symmetric matrix, indicated as distance matrixDM(N × N) DM=d(1,1)d(1,2)…d(1,N)d(2,1)d(2,2)…d(2,N)…………d(N,1)d(N,2)…d(N,N)Equation 4 The recurrences between the states are evaluated from the matrixDM. A recurrence occurs if the distance d(i,j) is smaller than a threshold (referred to as radius R), i.e., if v(i) and v(j) fall within a sphere of radius R centered at v(i). The comparison of the distances in DM with R provides the recurrence map RM as the collection of the pixels b(i,j), defined as follows . The value of R was always set, in accordance with Ref. 46, smaller than 10% of the normalized mean distance (its value was adjusted depending on the data set). Because RM is symmetric, for the computation of the quantitative indexes of the recurrence map, only the upward or downward portion of it is considered.Often nonlinear plots contain subtle patterns that are difficult to detect by simple visual inspection. Thus quantitative descriptors that emphasize different features of the map have been proposed (42). Pixels ON are defined as recurrent (Re) if they share local neighborhoods in higher dimensional space (42). The percentage of plot occupied by Re points is defined as the %Rec. The percentage of Re forming lines parallel to diagonal is defined as %Det. Lines are constituted by two or more points that are diagonally adjacent with no intervening white spaces. Thus %Det is defined as %Det = L/Re × 100, where L is the number of points forming lines. Figure 3 shows examples of recurrence maps computed from simulated signals. The presence of diagonal structures can be clearly observed for the two simulated signals shown. %Det and %Rec quantify this difference, as it appears from Fig. 3. FOOTNOTESREFERENCES
Page 18morphine is often administered to patients for pain management, but it is also recommended for acute myocardial infarction patients to reduce the elevated resting arterial blood pressure and sympathetic tone that is associated with cardiac ischemia. Although morphine is often recognized for its hypotensive effect, recent evidence suggests it may actually increase arterial blood pressure (16). The mechanism underlying morphine-induced increases in arterial blood pressure at rest remains equivocal. Because activation of the sympathoadrenal (11) and renin-angiotensin (1) systems has been suggested to contribute to increases in resting arterial blood pressure with morphine, it is possible that central sympathetic outflow is increased. To our knowledge, only one study has examined the effects of morphine on muscle sympathetic nerve activity (MSNA) at rest (13). Kirno et al. (13) report no change in either MSNA or arterial blood pressure. Because resting arterial blood pressure did not change, it is unknown whether changes in arterial blood pressure induced by morphine are associated with changes in MSNA. The effect of morphine on cardiovascular and sympathetic responses during exercise is also not well documented. Exercise elicits marked increases in MSNA (15, 22) and also purportedly releases various endogenous opioids for analgesic actions (3). Several centers in the brain stem contain opioid receptors (21), including the ventrolateral medulla, which also receives skeletal muscle afferent input (17). Therefore, it is reasonable to speculate that endogenous opioids might modulate MSNA during exercise. Evidence is limited for endogenous opioid modulation of MSNA during exercise in humans. Most studies have used opioid receptor blockade to examine MSNA responses during exercise. The opioid-receptor antagonists naloxone and naltrexone have been reported to either increase (6) or not change (4, 23) MSNA responses during exercise. The effect of opioid receptor agonists on MSNA responses to exercise is less established. Only one study using the opioid-receptor agonist codeine reported no change in MSNA responses during dynamic handgrip (4). Moreover, this study did not examine MSNA responses during isolated activation of the muscle metaboreflex [i.e., postexercise muscle ischemia (PEMI)]. The effect of morphine, a potent opioid-receptor agonist, on MSNA during exercise has not been investigated in humans, but animals have demonstrated an attenuation of cardiovascular and sympathetic responses to exercise with morphine or the opioid-receptor agonist Met-enkephalin (10, 20). Therefore, the purpose of this study was to examine the effect of the opioid-receptor agonist morphine on arterial blood pressure and MSNA responses at rest and during isometric handgrip (IHG) and PEMI. We hypothesized that morphine would increase arterial blood pressure and MSNA at rest and would attenuate sympathetic and cardiovascular responses during IHG and PEMI. Our results indicate that morphine modulates arterial blood pressure and MSNA at rest but not during exercise. METHODSTwelve healthy men (age 18–35 yr) volunteered to participate in the study. Subjects abstained from nicotine, alcohol, and caffeine for a minimum of 8 h before the experiment. After verbal explanation of the testing procedures, all participants signed a written informed consent approved by the Institutional Review Board at the University of Iowa. On each experimental day, subjects performed two bouts of exercise. The first exercise bout was designated as the control trial because no drug intervention was performed. During the second exercise bout, morphine or saline was administered as an intravenous bolus infusion into the nonexercising arm over a 10-min period. The two exercise bouts were conducted in the same order on each day and were separated by at least 35 min of rest. Morphine and saline were administered on separate days, and both the investigator and the subjects were blinded with regard to the drug intervention until analysis of data was completed. During each exercise bout, subjects performed 2 min of IHG (30% maximum voluntary contraction) followed by 2 min of PEMI before (control) and after systemic infusion of morphine (0.075 mg/kg loading dose + 1 mg/h maintenance) or placebo (saline). This morphine dose has been shown to significantly elevate arterial blood pressure (16). The maximal voluntary contraction level was established on each test day by using the peak force generated from three maximal efforts. PEMI was induced 5 s before the cessation of exercise by inflating a blood pressure cuff on the arm to 250 mmHg. Each exercise trial began and ended with a 5-min baseline and 2-min recovery. Multifiber recordings of MSNA were made by inserting a tungsten microelectrode into the peroneal nerve at the head of the fibula of a resting leg; a different leg was used for the morphine and placebo trials. A reference electrode was inserted subcutaneously 2–3 cm from the recording electrode. Both electrodes were connected to a differential preamplifier and then to an amplifier (total gain between 40,000–80,000), where the nerve signal was band-pass filtered (700–2,000 Hz) and integrated (time constant, 0.1 s) to obtain a mean voltage display of the nerve activity. Satisfactory recordings of MSNA were defined by spontaneous, pulse-synchronous bursts that increased during end-expiratory apnea and did not change during stroking of the skin or auditory stimulation (yell). Continuous heart rate (HR) was recorded with a three-lead electrocardiogram. A pneumograph bellows was wrapped around the subject's chest to monitor respiratory rate and to ensure subjects avoided a Valsalva maneuver during IHG. After local anesthesia, a 20-gauge catheter was inserted into a forearm vein for systemic infusion of morphine or saline. Mean arterial pressure (MAP) was derived by using a Finapres positioned on the middle digit of the subject's nonexercising hand. The mean voltage neurograms were displayed together with an electrocardiogram and respiratory pattern on a chart recorder (model ES2000, Gould) at a paper speed of 5 mm/s. The nerve traffic was also routed to a storage oscilloscope and a loudspeaker for monitoring during the study. Successful nerve recordings were obtained in 11 subjects during both the morphine and placebo trials. All data were analyzed in 1-min segments. Baseline data for the morphine and placebo trials were compared by using a paired t-test, and the exercise and PEMI data were analyzed by using a two-within-factor (drug × exercise bout) repeated analysis of variance. Significance was accepted at theP < 0.05 level. All data are presented as means ± SE. RESULTSPreexercise values for all measured variables are presented in Table 1. Preexercise values between the control and saline trials were not different. However, morphine significantly increased MSNA (17 ± 2 to 22 ± 2 bursts/min; P < 0.01) and MAP (87 ± 2 to 91 ± 2 mmHg; P < 0.02) at rest, whereas HR (61 ± 4 to 59 ± 3 beats/min; P < 0.01) decreased. Morphine did not elicit any noticeable side effects.
There were no significant drug × exercise interactions for either the morphine (P = 0.43) or placebo trial (P = 0.63) for total MSNA. Similarly, there was no significant interaction observed for MAP and HR. Exercise did significantly increase MSNA, MAP, and HR (all P < 0.01) across all exercise bouts (Fig. 1). PEMI elicited comparable responses for all variables during both the morphine and placebo trials. Increases in MSNA and MAP during exercise were maintained during PEMI, whereas HR returned to baseline levels. Repeated-measures ANOVA revealed a main effect for drug during the morphine trial for burst frequency (P = 0.02) but not for any other variable. Figure 2 shows that MSNA responses during the first minute of IHG were less with morphine than without (P < 0.05), but responses during the second minute of IHG and PEMI were not different. During the placebo trial, MSNA, MAP, and HR responses to IHG and PEMI were not different between the two exercise bouts. All variables returned to preexercise levels during recovery (not shown). Fig. 1.Muscle sympathetic nerve activity (MSNA), mean arterial pressure (MAP), and heart rate (HR) responses during baseline, the first and second minutes of isometric handgrip (IHG-1 and IHG-2, respectively), and the first and second minutes of postexercise muscle ischemia (PEMI-1 and PEMI-2, respectively) for the morphine and placebo trials. Solid symbols are significantly different from the corresponding baseline value, P < 0.05. Fig. 2.Change (Δ) in MSNA from baseline during IHG-1, IHG-2, PEMI-1, and PEMI-2 during the morphine and corresponding control trial. * Response is significantly different from the morphine trial,P < 0.05. DISCUSSIONThe primary findings of this study are that 1) morphine modulates MSNA, MAP, and HR at rest in humans and 2) cardiovascular and MSNA responses to IHG and PEMI are not altered by the opioid-receptor agonist morphine. The present study provides the first evidence of concurrent increases in resting arterial blood pressure and MSNA with morphine in humans. Furthermore, our results suggest that morphine does not modulate cardiovascular or MSNA responses during exercise. Morphine has been recognized for its hypotensive effect (2,12). Previous studies suggest that morphine decreases arterial blood pressure as a result of multiple mechanisms, including decreases in cardiac and renal sympathetic nerve activity (7, 18), an increase in vagal tone (26), histamine release (5, 19), and venous and arterial vasodilation (14). Although morphine has been reported to predominantly decrease arterial blood pressure, morphine has also been reported to elevate arterial blood pressure in humans (1, 11, 16). Mildh et al. (16) recently reported that intravenous injection of morphine increased MAP, but in this study morphine was injected during ischemia-induced pain (300-mmHg upper arm tourniquet), not at rest. Bailey et al. (1) and Hoar et al. (11) demonstrated that morphine increased arterial blood pressure at rest, but these studies were performed with surgical subjects who were under the influence of other drugs (e.g., diazepam and nitrous oxide). Moreover, the mechanisms responsible for increases in MAP during morphine administration remain equivocal. Previous studies have attributed increases in arterial blood pressure during morphine administration to activation of the sympathoadrenal (11) and renin-angiotensin (1) systems. Another mechanism that could contribute to the rise in arterial blood pressure with morphine is a direct increase in central sympathetic outflow. To our knowledge, only one study has previously examined the effect of morphine on resting MSNA in humans (13). These authors reported that baseline MSNA did not change with intrathecal administration of morphine (0.4 mg) with no change in arterial blood pressure or heart rate. Because of these findings, the mechanisms responsible for the elevated arterial blood pressure with morphine could not be ascertained by the authors. In the present study, morphine increased baseline MAP and MSNA and reduced HR. Differences between our study and the findings of Kirno et al. (13) might be due to different administration routes (intravenous vs. intrathecal). Neural and cardiovascular effects of exogenous opioids appear to be dependent on the dose, the route of administration, and the type of opioid administered (2,12). However, results from the present study are enhanced by several aspects of our research design. First, both the investigators and subjects were blinded with regard to the drug intervention to minimize bias. Second, the same subjects were used for both the placebo and morphine trials to reduce variability that might be observed by using a second random sample of subjects. Therefore, the results from the present study provide new evidence for a possible mechanism responsible for morphine-induced increases in arterial blood pressure at rest. The role of endogenous opioids on sympathetic responses during exercise remains equivocal. Farrell et al. (6) reported that the endogenous opioid-receptor antagonist naloxone augments MSNA responses during IHG in humans. In contrast, several studies have failed to report differences in cardiovascular and sympathetic responses to dynamic or static exercise after administration of an opioid-receptor antagonist (4, 8, 9, 23, 25). Specifically, Ray and Pawelczyk (23) and Cook et al. (4) found no effect of either naloxone or naltrexone on MSNA during isometric and dynamic handgrip in humans. One limitation of previous opioid antagonistic studies examining MSNA during exercise is that the exercise muscle mass (small) and duration (brief) may have prevented activation of the endogenous opioid system. Because opioid-receptor antagonists block, rather than activate, opioid receptors, an interaction between sympathetic responses during exercise and the endogenous opioid system would only be observed if the exercise stimulus were sufficient to activate the opioid system. In humans, it is not currently possible to artificially evoke endogenous neurotransmitter release during exercise. However, exogenous administration of an opioid agonist, such as morphine, should activate opioid receptors during exercise. Currently, the influence of opioid-receptor agonists on sympathetic responsiveness during exercise is not well documented. Using animals, Pomeroy et al. (20) and Hill and Kaufman (10) demonstrated that intrathecal administration of the opioid-receptor agonists morphine and Met-enkephalin analog attenuated sympathetic and cardiovascular responses to exercise. In contrast, Cook et al. (4) demonstrated that codeine (60 mg) does not alter MSNA responses during dynamic handgrip in humans. Because of this apparent conflict in data, we examined MSNA responses during exercise in humans after the intravenous administration of the more potent opioid-receptor agonist morphine. Morphine predominantly activates μ-receptors, which are distributed throughout the rostral and caudal ventrolateral medulla (21). Because the ventrolateral medulla also receives skeletal muscle afferents (17), it is possible that the neural interaction during simultaneous stimulation of μ-receptors and skeletal muscle afferents may influence ventrolateral medulla output of MSNA. However, our results suggest that morphine does not alter MSNA responses during IHG. Although changes in MSNA appear to be attenuated during the first minute of IHG during the morphine trial, MSNA responses during the second minute were not different. One explanation for the attenuation observed during the first minute could be related to the elevated baseline MSNA and MAP during the morphine trial. The elevated MAP may have prevented the small increase in MSNA as observed in the control trial (24). It is also possible that morphine had a specific effect on skeletal muscle reflexes. However, if morphine did have a specific effect on the skeletal muscle reflexes, it would be expected that the effect would have persisted during the second minute of IHG, but this was not observed. Therefore, we conclude that MSNA responses during brief periods of IHG are not altered by stimulation of opioid receptors. It should be recognized that we cannot exclude a role for δ-receptors during exercise because morphine does not bind to this receptor (2). Because IHG engages both mechanoreceptors and metaboreceptors, PEMI was performed to isolate the effects of the muscle metaboreflex. To our knowledge, the influence of an opioid agonist during PEMI has not been reported previously. Morphine did not affect MSNA responses during PEMI. This finding indicates that the muscle metaboreflex, the primary mechanism for MSNA increases during exercise, is unaltered by activation of opioid receptors. In summary, intravenous injection of the opioid-receptor agonist morphine increases arterial blood pressure and MSNA at rest but does not modulate cardiovascular and sympathetic responses to IHG and PEMI. These findings suggest that central sympathetic outflow contributes to increases in arterial blood pressure with morphine. Moreover, we conclude that the activation of endogenous opioid receptors does not modulate cardiovascular and sympathetic responses during forearm exercise in humans. We thank C. A. Sinkey, E. A. Anderson, and G. A. Hill for assistance with this project. FOOTNOTESREFERENCES
Page 19there is a strong demand for effective pharmacological treatment for obesity and few candidate agents. Dexfenfluramine (Dex), an anorexigenic agent acting as a serotonin reuptake inhibitor, is effective while taken but also has been reported to cause a 30-fold increase in the relative incidence of pulmonary hypertension (PH) (1, 17). This, coupled with the association between another anorexigenic compound, aminorex fumarate, and PH could compromise future development of these agents. Although the mechanism is unknown, development of PH only in a small subgroup of patients exposed to anorexigens suggests a genetic predisposition and/or additional risk factors (1,17). This may explain why it has been so difficult to reproduce the disease by feeding Dex to experimental animals (15). Additional factors such as female sex, obesity, or endothelial injury could be required to induce PH in animal models (1, 17, 38,42). Alternatively, Dex could have deleterious as well as protective properties such as blockade of the serotonin transporter 5-HTT, and the protective properties are deficient in susceptible patients (13). Dex-associated PH is accompanied by structural lesions in pulmonary arteries that are similar to those found in patients with primary or advanced secondary PH (19, 30). These proliferative and obliterative changes are associated with stimulation of extracellular matrix glycoproteins, such as collagen, tenascin, and fibronectin (4, 23). We have observed a codistribution of tenascin with proliferating smooth muscle cells in pulmonary arteries from both patients with advanced pulmonary vascular disease and rats that develop PH in association with increased muscularization and loss of distal vessels after injection of the toxin monocrotaline (MCT) (23, 25). In the rats and in cultured cells, induction of elastase activity is related to the upregulation of tenascin production (24, 25) and appears critical to the progression of pulmonary vascular disease (7, 8, 26, 31, 44, 49,51, 52). The activity of endogenous vascular elastase is increased after injection of the toxin MCT in association with pulmonary endothelial injury (44, 49). Furthermore, administration of serine elastase inhibitors largely prevents development, retards progression (49), and even induces regression (8) of MCT-induced PVD. It would therefore seem feasible to hypothesize that, in the setting of an experimental endothelial injury induced in a rat by MCT, the vascular remodeling that occurs as sequelae of the elevated elastase activity would be worsened by Dex, particularly in a female and/or obese rat. Our results indicated that obese female rats lose weight with Dex but PH is not induced even with concomitant injection of MCT. Most surprising was a paradoxical protective effect of Dex on MCT-induced PH in Sprague-Dawley (S-D) rats. We could not attribute this effect to a reduction in elastase activity because Dex administration was associated with an increase in elastase activity that was further augmented by MCT. We therefore speculate that Dex has a protective effect in these rats that overrides the heightened elastase activity, but we could not attribute this to other factors that repress vascular remodeling, such as induction of NO synthase (NOS) (33) or bone morphogenetic protein (BMP) 2 expression (9, 27, 35, 36). Alternatively, Dex prevents the sequelae of elastase activity. MATERIALS AND METHODSThis protocol was approved by the Animal Care Committee of The Hospital for Sick Children, Toronto, Ontario, Canada. To investigate the effects of Dex in inducing or aggravating PH in rats under control conditions or after injection of MCT, 12-wk-old female JCR:LA-corpulent obese (cp/cp, n = 38) and lean (+/?,n = 16) rats (Department of Surgery, the University of Alberta) and 7-wk-old female S-D rats (n = 36, Charles River Breeding Laboratories, Montreal, Quebec, Canada) were used. Because the different strains were being studied in relation to their response to Dex and MCT, we optimized for the obese or lean phenotype by assessing those strains at 12 wk and for the effects of MCT by assessing the S-D rats at 7 wk. The JCR:LA-corpulent rat is one of several strains incorporating the autosomal recessive cpgene originally isolated by Koletsky and has been well characterized (5, 6, 37, 41, 48). Lean rats are bred as a 2:1 mixture of animals heterozygous (+/cp) or homozygous normal (+/+) forcp gene and are phenotypically lean, not distinguishable from the parent strain, and designated +/?. Rats that are homozygous for the cp gene (cp/cp) lack any functional leptin receptors and are phenotypically obese with hyperlipidemia and insulin resistance (41). S-D andcp/cp rats were assigned sequentially to one of four groups: those injected subcutaneously with 60 mg/kg of MCT or an equal amount of 0.9% saline and those treated with or without Dex (5 mg · kg body wt−1 · day−1) in the drinking water from 1 day before MCT or saline injection to day 16. An additional subgroup of S-D rats was included, in which Dex administration was terminated after day 10 to determine whether rebound PH might occur. Because the +/? rats served as a genetic control for the cp/cp strain, the combined Dex + MCT group was omitted. The present dose of Dex was chosen because it produced a sustained decrease in body weight gain and food intake and changed the metabolism of lipids and insulin in JCR:LA-cp rats in previous reports (5,6). The administration of Dex was initiated 1 day before MCT injection, so that body weight reduction induced by Dex would coincide with the early phase after MCT injection. Our previous reports showed that elastase activity is elevated as early as 2 days after MCT injection and is closely associated with the subsequent development of MCT-induced PH (44, 49). On day 14, rats were catheterized under pentobarbital sodium (33 mg/kg ip) anesthesia by a closed-chest technique, as described previously in detail (49). Briefly, a pulmonary artery (PA) catheter of Silastic tubing (0.31 mm ID and 0.64 mm OD) was inserted through the right external jugular vein into the PA. PA pressure was monitored by use of a physiological transducer (MS20, Electromedics, Englewood, CO), an amplifier system (Interface 4600, Gould, Mississauga, Ontario, Canada), and a monitor (V1000, Gould). The catheter was passed under the skin and exteriorized at the back of the animal's neck. At 48 h after catheterization, PA pressure was recorded in ambient air, while the rat was fully conscious. Systolic systemic blood pressure was determined by using a tail cuff attached to a blood pressure analyzer (Linear recorder mark VII WR3101, Graphtec). The hematocrit was determined from a 0.1-ml blood sample. MCT solutions were prepared from the crystalline compound (Trans World Chemicals, Rockville, MD) and dissolved in pH 7.0 buffer (44). Dex (Institut de Recherches Internationales Servier, Creteil, France) was added to the drinking water, as previously described (5). Food and water were provided ad libitum throughout the experiment. The rats were not disturbed until the conclusion of the study, other than for weighing and normal care. After the hemodynamic measurements were completed, lung tissue was prepared for morphometric analysis of the vasculature, as previously reported in detail (49). Briefly, under pentobarbital sodium anesthesia, a midline sternotomy was performed to expose the heart and lungs. The lungs were inflated via the trachea with preheated phosphate-buffered saline and perfused through a PA cannula with a hot (60°C) mixture of radiopaque barium and gelatin at 100 cmH2O pressure for 5 min. Then the lungs were fixed by perfusion through the trachea with 10% formalin until fully inflated. The lungs were then clamped and maintained in fixative for 72 h. A 10 × 10 × 5-mm tissue block, obtained from the midsection of the left lung, was processed for light microscopy. Sections from paraffin blocks were stained by the elastic Van Gieson method. The right ventricle (RV) was dissected from the left ventricle plus septum (LV+S) and weighed separately. The weight ratios RV/(LV+S) and RV/final body weight were calculated. Light-microscopic slides were analyzed blindly without knowledge of the treatment groups, as reported previously (49). Briefly, all barium-filled arteries >15 μm external diameter were assessed at ×400 magnification. Each artery was first categorized according to its accompanying airway (i.e., a terminal bronchiolus, respiratory bronchiolus, alveolar duct, or alveolar wall). The structural type of each artery was determined as muscular (i.e., with a complete medial coat of muscle), partially muscular (i.e., with only a crescent of muscle), or nonmuscular (i.e., no apparent muscle). The percentage of muscular and partially muscular arteries at alveolar wall and alveolar duct level was determined. For all muscular arteries with an external diameter of 50–100 or 101–200 μm, the wall thickness of the media (i.e., distance between external and internal elastic laminae) was measured at two points across the lumen along the shortest curvature and expressed as percent medial wall thickness, calculated as twice the average wall thickness divided by the external diameter. Elastase activity was monitored in isolated central pulmonary arteries by using a sensitive fluorogenic synthetic substrate assay (51), and serine elastase activity was confirmed as the amount of activity inhibited by recombinant human elafin (47) (a gift from Dr. J. Fitton of Zeneca Pharmaceuticals, Macclesfield, UK). One milliliter of Tris assay buffer (50 mmol/l Tris · HCl, 150 mmol/l NaCl, 10 mmol/l CaCl2 2 H2O, 0.02% Brij, pH 8.0) including 33 μg protein (sample) and 8 μmol/l of the synthetic substrate Suc-Ala-Ala-Ala-AMC (Bachem) was added to each cuvette. The samples in each group were assayed in triplicate at 37°C for 20 h. This assay condition was previously validated for serine elastase activity (49,51). The fluorescence of each sample was measured at 380–440 nm by a spectrophotometer (F-4000, Hitachi, Japan). To selectively monitor serine elastase activity, each assay was also performed in the presence of 2 μg/ml of the selective serine elastase inhibitor elafin. The dose of elafin was chosen on the basis of inhibition of human leukocyte and vascular elastase (51). Because previous studies showed that chronic Dex administration augmented endothelium-dependent relaxation of porcine pulmonary arteries (10), Western immunoblotting was performed to determine whether Dex upregulates any of three isoforms of NO synthases (28) expressed in rat lungs. Twelve rats were assigned at random to one of the four groups with or without MCT (60 mg/kg sc) injection or Dex (5 mg · kg−1 · day−1) administration. Dex was initiated 1 day before MCT or saline injection, and lungs were harvested 2 days after MCT or saline injection, on the basis of a previous study showing the amelioration of PH by NO donors (l-arginine was administered for the first week in MCT-injected rats) (33). Crude lung homogenates were prepared as previously reported (31). Briefly, lung tissue was homogenized in 25 mmol/l Tris · HCl, pH 7.4, containing 1 mmol/l EDTA, 2 μmol/l EGTA, 0.1% (vol/vol) 2-mercaptoethanol, 1 mmol/l phenylmethylsulfonyl fluoride, 2 μmol/l leupeptin, and 1 μmol/l pepstatin A on ice with a homogenizer (Polytron, Switzerland). The homogenate was centrifuged at 1,500 g at 4°C for 10 min to remove cell debris. Aliquots of 40 μg of total protein, as determined by Bradford protein assay (BioRad Laboratories, Hercules, CA), were electrophoresed under reducing conditions by SDS-PAGE on 8–16% polyacrylamide Tris-glycine gel (Helix, Mississauga, Ontario, Canada) and transferred onto a polyvinylidene difluoride membrane (Millipore, Bedford, MA). Nonspecific binding was blocked by incubating the blot in blocking buffer (5% dry nonfat milk in 10 mmol/l Tris, pH 7.4, 50 mmol/l NaCl, and 0.5% Tween 20) for 1 h at room temperature. The blot was then incubated with primary antibodies against three isoforms of NOS in blocking buffer at room temperature for 1 h. The primary antibodies used were anti-endothelial NOS monoclonal antibody (diluted 1:500, Transduction Laboratories, Lexington, KY), anti-neuronal NOS polyclonal antibody (diluted 1:1,000, Affinity BioReagents), and anti-inducible NOS polyclonal antibodies (diluted 1:1,000, Transduction Laboratories, Lexington, KY; diluted 1:2,000, Affinity BioReagents; and diluted 1:200, Santa Cruz, CA). Blots were washed with TBS-T (10 mmol/l Tris · HCl, pH 7.4, 50 mmol/l NaCl and 0.5% Tween 20) for 30 min, incubated with horseradish peroxidase-conjugated goat anti-murine (1:5,000) or anti-rabbit (1:5,000) antibodies (Sigma Chemical) at room temperature for 1 h, and washed with TBS-T for 30 min, and protein bands were visualized by enhanced chemiluminescence (ECL, Amersham) and quantified by scanning soft-laser densitometry (Bio-Rad Gel Doc 1,000). Equal loading and transfer of proteins were confirmed by visualizing proteins after staining the gel with Coomassie blue. Paraffin-embedded lung tissue samples from S-D rats treated with MCT and/or Dex or control untreated rats were used. Slides were deparaffinized and hydrated with xylene and a graded ethanol series (100, 90, 70, or 50% ethanol and double-distilled water). Endogenous peroxidase activity was quenched with 3% hydrogen peroxide in methanol for 30 min. Immunostaining was performed by using a Vectastain anti-goat kit. Incubation of primary antibody for BMP-2 (Santa Cruz) was in a 1:250 dilution for 1 h in a humidified chamber at room temperature. Antigens were visualized by using diaminobenzidine substrate. Data are presented as means ± SE. Differences between treatment groups were determined by a one-way ANOVA, followed by Student-Newman-Keuls test for food intake and weight gain or the Scheffé test for hemodynamic and morphometric studies. A level ofP < 0.05 was statistically significant. RESULTSThe cp/cp rats in all experimental groups had similar initial body weights. Control and MCT-injected rats gained weight gradually over the experimental period, related to a steady food intake. Dex administration was associated with a decrease in both weight gain and food intake evident on day 3, which persisted through the experimental period (P < 0.05 for all comparisons at each time point) (Fig.1). In Dex-treated rats, MCT injection caused a further reduction in weight without a further decrease in food intake (the difference was significant on day 3, but a trend persisted throughout the experimental period). The +/? rats in all experimental groups showed only a trivial increase from similar initial body weights over the experimental period attributed to a low food intake, and that was unaffected by MCT or Dex (Fig. 1). Fig. 1.Effect of administering dexfenfluramine (Dex) on the body weight gain (left) and food intake (right) incp/cp (A), +/? (B), and Sprague-Dawley (S-D) rats (C) under control conditions or after injection with monocrotaline (MCT). Dex administration significantly reduced body weight gain and food intake from day 3 in cp/cp and S-D rats, but not in +/? rats. MCT significantly reduced body weight gain and food intake fromday 7 in S-D rats, but not in cp/cp or +/? rats. Dex and MCT had no additive effects on body weight gain or food intake in any of the groups, except day 3 incp/cp rats with respect to body weight gain. Number of rats is indicated in parentheses after the group description. Values are means ± SE. Dex+MCT:long, MCT-injected rat group to which Dex is administered for the entire experimental period; Dex+MCT:short, MCT-injected rat group to which Dex is administered until day 10. * P < 0.05 vs. control rats; † P < 0.05 vs. MCT-injected rats; ‡ P < 0.05 vs. Dex-treated rats. Normal S-D rats gained weight steadily, but on days 7–14 those treated with MCT showed a reduced growth curve consistent with a decrease in food intake (P < 0.05 for all comparisons at each time point). Dex decreased body weight and food intake from the earliest time point after administration in both control and MCT-treated rat groups (P < 0.05 at each time point). Despite the shorter time interval of Dex administration in one of the MCT subgroups, there was no tendency for catch-up in weight gain. Although there appeared to be an additive effect on weight reduction in the MCT rats treated with Dex, this was only evident in one of the subgroups and unlikely to be of significance. Among the groups of cp/cp rats or cp/? rats, values for mean PA pressures were similar (Fig.2) regardless of treatment. In S-D rats, mean PA pressure was similar in control rats with or without Dex treatment (17.2 ± 0.3 and 17.2 ± 0.2 mmHg, respectively). In MCT-injected rats, the PH observed (mean PA pressure 28.3 ± 0.5 mmHg, P < 0.05 vs. controls) was reduced to control levels by Dex given during the entire experimental period and partially reduced by Dex given until day 10 (19.8 ± 0.7 and 20.1 ± 1.0 mmHg, respectively, P < 0.05 vs. MCT rats). The systolic aortic pressure was similar in all treatment groups, as was the hematocrit value (data not shown). Fig. 2.Effect of Dex administration on mean pulmonary arterial pressure (left) and right ventricle (RV) to left ventricle plus septum [(LV+S); RV/(LV+S)] weight ratio (right) incp/cp (A), +/? (B), and S-D rats (C) under the control conditions or after injection with MCT. Neither Dex administration nor MCT injection had any effect on pulmonary arterial pressure or RV/(LV+S) ratio incp/cp or +/? rats. In S-D rats, MCT caused an increase in pulmonary arterial pressure and RV/(LV+S) ratio, which were reduced by Dex administration. Groups administered with Dex for the entire period or until day 10 showed the same effects. Number of rats is indicated in parentheses after the group description. Values are means ± SE. * P < 0.05 vs. control rats; † P < 0.05 vs. MCT-injected rats; ‡ P < 0.05 vs. Dex-treated rats. Indexes of RV/(LV+S) (Fig. 2) and RV/final body weight (not shown) were similar in all control and Dex-treated cp/cp, +/? and S-D rats. In S-D rats, MCT induced RV hypertrophy, as indicated by a higher RV/(LV+S) ratio and RV/final body weight ratio (P < 0.05 vs. controls) and consistent with the PA pressure data. Also, in keeping with the PA pressure, RV hypertrophy was not apparent in the groups treated with MCT + Dex for the entire experimental period or until day 10. Obese (cp/cp) and lean (+/?) rat groups exhibited similar degrees of medial wall thickness regardless of treatment with Dex or MCT. Although values for muscularization of peripheral arteries at the alveolar wall level (Fig. 3) or alveolar duct level (data not shown) tended to be higher with MCT, no significant differences were found. Control S-D rats with or without Dex had similar degrees of medial wall thickness for vessels 50–100 μm (Fig. 3) and 100–150 μM (data not shown) and muscularization of peripheral arteries at alveolar wall level and alveolar duct level. In the MCT-treated S-D rat groups, the percent wall thickness at 50–100 and 100–150 μm increased, as did the degree of muscularization of peripheral arteries at alveolar wall and alveolar duct level (P < 0.05 vs. controls). MCT + Dex treatment both for the entire and 10-day experimental periods similarly reduced the medial wall thickness at both 50–100 and 100–150 μm as well as the percent muscularization at alveolar wall and alveolar duct level (P < 0.05 vs. MCT rats). It is interesting that, in the group with MCT + Dex treatment for entire experimental period, values were similar to those in control rats that were not injected with MCT. In the MCT-injected group treated with Dex until day 10 only, there was a trivial but significant increase in muscularization of arteries compared with control saline-injected rats. Fig. 3.Effect of administering Dex on %medial wall thickness (arterial diameter: 50–100 μm) (left) and %muscularization (alveolar wall level) (right) incp/cp (A), +/? (B), and S-D rats (C). Neither Dex administration nor MCT injection had significant effects on %medial wall thickness and %muscularization incp/cp or +/? rats, although a trend for an increase was observed in %muscularization in both groups injected with MCT. In S-D rats, MCT caused an increase in %medial wall thickness and %muscularization, which was inhibited by concomitant Dex administration. Groups administered with Dex for the entire period or until day 10 had the same effects. Number of rats is indicated in parentheses. Values are means ± SE. * P < 0.05 vs. control rats; † P < 0.05 vs. MCT-injected rats; ‡ P < 0.05 vs. Dex-treated rats. Dex:long, MCT-injected rat group to which Dex is administered for the entire experimental period; Dex:short, MCT-injected rat group to which Dex is administered until day 10. To investigate the mechanism whereby Dex ameliorated MCT-induced PH in S-D rats, we pursued the possibility that Dex inhibits elastase activity stimulated 2 days after MCT injection in intact pulmonary arteries. Dex alone increased elastase activity more than twofold, whereas MCT increased elastase activity by approximately fourfold (P < 0.05 vs. control), and there was an additive effect with both Dex and MCT (Fig. 4). Fig. 4.Effect of administering Dex on elastase activity in pulmonary arteries isolated from rats under control conditions or 2 days after injection with MCT. Dex administration increased elastase activity 3.5-fold, compared with controls. MCT increased elastase activity in control and Dex-treated rats. Number of rats is indicated in parentheses. Values are means ± SE. * P < 0.05 vs. control rats; † P < 0.05 vs. MCT-injected rats; ‡ P < 0.05 vs. Dex-treated rats. To explain the apparent contradictory effect of Dex in suppressing the PH, RVH, and PA changes resulting from MCT, while at the same time increasing elastase activity, we addressed whether there might be induction of a concomitant overriding protective effect. Because Dex administration has been associated with improved endothelial-dependent relaxation, we monitored expression of NOS isoforms in lung tissues (28). We were unable to show either a Dex or MCT effect on endothelial NOS protein expression. Inducible NOS immunoreactivity was not detected in lungs from any of the rat groups when using three different antibodies. nNOS expression was about 1.5-fold increased by Dex but also by MCT alone (P < 0.05 vs. control for both), with no additive effects noted (Fig.5). Fig. 5.Effects of Dex on endothelial (e) or neuronal (n) nitric oxide synthase (NOS) expression in lungs harvested from S-D rats under control conditions or 2 days after injection with MCT. Neither Dex administration nor MCT injection affected eNOS expression, and nNOS expression was similarly increased by Dex and MCT, without additive effects. Number of rats is indicated in parentheses after the group description. Values are means ± SE. * P < 0.05 vs. control rats. No increase in BMP-2 immunoreactivity that could be related to vascular smooth muscle cell growth suppression (35) was observed in response to Dex treatment of rats under control conditions or after injection of MCT (data not shown). DISCUSSIONOur study was designed to show whether chronic administration of Dex could cause or aggravate PH and PVD when combined with female gender in an animal that was obese and induced to lose weight or in which the pulmonary vascular endothelium was injured by previous exposure to a toxin. Neither obesity per se nor female gender appeared to interact with Dex in producing PH even when combined with pulmonary vascular endothelial injury. In fact, the obese (cp/cp) as well as the counterpart control lean (+/?) rat appeared to be insensitive to the PH-producing effects of the MCT toxin. It is possible that a metabolic alteration that results from the abnormal weight gain or lack thereof interferes with the activity of MCT, in which case a measurement of elastase activity would be of interest. This resistance to the development of PH may also be a strain specific effect and may be related to the vascular smooth muscle cells of the cp/cp rats, which are reported to show a decreased mitogenic response (2). Unexpected was the paradoxical protective rather than deleterious effect of Dex on MCT-induced PH and PVD in the normal S-D rats, with lack of any rebound PH on withdrawal of Dex. Further investigation of the mechanism resulted in the observation that Dex alone induced and, in combination with MCT, augmented the activity of elastase, the enzyme linked pathobiologically to the development and progression of experimental PH in rats. This suggested that Dex might have an overriding protective effect that could counteract or block the sequelae of elastase activity. We suspected a Dex-mediated induction of NOS on the basis of previous studies showing the amelioration of PH with NO donors (l-arginine) (33), but this did not appear to be the case. We also evaluated the expression of BMP-2 by immunohistochemistry because this agent represses vascular smooth muscle cell proliferation (35) and because aberrant signaling through this pathway has been established as a genetic basis for familial PH (9, 27, 36). We found no evidence for an upregulation of this protein in response to Dex. This does not exclude the possibility that Dex may still positively influence signaling through the BMP receptor II pathway in a manner independent of increasing ligand expression. It is difficult to compare the pathobiology in patients with anorexigenic PH with that of rats with MCT-induced PH. Our studies may, however, explain why Dex causes PH in only a subgroup of patients, in which it is conceivable that the deleterious effects are left unbalanced by deficiency of a still unknown protective mechanism. One possibility is Dex-induced blockade of serotonin transport, discussed below. Alternatively, elastase may be necessary but not sufficient, and additional undetermined genetic abnormalities may be required to induce PH in Dex-susceptible patients. A variety of mechanisms have been explored to identify deleterious effects of Dex that could result in PH. Because Dex inhibits serotonin reuptake by interacting with its transporter in platelets and endothelial cells, serotonin could be involved in Dex-induced PH, if, as has been recently shown, there is concomitant heightened function of the serotonin transporter (14, 29). It is of interest that mice lacking the serotonin transporter (13) did not develop PH induced by hypoxia, although mice treated with Dex, which blocks serotonin uptake, were not protected against PH during exposure to chronic hypoxia (15). It could be that, in the hypoxia model, the transporter is induced or activated to offset the suppressant effects of Dex (Fig. 6). Serotonin induces platelet aggregation and is a powerful pulmonary vasoconstrictor with mitogenic effects on smooth muscle cells. In fact, a case of platelet storage disease with primary PH was reversed by a serotonin antagonist ketanserin (20), and elevated plasma serotonin is observed in patients with primary PH (21). The Fawn-hooded rat, which has a genetic defect in serotonin platelet storage, develops PH on exposure to moderate hypoxia (43). Because serotonin stimulates collagenase activity in rat smooth muscle cells (11), serotonin could also be associated with augmentation of the proteolytic activity contributed by elastase in PH. Fig. 6.Schema adapted after Ref. 29 speculating how in pulmonary artery smooth muscle cells, Dex, by inhibiting serotonin (5-HT) reuptake increases 5-HT interaction with the receptors 5-HT2A and 5-HT1B, inducing an intracellular signal that is reflected in elevated elastase activity. However, by inhibition of the transporter, Dex might prevent intracellular accumulation of 5-HT. The latter, we speculate, might mediate signaling that is relevant to the sequelae of elastase activity, e.g., tenascin induction, responsivity to growth factors, etc. Dex inhibits K+ currents and induces pulmonary vasoconstriction in rats (46). The vasoconstrictive effects of Dex per se in isolated rat pulmonary arteries have, however, been documented only at concentrations >100 times higher than plasma levels measured in humans (45, 46). In the intact lung, however, vasoconstrictive effects have been documented (46) at a concentration similar to that in human plasma, i.e., 10−7. In humans, Dex, however, similarly inhibits the voltage-gated K+ channel, which is dysfunctional in PA smooth muscle cells harvested from patients with primary PH (50). It is possible that, in susceptible patients, a combination of marked alteration in K+ currents coupled with elastase activity could lead to PH. The Dex-induced pulmonary vasoconstriction attributed to decreased K+ channel activity was enhanced by NOS inhibitors (32). Further clinical studies showed that ventilatory NO production is significantly deficient in patients with Dex-associated PH compared with Dex-unrelated PH (3). This suggested that a genetically determined deficiency in NO production, observed in patients with Dex-associated PH, could underlie the predisposition to severe PH. Consistent with this, our recent studies have shown that NO donors inhibit elastase activity in cultured PA smooth muscle cells (34). Taken together, Dex might require concomitant NO deficiency to induce elastase activity of the magnitude sufficient for the induction of PH. How Dex induces protection against PH in response to MCT could not, however, be explained by an inductive effect on NO, at least as judged by NOS expression, and remains to be understood. The protective effects of Dex on PH in this experimental model suggest that although this agent induces elastase it might concomitantly inhibit the sequelae of this enzymatic activity that leads to vascular disease. For example, in cultured smooth muscle cells, elastase liberates growth factors from the extracellular matrix in an active form and also, either directly or via metalloproteinases, upregulates the glycoprotein tenascin-C expression, which amplifies the proliferative response (24, 25). In addition, the products of elastase activity, elastin peptides, stimulate smooth muscle cell migration through induction of fibronectin (22). It is possible that Dex alters the response of smooth muscle cells to liberated growth factors or in some way prevents the induction or the response to tenascin-C or fibronectin. Inhibitors of serotonin reuptake, including fluoxetine, have been shown to decrease the mitogenic activity of serotonin in rat smooth muscle cells (13). To support this, the increase in muscularization of pulmonary arteries elicited in chronic hypoxic rats that is aggravated by administration of serotonin is suppressed by coadministration of Dex (12). It is tempting to speculate that a cell, which is genetically transformed so that it lacks a protective mechanism, may proliferate in response to Dex in a manner similar to the transformed fibroblast cell line in which there is activation of the serotonin receptor (2B) and stimulation of the mitogen-activated protein kinase pathway (16). Another possibility that could be considered is an interaction between Dex and MCT that negates the effects of MCT because both agents are metabolized by the cytochrome P-450 pathway in liver microsomes. This is unlikely, however, because Dex does not inhibit cytochrome P-450 3A activity (18) that is induced by MCT (40). In summary, the present study, albeit in an experimental animal model of PH, suggested that Dex could induce PH by aggravating elastase activity, if there was a concomitant loss of a protective mechanism. Future studies elucidating mechanisms of serotonin transport and BMP signal transduction might provide a clue to the nature of the protective mechanism that could be lacking in the patient subgroup that develops fatal PH after Dex. We thank Dr. Kazuo Maruyama (Department of Anesthesiology, Mie University, Mie, Japan) for the supply of the Silastic tubing used for rat catheterization. We thank Lily Morikawa and other members of the Department of Pathology at The Hospital for Sick Children for assistance with preparation of tissues for histology. We are indebted to the staff of the Animal Care Facility at the Hospital for Sick Children for support with animal care. We are grateful to Claire Coulber for technical help, and to Joan Jowlabar, Judy Matthews, Jeannie Carveth, and Judy A. Edwards for administrative and secretarial support. FOOTNOTESREFERENCES
Page 20sensations associated with breathing against external mechanical loads have been studied by using psychophysical methods (12, 21, 22, 36,38). Load detection is one of the two perceptual processes of respiratory mechanosensation (13). It has been shown that the threshold for detection of resistive loads was a constant fraction of the baseline resistance (36), which is known as the Weber fraction. The role of afferent feedback from the lung and lower airways, which is one of the sensory systems that may be involved in load perception, remains controversial. Two strategies have been adopted to determine the role of pulmonary receptors in respiratory load perception: either the principal afferent nerve (vagus nerve) is selectively blocked (17); or alternatively, all other possible sources are eliminated leaving only the vagi intact (4). High-level quadriplegic subjects with a tracheostomy provide indirect evidence about the role of pulmonary afferents in respiratory sensation, because both respiratory muscle afferents and upper airway receptors are bypassed, leaving only the pulmonary receptors intact (4). It was reported that these patients could reliably detect changes in tidal volume as little as 100 ml, which was comparable to that of normal subjects (4). These data suggest that pulmonary stretch receptors can provide conscious perception of volume, at least in the absence of all other signals. In contrast, other studies reported that the detection threshold of mechanical loads was not affected either by bilateral block vagus nerve (17) or after upper and lower airway anesthesia (7,9, 10) in normal subjects. However, it is possible that some pulmonary stretch receptors may escape anesthesia because the anaesthetic agents could not penetrate to the smooth muscle. Lung transplantation recipients provide a good model to study the role of lung and lower airway receptors in respiratory sensation because the afferent information from receptors located distal to the surgical anastomosis are interrupted. Tapper et al. (34) compared the detection threshold of inspiratory resistive load in heart-lung transplant recipients and normal subjects, and they found no significant difference in Weber fraction between lung transplant recipients and healthy control subjects. On the other hand, Peiffer et al. (27) found that the slope of the linear relationship between the Borg scores and peak inspiratory mouth pressure (Pm) associated with breathing against resistive loads was significantly lower in lung transplant recipients. However, magnitude estimation and load detection are two different perceptual processes, and they may involve different neural mechanisms. The objective of this study was to investigate the role of afferent input from lung and lower airways in detection of external inspiratory resistive loads by recruiting double-lung transplant (DLT) recipients as a lung denervation model. Unloaded and loaded breathing patterns were also compared between DLT recipients and matched normal (Nor) subjects. We hypothesized that the absence of pulmonary afferents in those DLT recipients would result in a higher detection threshold and Weber fraction compared with those in Nor subjects. METHODSA total of 10 DLT patients and 12 Nor subjects were recruited in this study. All subjects were Caucasian. The DLT subjects were recruited from the University of Florida Medical Center. None of the DLT subjects had any evidence of current respiratory or neurological disease. The time since the DLT patients received transplant surgery varied from 1.5 to 5.5 yr. None of these patients had evidence of rejection when they participated in this study. All the DLT subjects were on immunosuppressive (Imuran, Prograf, etc.) and steroid medications (prednisone, etc.) when they participated in the study. Forced vital capacity (FVC) and forced expiratory volume in 1 s (FEV1) were tested for each subject. Subjects with a FVC or FEV1 <70% of predicted values were excluded from this study. Only one DLT subject was excluded from this study because of abnormal lung function (FVC: 60.8% of predicted value; FEV1: 41.9% of predicted value). The Institutional Review Board of University of Florida reviewed and approved this study. All participants provided informed consent before participating in this study. Subjects were asked to refrain from strenuous physical activity, large meals, and caffeine for at least 4 h before the test. All subjects performed pulmonary function testing in the sitting position. Standard instructions according to American Thoracic Society Standards for spirometry testing were given to each subject. All subjects performed a FVC maneuver. Each test was repeated two to four times with at least a 1-min rest between each repetition. Values were also contrasted with age- and gender-predicted values. Background respiratory resistance was measured by using the forced oscillation method. The subject was seated in front of the apparatus and breathed “normally” through the mouthpiece, with his or her cheeks supported by both hands. Approximately 10 tidal breaths were collected continuously to analyze the resistance by computer (Jaeger Toennies, Medizintechnikmit System, version 4.5). The test was repeated at least three times for each subject with a 1-min rest between repetitions. The average of three measures was used as the subject's respiratory system resistance. Inspiratory muscle strength was measured as the maximal inspiratory pressure (MIP). Subjects were in a standing position when they performed the test. After exhaling to residual volume, subjects were instructed to place their lips around the mouthpiece and inspire as forcefully as possible with their nose clamped. The test was repeated until three measurements within 10% variation were obtained. There was at least a 1-min rest between repetitions. The maximal value obtained was recorded as the subject's MIP. During the detection experiment, the subject was seated in a lounge chair in a sound isolated chamber, separated from the experimenter and the experimental apparatus. The subjects were instructed to breathe normally through a mouthpiece connected to a non-rebreathing valve (2600 series, Hans Rudolph) with their nose clamped. The inspiratory port of the valve was connected to the resistive loading manifold (model 4813, Hans Rudolph). Pm was measured at the center of the non-rebreathing valve and recorded on a polygraph. Inspiratory airflow was measured by a differential pressure transducer (model MP45, Validyne) and signal conditioner (model CD316, Validyne) connected to a pneumotachograph. Inspired volume (Vi) was obtained by electrical integration of the airflow signal. Pm, inspiratory airflow, and Vi were recorded on a polygraph (model 7, Grass Instruments) and stored and analyzed on a computer (Chart, Powerlab AD Instrument). The resistive loads were sintered bronze disks placed in series in the loading manifold and separated by stopped ports. The load was applied for the entire inspiration and then removed. The subjects were asked to press the signal button held in their dominant hand as soon as they sensed the presence of a load. A series of test loads were presented in a practice session to familiarize the subject with the load sensation and the range of loads. During the subsequent experimental session, the subject listened to music of their choice to mask experiment sounds. A series of resistive loads (0.2, 0.8, 1.24, 1.64, 2.48, 3.26, 6.95, and 11.46 cmH2O · l−1 · s) were presented in a randomized block design, with each loaded breath separated by two to six unloaded breaths. A total of 10 presentations of each load magnitude were presented in two experimental trials with a 5-min break between trials. The subject was monitored by video camera throughout the experiment. The number of detections for each load was summed and divided by the number of total presentations to obtain the detection percent for each load. The detection percent was plotted against the magnitude of added load. The detection threshold (ΔR50) was determined as the magnitude of load corresponding to a detection percent of 50%. The Weber fraction (ΔR50/R0) for each load was computed by dividing ΔR50 by the sum of the subject's respiratory background resistance (R0) and the resistance of the apparatus. The resistance of the apparatus was 1.6 cmH2O · l−1 · s. Detection latency for each load was also computed by measuring the time from the start of inspiratory flow to the onset of the detection signal. A two-tailed t-test was used to compare ΔR50 and Weber fraction between the DLT and the Nor groups. A two-way repeated-measures ANOVA was performed to study the effects of group and load on detection percent and detection latency. Contrast analysis was performed to compare the effect of different loads. The P value for each contrast test was corrected by dividing 0.05 by the total number of contrasts. Peak Pm, Vi, peak inspiratory airflow, inspiratory duration (Ti), expiratory duration (Te), time to peak airflow (TP), breathing frequency (f), and minute ventilation (V˙e) were recorded for each loaded breath and the preceding unloaded breath, which was the control breath. The breathing pattern was compared by using two-way repeated-measures ANOVA to study the effects of group and load. Contrast analysis was performed to compare the effects of different resistive loads. The Pvalue for each contrast test was corrected by dividing 0.05 by the total number of contrasts. The descriptive statistics of all the variables were calculated and expressed as means ± SE. Significance level was set at 0.05, unless multiple contrast analysis was used. RESULTSThe group mean demographic characteristics and pulmonary functions of all the subjects who participated in this study are shown in Table1. The DLT and the Nor group were comparable in age, height, and weight. Background respiratory resistance and MIP between the two groups were not significantly different. Both FVC and FEV1 were significantly lower in the DLT group than in the Nor group (97.0 ± 4.9% of predictive value vs. 119.5 ± 5.0, and 83.8 ± 6.1 vs. 108.3 ± 3.7% of predictive value, respectively). However, both FVC and FEV1 were still within normal range in the DLT group. Furthermore, FEV1/FVC ratio was not significantly different between the two groups (88.1 ± 6.0% of predicted value for DLT vs. 93.9 ± 2.8% of predicted value for Nor; P = 0.371).
During the load detection experiment, all the subjects were asked to breathe “normally.” There was no cue or airflow targeting. Two-way repeated-measures ANOVA found no significant group, load, and interaction effects on Pm, Vi, peak inspiratory airflow, Ti, Te, TP, f, and V˙e during control breathing. During loaded breathing, group and interaction effects were also not significant, except for the interaction effect ofV˙e. The main effects of load were significant for all the above breathing pattern measures. As the magnitude of the resistive loads increased, Pm, Ti, and TP increased, whereas Vi, airflow, Te, f, and V˙edecreased. The Pm, peak inspiratory airflow, and Tiresponse during control breathing and loaded breathing are shown in Fig. 1. Fig. 1.Ventilatory responses during both control breathing and breathing against different level of resistive loads in terms of peak inspiratory mouth pressure (A), peak inspiratory flow (B), and inspiratory duration (C) in the double-lung transplant (DLT) group (●, loaded breathing; ▾, control breathing) and the normal (Nor) group (○, loaded breathing; ▿, control breathing). Values are means ± SE; n, no. of subjects. The total number of false-positive responses during the detection experiment were compared between the DLT and Nor groups, and no significant difference was found (4.8 ± 2.7 for DLT vs. 4.2 ± 1.1 for Nor; P = 0.621). The detection threshold (ΔR50) was significantly higher for the DLT subjects (2.91 ± 0.5 cmH2O · l−1 · s for DLT vs. 1.55 ± 0.3 cmH2O · l−1 · s for Nor; P < 0.05). The Weber fraction was also significantly elevated in the DLT group (0.502 ± 0.1 in DLT vs. 0.295 ± 0.05 in Nor; P < 0.05). The results of detection threshold and Weber fraction are shown in Figs.2 and 3, respectively. Fig. 2.Detection thresholds of the DLT group and the Nor group. * DLT group had a significantly higher detection threshold than the Nor group (P < 0.05). Fig. 3.Scatterplot of Weber fraction in the DLT group and the Nor group. * Group mean value of Weber fraction for the DLT group was significantly higher than that of the Nor group (P< 0.05). Detection latency and detection percent in response to different level of resistive loads are shown in Figs. 4and 5, respectively. Detection latency decreased and detection percent increased as the magnitude of the resistive load increased. Two-way repeated-measures ANOVA found significant group (P < 0.05), load (P< 0.001), and interaction effects (P < 0.05) on detection percent. The DLT group had significantly lower detection percent than the Nor group. Detection latency did not display significant effects of group (P = 0.67) or load (P = 0.084), nor was there significant interaction (P = 0.422) between the two factors. Fig. 4.Detection latency of the DLT group and the Nor group at different levels of resistive loads during loaded breathing. Values are as means ± SE;n, no. of subjects. Fig. 5.Detection percent of the DLT group and the Nor group at different levels of resistive loads during loaded breathing. Values are means ± SE;n, no. of subjects. * DLT group had a significantly higher detection percent than the Nor group (P < 0.05). DISCUSSIONThe results of the present study showed that both loaded and control breathing patterns were similar in the DLT group and the Nor group during breathing against inspiratory resistive loads. Despite a similar ventilatory response to resistive loads, detection threshold and Weber fraction were significantly elevated in the DLT group compared with the Nor group. These results suggest that the lung vagal afferent inputs are not essential to the regulation of resting breathing pattern and load compensation response. In addition, resistive load detection occurs in the absence of lung vagal afferents, yet these afferents contribute to load detection. The absence of pulmonary vagal afferents in many mammals has been found to be associated with an increased tidal volume and reduced f (11, 28), which were believed to be due to abolition of pulmonary stretch receptor input to the Hering-Breuer inflation reflex. However, the inflation reflex is relatively weak in humans compared with animals and is demonstrable only with large inflations. It would thus be expected that the vagal influence on breathing pattern might be substantially less in humans. Vagal blockade (18) and airway anesthesia (37) experiments have failed to show any significant effects of pulmonary afferents on resting breathing patterns in humans. However, the results from those studies may be confounded by the technical limitation on the completeness of lung deafferentation. Lung transplantation interrupts afferent traffic from receptors located distal to the surgical anastomosis, thus providing a model to investigate the role of lung vagal afferents in regulation of breathing in human. In this study, the breathing pattern recorded for both loaded breathing and the breathing cycle before each load (control breath) was used to evaluate the subjects' spontaneous resting breathing pattern and load compensation response. All subjects were instructed to breathe normally throughout the experiment, and no visual or auditory cue was provided to indicate the loaded inspiration. We found that there were no significant differences in peak Pm, Vi, peak inspiratory airflow, Ti, Te, TP, f, andV˙e during both control breathing and loaded breathing between DLT subjects and Nor subjects. Our results suggest that lung vagal afferents are not essential to the regulation of resting breathing pattern and load compensation responses in humans. Our findings were consistent with other lung transplant studies (23, 33). Shea and co-workers (33) compared resting breathing pattern in heart-lung transplant patients, heart transplant patients, as well as normal subjects. They found no difference in ventilation, tidal volume, f, Ti, and Te among all three groups during wakefulness and sleep. Kimoff et al. (23) also failed to find any major differences in ventilatory level or pattern between heart-lung transplant patients and normal subjects. In contrast, other studies have reported elevated f and reduced Ti in lung transplant recipients (24, 32). However, the lung transplant subjects in those studies had a restrictive spirometric pattern. A relationship between increased lung elastance and increased f has been reported by Renzi and colleagues (29, 31). Furthermore, Sanders et al. (31) observed that the heart-lung transplant patients recipients with a restricted spirometric pattern had a higher f during wakefulness and sleep compared with those recipients without a restrictive pattern. Therefore, the reduced TI and increased f are probably related to the presence of underlying pulmonary restriction rather than to lung-lower airway deafferentation. In the present study, none of the DLT subjects displayed a restrictive spirometric pattern. The difference in our result and those of Kinnear et al. (24) and Sanders et al. (32) is probably due to the difference in the pulmonary function of lung transplant patients recruited. The ventilatory response to added mechanical loads can be regarded as the sum of two components: one representing the effect of the passive respiratory system and one representing the effect of neural load-compensating mechanisms (2). The load-compensating component represents the action of neural mechanisms that modify the pressure developed by loaded respiratory muscles. Receptors in lung and lower airway could potentially contribute to these neural adjustments. However, our results showed no significant group difference. For both the DLT and Nor subjects, as the magnitude of the resistive load increased, Pm, Ti, and TP increased, whereas Vi, airflow, Te, f, and V˙edecreased. These results indicate that load compensation can occur in the absence of vagal afferent input, as long as the remaining afferent pathways are intact. Our results were similar to those of Forster et al. (15), who reported that first-breath load compensation remained after pulmonary vagal denervation in ponies. Load compensation response in lung transplant patients was also studied by Peiffer et al. (27). In contrast to our findings, they reported that the lung transplant recipients produced higher peak Pm and inspiratory flow rate. However, in their protocol, the load was applied after a short vocal cue. Subjects' breathing patterns might change in response to the cue. It is not known whether there is a difference in their reaction to the cue between DLT and the Nor groups, but it is likely that the cue allowed the subject to prepare for the load, thus adding a voluntary component to the load compensation response. Moreover, the inspiratory resistive loads were presented for the duration of two consecutive inspiratory breaths according to their methods. It is not known whether their data came from the first or the second loaded breath. Load compensation responses will be different for a first-breath response compared with a second-breath response. Finally, muscle strength was not compared between their lung transplant recipients and controls. Lung transplant recipients usually have weakened respiratory muscles because of the use of steroid medications and deconditioning after surgery. Most studies demonstrated a close relationship between weak muscle strength and increased respiratory drive (1, 16). The changes in loaded breathing pattern might be a result of changes in respiratory drive and respiratory muscle force in those patients (5, 35). The present study did not show any evidence of muscle weakness in the DLT patients. Therefore, the impact of respiratory muscle strength and drive on load compensation would be minimal in this study. An important assumption of this study is that DLT recipients are, and remain, lung vagally denervated after surgery. The results of several investigations performed in animals found reappearance of a weak Hering-Breuer inflation reflex as early as 5 mo after pulmonary autotransplantation (14, 25). However, reinervation would be less likely in the context of human allotransplantation than with simple reimplantation of an excised lung as in the canine model because no attempt is made to approximate nerves in DLT patients (23). In a study investigating the integrity of the cough reflex, which is mediated mostly by pulmonary receptors, after lung transplant, Higenbottam and co-workers (19) observed a significantly diminished cough response to ultrasonically nebulized distilled water for up to 3 yr after lung transplant. More compelling evidence for persistent lung denervation after human lung transplant has been provided by Iber et al. (20). They recently reported persistently absent expiratory prolongation after passive lung inflation during sleep in bilateral lung transplant recipients for a period of 49 mo after surgery. In contrast to the above findings, Butler et al. (8) recently reported early respiratory events (cough or apnea) and noxious sensations evoked by injections of lobeline (>30 μg/kg) occurred in a few bilateral lung transplantation subjects who were studied more than 1 yr after transplantation. Their results suggested that there might be functional reinnervation of the lungs after bilateral lung transplantation. However, changes in nonpulmonary receptors may have occurred over time to recover the sensitivity to lobeline in those patients. In this study, the time since the patients received DLT surgery varied from 1.5 to 5.5 yr, with an average of 3.45 yr. Although we did not test the reinnervation in our patients, it seems unlikely that reinnervation had occurred on the basis of previous findings (19, 20). Although the detection of inspiratory loads has been studied extensively, the site at which such detection occurs is still not known. There are a variety of mechanoreceptors located in lung and lower airway that are innervated by the vagus nerves. Afferent information from those receptors related to respiratory mechanical changes during loaded breathing may contribute to the detection of external loads. The present study showed that despite a similar loaded breathing pattern, the DLT group had a significantly higher detection threshold (2.91 ± 0.5 cmH2O · l−1 · s;P < 0.05) and Weber fraction (0.50 ± 0.1;P < 0.05) than the Nor group. The group effect on detection percent was significant, with a lower detection percent found in the DLT group. These results suggest that pulmonary vagal afferents may contribute to load detection. Two strategies have been adopted to determine the role of pulmonary receptors in respiratory load perception: either their principle afferent nerve (vagus nerve) is selectively blocked, or alternatively, all other possible sources are eliminated, leaving only the vagal input intact. High-level quadriplegic subjects with tracheostomies provide indirect evidence about the role of pulmonary afferents in respiratory sensation, because both respiratory muscle afferents and upper airway receptors are bypassed, leaving only the pulmonary receptors intact. It has been reported that these patients could reliably detect changes in tidal volume as little as 100 ml, which was comparable to that of normal subjects (4). Similarly, other studies (26,39) also found that detection of external loads did not appear to be impaired in quadriplegic patients in whom afferent pathways from the chest wall are disrupted. These finding suggest the possibility that pulmonary receptors may contribute to load detection. Contradictory results on role of lung vagal afferents in load detection were reported by Guz et al. (17). They studied the effect of bilateral block of the vagus and glossopharyngeal nerves in two healthy subjects. The difference threshold for elastic load detection was not affected by the nerve block. Furthermore, there was also no change in the sensation associated with a high resistive load in one subject. Burki (6) and Chaudhary and Burki (9,10) showed that upper and lower airway anesthesia in normal subjects did not alter the detection thresholds of either resistive or elastic loads. Nonetheless, it is possible that some pulmonary stretch receptors may escape topical anesthesia because the anaesthetic could not penetrate to the smooth muscle or because the drug was carried away rapidly by the rich blood flow (3). Moreover, because both upper and lower airway receptors were blunted in their methods, it is not possible to make a conclusion about the specific role of lung and lower airway afferents in load detection. Lung transplantation, through an interruption of afferent nerve fibers from the lung and lower airways, provides an opportunity to study the contribution of neural feedback from lung and lower airways to respiratory sensation. Besides the present study, there is only one another investigation of resistive load detection in lung transplant recipients by Tapper and his co-workers (34). They compared the detection threshold of inspiratory resistive loads in heart-lung transplant recipients, heart transplant recipients, and normal subjects, and they found no significant difference in the Weber fraction associated with a 50% probability of load detection between the heart-lung transplant recipients and the normal group. Therefore, they concluded that lower respiratory tract afferents did not play a significant role in the perception of respiratory resistive loads. The difference between the present study and the study of Tapper et al. might be due to the difference in lung transplant patients and the method applied to determine detection threshold. The DLT subjects in the present study are older than the heart-lung transplant patients in the study of Tapper et al. (46.5 ± 4.4 vs. 33.7 ± 1.5 yr, respectively). The effect of age on resistive load detection ability is unclear. Tapper et al. used a tracking procedure to determine the detection threshold. The tracking procedure causes more false-positive responses. The relationship between false-positive response rate and detection threshold is not known. Moreover, the imposition of resistance or shams was signaled by an audible cue during the preceding exhalation. It is reasonable to believe that a cue would improve a person's detection performance, which might result in the lower Weber fraction found in their heart-lung transplant patients. In the present study, the DLT patients' FEV1 and FVC values were significantly lower than those of the Nor subjects (83.8 ± 6.1 vs. 108.3 ± 3.7% of predictive values, and 97.0 ± 4.9 vs. 119.5 ± 5.0% of predictive values, respectively). However, both FEV1 and FVC are well within normal limits. It is unlikely that an increased detection threshold found in the DLT group was due to their lung function. The Weber fraction, which controlled for the effect of background resistance, was significantly higher in the DLT group than the Nor group. Furthermore, no significant correlation has been found between either FVC or FEV1 and Weber fraction in both DLT and Nor subjects (P > 0.05). Two DLT subjects had higher detection threshold and Weber fraction values compared with the remainder of the DLT group (Figs. 2 and 3). Their pulmonary function and demographics were similar to those of other DLT subjects. The two subjects were 3.5 and 2.0 yr posttransplantation. There was no significant correlation between time postsurgery and detection threshold for the DLT group. The cause of their high detection threshold was not clear either. It should also be noted that in the present study, all DLT subjects were on immunosuppressive agents and steroid medications. It is possible that the present results could be affected by the medications taken by the DLT recipients. Further studies are necessary to investigate the effect of these medications on resistive load detection. CONCLUSIONSIn summary, we found that both unloaded and loaded breathing patterns were similar in the DLT group and the Nor group, suggesting that lung vagal afferents are not essential to the regulation of resting breathing pattern and load compensation response. Furthermore, resistive load detection did occur in DLT patients. This means that nonvagal afferents are activated during breathing against resistive loads and do elicit a load detection response. However, the DLT recipients had a significantly higher detection threshold and Weber fraction than the Nor group. The impaired detection capability is likely due to the loss of lung vagal afferent inputs in those lung-denervated patients. The results of this study suggest that vagal afferents play a role in resistive load detection. The detection threshold is increased with the loss of vagal afferents. However, the effect of DLT medications on load detection cannot be ruled out. This study was supported by National Heart, Lung, and Blood Institute Grant HL-47892. FOOTNOTESREFERENCES
Page 21glycine is a major inhibitory neurotransmitter in the central nervous system (e.g., for review, see Ref. 30). The role of glycine in the motor control of upper airway dilator muscles, such as the genioglossus (GG) muscle of the tongue, remains incompletely understood. There is immunohistological (31, 43) and autoradiographic (29, 47) evidence for the presence of both glycine and its receptor in the hypoglossal motor nucleus, the source of motor outflow to the GG. From in vitro studies of neonatal rat brain stem, there is also evidence for glycine-mediated inhibition of hypoglossal motoneurons with effects antagonized by strychnine (5, 6, 24,25). There is also evidence in vivo for strychnine-sensitive inhibitory postsynaptic potentials recorded from hypoglossal motoneurons after electrical stimulation of the lingual nerve in cats (40). In addition, application of glycine by iontophoresis to the hypoglossal motor nucleus suppresses the antidromic field potential elicited in the hypoglossal motor nucleus by electrical stimulation of the whole hypoglossal nerve, although it has not been established whether these phenomena were recorded from retractor or protruder motoneurons (41). Accordingly, few studies have determined the effects of glycine at the hypoglossal motor nucleus on motor outflow to the GG muscle in vivo, and, to our knowledge, no studies have determined the interaction of such glycine mechanisms with integrative reflex respiratory control, such as the GG muscle responses to systemic hypercapnia. Glycine-immunoreactive fibers in the hypoglossal motor nucleus are also intermingled with fibers containing GABA, another major inhibitory neurotransmitter (18). GABA inhibits hypoglossal motoneurons via the GABAA receptor in vitro (6,24) and decreases motor outflow to GG muscle in vivo (19,20, 41). Because glycine and GABA are likely to be released together within the hypoglossal motor nucleus and individual hypoglossal motoneurons contain receptors for both neurotransmitters (24), it is important to determine whether coapplication of glycine and a GABAA-receptor agonist into the hypoglossal motor nucleus is additive in suppressing GG muscle activity. Accordingly, the present study tests the hypothesis that increasing glycine at the hypoglossal motor nucleus will suppress GG muscle activity, even in the presence of reflex respiratory stimulation by systemic hypercapnia, and that the suppressant effects of glycine will be antagonized by strychnine. We also hypothesize that a combination of glycine and the GABAA-receptor agonist muscimol at the hypoglossal motor nucleus will produce additive suppression of GG activity. Understanding the potential for inhibitory neurotransmitters such as glycine and GABA to suppress GG activity and modulate integrative reflex respiratory responses, such as those produced by CO2 stimulation, is important because major suppression of GG activity occurs in certain behaviors, such as rapid eye movement (REM) sleep, both in animals (11, 22, 27) and humans, especially during phasic REM sleep events (36, 45). REM sleep also effectively abolishes GG responses to CO2(11). However, although inhibitory postsynaptic potentials have been recorded at hypoglossal motoneurons during the pharmacological model of REM sleep induced by pontine carbachol (9, 46), the potential role of glycine and GABA in the suppression of GG activity remains open to question (17). However, before determination of whether such inhibitory mechanisms are recruited or not in REM sleep, it is important to determine the presence (or absence) of such mechanisms and to characterize their interaction with integrative respiratory control mechanisms. This study has direct relevance to the neural control of GG activity as well as clinical disorders, where suppression of GG muscle activity can lead to airway occlusion and obstructive sleep apnea (33). METHODSTwenty-four male Wistar rats (Charles River, mean body weight, 293 ± 7 g: range, 242–388 g) were studied. All procedures conformed to the recommendations of the Canadian Council on Animal Care, and the University of Toronto Animal Care Committee approved the experimental protocols. The rats were anesthetized with urethane (0.5 g/ml, 1 g/kg ip), and surgical levels of anesthesia were then maintained, as necessary, by inhalation of halothane (typically 0.2–2%). Halothane was administered via an anesthetic mask and then through a tracheal tube after tracheotomy (see below). The rats spontaneously breathed a mixture of 50% room air and 50% oxygen throughout all procedures. The rats were also given atropine sulfate (0.5 g/ml, 1 mg/kg ip) to minimize airway secretions. Body temperature was monitored with a rectal probe and maintained between 36 and 38°C with a water pump and heating pad (T/Pump-Heat Therapy System, Gaymar, NY). After the onset of effective surgical anesthesia, as judged by abolition of hindlimb withdrawal and corneal blink reflexes, the rats were tracheotomized and bilaterally vagotomized. Vagotomy was performed to prevent reflex inhibition of GG muscle originating in vagal afferents (1, 37) and to eliminate the confounding effect of vagal reflex responses to the changes in breathing rate and depth elicited by hypercapnic stimuli (see results). A polyethylene catheter was inserted in the femoral vein for continuous infusion of a solution containing 7.5 ml of 0.9% saline, 2.0 ml of 5% dextrose, and 0.4 ml of 1 M NaHCO3 at a flow rate of 0.4 ml/h. A bolus dose of dexamethasone (0.1 ml at 2 mg/ml) was also given to minimize brain edema. A catheter inserted into the femoral artery was used to record arterial blood pressure. To record the diaphragm electromyogram (EMG), multistranded stainless steel wires (AS636, Cooner Wire, Chatsworth, CA) were sutured into the costal diaphragm with an abdominal approach. The rats were then placed in a stereotaxic apparatus (model 962, Kopf, Tujunga, CA) in the flat skull position by using an alignment tool (Kopf model 944). Stainless steel wire electrodes were inserted bilaterally, under direct vision, into the GG muscle. In two additional rats, tongue EMG recordings were made before and after bilateral section of the medial and then lateral branches of the hypoglossal nerve. Two stainless steel screws (1.5 mm diameter) attached to insulated wire (30 gauge) were implanted in the skull over the frontal-parietal cortex to record the electroencephalogram (EEG). The EEG electrodes were placed ∼2 mm anterior and 2 mm to the right of bregma, and 3 mm posterior and 2 mm to the left of bregma (12). All electrical signals were amplified and filtered (Super-Z head-stage amplifiers and BMA-400 amplifiers/filters, CWE, Ardmore, PA). The EEG was amplified by 1,000 and filtered between 1 and 100 Hz, whereas the GG and diaphragm EMGs were amplified by 2,000 and filtered between 100 and 1,000 Hz. The electrocardiogram was removed from the diaphragm EMG by using an oscilloscope and an electronic blanker (model SB-1, CWE). The moving-time averages (time constant = 200 ms) of the GG and diaphragm EMGs were also obtained (S76-01, Coulbourn, Lehigh Valley, PA). The EEG and EMG signals were calibrated by using the built-in microvolt calibrator (20 μV to 1 mV) on the head-stage amplifiers. Blood pressure was measured with a transducer (DT-XX, Ohmeda, Madison, WI) and an appropriate amplifier (PM-1000, CWE). Inspired CO2 concentration was measured with a CO2 analyzer (CAPStar-100, CWE). All raw signals, along with the moving-time averages of the GG and diaphragm EMGs, were recorded on chart paper (TA11, Gould, Valley View, OH) and computer (1401 interface, Spike 2 software, CED, Cambridge, UK). The microdialysis probes were 240 μm in diameter with a 1-mm cuprophane membrane and a 6,000-Da cutoff (CMA/11 14/01, CSC, St. Laurent, Quebec). The probes were lowered slowly through a small hole drilled at the junction of the intraparietal and occipital bones and aimed at the hypoglossal motor nucleus by using the following coordinates: 13.86 ± 0.16 (SE) mm posterior to bregma (range, 12.90–16.85 mm), 0.31 ± 0.02 mm lateral to the midline (range, 0.15–0.55 mm), and 9.46 ± 0.15 mm ventral to bregma (range, 7.00–10.50 mm). When the microdialysis probe initially penetrated the hypoglossal motor nucleus, a transient burst of GG EMG activity was observed, and the probe was then lowered a further 0.5 mm to these final coordinates. This burst of GG activity during insertion of the probe was transient (lasting an average of 3 min and 40 ± 35 s in the group of rats) and was useful as a preliminary indication of probe placement at the beginning of the experiment (e.g., Ref. 14). The insertion of the probe did not affect the diaphragm EMG, respiratory rate, or blood pressure. The microdialysis probes were connected to FEP Teflon tubing (inside diameter = 0.12 mm) that in turn was connected to 1.0-ml plastic syringes via a zero-dead space switch (Uniswitch, BAS, West Lafayette, IN). The lag time for fluid to travel to the tip of the probe from the switch was 3 min and 38 s. The probes were perfused with artificial cerebrospinal fluid (aCSF) at a flow rate of 2.1 μl/min. The aCSF was made fresh on the day of each experiment with a composition (in mM) of 125 NaCl, 3 KCl, 1 KH2PO4, 2 CaCl2, 1 MgSO4, 25 NaHCO3, and 30 d-glucose. The aCSF was warmed to 37°C before CaCl2 was added (23). The aCSF was then bubbled with 100% CO2to a physiological pH of 7.35–7.45. The rats stabilized for at least 30 min after insertion of probes before any interventions. All interventions were typically performed between 1300 to 1900 during steady-state periods with stable breathing, blood pressure, and high-voltage and low-frequency EEG activity. In seven rats, all signals were recorded during microdialysis perfusion of aCSF (i.e., control) and glycine (glycine hydrochloride, FW: 111.5, Sigma Chemical, St. Louis, MO) dissolved in aCSF at doses of 0.0001, 0.001, 0.01, 0.1, 1.0, and 10 mM. Doses were applied in ascending order for a total of ∼40 min at each dose. After 20 min after an increment in dose, signals were recorded before, during, and after steady-state (6 min) application of 7% inspired CO2. For aCSF and each level of glycine delivered to the hypoglossal motor nucleus, analyses were performed over 1-min periods immediately before CO2application, in the 6th min of CO2 application, and 6 min after removal of CO2. To determine whether prior exposure of the hypoglossal motor nucleus to glycine affected subsequent responses to the higher doses, a further study was performed in four additional rats in which only the three highest doses of glycine (i.e., 0.1, 1.0, and 10 mM) were applied by using the same protocol described above. A second study in six rats was performed to determine whether strychnine, a glycine-receptor antagonist, could reverse the effects of glycine at the hypoglossal motor nucleus. The protocol involved perfusion of the hypoglossal motor nucleus for 25 min with each of the following agents in the following order: aCSF, glycine (1 mM), strychnine (0.1 mM; strychnine hydrochloride, FW: 370.9, Sigma Chemical), a return to glycine, and, finally, a switch back to aCSF. Data were analyzed 20 min after a switch between drugs. The focus of the study was the change in GG activity from the first glycine-to-strychnine condition compared with the second glycine-to-aCSF condition. A third study was performed to determine the effects on GG muscle activity of microdialysis perfusion of glycine and the GABAA-receptor agonist muscimol (muscimol hydrobromide, FW: 195.0, Sigma Chemical) into the hypoglossal motor nucleus, first alone and then together. For each experiment, aCSF was first perfused into the hypoglossal motor nucleus followed by either glycine (four rats, 0.01 mM) or muscimol (three rats, 0.1 μM). After this initial response to the first agent applied alone, a switch was made to perfuse both agents together. As such, a total of seven rats ended the protocol with glycine and muscimol applied together. Each drug was perfused for 35 min. For aCSF, or when glycine and muscimol were applied together, data were also recorded before, during application of steady-state (>6 min) CO2, followed by a 10-min recovery period. Preliminary studies showed that the doses of glycine (see results) and muscimol (19, 20) used were sufficient to clearly decrease GG muscle activity when applied individually to the hypoglossal motor nucleus. Data were analyzed 30 min after a switch between drugs. Breath-by-breath measurements of GG and diaphragm activities were calculated and averaged in consecutive 5-s epochs. The EMGs were analyzed from the moving-average signal (above electrical zero) and were quantified in arbitrary units. Electrical zero was the voltage recorded with the amplifier inputs grounded. GG activity was quantified as mean tonic activity (i.e., difference between basal activity at end expiration and electrical zero), peak inspiratory activity, and phasic respiratory-related activity (the difference between peak inspiratory and tonic end-expiratory activity). In practice, there was no tonic GG activity in this anesthetized preparation; therefore, only phasic respiratory activity is presented. The amplitude of the diaphragm EMG was quantified as the difference between the peak inspiratory and prior end-expiratory values' moving-time-averaged signal, and the mean values of diaphragm amplitude, respiratory rate, and blood pressure were also calculated for each 5-s epoch. Diaphragm minute activity was calculated as the product of diaphragm amplitude and respiratory rate. The EEG was sampled by computer at 500 Hz and analyzed on overlapping segments of 1,024 samples, windowed using a raised cosine (Hamming) function, and subjected to a fast-Fourier transform to yield the power spectrum. The window was advanced in steps of 512 samples, and the mean power spectrum of the EEG signal over each 5-s analysis epoch was calculated. The power contained within six frequency bands was recorded both as absolute power and as the percentage of the total power of the EEG signal. The band limits were δ2 (0.5–2 Hz), δ1 (2–4 Hz), θ (4–7.5 Hz), α (7.5–13.5 Hz), β1 (13.5–20 Hz), and β2(20–30 Hz). All respiratory, blood pressure, and EEG values were written to a spreadsheet and matched to the corresponding intervention at the hypoglossal motor nucleus, and the respective level of CO2, to provide a grand mean for each variable, for each intervention, in each rat. On completion of the experiments, the rats were overdosed with urethane and perfused intracardially with 40 ml of 0.9% saline followed by 20 ml of 10% formalin. Brains were then removed and fixed in 10% formalin. The medullary regions were blocked, transferred to 30% sucrose, and cut in 50-μm coronal sections with a cryostat (Leica, CM 1850, Nussloch, Germany). Each section containing the hypoglossal motor nucleus was mounted and stained with neutral red. Microdialysis sites were localized from the neutral red-stained sections and marked on standard brain maps (28). For all comparisons, differences were considered significant if the null hypothesis was rejected at P < 0.05 by using a two-tailed test. Data were analyzed by using either repeated-measures ANOVA (RM-ANOVA) or paired t-tests, as indicated in the text. Except where noted, the factors were CO2 level and dose of drug for two-way RM-ANOVA and dose of drug for one-way RM-ANOVA. For post hoc t-tests, Dunnett's test for comparisons with a single control (i.e., aCSF) was used. Analyses were performed by using Sigmastat (SPSS, Chicago, IL). All data are expressed as means ± SE. RESULTSFigure 1 shows an example of the tongue EMG recordings before and after bilateral section of the medial and then lateral branches of the hypoglossal nerve. Note that the tongue EMG activity was markedly decreased, but not completely abolished, after section of the medial branches of the hypoglossal nerve, showing that recordings were predominantly from the GG muscle. Tongue muscle activity was fully abolished after additional section of the lateral branches of the hypoglossal nerve, indicating that the retractor muscles also contributed a small component to the whole signal. These effects of hypoglossal nerve section were selective for reducing tongue muscle activity, as diaphragm activity was unaffected by the interventions. Fig. 1.Tongue electromyographic (EMG) recordings made before and after section of the right and left medial branches of the hypoglossal (XII) nerve and then with additional section of the lateral branches. The genioglossus (GG) and diaphragm (Dia) activities are also displayed as their moving-time averages (MTA) in arbitrary (Arb) units. Note that the tongue EMG activity was markedly decreased, but not completely abolished, after section of the medial branches of the XII nerve and then fully abolished after additional section of the lateral branches. Dia activity was unchanged. Figure 2A shows an example of the lesion site made by the microdialysis probe in the hypoglossal motor nucleus. The locations of all lesion sites from all rats were within the hypoglossal motor nucleus, as shown in Fig.2B. Fig. 2.A: middle histological section shows an example of a lesion site left by the microdialysis probe in the XII motor nucleus. Intact cells in the XII motor nucleus can be seen on this section as well as those rostral (left) and caudal (right). B: distribution of individual microdialysis sites from all rats are shown. The size of the bar represents the apparent size of the lesion from the histological sections. Cer, cerebellum; 4V, fourth ventricle; Sol, nucleus tractus solitarius; 12, XII motor nucleus; Gi, gigantocellular reticular nucleus; ROb, raphe obscurus, Py, pyramidal tract; AP, area postrema; MdV, medullary reticular nucleus, ventral part. Figure 3A shows an example of the responses of GG muscle to perfusion of glycine into the hypoglossal motor nucleus. This trace shows that increasing glycine caused progressive reductions in GG muscle activity. In this example, there were no changes in diaphragm or EEG activity, but blood pressure decreased at the higher doses of glycine. The group data in Fig.3B also show that increasing glycine at the hypoglossal motor nucleus produces graded suppression of GG muscle activity. Analysis confirmed that there was a significant effect of glycine on GG muscle activity (F6,36 = 33.40,P < 0.0001, one-way RM-ANOVA) with a mean decrease of ∼80% at 10 mM glycine. Compared with aCSF, a significant decrease in GG activity with glycine occurred at 0.001 mM (P < 0.05, Dunnett's test). Fig. 3.A: decrease in GG muscle activity with increasing glycine (Gly) at the XII motor nucleus. Blood pressure, the electroencephalogram (EEG), and the EMG activity of the GG muscle are shown. The GG and Dia activities are also displayed as their MTA in Arb units. The baseline of the integrator (i.e., electrical zero) is shown for the GG MTA. The arrows on the Dia and GG signals denote an increase in activity associated with inspiration. B: group data showing GG responses to Gly at the XII motor nucleus expressed as a percentage of that recorded with artificial cerebrospinal fluid (aCSF). Values are means ± SE. * Significant differences compared with the aCSF controls, P < 0.05. See text for further details. Figure 4A shows the effects of glycine at the hypoglossal motor nucleus on the GG muscle responses to CO2. Analyses showed that, despite the progressive decline in GG activity with glycine at the hypoglossal motor nucleus (F6,36 = 13.84; P < 0.001, two-way RM-ANOVA), there remained a significant stimulating effect of CO2 (F2,12 = 11.58,P = 0.002) that occurred independently of glycine dose (F12,72 = 1.44, P = 0.170). Although this result showed that the hypoglossal motor nucleus could still respond to excitatory inputs, despite the presence of glycine, further analyses were performed to determine whether glycine altered the percent increase in GG activity with CO2 stimulation. Analyses of the data shown in Fig. 4B confirmed that there was no effect of glycine on the percent increase in GG activity in response to CO2 (F6,36 = 1.93,P = 0.102, one-way RM-ANOVA). For these data, it should be noted that one rat had relatively low-baseline GG activity but large responses to CO2, which explains why the calculated percent increase in GG activity for the group appears larger than would be anticipated from the group mean values for the raw data (Fig.4A vs. 3B). Fig. 4.A: group data (n = 7) showing phasic respiratory GG activity during control microdialysis of aCSF into the XII motor nucleus and at each level of Gly. Data are shown before (open bars), during (solid bar), and after (shaded bar) steady-state stimulation with 7% inspired CO2.B: percent increase in GG activity with CO2stimulation was similar across all doses of Gly. Values are means ± SE. See text for further details. Figure 5 shows that there was no effect of glycine at the hypoglossal motor nucleus on phasic diaphragm activity (Fig. 5A; F6,36 = 0.914, P = 0.496, two-way RM-ANOVA). However, there was a significant effect of glycine on respiratory rate (Fig.5B; F6,36 = 5.52,P < 0.001), although this only became statistically significant compared with aCSF at 1 mM glycine (P < 0.05, Dunnett's test), i.e., a glycine level that was much higher than that which caused suppression of GG activity (0.001 mM; Fig.3B). Despite the slight decline in respiratory rate at the higher glycine doses, overall diaphragm minute activity was unaffected by glycine at the hypoglossal motor nucleus (Fig. 5C;F6,36 = 1.13, P = 0.365). Similarly, although there was a significant effect of glycine on blood pressure (F6,36 = 6.48, P< 0.001), this became significant compared with aCSF at 0.1 mM glycine (P = 0.05, Dunnett's test), i.e., again at a dose higher than the effects on GG activity. There was no effect of glycine at the hypoglossal motor nucleus on the distribution of frequencies in the EEG signal (all F6,36 < 1.92, allP > 0.100), although total EEG power declined with glycine (F6,36 = 7.45, P < 0.0001) with significant changes at 0.001 mM (P < 0.05, Dunnett's test). Fig. 5.Group data (n = 7) showing Dia amplitude (A), respiratory rate (B), Dia minute activity (C), and blood pressure (D) during control microdialysis of aCSF into the XII motor nucleus and at each level of Gly. Data are shown before (open bars), during (solid bars), and after (shaded bars) steady-state stimulation with 7% inspired CO2. Values are means ± SE. See text for further details. As expected, there was a significant effect of CO2 on diaphragm activity, with breathing becoming deeper (F2,12 = 14.13, P < 0.001 from two-way RM-ANOVA) and slower (F2,12 = 5.55, P = 0.02) with CO2 stimulation. However, unlike the significant effects of CO2 on GG activity (Fig. 4), the effect of CO2 on overall diaphragm minute activity was not statistically significant (F2,12 = 3.04, P = 0.085). Total EEG power was not affected by CO2 stimulation (F2,12 = 1.04, P = 0.385), although there were slight decreases in blood pressure with CO2 (Fig. 5D; F2,12= 5.26, P = 0.02). The decreases in GG muscle activity observed with 0.1, 1.0, and 10 mM glycine at the hypoglossal motor nucleus in study 1 (i.e., after exposure to the preceding lower doses) were similar to the decreases in GG activity measured when 0.1, 1.0, and 10 mM glycine were applied directly without such previous exposure (F1,2 = 8.17, P = 0.104, two-way RM-ANOVA). The application of these three highest doses of glycine alone to the hypoglossal motor nucleus was not associated with any changes in phasic diaphragm activity, respiratory rate, diaphragm minute activity, blood pressure, or total EEG power (all F3,9 < 0.90, all P > 0.480, one-way RM-ANOVA). Figure 6A shows an example of the typical suppression of GG muscle activity after a switch from aCSF to glycine at the hypoglossal motor nucleus and then reversal of this suppression with application of strychnine. In contrast, a switch back to aCSF from glycine does not reverse the suppression when measured over the same time frame, showing that the responses with strychnine were not a time effect due to washout of glycine. Data for the group of six rats are shown in Fig. 6B. Analysis showed that there was a significant effect of drug treatment on GG activity (F4,16 = 5.55, P = 0.005, one-way RM-ANOVA). Further post hoc analyses showed that a switch from aCSF to glycine caused a significant decrease in GG activity [t(5) = 5.50, P = 0.003, paired t-test], whereas application of strychnine returned GG activity to levels indistinguishable from that of the aCSF controls [t(5) = 1.25, P = 0.268]. Reapplication of glycine returned GG activity to values significantly below aCSF [t(5) = 3.63,P = 0.015], and activity remained significantly below this control after a subsequent switch back to aCSF [t(5) = 5.36, P = 0.003], confirming that the strychnine effect was due to drug rather than washout of glycine. Fig. 6.Raw traces (A) and group data (B) from 6 rats showing reversal of Gly-mediated suppression of GG activity with strychnine at the XII motor nucleus. A switch to aCSF after Gly produces no such return to control GG activity when measured over the same time scale. A: data for strychnine and the second exposure to aCSF were recorded 20 min after switching from aCSF microdialysis. B: values are means ± SE. * Significant differences compared with the initial aCSF control,P < 0.05. Figure 7A shows examples of the combined administration of glycine and muscimol to the hypoglossal motor nucleus in two of the seven rats. In the first rat (Fig. 7A, left), GG muscle activity is shown for aCSF, glycine, and then combined glycine and muscimol at the hypoglossal motor nucleus, whereas responses to aCSF, muscimol, and then combined muscimol and glycine are shown for the second rat (Fig.7A, right). Group data from all rats (Fig.7B) showed that coapplication of muscimol and glycine to the hypoglossal motor nucleus caused a significant decrease in GG activity [mean decrease = 59.2 ± 7.7% from aCSF, 95% confidence interval for the change = 40.3–78.0%,t(6) = 7.69, P < 0.001, paired t-test]. Individual application of muscimol or glycine caused suppression of GG activity by 42.5 ± 8.3 and 28.4 ± 15.1%, respectively, from aCSF, such that the algebraic sum of their individual effects (i.e., 70.9%) was within the 95% confidence interval for the actual effect observed when they were applied together. Fig. 7.Raw traces from 2 rats (A) and group data from 7 rats (B) showing major suppression of GG muscle activity with combined administration of Gly and muscimol (Mus) into the XII motor nucleus compared with when either agent was administered alone.B: values are means ± SE. In the presence of combined administration of glycine and muscimol to the hypoglossal motor nucleus, GG activity still increased with CO2 stimulation (Fig. 8). Analyses showed that the increase in GG activity with CO2was statistically significant (F2,12 = 3.94, P = 0.048, two-way RM-ANOVA), with this stimulating effect occurring independently of whether aCSF or combined glycine and muscimol were present at the hypoglossal motor nucleus (F2,12 = 0.62, P = 0.556). Furthermore, the percent increase in GG activity with CO2was similar whether aCSF or combined glycine and muscimol was applied to the hypoglossal motor nucleus [24.6 ± 6.8 vs. 24.8 ± 8.6%, respectively, t(6) = 0.03,P = 0.976, paired t-test], although there was a clear suppression of overall GG activity with glycine and muscimol (Fig. 8; F1,6 = 17.19,P = 0.006, two-way RM-ANOVA). Fig. 8.Group data from 7 rats showing phasic respiratory GG activity during control microdialysis of aCSF into the XII motor nucleus and combined perfusion of Gly and Mus. Data are shown before (open bars), during (solid bars), and after (shaded bars) steady-state stimulation with 7% inspired CO2. See text for further details. DISCUSSIONThis study shows that increasing glycine at the hypoglossal motor nucleus in vivo produces graded suppression of GG muscle activity, with application of strychnine reversing this suppression. The glycine-mediated suppression of GG activity occurred both in the absence of, and in the presence of, reflex respiratory stimulation with systemic hypercapnia. Nonetheless, at each level of glycine, the GG muscle was still able to respond to CO2 stimuli, and the proportional increase in GG activity was not affected by glycine at the hypoglossal motor nucleus. Overall, these results confirm previous in vitro studies (5, 6, 24) and extend them to the in vivo preparation where the interaction with integrative reflex respiratory control could also be determined. In the present study, we also showed that the combined effect of glycine and muscimol, a GABAA-receptor agonist, at the hypoglossal motor nucleus was additive in the suppression of GG activity. Moreover, with combined administration of glycine and muscimol, the GG muscle was also able to respond to CO2 stimuli, although overall activity was decreased compared with aCSF. These results are relevant to the neural control of hypoglossal motor outflow, because recent data show that glycine and GABA are released together onto hypoglossal motoneurons, with individual motoneurons also containing receptors for both transmitters (24). Such corelease of GABA and glycine also occurs onto neurons in the spinal cord (2, 15, 21). Overall, the results of this study have direct relevance to the neural mechanisms and respiratory control of hypoglossal motor outflow to GG muscle. These results may also be relevant to the periods of major suppression of GG muscle activity that occur in REM sleep (22,27, 36, 45), even in the presence of strong respiratory stimulation by CO2 (11). The extrapolation of these results to a behaving preparation is, however, necessarily limited because studies were performed in anesthetized rather than sleeping rats. The experiments were performed in anesthetized rats to allow for delivery of the varying levels of glycine to the hypoglossal motor nucleus in a controlled and systematic fashion, with and without steady-state CO2 stimuli. Microdialysis was used to allow repeated switches between different drugs and doses without the need for removing and replacing a microinjection cannula. Although it is unlikely that the use of anesthesia would have altered the direction of change in GG activity to the applied glycine, it is possible that anesthesia may have altered the magnitude of responses (13, 26). For example, halothane has been shown to potentiate the effects of both GABAA- and glycine-receptor stimulation (42), although it does not alter extracellular glycine concentration as measured in the cerebral cortex (34). Nevertheless, this initial characterization of the GG muscle responses to glycine, and its modulation of reflex respiratory responses to CO2 and the effects of combined application of muscimol, was performed in this reduced and controlled preparation because of the added complexity in a behaving rat of varying sleep-wake states with their own different effects on GG muscle activities (14) and responses to CO2 (11). Extrapolation to the intact behaving preparation may also be somewhat limited because the rats were vagotomized in these studies. Vagotomy was performed to prevent reflex inhibition of GG muscle originating in vagal afferents, particularly strong in the rat (1, 37), and to eliminate the confounding effect of vagal reflex responses to the changes in breathing rate and depth elicited by hypercapnic stimuli. Nevertheless, because the conditions of the experiments favored the presence of strong inspiratory GG activation and the absence of tonic activity, the balance of tonic and phasic components within the respiratory control network may have been altered to some degree by vagotomy because more tonic activity and less phasic inspiratory GG activity are typically present in intact rats (11). It remains to be tested whether application of glycine and muscimol to the hypoglossal motor nucleus produces similar suppression of hypoglossal motor output to the GG muscle in intact rats. We believe, however, that it is important to initially characterize the potential effects of glycine at the hypoglossal motor nucleus on GG muscle activity in a reduced and controlled preparation because controversy exists as to the potential role (if any) of such inhibitory neurotransmitters in the suppression of hypoglossal motor outflow. For example, although glycine has been implicated in the suppression of lumbar (39), trigeminal (38), and masseter (16) motoneuron activity, both in natural REM sleep or in the REM-like state produced by pontine carbachol, such mechanisms have not been demonstrated at the hypoglossal motor nucleus (17). Although, pontine carbachol does not produce the whole range of electrocortical and respiratory events characteristic of natural REM sleep (10), transient inhibitory postsynaptic potentials have been recorded at hypoglossal motoneurons in such a preparation, but the neurotransmitters responsible are currently unidentified (9, 46) and may be mediated by transmitters other than glycine or GABA. Nevertheless, before adequate interpretation of whether or not postsynaptic inhibitory mechanisms play a role (or not) in the suppression of motor outflow to GG muscle, either in behaviors such as REM sleep or otherwise, it is first necessary to characterize the presence of such inhibitory mechanisms at the hypoglossal motor nucleus and the effects of the applied neurotransmitters. Accordingly, this study provides such information in vivo for glycine and its interactions with GABAA-receptor-mediated effects at the hypoglossal motor nucleus and the ability of these mechanisms to suppress GG muscle activity in normocapnia and during reflex hypercapnic stimulation. After this initial characterization, future studies using in vivo microdialysis of the hypoglossal motor nucleus in freely behaving rats (14) can be used to determine the potential role of such inhibitory mechanisms in the major suppression of GG activity and responses to CO2 in natural REM sleep (11). Such mechanisms may indeed be operative, because there is an interesting case report of increased pharyngeal muscle activity and improvements of obstructive sleep apnea after blockade of putative glycinergic inhibition of pharyngeal motoneurons with systemically applied strychnine (32). However, the effects on GG muscle activity in REM sleep were not specifically distinguished in that study, and responses were observed predominantly in the tensor palatini muscle innervated by trigeminal motoneurons (32). The tongue electrodes were placed under direct vision into GG muscle, but there is the possibility that other muscles were being recorded as well as GG. For example, tongue retractors are coactivated with GG during respiratory stimulation (7, 8). In addition, it is possible that retractor motoneurons such as those innervating the styloglossus and hyoglossus muscles would be influenced by the interventions at the hypoglossal motor nucleus, although this may be to a lesser degree than GG motoneurons because the probe sites were located in rostral regions of the hypoglossal motor nucleus (Fig.2 and Ref. 28). Protruder motoneurons, such as those innervating GG, are predominantly located in rostral regions of the hypoglossal motor nucleus, whereas retractor motoneurons are clustered in more caudal regions (4). It is also a concern that neural structures close to the hypoglossal motor nucleus may have been influenced by diffusion of drug from the microdialysis probe and that this would complicate data interpretation. Significant suppression of GG muscle activity with glycine at the hypoglossal motor nucleus occurred at a threshold dose of 0.001 mM (Fig. 3). Although glycine also caused a significant decrease in respiratory rate, possibly indicating diffusion to nearby respiratory neurons (e.g., see Ref. 3), the glycine dose that caused this change (1 mM) was much higher than that which produced a change in GG activity. Moreover, there was no effect on phasic diaphragm activity or overall diaphragm minute activity with any dose of glycine. Similarly, although there was a statistically significant decrease in blood pressure with glycine at the hypoglossal motor nucleus, this decrease of 5–10 mmHg was observed at a dose that was also higher (0.1 mM) than that which produced a change in GG activity. Moreover, a decrease in blood pressure would be expected to cause a reflex increase in GG activity (35, 44), i.e., the opposite of the GG suppression that was observed with glycine (Fig. 3). Although there was no effect of glycine at the hypoglossal motor nucleus on the distribution of frequencies in the EEG signal, total EEG power declined with glycine at a dose similar to that which produced a change in GG activity. However, a decrease in total EEG power is indicative of an arousal response, an effect that would be expected to produce an increase in GG activity (13, 26) rather than the decrease that was observed with glycine. Overall, these observations suggest that the effects of glycine on GG muscle activity were likely primary responses to modulation of hypoglossal motor outflow via local inhibitory mechanisms similar to those observed in vitro (5, 6, 24). However, the changes in other physiological variables (e.g., blood pressure and respiratory rate), albeit at higher doses, suggest the likelihood of spread of the perfusate because of prolonged exposure. This suggestion is supported by the additional results with shorter exposures that showed that when the three higher doses of glycine were applied to the hypoglossal motor nucleus without the preceding lower doses, the changes in GG activity occurred without any associated changes in phasic diaphragm activity, respiratory rate, diaphragm minute activity, blood pressure, or total EEG power. This study shows that glycine receptor stimulation at the hypoglossal motor nucleus in vivo suppresses GG muscle tone and activity during hypercapnia. Moreover, the combined effects of glycine and the GABAA-receptor agonist muscimol at the hypoglossal motor nucleus was additive in the suppression of GG muscle activity. It remains to be determined whether these inhibitory neural mechanisms are recruited in natural REM sleep to explain the major decrease in GG muscle activity and reflex GG responses to systemic hypercapnia observed in that sleep state (11). This work was supported by funds from Canadian Institutes of Health Research (CIHR) Grant MT-15563. J. L. Morrison is a recipient of Post-Doctoral Fellowships from the Department of Medicine at the University of Toronto and Merck Frosst. R. L. Horner is a recipient of a CIHR Scholarship. FOOTNOTESREFERENCES
Page 22dietary influence on muscle metabolism during exercise has been studied since the 1930s when Christensen and Hansen (7) observed that the preexercise diet was of major importance for substrate metabolism as well as endurance during exercise. We have previously shown (16,17) that adaptation to a fat-rich diet during physical training for 7 wk increases utilization of lipids (16, 17) and decreased rate of muscle glycogen breakdown during submaximal exercise (17) compared with a group of similar subjects training on a carbohydrate-rich diet (16, 17). Interestingly, a prior study demonstrated (16) that, despite very high glycogen stores, induced by consuming a carbohydrate-rich diet after prolonged fat diet adaptation, the subjects were unable to utilize the glycogen stores efficiently, and exhaustion during exercise occurred despite very high residual muscle glycogen stores of ∼500 μmol/g dry wt. Thus prolonged fat adaptation seems to induce an inability to efficiently use muscle glycogen during exercise, and this adaptation is not overcome by 1 wk of carbohydrate diet. However, we did observe that toward the end of exercise the venous plasma glucose concentration in contrast to what might have been expected was higher in the group switching from fat to carbohydrate for the last week of training than in the group ingesting a carbohydrate-rich diet all along and from this we speculated that glucose utilization was attenuated. One goal of the present study therefore was to investigate whether glucose uptake may be attenuated under such circumstances. In a recent study, Burke and colleagues (5) demonstrated that 5 days of fat adaptation followed by 1 day of carbohydrate-rich diet resulted in muscle glycogen sparing and a higher fat oxidation during submaximal exercise compared with subjects continuously consuming a carbohydrate-rich diet. In contrast, neither a higher fat oxidation nor a difference in muscle glycogen breakdown rate was observed in our laboratory's previous study after a dietary switch to a carbohydrate-rich diet for 1 wk after 7 wk on a fat-rich diet, compared with subjects consuming a carbohydrate-rich diet all along (16). Still, because higher preexercise muscle glycogen stores in the group switching from fat to carbohydrate diet by itself would be expected to increase muscle glycogen breakdown (15), it could be argued that identical muscle glycogen breakdown rates in the two groups in fact represent a relative impairment of breakdown after the dietary switch. However, in neither of these latter studies was a detailed characterization of muscle metabolism during exercise made. The aim of the present study was to investigate substrate metabolism after prolonged adaptation to training and a fat-rich diet followed by a carbohydrate-rich diet and to compare muscle metabolism to that observed when a carbohydrate-rich diet was taken throughout the period. METHODS AND MATERIALSThirteen healthy, untrained male subjects, age 27 ± 1 yr, height 182 ± 2 cm, weight 87 ± 3 kg and maximal oxygen uptake 3.9 ± 1 liters O2/min participated in the study. The fiber type composition in the vastus lateralis muscle was 55 ± 3% type I, 34 ± 2% type IIA, and 11 ± 2% type IIX fibers (type IIB in old nomenclature). Subjects were fully informed of the nature and the possible risks associated with the study before they volunteered to participate. The study was approved by the Copenhagen Ethics Committee, and all subjects gave written consent. The experiment was a longitudinal diet-training intervention study. Initially, subjects were randomized into two groups. Over 8 wk, both groups followed a training program. Over the first 7 wk, one group consumed a fat-rich diet and the other group a carbohydrate-rich diet (see below for details). During the last week, both groups consumed a carbohydrate-rich diet. Through the remainder of this paper, the group that switched diet will be referred to as Fat-Carbohydrate (Fat-CHO in tables and figures) and the other group as Carbohydrate (CHO in tables and figures). After 8 wk of the diet and training regimen, substrate metabolism was investigated in a 60-min exercise bout performed at ∼70% of maximal oxygen uptake on a modified Krogh bicycle ergometer. Over the experimental period, the subject's maximal oxygen uptake was determined before diet change or training, after 3.5 and 6.5 wk of the experiment. An exercise test was also performed after 7 wk, the results of which have been published separately (17). The experimental diets are similar to those applied and described in our laboratory's previous studies (16). In brief, the energy composition of the fat-rich diet was 21% carbohydrate, 17% protein, and 62% fat, and the carbohydrate-rich diet was 65% carbohydrate, 15% protein, and 20% fat. Thus the diets had a markedly different fat and carbohydrate content but a similar protein content. The habitual diet and energy intake were determined from 4-day diet records, and, in addition, individual energy intakes were estimated from the World Health Organization's equation for calculation of energy needs (34). On the days of training, the calculated energy expenditure during training was added to the daily energy intake and consumed immediately after training. During the experimental period all food intake was strictly controlled and weighed to within 1 g. The subjects weighed themselves every morning, and the individual dietary energy intake was adjusted such that body weight changes were minimized. [1-13C]palmitate (99% enriched) and NaH13CO3 were purchased from Tracer Technologies (Newton, MA). The palmitic acid tracer in solution was added to methanolic potassium hydroxide to form the potassium salt, which was then dried under nitrogen, redissolved in sterile water, and passed through a 0.22-μm sterile filter. It was then mixed with sterile 20% (wt/vol) human albumin (State Serum Institute, Copenhagen, Denmark), to which it was bound. Subjects were asked to refrain from physical activity 2 days before the start of the studies. The subjects reported to the laboratory in the morning after a 12-h fast after traveling to the laboratories either by car or bus. After subjects spent 30 min in a supine position, a needle biopsy was taken with suction from the vastus lateralis muscle by using local anesthesia with 5 ml of 1% lidocaine (3). After this, the training and diet regimen was begun. Over 8 wk, both dietary groups followed an identical, supervised training program using a bicycle ergometer. During the whole period, physical training was performed four times a week. Each training session lasted between 60 and 75 min, and exercise intensity, which was carefully controlled, varied from 50 to 85% of maximal oxygen uptake. The training program consisted of four different protocols with exercise intervals varying in length and duration interspersed with breaks of active recovery. Training intensity was adjusted to changes in maximal oxygen uptake measured after 3.5 wk of the training period. At every training session, heart rate was monitored; pulmonary oxygen uptake was measured frequently. Thus the training regime could be monitored and adapted as required. After 7 wk, all subjects arrived overnight fasted, and a muscle biopsy was obtained from the vastus lateralis for substrate concentration measurement, described in more detail in a separate paper (17). However, these muscle substrate data are also presented here. After the muscle biopsy, the subjects participated in an exercise test (17). At this stage, the subjects on the fat diet switched to the carbohydrate-rich diet, and the group that had consumed the carbohydrate-rich diet continued on that diet. Both groups continued the training over the ensuing week. After a week, the subjects again reported to the laboratory, overnight fasted and having not trained for 2 days. After 30 min of rest in the supine position, Teflon catheters were placed by an aseptic technique in the femoral artery and vein of the same leg under local anesthesia (1% lidocaine), and the tips were advanced to ∼2 cm above (arteries) and below (vein) the inguinal ligament. A thermistor (Edslab Probe 94-030-2.5-F, Baxter Healthcare) for measuring venous blood temperature was advanced 8 cm proximal to the catheter tip. A catheter was also inserted into an antecubital vein for the infusion of [1-13C]palmitate. The catheters were flushed with sterile sodium citrate (∼3.6 mM). A needle biopsy was obtained with suction from the vastus lateralis muscle. Subjects were then placed in a semisupine position, where they rested for 30 min. Blood was sampled simultaneously from the femoral artery and vein, and femoral venous blood flow was measured by the thermodilution method by use of bolus injections of 3 ml ice-cold saline into the femoral vein (1). Resting oxygen uptake and background breath enrichment samples were then collected. Subsequently, the bicarbonate pool was primed with a bolus of NaH13CO3 (0.1 mg/kg), and a continuous infusion of [1-13C]palmitate (99% enriched) was started by use of a calibrated syringe pump (Vial Medical SE 200B, Simonsen & Weel, Copenhagen, Denmark) set at a constant rate of 0.06 μmol · kg−1 · min−1. Blood samples were taken after 50, 55, and 60 min of infusion. Subjects were then positioned on the Krogh cycle ergometer, and exercise was begun at ∼80% of their individual pretraining maximal oxygen uptake without a prior warm-up. During exercise, blood samples were taken every 15 min. For the analyses, very-low-density lipoprotein-triacylglycerol (VLDL-TG) concentration was only measured at 0, 30, and 60 min, and the hormone concentrations were not measured in the 45-min sample. Venous blood flow was measured frequently via continuous infusions of ice-cold saline according to the thermodilution principle (1). Expired air was sampled into Douglas bags at the same time, and subsequently small aliquots were collected into evacuated glass tubes (Vacutainer, Becton Dickinson, Meylan, France) for analysis of 13CO2 enrichment. Exercise heart rate was recorded continuously with a PE 3000 Sports Tester (Polar Electro). Water intake during exercise was standardized with subjects drinking 200 ml of water every 20 min. Immediately after 60 min of exercise, a further biopsy was taken from the vastus lateralis muscle of the opposite leg through a new incision, suitably anesthetized. Fatty acids (FA) were extracted from plasma, isolated by thin-layer chromatography, and converted to their methyl esters. The arterial and venous isotopic enrichment of plasma [13C]palmitate was determined by gas chromatography-mass spectrometry (GC-MS, INCOS XL, Finnigan Mat, Hemel, Hempstead, UK) by selected ion monitoring of ions at mass-to-charge ratio of 270 and 271. Heptadecanoate (C17) was used as an internal standard for quantification of total palmitate. Enrichment of 13CO2 in expired air was analyzed by isotope ratio-mass spectrometry (Europa Scientific 2020 IRMS) as previously described (28). The concentration of isotope in the infusate was determined, so that the exact infusion rate could be calculated. Blood glucose and lactate were analyzed on a glucose and lactate analyzer (Yellow Springs Instruments, Yellow Springs, OH). Plasma glycerol was analyzed as described by Wieland (33). Total plasma FA was measured fluorometrically as described by Kiens et al. (19). VLDL-TG in serum was isolated by ultracentrifugation at a density of 1.006 g/cm3 and then analyzed as described by Kiens et al. (20). Insulin in arterial plasma was determined by use of a radioimmunoassay kit (Insulin RIA100, Pharmacia, Sweden), and catecholamines in arterial plasma were determined by a radio enzymatic procedure (8). Blood oxygen saturation was measured on an OSM-2 hemoximeter (Radiometer Copenhagen). Hemoglobin was determined spectrophotometrically on the hemoximeter by the Drabkin and Austin cyan-methemoglobin method (11a). Blood gases (Pco2, Po2) and pH were measured with the Astrup technique (ABL 30, Radiometer, Copenhagen, Denmark). Hematocrit was determined in triplicate from microcapillary tubes. The biopsies were frozen in liquid nitrogen within 10–15 s of sampling. Before freezing, a section of the samples was cut off, mounted in embedding medium, and frozen in isopentane cooled to its freezing point in liquid nitrogen. Both parts of the biopsy were stored at −80°C until further analysis. Before biochemical analysis, muscle biopsy samples were freeze-dried and dissected free of connective tissue, visible fat, and blood with the use of a stereomicroscope. Muscle glycogen concentration was determined as glucose residues after hydrolysis of the muscle sample in 1 M HCl at 100°C for 2 h (22). Muscle triacylglycerol was determined as previously described (21). In brief, 20 mg wet wt of muscle tissue were freeze-dried and dissected free of all visible adipose tissue, connective tissue, and blood by the use of a stereomicroscope, leaving the muscle fibers for further analysis. The muscle fibers were mixed, and ∼1 mg dry wt of the pooled fibers was used for measurement of the intramyocellular triacylglycerol concentration. Glycerol from the degraded triacylglycerol was assayed fluorometrically as described previously (21). Serial transverse muscle sections were stained for myofibrillar ATPase to identify fiber type composition (4). Total muscle GLUT-4 protein content was assayed by Western blotting using a primary antibody against the 13 COOH-terminal amino acids of GLUT-4. Whole body oxygen uptake and carbon dioxide excretion at rest and during exercise were determined by collection of expired air in Douglas bags. The volume of air was measured in a Collins bell-spirometer (Tissot principle, Collins W.E., Braintree, MA), and the fractions of oxygen and carbon dioxide were determined with paramagnetic (Servomex) and infrared (Beckmann LB-2) systems, respectively. Gases of known composition were used to calibrate each system regularly. Uptake and release of substrates and metabolites across the leg were calculated from femoral arterial and venous differences multiplied by plasma or blood flow, according to Fick's principle. Indirect calorimetry calculations were performed according to the stoichiometric equations given by Frayn (12). Substrate balance across the leg was calculated using an active muscle mass of 4.6 kg for each leg, which was estimated as the difference between total carbohydrate oxidation and measured glucose uptake and lactate release divided by the measured muscle glycogen breakdown. To assess the contribution of protein oxidation to exercise, a nitrogen excretion rate of 135 μg · kg−1 · min−1was used, as described by Romijn et al. (27). FA oxidation was determined by converting the rate of triacylglycerol oxidation (in g · kg−1 · min−1) to its molar equivalent, with an average molecular weight of triacylglycerol assumed to be 860 g/mol (12). The calculations of rate of appearance (Ra) and disappearance (Rd), were performed by the use of Stele's non-steady-state equations (31) modified for the use of stable isotope tracer infusion (9, 27) Ra= F−V [(Cpalm t2+Cpalm t1)/2][(Et2−Et1)(t2−t1)](E2+E1)/2 Rd= Ra−V (Cpalm t1−Cpalm t2)/(t2−t1) where F is infusion rate, F13co2 is the content of13CO2 in the breath, V the effective volume of distribution (40 ml/kg), Cpalmt1 and Cpalmt2 are the concentrations of plasma palmitate at time t1 andt2, and Et1 and Et2 are the palmitate enrichments, respectively. The Ra of FAs was determined as the product of the fractional contribution of palmitate to total FA concentration and the Ra palmitate. The percentage of the tracer infused that was oxidized was calculated as the%Tracer oxidized=FCO2×(C)−1×(F)−1×100 where C is the acetate correction factor (C = 0.9) as reported by Sidossis and colleagues (29). In the present experiment, it was not ethically feasible to do acetate correction trials for each individual; thus we choose to apply an exercise correction at 0.9 as described and utilized previously (29).The plasma FA oxidation was determined as PlasmaFAoxidation=RdFA×%tracer oxidized The estimated oxidation of FAs originating from sources other than plasma FA was calculated asOxidation of nonplasmaFA =total fat oxidation−plasmaFAoxidation Across-the-leg extraction can be calculated asFractional extraction =[(Cpalm art×Epalm art−Cpalm vein×Epalm vein]Cpalm art×Epalm art The calculated tracer-derived fractional extraction was used to calculate actual tracer determined FA uptake over the leg as the measured arterial FA delivery multiplied by the fractional extraction.Results are given as means ± SE. Two-way ANOVA with repeated measures for the time factor was performed to test for changes due to diet and/or time. In the case of significant main effects or interactions, a Student-Newman-Keuls post hoc test was performed to discern statistical differences. In all cases, an α of 0.05 was taken as the level of significance. RESULTSThe habitual dietary energy and nutrient intake were similar in the two groups (Table 1). Over the experimental period, energy intake was also similar in the two groups and significantly higher than the habitual daily energy intake. Throughout the 8 wk, the subjects adhered to the prescribed dietary intake both during the first 7 wk (17) and during the 8th week, as evidenced by the close resemblance between the actual intakes and the prescribed dietary contents (Table 1). After the switch from a high-fat to a high-carbohydrate diet, the Fat-Carbohydrate group increased (P < 0.05) their daily intake of carbohydrates (in g) by 221% and vice versa decreased (P < 0.05) their fat intake by 65%. Protein intake was slightly lower (P < 0.05) in the Fat-Carbohydrate group than in the Carbohydrate group. Over the experimental period body weight decreased (P < 0.05) similarly in both groups from 87.4 ± 2.9 to 86.1 ± 2.9 kg.
Subjects trained under supervision a total of 31 ± 1 times. Before the experimental period, maximal oxygen uptake was similar in the two groups at 3.8 ± 0.1 and 4.1 ± 0.2 l/min, and after the training period it was similarly increased (P < 0.05) to 4.1 ± 0.1 and 4.4 ± 0.2 l/min in the Fat-Carbohydrate group and in the Carbohydrate group, respectively. After 8 wk, subjects exercised at an oxygen uptake of 2.9 ± 0.1 l/min, which was equivalent to a workload of 72 ± 3 and 70 ± 2% of posttraining maximal oxygen uptake in the Fat-Carbohydrate group and in the Carbohydrate group, respectively. Over the first 15 min of exercise, heart rate increased (P < 0.05) from rest to 155 ± 5 and 150 ± 2 beats/min, whereafter a progressive increase (P < 0.05) to 165 ± 5 and 160 ± 4 beats/min was observed in the Fat-Carbohydrate and the Carbohydrate groups, respectively. During exercise, respiratory exchange ratio (RER) values remained constant throughout the exercise period in both Fat-Carbohydrate (0.94 ± 0.01) and Carbohydrate (0.92 ± 0.02) groups, and RER was similar between the groups. During exercise, whole body fat oxidation (Table2) and carbohydrate oxidation (175 ± 4 and 186 ± 5 μmol · min−1 · kg−1in the Fat-Carbohydrate and the Carbohydrate group, respectively) were similar between groups.
Resting leg blood flow was similar in the two groups: 0.29 ± 0.04 and 0.36 ± 0.04 liters blood/min in the Fat-Carbohydrate group and in the Carbohydrate group, respectively. The blood flow increased (P < 0.05) similarly during the first 15 min of exercise to 6.3 ± 0.3 and 6.0 ± 0.2 liters blood/min in the Fat-Carbohydrate and Carbohydrate group, respectively, after which no further increases were observed. In the Fat-Carbohydrate group, arterial blood glucose concentrations increased (P < 0.05) progressively and continuously through exercise, whereas in the Carbohydrate group the arterial glucose increased (P < 0.05) until 30 min, after which point it decreased (P< 0.05) at the termination of exercise down to the initial value (Fig.1A). After 60 min, arterial glucose concentration was borderline significantly higher (P = 0.07) in the Fat-Carbohydrate group diet than in the Carbohydrate group. Glucose delivery was not significantly different between groups and was on average through exercise 30.2 ± 1.6 and 29.9 ± 1.4 mmol/min in the Fat-Carbohydrate group and in the Carbohydrate group, respectively. In both groups, during the exercise bout, glucose uptake across the leg increased similarly, and at all time points the uptake was lower (P < 0.05) in the Fat-Carbohydrate group than in the Carbohydrate group (Fig.1B). Glucose clearance increased (P < 0.05) across the exercise bout to 0.20 ± 0.02 l/min in the Fat-Carbohydrate group and to 0.35 ± 0.06 l/min in the Carbohydrate group, and during the later 30 min glucose clearance was lower (P < 0.05) in the Fat-Carbohydrate group than in the Carbohydrate group. Arterial blood lactate concentrations increased (P < 0.05) similarly from rest to 15 min to 2.3 ± 0.4 mmol/l in both groups, and no further changes were observed. Throughout the exercise, lactate release was similar between the groups, and, after an initial increase to 0.72 ± 0.24 mmol/min after 15 min (P < 0.05), a continuous decrease (P < 0.05) was observed across the rest of the exercise bout to 0.30 ± 0.11 mmol/min. Fig. 1.Arterial blood glucose concentrations (A) and glucose uptake across the leg (B) at rest and during exercise after 8 wk of training and adaptation to 7-wk fat-rich followed by 1-wk high-carbohydrate diet (Fat-CHO) or 8 wk carbohydrate-rich diet (CHO). Values are means ± SE. * P < 0.05 compared with resting values; † P < 0.05, Fat-CHO vs. CHO Arterial plasma FA concentrations were not significantly different during exercise in the two groups (Fig.2A). After 15 min of exercise, arterial plasma FA concentrations were at a nadir, after which a continuous increase (P < 0.05) was observed until the end of exercise in both groups (Fig. 2A). During the 60 min of exercise, FA delivery was not significantly different between groups, averaging 1.18 ± 0.24 and 1.52 ± 0.22 mmol/min in the Fat-Carbohydrate group and the Carbohydrate group, respectively. During exercise, plasma net FA uptake across the leg was lower (P < 0.05) in the Fat-Carbohydrate group than in the Carbohydrate group, averaging 99 ± 24 and 166 ± 28 μmol/min, respectively. The FA clearance was not significantly different between groups and was on average 0.24 ± 0.07 l/min in the Fat-Carbohydrate group and 0.34 ± 0.04 l/min in the Carbohydrate group during exercise. During the first 30 min of exercise, the average arterial serum VLDL-TG concentration was 1.06 ± 0.08 and 0.76 ± 0.15 mmol/l in the Fat-Carbohydrate and Carbohydrate groups, respectively, and a significant decrease, to 0.96 ± 0.13 and 0.70 ± 0.29 mmol/l for Fat-Carbohydrate and Carbohydrate, respectively, was observed at 60 min. No measurable VLDL-TG-uptake was observed across the leg during the exercise, −0.02 ± 0.07 and −0.02 ± 0.06 mmol/min in the Fat-Carbohydrate group and in the Carbohydrate group, respectively. The arterial plasma glycerol concentration was similar in the two groups and increased (P < 0.05) continuously from 40 ± 4 μmol/l at rest to 136 ± 20 μmol/l at the end of exercise. During exercise, plasma glycerol release across the leg was similar between groups and increased (P < 0.05) from rest 17 ± 2 to 84 ± 23 μmol/min after 15 min. After 30 min of exercise, a small decrease (P < 0.05) was observed to 31 ± 17 μmol/min, whereafter plasma glycerol release was not further changed. Fig. 2.Arterial plasma fatty acid (FA) concentrations (A) and total tracer-determined FA uptake across the leg (B) during exercise after 8 wk of training and adaptation to Fat-CHO or CHO diets. Values are means ± SE. * P < 0.05 compared with rest. The arterial epinephrine concentration increased (P < 0.05) similarly from 0.71 ± 0.21 and 0.52 ± 0.09 nmol/l at rest to 1.79 ± 0.21 and 1.93 ± 0.41 nmol/l after 60 min in the Fat-Carbohydrate group and in the Carbohydrate group, respectively. Likewise, the arterial norepinephrine concentration increased (P < 0.05) similarly from 4.2 ± 1.6 and 3.7 ± 1.5 nmol/l at rest to 14.4 ± 1.3 and 15.8 ± 2.3 nmol/l after 60 min of exercise in Fat-Carbohydrate and Carbohydrate, respectively. The arterial plasma insulin concentration decreased (P < 0.05) from 10.7 ± 3.7 and 6.2 ± 1.1 μU/ml at rest to 6.8 ± 1.3 and 3.5 ± 0.4 μU/ml after 15 min of exercise in Fat-Carbohydrate and Carbohydrate, respectively, after which no further changes were observed. At rest, the plasma insulin concentration was significantly higher in the Fat-Carbohydrate group than in the Carbohydrate group; however, during exercise no significant difference between groups was discernable. Muscle glycogen concentration before training was similar in the two groups (Table 3). After 7 wk, the glycogen concentration was unchanged after fat diet adaptation and significantly lower (P < 0.05) than after 7 wk of carbohydrate diet, whereas after the 7-wk carbohydrate diet the glycogen concentration was increased (P < 0.05) by 46% compared with the initial values (reported in Ref.17). After 8 wk, the glycogen concentration was significantly higher (P < 0.05) in the Fat-Carbohydrate group than in the Carbohydrate group and higher (P < 0.05) than the initial values (Table 3). In the Carbohydrate group, glycogen storage remained at the level observed after 7 wk. Muscle glycogen breakdown was similar in the Fat-Carbohydrate group and in the Carbohydrate group across the 60 min of exercise (Table 3). After 8 wk, muscle triacylglycerol concentrations were similar to the initial values in both groups, and in the Fat-Carbohydrate group the values were significantly lower (P < 0.05) than those reported after 7 wk (reported in Ref. 17). This finding clearly implies that muscle triacylglycerol stores are prone to rather large fluctuations when nutrient composition and physical activity are markedly altered. The total muscle GLUT-4 protein content was significantly increased (P < 0.05) with training after 7 wk in both groups (Table 3). After 8 wk, total muscle GLUT-4 protein content was not further changed from the 7-wk values.
The enrichment of plasma [13C]palmitate to [12C]palmitate, the tracer-to-tracee ratio, decreased (P < 0.05) from rest to 15 min into the exercise bout, after which a stable plateau was maintained until the termination of exercise (Table 2). During exercise, the arterial palmitate enrichment was similar between groups, and, as expected, the arterial enrichments were higher (P < 0.05) than the venous enrichments during exercise. This resulted in an average fractional extraction over the exercise calculated to 28 ± 14 and 22 ± 3% and a net extraction of 9 ± 2 and 10 ± 1% in the Fat-Carbohydrate group and the Carbohydrate group, respectively. The tracer-derived total leg FA-uptake during exercise from 30 min and onward was not different between groups (Fig. 2B). The release of FA, calculated as the total leg FA uptake (tracer derived) minus the net FA uptake, was similar between groups (157 ± 37 μmol/min Fat-Carbohydrate; 155 ± 61 μmol/min Carbohydrate). The whole body Ra and Rd of palmitate and thus FA was stable across the last 30 min of exercise in both groups, and no differences were observed between groups (Table 2). Through the exercise bout, the plasma FA oxidation was similar in the two groups (Table 2), as was the proportion of FA oxidized. The estimated oxidation of lipid sources other than plasma FA remained constant through exercise and was not significantly different between groups (Table 2). The substrate oxidation across the leg through the exercise bout is depicted in Fig. 3, and the muscle glycogen contribution is based on an estimated active muscle mass of 4.6 kg. The muscle mass active during exercise was estimated from plasma glucose uptake, the measured glycogen breakdown, and total carbohydrate oxidation, determined by indirect calorimetry. The substrate that is not covered in the sum of measured muscle substrates amounts to 15 ± 6% in the Fat-Carbohydrate and 6 ± 13% in Carbohydrate group. Fig. 3.Leg substrate utilization during 60 min bicycle exercise in the fasted state after 8 wk of training and adaptation to Fat-CHO or CHO diets. The contribution from glycogen was determined as net breakdown multiplied by the estimated active muscle mass of the leg (4.6 kg). Plasma FA contribution was calculated as the leg total FA uptake multiplied by the whole body %FA oxidized assessed by stable isotope tracer methodology, and contribution of blood glucose and lactate was calculated from blood flow multiplied by the arteriovenous difference through the exercise. The partition between nonprotein fat-carbohydrate oxidation calculated from whole body respiratory exchange ratio is indicated on the right side of each dietary treatment. Values are means. * P < 0.05 Fat vs. CHO. DISCUSSIONIn the present study, fuel utilization in the exercising leg was studied in two groups after prolonged adaptation to either fat- or carbohydrate-rich diet combined with training, followed by a carbohydrate-rich diet for an additional week in both groups. The main findings were that preexercise muscle glycogen stores were increased and leg glucose uptake during submaximal exercise decreased in the group that switched to a carbohydrate diet after the fat adaptation (Fat-Carbohydrate) compared with the group that remained on the carbohydrate-rich diet (Carbohydrate). The tracer-determined uptake of plasma long chain FAs (LCFA) across the leg was not significantly different between the groups and neither was whole body tracer determined Rd of plasma LCFA for palmitate. The RER values demonstrated that the overall proportions of carbohydrate and lipid utilization were similar in the two groups during exercise. The similar RER values during exercise in the two groups are in agreement with our laboratory's previous findings (16), in which a switch to a carbohydrate-rich diet after prolonged fat diet adaptation completely abolished the increased fat oxidation during exercise that was observed after prolonged fat adaptation (17). Burke and colleagues (5) found that RER values during 2 h cycling at 70% of peak oxygen uptake in well-trained athletes were lower after short-term fat adaptation (5 days) followed by 1 day of carbohydrate loading (70–75% energy % carbohydrate) compared with 6 days of adaptation to the carbohydrate-rich diet. In addition, they reported that tracer-determined whole body glucose uptake during exercise was similar in the two groups. Thus Burke and colleagues demonstrated that the observed decrease in carbohydrate oxidation was solely due to muscle glycogen sparing in the fat-adapted group. Intake of a fat-rich diet followed by carbohydrate loading for 1 day also lowered carbohydrate oxidation during a 4-h bicycle exercise trial at 65% of peak oxygen uptake compared with when carbohydrates were ingested during the whole period (6). In contrast, we found that muscle glycogen utilization in absolute numbers was similar between groups, but blood glucose uptake across the leg was lower during exercise in Fat-Carbohydrate than in Carbohydrate. The latter finding is most likely due to the increased muscle glycogen concentration in the Fat-Carbohydrate group (Table 3) because studies have demonstrated that glucose uptake during muscle contractions is inversely related to the muscle glycogen level (11, 13). The molecular mechanism behind the effect of glycogen on glucose uptake has been shown to involve impaired GLUT-4 translocation to the surface membrane (11). Increased muscle glycogen stores have also been demonstrated to be associated with decreased activation of 5′-AMP-activated protein kinase during contractions (10,25), but whether this is related to decreased muscle glucose uptake during exercise is uncertain (10, 23, 25). The difference in muscle glucose uptake between the two groups could not be ascribed to differences in total muscle GLUT-4 protein content because this was similar in the two groups (Table 3). However, it does not rule out the possibility of either recruitment differences in GLUT-4 transporters or differential distribution of the pools of transporters between the muscle membrane and intracellular stores. High plasma concentrations of LCFA have been shown to decrease muscle glucose uptake during exercise (14), but, because these were similar in the two groups during exercise (Fig. 2), group differences in leg glucose uptake in the present study could not be ascribed to differences in plasma LCFA. Despite the higher resting muscle glycogen concentration in the Fat-Carbohydrate group, the actual muscle glycogen breakdown during exercise was not significantly different between groups. It has been demonstrated that muscle glycogen breakdown rate normally is related to preexercise muscle glycogen concentrations both in vitro (18,24) and in vivo (15). Thus the absence of increased muscle glycogen breakdown in the Fat-Carbohydrate group suggests that muscle glycogen breakdown may be attenuated after fat diet adaptation followed by carbohydrate loading. This interpretation is supported by our laboratory's previous study (16) in which exhaustion in subjects on a similar diet as the present Fat-Carbohydrate group occurred in spite of very high (∼500 μmol/g dry wt) muscle glycogen levels. It thus appears that prolonged adaptation to a fat-rich diet, even when switching to a carbohydrate-rich diet for an additional week, affects muscle metabolism during exercise in such a way that muscle glycogen breakdown is impaired. The molecular mechanisms behind this phenomenon remain to be established. It might be argued that if glucose utilization is decreased and the overall oxidation of carbohydrates is unchanged in the Fat-Carbohydrate group compared with the Carbohydrate group, then glycogen breakdown must be increased. This was not found. However, the decrease in leg glucose uptake in terms of energy is rather small (Fig. 3), and it is quite possible that a similarly small increase in muscle glycogenolysis is missed. Furthermore, the overall combustion of carbohydrates and fat is calculated from whole body RER values that may not be exactly the same as respiratory quotient in the muscle and therefore small changes in the balance of carbohydrate vs. fat combustion in the leg may not necessarily be picked up at the level of pulmonary gas exchange. However, the similar isotope-derived calculated leg and whole body uptake of LCFA and absence of measurable breakdown of intramyocellular triacylglycerol in the both groups supports the concept of no significant difference in overall carbohydrate and fat utilization in the two groups. In our previous study (17), breakdown of VLDL-TG across the leg during exercise could account for a significant part of the lipid utilization across the leg after prolonged fat diet adaptation. However, in the present study there was no measurable breakdown of VLDL-TG across the leg in either group. Apparently, 1 wk of carbohydrate-rich diet after the fat adaptation is enough to abolish the significant contribution of VLDL-TG to energy provision during exercise in the fat-adapted state. Intramyocellular triacylglycerol breakdown during exercise was not measurable in the present and several previous studies performed in men (2, 19, 26, 30, 32). When calculating the relative contribution of oxidized substrates in the two groups, there is an apparent lack of substrates (Fig. 3). The extent to which this may be accounted for by uptake of VLDL-TG or breakdown of intramyocellular triacylglycerol, which were both too small to be measurable, or by FAs liberated from adipocytes adherent to muscle cells, as suggested by Kiens et al. (19), is not possible to determine from the available data. In any case, the contribution of these potential sources of energy substrates to oxidation is likely to be minor. Finally, it should be considered that because the pulmonary RER may not be exactly the same as leg respiratory quotient, the balance between carbohydrate and lipid combustion may not be entirely correct and therefore the apparent missing substrate may not necessarily be lipid. In conclusion, when switching to a carbohydrate diet for a week after prolonged fat adaptation, overall fat and carbohydrate utilization during submaximal exercise was not different compared with the group that consumed the carbohydrate-rich diet all along. However, muscle glycogen concentration was increased by 27% on average in the group switching from fat adaptation to a carbohydrate-rich diet. This did not lead to increased muscle glycogen utilization during exercise but to a decrease in utilization of blood glucose. Thus it seems as if prolonged fat adaptation leads to impaired muscle glycogen utilization, which is not remedied by 1 wk of carbohydrate-rich diet. The skilled technical assistance of Martin Rollo, Irene Beck Nielsen, Betina Bolmgren, Inge Reich, Winnie Tågerup, and Nina Pluzek is acknowledged. FOOTNOTESREFERENCES
Page 23we have, recently, provided evidence for transcytosis from the luminal to the interstitial side of specimens of parietal pericardium of rabbit by showing that the unidirectional flux of albumin or dextran 70 from lumen to interstitium is greater than that in the opposite direction and that this difference disappears at 12°C or with a transcytosis inhibitor (8). Our findings agree with earlier morphological evidence for transcytosis of macromolecules from the luminal to the interstitial side of the mesothelium in rat parietal pericardium (18) and mouse parietal peritoneum (13,14, 16). In the present research, we tried to provide evidence for transcytosis from the pleural space in vivo. Under physiological conditions, protein removal from the pleural space occurs by lymphatic drainage through the stomata of the parietal mesothelium (2, 5, 20, 21, 23,25), by solvent drag because of liquid absorption through the visceral mesothelium by Starling forces (1-3), and probably by transcytosis. On the other hand, proteins enter the pleural space by solvent drag because of liquid filtration through the parietal mesothelium by Starling forces (1, 2, 5, 21, 23, 25) and by diffusion because protein concentration in the pleural liquid (22) is smaller than that in the interstitium adjacent to the mesothelium (23). The data obtained on the parietal pericardium in vitro (8) suggest that the removal of albumin from the pleural space by transcytosis might be appreciable even if transcytosis in the pleura were smaller than in the pericardium. One may, therefore, formulate the hypothesis that, in the case of transcytosis, the quantity of labeled albumin left in the pleural liquid a few hours after injection into the pleural space of a bolus with labeled albumin plus nocodazole (a transcytosis inhibitor; Refs. 12, 17) should be greater than that left after a similar bolus injection without nocodazole. Moreover, taking into account that the albumin removed from the pleural space by transcytosis is eventually drained into blood by the lymphatics of the connective tissue of the pleura, one may formulate a second hypothesis: namely, that in case of transcytosis the quantity of labeled albumin occurring in plasma a few hours after injection into the pleural space of a bolus with labeled albumin plus nocodazole should be smaller than that found after a similar bolus injection without nocodazole. To verify these hypotheses, we injected 0.3 ml of albumin-Ringer solution with a small amount of labeled albumin (or dextran 70) into the right pleural space of anesthetized rabbits. Two kinds of experiments were performed: in one the injected bolus contained nocodazole, in the other it did not. After 3 h, we measured the volume of the pleural liquid on the right side and the concentration of labeled albumin (or dextran 70) to determine the quantity of labeled molecules left in the right pleural liquid under the two conditions. Moreover, we measured the concentration of labeled albumin (or dextran 70) in plasma, and, assuming its volume to be 4% of body weight (6, 10, 15), we determined the quantity of labeled macromolecule in plasma under both conditions. METHODSThe experiments were performed on 39 rabbits (2.45–2.9 kg body wt). The animals were anesthetized with pentobarbital sodium (Sigma Chemical, 18 mg/ml); the initial dose was 2 ml/kg intravenously, and small additional injections (0.2 ml/kg) were delivered when required to maintain an adequate level of anesthesia. The trachea was cannulated, and air flow and tidal volume (obtained by electronic integration of the flow signal by means of a Hewlett-Packard 8815A respiratory integrator) were recorded on a 7418 Hewlett-Packard thermopaper oscillograph throughout the experiment. The left jugular vein was exposed, and 2.5 ml of blood were sampled for background fluorescence measurement (see below). With the rabbit in the left lateral posture, the sixth right intercostal space was cleared from skin and muscles down to the intercostal muscles. A double thread loop was prepared in these muscles, and a stainless-steel cannula (1.4 mm OD; 0.9 mm ID) connected to a 1-ml glass syringe containing the bolus was inserted in the pleural cavity and tightened to the muscles by one thread loop. The bolus (0.3 ml) was then injected into the pleural space, and the cannula was removed while the second thread loop was tightened. During the procedure, a layer of Ringer solution was maintained over the site of injection to prevent air entrance into the pleural space during insertion and removal of the cannula. The injected bolus consisted of a Ringer solution (composition, in mM: Na+ 139, K+ 5, Ca2+ 2.5, Mg2+ 1.5, Cl− 119, HCO3− 29,d-glucose 5.6) containing bovine serum albumin (8 mg/ml, Sigma Chemical). Macromolecules labeled with a fluorescent marker were added to the albumin-Ringer solution: bovine serum albumin conjugated with Texas red (2.5 mg/ml, Molecular Probes) in one series of experiments (n = 16) and dextran 70 kDa conjugated with Texas red (2.0 mg/ml, Molecular Probes) in another series (n = 16). Different concentrations were used for the two tracers because fluorescence intensity was somewhat less for the albumin conjugate. For each series, two groups of experiments (n = 8 each) were performed: one as control, and in the other one an inhibitor of transcytosis, nocodazole (Sigma Chemical; Refs. 12, 17), was added to the injectate. The quantity of nocodazole added to the injectate was such that its concentration in the pleural liquid at the beginning of the experiment was ∼60 μM. This concentration should decrease substantially during the course of the experiment because of turnover of pleural liquid and diffusion of nocodazole (molecular mass 301 Da; as a rough analogy, see the time course of labeled mannitol concentration in a 2-ml hydrothorax; Ref. 28). After injection, the rabbit was turned supine. Before cannula insertion and every 30 min during the experiment, the respiratory system was passively expanded with a volume two to three times the spontaneous tidal volume by means of a large syringe. Three hours after the injection, a 5-ml blood sample was withdrawn from the jugular vein or the left carotid artery; then the rabbit was killed by an overdose of anesthetic and was placed in the left lateral posture 80° head up. The right pleural space was opened, its liquid was carefully collected through a polyethylene tube connected to a 1-ml glass syringe, and its volume was measured (4, 22). The liquid of the left pleural space, where no injection was made, was also collected and its volume was measured. This volume was used as an index of the volume of liquid present in the right pleural space before injection (see results). The experiments were discarded when the liquid was macroscopically contaminated by blood, when pathological signs were evident, or when the volume of liquid collected from the left space was >0.6 ml (4, 22). To approximately assess the time course of labeled albumin concentration during the experiments, in an additional series of experiments (n = 7) with albumin Texas Red without nocodazole, the experiment was ended 1 h, instead of 3 h, after the injection. In all experiments, breathing frequency and tidal volume were measured during the initial, middle, and final part of each experiment, and minute ventilation was computed. The three values of ventilation were averaged to obtain the mean ventilation in a given experiment. Fluorescence intensity in the liquid collected from the right pleural space at the end of the experiment was measured by a spectrofluorophotometer (Shimadzu RF1501) (excitation 596 nm; emission 615 nm). A calibration curve was made for each experiment by diluting the solution prepared for injection into the pleural space. The background fluorescence of albumin-Ringer solution was measured and subtracted. The calibration was linear over the range of labeled molecule concentration between 0 and 15 μg/ml. A calibration factor was computed for each curve by linear regression through the points. Samples (25 or 50 μl) of the liquid collected from the right pleural space were diluted 120 or 60 times, respectively, with albumin-Ringer solution, so that the reading on the spectrofluorophotometer fell on the linear portion of the calibration curve. Readings were also made of1) similarly diluted samples of liquid collected from right and left pleural space of two extra rabbits, in which no bolus was injected, to obtain the background fluorescence of pleural liquid;2) similarly diluted samples of the liquid collected from the left pleural space (where no injection was made), to determine the concentration of labeled molecules entered into this liquid during the experiment; 3) plasma obtained from the blood sampled before injection, to obtain the background fluorescence in plasma; and4) plasma obtained from the blood sampled at the end of the experiment, to determine the concentration of labeled molecules entered in plasma during the experiment. Readings obtained in pleural liquid and plasma samples at the end of the experiment were corrected for background fluorescence. Checks for the possible occurrence of unbound tracer in the injectate were made by measuring fluorescence intensity in the dialysate (for ∼16 h at ∼37°C) or ultrafiltrate (by centrifuging at 5,000 g for 30 min through low-binding cellulose ultrafiltration membranes with a 10-kDa nominal molecular weight cutoff, PLTK, Millipore) of diluted solutions of the labeled molecules. The fluorescence found both in the dialysate and in the ultrafiltrate, corresponding to that due to unbound Texas Red, was <0.1% of that in the original solution, both with labeled albumin and with labeled dextran 70. To assess whether macromolecule catabolism occurred in the pleural space, in two experiments with albumin-Texas red and in two experiments with dextran 70-Texas red, diluted samples of the pleural liquid collected at the end of the experiment were ultrafiltered by the above procedure through ultrafiltration membranes with a 30-kDa nominal molecular weight cutoff (PLTK, Millipore). Fluorescence in the filtrate was <0.1% of that in the solution before filtration. Hence, an appreciable catabolism of albumin or dextran 70 does not seem to occur in the pleural space during the experiments. The quantity of labeled macromolecule present in the pleural liquid of each space at the end of the experiment was obtained by multiplying the corresponding concentration by the overall volume of liquid in the pleural space. This is given by the volume collected plus that remaining adherent to the walls of the space when it is opened. The former was measured, and the latter was computed from the mean value found in previous ad hoc experiments (0.25 ml in 2.23-kg rabbits; Ref.22). Because the volume of liquid adherent to the walls is related to their surface area, this volume of liquid was computed as 0.25 ml × (body wt2/3/2.232/3). The quantity of labeled macromolecule present in plasma at the end of the experiment was obtained by multiplying the corresponding concentration by the volume of plasma in that rabbit. The volume of plasma was assumed to be 4% of the body weight (Refs. 6,10, 15). The surface area of the parietal and of the visceral pleura was obtained from previous measurements (22) by normalizing to body weight2/3. Because the surface area of the parietal pleura previously measured was only that facing the lung, the surface area of the costophrenic sinus was added by assuming it to be ¼ of the parietal pleura facing the lung. The results are reported as means ± SE. Statistical difference between groups was assessed by unpaired t-test. RESULTSThe volume of liquid in the left pleural space (where no injection was made) was similar in both kinds of experiments of both series, that with labeled albumin and that with labeled dextran 70 (Tables1 and 2). This volume of liquid was also similar to that previously determined in ad hoc experiments (4, 22), taking into account the small difference in body size by normalizing to body weight2/3(which yields 0.51 ml for the present rabbits). Therefore, the volume of liquid in the right pleural space at the beginning of the experiment should be ∼0.81 ml, 0.51 ml being the volume under physiological conditions and 0.3 ml being the injected volume. Because of the variance in the volume of pleural liquid among rabbits and of the error involved in its measurement, to get more reliable data it is better to pool together the values of liquid volume in the right pleural space of both series of experiments in a given condition, i.e., control or with nocodazole, so that each group consists of 16 experiments. By so doing, the volume of the right pleural liquid was smaller (P< 0.01) at the end of control experiments, 0.66 ± 0.04 ml, than at the end of experiments with nocodazole, 0.81 ± 0.04 ml. Therefore, transcytosis seems to remove 0.15 ml of liquid from the pleural space in 3 h, i.e., 0.05 ml/h or 26 μl · h−1 · kg−2/3. Moreover, the data suggest that in control experiments approximately half of the liquid injected was removed, whereas in the experiments with nocodazole the net removal of liquid was approximately nil.
The quantity of labeled albumin in the right pleural liquid was smaller (P < 0.01) at the end of control experiments, 318 ± 27 μg, than at the end of experiments with nocodazole, 419 ± 17 μg (Table 1). The quantities of labeled albumin left in the right pleural space should be a little greater than those in the right pleural liquid because of the labeled albumin adsorbed to the walls, which is not measured. This, however, does not represent a problem, because the quantity of labeled albumin adsorbed should be similar in both kinds of experiments and because what matters for our aims is the difference in quantity of labeled albumin left in the liquid between the two kinds of experiments. The difference in quantity of labeled albumin left in the right pleural liquid between nocodazole and control experiments, 101 μg, should provide the quantity of labeled albumin removed by transcytosis in 3 h. The quantity of labeled albumin in plasma was greater (P < 0.01) at the end of control experiments (64.3 ± 6.9 μg) than at the end of experiments with nocodazole (32.5 ± 5.6 μg, Table 1). The finding that the quantity of labeled albumin in plasma is greater in control than in nocodazole experiments fits with the occurrence of transcytosis because, when transcytosis is operating, more labeled albumin is removed from the pleural space, and therefore more labeled albumin reaches the blood. The quantity of labeled dextran 70 in the right pleural liquid was smaller (P < 0.01) at the end of control experiments, 283 ± 10 μg, than at the end of experiments with nocodazole, 381 ± 21 μg (Table 2). The same consideration made above on the labeled albumin adsorbed to the walls of the space also applies to the labeled dextran 70 adsorbed to the walls (which might be relatively greater in quantity, because dextran is not normally present in the pleural space). The difference in quantity of labeled dextran left in the right pleural liquid between nocodazole and control experiments, 98 μg, should provide the quantity of labeled dextran removed by transcytosis in 3 h. The quantity of labeled dextran in plasma was greater (P < 0.01) at the end of control experiments (22.3 ± 1.8 μg) than at the end of experiments with nocodazole (14.4 ± 1.5 μg, Table 2). The finding that the quantity of labeled dextran in plasma is greater in control than in nocodazole experiments fits with the occurrence of transcytosis for the reason mentioned above for labeled albumin. Pulmonary ventilation, as well as that normalized to body weight, was similar in both kinds of experiments of a given series (Tables 1 and2). DISCUSSIONThe finding that the quantity of labeled albumin or labeled dextran removed from the right pleural liquid is smaller in the experiments with nocodazole than in control ones suggests the occurrence of transcytosis from lumen to interstitium in the pleural mesothelium in vivo, in line with our in vitro evidence on the parietal pericardium (8). Actually, the quantity of labeled macromolecules removed by transcytosis should be somewhat underestimated because of the following reason. The concentration of labeled macromolecules in the right pleural liquid decreases markedly during the experiment because the liquid entering the pleural space during the experiment is essentially free of labeled macromolecules, whereas these leave the space by lymphatic drainage through the stomata of the parietal pleura (2, 5, 19, 21, 23, 25), by convection owing to the Starling forces through the visceral pleura (1-3), by diffusion (7), and by transcytosis (8). In the experiments with nocodazole, the concentration of labeled macromolecules should decrease more slowly than in control experiments because one mechanism removing macromolecules has been blocked. If the concentration of macromolecules in the experiments with nocodazole is a little higher than in control experiments, the quantity of labeled macromolecules leaving the right pleural space with the other mechanisms should be a little greater in the experiments with nocodazole than in control ones. Because pulmonary ventilation is similar in the two kinds of experiments of a given series, one can rule out the possibility that our results are affected by a different degree of lymphatic drainage from the pleural space caused by a different degree of ventilation. This control has been done because an increase in lymphatic drainage from large hydrothoraxes has been shown to occur when ventilation is increased (10). Moreover, the finding that the quantity of labeled albumin or labeled dextran 70 in plasma is smaller in the experiments with nocodazole than in control ones fits with the occurrence of transcytosis in pleural mesothelium in vivo. One could argue that nocodazole by disrupting the microtubules could modify the shape of the mesothelial (and endothelial) cells, and this, in turn, might reduce the size of the stomata of the parietal pleura, and, hence, reduce the lymphatic drainage through them. On the other hand, the electron micrographs by Hastings et al. (17) show that the shape of the alveolar and endothelial cells 2 h after instillation of nocodazole in rabbit alveoli was not altered. Moreover, the light micrographs by the same authors show that the shape of rabbit alveoli 2 h after instillation of nocodazole was not changed. The alveolar epithelium lacks a strong support like the connective tissue layers of the pleura: hence, after disruption of the microtubules, it should undergo a change in shape more easily than the mesothelium. Therefore, the finding that the shape of the alveoli is not altered by nocodazole makes it unlikely that the disruption of microtubules alters the shape of the mesothelial cells in such a way as to reduce the size of the stomata of the parietal pleura. The vesicular transport of albumin in the parietal pericardium of rabbits in vitro is only fluid phase, although this transcytosis seems triggered by albumin concentration (8). The situation might be different in vivo, where one cannot rule out the occurrence of receptors for albumin in the vesicular membrane because one cannot perform experiments with very low concentrations of unlabeled albumin, which are required to detect the competition for receptors between labeled and unlabeled albumin. In case the vesicular transport of albumin from the pleural space in vivo is only fluid phase, it may be computed from the vesicular liquid flow (0.05 ml/h, seeresults) times the concentration of albumin in the pleural liquid under physiological conditions (∼10 mg/ml; Ref.22) times the lumen-vesicle partition coefficient for albumin (∼0.78). This coefficient is given by (1 −a/r)3 (11), wherea is the hydrodynamic radius of the solute (3.55 nm for albumin; Ref. 7) and r is the radius of the vesicle (∼45 nm; Refs. 14, 26). This computation yields a value of 390 μg/h, or 203 μg · h−1 · kg−2/3. On the other hand, one may attempt to estimate the vesicular transport of albumin from the pleural space in vivo in a different way that does not require the assumption of a mere fluid-phase transport. This computation is based on the vesicular transport of labeled albumin (seeresults) and the ratio between the concentration of unlabeled albumin in the pleural liquid under physiological conditions (∼10 mg/ml; Ref. 22) and the mean concentration of labeled albumin in the right pleural liquid during the experiment. To estimate the latter, the time course of the concentration of labeled albumin in the right pleural liquid (Calb*) during the experiment has to be approximately drawn. To this end, in a separate group of seven rabbits we measured 1 h after injection of the control bolus the same parameters previously measured after 3 h. These data are reported in Table 3. Because of the mechanisms involved, the decay of concentration of labeled albumin should be a curve with upward concavity. The approximate time course of Calb* during the experiment was obtained by interpolation through the initial value (see results), the value after 1 h (Table 3), and the value after 3 h (Table 1). It is shown in Fig.1 (○), along with the time course of the quantity of labeled albumin in the pleural liquid during the experiment (●). Mean Calb* during the experiment was obtained by averaging the readings taken on the corresponding line every 12 min. Its value, 692 μg/ml, is 14.5 times smaller than the concentration of albumin in the pleural liquid under physiological conditions (10 mg/ml). Therefore, the vesicular transport of albumin from the pleural space under physiological conditions should be given by the vesicular transport of labeled albumin during the experiment (101 μg) times 14.5 divided by 3 = ∼488 μg/h, or ∼254 μg · h−1 · kg−2/3. This value is 25% greater than that computed from the vesicular liquid flow. This difference should be even greater because the rate of vesicular transport of labeled albumin (as it has been measured) is somewhat underestimated (see above). This difference suggests that the vesicular transport of labeled albumin from the pleural space in vivo is not only fluid phase, but the data on which it is based are not precise enough to afford a conclusion. For the rest of thisdiscussion, the vesicular transport of albumin in vivo will be taken as the mean between the values obtained by the two procedures, i.e., 439 g/h or 228 μg · h−1 · kg−2/3.
Fig. 1.Time course of concentration (○) and quantity (●) of labeled albumin in pleural liquid during the experiment. For further information, see text. The surface area of the parietal pleura in our 2.65-kg rabbits should be ∼115 cm2 (see methods). Therefore, if transcytosis occurred only through the parietal pleura, the vesicular transport of albumin from the pleural space would be ∼3.8 μg · h−1 · cm−2 and the vesicular liquid flow ∼0.4 μl · h−1 · cm−2. Though the morphological features of mesothelial cells seem similar on the parietal and visceral side, except for a greater density of microvilli on the visceral side (27), no direct evidence of transcytosis has been yet provided for the visceral mesothelium. The surface area of the visceral pleura in our rabbits should be ∼132 cm2 (see methods). Therefore, the overall surface area of the parietal and visceral pleura in our rabbits should be ∼247 cm2. If the rate of transcytosis in the visceral pleura were similar to that in the parietal one, albumin removal from the pleural space by transcytosis would be ∼1.8 μg · h−1 · cm−2 and the vesicular liquid flow ∼0.2 μl · h−1 · cm−2. In vitro albumin transcytosis through specimens of parietal pericardium of rabbits, with an albumin concentration in the solution similar to that occurring under physiological conditions (10 mg/ml), was 5 × 10−4μmol · h−1 · cm−2or 36 μg · h−1 · cm−2(8). Therefore, the rate of albumin transcytosis found in the parietal pericardium in vitro is one order of magnitude greater than that found in the pleural space in vivo in the present research. Part of this difference was expected because morphological studies showed a greater concentration of vesicles in the cytoplasm of the pericardial mesothelium (18, 19) than in that of the pleural mesothelium (26, 27). Moreover, the rate of albumin transcytosis measured in the present research should be somewhat underestimated for the reason indicated above. This underestimation, however, should not be such as to fill the gap. The lymphatic drainage from the pleural space of dogs has been determined by Miniati et al. (20), who made a 0.5-ml intrapleural injection of Ringer solution with131I-labeled albumin and a simultaneous intravenous injection of 125I-labeled albumin. Plasma activity of both tracers was followed for 24 h: the 131I-labeled albumin curve provided the output function from the pleural space, whereas the 125I-albumin curve served as input function for the interstitial space (including the serosal cavities). They found a lymphatic drainage from the pleural space of 0.02 ml · h−1 · kg−1. This lymphatic drainage of albumin was considered to occur mostly through the stomata of the parietal pleura. Therefore, taking into account that the concentration of albumin in the pleural liquid of dogs under physiological conditions is 6.1 mg/ml (22), the lymphatic drainage of albumin from the pleural space of their 17.5-kg dogs should be ∼122 μg · h−1 · kg−1, i.e., ∼314 μg · h−1 · kg−2/3. On the other hand, the lymphatic drainage from the pleural space does not necessarily occur mostly through the stomata, because albumin leaving the pleural space outside the stomata is also eventually drained into blood by the lymphatics of the connective tissue of the pleura (3). Miniati et al. pointed out that the initial upward concavity of the time course of plasma recovery of labeled albumin injected into the pleural space (Fig. 9 of their article) is due to a large liquid volume for albumin distribution interposed between pleura and plasma. Because they believed that most of albumin removal from the pleural space occurs through the stomata of the parietal pleura, they concluded that the lymphatic network was adequate to explain the delayed appearance of labeled albumin in plasma. The present finding of albumin transcytosis by the pleural mesothelium suggests that the lymphatics contribute only part of the initial upward concavity of the above-mentioned curve, the rest being due to the interstitial liquid of the pleural connective tissue, where albumin removed by transcytosis (parietal and, perhaps, visceral pleura) and by convection (visceral pleura) is distributed before being drained into blood by lymphatics. The possibility that a substantial part of albumin leaves the pleural space outside the stomata was considered by Broaddus et al. (9) after they found in sheep that the removal rate of hydrothorax computed from the clearance of labeled erythrocytes (which may leave the space only through the stomata) was 89% of that computed from the clearance of labeled albumin. Their hydrothoraxes, however, were very large (10 ml/kg body wt), and under this condition the lymphatic drainage from the pleural space may increase more than 10 times (9). Convection too increases (because of the increase in pleural liquid pressure), but to a smaller extent. No information is available on transcytosis, but its increase should be smaller than that of the lymphatic drainage. Consequently, the above difference of 11% between the overall (direct plus indirect) lymphatic drainage from the pleural space and that through the stomata of the parietal pleura (direct) should be smaller than that occurring under physiological conditions. Unfortunately, no data are available on the lymphatic drainage of albumin from the pleural space of rabbits. It has been stated that the turnover rate of pleural liquid per unit pleural surface area is ∼3 times greater in rabbits than in dogs (23). This statement, however, is based on a comparison between an estimate of liquid filtration through the parietal pleura of rabbits (24) and the lymphatic drainage of liquid from the pleural space of dogs, computed from albumin clearance from this space (20). This comparison, in turn, implies the assumption that the lymphatic drainage of liquid from the pleural space represents most of the liquid outflow from the pleural space, and that it occurs through the stomata of the parietal pleura (23). This is not the case because of transcytosis and because of the liquid outflow caused by Starling forces through the visceral pleura and that coupled to the active absorption of NaCl, which are not negligible under physiological conditions (5). Hence the turnover rate of pleural liquid per unit pleural surface area does not seem to be so much greater in rabbits than in dogs. Therefore, our results suggest that under physiological conditions the vesicular transport of liquid and albumin from the pleural space might contribute a substantial part of the overall removal of liquid and albumin from this space. We are grateful to Prof. D. Cremaschi for stimulating discussion and for critically reading the paper. Moreover, we thank R. Galli for skillful technical assistance. FOOTNOTESREFERENCES
Page 24muscle fatigue can be defined as the fall in maximum force-generating capacity of the muscle. During exercise, the magnitude and mechanisms of human skeletal muscle fatigue vary widely and depend to a large extent on the individual, the type of muscle, and the exercise stimulus or task. In general, fatigue may arise during muscular contractions due to failure at one or more sites along the pathway of force production from the central nervous system to the contractile apparatus (16). There is reason to believe that both age and gender can affect the fatigue process, although our understanding of these effects is hampered by a lack of consensus in the literature. Although it has been reported that older adults fatigue relatively more than young adults (12, 33) and that men fatigue more than women (18,20, 36), some investigators have found no effect of age (31, 35, 44) or gender (14) on fatigue. Still others have found that older subjects fatigue relatively less than younger subjects (3, 14). Along with the lack of clarity regarding the effects of age and gender on the magnitude of muscle fatigue, the mechanisms of these differences have not been established. Differences in fatigability across age or gender could occur as a result of differences in neural drive, fiber-type composition, contractile function, muscle membrane excitability, metabolic capacity, or muscle mass and blood flow. For example, it was recently suggested that central activation failure may play a relatively larger role in the fatigue of older compared with younger adults (2, 44). Other investigators have reported impairments in excitation-contraction coupling in the muscle of older adults (13), although the possible role of this impairment in fatigue has not been established. The results of some (37), but not all (9, 28), studies suggest that oxidative capacity may be impaired with aging, despite a general shift toward a more oxidative fiber-type profile in older compared with younger muscle (30, 34). An impaired oxidative capacity in the muscle of older adults might contribute to fatigue in this group. Finally, it is unclear how a gender-based difference in fatigue might interact with the aging process. In addition to the effects of activation, contractile function, and metabolism on muscle performance, the degree of fatigue that develops during exercise may be affected by muscle size and, consequently, vascular constriction during contraction. The impact of larger muscle mass, greater strength, and higher target tensions during exercise in men compared with women has been addressed in several studies. In the adductor pollicis, a gender-based difference in endurance time during a submaximal contraction persisted despite matching subjects to similar strengths (18). More recently, Hunter and Enoka (21) showed a gender difference in endurance (time to failure to maintain target tension) of the elbow flexor muscles during a contraction sustained at 20% maximal voluntary contraction (MVC) force but similar fatigue (fall in MVC) in men and women at the end of this exercise. Notably, the gender difference in endurance was negated by accounting for preexercise differences in muscle strength. These and other (14) results suggest that the relationship between muscle strength and fatigue should be examined in studies of the effects of age or gender on fatigue. The purpose of this study was to investigate the magnitude and mechanisms of fatigue (i.e., fall in MVC of the ankle dorsiflexor muscles) in healthy young and older men and women during a progressive, intermittent isometric dorsiflexion exercise protocol that proceeds from a steady-state oxidative phase to a more glycolytic, fatiguing phase (26). The dorsiflexor muscles are functionally important for locomotion, posture, balance (49), and the prevention of falls in older adults (7). Furthermore, habitual use of the dorsiflexor muscles may make men and women less susceptible to disuse deconditioning than use of muscles more typically involved in high-power activities (e.g., quadriceps femoris), which are not often used by older adults. To examine the mechanisms of fatigue, we obtained simultaneous measures of central and peripheral muscle activation, muscle contractile properties, and intramuscular energy metabolism by using a unique combination of voluntary and electrically stimulated muscle contractions, electromyography (EMG), and phosphorus-31-magnetic resonance spectroscopy. The relationship between strength and fatigue was also examined. To control for the effects of varying levels of physical activity on these measures, we studied individuals with similar, relatively sedentary habitual activity levels. By simultaneously measuring many of the factors that have been suggested to explain age- and gender-based differences in fatigue, we hoped to resolve some of the current discrepancies in this area. METHODSWe used a cross-sectional design to make comparisons across age and gender. Measures of MVC force were made before, every 2 min during, and 0, 2, 5, and 10 min after the exercise protocol. Central activation was measured before and immediately after exercise. Measures of peripheral activation and contractile function were made before and 0, 2, 5, and 10 min after exercise. The metabolic measures were made before and continuously throughout exercise. Forty-one healthy, nonsmoking men and women aged 25–45 (10 men, 10 women) or 65–85 yr (11 men, 10 women) were recruited from the community. Individuals with chronic disease or those taking medications that might affect muscle function were excluded. All subjects were relatively sedentary in that they participated in two or less periods of continuous (>20 min) activity per week. To minimize the chances of including individuals with latent peripheral vascular disease, any subject with resting supine ankle and/or brachial systolic blood pressure of <1.0 was eliminated from the study. All subjects provided written, informed consent, as was approved by the Committee on Human Research at the University of California, San Francisco. All studies were conducted in San Francisco. To establish that our study groups were similarly active, physical activity was quantified by using a three-dimensional accelerometer (Tritrac R3D, Professional Products, Madison, WI). Each subject wore an accelerometer at the waist during waking hours for a period of 1 wk, as previously described (27). All subjects maintained a brief written log of activities that was examined and discussed on returning the monitor to the laboratory. The vector magnitude (arbitrary units) for all three dimensions was averaged over the 7 days and divided by 1,000 (for ease of expression) and used as the measure of physical activity. All measures of force, contractile properties, activation, and metabolism were acquired with the subject seated with one leg extended (knee fixed at ∼170° extension) into the 30-cm-bore superconducting magnet, as described in detail elsewhere (24, 25, 28). The foot was secured to a platform (ankle angle 120°) under which was mounted a nonmagnetic force transducer, which in turn was coupled to a personal computer. The stimulating electrode (pair of 10-mm nonmagnetic disks, Grass Instruments, West Warwick, RI; mounted on plastic) was placed over the peroneal nerve, ∼1 cm distal to the fibular head. A copper ground plate was placed distally between the stimulating electrode and the EMG recording electrodes (see below). For each subject, supramaximal intensity [15% greater than that necessary to elicit a maximal compound muscle action potential (CMAP)] was determined and then used for all subsequent stimuli. Twitch (0.1-ms pulse) and tetanic (50-Hz, 500-ms train) forces were obtained at sampling rates of 2,500 and 500 Hz, respectively. CMAP was recorded at 2,500 Hz with nonmagnetic surface electrodes (10-mm disks) taped over the belly and distal tendon of the tibialis anterior muscle, as previously described and used (24, 25, 29, 39). Force and EMG data were acquired and transferred to spreadsheet for analysis. Phosphorus magnetic resonance spectroscopy was used to acquire information regarding intramuscular energy metabolism, as performed previously (24). Data were collected in the 1.9-T superconducting magnet by using a 3 × 5-cm elliptical surface coil taped over the belly of the tibialis anterior muscle, just proximal to the EMG recording electrode. After collection, data were transferred to personal computer for analysis, described in Force and contractile measurements. Before the fatigue protocol, the following measures were made, in order: CMAP and accompanying twitch, MVC force, central activation ratio (CAR), potentiated CMAP + twitch, and stimulated tetanus. Each measure was separated by 1 min of rest. The CMAP + twitch measure and the MVC measure were each repeated three times at 1-min intervals. Peak MVC force was determined from the best of the three 3- to 4-s trials. To ensure optimal performance by the subject, any MVC trial that resulted in a force of <90% of the other trials was repeated. Twitches were acquired before and 0, 2, 5, and 10 min after exercise, and tetani were acquired before and 0, 5, and 10 min after exercise. The postexercise and recovery twitch forces were scaled to the potentiated twitch, which was obtained immediately after the third baseline MVC. Because contractile failure is often a source of fatigue (17) and because it has been suggested that excitation-contraction coupling may be impaired with age (13), we measured several indexes of contractile function by using stimulated twitch and tetanic contractions of the dorsiflexor muscles before, immediately after, and during recovery from exercise, as performed previously (27, 39). Contractile function may be quantified by the peak forces elicited during stimulated contractions as well as by the speeds of contraction and relaxation. Together, these provide indirect information related to changes in the periphery before (e.g., due to differences in fiber type) and during fatiguing exercise, in particular excitation-contraction coupling and calcium resequestration (5, 16, 17, 47). To fully represent the contraction and relaxation characteristics of the muscle, both the maximum rates of force development (dF/dt) and relaxation (−dF/dt), and the more global measures of twitch contraction time and tetanic half-relaxation time, were determined. The dF/dt and −dF/dt were calculated for both the twitch and tetanic contractions. Because the rate of force development is faster with higher force production (38), dF was expressed as a percentage of the peak force achieved during each contraction. Thus these rates are expressed as percent peak force per millisecond. This approach allows comparisons of rates across individuals with differing torque-producing capacities. For the tetanus, the half time of force relaxation was calculated as the time (in ms) from the last CMAP in the train to the point at which force fell to 50%. For the twitch, the time to peak force (in ms) was calculated, with the use of the differential of the force trace, from the time of force onset to the time at which dF/dt = 0. These calculations were performed in an Excel spreadsheet (Microsoft, Redmond, WA), as previously reported (24, 27). Central activation, measured here as that portion of neuromuscular activation located proximal to the stimulating electrode, was quantified by using the CAR [CAR = MVC/(MVC + superimposed stimulated force); Ref. 25]. The stimulated force was elicited with a supramaximal train (50 Hz, 500 ms) that was superimposed on the voluntary contraction when force had reached maximal and plateaued. CAR was determined before and at the end of exercise. Peripheral activation was assessed from CMAP, which reflects the excitability of the neuromuscular junction and muscle membrane (1). CMAP peak-to-peak amplitude (in mV) and duration of the negative peak (in ms) were determined. After acquisition of the baseline force and contractile and activation measures, the subject sat quietly while the magnet was shimmed and phosphorus data were acquired from the resting muscle. The repetition time for all acquisitions was 1.25 s. The data were averaged over 1 min for the rest spectrum (48 acquisitions) and every 30 s (24 acquisitions) during exercise. To ensure accurate quantification of overlapping peaks, all peaks in the spectra [bone broad component, phosphomonoesters, Pi, phosphodiesters, phosphocreatine (PCr), 3 peaks of adenosine triphosphate] were fit by using NMR1 software (New Methods Research, White Plains, NY). The data were then imported into a spreadsheet, corrected for partial saturation, and used to calculate Pi/PCr, Pi (in mM), diprotonated Pi (H2PO4−, in mM), and pH, according to standard equations (46). Corrections for partial saturation of PCr and Pi were made by using values obtained experimentally and from the literature (4), respectively. Metabolic data were acquired before and continuously during exercise. After acquisition of baseline measures of force, contractile properties, activation, and metabolism, each subject practiced several contractions at 10% MVC to become familiar with the target intensity and duty cycle of the exercise protocol. The subject then performed 16 min of isometric contractions (4-s contraction, 6-s relaxation). Exercise began at 10% of MVC and was incremented by 10% every 2 min. To determine the time course of fatigue, an MVC was performed at the beginning of each 2-min stage. Immediately postexercise, MVC with superimposed train (CAR), tetanic force, and CMAP with accompanying twitch force were measured. The primary fatigue variable was the fall of MVC at end of exercise, i.e., postexercise MVC/preexercise MVC. This protocol typically causes MVC to fall to ∼70–75% of initial levels (26, 29). Two-factor (age, gender) ANOVAs were used to examine differences between groups in preexercise force (MVC, tetanic force, twitch force), contractile properties (maximum rates of twitch and tetanic force development and relaxation, twitch contraction time, and tetanic half-relaxation time), peripheral activation (CMAP amplitude and duration), and metabolic variables at rest (Pi/PCr, pH). Due to the ceiling effect of the CAR measure, Mann-Whitney and Wilcoxon nonparametric procedures were used to detect differences across groups in pre- and postexercise CAR values. Two-factor (age, gender) repeated-measures (pre-, postexercise) ANOVAs were used to compare changes in force, contractile properties, and activation before vs. immediately after exercise. Changes in metabolites throughout exercise and the recoveries of MVC, tetanic force, twitch force, CMAP, and contractile properties were also compared across groups by using two-factor, repeated-measures ANOVA. To investigate the role of muscle mass in fatigue, the association between preexercise MVC (a surrogate for mass, given complete activation and a consistent ankle angle) and fatigue was determined by using univariate linear regression analysis for all subjects combined. Likewise, to determine whether initial strength (mass) was related to the metabolic response to exercise, the relationship between preexercise MVC and end-exercise H2PO4−was also determined by univariate linear regression analysis. To explore the relationship between the accumulation of metabolites typically associated with fatigue (11, 40, 48) and the development of fatigue (i.e., time course of the fall of MVC), linear regression analyses of the relationships between the change in MVC during exercise and Pi, pH, and H2PO4− were performed for each subject group. For these three analyses, the MVCs obtained at the end of each 2-min stage were plotted against the metabolite value obtained during the final 30 s of each stage. For all analyses, significance was established when P< 0.05. Descriptive data in Table 1 are presented as means ± SD; all other data are presented as means ± SE. There were no age-by-gender interactions for any of the measured variables, and these P values are not presented.
RESULTSDescriptive data are provided in Table 1. To aid interpretation, data are grouped by age and gender throughout the paper. The men were taller and heavier than the women, with no age effect. Seven of the older women were on estrogen replacement therapy. There were no age or gender main effects (P > 0.05) for physical activity level. Group sizes for the activity measurement were 10 young women, 8 older women, 9 young men, and 11 older men. For each category of variables in the following sections, the data from measurements taken before the fatigue protocol are presented first, followed by comparisons of the exercise and recovery data. Preexercise values for MVC, tetanic force, and twitch force are provided in Table 2. Overall, men were stronger than women, with no significant effect of age.
Figure 1 shows the changes in target force and MVC for each group during exercise. All groups performed the progressive exercise protocol similarly, up to ∼60% MVC (Fig.1A). Thereafter, it became increasingly difficult to achieve target force, particularly for the young. As shown in Fig.1B, there was an effect of age on fatigue, with the older subjects showing less fatigue (postexercise MVC/preexercise MVC) at the end of exercise compared with the young subjects (P < 0.01). There was no effect of gender on fatigue (P = 0.24). Fatigue was associated with preexercise strength (r = 0.49, n = 41, P < 0.01), which suggests that ∼25% of the variability in fatigue was related to muscle strength. Fig. 1.Target force (A) and maximal voluntary contraction (MVC; B) during 16 min of intermittent isometric exercise in young and older men and women (means ± SE). OM, older men; OW, older women; YM, young men; YW, young women. Note that all groups performed similarly through the first 10–12 min of exercise (A), beyond which time fatigue began to interfere with their ability to achieve target force. MVCs obtained at the end of every 2-min stage (B) indicated that young subjects fatigued more than older subjects (P < 0.01), with no effect of gender. MVC and tetanic and twitch force data at end of exercise are presented in Table 3. As discussed above, there was a significant fall in MVC during exercise. At the end of exercise, tetanic force had also fallen in all groups (P < 0.01), with no effect of age or gender on this change. At the end of exercise, twitch force had fallen relatively more in men compared with women (P = 0.04), with no effect of age. During the recovery period, there were no differences across groups in the recovery of MVC, tetanic force, or twitch force.
Before the fatigue protocol, older subjects showed the expected age-related slowing of contractile properties in response to twitch and tetanic stimuli (Table 2). Twitch contraction time was longer, and the maximum rates of twitch force development and relaxation were lower, in older compared with young subjects (P < 0.01, all). Likewise, the maximum rates of tetanic force development and relaxation were lower (P < 0.01, both), and the half-time of force relaxation tended to be longer (P = 0.07) in older compared with young subjects. There was no effect of gender on these variables. Before the exercise protocol, the potentiation of twitch force after baseline MVCs was greater in men (161 ± 12%) than in women (133 ± 7%, P = 0.02), with no effect of age (P = 0.34). The ratio of potentiated twitch to tetanic force was higher in older (0.17 ± 0.01) compared with young subjects (0.13 ± 0.02, P < 0.01). There was no effect of gender on this variable. Exercise had no effect on twitch contraction time or the maximum rates of twitch force development and relaxation in any group. However, there was an increase in the maximum rate of tetanic force production after exercise (P < 0.01), with a significant gender effect indicating that women had a greater increase in the speed of tetanic force production compared with men (P < 0.01; Fig.2). In contrast, there was a significant slowing of both the maximum rate and the half-time of tetanic force relaxation in response to exercise (P < 0.01, all; Fig. 2). The recovery of all twitch and tetanic contractile variables was similar across groups. Fig. 2.Contractile properties (means ± SE) of the electrically evoked tetanus (50 Hz, 500 ms) before and at the end of 16 min of exercise. Before exercise, the maximum (max) rates of tetanic force development and relaxation (A) were significantly slower in older compared with young subjects (P < 0.01), whereas the half-time (t1/2) of force relaxation (B) tended to be longer in the older group (P = 0.07). At the end of exercise, the maximum rate of tetanic force development was increased (A, top), and women showed a greater increase than men (P < 0.01). Both the maximum rate of force relaxation (A,bottom) and the t1/2 of force relaxation (B) slowed as a result of exercise, with no effect of age or gender. □, Older men; ○, older women; ■, young men; ●, young women. There was a significant gender effect (P < 0.01) on the change in the twitch-to-tetanic force ratio from pre- to postexercise, such that this ratio increased by 0.05 ± 0.06 in women, whereas in men it decreased by 0.01 ± 0.06. There was no effect of age on the change in the twitch-to-tetanic force ratio. As shown by the CAR data in Table 2, there was no difference between groups in the ability to maximally activate the dorsiflexor muscles before the exercise protocol. Likewise, CAR was unchanged in all groups at the end of the exercise protocol (Table 3). The amplitude of the unpotentiated CMAP was significantly lower in older compared with young subjects (P < 0.01; Table2). Before the start of the exercise protocol, CMAP amplitude increased in all groups after potentiation by the three baseline MVCs (P < 0.01). The duration of the unpotentiated CMAP was similar in all groups (Table 2), and there was a similar increase of CMAP duration in all groups after the MVCs (P < 0.01). All postexercise CMAP values were compared with the potentiated CMAP. At the end of exercise, there was no further change in CMAP amplitude from the potentiated level in any group. However, there was a significant shortening of CMAP duration (P < 0.01) immediately after exercise, which was similar in all groups. There were no gender main effects for CMAP amplitude or duration, either before or after exercise. During the recovery period, CMAP amplitude increased in young relative to older adults (P = 0.04). At rest, Pi/PCr was higher in young compared with older (P < 0.01) subjects, with no effect of gender. Resting pH was similar in all groups (Table 2). There was no age effect on the change in Pi/PCr during exercise (P= 0.76; Fig. 3), but there was a significant gender effect in that women had a smaller increase in Pi/PCr compared with men (P < 0.01; Fig. 3). As shown in Fig.3, the rate of change in Pi/PCr increased in the young and older men after ∼8 min of exercise, indicating the end of the steady-state, oxidative phase of this exercise protocol (8, 26). In contrast, women showed little change in the rate of increase of Pi/PCr during exercise, suggesting that oxidative metabolism was able to keep pace with energy needs in the women throughout exercise. Fig. 3.Intracellular Pi-to-phosphocreatine ratio (Pi/PCr) during exercise (means ± SE). There was a similar, gradual increase in Pi/PCr during the first 6–8 min of exercise, which indicates that all groups had a similar capacity for oxidative metabolism during this steady-state portion of the protocol. Beyond ∼8 min, there was a loss of the steady state, particularly in the young and older men. There was no effect of age on the change in Pi/PCr during exercise. At the end of exercise, there was a significant gender effect in that men had a higher Pi/PCr compared with women (P < 0.01). A: older men. B: older women.C: young men. D: young women. Elevations in the concentration of Pi, H+, and H2PO4− have each been implicated as sources of fatigue during exercise (11, 40, 48). In the present study, there were both age and gender effects for the changes in these metabolites, as shown in Fig. 4. During exercise, Pi increased more in young compared with older subjects and in men compared with women (P < 0.01, both). Intracellular pH decreased more in older compared with young subjects and in men compared with women (P < 0.01, both). Finally, H2PO4− increased more in young compared with older subjects and in men compared with women (P < 0.01, both). The concentration of H2PO4− during the final minute of exercise was linearly related to preexercise MVC or strength (r = 0.53, P < 0.001). Fig. 4.Intracellular Pi (A), pH (B), and H2PO4− (C) during exercise (means ± SE). Beyond ∼8 min of exercise, the accumulation of Pi, pH, and H2PO4− begins to occur at a more rapid rate, as oxidative metabolism is no longer able to keep pace with increasing energy demands, and anaerobic processes begin to contribute relatively more to the energy supply. Intracellular pH decreased, and Pi and H2PO4− increased more in young compared with older subjects and in men compared with women (P < 0.01, all), suggesting a greater reliance on glycolytic metabolism in young subjects and men, respectively. Although only age affected fatigue, there were both age and gender effects on the metabolic response to exercise. To investigate whether there might be a different role for metabolic inhibition of contraction in the fatigue across genders, we examined the relationship between H2PO4−, a putative fatigue agent (48), and the development of fatigue during exercise for each group, as shown in Fig. 5. Although the range of this relationship was smaller in the women due to lower H2PO4− production, the slope of their relationship between fatigue and H2PO4−appeared steeper than that of the men. The relationships between fatigue and both Pi and pH were qualitatively similar to those for H2PO4− (data not shown). Fig. 5.Relationships between fatigue and H2PO4− during exercise (means ± SE) in older men (A), older women (B), young men (C), and young women (D). The fall of MVC during exercise is related to the increase in H2PO4− in each group, regardless of the magnitude of fatigue or degree of metabolite accumulation. Although men had a nearly twofold greater increase in H2PO4− during exercise, they developed no greater fatigue than women. As a result, the slope of the relationship between fatigue and H2PO4− appears to be steeper for women. See text for details. DISCUSSIONThe primary results of this study were that, during incremental isometric exercise, 1) older subjects exhibited less fatigue compared with young subjects, 2) there was no effect of gender on fatigue, and 3) the metabolic response to exercise varied with age and gender in a manner that suggests a greater reliance on nonoxidative sources of ATP in young compared with older subjects and in men compared with women. Of note is the fact that these results were obtained in groups with similar habitual physical activity levels, which minimizes the effect of activity on our measures. There were three main findings related to contractile function. First, contraction and relaxation rates were slowed in the unfatigued muscles of the older compared with young subjects, as expected. Second, there were no age-based differences in the degree of twitch potentiation before exercise or in the change in twitch-to-tetanus ratio, the increase in the rate of tetanic force development, and the slowing of force relaxation of the tetanus after exercise. Third, there was earlier potentiation of twitch force in men compared with women, with no effect of age. Overall, alterations in contractile function did not explain the age-related difference in fatigue that we observed. Our finding of an age-related slowing of electrically evoked twitch and tetanic contractile properties in the unfatigued muscle is similar to the results from previous studies of the dorsiflexor (10,39) and other (15) muscles. This slowing is consistent with the age-related shift toward a higher percentage of type I fiber content reported by others (22, 34). Exercise caused a similar slowing of tetanic force relaxation in all groups, which is often a consequence of fatigue (5). Slowed force relaxation is likely due to the slowing of calcium resequestration by the sarcoplasmic reticulum in fatigued muscle (47). The fact that the older subjects showed no excessive slowing in either force development or relaxation during fatigue suggests that neither excitation-contraction coupling nor calcium kinetics were altered in this group compared with the young. The lack of an age-related effect on excitation-contraction coupling is further supported by the lack of an effect of age on the change in the twitch-to-tetanus ratio after exercise and by the similar recoveries of all force and contractile variables in all groups (17). Before the exercise protocol, the potentiation of twitch force after three MVCs was greater in men compared with women, with no effect of age. This result suggests that there was a greater increase in actin-myosin Ca2+ sensitivity in response to three 3- to 4-s MVCs in the men and no impact of age on this system (41). At the end of exercise, tetanic force fell similarly in all groups, whereas twitch force fell more in men compared with women (Table 3). As a result, the twitch-to-tetanic force ratio increased in women but decreased in men during fatigue. Furthermore, although the maximum rate of tetanic force development increased in all groups in response to the exercise protocol, this increase was significantly higher in women compared with men (Table 3). Taken together, these observations suggest that the majority of force potentiation occurred very rapidly (i.e., after baseline MVCs) in men, whereas potentiation reached its maximum later, during the exercise protocol, in women. These results indicate a gender-based difference in the magnitude and timing of force potentiation that may reflect differences between men and women in actin-myosin Ca2+sensitivity (41). Neural activation may be separated broadly into central and peripheral components. In the present study, these are delineated by the location of the stimulating electrode, with all elements proximal to the electrode representing central activation and all elements distal to the electrode comprising peripheral activation. The main findings related to activation in this study were that neither central nor peripheral activation failure contributed to fatigue in any group in response to this incremental isometric protocol. In contrast to some reports (2, 44), we observed no age-related impairment of central activation, either before (CAR, Table2) or at the end (Table 3) of fatiguing exercise. Likewise, there was no difference between men and women in the ability to fully activate the dorsiflexor muscles. These results are consistent with our previous work in this muscle group (27), as well as with the work of others (20). It is likely that the moderate degree of fatigue observed with this protocol precluded the development of central fatigue, as central activation failure is often associated with more severely fatiguing exercise (e.g., Ref. 24). Although there was no failure of central activation, per se, it is possible that lower motor unit discharge rates may have played a role in the greater fatigue resistance of the older subjects. It has been reported that discharge rates are reduced in older compared with young adults during both submaximal and MVCs (10, 23). During a progressive exercise protocol such as that used here, lower discharge rates in older adults could serve to 1) acutely limit the extent to which Pi/PCr increases and pH decreases during exercise as the ability to drive the muscle at higher frequencies is limited and 2) shift the muscle toward a more oxidative profile (i.e., the prolonged exposure of all fibers to lower discharge rates would result in adaptation toward a slower, more oxidative muscle). Precedence for the first possibility exists from a study of muscle fatigue in people with multiple sclerosis (MS). Fatigue in the MS and control groups was similar during the same incremental isometric exercise protocol reported here (29). However, the metabolic response to exercise (i.e., Pi/PCr, pH) was markedly smaller in MS (29). The metabolic difference could not be explained by differences in motor unit recruitment in the MS group. Instead, this difference was likely because of an inability of MS patients to generate high discharge rates during exercise, as shown previously by Rice et al. (42). The second possibility, related to a morphological adaptation to chronically reduced activation rates, is consistent with the age-related increase in type I fiber area reported in the tibialis anterior muscle, from 76% in young adults to 84% in older adults (22). The observation of slower contractile properties in the unfatigued muscle of the older adults in this study is compatible with such a fiber-type shift. This shift might arise both from the loss of type II fibers due to the denervation-reinnervation process that occurs with aging (6) as well as from the lower discharge rates experienced by the muscle of older adults. Regardless of the mechanism, the shift toward a slower muscle is consistent with the greater fatigue resistance observed in the older group in this study. Although CMAP amplitude was smaller in older compared with young subjects before voluntary muscle activations (Table 2), the degree of potentiation of the CMAP in response to baseline MVC contractions was similar in all groups. This result suggests that, before fatigue, the enhancement of sarcolemmal Na+-K+ pump activity in response to contraction (19) is unaffected by age or gender. At the end of exercise, there was no change from baseline in CMAP amplitude in any group (Table 3), which suggests that there was no decrease in peripheral excitability during fatigue. The duration of the CMAP was shorter in all groups after exercise, suggesting that conduction velocity across the neuromuscular junction or along the muscle membrane had increased during this submaximal protocol. An increase in conduction velocity may have occurred due to a “warm-up” effect in the muscle. Overall, peripheral activation failure did not appear to play a role in the development of fatigue in any group during this protocol. Furthermore, there was no evidence to suggest that differences in peripheral excitability across age had an impact on the age-related difference observed in fatiguability in this study. In contrast to the activation data, the metabolic data showed significant age- and gender-related differences in response to exercise. The exercise protocol used in this study begins with a low-intensity, metabolically steady-state portion and ends with relatively high-intensity contractions that produce greater changes in energy metabolites and pH (8, 26). During steady-state exercise, Pi/PCr reflects the ability of the muscle to respond oxidatively to the need for ATP (8, 26). In the present study, steady state was maintained similarly in all groups through the first half of the exercise protocol, which suggests a similar potential for oxidative metabolism in all groups. Beyond ∼8 min, which corresponded to an intensity of ≥50% MVC, Pi/PCr increased at a more rapid rate in men than in women (Fig. 3). This observation suggests that women continued to keep pace with the energy demand via oxidative phosphorylation throughout the exercise protocol, whereas men were less able to do so as the exercise progressed. There was no effect of age on the change in Pi/PCr during exercise, suggesting that the gender-based difference in metabolic pathway “preference” persists with aging. Interestingly, this difference is consistent with reports indicating a relatively greater reliance on carbohydrate as a fuel in men compared with women (45). In contrast to the Pi/PCr data, the changes in pH, Pi, and H2PO4− showed both age and gender effects (Fig. 4). Each of these metabolites has, at various times, been implicated in the fatigue process (11, 24, 40,48). At a pH of 6.75, Pi exists in equal proportions of its mono- and diprotonated species. As pH drops below 6.75, the diprotonated species predominates. Thus the concentration of diprotonated Pi (H2PO4−) reflects both the increase in Pi and the decrease in pH. For all groups, there was a very strong relationship between the time course of fatigue and the accumulation of H2PO4− (Fig. 5). Interestingly, however, the slope of this relationship appeared to be steeper in the women, suggesting a greater “sensitivity” of force production to the accumulation in H2PO4− in women compared with men. This observation may, in part, explain the lack of a gender effect on fatigue despite the greater metabolic changes in men compared with women. The mechanism of this greater sensitivity in women is not clear but may be related to a generally lower reliance on glycolytic metabolism in female muscle. That is, if female muscle is typically less likely to encounter high concentrations of H+, Pi, and H2PO4−, it may be more sensitive to these metabolites when they do accumulate. The smaller change in H+ concentration in the older group suggests that older individuals may have a smaller capacity for glycolytic metabolism compared with young subjects. This possibility finds support in the literature from Larsson et al. (32), who reported lower lactate dehydrogenase activity in the vastus lateralis muscle of older compared with young men. As noted, it has been suggested that differences in muscle mass might account for some of the differences in fatigue observed across gender or age via the impact of intramuscular pressure on muscle perfusion during the contractions. Higher absolute forces will produce higher intramuscular pressure and, therefore, relatively less perfusion. This concept is supported by the recent report from Hunter and Enoka (21) in which gender differences in elbow flexor endurance were nullified after adjustment for strength. Similarly, differences in strength and absolute target tension were likely important factors in an earlier study of fatigue in which gender-based differences in the endurance response of the elbow flexors to immobilization were reported (43). In the present study, preexercise MVC was associated with ∼24% of the fatigue that developed during exercise. This result provides some support for the possibility that the intramuscular pressure developed during each contraction may be relatively higher in the stronger subjects, thus leading to greater occlusion of blood flow to the working muscle during each contraction. A difference in blood flow would be particularly evident at the higher contraction intensities in our protocol; interestingly, it is at these intensities that the metabolic response to exercise diverges in young compared with older subjects and in men compared with women (Figs. 3 and 4). The significant relationship between strength (preexercise MVC) and end-exercise H2PO4− concentration is also consistent with the possibility that oxygen delivery may have been relatively better in the weaker subjects, thereby allowing them to rely on oxidative pathways for longer periods during the exercise protocol. It should be noted that the measure of fatigue selected by various investigators may be complicating our understanding of the effects of age and gender on muscle function. As mentioned previously, Hunter and Enoka (21) recently observed a difference in endurance time but no difference in fatigue (i.e., fall of MVC) in the elbow flexors after a contraction sustained at 20% MVC. A similar example is provided for aging by Bilodeau et al. (3), who reported similar fatigue in young and older subjects after a sustained 35% MVC elbow flexion contraction but a longer endurance time in the older subjects. The observations of similar fatigue but different endurance times across these study groups suggest that endurance and fatigue are, to some degree, physiologically distinct tasks. Thus, in addition to the need for accounting for differences in strength across study groups, there is a need for care when interpreting and comparing protocols with different criteria for fatigue. In conclusion, the results of this study indicate that the dorsiflexor muscles of older adults have a greater resistance to fatigue during intermittent, submaximal isometric contractions compared with young adults of similar physical activity habits. There was no effect of gender on fatiguability, nor were there significant age- or gender-based differences in central or peripheral activation during exercise. Whereas the older group showed the expected slowing of muscle contractile properties before exercise, there was no evidence of an age-based difference in contractile function as a result of the fatiguing exercise. In contrast, the metabolic data suggest that both age and gender affect the response to intermittent isometric exercise. It appears that older compared with young subjects, and women compared with men, rely less on anaerobic pathways for the supply of ATP during muscular contractions. The authors thank the volunteers who participated in this study. We also thank Dr. Milton Hollenberg for medical consultations; Hung Dao, Danielle Bartholomew, and Ian Lanza for assistance with data acquisition and analysis; Drs. John Neuhaus and John Buonoccorsi for statistical advice; and Drs. Kirsten Johansen and David Russ for comments on the manuscript. FOOTNOTESREFERENCES
Page 25the anabolic and lipolytic effects of the β2-adrenergic receptor (AR) agonist clenbuterol have been widely investigated, principally at doses ranging from 1 to 5 mg/kg in a variety of sedentary laboratory and livestock animals (6, 12, 25, 29). In response to such doses, the size of the heart, skeletal muscle, bone, lung, and kidney all increase, whereas the liver (35) and adipose tissue decrease in mass (6, 30). However, not all investigations have been able to demonstrate clenbuterol's anabolic effects, particularly when low doses are used (5), which suggests a dose-response relationship between clenbuterol and muscle hypertrophy. Clenbuterol administration has been shown to be beneficial in some animal models of Duchenne muscular dystrophy (14, 15) but not in others (21). It has yet to be proven, however, that clenbuterol-induced muscle hypertrophy is of any functional significance in normal populations, whether the dose administered is low (5) or high (20). More recent animal studies using doses (1–5 mg/kg) known to promote anabolism and investigating the combined effects of clenbuterol administration and exercise have shown a decrease in exercise performance (17,26) and a high incidence of sudden cardiac failure (11). This suggests that clenbuterol administration may be antagonistic to the muscular and/or cardiovascular adaptation to exercise, although the mechanism by which this occurs is not yet understood. It is conceivable that, like the less selective β-AR agonist isoproterenol (3, 27), clenbuterol may induce cell death and necrosis in the heart. Indeed, clenbuterol has previously been shown to induce general histological damage in the soleus muscle of the rat in response to a dose of 2 mg/kg administered via drinking water (36). We have, therefore, tested the hypothesis that clenbuterol administration may induce myocyte damage in the heart as well as the soleus. This may then provide a possible mechanism for the adverse effects of clenbuterol on the adaptation to exercise (17,26) and for the increased collagen deposition found in the heart after its long-term administration (11). Not surprisingly, the aforementioned anabolic and lipolytic (i.e., repartitioning) effects of clenbuterol have attracted the attention of many athletes, despite clenbuterol being banned by the World Anti-Doping Agency. Body builders in particular take high doses of this β2-AR agonist (28). The protocol for determining an individual's optimal dose is crude and involves ever-increasing daily doses until the side effects can no longer be tolerated (10). The doses employed during clenbuterol abuse, therefore, vary widely, with men being better able than women to tolerate the side effects, which include tachycardia, hypokalemia, arrhythmia, muscle cramps, and muscle tremors (19). An average daily dose for males can be eight tablets or ∼2 μg clenbuterol/kg body wt. In addition, because of its lack of androgenic side effects, clenbuterol is also popular among sedentary as well as athletic women for use as a repartitioning agent. Scientific investigations into the effects of clenbuterol in humans are far less numerous than those pertaining to livestock or laboratory animals. Nonetheless, clenbuterol has been shown to have some therapeutic potential in speeding up the rehabilitation of postoperative muscle wasting in humans (23) and has been proposed for the pharmacological amelioration of cachexia in chronic diseases such as cancer (4) and Duchenne muscular dystrophy (15). The present finding that even a single administration of clenbuterol induces necrosis in cardiac and skeletal myocytes demonstrates that, before clenbuterol can be properly considered for clinical use, its potential myotoxic effects need to be more closely investigated. The pharmacokinetics of chronic clenbuterol administration are likely to be complex, with the possibility of tachyphlaxis or the accumulation of unmetabolized clenbuterol in the plasma. To avoid these complications, we have used only a single administration (enteral or parenteral) of clenbuterol and investigated the incidence of myocyte-specific necrosis in response to this in the heart and soleus muscle of the rat. Clenbuterol is lipophilic and is known to have direct intracellular actions (1). Although the anabolic effects of clenbuterol administration have previously been shown to be β2-AR mediated (6), the same cannot be assumed for clenbuterol's myotoxic effects. To investigate whether clenbuterol-induced necrosis was receptor mediated or a direct action of clenbuterol, selective β-AR antagonists have been used. In addition, the norepinephrine (NE) depleting agent reserpine was also used to investigate the possible neuromodulation of the sympathetic nervous system (SNS) by clenbuterol. The present study presents data on clenbuterol-induced necrosis in the heart and soleus muscle of the rat at doses commonly employed to demonstrate its anabolic effects. The dose dependency and time course of cardiac and skeletal myocyte necrosis has been investigated in detail. To achieve meaningful quantification in the heart, a study of the topographical distribution of this cellular damage was crucial. Furthermore, by scaling the data provided in the current rat model, it was found that doses commonly abused by athletes fall within the range capable of inducing myocyte death and loss. METHODSAll experimental procedures were carried out under the British Home Office Animal (Scientific Procedures) Act 1986. Male Wistar rats weighing 298 ± 22 g were bred in-house in a conventional colony, housed in controlled conditions of 25°C, 50% relative humidity, and a 12-h light (0600–1800) and 12-h dark cycle, with water and food (containing 18.5% protein) available ad libitum. After the appropriate experimental procedures, rats were killed by cervical dislocation, and the heart and soleus muscles were quickly isolated. The atria were dissected from the ventricles and mounted with a piece of liver as a support. The remaining great vessels were removed directly superior to the coronary sulcus, and the ventricles were mounted apex uppermost. Soleus muscles were mounted in transverse section and supported with liver. Tissues were immediately snap frozen in super-cooled isopentane and stored at −80°C before cryosectioning (5 μm) and storage at −20°C. Necrosis was detected in skeletal and cardiac myocytes by using an anti-myosin monoclonal antibody (Ab) in vivo (3, 27). This Ab, administered before clenbuterol challenge, can only permeate the disrupted sarcolemmal membranes of irreversibly damaged necrotic myocytes. All animals (experimental and control) received an intraperitoneal injection of Ab (0.9 mg/kg) 1 h before either administration of clenbuterol (experimental groups) or saline vehicle (control group). The anti-myosin Ab was then immunohistochemically detected on the cryosections of the harvested tissues by using a horseradish peroxidase-conjugated secondary Ab and visualized with 3,3′-diaminobenzidine. Sections were then counterstained with hematoxylin and permanently mounted before being examined by using light microscopy (×100 magnification). To quantify the necrosis in the soleus, randomized fields of view across each traverse section taken at four points about the midbelly of the muscle were investigated. Both necrotic and viable myocytes were counted (>500), and the number of necrotic fibers was expressed as a percentage of the total. In the heart, randomized fields of view within discrete areas of the ventricles (i.e., subendocardium, subepicardium, etc.) were digitized, and image analysis was used to measure the percent area of positive staining (cardiomyocyte damage) within each region. To precisely control the dose of clenbuterol received by each animal and the time of its administration, a single dose of clenbuterol (ICN Biomedical) was subcutaneously administered in a saline vehicle. The only exception to this was a comparative investigation into the effect of the route of administration, in which a precisely controlled dose of clenbuterol was administered either by gavage or by subcutaneous injection. Clenbuterol was administered over the range 1 μg/kg to 5 mg/kg (n = 5 in each group). Animals were killed 18 h after clenbuterol challenge, according to the procedure described above. The incidence of myocyte necrosis was then quantified in the heart and soleus muscle from each of these animals. The peak-damaging dose of clenbuterol (5 mg/kg) was administered to seven independent groups (n = 3 in each group) of rats. Each group was then killed at selected time points from 0 (saline vehicle controls) to 24 h after the administration of clenbuterol, and the incidence of myocyte necrosis was then quantified in the heart and soleus muscles. Throughout the investigations into the dose dependency and time course of clenbuterol-induced necrosis, the myocyte necrosis in the ventricles was consistently sampled at a point 2 mm from the apex (based on preliminary studies). To further strengthen the model and characterize the clenbuterol-induced necrosis, its distribution was investigated along the longitudinal and transverse planes of the ventricles. Rats (n = 3) were administered a peak damaging dose of 5 mg/kg, and the hearts were harvested 12 h (peak time) later. Each heart was sampled at 400-μm intervals along the longitudinal axis from apex to base, and the incidence of cardiomyocyte necrosis was quantified in the subendocardium and subepicardium of the left ventricles. In the atria and papillary muscles of the left and right ventricles of the same hearts, random samples were taken, and the incidence of clenbuterol-induced necrosis was quantified. Rats were randomly assigned into four independent groups: parenteral clenbuterol, enteral clenbuterol, parenteral control, and enteral control (n = 5–6 rats in each group). The parenteral clenbuterol group received a single subcutaneous injection of 5 mg clenbuterol/kg body wt, whereas the parenteral control group received an equivolume subcutaneous administration of the saline vehicle only. Rats in the enteral clenbuterol and enteral control groups received 5 mg clenbuterol/kg body wt in 1 ml of saline or the saline vehicle, respectively, and both were administered by gavage. All animals were killed 12 h after clenbuterol (experimental groups) or saline (control groups) administration, and the incidence of myocyte-specific necrosis was investigated in the heart and soleus muscles. Rats were randomly assigned to the following groups: negative or positive controls, and those undergoing β2-AR blockade, β1-AR blockade, or NE depletion (n = 4–10 rats in each group). The negative control group received a subcutaneous injection of the saline vehicle only; the positive control group received a subcutaneous injection of 5 mg clenbuterol/kg body wt. The β2-AR blocked group received 10 mg ICI 118,551/kg body wt, the β1-AR blocked group received 10 mg bisoprolol/kg body wt, and each β-AR antagonist was administered subcutaneously 1 h before a peak damaging dose of 5 mg clenbuterol/kg body wt. The NE-depleted group received an intraperitoneal injection of 2 mg reserpine/kg body wt 24 h (8, 32) before subcutaneous administration of a peak damaging dose of 5 mg clenbuterol/kg body wt. All animals were killed 12 h (peak time) after administration of clenbuterol, and the heart and soleus muscles were harvested and analyzed. All data are presented as means ± SE. Experiments were analyzed by using either one-way analysis of variance with multiple post hoc analyses or Student's independent t-test. Pvalues of <0.05 were taken to indicate statistical significance. RESULTSNo necrotic damage was evident in either cardiac or skeletal muscle from control animals receiving the myosin Ab and saline vehicle only (Fig. 1, A and C). Clearly discernible myocyte necrosis was found in both the heart (Fig.1B) and soleus (Fig. 1D) after clenbuterol administration. The onset of this myocyte necrosis in the left ventricular subendocardium of clenbuterol-treated rats occurred at a dose of 0.1 mg clenbuterol/kg body wt (Fig.2A). In the soleus from the same animals, a dose of only 0.01 mg clenbuterol/kg body wt was sufficient to initiate similar damage (Fig. 2B). Necrosis in the left ventricular subendocardium seemed positively correlated to the dose administered, whereas the degree of necrosis in the soleus appeared maximal and was maintained throughout the dose range of 0.01 to 5.0 mg clenbuterol/kg body wt (Fig. 2) Fig. 1.Immunohistochemical identification of myocyte necrosis in the heart and soleus muscle. Control sections from animals receiving only the myosin antibody and saline vehicle in left ventricular (LV) subendocardium (×400 magnification; a) and soleus muscle (×200 magnification; c) cross sections are shown. No damage was found in these tissues. Typical examples of myocyte necrosis 12 h after being exposed to a single in vivo administration of 5 mg of clenbuterol/kg body wt in LV subendocardium (b) and soleus (d) cross sections are also shown. Diaminobenzidine brown stain represents secondary (in vitro) immunoperoxidase detection of the primary myosin antibody administered in vivo and, therefore, represents necrosis. Tissue preparation: 5-μm cryosections with hematoxylin counterstain. Fig. 2.Dose dependency of myocyte necrosis in response to a single parenteral administration of clenbuterol. Myocyte-specific necrosis was quantified as percentage area in the LV subendocardium (A) and percent number of fibers in the soleus (B) 18 h after subcutaneous administration of clenbuterol. No necrosis was found in the control tissues receiving saline vehicle only. Data are presented as means ± SE ofn = 5 samples. * P < 0.05. The time course of necrosis was investigated in detail over 0–24 h in response to a single injection of 5 mg clenbuterol/kg body wt (Fig.3). No necrosis was found in the zero time controls, which received the myosin Ab and saline vehicle only. In both the subendocardium (Fig. 3A) and soleus (Fig.3B), necrosis was first detected at 4 h, with peak necrosis reached at 12 h in the subendocardium and 15 h in the soleus. Fig. 3.Time course of myocyte-specific necrosis in response to a single administration of clenbuterol. Incidence of myocyte necrosis in the LV subendocardium (A) and soleus muscle (B) at specific time points after a single subcutaneous administration of 5 mg clenbuterol/kg body wt is shown. No necrosis was found in the control group receiving the myosin antibody and saline vehicle only. Data are presented as means ± SE of n = 3 samples. Necrosis in response to a single dose of clenbuterol was heterogeneously distributed throughout the heart. To investigate this in detail and permit the standardization and quantification of peak necrosis, the damage was sampled along the entire length of the ventricles. Necrosis in the left ventricular subendocardium was followed and found to peak 2.4 mm from the apex, i.e., approximately one-quarter of the way along the axis from the apex to the base (Fig.4). In any given cross section at this point of maximal injury, more damage was found in the left ventricular subendocardium (0.87 ± 0.05%), with damage in the right ventricular subendocardium (0.5 ± 0.1%) being greater than the right and left subepicardia (0.2 ± 0.03%). Fig. 4.Topographical distribution of necrosis along the longitudinal axis of the left ventricle. Incidence of cardiomyocyte-specific necrosis was investigated at 400-μm intervals along the longitudinal axis of the heart. Peak cardiomyocyte necrosis (0.87 ± 0.05%) in the LV subendocardium was found 2.4 mm from the apex. Data are presented as means ± SE of n = 3 samples. Although not studied in the same degree of detail, significant (P < 0.05) necrosis was also found in the papillary muscles of the left ventricles (1.1 ± 0.2%) and left (0.17 ± 0.04%) and right (0.08 ± 0.03%) atria. The subcutaneous administration of the saline vehicle only (parenteral control group) did not induce any necrosis in the myocytes of either the heart or the soleus (Fig. 5). In contrast, administration of the saline vehicle by gavage (enteral control group) did induce some necrosis in both the heart and soleus, presumably due to the increased stress of this procedure. Both parenteral and enteral administration of 5 mg clenbuterol/kg body wt induced myocyte-specific necrosis in the heart (Fig. 5A) and soleus (Fig. 5B); in both cases (parenteral and enteral), this necrosis was significantly greater than that found in their respective control groups. In the heart, there was no significant difference in the incidence of the clenbuterol-induced necrosis by either route of administration (Fig. 5A), whereas in the soleus the enteral administration of clenbuterol induced significantly more myocyte necrosis than parenteral administration (Fig.5B). Subcutaneous injection was therefore retained as the most precise means of administering clenbuterol in the investigation into the receptor pathway mediating clenbuterol-induced necrosis. Fig. 5.Effect of route of administration on the incidence of clenbuterol-induced myocyte necrosis. Myocyte-specific necrosis in the heart (A) and soleus (B) in response to 5 mg clenbuterol/kg body wt administered by either subcutaneous injection (closed bars) or gavage (open bars) is shown; all tissues were harvested 12 h after clenbuterol administration. All data are presented as means ± SE (n = 5–6 samples). * P < 0.05, ** P < 0.01.aExperiment group significantly different from respective control; badministration by gavage significantly different from subcutaneous administration. The effects of prior β-AR antagonism are shown in Table1. No necrosis was found in the negative control group, which received the myosin Ab and saline vehicle only. In the soleus, only prior β2-AR antagonism had any significant effect, whereas in the heart both β1- and β2-AR antagonism significantly reduced the clenbuterol-induced necrosis compared with the positive control group. Reserpine was administered to investigate clenbuterol's neuromodulation of the SNS and concomitant release of NE. Depletion of the NE-releasing capacity of the SNS by reserpine significantly prevented clenbuterol-induced necrosis in the heart but not in the soleus.
DISCUSSIONThe use of a high avidity anti-myosin Ab has allowed us to investigate myocyte-specific necrosis. This, in combination with the carefully controlled in vivo protocol, ensured that this technique only identifies those myocytes with a ruptured sarcolemmal membrane, a key indicator of the transition from reversible (oncosis) to irreversible (necrosis) cell death (22, 34). Using this model, we have demonstrated that clenbuterol administration induces necrosis in the heart and soleus muscle of the rat. The finding that clenbuterol induces myocyte-specific necrosis in the heart is novel. It may be speculated that, in the absence of a functional satellite cell system, all necrosis in the heart will lead to reparative fibrosis. Our acute data may therefore provide etiological support to those of Duncan et al. (11), who showed an increase in collagen infiltration (possibly reparative fibrosis) in the heart after chronic clenbuterol administration. A possible mechanism for clenbuterol's cardiotoxicity is its adverse effect on taurine levels in the heart (9, 36). This amino acid is known to have a protective role in some tissues, particularly in the heart and lungs, with one of its possible roles being the modulation of calcium levels (16). Doheny et al. (9) showed that taurine levels in the heart are depressed in response to a single subcutaneous administration of clenbuterol. Furthermore, the dose of clenbuterol (125 μg/kg body wt) used and the time point (5 h) after clenbuterol administration at which the taurine levels in the heart become significantly depressed almost exactly match those found for the onset of necrosis in the heart in our investigation (Figs. 2A and 3A). Doheny et al. (9) did not investigate taurine levels in the soleus but only in the gastrocnemius muscle, where taurine levels increased 6 h after clenbuterol administration. In light of the present findings, it may also be of interest to investigate changes in the levels of taurine in the soleus after controlled administration of clenbuterol. This is the first time that myocyte-specific damage has been rigorously investigated and quantified in either the heart or the soleus in response to controlled doses of this β2-AR agonist. Waterfield et al. (36) have previously demonstrated generalized histological damage in the soleus in response to a dose of 2 mg clenbuterol/kg body wt given via drinking water. Although our data clearly support those of Waterfield et al., they further our knowledge by revealing that clenbuterol-induced necrosis occurs in the myocytes and may therefore directly affect muscle function. In addition, by rigorously controlling clenbuterol administrations, we have also been able to advance our knowledge of the dose dependency and time course involved in clenbuterol-induced necrosis. Throughout our initial investigations (dose dependency, time course, and topographical distribution) of clenbuterol-induced necrosis, clenbuterol was administered parenterally. Many of the previous studies investigating the effects of clenbuterol have administered clenbuterol in the drinking water. Such an approach, although easy to use and effective in inducing anabolism, has several possible shortcomings that cannot always be sufficiently well controlled. Clenbuterol is readily oxidized and needs to be protected from light. The practice of making up fresh solutions on a daily or more often a weekly basis means that the actual dose received by each animal in a communal cage cannot be measured with any precision. This problem is further complicated by the clenbuterol-induced increase in thirst, which we have found to be dependent on the dose administered (unpublished observation). To establish that ingestion of clenbuterol is also myotoxic, we administered a single dose of clenbuterol enterally, the only controllable way to accurately achieve this is by gavage administration. Although this method does not exactly match that of administration via the drinking water, it does replicate the method of administration chosen by most humans, i.e., ingestion of clenbuterol in tablet form. The data (Fig. 5) clearly demonstrate clenbuterol's myotoxic effects when administered enterally. Necrotic damage in the heart was virtually the same whether clenbuterol was administered parenterally or enterally (Fig. 5A). Interestingly, in the soleus, enteral administration of clenbuterol appeared even more damaging than parenteral administration (Fig.5B). The finding that enteral administration of the saline vehicle control also induced necrosis is presumed to be stress related, although the animals were compliant and apparently relaxed during administration. Hence, we have favored parenteral administration of clenbuterol. This route of administration provides certainty of the dose received by each animal and the time at which the dose was administered. It is quick and achievable under stress-free conditions such that no necrosis is found in the tissues from control animals. We have shown that cardiomyocyte-specific necrosis in the myocardium is not uniformly distributed. Of the possible factors mediating this heterogeneity, the principal ones may be regional differences in β-AR distribution, taurine metabolism, other metabolic requirements, or hemodynamic stresses. Unfortunately, present data relating to the distribution of β-ARs and in particular β2-ARs in the heart are scarce. Beau et al. (2) found the transmural distribution of β-ARs to be uniform in the nonfailing human heart. Although generally true for the rat heart, in some cases a greater density of β-ARs has been found in the papillary muscles and subendocardium (33). This is consistent with the pattern of damage found in the present investigation. However, in contrast, the same group of workers (33) found a homogeneous distribution of β-ARs along the longitudinal axis of the heart, which is not consistent with the pattern of damage found in the present investigation (Fig. 4). A possible explanation for this disparity is that Tofukkji et al. (33) only sampled the heart at three points (base, midventricular, and apex). If the same sampling frequency had been applied to the data in Fig. 4, then it is easy to see how important information would have been missed. These data (Fig. 4) serve to reaffirm the dangers of random sampling and the absolute requirement of in-depth topographical knowledge and standardized procedures when quantifying cell death, or anything else, in a complex organ such as the heart. In support of a possible hemodynamic contribution to the pattern of damage, we consistently found a greater degree of necrosis in the left, rather than right, side of the heart, with damage being most extensive in the left ventricular subendocardium. The inherently higher energy demands of the subendocardium may make this region of the heart more sensitive to hemodynamic perturbations. This compounded with clenbuterol-induced tachycardia and concomitant reductions in diastolic interval, and therefore blood supply, may explain the increased susceptibility of the subendocardium to clenbuterol-induced necrosis. Also of particular importance within the present investigation is the novel finding that clenbuterol administration induced a significant amount of necrosis in papillary muscles, with possible effects on valve function. It is conceivable that this damage, in combination with the irreversible loss of myocytes from the ventricular walls (and hence a reduction in the pumping capacity of the heart; Refs. 7,37), may play an etiologic role in the reduction in exercise capacity, and even cardiac failure, seen in clenbuterol-treated animals subjected to exercise (11). It was found that a significant (P < 0.01) proportion (89%) of clenbuterol-induced necrosis in the soleus was mediated through the β2-AR pathway. Therefore, only a small percentage of clenbuterol-induced necrosis can be attributed to the passage of clenbuterol through the lipid membrane. In contrast, in the heart, both prior β1- or β2-AR antagonism was highly effective in preventing clenbuterol-induced necrosis with little residual damage attributable to any direct intracellular action (Table 1). Because of clenbuterol's greater potency over many other common β2-AR selective agonists, it is difficult to extrapolate the findings of the present study to other agents. However, the finding that clenbuterol-induced myotoxicity is mediated through the β2-AR system suggests that over stimulation of this pathway per se would be toxic. The finding that clenbuterol induces damage in the heart through both the β1- and β2-AR is not in agreement with our previous findings that the less selective β1- and β2-AR agonist isoproterenol induces necrosis in the myocardium through the β1-AR pathway only (27). This, coupled with the knowledge that NE itself can be cardiotoxic (24), led us to investigate the possibility that clenbuterol may have a neuromodulatory effect over the SNS. Clenbuterol acting on the β2-AR of the sympathetic varicosities could facilitate the release of NE, which could then preferentially act on β1-ARs to induce necrosis through overstimulation of this pathway. Reserpine was administered to block the uptake of NE from the cytosol into the transport vesicles of the sympathetic varicosities. Thus, after a period of basal neuronal activity, the NE-releasing capacity of the neuron is depleted, effectively blocking this pathway. The results (Table 1) support this hypothesis, with the prior administration of reserpine significantly (P < 0.01) preventing clenbuterol-induced necrosis in the heart but not in the soleus. These data clearly show that the myotoxic effects of clenbuterol on the heart (β1-AR mediated) can be separated from its anabolic (β2-AR mediated) effects on the heart and skeletal musculature. This information may be of great value when proposing clenbuterol administration as a pharmacological aid for the amelioration of muscle wasting in chronically ill patients. The discovery that clenbuterol-induced necrosis in the heart is indirectly mediated through the SNS and β1-ARs, whereas that in the soleus is directly mediated through β2-ARs, may also account for the differences found between the two muscle types in the dose-dependency experiments (Fig. 2). It appears that the heart is not simply less sensitive to clenbuterol, but rather the indirect route of action of clenbuterol on the heart (i.e., β2-AR stimulated NE release, which then acts on cardiomyocyte β1-ARs) requires a higher dose to elicit comparable damage (Fig. 2). In the soleus, the data (Fig. 2B) suggest a threshold response with doses >10 μg/kg body wt possibly inducing receptor desensitisation and, hence, no further increase in the incidence of necrosis. The present finding that clenbuterol-induced necrosis is mediated through the β-AR pathway in vivo lends support to previous work in vitro showing that β-AR stimulation reduces the viability of cultured cardiomyocytes (24). Although the intracellular mechanisms of clenbuterol-induced myotoxicity have not been investigated here, the aforementioned work (24) in vitro elegantly demonstrated that loss of cardiomyocyte viability was preceded by an increase in intracellular cAMP followed by an increase in intracellular Ca2+. This is consistent with earlier work demonstrating that an increase in intracellular Ca2+ is a final common pathway in cell death (31), leading to the activation of proteases and phospholipases (18). The findings of this investigation show that the doses commonly employed to elicit clenbuterol's anabolic properties also induce significant myocyte necrosis in the heart and soleus muscle. It is surprising, therefore, that so little information exists on the myotoxic effects of clenbuterol. A possible explanation for this is that the anabolic effects of clenbuterol have been predominantly investigated in sedentary populations of livestock or caged laboratory animals whose daily activity levels do not make full use of their cardiac functional reserve (7, 37). It is conceivable that cumulative clenbuterol-induced cardiomyocyte necrosis would gradually reduce an animal's cardiac reserve. This could remain asymptomatic until such time when the animal is stressed or required to do work, i.e., vigorous exercise. This could explain the seminal finding of Duncan et al. (11) of a reduction in exercise capacity and a high incidence of sudden cardiac failure in swim endurance-trained rats when receiving clenbuterol. Although the anabolic or hypertrophic effects of clenbuterol have been quite widely demonstrated, studies investigating the functional significance of this anabolism in the form of increased isometric force (5, 20) or exercise capacity (17, 26) in normal populations have been ambiguous. The results from the current investigation could help to resolve this uncertainty, particularly in those studies that have used high doses of clenbuterol (17, 20). It is conceivable that the concomitant loss of myocytes incurred during such administration protocols would reduce the muscles' capacity to do work despite their increased cross-sectional area and wet weight. For example, some of the extra mass could be attributable to reparative fibrosis. Unfortunately, muscle histology was not investigated in these earlier studies (17, 20), and so any possible myocyte damage would have remained undetected. The present investigation has provided important information on the effects of acutely administered clenbuterol in the rat. In humans, a single dose of clenbuterol is generally self-administered as a 20-μg tablet. This is equivalent to 0.3 μg/kg body wt in a 70-kg male and is comparable to the dose administered in the only clenbuterol investigation using human subjects (23). To compare this with our 300-g rats, the dose needs to be scaled for differences in body weight and metabolic rate between the two species (Kleiber's Law, 0.75 exponent). The relative dose per kilogram in the rat is 60 times that of the human dose, i.e., 17.9 μg clenbuterol/kg body wt. As demonstrated in Fig. 2, this dose is sufficient to induce 3.8 ± 0.49% necrosis in the fibers of the soleus. Such a level of necrosis may appear small, but this is in response to a single administration, and this level of necrosis may underestimate the level induced by enteral administration (Fig. 5B). Individuals abusing clenbuterol often take several tablets and use the side effects of muscle tremors and tachycardia to judge their maximum dose. By using the above calculations, a daily dose of five to six tablets would be sufficient to reach the threshold (100 μg/kg body wt) for inducing damage in the heart and to induce 6.8 ± 1.9% necrosis in the soleus. An important additional factor to be considered is clenbuterol's long half-life within the body (38). Abusers of this substance often administer it by using an “on-off” cycle over several days. An accumulation of nonmetabolized clenbuterol during the on stage of the cycle may lead to chronically elevated plasma levels, which would further impact myocyte loss in both striated muscles. Although the present investigation has not investigated the compound effects of chronic clenbuterol administration, it does demonstrate that, at the very least, damage will be induced at the onset of each cycle of administration. These data support the conclusion that clenbuterol causes significant myocyte necrosis in the heart and soleus muscle of the rat at doses previously used to demonstrate its anabolic properties. These findings are in good agreement with the indirect evidence from previous investigations into the effects of clenbuterol administration and exercise training, i.e., finding an antagonistic relationship between the two interventions. Furthermore, by scaling these data from our animal model, tentative conclusions can be drawn about the deleterious effects of clenbuterol abuse within the sporting community. This is again supported by accounts of myocardial infarction in bodybuilders taking cocktails of anabolic agents, including clenbuterol (13). Clenbuterol abuse may therefore pose a long-term health risk. FOOTNOTESREFERENCES
Page 26airway remodeling processes in asthma may play an important role in modulating airway mechanics and provide an explanation for at least part of the airway hyperresponsiveness (AHR) seen in asthma (24). The remodeling comprises epithelial denudation, inflammatory cell infiltration, goblet cell metaplasia, subepithelial fibrosis, angiogenesis, and smooth muscle hyperplasia and hypertrophy in the airway wall of asthmatic subjects (19), of which increases in the volume of airway smooth muscle (ASM) appear to have the most significant impact on AHR. Limited numbers of studies in which the number of smooth muscle cells (SMCs) in airways from asthmatic and healthy individuals were counted have shown either that there was an increase in SMC number in the airways of asthmatic subjects compared with control airways (12, 18) or that there was no increase in asthmatic SMC number (35). The merit of the use of stereological techniques in the latter study notwithstanding, the use of airways from smokers as controls confounded the interpretation, as it is well established that chronic obstructive pulmonary disease is associated with airway wall remodeling and increases in the volume of ASM (5, 21). Increases in the volume of ASM have been detected in sensitized and antigen-challenged rats (10, 33, 37) and cats (25,26). However, it is still unclear whether the increased volume of muscle detected in these studies is due to hyperplasia or to hypertrophy of SMCs or to an increased volume of extracellular matrix. Recent studies examining DNA synthesis are strongly suggestive of ASM proliferation (28, 30-32), but this interpretation has not been formally tested. We have used the ovalbumin (OA)-challenged Brown Norway (BN) rat, in which chronic OA challenge is an established stimulus for airway wall remodeling (9, 20, 28, 30-33), to examine whether OA challenge resulted in increased SMC number in large airways. In addition, the effects of FCS alone and when combined with OA were examined, as FCS is known to enhance the proliferation of individual cell types that contribute to the remodeling response. It was anticipated that such exposure to FCS would increase the SMC number to a greater extent than OA alone. The use of FCS would, therefore, facilitate exploration of any relationship between SMC number and airway reactivity over a greater range of SMC number than generated by the standard chronic OA challenge. To investigate the SMC hyperplasia, we have applied a new approach to SMC counting with a simple, direct, and efficient method, instead of the tedious methods of SMC counting used in previous studies. With the use of the Cavalieri and the optical dissector methods (4), the total number of airway SMCs in the main bronchus has been estimated. Mast cell numbers were enumerated because increased numbers have been identified in the muscularis of asthmatic subjects (2), and many mast cell mediators are potentially significant in initiating airway SMC proliferation. OA challenge of BN rats resulted in a modest level of AHR, but there was no increase in the number of airway SMC in the left main bronchus. However, the incorporation of FCS into the antigen challenge protocol resulted in significant increases in SMC number and AHR that persisted for at least 7 days after the final OA and FCS exposure. METHODSIn the first series of experiments, 12 male BN rats (180–200 g, Animal Resource Centre) were sensitized by a single 1-ml intraperitoneal (ip) injection of OA (1 mg/ml in 100 mg/ml aluminium hydroxide in sterile saline). In addition, each rat received a 0.5-ml ip injection of the adjuvant Bordetella pertussis containing 6 × 109 heat-killed organisms. Fourteen days after sensitization, the rats were divided into two groups: one group was exposed to an aerosol of 5% (wt/vol) OA generated by a Hudson nebulizer (driven by a flow rate of 8 l/min of compressed air) for 15 min on three occasions with an interval of 5 days between each exposure (OA group, n = 6), whereas the other group was exposed to an aerosol of saline of the same duration (saline group,n = 6). In the second series of experiments, 32 male BN rats (220–290 g) were sensitized as described above. Fourteen days after sensitization, the rats were divided into four groups: one group was exposed to an aerosol of saline for 15 min on four occasions with an interval of 5 days between each exposure (saline group,n = 8), whereas the other three groups were exposed to an aerosol of FCS (10% vol/vol; FCS group, n = 7), OA (5% wt/vol; OA group, n = 7), or FCS and OA (FOA group, n = 6) for the same duration. The OA sensitization and challenge protocol was modified from that used by Sapienza et al. (33). The experiments described in this manuscript were conducted in compliance with the guidelines of the National Health Medical Research Council of Australia on animal experimentation. Airway reactivity to ACh was assessed in vivo 24 h (first series of experiments) or 7 days (second series of experiments) after the last aerosol challenge. The rats were anesthetized with an ip injection of pentobarbitone sodium (12 mg/kg) and ethyl carbamate (1 g/kg). A tracheal cannula, through which the rat was mechanically ventilated with room air, was inserted via a tracheotomy. Cannulas were also inserted into the external jugular vein for drug administration and into the carotid artery for blood pressure monitoring. The rat was then placed in a whole body plethysmograph and mechanically ventilated (10 ml/kg, 60 breaths/min) (rodent ventilator, UGO Basile). The transrespiratory pressure was measured with a differential pressure transducer (Validyne) with one port attached to the interior of the plethysmograph and the other port attached to the intratracheal cannula. The airflow rate was measured with a pneumotachograph. The respiratory resistance and compliance were calculated on-line with a Buxco pulmonary mechanics analyzer (Buxco Electronic) based on the principles of Amdur and Mead (1), and the outputs were displayed on a Macintosh LC III monitor (Apple) by using MacLab software (AD Instruments). After surgery, an initial dose of gallamine triethiodide (8 mg/kg) was administered intravenously to inhibit spontaneous respiratory movements, and further doses of 4 mg/kg were administered as required. After stabilization of cardiorespiratory parameters (15–20 min), the rat was given ACh intravenously at an initial dose of 25 μg/kg and then in increasing doses by doubling up to 400 μg/kg to obtain a response of >100% increase over baseline. The response was measured as the peak increase above the baseline immediately before ACh administration. The dose required to increase respiratory resistance by 100% (PC100) was estimated by log-linear interpolation of dose-response curves from individual animals. After assessment, the rats were killed by an overdose of anesthetic. Fixation in situ with 4% paraformaldehyde in 0.1 M phosphate buffer was achieved by perfusion via a tracheal cannula and via an abdominal aorta cannula. The pressure for tracheal infusion was 25 cmH2O and for the arterial perfusion was 100 mmHg. After 3 min of in situ fixation, the lung was removed from the rat and immersed in the same fixative for 24 h. Samples from the right main bronchus and lung were paraffin embedded. The paraffin sections (4-μm thickness) were used for hematoxylin and eosin and toluidine blue staining. The left main bronchus was used for stereological study of ASM. The specimen was defined anatomically as beginning at the tracheal bifurcation and continuing until the next bifurcation. The dissected left main bronchus was dehydrated through a series of increasing concentrations of ethanol. The sample was infiltrated with catalyzed ImmunoBed A solution and embedded with a mixture of 1 part of ImmunoBed B and 25 parts of catalyzed immunoBed A (Polysciences, Warrington, PA). The resin-embedded specimens were serially sectioned at 30 μm (Fig.1). Fig. 1.The left main bronchus was dissected (A) and embedded in ImmunoBed resin. The bronchus, defined anatomically as beginning at the tracheal bifurcation and continuing until the next bifurcation, was cut exhaustively from a random start at a thickness of 30 μm, and every 20th section was used for the next sampling step (B). A series of frames were selected with the aid of an autostage moving apparatus (C–E; see text for detailed explanation). Cell nuclei intersecting the forbidden lines (thick lines) were not counted (F). The 4-μm paraffin sections of 4% paraformaldehyde-fixed bronchial airway were stained with hematoxylin and eosin or toluidine blue by using standard techniques, and mast cells and eosinophils were enumerated by using Image Analysis software (see below). The 30-μm sections of resin-embedded right bronchus were stained with hematoxylin for 60 min and eosin for 15 min to allow permeation of the stains through the plastic resin. SMCs were enumerated by using stereological techniques as described below in detail. Eosinophil and mast cell numbers within the airway wall of the left bronchus were enumerated by using standard image analysis techniques. Briefly, the number of eosinophils (eosin-positive granulated cells) and mast cells (metachromatic appearance in toluidine blue-stained sections) were counted in a complete airway wall section from each animal in each of the subject groups with the aid of a light microscope (Leica DMIRB). The number of positive cells was normalized to the respective sample area (area of bronchus wall) by using a calibrated image analysis system comprising an RGB video camera (Hitachi) coupled to Image Pro Plus 4.0 image analysis software (Media Cybernetics). The cell numbers are expressed as the number of positive cells per millimeter square of bronchus wall. The bronchial wall is composed of four layers: mucosa (epithelium, basement membrane, lamina propria), submucosa, cartilage, and adventitia (3). The bronchial wall area measured in this stereological study did not include the adventitial layer because of the difficulty in defining its outer border on the sections of bronchial wall. A systematic, random sampling of thick (30-μm) sections was used to evaluate SMC number (7, 17). The bronchus was exhaustively sectioned at 30-μm thickness by using a Reichert-Jung 1150 Autocut (Nussloch, Germany) fitted with a glass ralph knife. Beginning with a random start, every 20th section was sampled, mounted on a poly-l-lysine-coated glass slide, stained with hematoxylin (60 min) and eosin (15 min), and viewed by microscopy. The image was displayed on a high-resolution monitor at a final magnification of ×2,060 by a video camera, and a series of systematic fields were selected with the aid of an electronic stepping stage. The “optical dissector” method was employed to count SMCs (16) within the acceptance area of a three-dimensional (3D) computer-generated unbiased counting frame (or else brick) of 275-μm2 area and 10-μm depth. The topmost plane of the optical dissector was positioned at a minimum of 10 μm below the cut surface of the section, and all of the SMC nuclei that came into focus within a subsequent depth of 10 μm were counted. For each bronchus, a minimum of 10 sections was sampled and 150–200 SMCs counted, to accurately determine the total number of SMCs (10, 17). The numerical density of SMC in the bronchial wall was calculated by dividing the number of SMCs counted by the volume sampled such that Nv=∑Nsmc/∑Pdisector×V(P) where Nv is SMC density, Σ Nsmc is total number of SMC counted, Σ Pdissector is total number of points sampled, and V(P) is the volume associated with each point.The Cavalieri method was used to estimate the volume of the bronchus as defined for the optical dissector. First, the total cross-sectional area of the previously sampled bronchi sections was determined by point counting. The volume was then calculated by using the following equation Vbron=∑Pbron×A(P)×T ×total number of sections/number of sections sampled where Vbron is volume of bronchus, Σ Pbron is total number of points overlying the bronchial wall in all sampled sections, A(P) is the area of each test grid (point), and T is thickness of each section. The Nsmc in the bronchus was calculated by multiplying the Nv by the Vbron, where Nsmc = Nv× Vbron. The following chemicals were used: ACh (BDH), aluminium hydroxide (BDH), Bordetella pertussis (CSL), ethyl carbamate (AJAX, Sydney, Australia), gallamine triethiodide (Flaxedil, May & Baker), ImmunoBed A (Polysciences), ImmunoBed B (Polysciences), Nembutal (Boehringer Ingelheim), OA (Grade II, Sigma Chemical, St. Louis, MO), paraformaldehyde (Probing & Structure), poly-l-lysine (Sigma Chemical), and toluidine blue (Sigma Chemical). Results are presented as grouped data from n rats and are expressed as means ± SE; n represents the number of rats. Student's unpaired t-test was used to determine whether there were significant differences between pairs of means. In some cases, a two-way ANOVA was used to investigate whether there was an effect of OA or FCS treatment and whether there was an interaction between the two variables. All statistical analyses were performed by using GraphPad Prism for Windows (version 2.01). In all cases, probability levels < 0.05 (P < 0.05)were taken to indicate statistical significance. RESULTSMost of the animals that were exposed to an aerosol of OA developed an immediate response within 2–5 min. The responders sneezed and showed increased amplitude of breathing movements and wheezing. Signs of obstruction resolved quickly after the termination of OA exposure. Saline-exposed rats did not display these signs. There was no significant difference in weight (251 ± 1 g for saline vs. 254 ± 2 g for OA) or baseline respiratory resistance (0.44 ± 0.04 cmH2O · ml−1 · s for saline vs. 0.41 ± 0.04 cmH2O · ml−1 · s for OA) between sensitized, saline-challenged and sensitized, OA-challenged rats that had their airway mechanics assessment 24 h after the last of three OA aerosol challenges administered 5 days apart from day 14 (P > 0.05, Student's unpairedt-test). In animals that had their airway mechanics assessment 7 days after the last aerosol challenge, there was a small difference in weight between saline-challenged animals and those challenged with FCS and FOA (Table 1). The baseline respiratory resistance values for FCS- and OA-treated animals were significantly different from those of saline-treated animals (P < 0.05, two-way ANOVA), but there was no significant interaction between FCS and OVA treatment (P > 0.05, two-way ANOVA) (Table 1).
In the first series of experiments in which airway mechanics were assessed 24 h after the last aerosol challenge, both sensitized saline-challenged and sensitized OA-challenged animals showed a dose-dependent increase in respiratory resistance to intravenous ACh. However, OA-challenged rats had significantly greater responses to ACh at doses of 100 and 200 μg/kg compared with sensitized saline-challenged rats (Fig. 2). The PC100 was significantly lower for OA-challenged animals [log(PC100): 2.12 ± 0.02] than for the saline group [log(PC100): 2.48 ± 0.07; P < 0.01]. Fig. 2.Effect of intravenous ACh on respiratory resistance in sensitized Brown Norway rats that were challenged with either saline (open bars) or ovalbumin (OA; solid bars). Values are means ± SE. * Significant difference between saline and OA-challenged rats,P < 0.001 (Student's unpaired t-test). In the second series of experiments in which airway mechanics were assessed 7 days after the last aerosol challenge, the PC100was significantly lower in FCS-treated than in saline-treated animals (P > 0.05, two-way ANOVA) but was not different in animals treated with OA alone (P > 0.05, two-way ANOVA) (Table 1). In addition, two-way ANOVA revealed that there was no interaction between FCS and OA (P = 0.6179). The OA challenge protocol used in both series of experiments produced an obvious airway inflammation (Fig. 3). In the first series of experiments, marked eosinophil infiltration and lymphocytic nodules in the airways and large mononuclear cell-dominated, mixed cell nodules in the alveolar spaces were observed 24 h after exposure to the last of three OA challenges (each separated by 5 days). In saline-challenged animals, there was no consistent inflammatory cell infiltration, although isolated lymphocytic nodules and a small number of eosinophils were observed in some specimens (Fig. 3). Eosinophil number increased significantly (P < 0.05, unpaired t-test) from 22 ± 6 (number/mm2 bronchus wall area) in saline-treated rats to 362 ± 102 in OA-exposed rats. In contrast, mast cell numbers in saline-exposed rats (48 ± 7) were not different from those in OA-exposed rats (63 ± 18). Moreover, the distribution of mast cells did not appear to be affected by OA challenge (data not shown). Fig. 3.Representative sections of large airway from animals exposed to saline (a) and OA (b) 24 h after the last aerosol exposure or saline (c), OA (d), FCS (e), or FCS and OA (f) 7 days after the last aerosol exposure (hematoxylin and eosin stained). Structural and cellular features of note are shown at both low (×40, leftpanels) and high (×100, right panels) magnification and include a mononuclear cell infiltration (m), goblet cell hyperplasia (gob), eosinophils (eos), smooth muscle cells (smc), and epithelial cells (epi). The degree of inflammation in airways harvested 7 days after the last exposure to OA appeared to be less than that observed 24 h after the final challenge (Fig. 3). However, the eosinophilia in response to OA was not statistically different from that observed in tissues harvested 24 h after the last OA exposure (Table2). Furthermore, FCS alone had no significant effect on eosinophil number but appeared to reduce the response to OA (Table 2). Mast cell numbers were unchanged (P > 0.05) by OA, FCS, or FOA exposure when measured 7 days after the last aerosol exposure (Table 2).
In the first series of experiments in which airway tissue was harvested 24 h after the last challenge, OA aerosol challenge had no effect on SMC number (saline exposed, 5.17 ± 0.29 × 105; OA exposed, 5.54 ± 0.47 × 105) or left main bronchus volume (saline exposed, 5.98 ± 0.23 mm3; OA exposed, 6.41 ± 0.28 mm3). However, in animals that had airway mechanics assessed 7 days after the last aerosol challenge, OA alone and the combination of FCS and OA treatment increased SMC numerical density (P < 0.05, two-way ANOVA) and the total number of SMCs in the left main bronchus (P < 0.001, two-way ANOVA) (Table3). Although there was a significant interaction between FCS and OA with respect to increases in SMC numerical density (P = 0.0314, two-way ANOVA) and total number of SMC (P < 0.0001, two-way ANOVA) (Table 3), FCS alone had no effect on either of these measures of remodeling. The combination of FCS and OA treatment increased bronchus volume (P < 0.05, two-way ANOVA), but neither stimulus alone had an effect (P > 0.05, two-way ANOVA) (Table 3).
DISCUSSIONIn this study, the existence of smooth muscle hyperplasia was investigated in a well-established model of antigen-induced airway wall remodeling in OA-sensitized BN rats. Previous studies have established an increase in the mass of ASM and an increase in the fraction of cells incorporating bromodeoxyuridine (BrdU), a marker of cells in, or having been in, S phase (28, 30-32). These observations suggest, but do not establish, that the increase in ASM mass is at least partly the result of ASM hyperplasia, with hypertrophy and increased extracellular matrix also having the potential to increase the volume of bronchus wall occupied by smooth muscle. The number of SMC in the bronchus was investigated by using methods of unbiased stereology to directly ascertain whether the reported increased volume of smooth muscle observed in BN rat airways may be explained by an increase in SMC number. There was no increase in SMC number in the left main bronchus of sensitized BN rats that were exposed to three aerosols of OA, despite increased airway reactivity and inflammation. However, when animals were exposed to four aerosols of OA or when FCS was incorporated into the OA challenge protocol, there was a significant increase in SMC number. The reasons for the difference in the outcome between three and four OA exposures may relate to the additional time for remodeling to occur in the latter protocol (13 days extra). Although a number of studies report increased ASM mass after only three challenges (28, 30-32), the differences in the efficiency of delivery of the OA challenge in different protocols precludes meaningful direct comparison of the extent of OA challenge required for remodeling. The BN rat model of OA-induced inflammation has been used extensively to evaluate the mechanisms underlying airway inflammation and AHR. Sensitization alone has no effect on airway responsiveness and does not elicit measurable airway inflammation (14, 15, 20, 29,36). A single OA challenge appears to be adequate to induce transient AHR, but multiple OA aerosol challenges are required for persistent AHR (15, 33). Cysteinyl leukotrienes (31) and endothelin (30) antagonists reduce the SMC DNA synthesis induced by repeated OA exposure without reducing the AHR in this model. In the present study, sensitized BN rats developed a persistent AHR (7 days) after a combination of OA and FCS challenges, but not after OA alone, even though OA alone was associated with a transient AHR. There was an obvious eosinophilic airway inflammation 24 h after the last OA challenge that had not resolved after 7 days. Recently, Palmans et al. (27) showed that, after 2 wk of OA exposure, there was an increase in AHR and eosinophilic inflammation, but repeated OA exposure for 4–12 wk was accompanied by a resolution of the AHR, despite persistent eosinophilic inflammation. Thus eosinophilic airway inflammation may not be the only factor regulating the development of AHR. Studies using different protocols for OA exposure have provided evidence consistent with a dissociation of eosinophilia and AHR by showing that AHR persists beyond the eosinophilic response. Repeated, but not single, OA challenge of BN rats induced a persistent AHR at a time when eosinophilic airway inflammation was resolving (15). Cui et al. (9) also demonstrated that 9 wk of antigen challenge resulted in AHR without concurrent eosinophilia. Thus, whereas the current evidence does not preclude a role for eosinophils in either airway wall remodeling or AHR, persistent eosinophilia does not appear to be required for persistent AHR in the BN rat model. Recent studies in atopics and patients with mild asthma support the notion that eosinophils alone are not responsible for allergen-induced AHR or the late asthmatic response (6, 22). Systemic administration of recombinant human interleukin-12 to patients with mild allergic asthma caused a decrease in blood and sputum eosinophils but had only minor effects on airway responsiveness to histamine and no effect on airway responsiveness to inhaled allergen (6). Intravenous administration of an interleukin-5-blocking antibody to patients with allergic asthma caused a marked suppression of the increase in blood and sputum eosinophil numbers after allergen challenge but did not protect against the allergen-induced late asthmatic response or the increase in airway responsiveness to histamine (22). T cells adoptively transferred from OA-sensitized rats confer increased reactivity on OA-challenged recipient rats (23). The role of T cells in AHR in asthma has yet to be defined. A digitized tracing method has been widely used to show that airway SMC volume is increased in asthmatic airways. Only three studies have attempted to resolve the contributions of smooth muscle hyperplasia and/or hypertrophy to this increase in SMC volume in asthmatic airways by directly counting the number of SMCs (12, 18, 35). Two of these studies showed a threefold increase in SMC number in the large airways of human asthmatic subjects compared with healthy individuals (12, 18), whereas the third failed to find an increase in SMC number, possibly because of inclusion in the “nonasthmatic” control group of tobacco smokers and patients with emphysema (35). Airways from smokers and chronic obstructive pulmonary disease patients may be thickened and show SMC volume increases (5, 21). Ebina and colleagues (12) used 3D reconstruction of a series of sections from control and asthmatic subjects to demonstrate hyperplasia as well as hypertrophy of ASM in asthma. Asthmatic patients were classified into two types: type I asthmatic airways showed SMC hyperplasia only in the larger bronchi, and there was no SMC hypertrophy in either the large or small airways; type II asthmatic airways, on the other hand, showed only mild hyperplasia in the larger bronchi, but SMC hypertrophy was evident in large and small airways. We have used a new approach to SMC counting by a combination of unbiased, systematic, random-start sampling and the use of an unbiased probe: the optical dissector under high magnification. This sampling method is unbiased because there is no design-derived error (8,17). The optical dissector method allows the use of thick sections and the counting of SMC directly in a series of 3D counting frames. This approach is much simpler and more efficient than the 3D reconstruction method of Ebina et al. (12), although similar principles are applied. In addition, tissue sections were embedded in plastic rather than paraffin, because tissue shrinkage is <4% in plastic sections, but up to 14% in paraffin sections (35). Sensitized BN rats that had been exposed to four aerosols of OA showed a significant increase in SMC number in the left main bronchus compared with sensitized, saline-challenged animals. Interestingly, the incorporation of FCS into the antigen-exposure protocol resulted in a greater increase in SMC number than that observed in animals treated with OA alone. FCS is mitogenic for cultured airway SMCs (34). Thus it was not surprising to find an increase in SMC number in FCS-treated animals and synergy between FCS and OA. In addition, the bronchus volume and SMC numerical density were also significantly increased in animals treated with FOA. These results are consistent with earlier studies showing increased smooth muscle area in antigen-challenged BN rats: Sapienza et al. (33) demonstrated a twofold increase in the quantity of ASM in small, medium, and large airways of OA-challenged rats; Cui et al. (9) showed that chronic trimellitic anhydride (occupational allergen) exposure increased the thickness of smooth muscle in the small airways in sensitized BN rats; Wang et al. (37) also demonstrated an increase in ASM volume in large airways (internal perimeter > 2 mm) of OA-challenged rats; Panettieri et al. (28), using BrdU incorporation, indicated that DNA synthesis increased in SMCs after three OA challenges. Similarly, Salmon et al. (32) showed that repeated OA exposure increased the fraction of BrdU-positive ASM in the BN rat model. Our findings suggest that the increase in ASM area is explained in part by an increase in SMC number, but we have not estimated the area/volume of airway wall occupied by smooth muscle. A recent study by Palmans et al. (27) showed that the area of smooth muscle around the airways did not change in OA-challenged BN rats, despite an increase in the total wall area of small, medium, and large airways in sensitized rats exposed to OA for 2 wk. It remains possible that SMC number increases without detectable increases in area occupied by ASM. Increased reactivity 24 h, but not 7 days, after the final OA challenge in our chronic exposure study suggests that inflammatory factors rather than structural changes are key determinants of this acute AHR. Incorporation of FCS into the OA challenge protocol (FOA) resulted in a substantial increase in SMC number and a modest, persistent AHR. However, the AHR in FOA-treated animals was largely associated with an interaction between FCS and OA, because neither treatment alone showed marked increases in responsiveness. Interestingly, the persistent eosinophilia seen 7 days post-OA challenge was diminished by the addition of FCS to the challenge aerosol, providing further evidence of a dissociation between eosinophilia and persistent AHR. In animals treated with FOA, there was an increase in SMC number and AHR, raising the possibility that structural rather than inflammatory mechanisms may contribute to the persistence of AHR after repeated FOA challenge in this model. Nevertheless, it is also possible that FCS induced a reactivity change that was independent of the increase in SMC number, as FCS alone increased reactivity without influencing SMC number. In summary, stereological methods were used to study ASM hyperplasia in sensitized antigen-challenged BN rats. There was no hyperplasia of SMC in the left main bronchus in animals exposed to three aerosols of OA, despite an acute increase in AHR and airway inflammation. However, sensitized rats that received four aerosols of OA showed a significant increase in SMC number that was further increased when FCS was incorporated into the OA challenge protocol. These animals also displayed AHR, but this effect could be attributed to the actions of FCS. Thus chronic exposure to OA can cause acute reactivity changes, but additional stimuli are required for persistent changes in reactivity in these relatively short treatment protocols. Compared with traditional two-dimensional morphological studies, modern stereology provides a new, simple, and efficient approach to obtaining precise and reproducible information in a 3D volume. Future applications of this technique to the studies of airway remodeling will advance our knowledge of the structure-function relationship in airways that have undergone remodeling. We thank Anne Pirdas-Zivcic for technical assistance. FOOTNOTESREFERENCES
|