The control of arteriolar diameter reflects the sum of multiple inputs to vascular smooth muscle cells. In skeletal muscle, activation of sympathetic nerves results in vasoconstriction.123 In contrast, increasing the metabolic activity of muscle fibers induces vasodilation.45 Arteriolar diameter also reflects changes in transmural pressure and luminal blood flow.67 The interaction between central (ie, neural) and local mechanisms of blood flow control is a long-standing question in cardiovascular and exercise physiology. Of particular interest in the present study is the apparent “competition” between sympathetic vasoconstriction and the vasodilation that occurs during muscular activity.5891011 Sympathetic outflow increases with exercise, yet there is a preferential increase in blood flow to active muscle. In effect, functional vasodilation “overrides” sympathetic vasoconstriction, giving rise to the concept of “functional sympatholysis.”59 Nevertheless, sympathetic nerve activity can limit muscle blood flow by constricting resistance vessels.111213 Show The mechanisms of interaction between sympathetic vasoconstriction and peripheral vasodilation have been studied in some detail. For example, substances released by muscle fibers1 and motor nerves14 can inhibit norepinephrine (NE) release and dilate arterioles. Vasomotor responses can also result from stimuli delivered to sites well removed from a particular branch of the arteriolar network. For example, acetylcholine (ACh) microiontophoresis induces a dilation that is conducted rapidly from cell to cell along the arteriolar wall over distances encompassing several millimeters and multiple branch orders.151617 However, the influence of sympathetic nerve activity on cell-to-cell conduction in arteriolar networks has not been determined. In the present study, we tested the hypothesis that sympathetic nerve activity can influence the conduction of vasodilation along the arteriolar wall. Our findings in striated muscle show that stimulation of sympathetic nerves substantially depressed conducted vasodilation via activation of α-adrenoceptors. Furthermore, the conduction of vasodilation significantly attenuated sympathetic vasoconstriction of arterioles, suggesting a novel explanation for functional sympatholysis.59 Materials and MethodsAnimal Care and Preliminary SurgeryAll procedures were approved by the Institutional Animal Care and Use Committee of the John B. Pierce Laboratory. Male golden hamsters (n=21, 109±4 g, Charles River) were maintained at 24°C on a 14-h/10-h (light/dark) cycle and provided food (Purina rodent chow) and water ad libitum. Hamsters were anesthetized with pentobarbital sodium (60 mg/kg IP) and tracheotomized to maintain a patent airway. The right carotid artery was cannulated for monitoring arterial pressure (CDX III transducer, Cobe Laboratories). Another cannula was secured in the left femoral vein to replace fluids and maintain anesthesia during experiments (10 mg pentobarbital per milliliter isotonic saline, infused at 0.41 mL/h). Cremaster PreparationThe right hamster cremaster muscle was prepared as recently described in detail.16 Briefly, by use of a stereo microscope (model DRC, Zeiss), the muscle was exposed and positioned onto a transparent acrylic pedestal. The cremaster muscle was opened from the apex to the inguinal canal along the ventral surface. The spermatic artery and vein were ligated, and then the testis and epididymis were carefully separated from the muscle and removed. Eight to 12 sutures (6-0 silk, Ethicon) were secured around the edge of the muscle and used to spread the tissue radially. The cremaster preparation was superfused continuously (5 mL/min) with a bicarbonate-buffered physiological saline solution (PSS) (34±1°C, pH 7.4) of the following composition (mmol/L): NaCl 131.9, KCl 4.7, MgSO4 1.2, CaCl2 2, NaHCO3 18 (Sigma Chemical Co). Superfusion solutions were gassed continuously with 5% CO2/95% N2 unless noted otherwise; in three experiments, a Clark-type electrode (model PHM 7/MK2, Radiometer Copenhagen) was used to measure Po2 of the PSS on the surface of the preparation. Esophageal temperature was maintained at 38±1°C by positioning the hamster on a copper coil through which warm water (43°C) was circulated. When all experimental procedures for the day were complete, the hamster was given an overdose of pentobarbital through the venous cannula. Video MicroscopyThe preparation was transferred to the stage of an intravital microscope (model ACM, Zeiss) and equilibrated for 45 minutes. Second-order (2A; resting diameter, 15 to 54 μm) and third-order (3A; resting diameter, 9 to 30 μm) arterioles were selected for study on the basis of optical clarity and resting tone, as demonstrated by a brisk and reversible dilation in response to topical application of adenosine (0.1 mmol/L). Microvessels were observed with bright-field microscopy using Köhler illumination (ACH/APL condenser [numerical aperture, 0.32]; objectives, Zeiss UD40 [numerical aperture, 0.41] or Leitz L25 [numerical aperture, 0.35]). A video camera (CCD model C2400, Hamamatsu) was positioned on a trinocular imaging tube and coupled to a video monitor (model PVM 1343MD, Sony). Final magnification on the monitor was ≈×1400 when either objective was used. Internal vessel diameters were recorded from the video monitor by using a video caliper. A stage micrometer (100×0.01=1 mm, Graticules Ltd) was used for calibration; spatial resolution was ≤1 μm. The direct effect of ACh was assessed at the micropipette tip (referred to as “local”), and conducted vasodilation was evaluated at the origin of the stimulated branch, which was typically ≈1 mm upstream from the ACh micropipette1617 (Fig 1). Control experiments eliminated the possibility of ACh diffusion to the upstream site or of nonspecific effects of iontophoretic current on arteriolar diameter.1516 Data were acquired at 100 Hz by using a MacLab system (AD Instruments) coupled to a Macintosh IIVX computer. Micropipettes and MicroelectrodesBorosilicate glass capillary tubes (Corning No. 7740; outer diameter, 1.2 mm; inner diameter, 0.68 mm; Warner Instrument Corp) were pulled (model P-87, Sutter Instruments) to produce micropipettes with tips (outer diameter) of 1 to 2 μm (for microiontophoresis) or 2 to 3 μm (for perivascular nerve stimulation [PNS]); tip dimensions were measured at ×630 optical magnification with bright-field microscopy. Micropipettes were backfilled with 1 mol/L ACh (Sigma) or 0.9% NaCl after filtering to remove particles >0.2 μm (Acrodisc, Gelman Sciences). MicroiontophoresisMicropipettes containing ACh were secured in a holder and connected to an iontophoresis programmer (model 160, World Precision Instruments) via a Ag/AgCl wire; the programmer was gated externally by the MacLab system. A second Ag/AgCl wire secured at the edge of the preparation served as the reference electrode. Micropipettes were positioned ≈1 mm distal to the vessel origin (Fig 1) with a micromanipulator (model M, Leitz); a large movable stage enabled the entire preparation and micromanipulators to be moved as a unit without disturbing the spatial relation between arterioles, micropipettes, and microelectrodes.16 The retaining current (0.1 to 0.2 mA) was adjusted to just prevent leakage (indicated by vasodilation) from the micropipette tip. Based on dose-response curves to ACh microiontophoresis (data not shown), the amplitude of ejection current was held at 1 mA; stimulus durations were selected to elicit maximal and half-maximal conducted responses (500 and 200 ms, respectively). To quantify the effect of ACh on arteriolar diameter, the magnitudes of local and conducted responses to microiontophoresis were calculated as follows: peak response diameter (in micrometers) minus preceding baseline diameter (in micrometers). Perivascular Nerve StimulationA distal segment of the first-order (1A) arteriole was exposed by microdissection of adjacent striated muscle fibers. A stimulating microelectrode was prepared by using a micropipette filled with 0.9% NaCl; this was secured in a Leitz micromanipulator and positioned adjacent to the exposed 1A segment (Fig 1). The microelectrode was connected via a Ag/AgCl wire to the negative terminal of a stimulator (model S48, Grass Instruments Co); the positive terminal of the stimulator was connected to the Ag/AgCl reference wire. With constant pulse duration (1 ms) and stimulation frequency (8 Hz), voltage for PNS was adjusted (average, 120 V) until maximal constriction was observed in an arteriolar branch located ≈5 mm proximal to the microelectrode; observations at such remote sites ensured that vasoconstriction was not due to direct depolarization of smooth muscle cells at the microelectrode tip. To determine the frequency-response characteristics of arterioles to PNS, diameter responses were characterized at seven stimulation frequencies (0.5, 1, 2, 4, 8, 16, and 32 Hz)18 in five hamsters. The train durations for stimulation were selected to provide stable diameter responses and varied inversely with stimulation frequency: 30 s for 0.5 and 1 Hz; 20 s for 2, 4, and 8 Hz; 16 s for 16 Hz; and 8 s for 32 Hz. At 16 Hz, for example, constriction began ≈3 s after initiating PNS and peaked within 16 s. To test for activation of sympathetic nerves, the above procedures were repeated after 30 minutes of exposure to phentolamine (1 μmol/L, Research Biochemicals Inc), an α-adrenoceptor antagonist, added to the superfusion solution. For summary frequency-response curves, changes in arteriolar diameter (peak diameter response [in micrometers] minus resting diameter [in micrometers]) at a given frequency of PNS were divided by the maximal diameter change and expressed as percentage of maximum for each arteriole. These summary data were used to ascertain the stimulus frequencies required to elicit 50% and 100% of the maximal PNS responses (4 and 16 Hz, respectively), which were used in subsequent experiments. Although 32 Hz was often the stimulus giving maximum response, preliminary experiments revealed sympathetic “escape” (ie, dilation after constriction during nerve stimulation) in 3A arterioles during stimulation at this frequency.1 Thus, responses to 16-Hz PNS were taken as maximal without escape. Interaction Between PNS and AChFor these experiments, eight hamsters were used to test whether PNS (ie, sympathetic vasoconstriction) would alter conducted vasodilation in arterioles. After positioning the ACh micropipette as shown in Fig 1, ACh was applied (1 mA, 500 and 200 ms) under control conditions; each ACh stimulus was then tested during 4- and 16-Hz PNS. Because PNS results in arteriolar constriction, control experiments were necessary to account for the change in diameter per se. Therefore, responses to identical ACh stimuli were also evaluated during vasoconstriction induced by elevating superfusate O2 concentration from 0% to 10%. For all experiments, diameter values were quantified at resting baseline and when responses to experimental manipulations (eg, PNS or 10% O2) had stabilized. Local and conducted responses to ACh microiontophoresis were recorded at the peak of the diameter response at rest and during PNS or 10% O2. Representative diameter tracings for individual 2A arterioles are presented in Fig 1 to illustrate experimental protocols. Conduction and PNS With Sympathetic BlockadeThese experiments tested whether inhibition of α-adrenoceptors with phentolamine (1 μmol/L) or blockade of nerve action potentials with tetrodotoxin (TTX, 1 μmol/L; Sigma) would alter the influence of PNS on conducted vasodilation. On the basis of the similarity of results between 2A and 3A branches in the above experiments (see “Results”), only 2A arterioles were studied with these protocols. Micropipettes containing ACh were positioned as in Fig 1. Responses to PNS (16 Hz) and ACh (1 mA, 500 ms) were studied before and after 30 minutes of exposure to either phentolamine or TTX in the superfusion solution (Fig 1). Eight hamsters were studied with these protocols (phentolamine, n=3; TTX, n=5). Arteriolar diameters at rest and in response to PNS were measured under control conditions and in the presence of phentolamine or TTX (Table). Responses to ACh microiontophoresis at rest and during PNS were recorded locally and at upstream (conducted) sites in the absence and presence of phentolamine or TTX as described above. StatisticsOne to three arterioles were studied in each preparation; each vessel was treated as a separate experiment.1617 Experimental treatments affecting the entire preparation (eg, PNS, 10% O2, TTX, and phentolamine) are referred to as “global.” Unpaired t tests were performed to determine whether responses to PNS or to ACh varied between 2A and 3A arterioles. Repeated-measures ANOVA was used to compare the effect of global treatments on resting diameter and on vasomotor responses to ACh. Post hoc comparisons of cell means were performed by multiple linear comparisons with the family-wide error rate adjusted to P≤.05. Thus, critical P values for individual comparisons were determined by dividing .05 by the number of comparisons of interest as determined a priori for each analysis. All statistical comparisons were performed with superanova (Abacus Concepts Inc). Summary data are presented as mean±SEM. ResultsPreparations were stable throughout each day’s experiments (duration, 3 to 5 hours) as assessed by the maintenance of vasomotor tone (2A and 3A arterioles typically increased diameter 50% to 100% with adenosine) and the stability of mean arterial pressure from the beginning (96±3 mm Hg, n=20) to the end (98±4 mm Hg) of the experiments. Mean arterial pressure during PNS (95±1 mm Hg) was not different from control pressure. Responses to 200-ms ACh and 4-Hz PNS were similar in direction yet reduced in magnitude when compared with 500-ms ACh and 16-Hz PNS; therefore, only the latter data are presented for clarity. Perivascular Nerve StimulationThe magnitude of vasoconstriction increased with the frequency of stimulation in both 2A and 3A arterioles (Fig 2); there were no differences between responses (percentage of maximal response) of 2A or 3A arterioles at either 4 or 16 Hz. The greatest vasoconstrictions occurred at 16 or 32 Hz, irrespective of vessel order. Responses to Global StimuliThe Po2 of control PSS (gassed with 5% CO2/95% N2) averaged ≈30 mm Hg on the surface of the preparation and increased to ≈100 mm Hg when gassed with 10% O2/5% CO2/85% N2. The vasoconstriction induced by 10% O2 was not different from that elicited with 16-Hz PNS in either 2A or 3A branches (Fig 3). Whereas phentolamine did not affect baseline diameter, arterioles stabilized at smaller (P<.05) diameters in the presence of TTX (Table). Arteriolar constriction to PNS was reduced by 65% in the presence of phentolamine (Fig 2) and eliminated completely during TTX exposure (P<.05). Vasoconstrictor responses to PNS returned after washout (45 minutes) of phentolamine or TTX with control PSS (data not shown). Local Responses to Vasomotor StimuliAt the tip of the ACh micropipette, diameter increased significantly in response to ACh in both 2A and 3A arterioles (Fig 4). During vasoconstriction with 16-Hz PNS or elevated O2, ACh elicited greater dilation of 2A arterioles than was elicited during control conditions. Conducted ResponsesRepresentative tracings of conducted vasodilation in 2A arterioles are presented in Fig 1. As shown previously,1516 a 2- to 3-s delay preceded dilation, which peaked ≈10 s thereafter. Conduction increased arteriolar diameter at the vessel origin under all conditions (P<.05); the amplitude of conducted responses did not differ between 2A and 3A branches (Fig 4). Conducted vasodilation was significantly less during 16-Hz PNS than during control conditions or equivalent vasoconstriction with 10% O2 (Fig 4); this effect of PNS on conduction was not different between branch orders. In four arterioles, vasoconstriction was not observed during PNS; nevertheless, PNS attenuated conducted vasodilation in each case. Phentolamine had no affect on conducted vasodilation under control conditions (Fig 5), yet it eliminated the PNS-induced depression of conducted responses. In the presence of TTX, conducted vasodilation was enhanced by ≈25% at rest and was more than twofold greater during PNS (Fig 5). DiscussionWe have investigated the interaction between perivascular nerve activity and conducted vasodilation in arterioles of striated muscle. In response to PNS, frequency-dependent vasoconstriction occurred in 2A and 3A arterioles, which was suppressed by phentolamine and abolished with TTX. Whereas local vasodilation to ACh (ie, at the site of direct action) increased during PNS, the corresponding conducted responses were diminished by nearly half (Fig 4). The suppression of conduction by PNS was eliminated with TTX or phentolamine; TTX also enhanced conduction at rest. The present data are the first to demonstrate an interaction between sympathetic nerve activity and the conduction of vasomotor responses along arterioles. Our findings indicate the presence of neural modulation of cell-to-cell communication in arterioles of striated muscle both at rest and during elevated sympathetic outflow. Perivascular Nerve StimulationPrevious studies in the rat have activated sympathetic nerves by stimulating paravertebral ganglia.218 In the hamster, we have found these ganglia to be extremely difficult to isolate. Therefore, an alternative approach was taken. Because sympathetic nerves run as a plexus around the arteriolar network of striated muscle23 (S.S. Segal and B.D. Walker, unpublished data, 1994), we reasoned that action potentials triggered distally in the network should propagate in a retrograde direction. Thus, a segment of the primary arteriole was exposed near the distal edge of the tissue and the stimulating microelectrode positioned adjacent to the exposed vessel. In response to PNS, vasoconstriction propagated into arterioles located in the central region of the muscle, which confirmed our reasoning. Arterioles constricted in response to PNS over the range of 0.5 to 32 Hz, which is consistent with previous functional studies218 and recordings of sympathetic activity from peripheral nerves.19 The attenuation of PNS-induced vasoconstriction by phentolamine confirmed the activation of sympathetic nerves. The slight dilation observed at the lowest PNS frequencies in the presence of phentolamine (Fig 2) may have reflected activation of β-adrenoceptors.14 Alternatively, this vasodilation may have resulted from other substances released by perivascular nerves.2021 Nevertheless, the effects of PNS observed in the present study primarily involved the release of NE (Figs 2 and 5). The frequency-response characteristics of rat cremaster arterioles to sympathetic stimulation were found to vary with network location: 2A branches were less sensitive to low-frequency stimulation (0.2 to 4 Hz) than 3A branches.18 These differences have been explained by corresponding variation in the distribution of α-adrenoceptors within the arteriolar network (1A and 2A, α1 and α2; 3A and 4A, α2)1822 and provided a rationale for our studying both 2A and 3A arterioles in the hamster cremaster muscle. Although the direct effect of ACh on 2A arterioles was enhanced by vasoconstriction (Fig 4), we found no difference between branch orders in sensitivity to PNS (Fig 2) or to the interaction between conduction and PNS (Fig 4). Hamster arterioles may have more uniform α-adrenoceptor distribution than observed in the rat1822 ; however, this remains to be ascertained. Responses to AChLocal responses to ACh were either maintained or increased by global treatments (eg, PNS and elevated O2); in no case was the direct effect of ACh attenuated. In contrast, PNS attenuated conducted vasodilation (Figs 4 and 5). This reduction was not dependent on vasoconstriction per se because equivalent constriction with 10% O2 did not affect conduction. Because phentolamine eliminated the attenuation of conduction during PNS, we conclude that the effect of PNS on conducted vasodilation is mediated via NE activation of α-adrenoceptors. In a reciprocal fashion, ACh-induced vasodilation overcame the PNS-induced vasoconstriction both directly and at sites of conduction. In addition to its direct action as a vasodilator, ACh could attenuate sympathetic vasoconstriction via presynaptic inhibition of NE release.14 However, this effect would occur only at the site of ACh release. Because NE is released throughout the perivascular nerve plexus,2323 conducted vasodilation must interact with sympathetic vasoconstriction by a mechanism other than presynaptic inhibition. The conduction of vasodilation occurs via coupling between endothelial cells and smooth muscle cells along the arteriolar wall; a key component appears to involve the spread of hyperpolarization triggered locally by ACh.1724 In the hamster cheek pouch, micropipette application of depolarizing KCl solution (137 mmol/L) or the microiontophoresis of NE onto arterioles was found to induce vasoconstriction that conducted along arterioles and attenuated conducted vasodilation.1517 Nevertheless, the cheek pouch microcirculation is devoid of sympathetic nerves,25 and the influence of sympathetic nerve activity on conduction in arterioles has not previously been investigated. Arteriolar smooth muscle cells depolarize in response to sympathetic nerve stimulation or exposure to NE.2326 Therefore, the present findings lead us to hypothesize that depolarization of arteriolar smooth muscle cells induced by NE release during PNS may underlie the attenuation of hyperpolarization and conducted vasodilation triggered by ACh. Alternatively, NE may alter cell-to-cell coupling in the arteriolar wall and thereby reduce the amplitude of conduction. In support of this argument are the findings that NE increases intracellular Ca2+ in smooth muscle cells through binding to α-adrenoceptors,27 which may reduce gap junctional conductance.28 Whereas the present results are the first to indicate that PNS depresses conducted vasodilation via α-adrenoceptor activation in vivo, further experiments are required to identify the subsequent event(s) that influence conduction. TetrodotoxinTTX blocks the fast voltage-sensitive Na+ channels and thereby inhibits the propagation of action potentials. In previous studies, administration of TTX to cheek pouch arterioles had no affect on conduction.1617 This finding argued against a role for nerves in conduction and contributed to the conclusion that cell-to-cell coupling was the basis of conduction in arterioles.1724 In contrast to the cheek pouch,25 arterioles of the hamster cremaster muscle are richly invested with perivascular nerve fibers1821 (S.S. Segal and B.D. Walker, unpublished data, 1994); therefore, TTX should inhibit neurotransmitter release. In the presence of TTX, the elimination of PNS-induced vasoconstriction and augmented conducted responses (Fig 5) are consistent with this interpretation. Arterioles developed sustained constriction during exposure to TTX. In fact, the magnitude of constriction to TTX was not different from that obtained with PNS (Table). Although the cause of this response is unclear, vasoconstriction by itself (eg, with 10% O2) does not influence conduction (Fig 4). In addition, TTX does not directly affect the membrane potential of vascular smooth muscle cells.29 Therefore, whereas enhanced conduction during TTX exposure may be explained by elimination of perivascular nerve activity, the inability of phentolamine to alter conduction at rest suggests than neuromodulators in addition to NE could influence cell-to-cell coupling in arterioles. SignificanceFunctional hyperemia occurs in the cremaster muscle in response to electrical stimulation30 and during physical exercise31 ; cremaster preparations have proven highly useful in studies of blood flow control in the microcirculation of striated muscle.271618212230 Muscular exercise also increases the activity of the sympathetic nervous system.511 The mechanism by which active muscle overrides this vasoconstrictor stimulus is unclear in spite of the volume of research on this topic1589101213 ; the products of muscle metabolism cannot completely account for this phenomenon. ACh release at neuromuscular junctions increases greatly during exercise. As shown in the present study and in previous studies,1516 this molecule is highly effective in triggering conducted vasodilation. Recent work suggests that neuromuscular junctions in striated muscle could provide a vasomotor stimulus to arterioles.32 Therefore, we speculate that conducted vasodilation triggered at neuromuscular junctions may contribute to functional sympatholysis. Although the present findings demonstrate that the direct effects of ACh are not impaired in the presence of increased sympathetic activity, conduction is clearly suppressed. Nevertheless, the persistence of conducted vasodilation during PNS indicates that it may still contribute to the rapid increase in capillary surface area that occurs with the onset of muscular exercise.916 Summary and ConclusionPerivascular nerve stimulation at the distal end of primary arterioles activated sympathetic nerves throughout arteriolar networks in hamster striated muscle. During PNS, ACh microiontophoresis reversed vasoconstriction in 2A and 3A arterioles locally and by triggering conducted vasodilation. The magnitude of conducted vasodilation was diminished similarly by PNS in both vessel orders, and this effect was reversed with phentolamine. In the presence of TTX, responses to PNS were eliminated, and the magnitude of conducted vasodilation was increased. These findings indicate that sympathetic nerves can influence cell-to-cell communication along the arteriolar wall, both at rest and during enhanced sympathetic activity. The attenuation of sympathetic vasoconstriction by conducted vasodilation suggests a novel explanation for functional sympatholysis. Figure 1. Top, Diagram of arteriolar networks studied. PNS indicates perivascular nerve stimulation; ACh, acetylcholine. The break in the first-order (1A) arteriole accounts for the distance between the PNS microelectrode and the sites where PNS and ACh responses were assessed (see “Materials and Methods” for details). The ACh micropipette was positioned ≈1 mm downstream from the vessel origin (*), where conducted responses were measured. Bottom, Two representative tracings of conducted responses at the origin of second-order arterioles to ACh microiontophoresis (1 mA, 500 ms; given at arrowheads), 16-Hz PNS, and 10% O2 or tetrodotoxin (TTX , 1 μmol/L) in the superfusion solution. Solid bars below each tracing indicate the duration of corresponding manipulation (for TTX, application was begun 30 minutes before); the 10 sec calibration bar pertains to both tracings. In the upper tracing, note vasoconstriction and suppression of conducted vasodilation during 16-Hz PNS; TTX augmented conducted vasodilation above the control level and eliminated both vasoconstriction and suppression of conduction induced by PNS. In the lower tracing, whereas 16-Hz PNS suppressed conducted vasodilation, equivalent vasoconstriction with 10% O2 had no effect on the magnitude of conduction. Figure 2. Frequency-response curves of arteriolar diameter to perivascular nerve stimulation (PNS). Responses to PNS were evaluated in second-order (2A, n=8, top panel) and third-order (3A, n=9, bottom panel) arterioles under control conditions and in the presence of 1 μmol/L phentolamine. Diameter change (% MAX) is defined in “Materials and Methods.” Resting diameters of these arterioles were 40.6±5.5 μm (2A) and 20.4±2.9 μm (3A); corresponding diameters with topical adenosine (0.1 mmol/L) were 61.2±5.8 and 30.7±4.4 μm, respectively. Figure 3. Bar graph showing the diameter of second-order (2A, open bars, n=12) and third-order (3A, hatched bars, n=10) arterioles adjacent to acetylcholine-filled micropipette tip (local) under control conditions (CNTL), during 16-Hz perivascular nerve stimulation, and during 10% O2 in the superfusion solution. These data were recorded 1028±45 and 998±37 μm distal to the origin of 2A and 3A arterioles, respectively; corresponding responses at the vessel origins were not different. Maximal diameters during dilation with adenosine were 50.2±2.9 and 33.0±2.5 μm for 2A and 3A branches, respectively, and did not differ from local responses to acetylcholine (1 mA, 500 ms). *Significantly different from CNTL (P<.05). Figure 4. Bar graphs showing local and conducted vasodilation to acetylcholine microiontophoresis (1 mA, 500 ms) in second-order (2A, open bars) and third-order (3A, hatched bars) arterioles. Corresponding diameters and n values are reported in Fig 3. Top, Local diameter changes under control conditions (CNTL), during 16-Hz perivascular nerve stimulation, and during 10% O2 in the superfusion solution. Each diameter increase was statistically significant (P<.05). *Significantly different from CNTL (P<.05). Bottom, Magnitude of conducted vasodilation (ie, diameter change) at the origin of arterioles. Each diameter increase was significant (P<.05). *Significantly different from CNTL and from 10% O2 (P<.05). Figure 5. Bar graph showing the magnitude of conducted vasodilation at the origin of second-order (2A) arterioles under control conditions (CNTL) and in the presence and absence of tetrodotoxin (TTX, 1 μmol/L) or phentolamine (PHEN, 1 μmol/L); local responses to acetylcholine were not affected by these treatments (data not shown). Acetylcholine-filled micropipettes were positioned 956±38 and 1024±47 μm from the vessel origin for PHEN and TTX experiments, respectively. Corresponding n values and diameters at rest (baseline) and during perivascular nerve stimulation (PNS) are shown in the Table. *Significantly different between rest and PNS (P<.05). #Significantly different between TTX or PHEN and corresponding CNTL response (P<.05).
This study was supported by National Institutes of Health grant R29-HL-41026 (Dr Segal) and by predoctoral and postdoctoral fellowships (Dr Kurjiaka) from the American Heart Association, Pennsylvania and Connecticut Affiliates, Inc, and was performed during the tenure of an Established Investigatorship Award (Dr Segal) from the American Heart Association and Genentech, Inc. FootnotesReferences
Page 2Epidemiological studies have suggested that patients with left ventricular hypertrophy (LVH) secondary to systemic hypertension are at a significantly greater risk of sudden cardiac death (SCD).12 In general, SCD is considered to be caused by acute cardiac arrest secondary to lethal ventricular tachyarrhythmias such as ventricular tachycardia (VT) and ventricular fibrillation (Vf). Clinical experiences have shown that patients are most susceptible to lethal ventricular tachyarrhythmias in the acute phase of myocardial infarctions, particularly within 90 minutes from the onset.3 Thus, a possible cause of the high incidence of SCD in patients with LVH may be a greater susceptibility to VT/Vf during acute ischemia in addition to a high incidence of ischemic cardiac events. Experimental studies have revealed that the incidence of SCD in LVH dogs after coronary artery occlusion and the occurrence of Vf in Langendorff-perfused hypertrophied rat hearts after regional ischemia are significantly higher.45 Therefore, we can expect to reduce the incidence of SCD in patients with LVH if we prevent the occurrence of lethal ventricular tachyarrhythmias associated with acute myocardial infarction. It has been reported that hypertensive LVH regresses after chronic treatment with antihypertensive agents such as angiotensin-converting enzyme (ACE) inhibitors, calcium antagonists, centrally acting sympathetic inhibitors, β-blockers, and α-blockers.67 However, it remains unknown whether a reduction of LVH will also reduce the excessive risk of SCD. In the present study, to test the hypothesis that regression of LVH prevents SCD, we examined the incidence of lethal ventricular tachyarrhythmias induced by acute ischemia in hypertrophied rat hearts and the effect of regression of LVH caused by chronic antihypertensive treatment on ischemia-induced lethal arrhythmias. In addition, we examined a possible mechanism underlying the enhanced arrhythmogenesis in hypertrophied hearts during acute ischemia by electrophysiological study using microelectrode techniques. Materials and MethodsExperimental AnimalsTwelve-week-old male spontaneously hypertensive rats (SHR), which were reported to show LVH,8 were chronically treated with antihypertensive agents for 6 weeks. These 18-week-old rats were then used for the experiments. In each treated group, drugs were withdrawn ≈24 hours before the experiments in order to eliminate their pharmacological actions. Hearts from age-matched SHR without treatment and Wistar-Kyoto (WKY) rats served as hypertensive hypertrophied hearts and normotensive hearts, respectively. Experimental animals were classified into the following five groups: (1) SHR without treatment (SHR-N), (2) SHR chronically treated with captopril (SHR-C), (3) SHR chronically treated with the angiotensin II receptor antagonist TCV-116 (SHR-A), (4) SHR chronically treated with hydralazine (SHR-H), and (5) WKY rats. Antihypertensive TreatmentDuring the entire 6-week treatment period, SHR-C, SHR-H, and SHR-A received captopril (2 g/L), hydralazine (12 mg/kg per day) in drinking water, or TCV-116 (30 mg/kg per day) mixed with the chow, respectively. The doses of captopril and hydralazine were determined from previous reports.910 To examine the dose-dependent effect on both blood pressure and regression of LVH, in addition to SHR-A (receiving TCV-116 at 30 mg/kg per day), two additional SHR groups received antihypertensive treatment with TCV-116 at doses of 1 mg/kg per day (A1 group) and 10 mg/kg per day (A10 group). Moreover, another group of SHR received a 1-week treatment (from 12 to 13 weeks of age) with TCV-116 at a dose of 30 mg/kg per day to dissociate the antihypertensive effect from LVH regression; this short-term treatment group was classified as SHR-AS. Measurement of Blood PressureAt 12 (before treatment), 14, 16, and 18 weeks of age (during treatment), SHR blood pressure was measured at the same time of day by the tail-cuff method. WKY blood pressure was measured at 12 and 18 weeks of age, and that of SHR-AS was measured at 12 and 13 weeks of age. Langendorff-Perfused Heart StudiesAnimals 18 weeks old were weighed and anesthetized with diethyl ether. After thoracotomy, the hearts were quickly excised and placed in oxygenated Tyrode’s solution. Each heart was mounted on a Langendorff perfusion apparatus by cannulation of the aorta and retrogradely perfused with Tyrode’s solution from a reservoir at a constant pressure of 70 cm H2O in WKY rats and of 100 cm H2O in SHR-N because of differences in blood pressure. Since the blood pressures in the SHR-C, SHR-H, SHR-A, and SHR-AS treatment groups were all lowered to the normal (WKY) level, the perfusion pressure used in these groups was 70 cm H2O. The perfusion pressures used in the A1 and A10 groups were 100 and 85 cm H2O because the blood pressures in both groups were not sufficiently lowered to the normal level. In a part of the experiments, SHR-N hearts were perfused at 70 cm H2O and 130 cm H2O. Flow was measured by collecting the coronary efflux for 1 minute. Wet weight of the heart was measured after each experiment. The composition of Tyrode’s solution was (mmol/L) NaCl 129, KCl 4, MgCl2 0.5, NaH2PO4 1.8, CaCl2 2.7, NaHCO3 20, and glucose 5.5. The perfusate was equilibrated with 95% O2 and 5% CO2 (pH 7.30, 37°C). An ECG was recorded with fine silver wire electrodes, one implanted in the right ventricle and the other fixed to the iron aortic cannula. The ECG was continuously monitored and recorded on a polygraph (model CM-616G, Nihon Kohden). After a 30-minute equilibration period, regional ischemia was produced by ligation of the left anterior coronary artery; the subsequent study period was also 30 minutes. Spontaneous arrhythmias were recorded in 19 WKY rats, 19 SHR-N, 11 SHR-C, 11 SHR-H, 10 SHR-A, 9 A1, 10 A10, and 9 SHR-AS during sinus rhythm. Heart rate and arrhythmias were analyzed from the ECG recordings. Arrhythmias were defined as follows: Vf, rapid ventricular depolarizations with irregular and varying morphology that persisted for >5 seconds; VT, consecutive premature complexes lasting >10 seconds, whose morphology differed from that during sinus rhythm. The risk area was measured in 11 WKY rats, 10 SHR-N, 8 SHR-C, 9 SHR-H, and 9 SHR-A. At the end of each experiment, Evans blue dye (Sigma Chemical Co) was added to the perfusate. After the removal of the atria and right ventricle, the left ventricle was cut transversely into four slices of approximately equal thickness. Each slice was photographed with magnification, and the perfused area and nonperfused area (risk area) were traced onto a transparent sheet and measured with computerized planimetry. The percentage of the risk area was then calculated. Measurement of Action PotentialConventional microelectrode techniques were used to record transmembrane action potentials in 10 WKY rats, 11 SHR-N, 10 SHR-C, 7 SHR-H, and 7 SHR-A. After anesthesia with diethyl ether, the hearts were quickly removed, and the papillary muscles of the left ventricle were carefully dissected. The preparation was transferred to a tissue chamber of 5-mL volume and superfused at a rate of 10 mL/min with a Krebs-Henseleit solution of the following composition (mmol/L): NaCl 119, KCl 4.8, CaCl2 2.5, MgSO4 1.2, KH2PO4 1.2, NaHCO3 24.9, and glucose 10.0. The solution was gassed with 95% O2 and 5% CO2 and kept at a temperature of 33±1.0°C. One end of the muscle was hooked to the lever arm, and the other end was pinned to the bottom of the tissue chamber. The muscle was maintained under a constant resting tension of 0.2 g. The preparations were stimulated at a rate of 0.5 Hz through platinum field electrodes. Stimuli were rectangular pulses of 1-millisecond duration at twice the diastolic threshold, delivered from an electronic stimulator (model SEN-6100, Nihon Kohden) through an isolation unit (model SS-302J, Nihon Kohden). Transmembrane potentials were recorded by using glass microelectrodes filled with 3 mol/L KCl (tip resistance, 10 to 30 MΩ). The microelectrode was coupled via a Ag/AgCl junction to a high-impedance-capacitance neutralizing amplifier (model MEZ8201, Nihon Kohden). An agar bridge containing 3 mol/L KCl was used as a reference electrode. An electronic differentiator whose output was linear from 50 to 800 V/s was used to measure the maximum rate of rise (V̇max) of action potentials. These amplified signals were displayed on an oscilloscope (model VC-9, Nihon Kohden), photographed on 35-mm film, and recorded on a chart recorder (model WR3101, Graphtec). After an equilibration period of 90 to 120 minutes, control recordings of action potentials were made. Control measurements were obtained 30 minutes after stable recordings were achieved. The perfusate was then switched to a hypoxia/no-glucose solution (Krebs-Henseleit solution gassed with 95% N2 and 5% CO2 and containing no glucose). The Po2 of the hypoxia/no-glucose solution was 40 to 50 mm Hg compared with >500 mm Hg in the control solution. During superfusion with the hypoxia/no-glucose solution for 30 minutes, recordings of action potentials were made. Statistical AnalysisAll data are expressed as mean±SEM. When multiple subgroups were present, the P value for comparisons was adjusted by the Bonferroni method. Differences between the mean values of multiple subgroups were evaluated by ANOVA, and intergroup comparisons were performed by the adjusted t test within ANOVA (Bonferroni method). The χ2 test was used to compare the difference in the incidence of arrhythmias. Significance was established at P<.05. ResultsAntihypertensive Effects and Regression of LVHFig 1 shows changes in systolic blood pressures (SBP) over 6 weeks in SHR with or without antihypertensive treatment. SBP in WKY rats was 102±3 mm Hg at 12 weeks of age and 114±4 mm Hg at 18 weeks of age. SBP in SHR-N 12 weeks old was 186±6 mm Hg and was higher than that in WKY rats (P<.005 versus WKY rats). SBP in SHR-N increased more thereafter and reached 205±4 mm Hg at 18 weeks of age. SBP values for SHR-C, SHR-A, and SHR-H were 186±4, 194±4, and 183±6 mm Hg at the age of 12 weeks (before drugs were given). SBP in each treated group was equally lowered 2 weeks after treatment and showed no differences between subgroups before and during treatment. At 18 weeks of age, SBP was 108±4 mm Hg in SHR-C, 117±3 mm Hg in SHR-A, and 111±3 mm Hg in SHR-H and did not significantly differ from 114±4 mm Hg in age-matched WKY rats (Fig 1). Table 1 shows left ventricular weight (LVW), body weight (BW), and the ratio of LVW to BW (LVW/BW) in each group at 18 weeks of age. LVW and LVW/BW values for SHR-N and SHR-H were significantly higher than those for WKY rats (P<.005 versus WKY rats). These results indicate that hearts from SHR-N and SHR-H showed LVH and that treatment with hydralazine failed to cause LVH to regress sufficiently despite equally lowered blood pressure. LVW and LVW/BW values for SHR-C and SHR-A were lower than those for SHR-N (P<.005 versus SHR-N). Thus, SHR-C and SHR-A showed regression of LVH. Ischemia-Induced Arrhythmias in Langendorff-Perfused HeartsIschemia-induced arrhythmias during the first 30 minutes after left coronary artery ligation are illustrated in Fig 2, and the incidence of VT/Vf is summarized in Fig 3. The number of occurrences of VT and Vf per 10 rats is also summarized in Table 2. Most lethal arrhythmias in each group occurred 10 to 15 minutes after left coronary artery occlusion. Although premature complexes, bigeminy, or nonsustained VT sometimes occurred in WKY hearts, VT occurred in only 1 of 19 WKY hearts, and Vf never occurred. VT/Vf occurred frequently (63%) in SHR-N hearts with LVH. For SHR-N, the ratio of VT per 10 hearts was 5.3, and that of Vf per 10 hearts was 3.7. Since the performance of non–blood-perfused hearts may be highly dependent on coronary perfusion pressure, ischemia-induced arrhythmias in the SHR-N hearts were examined at two different coronary perfusion pressures of 70 cm H2O and 130 cm H2O in addition to coronary perfusion pressure of 100 cm H2O. VT/Vf occurred in 4 of 5 SHR-N hearts (80%) at a perfusion pressure of 130 cm H2O and in 4 of 4 SHR-N hearts (100%) at a perfusion pressure of 70 cm H2O. These results indicate that the SHR-N hearts are susceptible to ischemia-induced VT/Vf at various perfusion pressures in this non–blood-perfused preparation. In SHR-C and SHR-A hearts, which showed regression of LVH, arrhythmias (such as premature complexes, bigeminy, or nonsustained VT) sometimes occurred as in WKY hearts, but lethal arrhythmias such as VT/Vf rarely occurred. In SHR-H hearts, which did not show regression of LVH, the occurrence of lethal arrhythmias could not be prevented. The incidence of VT/Vf was high (45%) in SHR-H hearts (Fig 3, Table 2). Heart rate and coronary efflux in each group showed no differences before coronary occlusion (Table 3). Although heart rate and coronary efflux decreased after coronary occlusion, no differences in their extent were seen in the SHR groups. There was also no significant difference in risk area among groups, as shown in Table 3. Dose-Dependent Effect of Angiotensin II Receptor AntagonistSince the antihypertensive treatment with captopril (SHR-C) and TCV-116 (SHR-A) equally caused LVH to regress and reduced the incidence of ischemia-induced VT/Vf, it was suggested that the common pathway of these agents, ie, blocking of the effect of angiotensin II (Ang II), may be very important. In addition, results from Table 1 and Fig 3 indicate that enhanced arrhythmogenesis during acute ischemia in SHR-N might be prevented not by lowering blood pressure but by regression of LVH. In other words, regression of LVH appears to be more important than normalization of blood pressure. To test this hypothesis, we examined the effect of lower doses of TCV-116 (A1 and A10, 1 and 10 mg/kg per day, respectively) and compared that with results from SHR-N and SHR-A (TCV-116, 30 mg/kg per day). Furthermore, SHR were treated for 1 week (from 12 to 13 weeks of age) with TCV-116 at a dose of 30 mg/kg per day (SHR-AS), which would be expected to cause reduction in blood pressure and little regression of LVH. Fig 4 shows changes in SBP in SHR treated with TCV-116 at doses of 1, 10, and 30 mg/kg per day (A1, A10, and SHR-A, respectively) and in SHR without treatment (SHR-N). During treatment for 6 weeks, SBP in A1 was only a little lower than that in SHR-N. SBP in A10 was significantly lower than that in SHR-N during the course of treatment, although the extent of decrease in A10 was not as great as in SHR-A. SBP in SHR-AS was 194±5 mm Hg at the age of 12 weeks (before the drug treatment) and was reduced to 119±5 mm Hg after 1 week of treatment. This value was almost the same as the SBP in SHR-A at the age of 14 to 18 weeks. LVW/BW in each group at 18 weeks of age and in SHR-AS at 13 weeks of age is shown in Fig 5A. As shown in Fig 5A, TCV-116 caused LVH to regress in a dose-dependent manner. The incidence of VT/Vf (Fig 5B) and the occurrence ratios per 10 hearts for VT/Vf (Table 2) during the first 30 minutes after left coronary artery ligation in SHR treated with TCV-116 were parallel with the degree of LVH regression. Treatment with TCV-116 reduced the incidence of VT/Vf in a dose-dependent manner. However, in SHR-AS with short-term treatment, little regression of LVH and a high incidence of VT/Vf were observed. LVW/BW in SHR-AS was 3.04±0.05 mg/g (n=9) and was not significantly different from 3.06±0.08 mg/g in SHR without treatment at 13 weeks of age (n=6). Again, there were no significant differences in heart rate before coronary occlusion and coronary efflux before and after coronary occlusion among the subgroups. Action Potential Duration in Normal and Ischemia-Simulating ConditionsOriginal chart recordings of action potential and action potential duration (APD) at −50 mV (APD−50 mV) during 30 minutes of superfusion with hypoxia/no-glucose solution in the left ventricular papillary muscle of SHR-N are shown in Fig 6A. Resting membrane potential (RMP), action potential amplitude (APA), and APD began to decrease within a few minutes after exposure to the hypoxia/no-glucose solution. The actual recordings of the action potentials recorded before and 30 minutes after exposure to the hypoxia/no-glucose solution are shown in Fig 6B. Table 4 summarizes the action potential characteristics before and after exposure to the hypoxia/no-glucose solution. In the control condition, APDs in SHR-N were significantly longer than those in WKY rats, although there were no significant changes in other action potential parameters among subgroups. APDs in SHR-H, which failed to show regression of LVH, were as long as those in SHR-N. However, APDs in the LVH regression groups (SHR-C and SHR-A) showed no differences from those in the WKY group. After exposure to the hypoxia/no-glucose solution, APDs in each group shortened almost to the same level, and there were no significant differences in absolute values of APDs at 20% and 75% repolarization (APD20 and APD75, respectively) among the subgroups (Table 4). In other words, the extent of APD shortening during superfusion with the hypoxia/no-glucose solution in the LVH groups was significantly greater than that in the WKY group, as shown in Fig 7. Shortening for both APD20 and APD75 was greater in the LVH groups than in the WKY group during superfusion of the hypoxia/no-glucose solution from 5 to 30 minutes, whereas shortening in the regression groups was almost the same as that in the WKY group during superfusion of the hypoxia/no-glucose solution. There were no differences in changes of other action potential parameters including RMP, APA, and V̇max during superfusion of the hypoxia/no-glucose solution among subgroups (Table 4). DiscussionThe present study in the rat heart model demonstrates the following: (1) Hypertrophied hearts show a greater susceptibility to VT/Vf during acute myocardial ischemia because of greater APD dispersion between the normal and ischemic zones. (2) The reduction of electrical inhomogeneity in regressed hypertrophied hearts may prevent lethal ventricular tachyarrhythmias. (3) Blockade of the effect of Ang II plays an important role not only in the regression of LVH but also in the prevention of SCD. We should extrapolate these findings to the clinical setting with great caution, because there are the species differences in cardiac electrophysiological properties. However, our data lead us to suggest that regression of LVH can reverse enhanced arrhythmogenesis in LVH during acute myocardial ischemia and that chronic inhibition of the effect of Ang II in patients with LVH may prevent SCD caused by ischemia-induced lethal tachyarrhythmias. The susceptibility to ischemia-induced lethal ventricular tachyarrhythmias in SHR with LVH is consistent with the previous studies of hypertrophied rat hearts induced by pressure overload.5 These experimental studies indicate that in hypertensive LVH the risk of SCD caused by VT/Vf during the acute phase of myocardial infarction is very high. The findings are consistent with epidemiological data12 showing that the incidence of SCD in hypertensive patients with LVH is significantly higher than in those without LVH. Chronic treatment with captopril and the Ang II receptor antagonist TCV-116 caused LVH to regress in SHR and prevented ischemia-induced VT/Vf, although hydralazine failed to reduce the incidence of VT/Vf because LVH remained, despite equal reduction of blood pressure. These results support the idea that regression of LVH is very important in preventing lethal tachyarrhythmias during the acute phase of myocardial infarction and that it is necessary to use drugs with a highly potent effect on regression of LVH in treating hypertensive patients with LVH. Despite the equally lowered blood pressure, there was a difference in regressive effects achieved by captopril, TCV-116, and hydralazine. This means that in addition to blood pressure, other important factors may be involved in cardiac hypertrophy11 and its regression. It was recently shown that cardiac hypertrophy of aortic-constricted rats might be related to increased activity of the renin-angiotensin system and that the direct cardiac actions of Ang II included acceleration of protein synthesis that resulted in cardiac hypertrophy.1213 These previous reports strongly suggest that Ang II plays an important role in cardiac hypertrophy. In addition, there is increasing evidence indicating the existence of the tissue (local) renin-angiotensin system in several organs, including the heart,1415 and molecular biological measurements suggest that Ang II, particularly mediated by the type-1 Ang II (AT1) receptor, causes hypertrophy of cardiac myocytes.16 It was reported that due to chronic aortic banding, LVH rats showed regression of LVH with a nonantihypertensive low dose of the ACE inhibitor ramipril but did not show any significant decrease in left ventricular mass with the calcium antagonist nifedipine and the arterial vasodilator hydralazine, despite effective reduction in blood pressure.17 Dzau18 also suggests that ACE inhibitors have a potent effect on the decrease in left ventricular mass in hypertensive LVH. In the present study, to evaluate the effect of Ang II blockade on regression of LVH and prevention of SCD, we used captopril, an ACE inhibitor, and TCV-116, an AT1 receptor–specific antagonist. These two drugs effectively caused LVH to regress and reduced the incidence of VT/Vf during acute ischemia. Moreover, low doses of TCV-116 (1 mg/kg per day) inhibited the progression of LVH with only a slight antihypertensive effect. On the other hand, treatment with hydralazine failed to cause LVH to regress sufficiently, possibly because of augmented sympathetic nerve activity and sodium retention.19 Thus, blocking the effect of Ang II may play a beneficial role not only in the regression of LVH but also in the prevention of SCD. The main electrophysiological feature of experimental LVH induced by aortic banding or renal hypertension in rats is prolongation of APD at the plateau level.520 Ventricular myocytes from felines with right ventricular hypertrophy induced by pulmonary artery binding showed that the time course of inactivation of the Ca2+ current was delayed and that the magnitude of the delayed rectifier K+ current was reduced with slower activation and enhanced deactivation.2122 In feline left ventricular myocytes with hypertrophy induced by long-lasting pressure overload, the newly expressed T-type Ca2+ current23 and the long-lasting opening of the L-type Ca2+ channels24 have been described. These reports explain the prolongation of APD in felines with ventricular hypertrophy. However, we cannot simply extrapolate these ionic mechanisms to APD prolongation in hypertrophied rat hearts because there are some differences of the membrane current system between rat and feline heart cells. The transient outward current (Ito) plays an important role in the repolarization of rat ventricular cells.25 In rats with LVH, diminished Ito due to reduced functional channel density, which may result in the prolongation of APD, has been reported.2627 However, it has also been suggested that the functional expression of Ito is enhanced in hypertrophied feline right ventricular myocytes.28 The L-type Ca2+ current density in hypertrophied rat left ventricular myocytes has been reported to be unchanged29 or increased.30 Further studies are needed to determine electrophysiological alterations in the hypertrophied rat heart. Mechanical overload and humoral factors associated with cardiac hypertrophy were reported to promote protein synthesis not only quantitatively but also qualitatively, as an adaptive response.31 Rat cardiac muscles in the neonatal phase, when the thickness of the ventricular wall increases prominently, showed prolongation of APD.32 Thus, the ion channels or their modulatory mechanisms may be changed in cardiac myocytes in both hypertrophied hearts and neonatal hearts as a kind of adaptive response. Treatment with hydralazine did not cause LVH to regress sufficiently, and prolongation of APD remained. However, treatment with captopril or TCV-116 reversed LVH, and APD in these regression groups was not prolonged and did not differ from that in the WKY group. These results suggest that prolongation of APD was normalized by removal of both the mechanical overload and humoral factors, such as Ang II, and that a close relation exists between hypertrophied cardiac myocytes and prolongation of APD. The mechanism of lethal ventricular tachyarrhythmias during the acute phase of myocardial infarction is considered to be reentry, although abnormal automaticity arising from depolarized Purkinje fibers may be included.3334 In an inhomogeneous tissue affected by acute ischemia, in which the conduction velocity and duration of refractoriness greatly differ in various areas of myocardium, reentry can easily occur.35 In LVH groups such as SHR without treatment or those treated with hydralazine, APD was more prolonged than in WKY rats in the control condition and showed a greater shortening during superfusion of the hypoxia/no-glucose solution. These results indicate greater APD dispersion between the normal and ischemic zones in the hypertrophied myocardium. Greater APD dispersion may reflect enhanced dispersion of refractoriness and then cause an augmented injury current not only between the normal and ischemic zones but also among myocardial cells within the risk area because of the wave-front phenomenon.36 The previous report comparing SHR and WKY rats described that the action potential changes and the alterations of conduction and refractoriness were more prominent in hypertrophied than in normal endocardial tissue during simulated ischemia.37 These results also agree with our data showing the enhanced shortening of APD during ischemia in hypertrophied myocardium. Therefore, reentrant arrhythmias might occur easily in hypertrophied hearts. In regression groups such as SHR chronically treated with captopril or TCV-116, APD values varied little from those found in the WKY group and shortened slightly during superfusion with the hypoxia/no-glucose solution. At the same time, the incidence of VT/Vf was significantly reduced. It is suggested that the electrical inhomogeneity and enhanced arrhythmogenesis in LVH was reversed by the regression of LVH. The reduction of electrical inhomogeneity during acute ischemia in regressed LVH may lessen the likelihood of SCD resulting from ischemia-induced lethal arrhythmias. The enhanced shortening of APD during ischemia in hypertrophied hearts may be attributed to greater changes of the membrane currents responsible for ischemia-induced APD shortening. Experiments in hypertrophied feline left ventricular myocytes using the patch-clamp method revealed that the open-state probability of the ATP-sensitive K+ channel was significantly higher at various pH levels and depleted ATP conditions3839 and that the magnitude of the Ca2+ current was significantly reduced during metabolic inhibition induced by CN+.40 Therefore, depletion of intracellular ATP in hypertrophied myocytes might induce greater activation of ATP-sensitive K+ current at any given level of ATP and greater inhibition of the Ca2+ current. These results may partly explain the enhanced shortening of APD during ischemia in hypertrophied cardiac cells. However, the ischemia-induced changes in other membrane currents in hypertrophied myocytes still remain unclarified. Further study to determine the ischemia-induced changes in membrane currents in hypertrophied hearts may help to prevent ischemia-induced lethal arrhythmias. Figure 1. Graph showing systolic blood pressure (SBP) over 6 weeks in all rat groups: SHR-N indicates spontaneously hypertensive rats (SHR) without treatment; SHR-C, SHR treated with captopril; SHR-A, SHR treated with angiotensin II receptor antagonist TCV-116; and SHR-H, SHR treated with hydralazine. Values are mean±SEM. Figure 2. Representative tracings showing examples of ischemia-induced arrhythmias in hearts of Wistar-Kyoto (WKY) rats, spontaneously hypertensive rats (SHR) without treatment (SHR-N), and SHR treated with captopril (SHR-C). Regional ischemia was produced by ligation of the left coronary artery, and ischemia-induced arrhythmias often occurred between the first 10 and 15 minutes. In WKY hearts, only single (as shown above) or paired premature complexes, bigeminy, and nonsustained (lasting <10 seconds) consecutive premature complexes occurred, and these instances were infrequent. In SHR-N hearts with LVH, ventricular tachycardia or ventricular fibrillation (as shown above) occurred often. In SHR-C hearts with regressed LVH, ventricular tachycardia or ventricular fibrillation never occurred, although other nonsustained events such as bigeminy (as shown above) occurred infrequently. Figure 3. Bar graph showing the incidence of ventricular tachycardia (VT) or ventricular fibrillation (Vf) during the 30-minute period after left coronary ligation. WKY indicates Wistar-Kyoto rats; SHR-N, spontaneously hypertensive rats (SHR) without treatment; SHR-C, SHR treated with captopril; SHR-A, SHR treated with angiotensin II receptor antagonist TCV-116; and SHR-H, SHR treated with hydralazine. *P<.05 and **P<.005 vs WKY. Figure 4. Graph showing systolic blood pressure (SBP) over 6 weeks in spontaneously hypertensive rats (SHR) treated with various doses of angiotensin II receptor antagonist TCV-116. SHR-N indicates SHR without treatment (n=19); A1, SHR treated with TCV-116 at 1 mg/kg per day (n=9); A10, SHR treated with TCV-116 at 10 mg/kg per day (n=10); and SHR-A, SHR treated with TCV-116 at 30 mg/kg per day (n=10). Values are mean±SEM. *P<.05 and **P<.005 vs SHR-N. Figure 5. Bar graphs showing the ratio of left ventricular weight to body weight (LVW/BW) (A) and the incidence of ventricular tachycardia (VT) or ventricular fibrillation (Vf) (B) in spontaneously hypertensive rats (SHR) treated with angiotensin II receptor antagonist TCV-116. SHR-N indicates SHR without treatment (n=19); A1, SHR treated with TCV-116 at 1 mg/kg per day (n=9); A10, SHR treated with TCV-116 at 10 mg/kg per day (n=10); SHR-A, SHR treated with TCV-116 at 30 mg/kg per day (n=10); and SHR-AS, SHR receiving short-term treatment with TCV-116 at 30 mg/kg per day (n=9). *P<.05 and **P<.005 vs SHR-N. Figure 6. Changes in action potentials of left ventricular papillary muscles during 30 minutes of superfusion with hypoxia/no-glucose solution. WKY indicates Wistar-Kyoto rats; SHR-N, spontaneously hypertensive rats (SHR) without treatment; SHR-C, SHR treated with captopril; SHR-A, SHR treated with angiotensin II receptor antagonist TCV-116 at 30 mg/kg per day; and SHR-H, SHR treated with hydralazine. A, Original chart recordings of action potentials (APs) and action potential duration (APD) at −50 mV during 30 minutes of superfusion with hypoxia/no-glucose solution in an SHR-N papillary muscle. The arrow on the left indicates the beginning of exposure to the hypoxia/no-glucose solution, and the arrow on the right indicates the return to the oxygenated normal solution. B, Actual recordings of APs before (control) and 30 minutes after exposure to the hypoxia/no-glucose solution (hypoxia 30 min). Tracings in each panel represent zero potential (top) and membrane potential (bottom). APs for the SHR-N preparation were obtained from the same experiment as shown in panel A. Figure 7. Graphs showing changes in action potential duration at 20% and 75% repolarization (APD20 and APD75, respectively) during 30 minutes of superfusion with the hypoxia/no-glucose solution. WKY indicates Wistar-Kyoto rats; SHR-N, spontaneously hypertensive rats (SHR) without treatment; SHR-C, SHR treated with captopril; SHR-A, SHR treated with angiotensin II receptor antagonist TCV-116 at 30 mg/kg per day; and SHR-H, SHR treated with hydralazine. Percent changes from control values (before hypoxia) are indicated. Numbers of experiments are as follows: WKY, 10; SHR-N, 11; SHR-C, 10; SHR-A, 7; and SHR-H, 7. Values are mean±SEM. *P<.05 vs WKY.
FootnotesReferences
Page 3
Abstract Nitric oxide (NO) produced within the lungs maintains pulmonary vascular homeostatic properties, modulating leukocyte traffic, platelet aggregation, and vasomotor tone. Because reactive oxygen intermediates generated during reperfusion react rapidly with available NO, we hypothesized that the NO donor nitroglycerin (NTG) would enhance lung preservation for transplantation by improving graft blood flow and reducing graft neutrophil and platelet sequestration. By use of an orthotopic rat left lung transplant model, with ligation of the native right pulmonary artery to ensure that recipient survival and physiological measurements depend entirely on the transplanted lung, transplants were performed in 70 male Lewis rats after 6-hour 4°C preservation in Euro-Collins solution (EC) alone or EC with supplemental NTG. Compared with EC alone, supplemental NTG significantly increased pulmonary arterial flow (2.2±1.4 to 21.4±2.9 mL/min, P<.01), decreased pulmonary vascular resistance (7.4±2.0 to 1.4±0.1×103 Woods units, P<.05), improved arterial oxygenation (163±57 to 501±31 mm Hg, P<.01), and enhanced recipient survival (17% to 100%, P<.001). These beneficial effects of NTG were dose dependent over a range of 0.001 to 0.1 mg/mL. Although NTG caused significant pulmonary vasodilation during the harvest/flushing period, the direct-acting vasodilator hydralazine caused greater vasodilation than did NTG but was associated with poor graft function, elevated pulmonary vascular resistance, and poor recipient survival. To explore nonvasodilator protective mechanisms of NTG, graft neutrophil and platelet sequestration were studied; supplemental NTG significantly reduced both neutrophil and platelet accumulation compared with either hydralazine or EC alone. These findings suggest that vasodilation alone at the time of harvest is insufficient to protect the lungs. NTG, which produces antineutrophil and antiplatelet effects as well as harvest vasodilation, appears to be a simple and effective additive that will improve lung preservation for transplantation. There has been a recent burgeoning of clinical lung transplantation,1 prompted in part by the development of improved pulmonary preservation solutions, but the lungs remain among the organs most vulnerable to ischemia and reperfusion injury during the transplantation process. The inability to preserve lungs beyond 4 to 6 hours is a major impediment for immunologic cross-matching and hampers efforts at multiple or distant organ procurement. Even with optimal preservation techniques, the perioperative morbidity and mortality remain high, with early graft failure characterized by elevated pulmonary vascular resistance (PVR), poor gas exchange, and neutrophil infiltration.2 Recent studies have demonstrated that maintaining endothelial function within cardiac grafts is critical to successful cardiac preservation,345 but the current gold-standard clinical lung preservation solution, modified Euro-Collins solution (EC), consists of a simple electrolyte solution without additives to specifically address the maintenance of endothelial function.2 Because the lungs are among the most richly vascularized of organs, we hypothesized that maintaining normal endothelial properties during preservation and transplantation of the lungs would be critical to the ultimate success of lung transplantation, especially after periods of prolonged preservation. Of the numerous factors that influence vascular function, endothelium-derived relaxing factor (EDRF, whose identity appears to be nitric oxide [NO]678 ) has emerged as a key modulator of normal pulmonary vascular physiology. In addition to preventing neutrophil adherence to the endothelium,9 maintaining endothelial barrier properties,10 and inhibiting platelet aggregation,5 NO has an important role in modulating pulmonary vascular tone. Endogenous pulmonary NO production participates in the physiological regulation of pulmonary vasomotor tone, as has been shown in animal models by the use of inhibitors of NO synthase,1112131415 although the degree of importance is affected by species and experimental conditions under which observations are made. Even in humans, NO can be identified in exhaled air16 and is thought to regulate basal PVR.17 Models of cardiac ischemia and reperfusion have demonstrated that both EDRF bioactivity18 and NO levels5 fall within minutes of reperfusion because of the quenching of NO by superoxide generated during reperfusion; this reaction is rapid, with a rate constant of 108 (mol/L)−1 · s−1, forming peroxynitrite in the process.19 Because reactive oxygen intermediates are formed in especial abundance in the pulmonary reperfusion microenvironment,220 we hypothesized that endothelium-dependent vascular homeostatic properties might be perturbed by the lack of available NO and that pulmonary preservation might be enhanced by nitroglycerin (NTG), an NO donor. In this study, we used a recently developed model of rat orthotopic lung transplantation in which the native right lung supports the animal during surgery but is effectively removed from the circulation after transplantation so that physiological measurements and recipient survival represent function of the transplanted lung.21 Experiments were designed to test (1) whether vasodilation at the time of harvest is sufficient to improve lung preservation and (2) whether NTG supplementation would enhance NO-related mechanisms of vascular homeostasis after lung transplantation. By use of inbred male Lewis rats (250 to 300 g), hemodynamic and oxygenation measurements were made in an orthotopic left lung transplant model.21 In brief, the donor rat was heparinized (500 U IV), and the superior venae cavae were ligated. Thirty milliliters of 4°C preservation solution was infused into the inferior vena cava at a constant pressure (20 mm Hg) and vented out the left atrium (LA). The time required to deliver the 30-mL volume of preservation solution at constant infusion pressure was recorded as an index of PVR during harvest. The left pulmonary artery (PA) and pulmonary vein (PV) were then divided, the bronchus was ligated and divided with the lung partially inflated, and the lung was removed. A 14-gauge cuff was placed on each vascular stump, a 16-gauge grooved cylinder was inserted into the bronchus, and the lung was submerged in 4°C preservation solution for 6 hours. The recipient rat was anesthetized and intubated (ventilated with 100% oxygen); a left thoracotomy was performed; the left bronchus, PA, and PV were isolated, cross-clamped, and divided; and the native lung was removed. The cylinder (bronchus) and cuffs (PA and PV) were connected to the appropriate recipient structures, maintaining warm ischemic times below 10 minutes. The hilar cross clamp was released, reestablishing blood flow and enabling gas exchange. A snare was then passed around the right PA, and 2F catheters (Millar Instruments) were introduced into the main PA and the LA. A flow probe (Transonic) was then placed around the main PA. Preservation solutions consisted of modified EC, EC with supplemental NTG (5 mg/mL, intravenous formulation, DuPont Merck Pharmaceuticals), or EC with supplemental hydralazine (CIBA-GEIGY Limited). EC solution was purchased from Baxter Healthcare and consisted of Na+ (10 mEq/L), K+ (115 mEq/L), Cl− (15 mEq/L), HPO4−2 (85 mEq/L), H2PO4− (15 mEq/L), and HCO3− (10 mEq/L), modified by adding magnesium sulfate (10 mL of 10% solution) and glucose (50 mL of 50% solution) to each liter before use. After flushing the lungs with hypothermic preservation solutions as described, harvested lungs were preserved for 6 hours at 4°C in 50 mL of preservation solution with a composition identical to that used during harvest. On-line hemodynamic monitoring was accomplished by using a MacLab and a Macintosh IIci computer. The hemodynamic parameters that were measured included LA pressure (in millimeters of mercury), PA pressure (in millimeters of mercury), and PA blood flow (in milliliters per minute). Arterial oxygen tension (in millimeters of mercury) was measured during inspiration of 100% oxygen; Po2 was analyzed with a model ABL-2 gas analyzer (Radiometer). PVRs were calculated as follows: (mean PA pressure−LA pressure)/PA flow, expressed as Woods units×103. After baseline measurements were taken, the native (right) PA was ligated, and serial measurements were taken every 5 minutes until the time of euthanasia at 30 minutes (or until recipient death, if it preceded the 30-minute time point). Recipient death was identified by cessation of cardiac mechanical activity as viewed through the open thorax. To determine whether longer survival was possible in lungs preserved with supplemental NTG (0.1 mg/mL), for certain experiments, observation was continued until 8 hours after surgery without hemodynamic measurements, and death was identified as described above. Thirty minutes after ligation of the native right PA or at the time of recipient death, transplanted lungs were removed, rinsed briskly in physiological saline, and snap-frozen in liquid nitrogen until the time of myeloperoxidase assay. Tissue was homogenized in phosphate buffer (50 mmol/L, pH 5.5, 5 mL/g of tissue) containing hexadecyltrimethylammonium bromide (0.5%, Sigma Chemical Co). The assay was performed, as previously described,22 by thawing the sample, centrifuging at 40 000g for 15 minutes, and decanting the supernatant, which was assayed for myeloperoxidase activity by using a standard chromogenic spectrophotometric technique in which test sample (0.03 mL) was added to phosphate buffer (0.97 mL) containing o-dianisidine dihydrochloride (Sigma) and hydrogen peroxide (0.0005%) and change in absorbance at 460 nm was measured over 1 minute (increase in optical density was linear over this time interval). Graft platelet accumulation was determined by using 111In-labeled platelets, prepared as previously described.23 Blood (5.0 mL) was taken from a gender/strain-matched donor rat and heparinized (2500 U). Platelets were isolated by differential centrifugation, first at 300g for 5 minutes to obtain platelet-rich plasma, which was then washed three times at 2000g for 15 minutes in 10 mL of acid/citrate/dextrose anticoagulant (ACD-A, containing 38 mmol/L citric acid, 75 mmol/L sodium citrate, and 135 mmol/L glucose). The pellet was suspended in 5 mL of ACD-A and centrifuged at 100g for 5 minutes to remove contaminating red blood cells, and the supernatant was collected. 111In oxyquinoline (70 μL of 1 mCi/mL, Amersham Mediphysics) was added with gentle shaking for 30 minutes at room temperature. The radiolabeled platelets were washed three times in ACD-A and resuspended in PBS, and platelet number was adjusted to 5×107/mL. After completion of the vascular and bronchial anastomoses, 0.5 mL of 111In-labeled platelet suspension was injected intravenously into the recipient. One minute after platelet infusion (immediately before reperfusion), 0.5 mL of blood was taken from the LA to determine blood radioactivity to ascertain blood platelet concentrations to normalize for variations in blood loss during surgery. Five minutes after reperfusion, the native right PA was ligated. The graft was removed 10 minutes thereafter, and 111In-labeled platelet deposition was quantified by gamma counting. Platelet accumulation was expressed as the ratio of graft radioactivity to blood radioactivity normalized to dry weight. Platelet function was measured by evaluating bleeding time in pulmonary transplant recipients as previously described24 ; immediately after reperfusion, the rat tail was transected 5 mm proximal to the tip, and every 30 seconds, blood drops were collected onto a piece of filter paper placed tangentially to the tail. Bleeding time was recorded as the time from the initial transection to the cessation of bleeding. Recipient survival data were evaluated by using Fisher’s exact test. All other data were evaluated by using the Mann-Whitney U test. Because survival assessment was measured independent of hemodynamic measurements, all survival data are included in “Results,” even when equipment malfunction precluded obtaining hemodynamic measurements. Values are expressed as mean±SEM, with differences considered statistically significant at P<.05. To determine whether supplementation of the clinical standard pulmonary preservation solution with NTG would enhance pulmonary vascular function and improve recipient survival after lung transplantation, experiments were performed by using EC as the base solution. A preservation duration of 6 hours was chosen for these experiments on the basis of pilot studies demonstrating significant graft failure with EC alone at this preservation duration21 and the clinical relevance of 4 hours as the upper limit of acceptable pulmonary preservation in humans. At 6 hours of preservation in EC alone, graft failure occurred rapidly after pulmonary reperfusion, with marked increases in PVR accompanied by declines in PA flow and arterial oxygenation (Fig 1A). Although PA pressure rose initially upon ligation of the native (right) PA, this was followed by a rapid decline in PA pressure, followed by recipient death. In sharp contrast, when NTG was added to the preservation solution, hemodynamic and functional (arterial oxygenation) parameters were stabilized, and the recipient survived the 30-minute observation period (Fig 1B). Compared with the control preservation solution (EC alone), NTG (0.1 mg/mL) added to the preservation solution increased PA flow (2.2±1.4 versus 21.4±2.9 mL/min, P<.01) (Fig 2A), decreased PVR (7.4±2.0 versus 1.4±0.1×103 Woods units, P<.05) (Fig 2B), and improved arterial oxygenation (163±57 versus 501±31 mm Hg, P<.01) (Fig 2C). To ensure that these beneficial effects of NTG were due to NTG itself rather than the diluent found in the intravenous formulation that we used to prepare our preservation solution, EC was supplemented with an equivalent amount of diluent, and the effects on graft preservation were measured (n=3); for these experiments, preservation was not different from that found with the control solution (PVR was 6.1±3.3×103 Woods units, and arterial oxygenation was 127±47 mm Hg; P=NS versus control solution). Recipient survival was also improved significantly by supplementation of the preservation solution with NTG compared with the control solution (17% versus 100%, respectively; P<.001). The beneficial effects of NTG were dose dependent over a range of 0.001 to 0.1 mg/mL, with maximal beneficial effect obtained at 0.1 mg/mL (Fig 3). To determine whether recipients of grafts preserved with supplemental NTG (0.1 mg/mL) could survive beyond the 30-minute time point, at which we routinely assess survival and graft hemodynamics, six additional experiments were performed in which recipients were observed after the transplantation procedure for up to 8 hours after transplantation and ligation of the native PA. All six recipients survived beyond 30 minutes. One animal died at 43 minutes for unclear reasons, one died at 4 hours and one died at 6 hours because of bleeding complications at the anastomotic sites that could not be controlled, and three animals survived until they were euthanized at ≈8 hours. These empirical studies suggest that the benefit we observed in the NTG group at 30 minutes is likely to continue well beyond the 30-minute observation period. Because it has been suggested that vasodilators enhance pulmonary preservation by lowering PVR during harvest,225 resulting in more rapid and effective distribution of hypothermic preservation solution, we determined the flushing time required to deliver identical volumes of preservation solution at identical flushing pressure as a reflection of PVR during harvest. These experiments demonstrated that NTG did indeed lower PVR during flushing, resulting in more rapid preservation than in its absence (86.4±4.9 versus 192.7±9.6 seconds, P<.01 respectively; Fig 4). However, harvest vasodilation alone was insufficient to enhance pulmonary preservation, as hydralazine (0.02 μg/mL, a direct-acting vasodilator26 ) was even more effective at harvest pulmonary vasodilation than was NTG (flush time, 69.3±4.5 seconds; P<.05 versus NTG, P<.005 versus EC alone) but relatively ineffective at enhancing pulmonary preservation for transplantation (Figs 2 and 3). Grafts preserved with hydralazine (0.02 μg/mL) demonstrated poor function (reduced arterial oxygenation), poor PA blood flow, elevated PVR, and poor recipient survival (Figs 2 and 3). These data suggested that the beneficial effects of NTG were not exclusively due to its actions as a vasodilator. Because NTG inhibits platelet aggregation27 and neutrophil adherence to the reperfused coronary endothelium5 and because neutrophil aggregation and platelet plugging have been implicated in the no-reflow phenomenon,2728 we evaluated the effects of NTG added to the preservation solution on graft neutrophil and platelet accumulation after reperfusion. NTG (0.1 mg/mL) added to the preservation solution was associated with a significant decline in both neutrophil and platelet accumulation in the reperfused grafts, as quantified by graft myeloperoxidase activity (Fig 5A) and 111In-labeled platelet deposition (Fig 5B). In contrast, the addition of hydralazine to the pulmonary preservation solution did not alter the platelet or neutrophil deposition compared with the control solution (Fig 5). It is likely that these actions of NTG added to the pulmonary preservation solution are occurring within the confines of the transplanted lung and are not secondary to systemic effects, because measurements of bleeding times did not differ between animals transplanted with control or NTG-preserved lungs (Fig 6). Vascular endothelium plays a cardinal role in maintaining a homeostatic milieu both within blood vessels and the tissues they supply. NO serves as an important signaling molecule to reduce vasomotor tone of the subjacent vascular smooth muscle,6 maintain endothelial barrier properties,10 reduce platelet aggregation,5 and inhibit neutrophil adherence to the endothelial surface.9 After a period of ischemia, these physiological endothelial properties are perturbed, favoring vasoconstriction, thrombosis, and neutrophil adhesion, as the inflammatory response is activated. NO levels plummet after endothelial cell exposure to hypoxia and reoxygenation because of the rapid production of superoxide during reoxygenation.5 Because the loss of available NO may contribute to endothelial dysfunction and pulmonary graft failure during the immediate posttransplant period, the experiments presented here confirmed the hypothesis that an NO donor such as NTG would enhance NO-related mechanisms of vascular homeostasis within the pulmonary graft. Furthermore, these experiments used hydralazine to demonstrate that vasodilation alone at the time of harvest is insufficient to protect the lungs but that reducing neutrophil and platelet sequestration into the reperfused graft is important. NO produced within the lungs has important physiological functions,17 mediated by increases in intracellular cGMP within target cells.29 NTG is thought to act by way of intracellular S-nitrosothiol intermediates to directly stimulate guanylate cyclase or to release NO locally in effector cells29303132 and has recently been shown to increase NO in expired air,33 suggesting that NTG contributes to local levels of NO within the lungs. Other experiments in our laboratory demonstrate that supplementing NTG in the preservation solution augments tissue cGMP levels (unpublished radioimmunoassay data, 1994), suggesting that incorporating NTG into the pulmonary preservation solution is an effective means of delivery. The present set of experiments was designed to test whether NTG added to a pulmonary preservation solution might augment vascular homeostasis within the reperfused graft, thereby improving graft function and recipient survival. These experiments demonstrate that NTG is unequivocally effective in this regard, resulting in a marked stabilization of pulmonary hemodynamics and improved arterial oxygenation after transplantation. This is in contrast to hydralazine, which, although it is an effective vasodilator at the time of harvest, does not demonstrate the antineutrophil or antiplatelet effects that NTG does. The preservation of solid organs for transplantation has improved considerably over recent years, largely because of improvements in preservation techniques that enhance parenchymal function of the transplanted organs. However, the lungs remain among the most problematic organs for transplantation, for reasons that are not fully understood. In our studies, the beneficial effects of NTG in the setting of lung transplantation were not limited to vasodilation but included reduced neutrophil and platelet accumulation and improved gas exchange. These data are consistent with the observation of others that inhibiting platelet34 and neutrophil20353637 accumulation is also important after ischemia and reperfusion. Of the many different preservation strategies described in the experimental literature, only donor prostaglandin administration has been used in clinical lung transplant centers. However, the use of prostaglandins to improve donor preservation has remained sporadic because their effectiveness is controversial.2 Although prostaglandins per se were not tested in the present series of experiments, harvest vasodilation alone is insufficient to adequately preserve lungs, as demonstrated by our experiments in which an effective vasodilating dose of hydralazine during harvest failed to protect the lungs during reperfusion. Our data indicating that the vasodilator hydralazine is an ineffective pulmonary preservative are concordant with previously published data.38 In contrast to hydralazine, prostaglandins have a theoretical advantage in that they may improve vascular homeostasis by enhancing levels of the intracellular second messenger cAMP,3940 not only promoting vasodilation but inhibiting neutrophil adhesion and platelet aggregation as well.4142 This hypothesis is currently the subject of further investigation in our laboratory. The experiments presented here contribute to the growing body of evidence characterizing the detrimental role of neutrophils in pulmonary ischemia/reperfusion. Depletion of neutrophils from the perfusate20353637 has been shown to improve the function of reperfused lungs. In the present study, we have shown that attenuation of neutrophil accumulation within the pulmonary graft by NTG supplementation paralleled improved graft function and recipient survival after orthotopic transplantation. It is not surprising that NTG may interfere with neutrophil accumulation during reperfusion, as NO has been shown to interfere with neutrophil/endothelial adhesion,9 and local NO donors/analogues blunt myocardial injury and neutrophil accumulation during cardiac reperfusion.43 This attenuated neutrophil infiltration might contribute to the beneficial effects of NTG, because recruited neutrophils release numerous toxic compounds, including superoxide anion, chloramine, hypochlorous acid, hydroxyl radical, and hydrogen peroxide, as well as lysosomal contents, such as elastase, the metalloproteases (collagenase and gelatinase), neutral proteases, and heparinase.44 The initial source of superoxide after reperfusion in the lungs is not clear, although endothelial cells themselves have been shown in vitro to rapidly generate superoxide after hypoxia and reoxygenation.45 These initially formed oxygen radicals in the reperfusion milieu are potent neutrophil chemoattractants and activators46 that may compound subsequent neutrophil accumulation/activation, resulting in rapid graft failure. Grafts preserved with NTG appear to have a reduction in tissue oxidant stress after transplantation, as measured by the presence of thiobarbituric acid reactive substances (data not shown), supporting the potential beneficial outcome of reducing graft neutrophil infiltration. The initial inhibition of neutrophil recruitment into the reperfused graft by NTG may not only improve pulmonary parenchymal function but, by decreasing the reactive oxygen intermediate milieu, may result in greater local concentration of NO. As reactive oxygen intermediates induce prolonged expression of the neutrophil adherence molecule P-selectin on the endothelial surface, which mediates rapid neutrophil adhesion to the endothelium,47 initial reductions in neutrophil accumulation may be magnified by attenuating the production of reactive oxygen intermediates, further reducing neutrophil adhesion and activation. In this manner, the beneficial vascular effects of NTG may be magnified by its ability to attenuate the early phases of neutrophil adhesion. Reactive oxygen intermediates formed within lungs subjected to ischemia and reperfusion may rapidly combine with NO, forming highly toxic intermediates such as peroxynitrite and hydroxyl radical in the process.1948 This has caused reservations about the use of inhaled NO in the setting of pulmonary reperfusion. NTG may avoid this theoretical problem by directly activating guanylate cyclase via S-nitrosothiol intermediates.29303132 Although agents designed to limit the formation of reactive oxygen intermediates4950 have been studied in pulmonary preservation, none are used routinely for clinical lung transplantation. In pilot studies of our own, superoxide dismutase (conjugated to polyethylene glycol to extend its half-life in the circulation) administered to the pulmonary transplant recipient before reperfusion failed to protect the lungs compared with the control solution. This may relate to the extremely rapid kinetics of the reaction between superoxide and NO, with a rate constant of 108 (mol/L)−1 · s−1, which effectively competes with the dismutation of superoxide,19 or may relate to the relatively large size of the superoxide, limiting its accessibility to sites of superoxide formation. Because the reaction of superoxide with NO leads to the formation of highly toxic peroxynitrite and hydroxyl radicals, we performed limited experiments in which we tested whether blocking endogenous NO synthesis might, in combination with NTG, enhance pulmonary preservation. In these experiments, we tested the effects of NG-monomethyl-l-arginine (L-NMMA, a competitive inhibitor of NO synthesis) on graft preservation; addition of L-NMMA alone (5 μmol/L) to EC was associated with 100% recipient death (n=3), whereas L-NMMA plus supplemental NTG (0.1 mg/mL, n=3) was associated with 100% recipient survival. Further studies are currently under way to confirm whether inhibiting endogenous NO synthesis concomitant with the addition of supplemental NTG may have a pulmonary protective effect. The studies presented here demonstrate that a drug (hydralazine) that merely causes vasodilation at the time of pulmonary harvest but lacks other important vascular effects (such as the ability to reduce neutrophil and platelet sequestration after reperfusion) does not protect the lungs after transplantation. In contrast, NTG, which is an effective harvest vasodilator but which also has potent antineutrophil and antiplatelet actions, can improve gas exchange, reduce pulmonary vascular resistance, improve graft blood flow, and improve recipient survival after lung transplantation. These studies emphasize the pluripotent benefits of augmenting an important endogenous signaling pathway (NO), which may be depressed after pulmonary reperfusion. Figure 1. Graphs showing representative hemodynamic tracings of a lung transplant after hypothermic preservation for 6 hours in Euro-Collins solution (EC) alone (control, A) or EC supplemented with nitroglycerin (0.1 mg/mL, B) after the native lung was removed from the pulmonary circulation, as described in text. PAP indicates mean pulmonary arterial pressure; PAF, pulmonary arterial flow; PVR, pulmonary vascular resistance; and PA, pulmonary artery. Figure 2. Bar graphs showing the effects of nitroglycerin (NTG) or hydralazine on lung preservation for transplantation. All lung transplants were performed after 6 hours of hypothermic preservation in Euro-Collins solution alone (control, n=6 for panels A and B) or Euro-Collins solution supplemented with NTG (0.1 mg/mL, n=5) or hydralazine (0.02 μg/mL, n=6). Measurements were recorded at the final time at which the recipient was alive or at 30 minutes after ligation of the native right pulmonary artery (PA). A shows PA flow; B, pulmonary vascular resistance (PVR); and C, arterial oxygenation (n=10 for control, including oxygenation data from four pilot experiments in which hemodynamics were not measured). Values are mean±SEM. *P<.05 and **P<.01. Figure 3. Bar graph showing dose-dependent effects on recipient survival elicited by nitroglycerin (NTG) compared with hydralazine. Experiments were performed as described in Fig 2. The dose-response relation of NTG showed maximal ability to enhance lung preservation at 0.1 mg/mL (n=18 at 0 mg/mL, n=3 at 0.001 mg/mL, n=3 at 0.01 mg/mL, and n=7 at 0.1 mg/mL). *P<.05 and ****P<.001. Figure 4. Bar graph showing the effects of vasodilators added to the preservation solution on pulmonary vasodilation during harvest. Effect of preservation solution on pulmonary artery (PA) flushing time during harvest of the lung is shown: n=6 for control (Euro-Collins solution [EC] alone), n=5 for EC plus nitroglycerin (NTG, 0.1 mg/mL), and n=6 for EC plus hydralazine (0.02 μg/mL). *P<.05, **P<.01, and ***P<.005. Figure 5. Bar graphs showing effects of nitroglycerin (NTG) or hydralazine added to the pulmonary preservation solution on graft neutrophil and platelet accumulation. A, Myeloperoxidase activity (MPO) was used to quantify neutrophil deposition (n=6 for each group). ΔAbs indicates change in absorbance. B, 111In-labeled platelet accumulation, expressed as the ratio of graft radioactivity to blood radioactivity (n=6 for each group). *P<.05 and **P<.01. Figure 6. Bar graph showing the effect of nitroglycerin (NTG) added to the pulmonary preservation solution on systemic platelet function. Bleeding times (measured as the time required for hemostasis after uniform tail transection, as described in “Materials and Methods”) were used as an index of platelet function and were recorded immediately after reperfusion of lungs preserved with supplemental NTG (0.1 mg/mL, n=7) versus lungs preserved with control solution (n=10). This study was supported in part by grants from the Cystic Fibrosis Foundation (Dr Pinsky) and a Grant-in-Aid from the American Heart Association (Dr Pinsky). Dr Oz is the recipient of an Irving Scholarship, and Dr Pinsky is a Clinician Scientist of the American Heart Association. FootnotesReferences
|