- Email: [email protected]
Mechanisms of coronary vasomotor dysfunction in the transplanted heart
Mechanisms of Coronary Vasomotor Dysfunction in the Transplanted Heart David A. Fullerton, MD, Max B. Mitchell, MD, Robert C. McIntyre, Jr, MD, James M. Brown, MD, Xianzhong Meng, MD, David N. Campbell, MD, and Frederick L. Grover, MD Department of Surgery, University of Colorado Health Sciences Center, Denver, Colorado
The transplanted heart sustains both cold ischemic and reperfusion injuries. These can produce coronary vascular endothelial or smooth muscle injury or both, which, in turn, can produce coronary vasomotor dysfunction. Using a canine model of autologous heart transplantation, we examined the following coronary vasomotor control mechanisms in isolated coronary artery rings: (1) endothelial-dependent cyclic guanosine monophosphate (cGMP)-mediated vasorelaxation (response to acetylcholine); (2) endothelial-independent cGMP-mediated vasorelaxation (response to sodium nitroprusside); and (3) {l-adrenergic cyclic adenosine monophosphate (cAMP)mediated vasorelaxation (response to isoproterenol hydrochloride). Further, these mechanisms were related to 3 hours of cold ischemia alone and to 3 hours of cold ischemia plus 1 hour of reperfusion. Autologous heart
transplantation was performed in dogs, and isolated distal left anterior descending coronary artery rings were studied in individual organ chambers. Cold ischemia alone produced significant dysfunction of {l-adrenergic cAMP-mediated vasorelaxation, which was exacerbated after reperfusion. Neither endothelial-dependent nor endothelial-independent cGMP-mediated vasorelaxation was dysfunctional after cold ischemia alone, but both were significantly impaired after reperfusion. We conclude that cold ischemia and reperfusion each produce coronary vasomotor dysfunction in the transplanted heart. Cumulatively, such coronary vasomotor dysfunction can acutely impair coronary vasodilatation and potentially jeopardize myocardial blood flow in the transplanted heart. (Ann Thorae Surg 1994;58:86-92)
O
On the other hand, cGMP-mediated vasorelaxation can be either endothelial dependent or endothelial independent [3]. Agents such as acetylcholine produce coronary vascular smooth muscle relaxation by binding to muscarinic receptors on coronary vascular endothelium [4]. In response, the coronary vascular endothelium releases an endothelial-derived relaxing factor, which is thought to be nitric oxide [4]. Nitric oxide, in tum, activates guanylate cyclase within the coronary vascular smooth muscle cell [4] to produce cGMP and effect coronary vascular smooth muscle relaxation. On the other hand, sodium nitroprusside is a functional analogue of nitric oxide and directly activates coronary vascular smooth muscle guanylate cyclase to produce cGMP independently of the endothelium
ne of the distinguishing features of the coronary circulation is the fact that myocardial oxygen extraction is nearly maximal, even under resting conditions [1]. Normally there is a linear relationship between coronary blood flow and myocardial oxygen demand [1, 2]. Therefore, to increase myocardial oxygen supply in the setting of increased myocardial oxygen demand, coronary vasodilation is required. If the mechanisms of coronary arterial vasorelaxation are dysfunctional, the coronary circulation may be unable to appropriately vasodilate in response to an increased myocardial oxygen demand. In such a case, coronary reserve may be compromised. The principal intracellular mechanisms of coronary vascular smooth muscle relaxation are ultimately mediated through cyclic adenosine monophosphate (cAMP) or cyclic guanosine monophosphate (cGMP) [3]. In response to activation of receptors on coronary vascular smooth muscle cells (such as f3-adrenergic receptors), coronary vascular smooth muscle adenylate cyclase generates cAMP. The increased level of cAMP, in tum, effects coronary vasorelaxation [3]. In this study, we used isoproterenol hydrochloride as a tool to activate cAMP-mediated coronary vasorelaxation. Presented at the Fortieth Annual Meeting of the Southern Thoracic Surgical Association, Panama City Beach, FL, Nov 4--6, 1993. Address reprint requests to Dr Fullerton, Cardiothoracic Surgery, Box C310, University of Colorado Health Sciences Center, 4200 E Ninth Ave, Denver, CO 80262.
©
1994 by The Society of Thoracic Surgeons
[4].
Heart transplantation requires the transplanted heart to sustain both ischemic and reperfusion injuries. We hypothesized that these processes can produce injury to the coronary vascular endothelium, the vascular smooth muscle, or both and culminate in coronary vasomotor dysfunction. Specifically, we hypothesized that the processes of heart transplantation can produce dysfunction of the mechanisms of coronary vasorelaxation. Therefore, the first purpose of this study was to examine the following mechanisms of coronary vasomotor control in isolated coronary artery rings in a canine model of autologous heart transplantation: (1) endothelial-dependent cGMP-mediated vasorelaxation (response to acetyl0003-4975/94/$7.00
Ann Thorae Surg 1994;58:86--92
choline); (2) endothelial-independent cGMP-mediated vasorelaxation (response to nitroprusside); and (3) l3-adrenergic cAMP-mediated vas ore laxation (response to isoproterenol). The second purpose of this study was to relate each of these mechanisms to cold ischemia alone and to cold ischemia plus reperfusion.
Material and Methods All animals received humane care in compliance with the "Principles of Laboratory Animal Care" formulated by the National Society for Medical Research and the "Guide for the Care and Use of Laboratory Animals" published by the National Institutes of Health (NIH publication 86-23, revised 1985).
Surgical Protocol After endotracheal intubation and mechanical ventilation, mongrel dogs of either sex and weighing 18 to 25 kg underwent autologous heart transplantation under sodium pentobarbital and halothane anesthesia through a sternotomy. Blood pressure was continuously monitored with an intraarterial catheter, and mean aortic pressure remained greater than 40 mm Hg throughout the experimental protocol. After systemic heparinization (heparin sodium, 250 U/kg), total cardiopulmonary bypass was initiated via cannulas placed in the ascending aorta and superior and inferior venae cavae. A pediatric bubble oxygenator was employed (Cobe Laboratories, Denver, CO). The ascending aorta was cross-clamped, and St. Thomas' crystalloid cardioplegic solution (4°C, 1,000 mL) was infused into the proximal aortic root. The heart was excised, placed in cold saline solution (4°C), and stored in a cooler filled with ice for 3 hours. The heart was reimplanted using running monofilament suture material for all surgical anastomoses. Approximately 20 minutes was required to reimplant the heart. The heart was reperfused on release of the aortic cross-clamp (mean arterial pressure, 45 to 55 mm Hg at the time of reperfusion) and allowed to beat on bypass. After 1 hour of reperfusion, cardiopulmonary bypass was discontinued, and the heart was rapidly excised. The dogs were therefore on bypass for 4 hours. The coronary sinus was directly cannulated, and cold saline solution (4°C, 250 mL) was allowed to run in under the force of gravity (low pressure) to flush all blood from the coronary circulation. Dogs serving as controls were not placed on cardiopulmonary bypass. In these dogs, after general anesthesia and sternotomy, the aorta was cross-clamped and St. Thomas' crystalloid cardioplegic solution We, 1,000 mL) was infused into the proximal aortic root, after which the heart was quickly excised, and coronary artery rings were prepared.
Coronary Artery Ring Preparation Coronary artery rings were dissected from the distal left anterior descending coronary artery with the use of a dissecting microscope. The rings were studied in three
FULLERTON ET AL CORONARY DYSFUNCTION IN TRANSPLANTED HEART
87
groups of hearts at different points in time: control (immediately after harvest after cold cardioplegia) (n = 3 dogs, 6 rings); at the end of cold ischemia (after 3 hours in saline solution at 4°C) (n = 6 dogs, 12 rings); and at the end of cold ischemia plus reperfusion (1 hour after heart reimplantation) (n = 5 dogs, 18 rings). Under dissecting microscope magnification, the surrounding tissue was dissected from the coronary arteries. Each of the arteries was cut into rings 3 to 4 mm wide. Great care was taken during this process to avoid endothelial injury. The coronary artery rings were suspended on fine wire tensiometers in individual lO-mL organ chambers. The organ chambers were surrounded by water jackets and continuously warmed to 3rc. The organ chambers were filled with Earle's balanced salt solution and continuously bubbled with gas consisting of 21% oxygen, 5% carbon dioxide, and balanced nitrogen. Earle's balanced salt solution is a standard physiologic salt solution and contains CaCl z (1.80 mmol/L), MgS0 4 (anhydrous) (0.83 mmol/L), KCl (5.36 mmol/L), NaCl 116.34 mmol/L), NaP04 (dibasic) (0.40 mmol/L), D-glucose (5.50 mmol/L), NaHC0 3 (19.04 mmol/L), and phenol red Na (as pH indicator) (0.03 mmol/L). Ring tension was determined by use of a forcedisplacement transducer (Grass FT03) attached to each tensiometer apparatus. Ring tension was thereby recorded using a MacLab Data Interface Module (ADI Instruments, Milford, MA) on a Macintosh IIci computer (Apple Computer, Inc, Cupertino, CA). The resting tension of coronary artery rings of this size at which a maximal contractile response to KCl (80 mmol/L) was obtained was found to be 4 g in a separate series of experiments. Once the rings had been allowed to reach a steady state (approximately 90 minutes) at a baseline tension of 4 g, the vasorelaxing effects of the pharmacologic agents were tested. After the coronary artery rings were preconstricted with KCl (40 mmol/L) and in a steady state, the vasorelaxing effects of acetylcholine, nitroprusside, and isoproterenol were determined in each ring in a random order by generating cumulative concentration-response curves (10- 9 to 10- 5 mol/L). After the test of each agent, the organ chambers were flushed several times and the rings allowed to reach a steady state before they were preconstricted once again with KCl (40 mmol/L) to test the next vasorelaxing agent.
Histologic Studies Histologic observation and immunofluorescence microscopy were carried out in coronary arteries from two control hearts and from three hearts after cold ischemia plus reperfusion. Hematoxylin and eosin staining was performed on cryopreserved sections of distal left anterior descending coronary arteries using standard technique. For immunofluorescence microscopic studies, distal left anterior descending coronary arteries approximately 1 to 1.5 mm in diameter were embedded in OCT compound and frozen in dry-ice-cold isopentane. Tissue blocks were stored at -70°C before use. Transverse sections 5 ILm thick were cut with a cryostat (Reichert-Jung 2800,
88
FULLERTON ET AL CORONARY DYSFUNCTION IN TRANSPLANTED HEART
Ann Thorae Surg
1994;58:86-92
*
100
100
o
oZ Z
Cold Ischemia
80
and Reperfusion
:;: ::E
OIl ~
::E
OIl 50
Z
sa z '" OIl
Cold Ischemia and Reperfusion
b
40
25
""
b
""
75
~
eo
z
sa Z '" OIl
2S 2S «
Cold Ischemia
20
ontrol
Cold Ischemia
9
Control
8
7
6
CONCENTRATION SODIUM NITROPRUSSIDE (-log molar) CONCENTRATION ACETYLCHOLINE (-log molar)
Fig 1. Effects of cold ischemia alone and cold ischemia plus reperfusian on endothelial-dependent cyclic guanosine monophosphate-mediated coronary vasorelaxation. Concentration-response curves of acetylcholine in left anterior descending coronary artery rings preconstricted with KC!. Significant dysfunction was produced by cold ischemia plus reperfusion but not cold ischemia alone. Data are shown as the mean ± the standard error of the mean. (Op < 0.05 versus conirol.)
Frigocut, Germany). Sections were collected on polylysine-coated slides and incubated with monoclonal antibody against alpha smooth muscle actin (BioGenex Laboratories, San Ramon, CA) and polyclonal antibody against von Willebrand factor (Endotech Corp, Indianapolis, IN) in 1% bovine serum albumin-phosphate-buffered saline solution for 60 minutes at room temperature. After washing with phosphate-buffered saline solution, sections were incubated with fluorescein-conjugated anti-mouse immunoglobulin G and Cy3-conjugated anti-rabbit immunoglobulin G in 1% bovine serum albumin-phosphatebuffered saline solution for 45 minutes at room temperature. After a thorough rinse with phosphate-buffered saline solution, sections were mounted with glycerolbased antiquenching medium. Microscopic observation and photography were then carried out.
Fig 2. Effects of cold ischemia alone and cold ischemia plus reperfusion on endothelial-independent cyclic guanosine monophosphaie-mediated coronary vasorelaxation. Concentration-response curvesof sodium nitroprusside in left anterior descending coronary artery rings preconstricted with KCl. Significant dYsfunction was produced by cold ischemia plus reperfusion but not cold ischemia alone. Data are shown as the mean ± the standard error of the mean. (Op < 0.05 versus contro!.)
ated relaxation (response to isoproterenol) produced 98% ± 4% maximal relaxation (Fig 3). After 3 hours of cold ischemia alone, neither endothelial-dependent cGMP-mediated (response to acetylcholine) (see Fig 1) nor endothelial-independent cGMPmediated (response to nitroprusside) (see Fig 2) coronary vascular smooth muscle relaxation was dysfunctional. However, at the end of 3 hours of cold ischemia alone, f3-adrenergic cAMP-mediated vasorelaxation (response to isoproterenol) was significantly impaired, as isoprotere-
l00¢----O"
\ -,
\\"\J l:J
Statistical Analysis Data are presented as the percent relaxation of KCIinduced ring tension produced by the given concentration of vasorelaxing agent. Values are expressed as the mean ± one standard error of the mean. Analysis of variance with post hoc Scheffe's F test was used. A p value of less than 0.05 was considered significant.
*
I'. \...
T 1
.\.\ J +
1
Rep:esion
~+
\\J 20
Cold Ischemia
~
1
Cold Ischemia
lJ················
o
Control
Results Significant coronary vasomotor dysfunction was produced by the processes of heart transplantation. In control coronary artery rings (studied after cold cardioplegia alone), endothelial-dependent cGMP-mediated relaxation (response to acetylcholine) produced 97% ± 3% maximal relaxation (Fig I), endothelial-independent cGMP-mediated relaxation (response to nitroprusside) produced 99% ± 3% maximal relaxation (Fig 2), and l3-adrenergic cAMP-medi-
CONCENTRATION ISOPROTERENOL (-log molar)
Fig 3. Effects of cold ischemia alone and cold ischemia plus reperjusion on j3-adrenergic cyclic adenosine monophosphate-mediaied coronary vasorelaxation. Concentration-response curves of isoproterenol hydrochloride in left anterior descending coronary artery rings preconstricted with KCl. Significant dysfunction was produced with cold ischemia alone and was exaggerated by reperfusion. Data are shown as the mean ± the standard error of the mean. (Op < 0.05 versus control; "p < 0.05 versus cold ischemia plus reperjusion.)
FULLERTON ET AL CORONARY DYSFUNCTION IN TRANSPLANTED HEART
Ann Thorae Surg 1994;58:86--92
89
ischemia plus reperfusion, there is a reduction in the density of von Willebrand factor staining, a finding consistent with a loss of coronary vascular endothelial cells, and substantial irregularity in the coronary vascular endothelial basal lamina. Structural changes in coronary vascular smooth muscle are also demonstrated on hematoxylin and eosin-stained sections (Fig 5). Under highpower light microscopy, a loss of order and distinction of coronary vascular smooth muscle cells is suggested after cold ischemia plus reperfusion compared with controls.
Comment A
Heart transplantation requires the transplanted heart to sustain the injurious processes of both cold ischemia and reperfusion. The results of the present study demonstrate that these processes culminate in significant coronary artery vasomotor dysfunction in the transplanted heart and also demonstrate that cold ischemia and reperfusion each produce different patterns of coronary vasomotor dysfunction. Neither endothelial-dependent cGMP-
B
Fig 4. Immunofluorescent staining of (A) control left anterior descending coronary artery and (B) left anterior descending coronary artery after ischemia plus reperfusion. Von Willebrand factor stains red, and vascular smooth musclealpha actin stains green. After ischemia plus reperfusion, there is partial loss of von Willebrand factor staining and substantial irregularity produced in the endothelial basal lamina, findings consistent with injury to the coronary vascular endothelium. (x 160 before 56% reduction.)
nol produced only 78% ± 3% maximal relaxation (p < 0.05) (see Fig 3). After reperfusion, there was significant dysfunction of endothelial-dependent cGMP-mediated coronary vascular smooth muscle relaxation, as acetylcholine produced only 26% ± 8% maximal relaxation (p < 0.05) (see Fig 1). Likewise, endothelial-independent cGMP-mediated coronary vascular smooth muscle relaxation was dysfunctional after reperfusion, as nitroprusside achieved only 53% ± 1% relaxation (p < 0.05) (see Fig 2). Further, the dysfunction of ,B-adrenergic cAMP-mediated vasorelaxation was significantly exacerbated by reperfusion, as isoproterenol produced only 43% ± 6% maximal relaxation (p < 0.05) (see Fig 3). These functional changes related to structural changes in both coronary vascular endothelium and vascular smooth muscle. Figure 4 demonstrates immunofluorescent staining of coronary vascular endothelial von Willebrand factor and alpha-actin antigen staining in coronary vascular smooth muscle. As seen in Figure 4, after cold
B
Fig 5. Hematoxylin and eosin staining of (A) control left anterior descending coronary artery and (B) left anterior descending coronary arteryafter ischemia plus reperfusion. In addition to critical irregularity in the endothelial basal lamina, there is the appearance of vacuoles in the vascular smooth muscle cells and loss of distinct smooth muscle cell orientation. (x400 before 56% reduction.)
90
FULLERTON ET AL CORONARY DYSFUNCTION IN TRANSPLANTED HEART
mediated relaxation (response to acetylcholine) nor endothelial-independent cGMP-mediated relaxation (response to nitroprusside) was impaired by cold ischemia alone, but both were significantly impaired after reperfusion. On the other hand, f:l-adrenergic cAMP-mediated relaxation (response to isoproterenol) was impaired by cold ischemia alone, and this impairment was significantly exacerbated by reperfusion. Thus, both of the principal intracellular pathways of coronary vascular smooth muscle relaxation (cGMP- and cAMP-mediated) were dysfunctional in the canine transplanted heart. In this study, coronary vasomotor control mechanisms were examined in epicardial coronary arteries. It is possible that the influences of cold ischemia and reperfusion are different in smaller, intramyocardial coronary vessels. However, Dignan and colleagues [5] found no difference in the responses to ischemia of proximal compared with distal coronary artery endothelial cell or smooth muscle function. Further, as suggested by Quillen and coworkers [6], smaller coronary vessels may be even more susceptible to ischemia and reperfusion than larger arterial vessels. Some investigators have noted that endothelialdependent coronary vasomotor dysfunction can result from crystalloid cardioplegic solutions containing potassium. However, the present study corroborates the findings of Evora and colleagues [7] in demonstrating that no endothelial-dependent coronary vasomotor dysfunction was produced by cold cardioplegia and hypothermia alone. Thus, the dysfunction seen in the present study might be due to the injuries of cold ischemia and reperfusion rather than to the cardioplegic solution. The findings of this study are in concert with those of other investigators. In prior studies [6, 8-12], endothelialdependent but not endothelial-independent vasomotor dysfunction has been produced by relatively short periods of ischemia followed by reperfusion. Fewer studies have examined the influence of ischemia alone on coronary vasomotor control. Dignan and associates [5] demonstrated porcine coronary endothelial-dependent vasomotor control was intact for up to 160 minutes of cold ischemia alone but became dysfunctional after longer periods of cold ischemia. Selke and coauthors [13] found dysfunction of endothelial-dependent vasomotor control in porcine coronaries after cold cardioplegia and reperfusion but not after cold cardioplegia alone. The design of the present study attempted to reproduce the clinical circumstances of heart transplantation. In so doing, it differs from prior studies by examining the influences of a much longer period of ischemia (3 hours) and reperfusion. This design also allowed us to examine the influence of cold ischemia alone and cold ischemia plus reperfusion on coronary vasomotor control. Using the transplanted heart, truly global myocardial ischemia was achieved, thereby avoiding the rich coronary collateral circulation of the dog, which can provide flow distal to the point of coronary artery occlusion [14]. Last, heart transplantation requires the use of cardiopulmonary bypass, which may exaggerate the severity of reperfusion injury.
Ann Thorae Surg 1994;58:86--92
Perhaps because of these differences, the present study demonstrated dysfunction of both endothelial-dependent and endothelial-independent mechanisms of coronary vasomotor control. These results suggest that under conditions of heart transplantation, typically requiring several hours of ischemia before reperfusion, coronary vasomotor dysfunction is greater in the transplanted heart than after shorter periods of regional or global myocardial ischemia and reperfusion. Further, the findings suggest that protracted cold ischemia and reperfusion each produce coronary vasomotor dysfunction, albeit different patterns of dysfunction. Prior studies that have examined coronary vessels by scanning electron microscopy after ischemia and reperfusion have yielded conflicting results. VanBenthuysen and associates [10] did find structural changes in coronary endothelial cells after 1 hour of ischemia and 1 hour of reperfusion but not after ischemia alone. However, neither Quillin and co-workers [6] nor Selke and colleagues [13] found structural changes in coronary vascular endothelium after ischemia and reperfusion. The present study suggests structural changes are found in both coronary vascular endothelium and smooth muscle cells after protracted cold ischemia plus reperfusion in the transplanted heart. These structural changes were related to functional changes in coronary vasomotor control and at least suggest the possibility that the functional changes resulted from the structural injury produced by cold ischemia plus reperfusion. In the present study, coronary vasomotor dysfunction was demonstrated acutely after heart transplantation, and this dysfunction was related to histologic evidence of coronary vascular damage. Other investigators [15] have demonstrated reendothelialization of arteries denuded of endothelium. In rat aorta, endothelium-dependent relaxation has been shown to return by 8 weeks after mechanical denuding of the endothelium [16]. The design of the present study does not address whether the vasomotor dysfunction seen acutely will resolve in the transplanted heart or what the time course of this process might be. There are several potential mechanisms by which protracted cold ischemia and reperfusion may culminate in this significant coronary vasomotor dysfunction. Vascular endothelial expression of leukocyte adhesion molecules has been demonstrated after ischemia, thus setting the stage for neutrophil-endothelial interaction on reperfusion [17]. Cardiopulmonary bypass (which must be used in heart transplantation) is known to produce complement activation [18] and neutrophil activation [19-21] and to increase the circulating levels of cytokines such as endotoxin [21, 22], interleukins [23, 24], and tumor necrosis factor [22]. Endotoxin in particular has been shown to "prime" both neutrophils and vascular endothelial cells, thus creating the potential for significantly greater damage on reperfusion of an ischemic vascular bed [25]. Thus, in the setting of heart transplantation, the coronary vascular bed not only is subjected to protracted cold ischemia, but is reperfused with activated neutrophils, activated complement, and cytokines. Together, these may reasonably be expected to produce greater injury on
Ann Thorac Surg 1994;58:86--92
reperfusion than transient coronary artery occlusion and reperfusion without cardiopulmonary bypass. In summary, this study demonstrated significant coronary vasomotor dysfunction in the transplanted canine heart. By using a model of autotransplantation, we avoided the possible influences of rejection. Three hours of cold ischemia alone was found to impair l3-adrenergic cAMP-mediated coronary vasorelaxation, and this dysfunction was exacerbated on reperfusion. Neither endothelial-dependent nor endothelial-independent cGMPmediated mechanisms of vasorelaxation were impaired by protracted cold ischemia alone, but both were significantly impaired after reperfusion. We conclude that such coronary vasomotor dysfunction may impair coronary vasodilatation and jeopardize coronary reserve in the transplanted heart. By limiting coronary reserve, this coronary vasomotor dysfunction may, in turn, compromise myocardial function of the transplanted heart. Supported by grant R29HL49398 from the National Institutes of Health.
References 1. Feigl EO. Coronary physiology. Physiol Rev 1983;63:1-205. 2. Berne RM, Rubio R. Coronary circulation. In: Berne RM, ed. Handbook of physiology: the cardiovascular system. Bethesda, MD: American Physiological Society, 1979:873. 3. Vane JR, Anggard EE, Botting RM. Regulatory functions of the vascular endothelium. N Engl J Med 1990;323:27-36. 4. Ignarro LJ. Biological actions and properties of endotheliumderived nitric oxide formed and released from artery and vein. Circ Res 1989;65:1-21. 5. Dignan RJ, Dyke CM, Abd-Elfattah AS, et al. Coronary artery endothelial cell and smooth muscle dysfunction after global myocardial ischemia. Ann Thorac Surg 1992;53:311-7. 6. Quillen JE, Selke FW, Brooks LA, Harrison DG. Ischemiareperfusion impairs endothelium-dependent relaxation of coronary microvessels but does not affect large arteries. Circulation 1990;82:586-94. 7. Evora PRB, Pearson PJ, Schaff HV. Crystalloid cardioplegia and hypothermia do not impair endothelium-dependent relaxation or damage vascular smooth of epicardial coronary arteries. J Thorac Cardiovasc Surg 1992;104:1365-74. 8. Hashimoto K, Pearson PJ, Schaff HV, Cartier R. Endothelial cell dysfunction after ischemic arrest and reperfusion. A possible mechanism of myocardial injury during reflow. J Thorac Cardiovasc Surg 1991;102:688-94. 9. Pearson PJ, Schaff HV, Vanhoutte PM. Acute impairment of endothelium-dependent relaxations to aggregating platelets following reperfusion injury in canine coronary arteries. Circ Res 1990;67:385-93.
FULLERTON ET AL CORONARY DYSFUNCTION IN TRANSPLANTED HEART
91
10. VanBenthuysen KM, McMurtry IF, Horwitz LD. Reperfusion after acute coronary occlusion in dogs impairs endotheliumdependent relaxation to acetylcholine and augments contractile reactivity in vitro. J Clin Invest 1987;79:265-74. 11. Ku DD. Coronary vascular reactivity after myocardial ischemia. Science 1982;218:576-8. 12. Mehta JL, Nichols WW, Donnelly WH, Lawson DL, Saldeen TGP. Impaired canine coronary vasodilator response to acetylcholine and bradykinin after occlusion-reperfusion. Circ Res 1989;64:43-54. 13. Sellke FW, Shafique T, Shoen FJ, Weintraub RM. Impaired endothelium-dependent coronary microvascular relaxation after cold potassium cardioplegia and reperfusion. J Thorac Cardiovasc Surg 1993;105:52-8. 14. Harken AH, Simson MB, Haselgrove J, Wetstein L, Harden WR, Barlow CH. Early ischemia after complete coronary ligation in the rabbit, dog, pig and monkey. Am J Physiol 1981;241:H202-1O. 15. Haudenshild CC, Schwartz SM. Endothelial regeneration. II. Restitution of endothelial continuity. Lab Invest 1979;41: 407-18. 16. Cartier R, Pearson PI, Lin PI, Schaff HV. Time course and extent of recovery of endothelium-dependent contractions and relaxations after direct arterial injury. J Thorac Cardiovase Surg 1991;102:371-7. 17. Lefer AM, Ma X-L. Cytokines and growth factors in endothelial dysfunction. Crit Care Med 1993;21:59-14. 18. Kirklin JK, Westaby 5, Blackstone EH, Kirklin JW, Chenoweth DE, Pacifico AD. Complement and the damaging effects of cardiopulmonary bypass. J Thorac Cardiovasc Surg 1983;86:845-57. 19. Faymonville ME, Pincemain J, Duchateau J, et al. Myeloperoxidase and elastase as markers of leukocyte activation during cardiopulmonary bypass in humans. J Thorac Cardiovase Surg 1991;102:309-17. 20. Gu YJ, van Oeveren W, Boonstra PW, de Haan J, Wildevuur CRH. Leukocyte activation with increased expression of CR3 receptors during cardiopulmonary bypass. Ann Thorac Surg 1992;53:839-43. 21. Kharazmi A, Andersen LW, Baek L, Valerius NH, Laub M, Rasmussen JP. Endotoxemia and enhanced generation of oxygen radicals by neutrophils from patients undergoing cardiopulmonary bypass. J Thorac Cardiovasc Surg 1989;98: 381-5. 22. Jansen NJG, van Oeveren W, Gu YJ, van Vliet MH, Eijsman L, Wildevuur CRH. Endotoxin release and tumor necrosis factor formation during cardiopulmonary bypass. Ann Thorae Surg 1992;54:744-8. 23. Butler J, Chong GL, Baigrie RI, Pillai R, Westaby 5, Rocker GM. Cytokine responses to cardiopulmonary bypass with membrane and bubble oxygenation. Ann Thorac Surg 1992; 53:833-8. 24. Downing SW, Edmunds LH Jr. Release of vasoactive substances during cardiopulmonary bypass. Ann Thorac Surg 1992;54:1236-43. 25. Anderson BO, Bensard DD, Brown JM, et al. FNLP injures endotoxin-primed rat lung by neutrophil-dependent and -independent mechanisms. Am J Physiol 1991;260:R413-20.
DISCUSSION DR CURTIS G. TRIBBLE (Charlottesville, VA): I enjoyed this
paper, and I also enjoyed the opportunity to review the manuscript. I have a bit of perspective to add to what this paper is all about, ie, organ preservation. For a long time, most people interested in organ preservation have been focusing on parenchymal preservation, such as preservation of the myocardium or preservation of energy stores, and not on endothelial preservation. We in our laboratory, Dr Fullerton in his laboratory, and
others have become interested in endothelial preservation because it seems that endothelial dysfunction probably is related to organ dysfunction, especially in the case of the heart. We have seen this relationship in our laboratory and presented the data, which show a correlation between endothelial dysfunction and cardiac dysfunction, at the American College of Surgeons meeting last year. The vascular dysfunction that Dr Fullerton has discussed that
92
FULLERTON ET AL CORONARY DYSFUNCTION IN TRANSPLANTED HEART
is secondary to cold ischemia and reperfusion further highlights the challenge of how to better preserve vascular function or how to overcome this vascular dysfunction. I believe these issues will be the challenges for organ preservation research in the immediate future. This study was very well designed and very well presented, and represents an important area of research in organ preservation. It is also interesting to note that this type of research is not going to be done very often in this way by our physiology colleagues or our cardiology colleagues. It is the kind of research done primarily by surgical investigators. I have a minor technical question. I was interested that you used a root perfusate gas with only 21% oxygen in it, and I noticed that you pointed out that that gave an oxygen tension normal for oxygenated blood. In this type of work, we have used a 95% oxygen and 5% carbon dioxide mix on the premise that oxygenated crystalloid at an oxygen tension of 100 mm Hg or so is relatively ischemic. I am curious why you chose your gas concentration, and I bet you looked at both 21% and 95%. Perhaps the 21% oxygen is even better than the higher oxygen tension. Maybe there is a lesson in this that we should be using the more hypoxic gases, not just in our research but in the early stages of reperfusion also. In a more general sense, I would appreciate hearing your thoughts on how you think your results will lead to better preservation of organs in this setting or on what you might be thinking of doing clinically. For instance, should we preserve organs at a warmer temperature, 18° or 15°C, instead of 4°C? Should we reperfuse them with lower oxygen tensions? DR RICHARD E. CLARK (Pittsburgh, PAl: My comment in part echoes that of Dr Tribble. For 20 years, we have been talking about myocardial preservation, and we have gotten pretty good clinically. I think that if we are going to do any better, the future focus clearly has to be on the vascular smooth muscle and endothelium. I have two questions for you. Did you have an opportunity to do any scanning electron microscopy on your rings? The reason I ask is that we can very dramatically show under various preservation systems that the endothelial cells can separate and expose basement membrane. If this occurs prior to reperfusion, there is going to be a problem with the myocardium in the immediate reperfusion interval. The second question has to do with using this model to test various pharmacologic agents as a protective mechanism in your cold storage system. Have you had the opportunity to study thromboxane A2 agonists and antagonists? DR HENDRICK B. BARNER (St. Louis, MO): I believe there are data indicating that hyperkalemic crystalloid perfusion damages
Ann Thorae Surg 1994;58:86-92
the endothelium but that hyperkalemic blood cardioplegia does not, the critical range of potassium being 20 to 30 mEq. Could you comment? DR D. GLENN PENNINGTON (St. Louis, MO): Considering the drugs we know so well clinically, particularly nitroprusside and isoproterenol, could you give us some idea of the correlation between these dose ranges and what we might be using clinically? Would these doses be realistic in treating patients? DR FULLERTON: Dr Tribble, thank you very much for reviewing our manuscript. We do appreciate your comments, as we are aware of the nice work your group has done. Our standard ring preparation calls for a 21% oxygen gas. I wish I had some good reason to give you for that choice, but it is simply our standard model. It is chosen because at our altitude, that will give an oxygen tension of just under 100 mm Hg, about 98 mm Hg. Dr Clark, we did not have the opportunity to use scanning electron microscopy. However, we were pleased that our immunofluorescent and hematoxylin and eosin stainings are consistent with the findings shown by electron microscopy. Typically what is described, as you pointed out, is denuding of the endothelium or, of even more interest, the crypts that seem to form in the endothelial layer, which I believe we demonstrated as well. In our coronary ring model, we have not examined the influences of thromboxane, but we do have a line of investigation in our laboratory examining this in pulmonary artery rings. Dr Barner, as you know, there are several articles in the literature that attribute some endothelial dysfunction to crystalloid cardioplegia. We did not find that. The potassium in our cardioplegia is about 16 mEq/L, which is a bit on the low side compared with some of the other studies. However, our findings were actually quite similar to those of Dr Shaft's group, who found no dysfunction using a similar concentration of potassium. So I do not think our results are attributable to that. Although the data are not presented, when we have examined coronary rings without the influence of cardioplegia, we have found doseresponse curves that are identical to those of our control group. Dr Pennington, it is difficult to know exactly how these doses relate to blood levels of these drugs. Acetylcholine is not a very commonly used drug, although nitroprusside is. The half-life of nitroprusside is so rapid that it is very difficult to measure blood levels, and in fact, having sought this information, I have been unable to find it. The half-life of isoproterenol is about 2 minutes, and so at doses of about 0.01 /Jog' kg-I. min-I, which is relatively frequently used in our pediatric population, we have found, or the pharmacists tell us, that we are generating blood levels comparable with about 10- 7 mollL in our baths.
The transplanted heart sustains both cold ischemic and reperfusion injuries. These can produce coronary vascular endothelial or smooth muscle injury or both, which, in turn, can produce coronary vasomotor dysfunction. Using a canine model of autologous heart transplantation, we examined the following coronary vasomotor control mechanisms in isolated coronary artery rings: (1) endothelial-dependent cyclic guanosine monophosphate (cGMP)-mediated vasorelaxation (response to acetylcholine); (2) endothelial-independent cGMP-mediated vasorelaxation (response to sodium nitroprusside); and (3) {l-adrenergic cyclic adenosine monophosphate (cAMP)mediated vasorelaxation (response to isoproterenol hydrochloride). Further, these mechanisms were related to 3 hours of cold ischemia alone and to 3 hours of cold ischemia plus 1 hour of reperfusion. Autologous heart
transplantation was performed in dogs, and isolated distal left anterior descending coronary artery rings were studied in individual organ chambers. Cold ischemia alone produced significant dysfunction of {l-adrenergic cAMP-mediated vasorelaxation, which was exacerbated after reperfusion. Neither endothelial-dependent nor endothelial-independent cGMP-mediated vasorelaxation was dysfunctional after cold ischemia alone, but both were significantly impaired after reperfusion. We conclude that cold ischemia and reperfusion each produce coronary vasomotor dysfunction in the transplanted heart. Cumulatively, such coronary vasomotor dysfunction can acutely impair coronary vasodilatation and potentially jeopardize myocardial blood flow in the transplanted heart. (Ann Thorae Surg 1994;58:86-92)
O
On the other hand, cGMP-mediated vasorelaxation can be either endothelial dependent or endothelial independent [3]. Agents such as acetylcholine produce coronary vascular smooth muscle relaxation by binding to muscarinic receptors on coronary vascular endothelium [4]. In response, the coronary vascular endothelium releases an endothelial-derived relaxing factor, which is thought to be nitric oxide [4]. Nitric oxide, in tum, activates guanylate cyclase within the coronary vascular smooth muscle cell [4] to produce cGMP and effect coronary vascular smooth muscle relaxation. On the other hand, sodium nitroprusside is a functional analogue of nitric oxide and directly activates coronary vascular smooth muscle guanylate cyclase to produce cGMP independently of the endothelium
ne of the distinguishing features of the coronary circulation is the fact that myocardial oxygen extraction is nearly maximal, even under resting conditions [1]. Normally there is a linear relationship between coronary blood flow and myocardial oxygen demand [1, 2]. Therefore, to increase myocardial oxygen supply in the setting of increased myocardial oxygen demand, coronary vasodilation is required. If the mechanisms of coronary arterial vasorelaxation are dysfunctional, the coronary circulation may be unable to appropriately vasodilate in response to an increased myocardial oxygen demand. In such a case, coronary reserve may be compromised. The principal intracellular mechanisms of coronary vascular smooth muscle relaxation are ultimately mediated through cyclic adenosine monophosphate (cAMP) or cyclic guanosine monophosphate (cGMP) [3]. In response to activation of receptors on coronary vascular smooth muscle cells (such as f3-adrenergic receptors), coronary vascular smooth muscle adenylate cyclase generates cAMP. The increased level of cAMP, in tum, effects coronary vasorelaxation [3]. In this study, we used isoproterenol hydrochloride as a tool to activate cAMP-mediated coronary vasorelaxation. Presented at the Fortieth Annual Meeting of the Southern Thoracic Surgical Association, Panama City Beach, FL, Nov 4--6, 1993. Address reprint requests to Dr Fullerton, Cardiothoracic Surgery, Box C310, University of Colorado Health Sciences Center, 4200 E Ninth Ave, Denver, CO 80262.
©
1994 by The Society of Thoracic Surgeons
[4].
Heart transplantation requires the transplanted heart to sustain both ischemic and reperfusion injuries. We hypothesized that these processes can produce injury to the coronary vascular endothelium, the vascular smooth muscle, or both and culminate in coronary vasomotor dysfunction. Specifically, we hypothesized that the processes of heart transplantation can produce dysfunction of the mechanisms of coronary vasorelaxation. Therefore, the first purpose of this study was to examine the following mechanisms of coronary vasomotor control in isolated coronary artery rings in a canine model of autologous heart transplantation: (1) endothelial-dependent cGMP-mediated vasorelaxation (response to acetyl0003-4975/94/$7.00
Ann Thorae Surg 1994;58:86--92
choline); (2) endothelial-independent cGMP-mediated vasorelaxation (response to nitroprusside); and (3) l3-adrenergic cAMP-mediated vas ore laxation (response to isoproterenol). The second purpose of this study was to relate each of these mechanisms to cold ischemia alone and to cold ischemia plus reperfusion.
Material and Methods All animals received humane care in compliance with the "Principles of Laboratory Animal Care" formulated by the National Society for Medical Research and the "Guide for the Care and Use of Laboratory Animals" published by the National Institutes of Health (NIH publication 86-23, revised 1985).
Surgical Protocol After endotracheal intubation and mechanical ventilation, mongrel dogs of either sex and weighing 18 to 25 kg underwent autologous heart transplantation under sodium pentobarbital and halothane anesthesia through a sternotomy. Blood pressure was continuously monitored with an intraarterial catheter, and mean aortic pressure remained greater than 40 mm Hg throughout the experimental protocol. After systemic heparinization (heparin sodium, 250 U/kg), total cardiopulmonary bypass was initiated via cannulas placed in the ascending aorta and superior and inferior venae cavae. A pediatric bubble oxygenator was employed (Cobe Laboratories, Denver, CO). The ascending aorta was cross-clamped, and St. Thomas' crystalloid cardioplegic solution (4°C, 1,000 mL) was infused into the proximal aortic root. The heart was excised, placed in cold saline solution (4°C), and stored in a cooler filled with ice for 3 hours. The heart was reimplanted using running monofilament suture material for all surgical anastomoses. Approximately 20 minutes was required to reimplant the heart. The heart was reperfused on release of the aortic cross-clamp (mean arterial pressure, 45 to 55 mm Hg at the time of reperfusion) and allowed to beat on bypass. After 1 hour of reperfusion, cardiopulmonary bypass was discontinued, and the heart was rapidly excised. The dogs were therefore on bypass for 4 hours. The coronary sinus was directly cannulated, and cold saline solution (4°C, 250 mL) was allowed to run in under the force of gravity (low pressure) to flush all blood from the coronary circulation. Dogs serving as controls were not placed on cardiopulmonary bypass. In these dogs, after general anesthesia and sternotomy, the aorta was cross-clamped and St. Thomas' crystalloid cardioplegic solution We, 1,000 mL) was infused into the proximal aortic root, after which the heart was quickly excised, and coronary artery rings were prepared.
Coronary Artery Ring Preparation Coronary artery rings were dissected from the distal left anterior descending coronary artery with the use of a dissecting microscope. The rings were studied in three
FULLERTON ET AL CORONARY DYSFUNCTION IN TRANSPLANTED HEART
87
groups of hearts at different points in time: control (immediately after harvest after cold cardioplegia) (n = 3 dogs, 6 rings); at the end of cold ischemia (after 3 hours in saline solution at 4°C) (n = 6 dogs, 12 rings); and at the end of cold ischemia plus reperfusion (1 hour after heart reimplantation) (n = 5 dogs, 18 rings). Under dissecting microscope magnification, the surrounding tissue was dissected from the coronary arteries. Each of the arteries was cut into rings 3 to 4 mm wide. Great care was taken during this process to avoid endothelial injury. The coronary artery rings were suspended on fine wire tensiometers in individual lO-mL organ chambers. The organ chambers were surrounded by water jackets and continuously warmed to 3rc. The organ chambers were filled with Earle's balanced salt solution and continuously bubbled with gas consisting of 21% oxygen, 5% carbon dioxide, and balanced nitrogen. Earle's balanced salt solution is a standard physiologic salt solution and contains CaCl z (1.80 mmol/L), MgS0 4 (anhydrous) (0.83 mmol/L), KCl (5.36 mmol/L), NaCl 116.34 mmol/L), NaP04 (dibasic) (0.40 mmol/L), D-glucose (5.50 mmol/L), NaHC0 3 (19.04 mmol/L), and phenol red Na (as pH indicator) (0.03 mmol/L). Ring tension was determined by use of a forcedisplacement transducer (Grass FT03) attached to each tensiometer apparatus. Ring tension was thereby recorded using a MacLab Data Interface Module (ADI Instruments, Milford, MA) on a Macintosh IIci computer (Apple Computer, Inc, Cupertino, CA). The resting tension of coronary artery rings of this size at which a maximal contractile response to KCl (80 mmol/L) was obtained was found to be 4 g in a separate series of experiments. Once the rings had been allowed to reach a steady state (approximately 90 minutes) at a baseline tension of 4 g, the vasorelaxing effects of the pharmacologic agents were tested. After the coronary artery rings were preconstricted with KCl (40 mmol/L) and in a steady state, the vasorelaxing effects of acetylcholine, nitroprusside, and isoproterenol were determined in each ring in a random order by generating cumulative concentration-response curves (10- 9 to 10- 5 mol/L). After the test of each agent, the organ chambers were flushed several times and the rings allowed to reach a steady state before they were preconstricted once again with KCl (40 mmol/L) to test the next vasorelaxing agent.
Histologic Studies Histologic observation and immunofluorescence microscopy were carried out in coronary arteries from two control hearts and from three hearts after cold ischemia plus reperfusion. Hematoxylin and eosin staining was performed on cryopreserved sections of distal left anterior descending coronary arteries using standard technique. For immunofluorescence microscopic studies, distal left anterior descending coronary arteries approximately 1 to 1.5 mm in diameter were embedded in OCT compound and frozen in dry-ice-cold isopentane. Tissue blocks were stored at -70°C before use. Transverse sections 5 ILm thick were cut with a cryostat (Reichert-Jung 2800,
88
FULLERTON ET AL CORONARY DYSFUNCTION IN TRANSPLANTED HEART
Ann Thorae Surg
1994;58:86-92
*
100
100
o
oZ Z
Cold Ischemia
80
and Reperfusion
:;: ::E
OIl ~
::E
OIl 50
Z
sa z '" OIl
Cold Ischemia and Reperfusion
b
40
25
""
b
""
75
~
eo
z
sa Z '" OIl
2S 2S «
Cold Ischemia
20
ontrol
Cold Ischemia
9
Control
8
7
6
CONCENTRATION SODIUM NITROPRUSSIDE (-log molar) CONCENTRATION ACETYLCHOLINE (-log molar)
Fig 1. Effects of cold ischemia alone and cold ischemia plus reperfusian on endothelial-dependent cyclic guanosine monophosphate-mediated coronary vasorelaxation. Concentration-response curves of acetylcholine in left anterior descending coronary artery rings preconstricted with KC!. Significant dysfunction was produced by cold ischemia plus reperfusion but not cold ischemia alone. Data are shown as the mean ± the standard error of the mean. (Op < 0.05 versus conirol.)
Frigocut, Germany). Sections were collected on polylysine-coated slides and incubated with monoclonal antibody against alpha smooth muscle actin (BioGenex Laboratories, San Ramon, CA) and polyclonal antibody against von Willebrand factor (Endotech Corp, Indianapolis, IN) in 1% bovine serum albumin-phosphate-buffered saline solution for 60 minutes at room temperature. After washing with phosphate-buffered saline solution, sections were incubated with fluorescein-conjugated anti-mouse immunoglobulin G and Cy3-conjugated anti-rabbit immunoglobulin G in 1% bovine serum albumin-phosphatebuffered saline solution for 45 minutes at room temperature. After a thorough rinse with phosphate-buffered saline solution, sections were mounted with glycerolbased antiquenching medium. Microscopic observation and photography were then carried out.
Fig 2. Effects of cold ischemia alone and cold ischemia plus reperfusion on endothelial-independent cyclic guanosine monophosphaie-mediated coronary vasorelaxation. Concentration-response curvesof sodium nitroprusside in left anterior descending coronary artery rings preconstricted with KCl. Significant dYsfunction was produced by cold ischemia plus reperfusion but not cold ischemia alone. Data are shown as the mean ± the standard error of the mean. (Op < 0.05 versus contro!.)
ated relaxation (response to isoproterenol) produced 98% ± 4% maximal relaxation (Fig 3). After 3 hours of cold ischemia alone, neither endothelial-dependent cGMP-mediated (response to acetylcholine) (see Fig 1) nor endothelial-independent cGMPmediated (response to nitroprusside) (see Fig 2) coronary vascular smooth muscle relaxation was dysfunctional. However, at the end of 3 hours of cold ischemia alone, f3-adrenergic cAMP-mediated vasorelaxation (response to isoproterenol) was significantly impaired, as isoprotere-
l00¢----O"
\ -,
\\"\J l:J
Statistical Analysis Data are presented as the percent relaxation of KCIinduced ring tension produced by the given concentration of vasorelaxing agent. Values are expressed as the mean ± one standard error of the mean. Analysis of variance with post hoc Scheffe's F test was used. A p value of less than 0.05 was considered significant.
*
I'. \...
T 1
.\.\ J +
1
Rep:esion
~+
\\J 20
Cold Ischemia
~
1
Cold Ischemia
lJ················
o
Control
Results Significant coronary vasomotor dysfunction was produced by the processes of heart transplantation. In control coronary artery rings (studied after cold cardioplegia alone), endothelial-dependent cGMP-mediated relaxation (response to acetylcholine) produced 97% ± 3% maximal relaxation (Fig I), endothelial-independent cGMP-mediated relaxation (response to nitroprusside) produced 99% ± 3% maximal relaxation (Fig 2), and l3-adrenergic cAMP-medi-
CONCENTRATION ISOPROTERENOL (-log molar)
Fig 3. Effects of cold ischemia alone and cold ischemia plus reperjusion on j3-adrenergic cyclic adenosine monophosphate-mediaied coronary vasorelaxation. Concentration-response curves of isoproterenol hydrochloride in left anterior descending coronary artery rings preconstricted with KCl. Significant dysfunction was produced with cold ischemia alone and was exaggerated by reperfusion. Data are shown as the mean ± the standard error of the mean. (Op < 0.05 versus control; "p < 0.05 versus cold ischemia plus reperjusion.)
FULLERTON ET AL CORONARY DYSFUNCTION IN TRANSPLANTED HEART
Ann Thorae Surg 1994;58:86--92
89
ischemia plus reperfusion, there is a reduction in the density of von Willebrand factor staining, a finding consistent with a loss of coronary vascular endothelial cells, and substantial irregularity in the coronary vascular endothelial basal lamina. Structural changes in coronary vascular smooth muscle are also demonstrated on hematoxylin and eosin-stained sections (Fig 5). Under highpower light microscopy, a loss of order and distinction of coronary vascular smooth muscle cells is suggested after cold ischemia plus reperfusion compared with controls.
Comment A
Heart transplantation requires the transplanted heart to sustain the injurious processes of both cold ischemia and reperfusion. The results of the present study demonstrate that these processes culminate in significant coronary artery vasomotor dysfunction in the transplanted heart and also demonstrate that cold ischemia and reperfusion each produce different patterns of coronary vasomotor dysfunction. Neither endothelial-dependent cGMP-
B
Fig 4. Immunofluorescent staining of (A) control left anterior descending coronary artery and (B) left anterior descending coronary artery after ischemia plus reperfusion. Von Willebrand factor stains red, and vascular smooth musclealpha actin stains green. After ischemia plus reperfusion, there is partial loss of von Willebrand factor staining and substantial irregularity produced in the endothelial basal lamina, findings consistent with injury to the coronary vascular endothelium. (x 160 before 56% reduction.)
nol produced only 78% ± 3% maximal relaxation (p < 0.05) (see Fig 3). After reperfusion, there was significant dysfunction of endothelial-dependent cGMP-mediated coronary vascular smooth muscle relaxation, as acetylcholine produced only 26% ± 8% maximal relaxation (p < 0.05) (see Fig 1). Likewise, endothelial-independent cGMP-mediated coronary vascular smooth muscle relaxation was dysfunctional after reperfusion, as nitroprusside achieved only 53% ± 1% relaxation (p < 0.05) (see Fig 2). Further, the dysfunction of ,B-adrenergic cAMP-mediated vasorelaxation was significantly exacerbated by reperfusion, as isoproterenol produced only 43% ± 6% maximal relaxation (p < 0.05) (see Fig 3). These functional changes related to structural changes in both coronary vascular endothelium and vascular smooth muscle. Figure 4 demonstrates immunofluorescent staining of coronary vascular endothelial von Willebrand factor and alpha-actin antigen staining in coronary vascular smooth muscle. As seen in Figure 4, after cold
B
Fig 5. Hematoxylin and eosin staining of (A) control left anterior descending coronary artery and (B) left anterior descending coronary arteryafter ischemia plus reperfusion. In addition to critical irregularity in the endothelial basal lamina, there is the appearance of vacuoles in the vascular smooth muscle cells and loss of distinct smooth muscle cell orientation. (x400 before 56% reduction.)
90
FULLERTON ET AL CORONARY DYSFUNCTION IN TRANSPLANTED HEART
mediated relaxation (response to acetylcholine) nor endothelial-independent cGMP-mediated relaxation (response to nitroprusside) was impaired by cold ischemia alone, but both were significantly impaired after reperfusion. On the other hand, f:l-adrenergic cAMP-mediated relaxation (response to isoproterenol) was impaired by cold ischemia alone, and this impairment was significantly exacerbated by reperfusion. Thus, both of the principal intracellular pathways of coronary vascular smooth muscle relaxation (cGMP- and cAMP-mediated) were dysfunctional in the canine transplanted heart. In this study, coronary vasomotor control mechanisms were examined in epicardial coronary arteries. It is possible that the influences of cold ischemia and reperfusion are different in smaller, intramyocardial coronary vessels. However, Dignan and colleagues [5] found no difference in the responses to ischemia of proximal compared with distal coronary artery endothelial cell or smooth muscle function. Further, as suggested by Quillen and coworkers [6], smaller coronary vessels may be even more susceptible to ischemia and reperfusion than larger arterial vessels. Some investigators have noted that endothelialdependent coronary vasomotor dysfunction can result from crystalloid cardioplegic solutions containing potassium. However, the present study corroborates the findings of Evora and colleagues [7] in demonstrating that no endothelial-dependent coronary vasomotor dysfunction was produced by cold cardioplegia and hypothermia alone. Thus, the dysfunction seen in the present study might be due to the injuries of cold ischemia and reperfusion rather than to the cardioplegic solution. The findings of this study are in concert with those of other investigators. In prior studies [6, 8-12], endothelialdependent but not endothelial-independent vasomotor dysfunction has been produced by relatively short periods of ischemia followed by reperfusion. Fewer studies have examined the influence of ischemia alone on coronary vasomotor control. Dignan and associates [5] demonstrated porcine coronary endothelial-dependent vasomotor control was intact for up to 160 minutes of cold ischemia alone but became dysfunctional after longer periods of cold ischemia. Selke and coauthors [13] found dysfunction of endothelial-dependent vasomotor control in porcine coronaries after cold cardioplegia and reperfusion but not after cold cardioplegia alone. The design of the present study attempted to reproduce the clinical circumstances of heart transplantation. In so doing, it differs from prior studies by examining the influences of a much longer period of ischemia (3 hours) and reperfusion. This design also allowed us to examine the influence of cold ischemia alone and cold ischemia plus reperfusion on coronary vasomotor control. Using the transplanted heart, truly global myocardial ischemia was achieved, thereby avoiding the rich coronary collateral circulation of the dog, which can provide flow distal to the point of coronary artery occlusion [14]. Last, heart transplantation requires the use of cardiopulmonary bypass, which may exaggerate the severity of reperfusion injury.
Ann Thorae Surg 1994;58:86--92
Perhaps because of these differences, the present study demonstrated dysfunction of both endothelial-dependent and endothelial-independent mechanisms of coronary vasomotor control. These results suggest that under conditions of heart transplantation, typically requiring several hours of ischemia before reperfusion, coronary vasomotor dysfunction is greater in the transplanted heart than after shorter periods of regional or global myocardial ischemia and reperfusion. Further, the findings suggest that protracted cold ischemia and reperfusion each produce coronary vasomotor dysfunction, albeit different patterns of dysfunction. Prior studies that have examined coronary vessels by scanning electron microscopy after ischemia and reperfusion have yielded conflicting results. VanBenthuysen and associates [10] did find structural changes in coronary endothelial cells after 1 hour of ischemia and 1 hour of reperfusion but not after ischemia alone. However, neither Quillin and co-workers [6] nor Selke and colleagues [13] found structural changes in coronary vascular endothelium after ischemia and reperfusion. The present study suggests structural changes are found in both coronary vascular endothelium and smooth muscle cells after protracted cold ischemia plus reperfusion in the transplanted heart. These structural changes were related to functional changes in coronary vasomotor control and at least suggest the possibility that the functional changes resulted from the structural injury produced by cold ischemia plus reperfusion. In the present study, coronary vasomotor dysfunction was demonstrated acutely after heart transplantation, and this dysfunction was related to histologic evidence of coronary vascular damage. Other investigators [15] have demonstrated reendothelialization of arteries denuded of endothelium. In rat aorta, endothelium-dependent relaxation has been shown to return by 8 weeks after mechanical denuding of the endothelium [16]. The design of the present study does not address whether the vasomotor dysfunction seen acutely will resolve in the transplanted heart or what the time course of this process might be. There are several potential mechanisms by which protracted cold ischemia and reperfusion may culminate in this significant coronary vasomotor dysfunction. Vascular endothelial expression of leukocyte adhesion molecules has been demonstrated after ischemia, thus setting the stage for neutrophil-endothelial interaction on reperfusion [17]. Cardiopulmonary bypass (which must be used in heart transplantation) is known to produce complement activation [18] and neutrophil activation [19-21] and to increase the circulating levels of cytokines such as endotoxin [21, 22], interleukins [23, 24], and tumor necrosis factor [22]. Endotoxin in particular has been shown to "prime" both neutrophils and vascular endothelial cells, thus creating the potential for significantly greater damage on reperfusion of an ischemic vascular bed [25]. Thus, in the setting of heart transplantation, the coronary vascular bed not only is subjected to protracted cold ischemia, but is reperfused with activated neutrophils, activated complement, and cytokines. Together, these may reasonably be expected to produce greater injury on
Ann Thorac Surg 1994;58:86--92
reperfusion than transient coronary artery occlusion and reperfusion without cardiopulmonary bypass. In summary, this study demonstrated significant coronary vasomotor dysfunction in the transplanted canine heart. By using a model of autotransplantation, we avoided the possible influences of rejection. Three hours of cold ischemia alone was found to impair l3-adrenergic cAMP-mediated coronary vasorelaxation, and this dysfunction was exacerbated on reperfusion. Neither endothelial-dependent nor endothelial-independent cGMPmediated mechanisms of vasorelaxation were impaired by protracted cold ischemia alone, but both were significantly impaired after reperfusion. We conclude that such coronary vasomotor dysfunction may impair coronary vasodilatation and jeopardize coronary reserve in the transplanted heart. By limiting coronary reserve, this coronary vasomotor dysfunction may, in turn, compromise myocardial function of the transplanted heart. Supported by grant R29HL49398 from the National Institutes of Health.
References 1. Feigl EO. Coronary physiology. Physiol Rev 1983;63:1-205. 2. Berne RM, Rubio R. Coronary circulation. In: Berne RM, ed. Handbook of physiology: the cardiovascular system. Bethesda, MD: American Physiological Society, 1979:873. 3. Vane JR, Anggard EE, Botting RM. Regulatory functions of the vascular endothelium. N Engl J Med 1990;323:27-36. 4. Ignarro LJ. Biological actions and properties of endotheliumderived nitric oxide formed and released from artery and vein. Circ Res 1989;65:1-21. 5. Dignan RJ, Dyke CM, Abd-Elfattah AS, et al. Coronary artery endothelial cell and smooth muscle dysfunction after global myocardial ischemia. Ann Thorac Surg 1992;53:311-7. 6. Quillen JE, Selke FW, Brooks LA, Harrison DG. Ischemiareperfusion impairs endothelium-dependent relaxation of coronary microvessels but does not affect large arteries. Circulation 1990;82:586-94. 7. Evora PRB, Pearson PJ, Schaff HV. Crystalloid cardioplegia and hypothermia do not impair endothelium-dependent relaxation or damage vascular smooth of epicardial coronary arteries. J Thorac Cardiovasc Surg 1992;104:1365-74. 8. Hashimoto K, Pearson PJ, Schaff HV, Cartier R. Endothelial cell dysfunction after ischemic arrest and reperfusion. A possible mechanism of myocardial injury during reflow. J Thorac Cardiovasc Surg 1991;102:688-94. 9. Pearson PJ, Schaff HV, Vanhoutte PM. Acute impairment of endothelium-dependent relaxations to aggregating platelets following reperfusion injury in canine coronary arteries. Circ Res 1990;67:385-93.
FULLERTON ET AL CORONARY DYSFUNCTION IN TRANSPLANTED HEART
91
10. VanBenthuysen KM, McMurtry IF, Horwitz LD. Reperfusion after acute coronary occlusion in dogs impairs endotheliumdependent relaxation to acetylcholine and augments contractile reactivity in vitro. J Clin Invest 1987;79:265-74. 11. Ku DD. Coronary vascular reactivity after myocardial ischemia. Science 1982;218:576-8. 12. Mehta JL, Nichols WW, Donnelly WH, Lawson DL, Saldeen TGP. Impaired canine coronary vasodilator response to acetylcholine and bradykinin after occlusion-reperfusion. Circ Res 1989;64:43-54. 13. Sellke FW, Shafique T, Shoen FJ, Weintraub RM. Impaired endothelium-dependent coronary microvascular relaxation after cold potassium cardioplegia and reperfusion. J Thorac Cardiovasc Surg 1993;105:52-8. 14. Harken AH, Simson MB, Haselgrove J, Wetstein L, Harden WR, Barlow CH. Early ischemia after complete coronary ligation in the rabbit, dog, pig and monkey. Am J Physiol 1981;241:H202-1O. 15. Haudenshild CC, Schwartz SM. Endothelial regeneration. II. Restitution of endothelial continuity. Lab Invest 1979;41: 407-18. 16. Cartier R, Pearson PI, Lin PI, Schaff HV. Time course and extent of recovery of endothelium-dependent contractions and relaxations after direct arterial injury. J Thorac Cardiovase Surg 1991;102:371-7. 17. Lefer AM, Ma X-L. Cytokines and growth factors in endothelial dysfunction. Crit Care Med 1993;21:59-14. 18. Kirklin JK, Westaby 5, Blackstone EH, Kirklin JW, Chenoweth DE, Pacifico AD. Complement and the damaging effects of cardiopulmonary bypass. J Thorac Cardiovasc Surg 1983;86:845-57. 19. Faymonville ME, Pincemain J, Duchateau J, et al. Myeloperoxidase and elastase as markers of leukocyte activation during cardiopulmonary bypass in humans. J Thorac Cardiovase Surg 1991;102:309-17. 20. Gu YJ, van Oeveren W, Boonstra PW, de Haan J, Wildevuur CRH. Leukocyte activation with increased expression of CR3 receptors during cardiopulmonary bypass. Ann Thorac Surg 1992;53:839-43. 21. Kharazmi A, Andersen LW, Baek L, Valerius NH, Laub M, Rasmussen JP. Endotoxemia and enhanced generation of oxygen radicals by neutrophils from patients undergoing cardiopulmonary bypass. J Thorac Cardiovasc Surg 1989;98: 381-5. 22. Jansen NJG, van Oeveren W, Gu YJ, van Vliet MH, Eijsman L, Wildevuur CRH. Endotoxin release and tumor necrosis factor formation during cardiopulmonary bypass. Ann Thorae Surg 1992;54:744-8. 23. Butler J, Chong GL, Baigrie RI, Pillai R, Westaby 5, Rocker GM. Cytokine responses to cardiopulmonary bypass with membrane and bubble oxygenation. Ann Thorac Surg 1992; 53:833-8. 24. Downing SW, Edmunds LH Jr. Release of vasoactive substances during cardiopulmonary bypass. Ann Thorac Surg 1992;54:1236-43. 25. Anderson BO, Bensard DD, Brown JM, et al. FNLP injures endotoxin-primed rat lung by neutrophil-dependent and -independent mechanisms. Am J Physiol 1991;260:R413-20.
DISCUSSION DR CURTIS G. TRIBBLE (Charlottesville, VA): I enjoyed this
paper, and I also enjoyed the opportunity to review the manuscript. I have a bit of perspective to add to what this paper is all about, ie, organ preservation. For a long time, most people interested in organ preservation have been focusing on parenchymal preservation, such as preservation of the myocardium or preservation of energy stores, and not on endothelial preservation. We in our laboratory, Dr Fullerton in his laboratory, and
others have become interested in endothelial preservation because it seems that endothelial dysfunction probably is related to organ dysfunction, especially in the case of the heart. We have seen this relationship in our laboratory and presented the data, which show a correlation between endothelial dysfunction and cardiac dysfunction, at the American College of Surgeons meeting last year. The vascular dysfunction that Dr Fullerton has discussed that
92
FULLERTON ET AL CORONARY DYSFUNCTION IN TRANSPLANTED HEART
is secondary to cold ischemia and reperfusion further highlights the challenge of how to better preserve vascular function or how to overcome this vascular dysfunction. I believe these issues will be the challenges for organ preservation research in the immediate future. This study was very well designed and very well presented, and represents an important area of research in organ preservation. It is also interesting to note that this type of research is not going to be done very often in this way by our physiology colleagues or our cardiology colleagues. It is the kind of research done primarily by surgical investigators. I have a minor technical question. I was interested that you used a root perfusate gas with only 21% oxygen in it, and I noticed that you pointed out that that gave an oxygen tension normal for oxygenated blood. In this type of work, we have used a 95% oxygen and 5% carbon dioxide mix on the premise that oxygenated crystalloid at an oxygen tension of 100 mm Hg or so is relatively ischemic. I am curious why you chose your gas concentration, and I bet you looked at both 21% and 95%. Perhaps the 21% oxygen is even better than the higher oxygen tension. Maybe there is a lesson in this that we should be using the more hypoxic gases, not just in our research but in the early stages of reperfusion also. In a more general sense, I would appreciate hearing your thoughts on how you think your results will lead to better preservation of organs in this setting or on what you might be thinking of doing clinically. For instance, should we preserve organs at a warmer temperature, 18° or 15°C, instead of 4°C? Should we reperfuse them with lower oxygen tensions? DR RICHARD E. CLARK (Pittsburgh, PAl: My comment in part echoes that of Dr Tribble. For 20 years, we have been talking about myocardial preservation, and we have gotten pretty good clinically. I think that if we are going to do any better, the future focus clearly has to be on the vascular smooth muscle and endothelium. I have two questions for you. Did you have an opportunity to do any scanning electron microscopy on your rings? The reason I ask is that we can very dramatically show under various preservation systems that the endothelial cells can separate and expose basement membrane. If this occurs prior to reperfusion, there is going to be a problem with the myocardium in the immediate reperfusion interval. The second question has to do with using this model to test various pharmacologic agents as a protective mechanism in your cold storage system. Have you had the opportunity to study thromboxane A2 agonists and antagonists? DR HENDRICK B. BARNER (St. Louis, MO): I believe there are data indicating that hyperkalemic crystalloid perfusion damages
Ann Thorae Surg 1994;58:86-92
the endothelium but that hyperkalemic blood cardioplegia does not, the critical range of potassium being 20 to 30 mEq. Could you comment? DR D. GLENN PENNINGTON (St. Louis, MO): Considering the drugs we know so well clinically, particularly nitroprusside and isoproterenol, could you give us some idea of the correlation between these dose ranges and what we might be using clinically? Would these doses be realistic in treating patients? DR FULLERTON: Dr Tribble, thank you very much for reviewing our manuscript. We do appreciate your comments, as we are aware of the nice work your group has done. Our standard ring preparation calls for a 21% oxygen gas. I wish I had some good reason to give you for that choice, but it is simply our standard model. It is chosen because at our altitude, that will give an oxygen tension of just under 100 mm Hg, about 98 mm Hg. Dr Clark, we did not have the opportunity to use scanning electron microscopy. However, we were pleased that our immunofluorescent and hematoxylin and eosin stainings are consistent with the findings shown by electron microscopy. Typically what is described, as you pointed out, is denuding of the endothelium or, of even more interest, the crypts that seem to form in the endothelial layer, which I believe we demonstrated as well. In our coronary ring model, we have not examined the influences of thromboxane, but we do have a line of investigation in our laboratory examining this in pulmonary artery rings. Dr Barner, as you know, there are several articles in the literature that attribute some endothelial dysfunction to crystalloid cardioplegia. We did not find that. The potassium in our cardioplegia is about 16 mEq/L, which is a bit on the low side compared with some of the other studies. However, our findings were actually quite similar to those of Dr Shaft's group, who found no dysfunction using a similar concentration of potassium. So I do not think our results are attributable to that. Although the data are not presented, when we have examined coronary rings without the influence of cardioplegia, we have found doseresponse curves that are identical to those of our control group. Dr Pennington, it is difficult to know exactly how these doses relate to blood levels of these drugs. Acetylcholine is not a very commonly used drug, although nitroprusside is. The half-life of nitroprusside is so rapid that it is very difficult to measure blood levels, and in fact, having sought this information, I have been unable to find it. The half-life of isoproterenol is about 2 minutes, and so at doses of about 0.01 /Jog' kg-I. min-I, which is relatively frequently used in our pediatric population, we have found, or the pharmacists tell us, that we are generating blood levels comparable with about 10- 7 mollL in our baths.