H centrations of 10 to 110 mmoyl have become standard

Transcription

1 Impairment of Vascular Endothelial Function by High-Potassium Storage Solutions Barry B. K. Chan, MD, Irving L. Kron, MD, Terry L. Flanagan, MPH, John A. Kern, MD, Charles E. Hobson, MD, and Curtis G. Tribble, MD Department of Surgery, University of Virginia Health Sciences Center, Charlottesville, Virginia High-potassium cold storage solutions are currently used to preserve myocardial function during heart transplantation. However, the effects of high potassium concentration on vascular endothelial function are not well known. We therefore tested vascular rings for endothelial-dependent and endothelial-independent relaxation during storage in normokalemic, normothermic buffers and then in buffers supplemented with 10 to 110 mmoyl KCl. Maximal endothelial-dependent relaxation was significantly reduced at all high potassium concentrations. Endothelial-independent relaxation was impaired only with 80 and 110 mmoyl KCl buffers. Both endothelialdependent relaxation and endothelial-independent relaxation returned to normal values after washout of excess potassium. Similarly, endothelial-dependent relaxation and endothelial-independent relaxation were assessed in rings after 24 hours of hypothermic storage in normokalemic Krebs buffer, and in buffers containing 20 and 110 mmoyl KC1. Maximal endothelial-dependent relaxation was significantly reduced after preservation in the high-potassium solutions, whereas endothelialindependent relaxation was not impaired. We conclude that there is significant impairment of endothelial function after cold storage in a high-potassium buffer. Inadequate washout of potassium during normothermic conditions may lead to further functional impairment of vascular responsiveness. A low-potassium storage medium is recommended for improved vascular protection. (Ann Thorac Surg 1993;55:940-5) igh-potassium storage solutions with potassium con- H centrations of 10 to 110 mmoyl have become standard preservation modalities for both heart and other solid-organ transplantation [l]. These storage solutions, such as Plegisol (Abbott Pharmaceutical, Chicago, IL) (16 mmol/l potassium) and the University of Wisconsin solution (110 mmoll potassium), have been demonstrated to have improved parenchymal cellular protection against prolonged ischemia [l, 21. However, until recently, relatively little attention has been directed toward the preservation of normal vascular function, particularly endothelial-dependent vascular function [3, 41. The vascular endothelium, through its release of endothelium-derived relaxing factor (EDRF), is now recognized to be a crucial mediator of vascular tone [5, 61. Furchgott and Zawadski [5] demonstrated that relaxation of isolated rabbit aorta by acetylcholine or its analogues (methacholine) required the presence of intact endothelial cells. Acetylcholine and many other vasoactive substances have been shown to act on specific cell surface receptors to stimulate endothelial cells to release EDRF [6]. Endothelial-derived relaxing factor then activates and increases the production of intracellular cyclic guanosine monophosphate, which ultimately leads to smooth muscle relaxation. Vasodilators that act directly on the vascular smooth muscle independent of the presence of endothelial cells include agents such as nitroprusside and papav- Accepted for publication July 23, Address reprint requests to Dr Tribble, Department of Surgery, Box 181, University of Virginia Health Sciences Center, Charlottesville, VA erine. Disruption of the endothelial cell layer during surgical manipulations or as a result of preservation techniques not only leads to thrombogenesis but also may impair EDRF release and contribute to clinically significant vasospasm [7-91. Although several studies have suggested that hyperkalemic crystalloid cardioplegia perfusion leads to morphological and functional endothelial injury, effects of high-potassium preservation on endothelial-dependent vascular responsiveness are not known [4, The present study was designed to test the effects of high potassium concentrations currently used in standard preservation solutions on endothelial function and vascular reactivity. The effect of potassium was first evaluated as an isolated variable, in the absence of ischemia and hypothermia, because both these factors have been shown to cause endothelial dysfunction [6, 81. Endothelial-dependent and endothelial-independent vascular relaxations were then determined after hypothermic storage for 24 hours in high-potassium solutions. Material and Methods Infrarenal abdominal aortas were isolated from 2- to 5-kg New Zealand White rabbits, carefully cleaned of fat and adventitia, and sectioned into 2- to 3-mm vascular rings. The rings were suspended in water-jacketed tissue baths containing 8 ml Krebs-Henseleit buffer (in mmovl: NaC1, 122; KC1, 4.7; MgSO,, 1.2; glucose, 11; NaHCO,, 22; KH2P0,, 1.2; and CaCl,, 0.5) at 3TC, aerated with 95% 0,/5% CO,, and connected to Grass FT03 force transduc- Q 1993 by The Society of Thoracic Surgeons %. 00

2 Ann Thorac Surg 1993;55:940-5 CHANETAL 941 POTASSIUM IMPAIRS ENDOTHELIAL FUNCTION Duncan s multiple range test for multiple comparisons. Data were computed using the Statistical Analysis System (Cary, NC) program. A probability value of less than 0.05 was considered statistically significant. All values were reported and graphed as mean * standard error of the mean. Care of animals was in compliance with the Guide for the Care and Use of Laboratory Animals established by the National Institutes of Health (NIH publication no , revised 1985) K CONCENTRATION (mm) Fig 1. Mean developed tension of vascular rings perfused in buffers supplemented with increasing potassium (KJ concentrations. ers (Grass Instruments Co, Quincy, MA) for isometric tension recordings. Optimal resting tension was determined by preliminary length-tension experiments. After 1 hour of equilibration at 2 g resting tension, preconstriction was applied with a median effective dose of phenylephrine to achieve approximately 75% maximal tension. Cumulative dose-dependent relaxation response curves were then determined to methacholine to evaluate for endothelial-dependent relaxation (1 x to 5 x lo- movl) and to sodium nitroprusside (5 x moyl) to test for endothelial-independent relaxation [ 141. All drugs were obtained from Sigma Chemical Company (St. Louis, MO) and were prepared daily in Krebs buffer. After control values were obtained, the baths were replaced with Krebs buffer containing either 10, 20, 80, or 110 mmovl KCI. Relaxation responses to the potassiuminduced contractions (without phenylephrine) were repeated for methacholine and nitroprusside after 1 hour of incubation in these high-potassium buffers. Excess potassium was then washed out and vascular relaxation to phenylephrine contraction was again determined after reequilibration in normokalemic buffer. Vascular rings were then subjected to 24 hours of hypothermic (4 C) storage in normokalemic, nonaerated Krebs, and in Krebs buffers supplemented with 20 mmol/l and 110 mmol/l KCI. At the completion of each cold storage period, the excess potassium was flushed out with Krebs buffer. Endothelial-dependent and endothelial-independent relaxations were then determined to phenylephrine contraction after equilibration in normothermic, oxygenated buffer. Scanning electron microscopy of endothelial surfaces was performed for each experimental group with the JEOL 35C microscope. Endothelial cell borders were stained with a 0.25% silver nitrate solution before tissue fixation [9]. The micrographs were interpreted by us, and were also reviewed by a blinded observer. Each experimental group consisted of 16 rings from 4 animals. Data are expressed as dose-dependent percent relaxation to either potassium or phenylephrine contraction as described for each experimental group. Statistical analysis was performed using analysis of variance with Results The effects of high potassium concentration on vascular reactivity were determined by the relative degree of contraction to potassium alone and in response to phenylephrine stimulation, endothelial-dependent relaxation with methacholine, and endothelial-independent relaxation with nitroprusside. There was an increase in relative developed tension of rings bathed in solutions supplemented with increasing potassium concentration (Fig 1). Ten millimoles per liter of KCl caused minimal constriction of 1.0 & 0.5 g. Vasoconstriction appeared to reach a maximal level of 29.9 * 1.5 g with 80 mmoyl KCl. Endothelial-dependent relaxation to potassium-induced contraction was significantly impaired for all the potassium concentrations tested at methacholine doses of 1 x to 5 x movl (Fig 2). Potassium concentrations as low as 10 mmoyl had adverse effects on relaxation response. Solutions containing 110 mmol/l allowed for virtually no relaxation. Vascular relaxation returned to control values ( p = not significant) after washout of the excess potassium. Endothelial-independent relaxation to nitroprusside (5 x lop6 movl) was significantly reduced only at 80 and 110 mmol/l KCI (Fig 3). These also returned to control values (p = not significant) after washout of excess potassium. Vascular rings tested after hypothermic storage for 24 hours in high-potassium buffers followed by washout of the excess potassium exhibited a significant increase in mean developed tension in response to 1 x lo- movl phenylephrine as compared with rings stored in z P 4 X -1 1u 8ot c 60 a Pc0.05 COMPARED TO CONTROL (4.7 mm KCLJ LOG METHACHOLINE (MJ CONTROL 10 mm K 20 mm K 80 mm K 110 mm K Fig 2. Endothelial-dependent relaxation of vascular rings perfused in buffers supplemented with increasing potassium (K) concentrations.

3 CHAN ET AL POTASSIUM IMPAIRS ENDOTHEUAL FUNCTION Ann Thorac Surg 1993;55:94%5 80 r 0 * PcO.05 COMPARED TO CONTROL (4.7 mm KCL) Cmbd K CONCENTRATION (mm) Fig 3. Endothelial-independent relaxation of vascular rings perfused in buffers supplemented with increasing potassium (K) concentrations. normokalemic Krebs solution (Fig 4). Furthermore, endothelial-dependent relaxation was sigruficantly reduced at the maximum dose of methacholine (Fig 5). There were no differences in percent relaxation at the lower doses as depicted in the dose-response curves (see Fig 5). Endothelial-independent relaxation was not affected by hypothermic storage in high-potassium solutions. Scanning electron microscopy revealed sigruficant endothelial disruption after hypothermic storage in buffers containing 110 mmol/l potassium (Figs 6, 7). Storage in low-potassium solutions adequately preserved endothelial cell architecture (Figs 8, 9). Comment The effects of high-potassium solutions on endothelial function and on vascular reactivity were examined in this series of experiments, which attempted to simulate organ preservation modalities whereby organ parenchymal cells and the vasculature are flushed and then bathed in standard high-potassium storage solutions. Significant z P + 60 < X 5 40 w K s CONTROL K-20 K-110 P10.05 COMPARED TO CONTROL (4.7 mm KCLI " LOQ METHACmlLlNE (MI Fig 5. Endothelial-depend.& relaxation after hypothermic storage for 24 hours in buffers supplemented with increasing potassium (K) concentrations. Control groups consisted of vascular rings stored for 24 hours in normokalemic Krebs buffer at 4 C. impairment of vascular reactivity was observed when relaxation responses were tested in rings during incubation in high-potassium buffers at 37 C. Vascular reactivity returned to normal after adequate washout of the excess potassium. Extreme hyperkalemia of 80 and 110 mmovl. KCl not only promoted intense vasoconstriction but markedly reduced endothelial-dependent and endothelialindependent relaxation responses. Potassium concentrations as low as 10 mmol/l had adverse effects on endothelial-dependent relaxation. Effects of potassium were initially examined as an isolated variable, and the vessels were not exposed to ischemia or hypothermia. Both hypoxia and cold storage are factors in clinical preservation techniques. However, they have also been implicated in endothclial injury [6, 8, 151. Our data also indicate that hypothermic storage in high-potassium solutions potentiates vascular constriction but without dramatic impairment of endothelialdependent and endothelial-independent relaxation. En * P <0.05 COMPARED TO COMRM (4.7 mm KCL) * " CONTROL K CONCENTRATION (mm) Fig 4. Mean developed tension to 1 x molll phenylephrine after hypothermic storage for 24 hours in buffers supplemented with increasing potassium n0 concentrations. Tension was determined after washout of excess potassium and reequilibration in normokalemic buffer. Fig 6. Scanning electron micrograph of a noml rabbit aorta endothelium with intact endothelial cell borders, as outlined by the arrows. (Line = 70 pn.)

4 Ann Thorac Surg 1993;55:940-5 CHANETAL 943 POTASSIUM lmpalrs ENDOTHELLAL FUNCTION Fig 7. Scanning electron micrograph of a representative endothelial surface ufter 24 hours of hypotherrnic storage in 110 mmolll KCI brt,fer. There is significant disruption of the endothelial cell structure with feza intact cell borders. (~2,200 before 41% reduction; line = 10 pm.) Fig 9. Scanning electron micrograph of a rcpreserrtative endothelid surface after 24 hours of hypotherrnic storage in 20 mm11lll KCI bufer. There is minimal disruption of eridothelinl cell structure. Intact cell borders arc indicated l7y arrows. (x 1,800 before 41 FJ reduction; line = 10 pi.) dothelial-dependent relaxation was reduced at the highest dose of methacholine only, suggesting that the maximum relaxing capacity after hyperkalemic storage was diminished. The degree of injury attributed to hypothermia alone is difficult to measure as cold storage in Krebs buffer alone caused minimal structural and functional impairment. As previously noted by other investigators, impairment of endothelial function may not parallel structural injury to endothelial cells (11, 131. The rationale for potassium cardioplegia was based on studies that showed conservation of cellular energy supplies with rapid induction of cardiac arrest [16]. Standard Fig 8. Scanning electron micrograph of a representative endothelial surface after 24 hours of hypothermic storage in Krebs hufer (4.7 mmolll KCI). There is minimal disruption of endothelial cell structure. Intact cell borders are indicated by arrows. 0(1,800 before 41% reduction; line = 10 pm.) cardioplegia solutions such as Plegisol and Stanford solution contain 16 to 30 mmol/l potassium. The argument for using high-potassium preservation solutions was based on the early studies by Collins and associates [17] in 1969 that suggested an intracellular electrolyte medium (high potassium, low sodium) would limit potassium leakage from cells, decrease the activity of the Na ' -K ' adenosine triphosphatase pump, and conserve high-energy phosphates. Euro-Coliins and University of Wisconsin solution contain 110 to 140 mmol/l potassium. As methacholine has been demonstrated to induce vascular smooth muscle relaxation via EDRF [14], the impairment of endothelial-dependent relaxation during incubation in the high-potassium medium is most likely due to disturbance of endothelial function by high potassium concentrations. Furchgott and colleagues [5, 61 observed that buffers supplemented with 12 to 18 mmol/l KC1 reversibly attenuated endothelial-dependent relaxation to acetylcholine and suggested that this may be due to an impaired release of endothelium-derived hyperpolarizing factor. The return of normal relaxation after washout of the excess potassium may be explained by the effect of membrane depolarization on endotheliumdependent responses during exposure to the high potassium concentrations [5]. The observed results could also be attributed to differing levels of precontraction. Although considerable controversy exists, several investigators contend that the degree of relaxation is inversely related to the degree of precontraction irrespective of the amount of EDRF secreted [18]. It is possible that attenuation in both endothelial-dependent and endothelialindependent relaxation was due in part to the increased developed tension ("precontraction state") in response to high potassium concentrations, and not due to a specific functional endothelial or smooth muscle injury. Although the mechanisms are debatable, it is important to recognize

5 944 CHANETAL POTASSIUM IMPAIRS ENDOTHELJAL FUNCTION Ann Thorac Surg 1993;55:94&5 that significant vascular dysfunction does occur when the vasculature is bathed in a hyperkalemic medium. Several other studies have provided support of morphologic and functional endothelial damage after hyperkalemic crystalloid cardioplegia perfusion. Follette and co-workers [lo] described the separation of endothelial cells from the surface lamina in femoral arteries perfused with crystalloid cardioplegia (30 meql KCl). Hoover and colleagues [ll] observed an intense adventitial fibrotic reaction in arterialized saphenous vein grafts 6 weeks after potassium (40 meql) cardioplegic perfusion, even though there was no functional impairment of prostacyclin production from the endothelium. Hajula and associates [ 121 acknowledged the importance of endothelial lesions in the development of atherosclerosis and found that crystalloid cardioplegia (20 meql KC1) perfusion could cause gross deformity of the coronary endothelium. De Caterina and co-workers [13] studied prostacyclin production by saphenous vein bypass grafts as a marker of endothelial cell function, and demonstrated that prostacyclin production decreased with 40 and 80 meql KCl cardioplegic perfusion, but not with 20 meql KCl. They found that the endothelial coverage was intact at all potassium concentrations tested. More recently, Saldanha and Hearse [4] observed that coronary vascular responsiveness to perfusion with 5-hydroxytryptamine (an endothelial-dependent vasodilator) was significantly diminished after hypothermic cardioplegic (25 meql KCl) infusion. On the other hand, Mankad and associates [19] provided evidence that cardioplegia containing 20 mmoe potassium maintained the vasodilatory response of the coronary vasculature to both 5-hydroxytryptamine and nitroglycerin, an endothelial-independent vasodilator. Although the physiologic role of EDRF remains to be completely clarified, it is clear that the vascular endothelium plays a significant role in hemostasis and in the control of vascular tone [6]. Endothelial cells have receptors for a large number of physiologic vasoactive substances that stimulate the synthesis and release of EDRF [5,6]. Disruption of the endothelium could lead to thrombosis, graft occlusion, atherosclerosis, vasospasm, and other forms of vascular dysfunction. The value of protecting the vascular endothelium is further supported by recent studies that correlated endothelial dysfunction with postischemic organ dysfunction [3, 201. We recognize the limitations of a rabbit aorta vascular ring model. Large vessels (such as the aorta) and small resistance vessels (the microvascular circulation) may not have analogous endothelial vasodilatory properties. This study does not allow for direct extrapolation of results from aortic rings to the microcirculation. Furthermore, there is considerable heterogeneity in relaxation responses among different vascular beds. However, our data suggest that further studies using vascular beds are necessary to examine the potentially detrimental effects of high potassium concentrations on endothelial function. This study provides evidence that the high potassium concentrations in currently available organ preservation solutions significantly impair endothelial function and vascular smooth muscle reactivity. The impetus to use a low-potassium medium is further supported by a recent report that showed that a high-sodium, low-potassium (140 mmol/l sodium, 9 mmol/l potassium) variation of the widely used University of Wisconsin solution is equally or even more effective for solid-organ preservation when compared with the original high-potassium, low-sodium (120 mmom potassium, 30 mmol/l sodium) version [21]. An explanation for the effectiveness of the low-potassium University of Wisconsin solution may be improved protection of vascular function. Our data also suggest that inadequate washout of a high-potassium solution may lead to impaired vascular function. Current preservation techniques for heart and other solid-organ transplantation include flushing of the vasculature with a cold preservation solution such as cardioplegia, Euro-Collins, or University of Wisconsin solution as the organ is harvested. The organ is then stored under hypothermic conditions with the same preservation solution bathing the vasculature. Efficient evacuation of potassium from the entire vasculature may be difficult but is probably necessary for normal vascular function. A low-potassium storage medium may provide improved protection of the vascular endothelium. References 1. Belzer FO, Southard JH. Principles of solid-organ preservation by cold storage. Transplantation 1988;45:67M. 2. Ploeg RJ, Goossens D, McAnulty JF, Southard JH, Belzer FO. Successful 72-hour cold storage of dog kidneys with UW solution. Transplantation 1988;46: McKeown CMB, Edwards V, Philips MJ, Harvey PRC, Petrunka CN, Strashrg SM. Sinusoidal lining cell damage: the critical injury in cold preservation of liver allografts in the rat. Transplantation 1988;46: Saldanha C, Hearse DJ. Coronary vascular responsiveness to 5hydroxytryptamine before and after infusion of hyperkalemic crystalloid cardioplegic solution in the rat heart. J Thorac Cardiovasc Surg 1989; Furchgott RF, Zawadski JV. The obligatory role of endothelial cells in the relaxation of arterial smooth muscle by acetylcholine. Nature 1980;288: Furchgott RF, Vanhoutte PM. Endothelium-derived relaxing and contracting factors. FASEB J 1989;3: Johns RA, Peach MJ, Flanagan TL, Kron IL. Probing of the canine mammary artery damages endothelium and impairs vasodilation resulting from prostacyclin and endotheliumderived relaxing factor. J Thorac Cardiovasc Surg 1989;97: Neil DAH, Pollock GA, Moiyneux GS, Hardie IR. Vascular changes following hypothermic preservation. Transplant Proc 1989;21: Barone GW, Conerly MJ, Farley PC, Flanagan TL, Kron IL. Endothelial injury and vascular dysfunction associated with the Fogarty balloon catheter. J Vasc Surg 1989; Follette DM, Buckberg GD, Mulder DG, Fonkalsrud EW. Deleterious effects of crystalloid hyperkalemic cardioplegic solutions on arterial endothelial cells. Surg Forum 1980;31: Hoover EL, Pett SB, Eldor A, et al. The effects of crystalloid potassium cardioplegic solution on arterialized canine vein grafts: assessment of chronic prostacych production and histopathologic alterations. Circulation 1981;64(Suppl 2): Hajula A, Mattila S, Mattila I, et al. Coronary endothelial damage after crystalloid cardioplegia. J Cardiovasc Surg 1984;25: De Caterina R, Weksler 8, Alonso D, Cruz-Bracho MR,

6 Ann Thorac Surg 1993; CHANETAL 945 POTASSIUM IMPAIRS ENDOTHELIAL FUNCTION Subramanian VA. Functional endothelial damage by highpotassium cardioplegic solutions to saphenous vein bypass grafts. Surgery 1985;98: Singer HA, Peach MJ. Calcium- and endothelial-mediated vascular smooth muscle relaxation in rabbit aorta. Hypertension 1982;4(Suppl 2): Furusho N, Araki H, Katsuhide N, Miyauchi Y. Effect of cold cardioplegic solution and hypothermia on response to acetylcholine in perfused epicardial coronary artery of pig. Ann Thorac Surg 1989;47: Hearse DJ, Stewart DA, Braimbridge MV. Hypothermic arrest and potassium arrest. Metabolic and myocardial protection during elective cardiac arrest. Circ Res 1975;36: Collins GM, Bravo-Shugarman M, Terasaki PI. Kidney preservation for transportation. Lancet 1969;2: Dainty IA, McGrath JC, Spedding M, Templeton AGB. The influence of the initial stretch and the agonist-induced tone on the effect of basal and stimulated release of EDRF. Br J Pharmacol 1990;100: Mankad PS, Chester AH, Yacoub MH. Role of potassium concentration in cardioplegic solutions in mediating endothelial damage. Ann Thorac Surg 1991;51: Yanaga K, Shimada M, Makowka L, Higashi H. The effect of preservation fluid on the blood flow of pediatric liver allograf ts. Transplantation 1989;48: Moen J, Claesson K, Pienaar H, et al. Preservation of dog liver, kidney, and pancreas using the Belzer-UW solution with a high-sodium and low-potassium content. Transplantation 1989;