- Email: [email protected]
Coronary vascular responsiveness to 5-hydroxytryptamine before and after infusion of hyperkalemic crystalloid cardioplegic solution in the rat heart
J THORAC
CARDIOVASC SURG
1989;98:783-7
Coronary vascular responsiveness to 5-hydroxytryptamine before and after infusion of hyperkalemic crystalloid cardioplegic solution in the rat heart Possible evidence of endothelial damage Using the isolated, Langendorff-perfused rat heart (n = 8 per group), we have studied the effects of 5-hydroxytryptamine and papaverine on coronary flow before and after a 30-rninute infusion of hypothermic (200 C), nonoxygenated cardioplegic solution containing potassium in a concentration of either 25 or 40 mmoljL. Before infusion of the 25 mmoljL potassium cardioplegic solution, both 5-hydroxytryptamine (1 X 10-7 moljL) and papaverine (5 x 10-6 moljL) caused similar increases in Bow (+21.2% ± 1.6% and +22.8% ± 1.6%, respectively). After cardioplegia, the vasodilatory response to 5-hydroxytryptarnine was completely lost and a slight vasoconstriction was observed (-0.2 % ± 1.2 %). However, there was no significant change in the response to papaverine, which maintained a +23.7% ± 1.4% vasodilation. With the cardioplegic solution containing a 40 mmoljL concentration of potassium, the initial responses to 5-hydroxytryptarnine and papaverine were again similar (+21.5% ± 2.5% and +24.1 % ± 3.0%, respectively). After cardioplegia, 5-hydroxytryptamine caused a significant vasoconstriction (-4.3% ± 1.1 %), whereas the response to papaverine was again maintained (+19.1 % ± 2.3%). The results of this study support the concept that hyperkalemic crystalloid cardioplegic solutions cause vascular damage possibly involving the endothelium or its f\DICtion, which may adversely affect vascular responsiveness.
Clyde Saldanha, BSc, MB, BS, and David J. Hearse, BSc, PhD, DSc, FACC, London. England
Over the past decade the use of cardioplegic solutions has gained wide acceptance as a method for intraoperativemanagement of the heart during cardiac operations. Their effectiveness in protecting the myocyte against prolonged ischemia has been widely established by extensive laboratory and clinical studies.' The rationale and aims in the use of cardioplegic solutions have been
FromCardiovascular Research, Rayne Institute, St. Thomas' Hospital, London SEI 7EH, England. Thiswork was supported in part by grants from STRUTH and the National Heart Research Fund. Received for publication Oct. 6, 1988. Accepted for publication Jan. 31, 1989. Address for reprints: Professor D. J. Hearse, Cardiovascular Research, Rayne Institute, St. Thomas' Hospital, London SEI 7EH, England. 12/1/11814
reviewed by several authors.i" Significantly, these reviews have mainly concentrated on the effects of cardioplegic solutions on the myocyte and hence on the mechanical performance of the heart. Very little attention has been focused on the effects of these solutions on the conduction or the vascular tissue. However, if the heart is to be protected optimally and good function achieved postoperatively, it is imperative that all tissues be protected. With regard to the vascular tissue, the endothelial lining is of importance because of the fundamental role it plays in hemostasis. Intact endothelium produces prostacyclin, antithrombin III, and plasminogen activator. 5,6 By such mechanisms, platelet aggregation and fibrin deposition on the vessel wall are prevented. Conversely, injury to the endothelium exposes collagen fibers and basement membrane, which are potent stimuli for platelet aggregation and adhesion." Disrupted endothelial cells may also release tissue procoagulant, which further facil783
The Journal of
784
Thoracic and Cardiovascular
Saldanha and Hearse
Surgery
Table I. Composition of cardioplegic solutions ~
o
Potassium
....
-'
25 mmol/L
>a:
...:
Cl
Sodium chloride (mmol) Potassium chloride (mmol) Magnesium chloride (mmol) Calcium chloride (mmol) Sodium bicarbonate (mmol)
...:
pH
z o a: o
u
~
w
z
:I: U
110.0
25.0 16.0 1.2 10.0 7.8
40 mmol/l. 110.0 40.0 16.0 1.2 10.0 7.8 .
.... ~ o a:
w
"'"
10
~
o
....
-'
>a:
...:
z
o
a:
o u
~
w
Cl
z
...:
:t: U
.... Z
w
u
a:
that damage to the endothelium may result in pathologic responses to normal stimuli, for example, vasospasm in response to 5-HT released from aggregating platelets. I I In a previous study, we I2 have shown that 5-HTinduced vasodilation in the isolated rat heart can be abolished by quinacrine, which is a recognized EDRF antagonist. This study and that ofothers 13 provide evidence that EDRF is released in the perfused isolated heart in response to 5-HT.ln view of the important role of the endothelium and the potentially serious consequences of damage to it, we were interested in studying the effect of cardioplegic solutions per se (in the absence of the injurious effects of ischemia) on responses thought to be mediated by the endothelium.
w
"'"
10
Fig. 1. Dose-response characteristics for 5-HT and papaverine (PAP). Hearts were subjected to perfusion with various concentrations of 5-HT (A)and PAP (B). Changes in coronary flow were measured and expressed as a percent of drug-free control hearts. Each result represents the mean of eight hearts and the bars represent the standard errors of the mean.
itates the cascade to thrombus formation," Injury to the endothelium is also a key step in the initiation of atherosclerosis.' Platelets stick to exposed areas and release factors that stimulate the proliferation of intimal smooth muscle. When the endothelium eventually regenerates to cover areas previously denuded, the intima underlying regenerated endothelium is substantially thicker and more likely to accumulate lipid than is the intima under undamaged endothelium.t-" Besides its role in hemostasis, the endothelium may influence vascular tone by the release of an endotheliumdependent relaxant factor (EDRF).8,9 Although EDRF has been shown to be released in vessels in vivo'? and in response to chemicals such as 5-hydroxytryptamine (5HT) and norepinephrine," its physiologic role remains to be clearly established. However, it has been postulated
Methods Animals and reagents. Hearts were obtained from male Sprague-Dawley rats (200 to 250 gm body weight) that had been maintained on a standard laboratory diet. 5-HT hydrochloride hemihydrate (serotonin) was obtained from Aldrich Chemical Company Ltd., Gillingham, Dorset, England, and papaverine hydrochloride was obtained from Sigma Chemical Company, St. Louis. The cardioplegic solutions were prepared by adding potassium chloride to the St. Thomas' Hospital cardioplegicsolution (Plegisol,Abbott Laboratories, North Chicago) to obtain final potassium concentrations of 25 or 40 rnrnol/ L. The final composition of both solutions is shown in Table I. Perfusion technique. Rats were anesthetized with diethyl ether, the right femoral vein was dissected out, and 200 IU of heparin was injected into the vein. After 30 seconds,the thorax was opened and the heart was excised and placed into a beaker containing cold (4° C) Krebs-Henseleit bicarbonate buffer.14A modifiedLangendorff preparation 15 was used;this wasdesigned to allow accurate measurement of coronary flow rates under conditions of constant perfusion pressure (60 em H 20 ). Hearts were perfused via the aorta with ultrafiltered (5.0 /lm porosity filter) buffer containing glucose (11.1 mmol/L) and gassedwith 95% oxygen plus 5% carbon dioxide (pH 7.4 at 37° C). Drugs were dissolved in the buffer to give optimally active concentrations (derived from dose-response studies) of 5-HT (I X 10- 7 moljL) and papaverine (5 X 10-6 moljL). Experimental time course. After 15 minutes of control perfusionwith oxygenated Krebs-Henseleit bicarbonate buffer (60 em H 20 , 37° C during which coronary flow was continually measured with a graduated cylinder), hearts were perfused aerobically with buffer containing either 5-HT or papaverine for a
Volume 98 Number 5. Part 1
Cardioplegia and vascular injury
November 1989
:: 0
30
....I
>-
z
PAPb
S-HTb
u,
c:: ocr
25 mM POTASSIUM CARDIOPLEGIA
.,..
PAPa
10
c: 0 o
z
ocr
0
W
o z
S-HTa
J:
C,,)
•~
10
Z
W
o
20
0
Z
Z
PAPb
S-HTb
U.
> c:: ocr
c:: C,,)
30
....I
20
0
0
:: 0
40 mM POTASSIUM CARDIOPLEGIA
785
0
ocr
.o J:
-10
~
Fig. 2. The effect of 30 minutes' infusion of potassium cardioplegic solution, 25 mmoljL, on coronary vascular responsivenessto 5-HT and PAP. Changes in coronary flow (expressed as a percent of drug-free control) were measured in response to 5-HT (I X 10.7 moljL and PAP (5 X 10-6 moljL) before (5-HTb and PAPb) and after (5-HTa and PAPa) 30 minutes' infusion of a 25 mmoljL concentration of potassium cardioplegic solution. Each point represents the mean of eight hearts and the bars represent the standard errors of the mean.
3-minute period and changes in coronary flow were measured. This was followed by a 5-minute washout period with buffer during which it was ensured that coronary flow returned to control values. Hearts were then subjected to 30 minutes of infusion with nonoxygenated, hyperkalemic (potassium concentration 25 or 40 mmol/L), hypothermic (20 C) cardioplegic solution at a perfusion pressure of 100 cm H 20 . The total infusion solution was collected and its volume was measured. Finally, hearts were perfused aerobically (at 60 cm H 20 ) for 15 minutes with Krebs-Henseleit buffer followed by repeat perfusion for 3 minutes with buffer containing either 5-HT or papaverine, and once again the resultant change in coronary flow was measured. Expression of results. Values used for determining changes in coronary flow were taken as the mean of the last 2 minutes of drug-free perfusion before .changing to drug-containing buffer (control) and the mean of the second and third minutes of perfusion with drug-containing buffer. Changes in coronary flow were expressed as a percentage of the control value. Eight hearts were used in each group, and all data were expressed as the mean ± standard error. Statistical analysis of the results was made by unpaired Student's t test, and statistical significance was assumed when the probability was less than 0.001. 0
Results Dose-response studies. To select appropriate doses for the main studies, we conducted preliminary studies to test the effect of 5-HT (1 X 10- 7 , 1 X 10-6, and 1 X 10- 5 moljL) and papaverine (1 X 10-6, 5 X 10-6, and 1 X 10- 5 moljL) on coronary flowover a 5-minute period of perfusion. The results (Fig. 1, A and B) showed that both 5-HT and papaverine exerted a dose-dependent vasodilatory effect. A 5-HT dose of 1 X 10- 7 moljL and a
-10
5-HTa
Fig. 3. The effect of 30 minutes' infusion of potassium cardioplegic solution, 40 mmoljL on coronary vascular responsivenessto 5-HT and PAP. Changes in coronary flow (expressed as a percent of drug-free control) were measured in response to 5-HT (I X 10-7 moljL) and PAP (5 X 10-6 moljL) before (5HTb and PAPb) and after (5-HTa and PAPa) 3D-minutes' infusion of a 40 mmolj L concentration of potassium cardioplegic solution. Each point represents the mean of eight hearts and the bars represent the standard errors of the mean.
papaverine dose of 5 X 10-6 moljL were selected for further studies, because they gave comparable, half maximal responses (+20.6% ± 1.8%and +21.4% ± 1.6%,respectively), without exerting any significant effects on heart rate or contractile function (as measured with a left ventricular balloon). In additional studies, hearts perfused with oxygenated Krebs-Henseleit bicarbonate buffer were challenged with these concentrations of 5-HT and papaverine at 3D-minute intervals. Responses to both 5HT and papaverine were maintained even after 2 hours of perfusion. Studies with cardioplegia. The results obtained for all four groups of hearts are shown in Figs. 2 and 3. Before the potassium cardioplegic solution (25 mmoljL) was administered, both 5-HT and papaverine caused a similar increase in flow. In the case of 5-HT, coronary flow increased from 7.0 ± 0.4 to 8.5 ± 0.4 mljmin (+21.2% ± 1.6%), whereas in the case of papaverine coronary flow increased from 7.0 ± 0.3 to 8.6 ± 0.4 mIj min (+22.8% ± 1.6%). After cardioplegia, the vasodilatory response to 5-HT was completely lost; coronary flow remained unchanged at 7.1 ± 0.5 from 7.1 ± 0.3 mlj min (-0.2% ± 1.2%). However, there was no significant change in the response to papaverine, which maintained a +23.7% ± 1.4% vasodilation (coronary flow increased from 6.7 ± 0.3 to 8.2 ± 0.4 mljmin). With a 40 mmolj L concentration of potassium cardioplegic solution, the initial responses to 5-HT and papaverine were also sim-
The Journal of Thoracic and Cardiovascular Surgery
786 Saldanha and Hearse
ilar. In the case of 5-HT coronary flow increased from 6.8 ± 0.3 to 8.3 ± 0.5 mljmin (+21.5% ± 2.5%), whereas in the case of pa paverine coronary flow increased from 7.3 ± 0.2 to 9.1 ± 0.3 mljmin (+24.1% ± 3.0%). After cardioplegia, 5-HT caused a significant (p < 0.001) vasoconstriction, as coronary flow fell from 7.2 ± 0.2 to 7.0 ± 0.2 mljmin (-4.3% ± 1.1%). The response to papaverine was again maintained as coronary flow increased from 7.0 ± 0.4 to 8.3 ± 0.4 mljmin (+19.1% ± 2.3%).
Discussion In the present studies we exposed hearts to a 30-minute period of cardioplegic infusion at 20° C. This duration and temperature were selected so as to approximate to the cumulative time that the heart may be exposed to cardioplegic infusion under clinical conditions with prolonged arrest and multidose cardioplegia. We did not expose the heart to ischemia, because this is another factor that might contribute to endothelial injury. Under these experimental conditions, we have demonstrated that use of a hyperkalemic crystalloid cardioplegic solution can result in a potentially detrimental change in vascular responsiveness. Thus before cardioplegia 5-HT provoked vasodilation, whereas after cardioplegia 5-HT induced vasoconstriction. Because 5-HT has been demonstrated to cause vasodilation by the release of EDRF,9, 12, 16 the change in response is most likely due to cardioplegia-induced damage to the endothelium. The observed effect cannot be attributed to damage to vascular smooth muscle or edema, because the response to papaverine was maintained. Our results suggested that the vasoconstriction was greater after perfusion with a potassium concentration of 40 mmoljL than 25 mmoljL. This observation is consistent with the fact that potassium is a recognized vascular irritant and can be used to denude vessels of endothelium. 17 Other studies have provided evidence that hyperkalemic crystalloid cardioplegic solutions damage vascular endothelium. 18-21 Rosenbaum and colleaguesl'' showed that platelet deposition occurred in the coronary microvasculature after multidose, hyperkalemic cardioplegic arrest. Because platelet deposition occurs at sites of endothelial damage, it can be concluded that either cardioplegia fails to protect the endothelium against prolonged ischemia or that the cardioplegic solution itself promotes damage to the endothelium. There is some evidence to support both of these possibilities. Using tissue culture methods, Carpentier, Murawsky, and Carpentier'? evaluated the cytotoxicity of twelve different cardioplegic solutions. They found that the percentage of dead endothelial cells after 3 hours of incubation with cardioplegic solutions at 10° C ranged from 4.2% to 13.1%, Moreover, they found that the percentage of dead
cells after 3-hours' incubation at 19° C, followed by incubation in culture media for 24 hours at 37° C, ranged from 4.2% to 64.7%. Other investigators have used morphologic means to study the effects of cardioplegic solutions on the endothelium (from a variety of vessels) and also concluded that damage occurs. Early examination of vessels by scanning and transmission electron microscopy showed loss of endothelium, abnormal endothelial cells, and marked platelet and fibrin deposition.V 21 Late examination of vessels showed an intact intima with normal surface production of prostacyclin.F However, there was intense adventitial fibrosis and atherosclerotic changes compatible with initial endothelial damage. 22-24 The theoretical implications of endothelial damage are manifold. In the early postoperative period, deposition of platelets in the microvasculature has been implicated as a cause of reperfusion injury." Endothelial damage may lead to thrombosis, early occlusion of grafts, and myocardial infarction. The response to circulating catecholamines may bealtered and pathologic vasospasm produced instead of vasodilation. I I All these factors will contribute to poor myocardial performance in the immediate postoperative period. Long-term effects of endothelial damage include atherosclerosis, which is a widely recognized complication of grafts. Clinically, the effects of endothelial damage are perhaps not so apparent because of the widespread use of antiplatelet drugs.F' In summary, although limited by its observation in the rat heart, evidence is presented that endothelial function may be impaired after 30 minutes' perfusion with hyperkalemic crystalloid cardioplegic solution, Our results support the concept that potassium is a possible mediator; thus it might be wise to advocate the lowest potassium content capable of achieving arrest under hypothermic conditions. Further experiments are needed to establish the optimal conditions for preserving the endothelium and to determine whether cardioplegic solutions or the manner in which they are used need to be modified to protect the endothelium. The assistance of Mrs, C. Erlebach is gratefully acknowledged. REFERENCES 1. Hearse DJ, Braimbridge MY, Jynge P, Protection of the ischemic myocardium, Cardioplegia, New York: Raven Press, 1981, 2. Hearse DJ, Cardioplegia: the protection of the myocardium during open heart surgery; a review, J Physiol Paris 1980; 76:751-68. 3, Bretschneider HJ, Myocardial protection. Thorac Cardiovasc Surg 1981;28:295-302. 4. Buckberg GO. Strategies and logic of cardioplegic delivery
Volume 98 Number 5, Part 1 November 1989
5.
6.
7. 8. 9.
10.
11. 12.
13.
14.
15.
to prevent, avoid, and reverse ischemic and reperfusion damage. J THORAC CARDIOVASC SURG 1987;93:127-39. Gerlach E, Nees S, Becker BF. The vascular endothelium: a survey of some newly evolving biochemical and physiological features. Basic Res Cardiol 1985;80:459-74. Sixma JJ. Role of blood vessel, platelet and coagulation interactions in haemostasis. In: Bloom AL, Thomas DP, eds. Haemostasis and thrombosis. Edinburgh: Churchill Livingstone, 1981:252-67. Born GVR. Platelets and blood vessels. J Cardiovasc Pharmacol I984;6:S707-13. Furchgott RF. Role of endothelium in responses of vascular smooth muscle. Circ Res 1983;53:557-73. Vanhoutte PM, Rimcle TJ. Role of endothelium in the control of vascular smooth muscle function. J Physiol (Paris) 1982-83;78:681-6. Holtz J, Forstermann U, Pohl U, Giesler M, Bassenge E, Flow-dependent, endothelium-mediated dilation of epicardial coronary arteries in conscious dogs: effects of cyclooxygenase inhibition. J Cardiovasc Pharmacol 1984;6: 1161-9. Vanhoutte PM, Houston DS. Platelets, endothelium and vasospasm. Circulation 1985;72:728-34. Hearse DJ, Saldanha CBR. Quinacrine inhibits vasodilatation caused by 5-HT but not papaverine in the rat isolated heart [Abstract). Br J Pharmacol 1986;87:12P. Stewart DJ, Pohl U, Bassenge E. Effect of free radicals on resistance vessels in isolated rabbit hearts: enhanced tone and loss of endothelium-dependent dilation [Abstract). Circulation 1987;76(Pt 2):JV55. Krebs HA, Henseleit K. Untersuchungen tiber die Harnstoffbilding im Tierkorper. Hoppe-Seyler Z Physiol Chern 1932;210:33-66. Langendorff O. Untersuchungen am uberlender Saugertierherzen. Pfluger Arch 1935;61:291-332.
Cardioplegia and vascular injury
787
16. Cohen RA, Shepherd JT, Vanhoutte PM. 5-Hydroxytryptamine can mediate endothelium-dependent relaxation of coronary arteries. Am J Physiol 1983;245:H I07780. 17. Griffiths TM, Edwards DH, Lewis MJ, et al. The nature of endothelium-derived relaxant factor. Nature 1984;308: 645-7. 18. Rosenbaum D, Levitsky S, Silverman N, et al. Cardioplegia does not prevent reperfusion injury induced by intracoronary platelet deposition. Circulation 1983;68 (Pt 2):11102-6. 19. Carpentier S, Murawsky M, Carpentier A. Cytotoxicity of cardioplegic solutions: evaluation by tissue culture. Circulation 1981;64:(Pt 2):1190-5. 20. Follette DM, Buckberg GD, Mulder DG, Fonkalsrud EW. Deleterious effects of crystalloid cardioplegic solutions on arterial endothelial cells. Surg Forum 1980;31:253-5. 21. Harjula A, Mattila S, Mattila J, et al. Coronary endothelial damage after crystalloid cardioplegia. J Cardiovasc Surg 1984;25:147-52. 22. Hoover EL, Pet! SB, Amiram E, et al. The effects of crystalloid potassium cardioplegic solutions on arterialized canine vein grafts: assessment of chronic prostacyclin production and histopathologic alterations. Circulation 1981;64(pt 2):1196-100. 23. Olinger GN, Boerboom LE, Bonchek LI, et al. Hyperkalemia in cardioplegic solutions causing increased cholesterol accumulation in vein grafts. J THORAC CARDIOVASC SURG 1983;85:590-4. 24. Singh AK, Capone RJ, O'Shea P, Karlson KE. Long-term changes in canine vein graft after infusion of cardioplegic solution. Circulation I983;68(pt 2):IIlI2-6. 25. Cheseboro JH, Clements JP, Fuster V, et al. A platelet inhibitor drug trial in coronary-artery bypass operations. N Engl J Med 1982;307:73-8.
CARDIOVASC SURG
1989;98:783-7
Coronary vascular responsiveness to 5-hydroxytryptamine before and after infusion of hyperkalemic crystalloid cardioplegic solution in the rat heart Possible evidence of endothelial damage Using the isolated, Langendorff-perfused rat heart (n = 8 per group), we have studied the effects of 5-hydroxytryptamine and papaverine on coronary flow before and after a 30-rninute infusion of hypothermic (200 C), nonoxygenated cardioplegic solution containing potassium in a concentration of either 25 or 40 mmoljL. Before infusion of the 25 mmoljL potassium cardioplegic solution, both 5-hydroxytryptamine (1 X 10-7 moljL) and papaverine (5 x 10-6 moljL) caused similar increases in Bow (+21.2% ± 1.6% and +22.8% ± 1.6%, respectively). After cardioplegia, the vasodilatory response to 5-hydroxytryptarnine was completely lost and a slight vasoconstriction was observed (-0.2 % ± 1.2 %). However, there was no significant change in the response to papaverine, which maintained a +23.7% ± 1.4% vasodilation. With the cardioplegic solution containing a 40 mmoljL concentration of potassium, the initial responses to 5-hydroxytryptarnine and papaverine were again similar (+21.5% ± 2.5% and +24.1 % ± 3.0%, respectively). After cardioplegia, 5-hydroxytryptamine caused a significant vasoconstriction (-4.3% ± 1.1 %), whereas the response to papaverine was again maintained (+19.1 % ± 2.3%). The results of this study support the concept that hyperkalemic crystalloid cardioplegic solutions cause vascular damage possibly involving the endothelium or its f\DICtion, which may adversely affect vascular responsiveness.
Clyde Saldanha, BSc, MB, BS, and David J. Hearse, BSc, PhD, DSc, FACC, London. England
Over the past decade the use of cardioplegic solutions has gained wide acceptance as a method for intraoperativemanagement of the heart during cardiac operations. Their effectiveness in protecting the myocyte against prolonged ischemia has been widely established by extensive laboratory and clinical studies.' The rationale and aims in the use of cardioplegic solutions have been
FromCardiovascular Research, Rayne Institute, St. Thomas' Hospital, London SEI 7EH, England. Thiswork was supported in part by grants from STRUTH and the National Heart Research Fund. Received for publication Oct. 6, 1988. Accepted for publication Jan. 31, 1989. Address for reprints: Professor D. J. Hearse, Cardiovascular Research, Rayne Institute, St. Thomas' Hospital, London SEI 7EH, England. 12/1/11814
reviewed by several authors.i" Significantly, these reviews have mainly concentrated on the effects of cardioplegic solutions on the myocyte and hence on the mechanical performance of the heart. Very little attention has been focused on the effects of these solutions on the conduction or the vascular tissue. However, if the heart is to be protected optimally and good function achieved postoperatively, it is imperative that all tissues be protected. With regard to the vascular tissue, the endothelial lining is of importance because of the fundamental role it plays in hemostasis. Intact endothelium produces prostacyclin, antithrombin III, and plasminogen activator. 5,6 By such mechanisms, platelet aggregation and fibrin deposition on the vessel wall are prevented. Conversely, injury to the endothelium exposes collagen fibers and basement membrane, which are potent stimuli for platelet aggregation and adhesion." Disrupted endothelial cells may also release tissue procoagulant, which further facil783
The Journal of
784
Thoracic and Cardiovascular
Saldanha and Hearse
Surgery
Table I. Composition of cardioplegic solutions ~
o
Potassium
....
-'
25 mmol/L
>a:
...:
Cl
Sodium chloride (mmol) Potassium chloride (mmol) Magnesium chloride (mmol) Calcium chloride (mmol) Sodium bicarbonate (mmol)
...:
pH
z o a: o
u
~
w
z
:I: U
110.0
25.0 16.0 1.2 10.0 7.8
40 mmol/l. 110.0 40.0 16.0 1.2 10.0 7.8 .
.... ~ o a:
w
"'"
10
~
o
....
-'
>a:
...:
z
o
a:
o u
~
w
Cl
z
...:
:t: U
.... Z
w
u
a:
that damage to the endothelium may result in pathologic responses to normal stimuli, for example, vasospasm in response to 5-HT released from aggregating platelets. I I In a previous study, we I2 have shown that 5-HTinduced vasodilation in the isolated rat heart can be abolished by quinacrine, which is a recognized EDRF antagonist. This study and that ofothers 13 provide evidence that EDRF is released in the perfused isolated heart in response to 5-HT.ln view of the important role of the endothelium and the potentially serious consequences of damage to it, we were interested in studying the effect of cardioplegic solutions per se (in the absence of the injurious effects of ischemia) on responses thought to be mediated by the endothelium.
w
"'"
10
Fig. 1. Dose-response characteristics for 5-HT and papaverine (PAP). Hearts were subjected to perfusion with various concentrations of 5-HT (A)and PAP (B). Changes in coronary flow were measured and expressed as a percent of drug-free control hearts. Each result represents the mean of eight hearts and the bars represent the standard errors of the mean.
itates the cascade to thrombus formation," Injury to the endothelium is also a key step in the initiation of atherosclerosis.' Platelets stick to exposed areas and release factors that stimulate the proliferation of intimal smooth muscle. When the endothelium eventually regenerates to cover areas previously denuded, the intima underlying regenerated endothelium is substantially thicker and more likely to accumulate lipid than is the intima under undamaged endothelium.t-" Besides its role in hemostasis, the endothelium may influence vascular tone by the release of an endotheliumdependent relaxant factor (EDRF).8,9 Although EDRF has been shown to be released in vessels in vivo'? and in response to chemicals such as 5-hydroxytryptamine (5HT) and norepinephrine," its physiologic role remains to be clearly established. However, it has been postulated
Methods Animals and reagents. Hearts were obtained from male Sprague-Dawley rats (200 to 250 gm body weight) that had been maintained on a standard laboratory diet. 5-HT hydrochloride hemihydrate (serotonin) was obtained from Aldrich Chemical Company Ltd., Gillingham, Dorset, England, and papaverine hydrochloride was obtained from Sigma Chemical Company, St. Louis. The cardioplegic solutions were prepared by adding potassium chloride to the St. Thomas' Hospital cardioplegicsolution (Plegisol,Abbott Laboratories, North Chicago) to obtain final potassium concentrations of 25 or 40 rnrnol/ L. The final composition of both solutions is shown in Table I. Perfusion technique. Rats were anesthetized with diethyl ether, the right femoral vein was dissected out, and 200 IU of heparin was injected into the vein. After 30 seconds,the thorax was opened and the heart was excised and placed into a beaker containing cold (4° C) Krebs-Henseleit bicarbonate buffer.14A modifiedLangendorff preparation 15 was used;this wasdesigned to allow accurate measurement of coronary flow rates under conditions of constant perfusion pressure (60 em H 20 ). Hearts were perfused via the aorta with ultrafiltered (5.0 /lm porosity filter) buffer containing glucose (11.1 mmol/L) and gassedwith 95% oxygen plus 5% carbon dioxide (pH 7.4 at 37° C). Drugs were dissolved in the buffer to give optimally active concentrations (derived from dose-response studies) of 5-HT (I X 10- 7 moljL) and papaverine (5 X 10-6 moljL). Experimental time course. After 15 minutes of control perfusionwith oxygenated Krebs-Henseleit bicarbonate buffer (60 em H 20 , 37° C during which coronary flow was continually measured with a graduated cylinder), hearts were perfused aerobically with buffer containing either 5-HT or papaverine for a
Volume 98 Number 5. Part 1
Cardioplegia and vascular injury
November 1989
:: 0
30
....I
>-
z
PAPb
S-HTb
u,
c:: ocr
25 mM POTASSIUM CARDIOPLEGIA
.,..
PAPa
10
c: 0 o
z
ocr
0
W
o z
S-HTa
J:
C,,)
•~
10
Z
W
o
20
0
Z
Z
PAPb
S-HTb
U.
> c:: ocr
c:: C,,)
30
....I
20
0
0
:: 0
40 mM POTASSIUM CARDIOPLEGIA
785
0
ocr
.o J:
-10
~
Fig. 2. The effect of 30 minutes' infusion of potassium cardioplegic solution, 25 mmoljL, on coronary vascular responsivenessto 5-HT and PAP. Changes in coronary flow (expressed as a percent of drug-free control) were measured in response to 5-HT (I X 10.7 moljL and PAP (5 X 10-6 moljL) before (5-HTb and PAPb) and after (5-HTa and PAPa) 30 minutes' infusion of a 25 mmoljL concentration of potassium cardioplegic solution. Each point represents the mean of eight hearts and the bars represent the standard errors of the mean.
3-minute period and changes in coronary flow were measured. This was followed by a 5-minute washout period with buffer during which it was ensured that coronary flow returned to control values. Hearts were then subjected to 30 minutes of infusion with nonoxygenated, hyperkalemic (potassium concentration 25 or 40 mmol/L), hypothermic (20 C) cardioplegic solution at a perfusion pressure of 100 cm H 20 . The total infusion solution was collected and its volume was measured. Finally, hearts were perfused aerobically (at 60 cm H 20 ) for 15 minutes with Krebs-Henseleit buffer followed by repeat perfusion for 3 minutes with buffer containing either 5-HT or papaverine, and once again the resultant change in coronary flow was measured. Expression of results. Values used for determining changes in coronary flow were taken as the mean of the last 2 minutes of drug-free perfusion before .changing to drug-containing buffer (control) and the mean of the second and third minutes of perfusion with drug-containing buffer. Changes in coronary flow were expressed as a percentage of the control value. Eight hearts were used in each group, and all data were expressed as the mean ± standard error. Statistical analysis of the results was made by unpaired Student's t test, and statistical significance was assumed when the probability was less than 0.001. 0
Results Dose-response studies. To select appropriate doses for the main studies, we conducted preliminary studies to test the effect of 5-HT (1 X 10- 7 , 1 X 10-6, and 1 X 10- 5 moljL) and papaverine (1 X 10-6, 5 X 10-6, and 1 X 10- 5 moljL) on coronary flowover a 5-minute period of perfusion. The results (Fig. 1, A and B) showed that both 5-HT and papaverine exerted a dose-dependent vasodilatory effect. A 5-HT dose of 1 X 10- 7 moljL and a
-10
5-HTa
Fig. 3. The effect of 30 minutes' infusion of potassium cardioplegic solution, 40 mmoljL on coronary vascular responsivenessto 5-HT and PAP. Changes in coronary flow (expressed as a percent of drug-free control) were measured in response to 5-HT (I X 10-7 moljL) and PAP (5 X 10-6 moljL) before (5HTb and PAPb) and after (5-HTa and PAPa) 3D-minutes' infusion of a 40 mmolj L concentration of potassium cardioplegic solution. Each point represents the mean of eight hearts and the bars represent the standard errors of the mean.
papaverine dose of 5 X 10-6 moljL were selected for further studies, because they gave comparable, half maximal responses (+20.6% ± 1.8%and +21.4% ± 1.6%,respectively), without exerting any significant effects on heart rate or contractile function (as measured with a left ventricular balloon). In additional studies, hearts perfused with oxygenated Krebs-Henseleit bicarbonate buffer were challenged with these concentrations of 5-HT and papaverine at 3D-minute intervals. Responses to both 5HT and papaverine were maintained even after 2 hours of perfusion. Studies with cardioplegia. The results obtained for all four groups of hearts are shown in Figs. 2 and 3. Before the potassium cardioplegic solution (25 mmoljL) was administered, both 5-HT and papaverine caused a similar increase in flow. In the case of 5-HT, coronary flow increased from 7.0 ± 0.4 to 8.5 ± 0.4 mljmin (+21.2% ± 1.6%), whereas in the case of papaverine coronary flow increased from 7.0 ± 0.3 to 8.6 ± 0.4 mIj min (+22.8% ± 1.6%). After cardioplegia, the vasodilatory response to 5-HT was completely lost; coronary flow remained unchanged at 7.1 ± 0.5 from 7.1 ± 0.3 mlj min (-0.2% ± 1.2%). However, there was no significant change in the response to papaverine, which maintained a +23.7% ± 1.4% vasodilation (coronary flow increased from 6.7 ± 0.3 to 8.2 ± 0.4 mljmin). With a 40 mmolj L concentration of potassium cardioplegic solution, the initial responses to 5-HT and papaverine were also sim-
The Journal of Thoracic and Cardiovascular Surgery
786 Saldanha and Hearse
ilar. In the case of 5-HT coronary flow increased from 6.8 ± 0.3 to 8.3 ± 0.5 mljmin (+21.5% ± 2.5%), whereas in the case of pa paverine coronary flow increased from 7.3 ± 0.2 to 9.1 ± 0.3 mljmin (+24.1% ± 3.0%). After cardioplegia, 5-HT caused a significant (p < 0.001) vasoconstriction, as coronary flow fell from 7.2 ± 0.2 to 7.0 ± 0.2 mljmin (-4.3% ± 1.1%). The response to papaverine was again maintained as coronary flow increased from 7.0 ± 0.4 to 8.3 ± 0.4 mljmin (+19.1% ± 2.3%).
Discussion In the present studies we exposed hearts to a 30-minute period of cardioplegic infusion at 20° C. This duration and temperature were selected so as to approximate to the cumulative time that the heart may be exposed to cardioplegic infusion under clinical conditions with prolonged arrest and multidose cardioplegia. We did not expose the heart to ischemia, because this is another factor that might contribute to endothelial injury. Under these experimental conditions, we have demonstrated that use of a hyperkalemic crystalloid cardioplegic solution can result in a potentially detrimental change in vascular responsiveness. Thus before cardioplegia 5-HT provoked vasodilation, whereas after cardioplegia 5-HT induced vasoconstriction. Because 5-HT has been demonstrated to cause vasodilation by the release of EDRF,9, 12, 16 the change in response is most likely due to cardioplegia-induced damage to the endothelium. The observed effect cannot be attributed to damage to vascular smooth muscle or edema, because the response to papaverine was maintained. Our results suggested that the vasoconstriction was greater after perfusion with a potassium concentration of 40 mmoljL than 25 mmoljL. This observation is consistent with the fact that potassium is a recognized vascular irritant and can be used to denude vessels of endothelium. 17 Other studies have provided evidence that hyperkalemic crystalloid cardioplegic solutions damage vascular endothelium. 18-21 Rosenbaum and colleaguesl'' showed that platelet deposition occurred in the coronary microvasculature after multidose, hyperkalemic cardioplegic arrest. Because platelet deposition occurs at sites of endothelial damage, it can be concluded that either cardioplegia fails to protect the endothelium against prolonged ischemia or that the cardioplegic solution itself promotes damage to the endothelium. There is some evidence to support both of these possibilities. Using tissue culture methods, Carpentier, Murawsky, and Carpentier'? evaluated the cytotoxicity of twelve different cardioplegic solutions. They found that the percentage of dead endothelial cells after 3 hours of incubation with cardioplegic solutions at 10° C ranged from 4.2% to 13.1%, Moreover, they found that the percentage of dead
cells after 3-hours' incubation at 19° C, followed by incubation in culture media for 24 hours at 37° C, ranged from 4.2% to 64.7%. Other investigators have used morphologic means to study the effects of cardioplegic solutions on the endothelium (from a variety of vessels) and also concluded that damage occurs. Early examination of vessels by scanning and transmission electron microscopy showed loss of endothelium, abnormal endothelial cells, and marked platelet and fibrin deposition.V 21 Late examination of vessels showed an intact intima with normal surface production of prostacyclin.F However, there was intense adventitial fibrosis and atherosclerotic changes compatible with initial endothelial damage. 22-24 The theoretical implications of endothelial damage are manifold. In the early postoperative period, deposition of platelets in the microvasculature has been implicated as a cause of reperfusion injury." Endothelial damage may lead to thrombosis, early occlusion of grafts, and myocardial infarction. The response to circulating catecholamines may bealtered and pathologic vasospasm produced instead of vasodilation. I I All these factors will contribute to poor myocardial performance in the immediate postoperative period. Long-term effects of endothelial damage include atherosclerosis, which is a widely recognized complication of grafts. Clinically, the effects of endothelial damage are perhaps not so apparent because of the widespread use of antiplatelet drugs.F' In summary, although limited by its observation in the rat heart, evidence is presented that endothelial function may be impaired after 30 minutes' perfusion with hyperkalemic crystalloid cardioplegic solution, Our results support the concept that potassium is a possible mediator; thus it might be wise to advocate the lowest potassium content capable of achieving arrest under hypothermic conditions. Further experiments are needed to establish the optimal conditions for preserving the endothelium and to determine whether cardioplegic solutions or the manner in which they are used need to be modified to protect the endothelium. The assistance of Mrs, C. Erlebach is gratefully acknowledged. REFERENCES 1. Hearse DJ, Braimbridge MY, Jynge P, Protection of the ischemic myocardium, Cardioplegia, New York: Raven Press, 1981, 2. Hearse DJ, Cardioplegia: the protection of the myocardium during open heart surgery; a review, J Physiol Paris 1980; 76:751-68. 3, Bretschneider HJ, Myocardial protection. Thorac Cardiovasc Surg 1981;28:295-302. 4. Buckberg GO. Strategies and logic of cardioplegic delivery
Volume 98 Number 5, Part 1 November 1989
5.
6.
7. 8. 9.
10.
11. 12.
13.
14.
15.
to prevent, avoid, and reverse ischemic and reperfusion damage. J THORAC CARDIOVASC SURG 1987;93:127-39. Gerlach E, Nees S, Becker BF. The vascular endothelium: a survey of some newly evolving biochemical and physiological features. Basic Res Cardiol 1985;80:459-74. Sixma JJ. Role of blood vessel, platelet and coagulation interactions in haemostasis. In: Bloom AL, Thomas DP, eds. Haemostasis and thrombosis. Edinburgh: Churchill Livingstone, 1981:252-67. Born GVR. Platelets and blood vessels. J Cardiovasc Pharmacol I984;6:S707-13. Furchgott RF. Role of endothelium in responses of vascular smooth muscle. Circ Res 1983;53:557-73. Vanhoutte PM, Rimcle TJ. Role of endothelium in the control of vascular smooth muscle function. J Physiol (Paris) 1982-83;78:681-6. Holtz J, Forstermann U, Pohl U, Giesler M, Bassenge E, Flow-dependent, endothelium-mediated dilation of epicardial coronary arteries in conscious dogs: effects of cyclooxygenase inhibition. J Cardiovasc Pharmacol 1984;6: 1161-9. Vanhoutte PM, Houston DS. Platelets, endothelium and vasospasm. Circulation 1985;72:728-34. Hearse DJ, Saldanha CBR. Quinacrine inhibits vasodilatation caused by 5-HT but not papaverine in the rat isolated heart [Abstract). Br J Pharmacol 1986;87:12P. Stewart DJ, Pohl U, Bassenge E. Effect of free radicals on resistance vessels in isolated rabbit hearts: enhanced tone and loss of endothelium-dependent dilation [Abstract). Circulation 1987;76(Pt 2):JV55. Krebs HA, Henseleit K. Untersuchungen tiber die Harnstoffbilding im Tierkorper. Hoppe-Seyler Z Physiol Chern 1932;210:33-66. Langendorff O. Untersuchungen am uberlender Saugertierherzen. Pfluger Arch 1935;61:291-332.
Cardioplegia and vascular injury
787
16. Cohen RA, Shepherd JT, Vanhoutte PM. 5-Hydroxytryptamine can mediate endothelium-dependent relaxation of coronary arteries. Am J Physiol 1983;245:H I07780. 17. Griffiths TM, Edwards DH, Lewis MJ, et al. The nature of endothelium-derived relaxant factor. Nature 1984;308: 645-7. 18. Rosenbaum D, Levitsky S, Silverman N, et al. Cardioplegia does not prevent reperfusion injury induced by intracoronary platelet deposition. Circulation 1983;68 (Pt 2):11102-6. 19. Carpentier S, Murawsky M, Carpentier A. Cytotoxicity of cardioplegic solutions: evaluation by tissue culture. Circulation 1981;64:(Pt 2):1190-5. 20. Follette DM, Buckberg GD, Mulder DG, Fonkalsrud EW. Deleterious effects of crystalloid cardioplegic solutions on arterial endothelial cells. Surg Forum 1980;31:253-5. 21. Harjula A, Mattila S, Mattila J, et al. Coronary endothelial damage after crystalloid cardioplegia. J Cardiovasc Surg 1984;25:147-52. 22. Hoover EL, Pet! SB, Amiram E, et al. The effects of crystalloid potassium cardioplegic solutions on arterialized canine vein grafts: assessment of chronic prostacyclin production and histopathologic alterations. Circulation 1981;64(pt 2):1196-100. 23. Olinger GN, Boerboom LE, Bonchek LI, et al. Hyperkalemia in cardioplegic solutions causing increased cholesterol accumulation in vein grafts. J THORAC CARDIOVASC SURG 1983;85:590-4. 24. Singh AK, Capone RJ, O'Shea P, Karlson KE. Long-term changes in canine vein graft after infusion of cardioplegic solution. Circulation I983;68(pt 2):IIlI2-6. 25. Cheseboro JH, Clements JP, Fuster V, et al. A platelet inhibitor drug trial in coronary-artery bypass operations. N Engl J Med 1982;307:73-8.