Cardiac-derived thromboxane A2

Cardiac-derived thromboxane A 2 An initiating mediator of reperfusion injury? After crystalloid cardioplegic arrest, cardiac-derived thromboxane A2 may be an important initiating mediator of no-rejiow and hemodynamic deterioration during reperfusion because of its potent vasoactive properties. Although previous studies have already documented the increased release of cardiac thromboxane A2 after ischemia, none have studied the effects of cardiac thromboxane A2 on hemodynamics. We therefore tested the ability of cardiac thromboxane A2 to mediate deterioration of coronary flow and functional recovery during reperfusion after global ischemia. Crystalloid-perfused rat hearts that had undergone Langendorff preparation (n = 30) were subjected to 2 hours of global ischemia at 15° C under cardioplegic protection with (n = 15) or without (n = 15) thromboxane A2 receptor antagonist SQ29548. In eight of 15 hearts in each group, preischemic and postischemic aortic flow, coronary flow, cardiac output, heart rate, and stroke work were determined. In the remaining seven hearts in each group, preischemic and postischemic coronary effluent levels of the stable hydrolysis product of thromboxane A2 and thromboxane B2 were determined with radioimmunoassay through the use of nonrecirculating perfusate. At the completion of the experiment, water content was determined by wet weight/dry weight calculations. In a separate group (n = 7) preischemic myocardial water content was determined. Within the group protected by cardioplegic solution alone, postischemic aortic flow, coronary flow, cardiac output, and stroke work were all significantly decreased (p < 0.05) compared with preischemic values (aortic flow, 50.8 ± 2.7 versus 29.4 ± 3.3 ml/rnin; coronary flow, 13.2 ± 1.3 versus 8.5 ± 1.3 ml/min; cardiac output, 64.0 ± 3.8 versus 38.0 ± 4.4 ml/min; stroke work, 12.5 ± 0.7 versus 7.1 ± 0.8 em H20· ml), In relation to the group with cardioplegic solution alone, postischemic aortic flow, coronary flow, cardiac output, and stroke work were all significantly greater (p < 0.05) in the group with the receptor antagonist (aortic flow: 49.5 ± 2.4 versus 29.4 ± 3.3 ml/min; coronary flow; 12.4 ± 1.2 versus 8.5 ± 1.3 ml/min; cardiac output, 62.0 ± 2.8 versus 38.0 ± 4.4 ml/min; stroke work, 12.6 ± 0.8 versus 7.1 ± 0.8 em H20. ml), Overall, postischemic coronary effluent thromboxane B2 levels were greater than preischemic values (105.6 ± 12.4 versus 69.6 ± 9.8, p < 0.05) and treatment with the receptor antagonist did not significantly affect postischemic thromboxane B2 levels (92.0 ± 7.3 versus 82.3 ± 15.5, p = not significant). Neither ischemia nor treatment with the receptor antagonist significantly affected heart rate. In both groups, ischemia and reperfusion resulted in a small but significant increase in myocardial edema (hearts with cardioplegic solution alone, 82.6 % ± 0.8 % versus 75.8 % ± 0.6 %, p < 0.05; hearts with receptor antagonist, 80.1 % ± 0.9 % versus 75.8 % ± 0.6 %, p < 0.05), which was unaffected by treatment with the receptor antagonist (82.6% ± 0.8% versus 80.1 % ± 0.9%, p = not significant). These data suggest that cardiac-derived thromboxane A2 mediates deterioration of coronary flow and hemodynamic performance during reperfusion. Also, the cardioprotective effects of thromboxane A2 blockade, without

John G. Byrne, MD, Robert F. Appleyard, PhD, Shu-Ching Sun, MD, Gregory S. Couper, MD, Jacob A. Sloane, BA, Rita G. Laurence, BS, and Lawrence H. Cohn, MD, Boston, Mass.

From the Divisionof Cardiac Surgery, Brigham & Women's Hospital and Harvard Medical School, Boston, Mass. Presented in part at the Surgical Forum, Seventy-sixth Clinical Congress of the American College of Surgeons, San Francisco, Calif., October 9, 1990. Supported by Grant HL33463 from the National Heart, Lung, and Blood Institute.

Received for publication Dec. 23, 1991. Accepted for publication June 25, 1992. Address for reprints: Lawrence H. Cohn, MD, Brigham & Women's Hospital, 75 Francis St., Boston, MA 02115. J THORAC Ci\RDIOVi\SC SURG 1993;105:689-93 Copyright

1993 by Mosby-Year Book, Inc.

0022-5223/93 $1.00+ .10

12/1/40523

689

690

Byrne et al.

The Journal of Thoracic and Cardiovascular Surgery April 1993

significant effects on postischemic myocardial edema, suggest that thromboxane A2-mediated injury may be a greater impairment to myocardial performance than myocardial edema. This study is the first to document the hemodynamic effects of cardiac-derived thromboxane A2 and provides evidence that it may be among mediators of reperfusion injury. (J THORAC CARDIOVASC SURG 1993;105:689-93)

Arter crystalloid cardioplegic arrest, cardiac-derived mediators may be among the agents that initiate ischemia-reperfusion injury. Recent studies have addressed chemotactic factors, 1 resident cardiac mast cell products.? and eicosanoids.v" Cardiac-derived thromboxane A 2 (TxA 2) is of particular interest because of its potential role in immediate local vasoconstriction and platelet aggregation, which promotes the no-ref/ow phenomenon and the progression of ischemia. Although previous studies have already documented increased release of cardiac TxA 2 during (I) ischemic arrest with cardioplegia in dogs." 6 (2) early reperfusion after cardioplegia and crossclamp release in patients undergoing coronary artery bypass grafting," and (3) prolonged cold storage for transplantation.t none has yet determined the effects of cardiac TxA2 on coronary flow and hemodynamic recovery during reperfusion. The present study, therefore, was conducted to determine the importance of cardiac-derived TxA2 in mediating deterioration of coronary flow and hemodynamic recovery after prolonged global ischemia and reperfusion.

Materials 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" prepared by the National Academy of Sciences and published by the National Institute of Health (NIH Publication No. 86-23, revised 1985). Perfusion system. Thirty male Sprague-Dawley rats, weighing 250 to 300 gm, were anesthetized with inhalation ether and heparinized (0.4 U /gm). Hearts were rapidly excised through bilateral thoracotomies and placed in 4 0 C normal saline solution. The hearts were then immediately perfused retrogradely by the Langendorff method at 70 ern H 20 with 37 0 C modified Krebs-Henseleit (NaC!, 115 mmol/L; KCl, 4.7 mmol/L; CaCh, 2.5 mmol/L; MgS0 4, 1.2 mrnol/L; KH 2P04 , 1.2 mrnol/L; NaHC0 3, 25 mmol/L; and glucose, 10 mmol/L) equilibrated with 95% oxygen and 5% carbon dioxide. Nonrecirculating retrograde perfusion was maintained for 10 minutes, which allowing for stabilization. The left atrium was then cannulated through the open pulmonary veins to establish antegrade recirculation. Left atrial preload was set at 15 ern H 20. The hearts were then converted to the working mode for an additional 20 minutes. Afterload was maintained at 70 em H 20. Protocol. After 20 minutes of antegrade working perfusion, baseline hemodynamic measurements were made. Hearts were then assigned to one of two groups. Control hearts (n = 15, CP) were arrested with 5 ml4 0 C potassium crystalloid cardioplegic solution (28 mEq KCl and 5 mEq NaHC03 per liter in 2.5% dextrose and 0.45 NaC!, osmolarity 330 mOsm/L, pH 7.8)

delivered via the aortic cannula at 40 to 50 mm Hg. Experimental hearts (n = 15, SQ) were arrested with the same cardioplegic solution plus the thromboxane A 2 receptor antagonist SQ29548 (final concentration in cardioplegic solution 30 ttmol/L or 11.6 ttgjml, F. W. = 387). This specific TxA2 receptor antagonist (kindly supplied by Squibb Pharmaceuticals, Princeton, N.J.) was chosen because of its highly selective inhibition of TxA2 receptors in rat coronary endothelium." The pharmacologic actions of the oxobicycloheptane SQ29548, [IS-[a,2fJ(5Z),3{3,4a]]-7-[3[[2-(phenylamino) carbonyl] hydrazino]-7-oxabicyclo [2.2.1]hept- 2-yl]-5-heptenoic acid, have been described in detail elsewhere. to In both groups cardioplegic solution was readministered every 30 minutes while the arrested hearts were maintained at 150 C in a temperaturecontrolled saline bath for 2 hours. All hearts were then reperfused at 37 0 C in an identical manner, as described earlier. After 20 minutes of working reperfusion, postischemic hemodynamic measurements were made. Because the dissociation constant of SQ29548 for rat coronary endothelium TxA 2 receptors is unknown, we cannot be certain that the SQ29548 administered at the aortic root in the cardioplegic solution remained bound to the coronary endothelial TxA 2 receptors during reperfusion. However, since the first dose of cardioplegic solution was allowed to recirculate in the 500 to 700 ml priming volume of the Langendorff apparatus, the final concentration ofSQ29548 in the Krebs perfusate during reperfusion in the SQ group was approximately 100 ng/ml (approximate range, 83 to 116 ng/rnl). Stahl? has demonstrated that SQ29548, in this dose range, ameliorates TxA 2 agonist U-46619-induced vasoconstriction in Krebs-Ringer solution-perfused rat coronary endothelium. In the two subgroups (n = 7 each), with the use of nonrecirculating perfusate, preischemic and postischemic coronary effluent samples were collected in chilled tubes, stored at -20 0 C, and analyzed with radioimmunoassay (RIA) for thromboxane B2 (TxB2), the stable hydrolysis product of TxA2. A nonrecirculating perfusate was used in these groups because a recirculating perfusate would have introduced a cumulative effect on TxB1levels, potentially leading to a false conclusion in the event that postischemic TXBl levels were different from preischemic levels. Our objective was to determine whether cardiac TxB 1 release at the time of postischemic hemodynamic measurements was different from that before ischemia. For a similar reason, we did not measure TxBl release immediately on reperfusion because of the potential for "washout" of TxB2, produced during ischemia, and the possible false conclusion that postischemic TxB 2leveis were different from preischemic levels. Because functional measurements in our model require recirculating perfusate, functional measurements were not possible in these groups. Although these groups therefore differ from those in which hemodynamic measurements were made, they were similar to the first two groups in that they were subjected to an identical injury of 2 hours of global ischemia at 15 0 C under the same cardioplegic protection and reperfused at 37 0 C for 30 minutes. Another group of seven hearts was excised and mounted on the perfusion system in an identical manner but,

The Journal of Thoracic and Cardiovascular Surgery Volume 105, Number 4

Byrne et al.

69 1

Table I. Hemodynamic parameters AF(ml/min) Pre CP (n SQ (n

= =

Post

CF (mlfmin}

Pre

Post

SW (em H20· ml)

CO imlfmin)

Pre

Post

Pre

Post

Pre

Post

8) 50.8 ± 2.7 29.4 ± 3.3* 13.3 ± 1.3 8.5 ± 1.3* 64.0 ± 3.8 38.0 ± 4.4* 283 ± 15 268 ± 12 12.5 ± 0.7 7.1 ± 0.8* 8) 55.5 ± 2.5 49.5 ± 2.4t 13.3 ± 0.8 12.4 ± 1.2t 68.8 ± 2.5 62.0 ± 2.8t 264 ± 11 277 ± 10 14.5 ± 0.9 12.6 ± 0.8t

Values (mean ± standard error of the mean): AF, Aortic flow; CF. coronary flow; CO. cardiac output; HR, heart rate; SW, Stroke work; CP, unmodified crystalloid cardioplegic solution; SQ, CP plus thromboxane receptor antagonist SQ29548; Pre, preischemic; Post, postischemic. *p < 0.05 versus preischemic. v p < 0.05 versus CPo

after 20 minutes of nonworking retrograde perfusion, the hearts were removed and analyzed for water content to determine the preischemic water values. Measurements Hemodynamics. Total coronary flow and aortic flow were determined by timed collections of the drainage from the pulmonary artery and aortic outflow tracts, respectively. A Statham PI OEZ pressure transducer (Viggo-Spectramed, Inc., Oxnard, Calif.) was connected to the aortic outflow tract from which heart rate (HR) was determined. Cardiac output (CO) was calculated as the sum of coronary flowand aortic flow, and stroke work (SW) was calculated as follows: SW

(Afterload - preload) . CO HR

55cmH 20 · CO HR

RiA. Cardiac TxB2 was measured by RIA of the coronary effluent before ischemia and after 30 minutes of reperfusion. The assays were performed in duplicate for each sample. Specificrabbit TxB2antisera were stored in 0.5 ml volumesat -80 0 C until use and then thawed. The antiserum was then diluted 1:10,000 with TBS (tris-buffered saline solution with gelatin) (0.1%gelatin, 9% NaCI, in tris HCI 0.1 rnol/L, pH 7.3). Coronary effluent aliquot (0.2 ml), diluted TxB 2 antisera (0.1 ml), and [3H]TxB2tracer (Amersham, Arlington Heights, IlI.) (0.1 ml) were incubated for I hour at 370 C. The TxB2 tracer had an activity of 6000 to 7000 counts per minute per sample. A control assay run in TBS (0.2 m!) was substituted for coronary effluent.A blank assay was also used in which TBS (0.2 ml) and rabbit plasma (0.1 ml, mixed 9:I with 0.1 mol ethylenediamine tetraacetic acid) were incubated with the TxB2 tracer (0.1 mI). After incubation for I hour, the reaction was stopped by the addition of rabbit plasma (0.1 ml). Goat antirabbit serum (0.1 ml) was added and the precipitation was allowed to occur overnight at 4 0 C. Then ethylenediamine tetracetic acid 0.1 ml (0.5 ml) was added. The precipitate was collected by centrifugation at 1200g for 30 minutes at 4 0 C to isolate it from the supernatant, dissolved in 0.2 ml NaOH 0.1 mol, and mixed with 2 ml scintillationfluid.Samples were counted on a Beckman LS-230 scintillationcounter (Beckman Instruments, Inc., Brea, Calif.) to obtain raw counts per minute. The duplicated raw valueswere averaged and converted with a standard curve to obtain absolute TxB2 concentrations. Myocardial water content. Wet and dry weights of hearts usedin hemodynamic studies were measured, and water content was calculated as follows:

Percentage of H 20

(Wet weight - Dry weight) . X 100 Wet weight

Statistical analysis. Hemodynamic and RIA data differences between preischemic and postischemic values and values between the two groups were tested by two-way analysis of

variance with repeated measures over the preischemic and postischemic measurements; the measurements of myocardial water content were contrasted by one-way analysis of variance. Differences in mean values measured before and after ischemia and between the two groups were considered significant when p was less than 0.05. All values are expressed as mean ± standard error of the mean.

Results There were no significant differences between groups in preischemic values for any measurement (Tables I and II). Within the CP group, postischemic aortic flow, coronary flow, cardiac output, and stroke work were all significantly decreased compared with preischemic values. However, compared with the CP group, postischemic aortic flow, coronary flow, cardiac output, and stroke work were all significantly greater in the SQ group. Heart rate was not significantly affected by ischemia or treatment with SQ29548 (Table I). Overall, postischemic coronary effluent TxB 2 was significantly greater than preischemic values. Further analysis determined that treatment with SQ29548 did not significantly affect postischemic TxB 2 values (Table II). In both groups ischemia-reperfusion resulted in a small but significant increase in myocardial edema, which was not significantly affected by treatment with SQ29548 (Table III).

Discussion This study is the first to document the effects of cardiac-derived TxA2 on coronary flow and hemodynamic recovery after prolonged global ischemia and reperfusion and provides evidence that cardiac-derived TxA2 may be among the initial mediators of reperfusion injury. In this study we measured a 50% increase in TxA2 release after 30 minutes of reperfusion in reperfused hearts after 2 hours of global ischemia at 15° C, and we demonstrated that this release was unaffected by treatment with the TxA2 receptor antagonist SQ29548. We further demonstrated that, in the absence of SQ29548, postischemic recovery of coronary flow and left ventricular stroke work was only approximately 60%. In contrast, postischemic recovery was approximately 90% in hearts treated with SQ29548-a significant improvement compared with controls. Because SQ29548 is known to pre-

692

The Journal of Thoracic and Cardiovascular Surgery April 1993

Byrne et al.

Table II. Coronary effluent levels of thromboxane B2 (pgjml) CP (n = 7)

SQ (n = 7) 'I-Mcan

CP

+ SQ

Preischemic

Postischemic

'I- Mean Pre + Post

57 ± 15 82 ± 9 69.6 ± 9.8

109 ± 22 102 ± 10 105.6 ± 12.4*

82.3 ± 15.5 92.0 ± 7.3

Values (mean ± standard error of the mean): CPo Unmodified crystalloid cardioplegic solution: SQ. CP plus thrornboxane receptor antagonist SQ29548; LMeanPre + Post, overall mean of preischemic plus postischernic values; LMeanCP + SQ. overall mean of CP plus SQ values. *p < 0.05 versus preischemic.

vent agonist-induced vasoconstriction in Langendorff rat hearts," these results suggest that a causal relationship exists between postischemic cardiac TxA2 release and the poor recovery of coronary flow and left ventricular stroke work observed in the control rats. We therefore hypothesize that postischemic cardiac TxA2 release directly impairs recovery of ventricular function through the mechanism of TxA 2-mediated coronary vasoconstriction. These results further imply that interventions directed against this mechanism can have a significant and practical benefit. Had our experiment been performed in a bloodperfused system, the effects of cardiac TxA 2 release after cardioplegia might have been amplified by circulating platelets and leukocytes. In such a milieu, cardiac-derived TxA 2 may be an even greater contributing factor to reperfusion injury. In support of this hypothesis are recent findings, conducted in a Langendorff model, which demonstrate that after ischemia and reperfusion under stimulation with f-Met-Leu-Phe, a potent chemotactic peptide, cardiac release of TxA 2 increases dramatically (approximately 35-fold).11 This finding not only confirms that the cellular and biochemical machinery capable of producing large amounts of TxA2 are present in cardiac tissue but also implies that in a blood-perfused system in which chemotactic agents are abundant, large amounts of cardiac TxA 2 may be produced. In this study, with an approximately 50% increase in cardiac TxA 2 release after ischemia-reperfusion, control hearts recovered only approximately 60% of coronary flow and hemodynamic function. With amplification of cardiac TxA 2 release by chemotactic agents, an even greater effect on hemodynamic recovery could be expected. Furthermore, through its activation of blood platelets and leukocytes, cardiacderived TxA2 may promote further injury. However, had our experiment been performed in a blood-perfused system, these confounding hematologic variables could not have been excluded, making it much more difficult to establish the potential cause-and-effect relationship between cardiac TxA2 and impaired hemodynamic recovery implied in our results. Establishment of this

Table III. Myocardial water content Myocardial water content (%) Preischemic (n = 7)

Postischemic CP (n = 8) Past ischemic SQ (n = 8)

75.8 ± 0.6 82.6 ± 0.8* 80.1 ± 0.9*

Values (Mean ± standard error of the mean): CP, Unmodified crystalloid cardioplegic solution; SQ. CP plus thromboxane receptor antagonist SQ29548. *p < 0.05 versus preischemic.

relationship will enable future investigations to extrapolate these results to blood-perfused systems. We acknowledge that whether this relationship exists in bloodperfused systems remains speculative. The mechanism by which ischemia-reperfusion stimulates cardiac TxA2 release is not answered in this study, but it may be related to the increase in intracellular free calcium caused by ischemia-reperfusion, Because the rate-limiting step in TxA2 production is the availability of its phospholipid membrane precursor arachidonic acid, and because arachidonic acid production is governed by the activity of phospholipase A 2, a calcium-sensitive enzyme, increased intracellular free calcium may be the critical stimulus for cardiac TxA2 release. 12 This concept is supported by studies on endothelial cells in culture, which have documented the inhibition of endothelial TxA 2 production by a calcium channel blocker.l' The severity of ischemic injury necessary for cardiac TxA2 release remains controversial. Early studies documented the augmentation of cardiac TxA2 release after brief (5 minutes) ischemia or hypoxia and its attenuation by nonsteroidal antiinflammatory agents or a TxA2 synthetase inhibitor. 14 However, a more recent study suggests that at least 60 minutes of ischemia is necessary for its release'>, a finding consistent with our results. The question as to whether SQ29548 has intrinsic myoprotective properties apart from its ability to inhibit TxA 2-mediated vasoconstriction cannot be fully dismissed in the present study. However, in a study including a Langendorff-perfused rat heart model, one similar to ours, Stahl and colleagues? demonstrated that a study including SQ29548 did not directly affect the coronary vasculature. They concluded that SQ29548 was both potent and specific as a TxA 2 receptor antagonist in the coronary vasculature of the rat heart. In another study, Grover and colleagues 16 reported that although SQ29548 significantly reduced infarct sizes in a canine model of90 minutes of coronary occlusion and reperfusion, a similar but inactive compound SQ29585 failed to reduce infarct sizes. They concluded that although the possibility of non-TxAj-related effects of SQ29548 could not be fully excluded, the mechanism of myoprotection provided by SQ29548 was at least partially due to its effects on TxA2

The Journal of Thoracic and Cardiovascular Surgery Volume 105, Number 4

receptor antagonism. We concur with this conclusion and suggest that the myoprotective effects of SQ29548 observed in the present study were at least in part due to its effects as a specific TxA 2 receptor antagonist in rat coronary endothelium." We acknowledge, however, that we cannot fully exclude subtle nonthromboxane effects. It is not indisputably known from which cellular elements cardiac TxA2 is released. Previous reports have implicated platelets deposited in the coronary vasculature during cardioplegia 17 or during reperfusion," 18 the coronary vasculature itself, 19-20 or intracardiac mononuclear elements.' However, the reports attributing TxA2 release to trapped intracoronary platelets were conducted in blood-perfused large animal models in which hematologic variables cannot be excluded. 7,17, 18 In the present study, although intracoronary platelet deposition during cardiectomy may be a potential source of TxA2, this phenomenon was disputed in crystalloid-perfused Langendorff systems. 1I We therefore suggest that the cellular origins of cardiac-derived TxA 2 in the present study were either endothelial cells l9-2o or intracardiac mononuclear elements, such as resident cardiac mast cells.' Another important finding in our study was that myocardial edema did not account for all the changes in coronary flow and recovery of function. Failure of SQ29548 treatment to significantly affect edema suggests that the deterioration of coronary flow and hemodynamic performance in control hearts was not entirely a consequence of myocardial edema, but rather, at least in part, a consequence of vasoconstriction mediated by cardiac-derived TxA2. Without a decrease in myocardial edema, improved coronary flow in SQ29548-treated hearts was most likely due to prevention of TxA 2-mediated vasoconstriction and its consequent effects on coronary flow and hemodynamic performance. In conclusion, this report has confirmed increased cardiac TxA2 release after ischemia-reperfusion and has documented its effects on coronary flow and hemodynamic recovery, thereby providing evidence that the TxA2 released from cardiac tissue after ischemia may be among the initiating mediators of reperfusion injury. REFERENCES I. Elgebaly S, Masetti P, Allam M, Foronhar F. Cardiac derived neutrophil chemotactic factors: preliminary biochemical characterization. J Mol Cell Cardiol 1989; 21:585-93. 2. Keller A, Clancy R, Barr M, Marboe C, Cannon P. Acute reoxygenation injury in the isolated rate heart: role of resident cardiac mast cells. Circ Res 1988;63:1044-52. 3. Karmazyn M. Synthesis and relevance of cardiac eicosanoids with particular emphasis on ischemia and reperfusion. Can J Physiol PharmacoI1989;67:912-21.

Byrne et al.

693

4. Garlick P, Mashiter G, DiMarzo V, Tippins J, Morris H, Maisey M. The synthesis,release and action ofleukotrienes in the isolated, unstimulated buffer-perfused rat heart. J Mol Cell Cardiol 1989;21:1101-10. 5. Kobinia G, LaRaia P, D'Ambra M, et al. Effect of experimental cardiopulmonary bypass on systemic and transcardiac thromboxane B2levels. J THORAC CARDIOVASC SURG 1986;91:852-7. 6. Kobinia G, LaRaia P, D'Ambra M, Peterson M, Buckley M. Influence of coronary artery obstruction on cardiac prostaglandin metabolism during experimental cardioplegic arrest. Thorac Cardiovasc Surg 1986;34:124-7. 7. Teoh K, FremesS, Weisel R, et al. Cardiac release of prostacyc1in and thromboxane A2 during coronary revascularization. J THoRAc CARDlOVASC SURG 1987;93:120-6. 8. vanRijk G, Foegy M, Ramwell P, Goldman M, Lower R. Long-term myocardial preservation:thromboxane production and coronary resistance. J Surg Res 1983;35:417-20. 9. Stahl G, Darium H, Lefer A. Antagonism ofthromboxane actions in the isolated perfused rat heart. Life Sci 1986;38:2037-41. 10. Ogletree M, Harris D, Greenberg R, Hasanger M, Nakane M. Pharmacologic actions of SQ29548: a novel selective thromboxane antagonist. J Pharmacol Exp Ther 1985; 234:435-41. II. Stahl R, Deutsch E, Fishes C, Warsaw D, Addonizio P. Cardiac ischemia and endothelial function in the isolated rat heart. J Surg Res 1989;47:94-104. 12. Gerritsen M. Eicosanoid production by the coronary microvascular endothelium. Fed Proc 1987;46:47-53. 13. Romadan F, Upchurch G, Keagy B, Johnson G. Endothelial thromboxane production and its inhibition by calcium channel blocker. Ann Thorac Surg 1990;49:916-9. 14. DeCaterinia R, Giannessi D, Lazzerini G, Nanni N, Pelosi G, Barsacchi R. Characteristics of thromboxane production by the isolated-perfused rat heart. Adv Prostaglandin Thromboxane Leukot Res 1987;17:384-7. 15. Engel W, VanBilsen M, Dergoot M, et al. Ischemia and reperfusion induced formation of eicosanoidsin isolated rat hearts. Am J Physiol 1990;258:H1865-71. 16. Grover G, Schumacher W. Effect of the thromboxane receptor antagonist SQ 29,548on myocardial infarct size in dogs. J Cardiovasc Pharmacol 1988;11:29-35. 17. Rosenbaum D, Levitsky S, Silverman N, et al. Cardioplegia does not prevent reperfusion injury induced by intracoronary platelet deposition. Circulation 1983;68(Pt 2):III02. 18. Teoh K, Christakis G, Weisel R, et al. Prevention of myocardial platelet deposition and thromboxane release with dipyridamole. Circulation 1986;74(Pt 2):IIII45. 19. Ingerrnan-Wojenski C, Silver M, Smith J. Bovine endothelial cells in culture produce thromboxane as well as prostacyc1ine. J Clin Invest 1981;67:1292-6. 20. Metha J, Roberts A. Human vascular tissue produces thromboxane as well as prostacyc1ine. Am J Physiol 1983; 244:R839-44.