Heparinoids with low anticoagulant potency attenuate postischemic endothelial cell dysfunction

Heparinoids with low anticoagulant potency attenuate postischemic endothelial cell dysfunction W. Charles Sternbergh III, M D , Michael Sobel, M D , and

R a y m o n d G. Makhoul, M D , Richmond, Va. Purpose: Although standard heparin has been demonstrated to reduce endothelial cell dysfimction in acute ischemia-reperfusion injury, its mechanism of action remains unknown. We hypothesized that heparin's salutary endothelial effects are independent of its conventional anticoagulant activity and are not caused by nonspecific polyanion effects. Methods: Isolated rat hindlimbs were perfused at constant pressure with an albuminenriched crystalloid buffer. After 60 minutes of normothermic ischemia, endothelial function was assessed by measurement of endothelial-dependent vasodilation by log increment infusion of acetylcholine. Endothelial-independent vasodilation was measured by exposure to nitroprusside. Some groups were pretreated with heparinoids possessing minimal or intermediate anticoagulant activity. Results: Treatment with heparinoids with low anticoagulant activity significantly increased endothelial-dependent vasodilation when compared with the nontreated ischemic group and were statistically indistinguishable from the nonischemic control. Treatment with dextran sulfate, a randomly sulfated polymer with size and charge characteristics similar to heparin, did not change postischemic vasodilation. Endothelial-independent vasodilation was largely unaffected by ischemia-reperfusion or drug treatment. Conclusions: A heparinoid with negligible antithrombin-binding activity (Astenose) attenuated postischemic endothelial dysfunction, suggesting that its mechanism of action was independent of anticoagulant activity. Failure of dextran sulfate to be protective implied that the effect was not caused by nonspecific polyanion action. (J VAsc SURG 1995;21:477-83.)

In acute ischemia, pharmaceutical heparin is frequently used to retard thrombus formation or propagation. Reduced endothelial and end-organ injury after heparin administration in experimental models of myocardial, 1 cerebral, 2 and extremity 3,4 ischemia-reperfusion injury appears to support this use. However, the mechanism of heparin's protective effect is unknown. Although heparin is used during these acute ischemic events for its antithrombotic effect, some of its salutary actions may be unrelated to anticoagulant activity. From the Department of Surgery, MedicalCollegeof Virginia and H. H. McGuire Veterans Affairs Medical Center, Richmond. Presented in part at the SurgicalForum during the Seventy-eighth Meeting of the American College of Surgeons, New Orleans, La., Oct. 11-16, 1992. Supported by a National ResearchServiceAward (F32 HL08491) from the National Institutes of Health. Reprint requests: Michael Sobel, MD, PO Box 980108, Richmond, VA 23298. 24/1/61350

We have previously demonstrated that standard porcine mucosal heparin attenuated postischemic endothelial cell dysfunction as measured by release of nitric oxide. 5 The ability of heparin to preserve postischemic nitric oxide release was dose dependent, providing a maximal protective effect at 4 Ixg/ml (0.5 U/ml). s These protective endothelial actions of heparin were believed to be independent of its anticoagulant activity because the rat hindlimb was perfused with a crystalloid-based buffer devoid of circulating plasma proteins or cells. Heparin can be chemically modified to have negligible anticoagulant activity while retaining many of its other biologic activities. We hypothesized that such a nonanticoagulant heparin might provide similar postischemic endothelial protective effects. An isolated rat hindlimb model of acute ischemia and reperfusion was used to study the effects of several heparin preparations on postischemic endothelial cell function. In addition to a heparinoid with 477

478

Sternbergh, Sobel, and Makhoul

negligible antithrombin-binding activity (Astenose), a low-molecular-weight heparinoid (ORG-10172) and standard heparin were separately evaluated for their ability to preserve postischemic endothelial cell function. Dextran sulfate was also studied to evaluate the effect of a randomly sulfated polymer with size and charge characteristics similar to heparin. Endothelial cell function was assessed by measurement of nitric oxide release from the intact vasculature of the postischemic rat hindlimb. MATERIAL A N D M E T H O D S Animal care complied with the "Principles of Laboratory Animal Care" (formulated by the National Society for Medical Research) and the Guidefor the Care and Use ofLaboratoryAnimals (NIH Publication No. 86-23, revised 1985). All studies were approved by the Animal Care and Use Committee of Virginia Commonwealth University/Medical College of Virginia. Model. A previously described isolated rat hindlimb model peffused with modified Krebs-Henseleit buffer was used in these studies, 6'7 Briefly, the perfusion apparatus was gravity fed and delivered a constant perfusion pressure of 70 m m Hg. The perfusate chambers and all tubing delivering perfusate to the limb were waterjacketed to maintain a constant temperature of 37 ° C. By halting arterial perfusate flow, the limb could be subjected to complete ischemia. Male Sprague-Dawley rats weighing 325 to 400 grams were anesthetized with pentobarbital 50 mg/kg intraperitoneally, and the hindlimb was skinned. The femoral artery and vein were isolated immediately distal to the inguinal ligament and sequentially cannulated with 24-gauge and 20-gauge Teflon catheters, respectively (Critikon, Tampa, Fla.). The femoral artery was flushed with standard pharmaceutical-grade heparin (Lyphomed, Rosemont, Ill.) 500 U in 0.5 ml normal saline solution before venous cannularion. The limb was then amputated, and perfusion was immediately begun with a crystalloid-based buffer (see below). Saline solutionsoaked gauzes were placed around the limb to prevent desiccation during the study. Limb temperature was maintained at 37 ° __. 0.2 ° C throughout the experiment, including the ischemic interval. Venous flow from the hindlimb was measured by timed collection over a 30-second interval and expressed per gram dry limb weight (ml/min/gm dry limb wt). Venous effluent was not recirculated. Perfusate and reagents. The perfiasate was a modified Krebs-Henseleit buffers'r of the following composition (mmol/L): NaCI 118.4, KCI 4.7, CaC12

JOURNAL OF VASCULAR SURGERY March 1995

2.5, KH2PO 4 1.2, MgSO4-7H20 1.2, NaHCO s 34.9, glucose 10.0 plus 4 gm/dl of serum bovine albumin (Sigma Chemical Co., St. Louis, Mo.). The peffusate was prepared daily and filtered through a 1.2 ~m nylon filter (MSI, Westboro, Mass.) before use. The buffer was constantly bubbled with 95% O2/5% CO 2 during the studies and was not recirculated. Phenylephrine HCI (1.0 to 2.5 t~mol/L; Winthrop-Breon, New York, N.Y.) was added directly to the perfusate to preconstrict the limb to a standard total vascular resistance. 7 Acetylcholine CI (0.001, 0.01, 0.1, 1.0 i~g/min, Sigma) and sodium nitroprusside (0.5, 5.0, 50 i~g/min, Elkins-Sirm, Cherry Hill, N.J.) were suspended in normal saline solution and delivered into the arterial perfusate line with an infusion pump (Harvard Model 901; Harvard Apparatus Co., Inc., South Natick, Mass.). Limb flow rates became steady after 10 minutes of perfusion with phenylephrine and 1 to 1.5 minutes with acetylcholine and nitroprusside. 7 Indomethacin (2.8 I~mol/L, Sigma Chemical Co.) was diluted in 0.1 mol/L Na2CO 3 and added directly to the perfusate to inhibit prostanoid production. 7's Three heparin preparations were individually added to both the preischemic and postischemic perfusate at a concentration of 4 I~g/ml (0.5 USP units/ml). Previous dose response studies with standard heparin showed this concentration to have maximal endothelial protective effects. 5 This pharmacologic dose correlates with the average plasma heparin level in a patient treated with therapeutic anticoagulafion (activated partial thromboplasrin time [aPTT] 2 x control).9 The following heparins were used: (1) a standard pharmaceutical-grade porcine intestinal mucosal heparin (lot no. 301182, Lyphomed); (2) a heparin whose antithrombin-binding activity had been chemically destroyed by periodate oxidation of standard porcine heparin (Astenose; Glycomed, Inc., Alameda, Calif.); (3) a low-molecular-weight heparinoid (ORG-10172 [Lomoparan], batch BT, Organon, Inc., West Orange, N.J.). Dextran sulfate (lot 116F-0296, Sigma) was also used in separate experiments at a concentration of 4 t~g/ml. Potency of heparins. In vitro anti-Xa activity was determined by use of a modification of the method of Teien. 1° For each heparin an aPTT dose-response curve was developed with use of semiautomated methods and reagents from Organon (Durham, N.C.). The relative concentration required to prolong the aPTT to twice the control value was expressed as a percentage of the potency of the standard porcine mucosal heparin.

JOURNAL OF VASCULAR SURGERY Volume

21, Number 3

Experimental protocol. All limbs were initially perfused for a 20-minute stabilization period. Control (nonischemic) limbs were then perfused for 60 minutes without ischemia. Native venous flow was measured after the 20-minute stabilization period, and phenylephrine was then added to the perfusate to preconstrict the vasculature. After 10 minutes of perfusion with phenylephrine, the preconstricted flow rate was recorded, which was used as the baseline for subsequent measurements. Endothelialdependent vasoreactivity (nitric oxide release) was then examined by measuring the change in flow in response to log increments of acetylcholine infused into the arterial perfusate line. 7 Flow measurements were taken 2 minutes after the initiation of each infusion. Endothelial-independent vasoreactivity was similarly assessed with nitroprusside after flow returned to its preconstricted baseline. Nitroprusside directly relaxes vascular smooth muscle. Postischemic limbs underwent an equivalent period of 20-minute stabilization perfusion and were then rendered totally ischemic for 60 minutes at 37 ° -+ 0.2 ° C. The limbs were then reperfused with the buffer containing phenylephrine. After 10 minutes, the preconstricted baseline was recorded, and vasoreactivity to acetylcholine and nitroprusside were tested as performed in the control limbs. Both control and postischemic groups underwent identical amounts of perfusion (30 minutes) before administration of the agonists. Experimental groups. Six groups of limbs were studied. The control (nonischemic) group (n = 7) underwent no ischemia. Postischemic limbs were either untreated (n = 12) or pretreated with standard heparin (n = 6), Astenose (n = 6), the lowmolecular-weight heparin fraction ORG-10172 (n = 5), or dextran sulfate (n = 7). These drugs were added to both the preischemic and postischemic perfusate. We have previously demonstrated that heparin had no effect on vasoreactivity of control (nonischemic) limb vasculature, s Data analysis and statistics. Total vascular resistance (TVR) was calculated by dividing the perfusion pressure (70 m m Hg) by the flow rate and expressed as m m Hg/ml/min/gm dry limb wt. 7 The effect of agonists on T V R was expressed as a percent change from the phenylephrine-preconstricted baseline. 7,11This baseline was measured before each dose response curve. Comparisons between the six experimental groups were made by one-way analysis of variance with Tuke~s studentized range test for comparison of variance between any two groups (SAS, Cary,

Sternbergh, Sobd, and 3~lakhoul 479

N.C.). Groups were considered significantly different i f p < 0.05. All data are presented as mean +__SE. RESULTS Baseline parameters. The native and preconstricted T V R for the six experimental groups are presented in Table I. There were no significant differences between the native TVRs in the control (no ischemia) or nontreated ischemia group when compared with any of the treated ischemia groups. The native T V R of the ORG-10172 group was higher than that of the heparin group. Preconstricted TVRs, however, were not significantly different among the six experimental groups. Endothelial-dependent vasodilation. Infusion of log increments of acetylcholine caused graded reductions in limb TVR in the control (nonischemic) group (Fig. 1). Sixty minutes of normothermic ischemia and 10 minutes of reperfusion (postischemic group) significantly impaired acetylcholineinduced vasodilation as reflected by a reduction in the percent change in TVR. In fact, during 0.01 wg/min acetylcholine infusion a paradoxic vasoconstriction was observed in this group. Groups pretreated with Astenose, ORG-10172, or standard heparin had a significant increase in postischemic endothelialdependent vasodilation when compared with the untreated postischemic group (p < 0.05), and no paradoxic vasoconstriction occurred. These three groups were not significantly different from one another. Pretreatment with dextran sulfate did not significantly preserve postischemic vasodilation to acetylcholine. Endothelial-independent vasodilation. Infusion of log increments ofnitroprusside caused graded reductions in limb T V R in the control group (Fig. 2). This endothelial-independent vasodilation was not significantly changed from control in the untreated postischemic, Astenose, ORG-10172, or standard heparin group. Vasodilation of the dextran sulfate group was not significantly different from any other postischemic group, but when compared with the control group, there was an isolated reduction of TVR, which was limited to the lowest dilution of nitroprusside (0.5 ~g/min). The dextran sulfate group was not significantly different from control when subjected to 5 or 50 wg/min of nitroprusside, however. DISCUSSION Ischemia-reperfusion-induced endothelial cell dysfunction was significantly reduced after pretreatment with Astenose, a heparin with negligible

JOURNAL OF VASCULAR SURGERY March 1995

480 Sternbergh, Sobel, and Makhoul

Acetylcholine 0.001

-5

0.01

(/~g/min) 0.1

1.0

~> --15

-z5

"~r.) - 3 5

..... T v ...........

11--10__ • IschemiaN° Ischemia

~

.

~ ............V Ischemia + Astenose TM O--O Ischemia + 0RG-10172 -45

Z~--A O--O

~R~I ~.Am

Ischemia + Heparin Ischemia + Dextran Sulfate

l p ( 0.05 vS

)

Ischemia

Fig. 1. Endothelial-dependent vasodilation to acetylcholineas reflected by percent change in TVR. Two control groups underwent no ischemias (n = 7) or 60 minutes of ischemias (n = 12). Four groups underwent 60 minutes of ischemia after pretreatment with Astenose (n = 6), ORG-10172 (n = 5), standard heparins (n = 6), or dextran sulfate (n = 7). Limbs treated with dextran sulfate were not significantlydifferent from ischemia control group. Table I. TVR Ischemia

Native TVR Preconstricted TVR

No ischemia (n=7)

No drug (n=12)

Heparin (n=6)

43.7 ± 3.7 95.5 ± 7.1

39.1 + 3.6 91.2 -+ 6.3

35.9 -+ 1.9 98.1 + 4.8

Astenose (n=6) 46.9 + 2.2 104.5 _+ 6.4

ORG-IO172 (n=5) 50.0 + 2.4* 112.0 + 7.6

Dextran sulfate (n=7) 48.0 -+ 1.8 106.2 + 3.1

*p < 0.05 versus native TVR in heparin group. Otherwise, p = NS between the six groups' native or preconstricted TVR. TVR is expressed in mm Hg/ml/min/gm dry limb weight.

conventional anticoagulant activity. These protective effects, including elimination of the paradoxical vasoconstriction observed, were statistically indistinguishable from that conferred by standard heparin. Astenose is produced by a proprietary process of periodate oxidation of standard porcine mucosal heparin. This produces a heparinoid with minimal antithrombin binding activity, as reflected in the low aPTT activity (Table II). In contrast, standard heparin has approximately 180 anti-Xa units of anticoagulant activity and binds antithrombin III avidly. We had previously hypothesized that heparin's salutary effects on postischemic endothelial cell function may be independent of its conventional anticoagulant activity,s The demonstration of comparable salutary effects of a heparin with negligible anticoagulant activity provide strong confirmatory evidence for this hypothesis. ORG-10172, a heparinoid preparation with intermediate conventional anticoagulant effects, was also examined for its potential endothelial pro-

tective effects. ORG-10172 is composed of a mixture of glycosaminoglycans, including heparan sulfate ( - 8 0 % ) , dermatan sulfate (8% to 16%), and chondroitin sulfate (< 9%). This compound has a much lower affinity for antithrombin III when compared with standard heparin but does have significant anti-Xa activity (Table II). Pretreatment with ORG10172 also significantly attenuated postischemic endothelial cell damage. Thus two glycosaminoglycan preparations with different anticoagulant potencies produced similar effects on postischemic endothelial cell function. These data add further credence to the notion that the endothelial protective effects of various heparin preparations were not a factor of their anticoagulant activity. Paradoxic vasoconstriction of vessels exposed to a compound that usually causes dilation is a welldescribed phenomenon. 12 Typically this deleterious event occurs in blood vessels that have been injured. Indeed, in the untreated control group undergoing ischemia and reperfusion, paradoxic vasoconstriction

JOURNAL OF VASCULAR SURGERY Volume 21, Number 3

Sternbergh, Sobel, and Makhoul

481

Nitroprusside (Ng/min) 0.5

-10

> -r--~

0) t:d3

5.0

50

?;-...

-20 -30-

~2 r..) - 4 0 ' -50

V ............~ O--O A--/X ~--~

Ischemia Ischemia Ischemia Ischemia

+ + + +

Astenose 0RG-10172 Heparin Dextran Sulfate

~

Fig. 2. Endothelial-independent vasodilation to nitroprusside as reflected by percent change in TVR. Two control groups underwent no ischemia ~ (n = 7) or 60 minutes of ischemia s (n = 12). Four groups underwent 60 minutes of ischemia after pretreatment with Astenose (n = 6), ORG-10172 (n = 5), standard heparin 5 (n = 6), or dextran sulfate (n = 7). There were no significant differences between any group at any nitroprusside dose with exception of ischemia + dextran sulfate group versus nonischemic control at 0.5 ~g/min nitroprusside (p < 0.05). Table II. Potency of heparins Heparin type

Anti-Xa activity (units~rag)

aPTT activity (%)

Average tool. wt. (Daltons)

Porcine mucosal heparin Astenose ORG-10172 (Lomoparan) Dextran sulfate

180 <5 15 <5

100 10 10 <5

15,000 11,000 6,000 8,000

aPTT, Activated partial thromboplastin time; mol. wt., molecular weight.

was observed. This phenomenon was not seen in the nonischemic control group, suggesting that the paradoxical vasoconstriction was a consequence of the ischemia and not the use of acetylcholine per se. All heparinoids were equally effective in obviating this deleterious vasoconstriction. The protective effects of heparin and related glycosaminoglycans on endothelial function have been studied in several divergent models of experimental cellular injury. Heibert et al. 13 exposed endothelial cell cultures to xanthine and xanthine oxidase with presumptive liberation of damaging oxygen free radicals. Pretreatment with heparin increased cell viability and reduced lactate dehydrogenase (LDH) release when compared with nontreated control subjects. Pretreatment with heparan sulfate, a heparin with reduced anticoagulant activity, also produced an increase in cell viability but no change in L D H release.~3 After balloon cathete r injury of rabbit

aortas, Light et al. 14 found that heparin treatment normalized the structure of subsequently formed neoendothelium and also increased its endotheliumderived relaxing factor (EDRF) release when compared with untreated controls. Treatment with lowmolecular-weight heparin had no beneficial effects, however. Hemorrhagic shock-induced cardiac, hepatic, and renal dysfunction has been attenuated after treatment with both standard is and nonanticoagulant heparin. 16 Preliminary work also suggested that systemic treatment with heparan sulfate attenuated gut dysfunction caused by hemorrhagic shock. 17 Thus heparin and related glycosaminoglycans appear to confer protective effects under a variety of experimental conditions. Heparins are strongly polyanionic compounds and as such could confer beneficial membranestabilizing effects to injured endothelium. ~8 To investigate the possibility that the salutary effects of

482

JOURNAL OF VASCULAR SURGERY March 1995

Sternbergh, Sobel, and Makhoul

heparin in this study might be due to nonspecific polyanionic effects, a separate experimental group was treated with dextran sulfate. This randomly sulfated polymer has size and charge characteristics similar to that of heparin (Table II). We found that dextran sulfate did not attenuate postischemic endothelial damage in this model, suggesting that the beneficial effects of heparin were not a nonspecific polyanion effect. However, others have arrived at the opposite conclusion. In endothelial cell cultures exposed to xanthine and xanthine oxidase, Heibert et al.ls found that pretreatment with dextran sulfate (mol wt 8000, 5 ~g/ml) did have salutary effects, producing increased cell viability and reduced L D H release. 13 The large differences between our ex vivo model and this in vitro study make comparisons of these data problematic. Further study with this compound may be warranted. Endothelial-independent vasodilation of the postischemic hindlimbs was largely unaffected by treatment with the various heparin preparations. Limbs treated with dextran sulfate had an isolated reduction in vasodilation when compared with the nonischemic control at the lowest dilution of nitroprusside. However, none of the drug treatment groups (heparin, Astenose, ORG-10172, dextran sulfate) were significantly different from one another. Thus the maintenance of postischemic vasoreactivity of the hindlimbs after treatment with Astenose, ORG-10172, or standard heparin did not appear to be caused by direct effects on the vascular smooth muscle. Measurement o f endothelial dysfunction. Endothelial cell dysfunction was assessed by measurement of EDRF release from the intact vasculature of the rat hindlimb. Previous work by us ~ and others 19 using competitive inhibitors of nitric oxide release such as NC-monomethyl-L-arginine in intact hindlimb models have confirmed that nitric oxide is the principle EDRF in these preparations. The ability of endothelium to release nitric oxide is a sensitive indicator of its functional status; reductions in nitric oxide release can actually precede end-organ injury. 2° Thus measurement of EDRF (nitric oxide) release is a valid parameter in the assessment of endothelial damage from ischemia-reperfusion injury. Limitations of study. Drug treatment was initiated before the onset of ischemia. It is not yet known whether delayed administration of these heparinoids would be as beneficial. In an in vivo canine hindlimb model, Ricci et al.21 have shown that standard heparin administered 30 minutes before reperfusion after 4 hours of ischemia reduced subse-

quent skeletal muscle necrosis. In this study only a single parameter of endothelial cell injury was examined. However, we have subsequently demonstrated significant reductions in vascular permeability with heparin treatment in this model (unpublished observations). Others have also demonstrated an attenuation of the rise in postischemic permeability, as well as reductions in skeletal muscle necrosis with standard heparin treatment.3 Finally, because of the small sample sizes, the possibility of a type II error could not be eliminated in the dextran sulfate groups. Heparinoid pretreatment provided significant protection from endothelial damage caused by ischemia and reperfusion. The inability of dextran sulfate to reduce endothelial dysfunction suggested that the beneficial effects were not a nonspecific polyanion effect. The protective effect of these heparinoids appeared to be unrelated to anticoagulant activity because they possessed only minimal (Astenose) or intermediate (ORG-10172) conventional anticoagulant activity. Endothelial dysfianction is a common feature of organ dysfunction in septic and hemorrhagic shock, myocardial and extremity ischemia, and transplanted organs. Accordingly, further study of these heparinoids are warranted for potential use as endothelial protective agents. We thank Glycomed, Inc., for providing Astenose and Organon, Inc., for supplying ORG-10172 for this study. REFERENCES

1. Saliba MJ, Coven JW, Bloor CM. Effects of heparin in large doses on the extent of myocardial ischemia after acute coronary occlusion in the dog. Am J Cardiol 1976;37:599603. 2. Smith DR, Ducker TB, Kempe LG. Temporary experimental intracranial vascular occlusion: effect of massive doses of heparin on brain survival. J Neurosurg 1969;30:537-44. 3. Hobson RW II, Wright JG, Fox D, Kerr JC, Heparinization reduces endothelial permeability and hydrogen ion accumulation in a canine skeletal muscle ischemia-reperfusion model. J VASC SURG 1988;7:585-90. 4. Wright JG, Kerr JC, Valeri CR, Hobson RW II. Heparin decreases ischemia-reperfusion injury in isolated canine gracilis model. Arch Surg 1988;123:470-2. 5. Sternbergh WC III, Makhoul RG, Adelman B. Heparin prevents postischemic endothelial cell dysfunction by a mechanism independent of its anticoagulant activity. J" VAsc SURG 1993;17:318-27. 6. Sternbergh WC III, Adelman B. The temporal relationship between endothelial cell dysfunction and skeletal muscle damage after ischemia and reperfusion. J VAsc SuRe 1992; 16:30-9. 7. Sternbergh WC III, Makhoul RG, Adelman B. Nitric oxide mediated endothelial-dependent vasodilation is selectively attenuated in the post-ischemic extremity. Surgery 1993;114: 960-7.

JOURNAL OF VASCULARSURGERY Volume 21, Number 3

8. Stewart AG, Piper PJ. Vasodilator actions of acetylcholine, A23187 and bradykinin in the guinea-pig isolated perfused heart are independent of prostacyclin. Br J Pharmacol 1988;95:379-84. 9. Brandt JT, Triplett DA. Laboratory monitoring of heparin: effect of reagents and instruments on the activated partial thromboplastin time. Am J Clin Pathol 1981;76(suppl): 530-6. 10. Teien AN, Lie M. Evaluation of an amidolytic heparin assay method: increased sensitivity by adding purified antithrombin III. Thromb Res 1977;10:399-410. 11. Mugge A, Lopez JAG, Piegors DJ, Breese KR, Heistad DD. Acetylcholine-induced vasodilation in rabbit hindlimb is not inhibited by analogues of L-arginine. Am J Physio11991;260: H242-H247. 12. Moncada S, Palmer RMJ, Higgs EA. Nitric oxide: physiology, pathophysiology, and pharmacology. Pharmacol Rev 1991;43:109-42. 13. Heibert L, Liu J. Protective action of polyelectrolytes on endothelium. Sere Thromb Hemost 1991;17(Suppl 1) :42-6. 14. Light JT, Belian JA, Roberts MP, et al. Heparin treatment enhances the recovery of neoendothelial acetylcholineinduced vascular relaxation after balloon catheter injury in the rabbit aorta. Circulation 1993;88(part II):413-9. 15. Wang P, Singh G, Rana MW, Ba ZF, Chaudry IH. Preheparinization improves organ function after hemorrhage and resuscitation. Am J Physiol 1990;259:R645-R650.

Sternbergh, Sobd, and 3,lakhoul 483

16. Wang P, Ba ZF, Chaudry IH. Chemically modified heparin improves hepatoceliular function, cardiac output, and microcirculation after tranma-hemorrhage and resuscitation. Surgery 1994;116:169-76. 17. Singh G, Chaudry KI, Chandry IH. Restoration of gut absorptive capacity following trauma-hemorrhagic shock by the adjuvant use of heparan sulfate [Abstract]. J Trauma 1992;33:163. 18. Srinivasan S, Aaron S, Chopra PS, Lucas T, Sawyer PN. Effect of thrombotic and antithrombotic drugs on the surface charge characteristics of canine blood vessels. Surgery 1968;64: 82737. 19. Bellan JA, Minkes ILK, McNamara DB, Kadowitz PJ. NW-nitro-L-arginine selectivelyinhibits vasodilator responses to acetylcholine and bradykinin in cats. Am J Physiol 1991;260:H1025-H1029. 20. Dauber IM, VanBenthuysen KM, McMurtry IF, et al. Functional coronary microvascular injury evident as increased permeability due to brief ischerrLiaand reperfusion. Circ Res 1990;66:986-98. 21. Ricci MA, Corbisiero RiM, Mohamed F, Graham AM, Symes JF. Replication of the compartment syndrome in a canine model: experimental evaluation of treatment. J Invest Surg 1990;3:129-40. Submitted May 11, 1994; accepted Oct. 17, 1994.