Animal models of catheter-induced intimal hyperplasia in type 1 and type 2 diabetes and the effects of pharmacologic intervention.

Diabetes is a complex disorder characterized by impaired insulin formation, release or action (insulin resistance), elevated blood glucose, and multiple long-term complications. It is a common endocrine disorder of humans and is associated with abnormalities of carbohydrate and lipid metabolism. There are two forms of diabetes, classified as type 1 and type 2. In type 1 diabetes, hyperglycemia is due to an absolute lack of insulin, whereas in type 2 diabetes, hyperglycemia is due to a relative lack of insulin and insulin resistance. More than 90% of people with diabetes have type 2 with varied degrees of insulin resistance. Insulin resistance is often associated with impaired insulin secretion, and hyperglycemia is a common feature in both types of diabetes, but failure to make a distinction between the types of diabetes in different animal models has led to confusion in the literature. This is particularly true in relation to cardiovascular disease in the presence of diabetes and especially the response to vascular injury, in which there are major differences between the two types of diabetes. Animal models do not completely mimic the clinical disease seen in humans. Animal models are at best analogies of the pathologic process they are designed to represent. The focus of this review is an analysis of intimal hyperplasia following catheter-induced vascular injury, including factors that may complicate comparisons between different animal models or between in vitro and in vivo studies. We examine the variables, pitfalls, and caveats that follow from the manner of induction of the injury and the diabetic state of the animal. The efficacy of selected antidiabetic drugs in inhibiting the development of the hyperplastic response is also discussed.

Effect of combination endothelial nitric oxide synthase gene therapy and sildenafil on erectile function in diabetic rats.

Erectile dysfunction associated with diabetes mellitus is caused in part by disordered endothelial smooth muscle relaxation, neuropathy, and a decrease in cavernosal nitric oxide synthase (NOS) activity. The purpose of this study was to determine whether a combination of sildenafil and adenoviral gene transfer of endothelial NOS (eNOS) could enhance the erectile response in diabetic rats. Five groups of animals were utilized: (1) age-matched control rats, (2) streptozotocin (STZ)-induced diabetic rats (60 mg/kg i.p.), (3) STZ-rats + sildenafil (2 mg/kg i.v.), (4) STZ-rats transfected with AdCMVbetagal or AdCMVeNOS, and (5) STZ-rats transfected with AdCMVeNOS +sildenafil (2 mg/kg i.v.). At 2 months after i.p. injection of STZ, groups 4 and 5 were transfected with the adenoviruses and 1-2 days after transfection, all animals underwent cavernosal nerve stimulation (CNS) to assess erectile function. Cyclic 3',5'-guanosine monophosphate (cGMP) levels were assessed in the cavernosal tissue. STZ-diabetic rats had a significant decrease in erectile function as determined by the peak intracavernosal pressure (ICP) and total ICP (area under the erectile curve; AUC) after CNS when compared to control rats. STZ-diabetic rats+AdCMVeNOS had a peak ICP and AUC, which were similar to control animals. STZ-diabetic rats administered sildenafil demonstrated a significant increase in peak ICP at the 5 and 7.5 V settings, while the AUC was significantly increased at all voltage (V) settings. The increase in both ICP and AUC of STZ-diabetic rats transfected with AdCMVeNOS at all V settings was greater than STZ-diabetic rats transfected with AdCMVbetagal. STZ-diabetic rats transfected with AdCMVeNOS and administered sildenafil had a significant increase in total ICP that was greater than eNOS gene therapy alone. Cavernosal cGMP levels were significantly decreased in STZ-diabetic rats, but were increased after transfection with AdCMVeNOS to values greater than control animals. In conclusion, overexpression of eNOS and cGMP in combination with sildenafil significantly increased both the peak ICP and total ICP to CNS in the STZ-diabetic rat, which was similar to the response observed in control rats. Moreover, the total erectile response was greater in STZ-diabetic rats receiving eNOS gene therapy plus sildenafil than STZ-rats receiving sildenafil or eNOS gene therapy alone.

Effect of pharmacological treatments for diabetes on homocysteine.

Hyperhomocysteinemia is a well-established risk factor for cardiovascular disease. The association of hyperhomocysteinemia with diabetes mellitus is complex and may explain some of the risk of CVD in diabetics not explained by traditional risk factors. Both modifiable and non-modifiable factors interact with homocysteine metabolism and determine the plasma homocysteine concentrations. These include genetic abnormalities, age, sex, and various nutritional and hormonal determinants, all of which play a role in atherosclerosis and accelerated peripheral and cardio-vascular disease (CVD). Several medications modulate homocysteine metabolism and hence may play a role in the pathogenesis of CVD. Changes in renal function and interference with the homocysteine metabolism account for some of these drug effects. While a few of these drugs raise plasma homocysteine concentrations, others are beneficial and may counter some of the deleterious effects of hyperhomocysteinemia. Treatment of hyperhomocysteinemia with vitamins lowers plasma homocysteine concentrations and also reverses many of these drug effects. Little data is available on the effect of this intervention on cardiovascular outcomes. This review briefly outlines the effect of various medications used in the management of type 2 diabetes mellitus and metabolic syndrome.

Hormones and homocysteine.

Hyperhomocysteinemia is a well-established risk factor for cardiovascular disease. Various factors, both modifiable and non-modifiable, interact with the homocysteine metabolism and determine the plasma homocysteine concentrations. These include genetic abnormalities, age, sex and various nutritional and hormonal determinants, all of which play a role in atherosclerosis and accelerated peripheral and cardio-vascular disease (CVD). Several hormones modulate homocysteine metabolism and hence may play a role in the pathogenesis of CVD. The mechanisms involved are unclear. The association of hyperhomocysteinemia with diabetes mellitus is complex and may explain some of the risk of CVD in diabetics not explained by traditional risk factors. Much conflicting data exists in the literature on the role of insulin on homocysteine metabolism, although insulin affects the enzymes regulating the homocysteine metabolism. Treatment of hyperhomocysteinemia with vitamins lowers plasma homocysteine concentrations. Little data is available on the effect of this intervention on cardiovascular outcomes. This review briefly outlines the homocysteine metabolism, summarizes its hormonal determinants, and discusses the role of hyperhomocysteinemia in diabetes, hyperlipidemia and other endocrine disorders.

Cyclic GMP-dependent protein kinase expression in coronary arterial smooth muscle in response to balloon catheter injury.

Arterial smooth muscle cells undergo phenotypic and proliferative changes in response to balloon catheter injury. Nitric oxide (NO) and cGMP have been implicated in the inhibition of vascular smooth muscle cell proliferation and phenotypic modulation in cultured-cell studies. We have examined the expression of the major cGMP receptor protein in smooth muscle, cGMP-dependent protein kinase I (PKG), in response to balloon catheter injury in the swine coronary artery. On injury, there was a transient decrease in the expression of PKG in neointimal smooth muscle cells when compared with medial smooth muscle cells. The decrease in PKG expression was observed in the population of proliferating cells expressing the extracellular matrix protein osteopontin but not in cells present in the uninjured portion of the media. Coincident with the suppression of PKG expression in neointimal cells after injury, there was a marked increase in the expression of type II NO synthase (inducible NOS [iNOS], NOS-II) in the neointimal cells. These results suggest that PKG expression is transiently reduced in response to injury in the population of coronary arterial smooth muscle cells that are actively proliferating and producing extracellular matrix proteins. The reduction in PKG expression is also correlated temporally with increases in inflammatory activity in the injured vessels as assessed by iNOS expression. Coupled with our current knowledge regarding the role of PKG in the regulation of cultured cell phenotypes, these results imply that PKG may also regulate phenotypic modulation of vascular smooth muscle cells in vivo as well.

Endothelial cell regrowth and morphology after balloon catheter injury of alloxan-induced diabetic rabbits.

Neointimal thickening after catheter injury has been reported to be influenced by the integrity of the vascular endothelium. We have previously shown that neointimal thickening is significantly reduced in alloxan-induced diabetic New Zealand White rabbits after catheter injury compared with euglycemic rabbits. In the present study, it was hypothesized that endothelial cell regrowth, morphology, and endothelium-dependent vasoreactivity after catheter injury are improved in the diabetic rabbit (glucose >/=400 mg/dl) compared with the euglycemic rabbit. Two weeks after catheter injury, the percent endothelial regrowth was significantly increased in diabetic animals compared with euglycemic animals (32.1 +/- 2 and 15.6 +/- 1, respectively; P < 0.05). The endothelial cell morphology analyzed by scanning electron microscopy was also restored 2 wk after catheter injury in thoracic aortas from the diabetic animals compared with vessels from euglycemic animals. Endothelium-dependent relaxation to ACh in vessels from diabetic and euglycemic rabbits was attenuated 2 wk after injury, and, although improved by 4 and 8 wk, relaxation remained significantly depressed. These results suggest that endothelial cell regrowth and morphology in diabetic animals was improved compared with euglycemic animals; however, endothelium-dependent vasoreactivity remained impaired. Thus the attenuated neointimal thickening seen in the diabetic rabbit may be a function of the rate and degree of regrowth rather than the normalization of ACh-induced relaxation.

The influence of candesartan and PD123319 on responses to angiotensin II in the hindquarters vascular bed of the rat.

The effects of the AT1 and AT2 receptor blockers candesartan and PD123319 on hemodynamic responses to angiotensin II (AngII) were investigated in the anesthetized rat. Injections of AngII caused dose-related increases in systemic arterial and in hindquarters perfusion pressure that were reduced by candesartan. The inhibitory effects of candesartan were insurmountable, and a vasodepressor or vasodilator response to AngII was not unmasked. The AT2 receptor antagonist PD 123319 had no effect on increases in systemic arterial and hindquarters perfusion pressure in response to AngII. The present results suggest that pressor responses to AngII are mediated by the activation of AT1 receptors, and that AT2 receptors do not appear to modulate hemodynamic responses to AngII in the anesthetized rat.

Analysis of the effects of candesartan on responses to angiotensin II in the hindquarters vascular bed of the cat.

The effects of the nonpeptide angiotensin II (AngII) AT1 receptor blocker candesartan on responses to AngII were investigated in the hindquarters vascular bed of the cat. Under constant-flow conditions, injections of AngII into the hindquarters perfusion circuit elicited dose-dependent increases in perfusion pressure. Candesartan in a dose of 3 microg/kg intravenously (i.v.) decreased vasoconstrictor responses to AngII in a surmountable manner. At doses of 30 and 300 microg/kg i.v., candesartan shifted the dose-response curve to AngII to the right in an insurmountable manner, indicating an insurmountable blockade of AT1 receptors. The inhibitory effects of the larger doses of candesartan on responses to AngII were long in duration, and the AT1 receptor blocker had little effect on baseline pressures. Candesartan was without effect on vasoconstrictor responses to norepinephrine, U46619, PGF2alpha, vasopressin, BAY K8644; biphasic responses to endothelin-1; or on vasodilator responses to acetylcholine, albuterol, or levcromakalim. These results indicate that candesartan is a potent and selective angiotensin AT1 receptor blocker that can induce both surmountable and insurmountable AT1 receptor blockade and provide support for the hypothesis that there are "spare" AT1 receptors in the hindquarters vascular bed of the cat.

Balloon catheter vascular injury of the alloxan-induced diabetic rabbit: the role of insulin-like growth factor-1.

Neointimal thickening following catheter injury is characterized, in part, by growth factor-induced vascular smooth muscle cell (VSMC) proliferation. It was hypothesized that a reduction in serum insulin-like growth factor-1 (IGF-1), characteristic of chemically-induced diabetes, would result in decreased VSMC proliferation and attenuate neointimal thickening. It was found that alloxan-treated New Zealand White rabbits exhibit varying degrees of glycemia. Rabbits classified as diabetic (glucose > or = 400 mg/dL) had significantly decreased serum concentration of IGF-1 (87.4+/-14 nmol/L vs. 170+/-14 nmol/L) and significantly decreased intimal/medial (I/M) ratios 2, 4, and 8 weeks after aortic injury compared to euglycemic rabbits (13.7+/-2, 21.1+/-2, 32.4+/-3 in euglycemics and 6.6+/-1, 14+/-2, 19+/-5 in diabetics, respectively). The I/M for high hyperglycemic animals (glucose 286-399 mg/dL) was comparable to diabetic animals yet their serum IGF-1 levels were normal rather than depressed. Vascular IGF-1 content similarly increased upon injury in both diabetic and euglycemic animals. In diabetic animals, proliferating cell nuclear antigen (PCNA) immunostaining was present by day 1 peaked by day 5 and returned to control by day 14. In euglycemic animals, staining by day 1 continued to increase through day 14. A similar increase in mitogen-activated protein kinase (MAPK) activity occurred from day 1 through day 5 in both diabetic and euglycemic animals. This is the first demonstration of an association between MAPK activity and VSMC proliferation following vascular injury in diabetic animals as previously reported in euglycemic animals. In conclusion, this study provides evidence against a direct effect of IGF-1 in the reduction in neointimal thickening, VSMC proliferation, and MAPK activity upon catheter injury in chemically-induced diabetic rabbits.

Pharmacologic management of peripheral vascular disease.

Although our understanding of the pathophysiology of atherosclerosis and peripheral vascular disease continues to grow, we have yet to discover a medication that can safely and efficaciously be given to most claudicants that will alleviate their symptoms to prevent disease progression. Many patients with intermittent claudication improve or remain stable without therapy if they attempt to alter their risk factors (e.g., control of diabetes, smoking cessation, lowering of cholesterol levels). However, many require concomitant drug therapy to alleviate symptoms of PVD, and some require surgical intervention. Even with the recent advances in therapeutic development and the promise of agents currently in clinical trials, the questions of who to treat, when treatment should begin, and which agent to use remain uncertain.

Analysis of responses to human synthetic adrenomedullin and calcitonin gene-related peptides in the hindlimb vascular bed of the cat.

Vasodilator responses to human adrenomedullin (hADM), a newly discovered hypotensive peptide, human calcitonin gene-related peptide-alpha (hCGRP-alpha) and hCGRP-beta, which share structural homology with hADM, were compared in the hindlimb vascular bed of the cat under constant flow conditions. Injections of hADM (0.003-1 nmol), hCGRP-alpha, and hCGRP-beta (0.003-0.3 nmol) into the perfusion circuit caused dose-related decreases in hindlimb perfusion pressure. Vasodilator responses to hCGRP-alpha and hCGRP-beta were similar in potency and duration, and the doses of hCGRP-alpha and hCGRP-beta required to reduce hindlimb perfusion pressure 40 mm Hg (ED40 mm Hg) were significantly lower than the ED40 mm Hg for hADM. The duration of the hindlimb vasodilator responses to hCGRP-alpha and hCGRP-beta were significantly longer than the duration of the response to hADM. Amylin, a peptide that shares structural homology with ADM and with CGRP, had no significant effect on hindlimb perfusion pressure when injected in doses up to 1 nmol. Decreases in hindlimb perfusion pressure in response to hADM, hCGRP-alpha, and hCGRP-beta were not altered by L-N5-(1-iminoethyl)-ornithine (L-NIO) in a dose of the nitric oxide synthase inhibitor that decreased the vasodilator response to acetylcholine or by the cyclooxygenase inhibitor, meclofenamate, in a dose that decreased the vasodilator response to archidonic acid. The present data demonstrate that hADM, hCGRP-alpha, and hCGRP-beta have potent, but relatively short-lasting, vasodilator activity, and that vasodilator responses are not dependent on the release of nitric oxide or vasodilator prostaglandins in the hindlimb vascular bed of the cat.

Effect of short-term treatment with a monoclonal antibody to P-selectin on balloon catheter-induced: intimal hyperplasia, re-endothelialization, and attenuation of endothelial-dependent relaxation.

The effects of an anti-P-selectin monoclonal antibody (MAb, PB1.3; Cytel Corporation) on neoendothelialization; neoendothelial function, as evidenced by acetylcholine-induced relaxation (nitric oxide formation); and intimal hyperplasia following embolectomy catheter-induced injury to the rabbit thoracic aorta were investigated. Catheter injury was induced in two groups of New Zealand White rabbits. One group received no treatment, while the second group received short-term treatment with the MAb (i.p., immediately before and 12 h after induction of catheter injury). A third group underwent a sham operation and served as uninjured controls. Following sacrifice at 2 weeks after injury, aortic rings were assessed for degree of intimal hyperplasia, neoendothelial morphology (scanning electron microscopy), and acetylcholine-induced relaxation. Aortic tissue from catheter-injured animals that received treatment exhibited improved neoendothelial morphology, as compared with tissue from untreated but catheterized animals; however, no statistically significant attenuation of the hyperplastic response or improvement in the attenuated neoendothelial-dependent acetylcholine-induced relaxant response that is characteristic of neoendothelium that forms after catheter denudation was observed. These data suggest that short-term attenuation of P-selectin-mediated polymorphonuclear leukocyte (PMN)/endothelium, PMN/platelet interactions, and/or thrombin formation beneficially affects neoendothelialization of the vascular wall following balloon catheter-induced injury.

Differential effects of L-N5-(1-iminoethyl)-ornithine on tone and endothelium-dependent vasodilator responses.

The effects of the nitric oxide (NO) synthesis inhibitor L-N5-(1-iminoethyl)-ornithine (L-NIO) on baseline tone and on responses to the endothelium-dependent vasodilator agents were investigated in the pulmonary vascular bed of the cat under constant-flow conditions. When administered in doses of 1 and 5 mg/kg i.v., L-NIO inhibited pulmonary vasodilator responses to acetylcholine, bradykinin, and substance P but did not alter vasodilator responses to adenosine, pinacidil, or adrenomedullin. L-NIO in doses of 1-10 mg/kg i.v. did not significantly affect baseline lobar arterial pressure, and when administered in doses of 10-30 mg/kg i.v. the inhibitory effect on responses to bradykinin and substance P was not greater than that observed when the lower doses of L-NIO were administered. L-NIO in doses of 5-30 mg/kg i.v. reduced plasma reactive nitrogen intermediate levels. The inhibitory effects of L-NIO were similar to the inhibitory effects of N omega-nitro-L-arginine, N omega-nitro-L-arginine methyl ester, and N omega-nitro-L-arginine benzyl ester. The highest dose of L-NIO studied (30 mg/kg i.v.) caused a significant increased in lobar arterial pressure, and the administration of N omega-nitro-L-arginine methyl ester (100 mg/kg i.v.) caused a significant increase in lobar arterial pressure in animals previously treated with L-NIO (1 mg/kg i.v.). The results of the present study show that the effects of L-NIO on endothelium-dependent vasodilator responses and on baseline tone can be separated and may be interpreted to suggest that basal release of NO does not play an important role in the maintenance of baseline tone in the pulmonary vascular bed of the cat.

Comparative effects of nabumetone, sulindac, and ibuprofen on renal function.

OBJECTIVE: Nonsteroidal antiinflammatory drugs (NSAID) have been associated with hemodynamically mediated acute renal failure. There appear to be differences among NSAID in producing this effect. We compare renal effects of ibuprofen, sulindac, and nabumetone. METHODS: Seventeen women over age 56 receiving hydrochlorothiazide and fosinopril for hypertension who had osteoarthritis requiring NSAID received 3 different NSAID to evaluate potential varying renal effects. In an investigator blinded randomized study, patients received nabumetone, sulindac, or ibuprofen for 1 month with intervening 2 week control periods. After each period renal function was assessed by inulin and para-aminohippurate clearances and urinary prostaglandins were measured. RESULTS: No overall statistical differences among the NSAID were observed. However, there were clinically meaningful differences during ibuprofen therapy: 4 patients developed a clinically significant decrease in renal function; during sulindac therapy one of these also developed a clinically significant decrease in renal function. During nabumetone there were 0 episodes of clinically significant decrease in renal function. Using Gomez equations, glomerular hydrostatic pressure and afferent and efferent arteriolar resistances were estimated. None changed overall during any intervention. However, the 4 patients who developed decreased renal function while taking ibuprofen were analyzed separately. Glomerular hydrostatic pressure decreased 15%; afferent arteriolar resistance increased 85%. These changes were associated with marked decreases in vasodilatory prostaglandins compared to patients receiving ibuprofen who did not develop decreases in renal function. CONCLUSION: There are differences in effect on renal function among NSAID. These can be correlated with specific alterations in suppression of the cyclooxygenase system cascade and related to changes in the hemodynamic control of glomerular filtration.

Analysis of cardiovascular responses to PACAP-27, PACAP-38, and vasoactive intestinal polypeptide.

Responses to pituitary adenylate cyclase polypeptide (PACAP)-27, PACAP-38, and vasoactive intestinal peptide (VIP) were compared in the peripheral and pulmonary vascular beds of the cat and in the isolated perfused neonatal pig heart. Intravenous injections of PACAP-27 and PACAP-38 produced biphasic changes in systemic arterial pressure whereas iv injections of VIP caused only decreases in arterial pressure. When blood flow to the hind limb and mesenteric vascular beds was maintained constant, PACAP-27 and PACAP-38 caused dose-related biphasic changes in perfusion pressure, whereas VIP only decreased perfusion pressure. PACAP-27 was approximately threefold more potent than PACAP-38, and the pressor component of the biphasic response was blocked by alpha-adrenergic antagonists and adrenalectomy. PACAP-27, PACAP-38, and VIP produced decreases in pulmonary vascular resistance, and all three peptides had significant vasodilator activity in the isolated perfused neonatal pig heart. Although all three peptides decreased coronary vascular resistance, only PACAP-27 and PACAP-38 increased left ventricular contractility, with PACAP-27 approaching isoproterenol in potency. The results of these experiments show that PACAP-27, PACAP-38, and VIP have significant effects on vasomotor tone that depend on the vascular bed studied and the contribution of adrenal catecholamines.

Analysis of responses of allicin, a compound from garlic, in the pulmonary vascular bed of the cat and in the rat.

Allicin, diallyl disulfide-oxide, an active ingredient released from garlic is a systemic vasodilator that acts by an unknown mechanism. In the present experiments, pulmonary vascular responses to allicin (0.1-1.0 mg) were studied in the intact-chest anesthetized cat and in the isolated lung of the rat under constant flow conditions. When baseline tone in the pulmonary vascular bed of the cat was raised with U46619 (11 alpha,9 alpha-epoxymethano-9 alpha,11 beta-dideoxyprostaglandin F2 alpha), intralobar injections of allicin produced dose-related decreases in pulmonary arterial pressure without changing left atrial pressure indicating that allicin had significant vasodilator activity in the pulmonary vascular bed when tone was increased experimentally. Allicin also decreased systemic arterial pressure in a dose-related manner. In terms of relative vasodilator activity in the cat, allicin was 100-fold less potent than sodium nitroprusside and many orders of magnitude less potent than isoproterenol. In the cat, vasodilator responses to allicin were unchanged by methylene blue or N omega-nitro-L-arginine methyl ester. Allicin also significantly diminished the pulmonary pressor response to ventilatory hypoxia in the isolated perfused rat lung. These data show that allicin has significant vasodilator activity in the pulmonary vascular bed of the cat and the rat. The present data suggest that pulmonary vasodilator responses to allicin are independent of the synthesis of endothelial-derived relaxing factor or the activation of soluble guanylate cyclase.

Synthetic human adrenomedullin and adrenomedullin 15-52 have potent short-lived vasodilator activity in the hindlimb vascular bed of the cat.

Responses to synthetic human adrenomedullin, a novel hypotensive peptide isolated from human pheochromocytoma cells, and the carboxy terminal 15-52 amino acid fragment of adrenomedullin (ADM15-52) were investigated in the hindlimb vascular bed of the cat under constant flow conditions. Intraarterial injections of the peptides in doses of 0.01-0.3 nmol caused dose-related decreases in hindlimb perfusion pressure. When compared on a nmol basis, adrenomedullin and ADM15-52 were similar to bradykinin in vasodilator potency and were approximately 10 fold less potent than acetylcholine. The half-life of the vasodilator response to adrenomedullin and ADM15-52 ranged from 55 to 80 sec and was greater than the half-life of vasodilator responses to bradykinin in doses of 0.01-0.3 nmol and acetylcholine in doses of 0.01-0.3 nmol. The present data demonstrate that synthetic human adrenomedullin and ADM15-52 have potent but relatively short-lasting vasodilator activity in the hindlimb vascular bed of the cat. These data suggest that amino acid residues 15-52 of adrenomedullin are important for the expression of vasodilator activity in the hindlimb vascular bed of the cat.

Heparin treatment enhances the recovery of neoendothelial acetylcholine-induced vascular relaxation after balloon catheter injury in the rabbit aorta.

BACKGROUND: After catheter injury, the neoendothelium that grows is abnormal in morphology and in acetylcholine-induced generation of endothelium-derived relaxing factor (EDRF). Heparin has been shown to have stimulatory effects on vascular endothelial growth in vitro. Its effect in vivo on neoendothelial cell morphology and metabolism after injury has not been described. We investigated the effect of heparin treatment on the neoendothelium formed after injury. METHODS AND RESULTS: Four groups of New Zealand White rabbits were studied. Group 1 rabbits underwent catheter denudation and were killed 4 weeks after injury without receiving treatment (NO Tx, n = 8). Groups 2 and 3 underwent similar aortic injury, received 2 weeks of treatment with either heparin (n = 7) or low molecular weight heparin (LMWH, n = 5), and were killed at 4 weeks. Group 4 underwent sham operation (SHAM, n = 8). EDRF generation was determined by the relaxation of precontracted aortic rings in an organ bath in response to acetylcholine. The heparin-treated group exhibited a significant improvement in acetylcholine-induced relaxation (27%) versus both LMWH-treated (14%, P = .035) and untreated groups (11%, P = .004), although relaxation was only 50% of that observed in the uninjured control vessels (52%, P = .001). The neoendothelium formed in the heparin-treated group exhibited a more normal histological appearance and was aligned with the direction of blood flow as compared with that observed in the untreated or LMWH-treated groups. CONCLUSIONS: These results demonstrate that in vivo heparin administration enhanced the recovery of EDRF generation and augmented normalization of the morphologic appearance of the neoendothelium.