There is considerable evidence that endothelium-dependent relaxation is impaired in experimental diabetic animals (1-3). This defect has recently been confirmed in both type I (4,5) and type II (6) diabetic patients. The factor or factors that contribute to the defect in patients with diabetes is unclear, but information derived in experimental models has suggested several possibilities including (a) concurrent release of an endothelium-derived constricting factor arising from the cyclooxygenase pathway (7,8); (b) increase in protein kinase C (9); (c) inappropriate use of substrate for nitric oxide synthesis because of arginine deficiency (10,11); (d) abnormal nitric oxide synthase activity because of inadequate tetrahydrobiopterin cofactor (12); (e) quenching of nitric oxide by advanced glycosylation end products (13); or (f) increased quenching of nitric oxide derived from the endothelium by interaction with increased concentrations of oxygen-derived free radicals particularly superoxide anion radicals (14-18).
In each case, these studies revealed concurrent factors that might compromise endothelium-dependent relaxation in the acute setting. In contrast, only a limited number of studies have focused on identifying the antecedent factors or pathways in vivo that contribute to the underlying pathology of diabetes-induced endothelial dysfunction. Of these, most have focused on the use of aldose reductase inhibitors (19,20).
Previously we observed that the antioxidant enzyme profile in a variety of diabetic blood vessels indicated significant increases in tissue catalase activity, which was rapidly reversed by pancreatic islet transplantation (21). This suggests that diabetic vasculature is likely to be chronically exposed to elevated levels of hydrogen peroxide and thereby to increased oxidative stress. Thus it would appear that pharmacologic interventions targeted to interrupt oxygen-free radicals and their action might prevent diabetes-induced endothelial dysfunction; however, direct studies in vivo are necessary to verify this hypothesis. In this study, we evaluated the efficacy of chronic treatment with dimethylthiourea (DMTU), a known hydroxyl radical scavenger (22), in preventing the endothelial dysfunction caused by diabetes.
Male, Lewis strain rats (aged 10-11 weeks) were anesthetized with intraperitoneal injections of sodium pentobarbital (60 mg/kg). Diabetes was induced by penile-vein injection of streptozotocin (55 mg/kg in pH 4.5 citrate buffer). Blood glucose was measured in diabetic animals at 3 days and 1 week after streptozotocin to verify hyperglycemia. Diabetic (n = 9) and age-matched control rats (n = 11) were housed for 8 weeks before experimentation. A randomly selected group of diabetic animals (n = 8) received daily intraperitoneal injections of 50 mg/kg DMTU (in sterile saline) beginning 72 h after streptozotocin administration and throughout the duration of diabetes. This dose was chosen based on our previous observation of its efficacy in prolonging heterotopic cardiac allograft survival in the rat (23). At the conclusion of the 8-week period, blood glucose was determined from a drop of tail blood by using an ExacTech glucometer and test strips (Medisense, Inc., Cambridge, MA, U.S.A.). Serum insulin was determined by using an 125I radioimmunoassay kit (Diagnostic Products Corporation, Los Angeles, CA, U.S.A.). To assess chronic glycemic control in each group, blood also was collected for analysis of total glycosylated hemoglobin by using a commercial kit (Sigma Diagnostics, St. Louis, MO, U.S.A.) and measured on a spectrophotometer.
On the day of experimentation, rats were anesthetized with an intraperitoneal injection of 65 mg/kg sodium pentobarbital. Thoracic and abdominal aortas were carefully isolated, removed from open-chest animals, and placed in 4°C Krebs bicarbonate buffer. The aortic segments were carefully cleaned of fat and loose connective tissue. The thoracic aortic segments were sectioned into 3-mm (in length) rings. Extreme care was always taken to avoid stretching and contact with the luminal surface of the endothelium to avoid inadvertent damage during isolation.
The abdominal aortas were processed in a similar manner for the determination of tissue catalase activity to determine the long-term effect of diabetes on oxidative stress status in the control, untreated diabetic, and DMTU-treated diabetic groups. We previously showed that catalase activity is nearly doubled, whereas superoxide dismutase activity is unaltered, and glutathione peroxidase is marginally (but not significantly) reduced in aorta, iliac artery, and femoral artery of diabetic rats (21). Thus we have used catalase measurements only as evidence of chronic peroxidative stress in these animals. Catalase was determined by the method of Aebi (24). Because the measure of catalase is known to be nonlinear, results were appropriately expressed in units of K per mg where K is the first-order rate constant calculated by the equation K = (2.3/t) × (log A1/A2), and A1 and A2 are the optical density values at the beginning and end of time interval t.
Isolated vascular-ring experiments
Aortic rings were suspended between parallel triangular hooks in 10-ml tissue baths maintained at 37°C. The medium consisted of a modified Krebs-Henseleit bicarbonate buffer containing (in millimoles) 118 NaCl, 4.7 KCL, 2.5 CaCl2, 1.2 MgSO4, 11 glucose, and 24 NaHCO3, which was maintained at pH 7.4 by oxygenation with 95% O2:5% CO2. The buffer also contained 0.8 μM imipramine and 0.9 μM propranolol to control for any potential differences in diabetes-induces changes in catecholamine uptake or β-adrenergic activity.
Rings were stretched to an optimal resting tension of 2.0 g for both control and diabetic blood vessels. Changes in isometric tension were recorded of a Gould TA6000 recorder by Radnoti force-displacement transducers and amplifiers.
After 90 min of equilibration, rings were precontracted with increasing concentrations of norepinephrine (1 nM to 30 μM). Stock solutions of norepinephrine contained ascorbate to prevent autooxidation (final concentration, 20 nM). The pD2 [-log 50% effective concentration (EC50)] for response of rings to norepinephrine was determined for each ring experiment. After contraction, each ring was serially washed and reequilibrated to baseline. Rings were then contracted with a submaximal concentration of norepinephrine (usually 1 μM). The concentration of agonist was varied in some instances so that equieffective concentrations of constrictor were used based on the tensions derived in the initial norepinephrine concentration-response measurements.
At the plateau of contraction, relaxation responses to cumulative concentrations of vasodilators were measured by using acetylcholine (endothelium dependent), calcium ionophore A23187 (endothelium dependent), or nitroglycerin (endothelium independent). We previously demonstrated that relaxation to acetylcholine and A23187 in control and diabetic rat aortas are caused by nitric oxide-dependent relaxation based on the complete blockade of relaxation by arginine analogs (10). Only one type of vasodilator was used in each ring preparation.
All chemicals were of the highest purity available. Most chemicals were obtained from Sigma Chemical Co. and dissolved in deionized water and diluted in buffer. Statistical analysis was performed by using analysis of variance followed by the Fisher Projected Least Squares Difference test for multiple group mean comparisons, unpaired t test when comparing two group means, or paired t test when comparing two group means in a repeated format. A value of p < 0.05 was set as indicating statistical significance. All data are expressed as the mean ± SEM.
Body weight and blood analysis
Body weight was significantly increased (p < 0.01) at the end of the study compared with starting weight in control animals (281 ± 5 and 366 ± 5 g for initial and final weights, respectively). In contrast, no weight gain was observed in untreated diabetic animals (282 ± 5 and 252 ± 8 g for initial and final weights, respectively) and in DMTU-treated diabetic animals (281 ± 2 and 261 ± 7 g for initial and final weights, respectively).
Before injection of streptozotocin, blood glucose was not different between control (65 ± 5 mg/dl) and diabetic rats (74 ± 2 mg/dl). At 48 h after streptozotocin and before drug administration, blood glucose increased (p < 0.001) to 246 ± 20 and 252 ± 14 mg/dl in the untreated diabetic and DMTU-treated diabetic groups compared with control rats (56 ± 8 mg/dl). At the end of the study, blood glucose levels were increased in untreated diabetic and DMTU-treated diabetic rats compared with aged-matched, control rats (Fig. 1, top). Serum insulin was decreased in untreated diabetic animals compared with control animals and also was not altered by DMTU treatment (Fig. 1, middle). Total glycosylated hemoglobin was significantly increased in untreated diabetic animals compared with control animals and was not altered by DMTU treatment (Fig. 1, bottom).
Tissue catalase activity
Catalase was significantly (p < 0.01) increased in diabetic aortas (0.324 ± 0.023 K/mg) compared with control aortas (0.192 ± 0.015 K/mg). Aortic catalase (0.334 ± 0.045 K/mg) was unaltered in DMTU-treated diabetic animals.
Concentration-response reactions of aortic rings to norepinephrine were performed. These studies revealed no significant difference in the maximum tension or pD2 values between control rings (2.52 ± 0.19 and 6.5 ± 0.1 g/mm2, respectively) and diabetic rings (2.22 ± 0.16 and 6.3 ± 0.1 g/mm2, respectively). In contrast, maximum tension and pA2 values were reduced in rings from the DMTU-treated animals (1.89 ± 0.15 and 6.2 ± 0.1 g/mm2, respectively; p < 0.01 each) compared with control rings but not compared with rings from untreated diabetic animals. For the vasodilator studies, rings were contracted with equieffective, submaximal concentrations of norepinephrine, which gave 66 ± 2% (control), 65 ± 4% (diabetic), and 60 ± 3% (DMTU-treated diabetic) of maximum tension development.
Acetylcholine produced concentration-dependent relaxations in control and diabetic aortic rings (Fig. 2, top). Acetylcholine-stimulated relaxation was reduced in diabetic aortas compared with those of age-matched control animals. In contrast, relaxation to the endothelium-dependent vasodilator A213187 was unaltered by diabetes (Fig. 2, middle). Similarly, endothelium-independent relaxation to nitroglycerin was not altered by diabetes (Fig. 2, bottom). Long-term treatment with DMTU prevented the diabetes-induced impairment in relaxation to acetylcholine although it had no effect on the relaxation produced by either A23187 or nitroglycerin.
Diabetic vasculature appears to be exposed to increased oxidative stress, as evidenced by increased concentrations of plasma lipid peroxides (25,26) and compensatory increases in tissue catalase activity in aorta, iliac artery, and femoral artery of diabetic rats (21). These data suggest that blood vessels in this condition are chronically exposed to increased concentrations of hydrogen peroxide. It is possible that exposure to this reactive oxygen species might lead to endothelial dysfunction vascular; however, this has never been directly tested.
It is possible to demonstrate in vitro a diminution in endothelium-dependent relaxation to acetylcholine in normal blood vessels exposed to free radical-generating systems. Indeed, we previously showed that xanthine plus xanthine oxidase (to generate superoxide anion radicals and hydrogen peroxide) and activated polymorphonuclear leukocytes both impair such responses in the short term (27,28). We subsequently determined that it is likely that the hydroxyl radical is the molecular oxygen species responsible for this injury (29).
The molecular oxygen species responsible for the development of endothelial dysfunction in diabetes is not yet known. That oxygen radicals may be important in the origin of this defect was recently shown in a single study in which long-term administration of the antioxidant vitamin E prevented the impaired relaxation response to acetylcholine in diabetic rat aortas without altering endothelium-independent relaxation to nitroglycerin (30). Although this strongly supports a role for reactive oxygen, no study has delineated which molecular oxygen species is the nascent event that leads to this dysfunction.
In our study, we concluded that the hydroxyl radical is the reactive molecular oxygen species responsible for this injury. This would be consistent with recent evidence for increased hydroxyl radical formation in the blood of diabetic rats (31). It appears from this study that DMTU improved relaxation to acetylcholine but not relaxation to either nitroglycerin or A23187. This suggests that the beneficial effect is not the result of a generalized increase in the responsiveness of diabetic vascular smooth muscle to all vasodilators. Rather it suggests a selective effect of DMTU in protecting the vascular endothelium and, in particular, receptor-dependent, endothelium-dependent relaxation.
The salient action of DMTU cannot be ascribed to alterations in the diabetogenic status or glycemic control because glucose was elevated to similar levels in untreated and treated groups before DMTU administration. Furthermore, blood glucose and serum insulin levels were comparable at the end of the study. The observation that total glycosylated hemoglobin levels were comparable suggests that long-term glycemic control also was not modified by DMTU treatment.
In addition, our determinations of tissue catalase levels suggest that both untreated and DMTU-treated diabetic animals continued to be exposed to the same degree of peroxidative stress (i.e., hydrogen peroxide production). Thus the precursor for hydroxyl radical formation that could be scavenged by DMTU was similar between the untreated and treated diabetic groups. This is important because it suggests that an action of reactive oxygen occurring distal to hydrogen peroxide formation is involved in the cause of diabetes-induced endothelial dysfunction.
The specific injury to receptor-mediated, endothelium-dependent relaxation mediated by hydroxyl radicals is likely caused, in part, by hyperglycemia itself. In this regard, we have shown that incubation of cultured endothelial cells with elevated glucose concentrations leads to a decrease in the calcium signal-transduction pathway in response to bradykinin, whereas responses to the receptor-independent agonist, ionomycin, are unaltered (32). We observed that this defect in glucose-exposed endothelial cells also was mediated by hydroxyl radicals because DMTU completely prevented this defect.
In conclusion, we demonstrated for the first time that hydroxyl radicals are the likely reactive oxygen species that leads to diabetes-induced endothelial dysfunction. Thus therapies designed to interfere with hydroxyl radical formation may be useful interventions to prevent the vascular complications associated with this disease.
Acknowledgment: This work was supported by a grant #HL47072 from the National Institutes of Health, Heart and Lung Institute.
1. Pieper GM, Gross GJ. Endothelial dysfunction in diabetes. In: Rubanyi GM, ed. Cardiovascular significance of endothelium-derived vasoactive factors
. Mount Kisco, NY: Futura, 1991;223-49.
2. Cohen RA. Dysfunction of vascular endothelium in diabetes mellitus. Circulation
3. Kamata K, Miyata N, Abiru T, Kasuya Y. Functional changes in vascular smooth muscle and endothelium of arteries during diabetes mellitus. Life Sci
4. Johnstone MT, Creager SJ, Scales KM, Cusco JA, Lee BK, Creger MA. Impaired endothelium-dependent vasodilation in patients with insulin-dependent diabetes mellitus. Circulation
5. McNally PG, Watt PAC, Rimmer T, Burden AC, Hearnshaw JR, Thurston H. Impaired contraction and endothelium-dependent relaxation in isolated resistance vessels from patients with endothelium-dependent diabetes mellitus. Clin Sci
6. McVeigh GE, Brennan GM, Johnston GD, et al. Impaired endothelium-dependent and independent vasodilation in patients with type 2 (non-insulin-dependent) diabetes mellitus. Diabetologia
7. Tesfamariam B, Jakubowski JA, Cohen RA. Contraction of diabetic rabbit aorta caused by endothelium-derived PGH2-TxA2. Am J Physiol
1989;57(Heart Circ Physiol
8. Mayhan WG. Impairment of endothelium-dependent dilatation of cerebral arterioles during diabetes mellitus. Am J Physiol
1989;256(Heart Circ Physiol
9. Pelligrino DA, Koenig HM, Wang Q, Albrecht RF. Protein kinase C suppresses receptor-mediated pial arteriolar relaxation in the diabetic rat. Neuroreport
10. Pieper GM, Peltier BA. Amelioration by L-arginine of a dysfunctional arginine/nitric oxide pathway in diabetic endothelium. J Cardiovasc Pharmacol
11. Pieper GM, Jordan M, Adams MB, Roza AM. Syngeneic pancreatic islet transplantation reverses endothelial dysfunction in experimental diabetes. Diabetes
12. Pieper GM. Acute amelioration of diabetic endothelial dysfunction with a derivative of the nitric oxide synthase cofactor, tetrahydrobiopterin. J Cardiovasc Pharmacol
13. Bucala R, Tracey KJ, Cerami A. Advanced glycosylation products quench nitric oxide and mediate defective endothelium-dependent vasodilatation in experimental diabetes. J Clin Invest
14. Langenstroer P, Pieper GM. Regulation of spontaneous EDRF release in diabetic rat aorta by oxygen free radicals. Am J Physiol
1992;263(Heart Circ Physiol
15. Pieper GM, Mei DA, Langenstroer P, O'Rourke ST. Bioassay of endothelium-derived relaxing factor in diabetic rat aorta. Am J Physiol
1992;263(Heart Circ Physiol
16. Diederich D, Skopec J, Diederich A, Dai FX. Endothelial dysfunction in mesenteric resistance arteries of diabetic rats: role of free radicals. Am J Physiol
1994;266(Heart Circ Physiol
17. Pieper GM, Moore-Hilton G, Roza AM. Evaluation of the mechanism of endothelial dysfunction in the genetically-diabetic BB rat. Life Sci
18. Tesfamariam B, Cohen RA. Free radicals mediate endothelial cell dysfunction caused by elevated glucose. Am J Physiol
1992;263(Heart Circ Physiol
19. Cameron NE, Cotter MA. Impaired contraction and relaxation in aorta from streptozotocin-diabetic rats: role of polyol pathway. Diabetologia
20. Tesfamariam B, Palacino JJ, Weisbrod RM, Cohen RA. Aldose reductase inhibition restores endothelial cell function in diabetic rabbit aorta. J Cardiovasc Pharmacol
21. Pieper GM, Jordan M, Dondlinger LA, Adams MB, Roza AM. Peroxidative stress in diabetic blood vessels: reversal by pancreatic islet transplantation. Diabetes
22. Wasil M, Halliweel B, Grootevel M, Moorhouse CP, Hutchison DCS, Baum H. The specificity of thiourea, dimethylthiourea
and dimethylsulfoxide as scavengers of hydroxyl radical. Biochem J
23. Roza AM, Pieper G, Moore-Hilton G, Johnson CP, Adams MB. Free radicals in pancreatic and cardiac rejection. Transplant Proc
24. Aebi H. Catalase in vitro. Methods Enzymol
25. Sato Y, Hotta N, Sakomoto N, Mutsuoka S, Ohishi N, Yagi K. Lipid peroxides in plasma of diabetic patients. Biochem Med
26. Gallou G, Ruelland A, Legras B, Maugendre D, Allannic A, Clourec L. Plasma malondialdehyde in type 1 and type 2 diabetic patients. Clin Chim Acta
27. Pieper GM, Gross GJ. Selective impairment of endothelium-dependent relaxation by oxygen-derived free radicals: distinction between receptor versus nonreceptor mediators. Blood Vessels
28. Pieper GM, Gross GJ. Priming by platelet-activating factor of neutrophil-induced impairment of endothelium-dependent relaxation. J Vasc Med Biol
29. Pieper GM, Langenstroer P, Gross GJ. Hydroxyl radicals mediate injury to endothelium-dependent relaxation in diabetic rat. Mol Cell Biochem
30. Keegan A, Walbank H, Cotter MA, Cameron NE. Chronic vitamin E treatment prevents defective endothelium-dependent relaxation in diabetic rat aorta. Diabetologia
31. Ohkuwa T, Sato Y, Naoi M. Hydroxyl radical formation in diabetic rats induced by streptozotocin. Life Sci
32. Pieper GM, Dondlinger L. Mechanism for elevated glucose-induced alteration in intracellular calcium signals to bradykinin in cultured endothelial cells [Abstract]. Circulation