Patients With Hypercholesterblemia Is Not Due to Increased Extracellular Nitric Oxide

Breakdown by Superoxide Anions Carlos E. Garcia, MD, Crescence M.

Richard 0. Cannon III, MD, At-shed A. Kilcoyne, RN, Carmine Cardillo, MD, Quyyumi, MD, and Julio A. Panza, MD

Patients with hypercholesterolemia have impaired en- dothelium-dependent vasodilation due to decreased nitric oxide activity. The present study aimed to deter- mine whether this form of endothelial dysfunction is related to enhanced extracellular breakdown of nitric oxide by superoxide anions. To this end, the vascular responses to acetylcholine (an endothelium-dependent vasodilator) and sodium nitroprusside (a direct smooth muscle dilator) were studied before and after combined administration of copper-zinc superoxide dismutase (a scavenger of superoxide anions with poor intracellular penetrance; 6,000 U/min) in 20 normal controls (11 men and 9 women, age 50 f 6 years) and in 20 hyper- cholesterolemic patients (10 men and 10 women, age 49 f 9 years). Drugs were inftised into the brachial artery and the response of the forearm vasculature was measured by plethysmography. The vasodilator re- sponse to ace

7 lcholine was significantly blunted in

hypercholestero emit patients compared with normal controls (maximal flow 8.8 f 2 vs 12.7 ?: 3 ml/min/lOO ml, respectively; p ~0.03); however, no difference was

observed in the res + 2 and 9.5 + 3

nse to sodium nitroprusside (9.7 m /min/lOO ml). In normal controls, p”

the infusion of superoxide dismutase did not signifi- cantly modify the response to acetylcholine (maximal flow 12.7 + 3 vs 12.1 f 3 ml/min/ 100 ml before and after superoxide dismutase, respectively). Similarly, in h (rp”

rcholesterolemic patients, the infusion of superox- i e dismutase did not alter the response to acetylcholine (maximal flow 8.8 + 2 and 8.9 f 2 ml/min/lOO ml). A subset of 19 subjects (8 normal and 11 patients) received a 60-minute infusion of superoxide dismutase at 24,000 U/min without alteration in their response to acetylcholine. Su the res

roxide dismutase did not modify

r nse to so r ium nitroprusside in either group.

These indings confirm previous observations of im- paired endothelium-dependent vasodilation in hyper- cholesterolemic cept of increase B

atients, but do not support the con- extracellular destruction of nitric oxide

by su eroxide anions as the mechanism responsible for this a%normality.

(Am J Cardiol 1995;76: 1157-l 161)

E ndothelial dysfunction has been demonstrated in sev- eral cardiovascular conditions associated with or pre-

disposing to atherosclerosis.1d In patients with hyper- cholesterolemia, an impaired endothelium-dependent vasodilator response has been reported in the forearm vasculature.5-8 Previous investigations have shown that this abnormality is largely related to decreased nitric oxide activity in response to endothelial agonists.6 How- ever, the precise mechanism responsible for the defect in the nitric oxide system that leads to endothelial dysfunc- tion remains unknown. Recent investigations in animal models of endothelial dysfunction secondary to dys- lipoproteinemia have suggested that nitric oxide produc- tion in this condition may be normal or even increased,9 and that an enhanced breakdown of nitric oxide by super- oxide anions may contribute to this form of endothelial dysfunction.‘O The purpose of this investigation, there- fore, was to determine whether the abnormal endotheli- urn-dependent vascular relaxation of patients with hyper- cholesterolemia is related to an increased extracellular breakdown of nitric oxide by superoxide anions.

From the Cardiology Branch, National Heart, Lung, and Blood Insti-

tute, National Institutes of Health, Bethesda, Maryland. Manuscript

received May 5, 1995; revised manuscript received and accepted

August 15, 1995.

Add&s for reprints: Julio A. Panza, MD, National Institutes of

Health, Building 10, Room 7B-1.5, Bethesda, Maryland 20892.

METHODS Study group: Twenty patients with hypercholesterol-

emia without any other apparent medical condition were recruited into the study. Each subject was screened by clinical history, physical examination, routine chemical analyses, electrocardiography, and chest radiography. Patients were included in the study if the plasma cho- lesterol level, measured at the time of initial screening after a 1Zhour fasting period, was >250 nig/dl, and they had no history or evidence of present or past hyperten- sion, cardiac disease, diabetes mellitus, peripheral vas- cular disease, coagulopathy, or any other disease predis- posing them to vasculitis or Raynaud’s phenomenon. There were 10 men and 10 women (mean age 49 f 9 years [range 22 to 661). The lipid profile showed: cho- lesterol 293 rt 53 mg/dl, high-density lipoprotein cho- lesterol 48 f 16 mg/dl, triglycerides 234 f 183 mg/dl, and low-density lipoprotein cholesterol 245 f 55 mg/dl. Patients had not taken any cholesterol-lowering agents within the previous 2 months or any antioxidant vitamin supplements in the preceding 6 months.

Twenty normal volunteers (11 men and 9 women), matched with patients for approximate age, were select- ed as a control group. Their mean age was 50 + 6 years (range 39 to 65). Each subject underwent a screening evaluation identical to that described for the hypercho- lesterolemic patients and had no evidence of present or

PREVENTIVE CARDIOLOGY/ENDOTHELILlM-DEPENDENT VASODllATlON IN IHYPERCHOLESTEROLEMIA 1157

past hyperlipidemia (plasma cholesterol was 5200 mg/dl), hypertension, cardiac disease, or any other systemic con- dition. None of these subjects was taking any medica- tion at the time of the study.

All participants gave written informed consent for all procedures. This study was approved by the National Heart, Lung, and Blood Institute Investigational Review Board.

Protocol: All studies were performed in the morning in a quiet room with a temperature of approximately 22°C. Participants were asked to refrain from drinking alcohol or beverages containing caffeine and from smok- ing for 224 hours before studies.

Each study consisted of the infusion of drugs into the brachial artery and the measurement’of the response of the forearm vasculature by means of forearm plethys- mography. All drugs utilized in this study were approved for human use by the Food and Drug Administration in the form of an Investigational New Drug, and were pre- pared by the Pharmaceutical Developmental Service of the National Institutes of Health following specific pro- cedures to ensure accurate bioavailability and sterility of the solutions.

While the participants were supine, a needle was in- serted in the brachial artery of the nondominant arm (left in most patients). This arm was slightly elevated above the level of the right atrium, and a mercury-filled Silas- tic strain gauge was placed on the widest part of the fore- arm.nJ2 The strain gauge was connected to a plethysmo- graph (model EC-4, D.E. Hokanson, Issaquah, Washing- ton)n calibrated to measure the percent change in volume; the plethysmograph in turn was connected to a chart recorder to record the forearm flow measurements. For each measurement, a cuff placed on the upper arm was inflated to 40 mm Hg with a rapid cuff inflator (model E- 10, Hokanson) to occlude venous outflow from the ex- tremity. A wrist cuff was inflated to suprasystolic pres- sures 1 minute before each measurement to exclude hand circulation. l4 Flow measurements were recorded for ap- proximately 7 seconds every 15 seconds; 7 readings were obtained for each mean value.

Basal measurements were obtained after a 3minute infusion of 5% dextrose solution at 1 ml/mm. Forearm flows were then measured after the infusion of sodium nitroprusside and acetylcholine. Sodium nitroprusside was used as an endothelium-independent substance, since its vasodilator effect is largely due to its direct action on smooth muscle cells.15,16 Acetylcholine, in contrast, in- duces vasodilation by stimulating the release of relaxing factors from the vascular endothelium. 17,18

Sodium nitropmsside was infused at 0.8, 1.6, and 3.2 kg/mm and acetylcholine chloride (Sigma Chemical, St. Louis, Missouri) at 7.5, 15, and 30 pg/min (infusion rates, 0.25, 0.5, and 1 ml/min, respectively, for each drug). Each dose was infused for 5 minutes and forearm flow was measured during the last 2 minutes of the infu- sion. A 3Ominute rest period was allowed and another basal measurement was obtained between the infusion of the 2 drugs.

After another 30minute rest period, flow measure- ments were obtained to corroborate return to basal val- ues. Then, bovine copper-zinc superoxide dismutase

(CuZn SOD; DDI Pharmaceuticals, Mountain View, California) was infused at 6,000 U/min (approximately 2 mg/min, infusion rate 1 ml/mm) for 10 minutes to achieve an intravascular concentration of 200 U/ml (approximately 67 p,g/ml), and forearm blood flow was measured during the last 2 minutes of the infusion. The antioxidant activity of CuZn SOD (measured in units) was determined by the cytochrome c reduction inhibi- tion assay.19 Superoxide dismutase is a scavenger of su- peroxide anion20 that reduces the rate of breakdown of nitric oxide.21

Subsequently, cumulative dose-response curves for acetylcholine and sodium nitroprusside were repeated using the same doses, infusion rates, and resting inter- val previously mentioned. The infusion of CuZn SOD was discontinued during the rest period, but reinstated before obtaining the second of these dose-response curves. The sequence of administration of acetylcholine and sodium nitroprusside, both before and after infusion of CuZn SOD, was randomized to avoid any bias relat- ed to the order of drug infusion.

In 8 normal controls and in 11 hypercholesterolemic patients, after measurements during the simultaneous infusion of CuZn SOD (6,000 U/mm) and the highest dose of acetylcholine (30 pg/min) were obtained, the dose of CuZn SOD was raised to 12,000 U/min for 15 minutes, and subsequently to 24,000 U/min for 15 min- utes. The response to acetylcholine (30 pg/min) was again measured during the infusion of CuZn SOD at these higher doses. These additional experiments were performed to determine whether (1) prolonged infusion (up to 60 minutes) and (2) higher doses (12,000 and 24,000 U/mm) of CuZn SOD would modify acetylcho- line-induced vasodilation.

During the studies, the participants did not know which drug was being infused. Blood pressure was re- corded from the intraarterial catheter before each mea- surement. Forearm vascular resistance was calculated as mean arterial pressure divided by the forearm blood flow.

Statistical analysis: Differences between 2 means were compared by paired or unpaired Student’s t test, as ap- propriate. The responses to sodium nitroprusside and acetylcholine were compared by analysis of variance for repeated measures using a multiple linear regression mo- del that included dummy variables to correct for be- tween-subjects variability. 22 All calculated p values are 2-tailed. All p values co.05 were considered to indicate significance. All group data are reported as mean f SD unless otherwise indicated.

RESULTS Vascular responses to acetylcholine and sodium nitro-

prusside: Similar to the findings of previous studies,5-8,23 the increase in blood flow and decrease in vascular resis- tance with acetylcholine were significantly reduced in hypercholesterolemic patients compared with normal controls (Figure 1). At the highest dose (30 pg/min), forearm blood flow was 12.7 f 3 and 8.8 + 2 ml/mm/100 ml in controls and patients, respectively (p ~0.05).

However, no significant differences were found be- tween the 2 groups in forearm blood flow and vascular resistance response to sodium nitroprusside. At the high-

1158 THE AMERICAN JOURNAL OF CARDIOLOGY@ VOL. 76 DECEMBER 1, 1995

est dose (3;2 p,g/min), ,forearm blood flow was 9.,5 f 3 and 9.7 & 2 ml/mitr/lOO ml in controls and hyper- cholesterolemic patients, respectively.

Effect of bovine copper-zinc superoxide dismutase on basal bldbd flow and vascular resistance: Basal forearm blood flow, measured at the beginning of the study, was similar in hypercholesterolemic patients and normal con- trols (2.4 f 0.7 vs 2.4 f 0.6 ml/min/lOO ml, respective- ly). Similarly basal vascular resistance was not signifi- cantly different between patients and controls (38.6 + 12 vs 36.1 f 9 mm Hg/ml-l l mini l 100 n-&l, respective- ly).

The infusion of CuZn SOD did not produce any sig- nificant change in blood flow or vascular resistance in either group. In hypercholesterolemic patients, blood flow was 2.3 f 0.7 and 2.3 + 0.7 ml/min/lOO ml (p = NS), and vascular resistance 38 +: 11 and 38 of: 10 mm Hg/n&l l m&r l 100 ml-’ (p = NS) immediately before and after CuZn SOD infusion, respectively. In normal controls, blood flow was 2.8 f 0.7 and 2.7 -I 0.7 ml/min/100 ml (p = NS), and vascular resistance was 31 f 8 and 33 + 12 mm Hg/n&l l min-’ l 100 ml-’ (p = NS) immediately before and after CuZn SOD infusion, respectively.

15 r C+O Normal Controls

W Hypercholesterolemic Patients

50-

8,

it a-

Ei 30- Ytj

Qi 20- >E

a EI g: lo- L-- 2

0 ’ I I I

Baseline 7.5 15 30

Acetylcholine (Irglmin)

FIGURE 1. Forearm blood flow and vascular resistance normal controls [open circles)

(closed circles). In this to comparison of the 2

for repeated measures. Values are mean f SEM.

Effect of bovine copper-zinc superoxide dismutase on vascular responses to acetyicholine and sodium nitro- prusside: In normal controls, the vasodilator response to acetylcholine was not significantly modified after infu- sion of CuZn SOD (Figure 2). At the highest dose of acetylcholine (30 p&mm), blood flow was 13 + 3 and 12 + 3 ml/min/100 ml before and after infusion of CuZn SOD, respectively (p = NS). The infusion of CuZn SOD did not modify the vasodilator response to sodium nitro- prusside in normal controls (maximal blood flow 9.5 f 3 vs 9.2 f 3 ml/min/lOO ml before and after CuZn SOD infusion, respectively; p = NS).

Likewise, in hypercholesterolemic patients, the re- sponse to acetylcholine was not significantly altered by the infusion of CuZn SOD (Figure 3). At the maximal dose of acetylcholine, blood flow was 8.8 f 2 and 8.9 f 2 ml/mm/100 ml before and after infusion of CuZn SOD, respectively (p = NS). Similarly, CuZn SOD did not pro- duce any significant difference in the response to sodi- um nitroprusside in hypercholesterolemic patients (max- imal blood flow 9.7 + 2 vs 10.6 f 4 ml/mm/100 ml before and after CuZn SOD infusion, respectively).

Prolonged infusion (60 minutes) of CuZn SOD at increasing doses (12,000 and 24,000 U/mm) did not pro- duce any significant change on endothelium-dependent vasodilation in either normal controls or hypercholes-

l5 r O-0 Before CuZn SOD

M After CuZn SOD

“01

Ofi Baseline 7.5 15 30

Acetylcholine @g/min)

FIGURE 2. Forearm blood flow and vascular resistance responses to acetylcholine in 20 normal controls before (opm circ/esJ and after (closed circles] infusion of bovine copper-zinc superoxide dismutase (CuZn SOD). Values are mean I SEM.

PREVENTIVE CARDIOLOGY/ENDOTHELIUM-DEPENDENT VASODllATlON IN HYPERCHOLESTEROLEMIA 1159

terolemic patients. In the 8 normal controls, at the high- est dose of acetylcholine (30 ~g/min), blood flow was 13 + 3 and 13 + 4 rnl/rnin/lOO ml before and after infu- sion of CuZn SOD (24,000 U/min), respectively (p = NS). In the 11 hypercholesterolemic patients, at the same dose of acetylcholine, blood flow was 8.5 + 4 and 9.6 f 5 ml/min/lOO ml before and after infusion of CuZn SOD (24,000 U/n-k-r), respectively (p = NS).

DISCUSSION The present study was designed to investigate

whether an enhanced breakdown of nitric oxide by su- peroxide anions is responsible for the impaired endothe- lium-dependent vasodilation previously demonstrated in patients with hypercholesteremia. To this end, we uti- lized CuZn SOD, a scavenger of superoxide anion with poor intracellular penetrance.20 We hypothesized that if extracellular nitric oxide destruction by superoxide anion was increased, then superoxide dismutase, by partially or totally removing superoxide anions, would allow a greater delivery of nitric oxide to the smooth muscle cells,21 and would therefore improve the response to acetylcholine in hypercholesterolemic patients.

Our results continned previous studies showing that patients with hypercholesteremia have impaired dilator re-

‘* r M Before Cdn SOD

H After CuZn SOD

0, 50r

?! Qi 20- >E a Es gE lo-

p--

0 ’ I I I

Baseline 7.5 15 30 Acetylcholine (pg/min)

FIGURE 3. Forearm blood flow and vascular resistance res onses to ace

bet (pe

choline in 20 hypercholesterolemic patients ore o r n circ es] and after /closed circles] infusion of bovine

copper-zinc superoxide dismutase (CuZn SOD). Values are mean 2 SEM.

sponse to acetylcholine with preserved response to the en- dothelium-independent agent sodium nitroprusside.5-8*23 These findings are compatible with abnormal endothe- lial dilator function. The infusion of CuZn SOD, how- ever, did not have any demonstrable effect on acetyl- choline-induced vasodilation. Therefore, our findings do not support the concept of increased extracellular de- struction of nitric oxide by superoxide anions as the mechanism responsible for the endothelial dysfunction of hypercholesterolemic patients.

Taken in conjunction with the findings of previous studies related to the abnormality of the nitric oxide sys- tem in these patients, 6~8,24 the present study may con- tribute to our understanding of endothelial dysfunction- in this condition. Thus, if the extracellular destruction of nitric oxide by superoxide anions is not increased, oth- er mechanisms must be operative. For example, our find- ings are consistent with the possibility of a specific defect of certain signal transduction pathways leading to nitric oxide production. This is suggested by the demonstra- tion that, while endothelium-dependent vasodilation to acetylcholine is impaired in hypercholesterolemic pa- tients, bradykinin-mediated responses are preserved.24 It has also been proposed that nitric oxide activity might be reduced due to the loss of formation of a more potent molecule that would normally be necessary for the final effect of nitric oxide on guanylate cyclase within the smooth muscle cells.9 Finally, it is possible that hyper- cholesterolemia is associated with an increased destruc- tion of nitric oxide by other forms of oxygen-free radi- cals different from superoxide anions and therefore not affected by the infusion of CuZn SOD.

That an increased destruction of nitric oxide may be responsible for the endotbelial dysfunction associated with dyslipidemia is supported by previous studies in animal models of this condition. For example, Minor et al9 showed that the production of nitrogen oxide may actually be increased in atherosclerotic aortic rings. In line with this observation, Ohara et allo showed that superoxide anion production is increased in endothelial cells from hypercholesterolemic vessels, and that the administration of oxypurinol (an inhibitor of superoxide anion production by xanthine oxidase) to isolated hyper- cholesterolemic vessels improved the relaxation to acetyl- choline, an effect not observed in normal vessels. In addi- tion, Mtigge et al25 showed that chronic administration of polyethylene-glycol superoxide dismutase improved the endothelium-dependent response to acetylcholine without affecting the endothelium-independent response to sodium nitroprusside. Thus, several lines of evidence obtained in animal studies strongly suggest that an in- creased breakdown of nitric oxide may play a role in the endothelial dysfunction associated with hypercholester- olemia.

In this regard, the results of the present study do not allow us to conclusively rule out the possibility of in- creased nitric oxide breakdown in hypercholesterolemia. For example, an increased destruction of nitric oxide by endogenous superoxide anions may occur intracellular- ly; however, because CuZn SOD has poor intracellular penetrance due to its negative charge,20 it is possible that it may not cross the cell membrane in sufficient amounts

1160 THE AMERICAN JOURNAL OF CARDIOLOGY@ VOL. 76 DECEMBER 1, 1995

to reduce the intracellular breakdown of nitric oxide. It is also possible that other types of reactive oxygen inter- mediates, such as hydroxyl radical or hydrogen peroxide, importantly contribute to nitric oxide destruction. In fact, the combination of superoxide dismutase with superox- ide anion leads to the formation of hydrogen peroxide. However, it is unlikely that hydrogen peroxide impor- tantly participates in nitric oxide destruction because catalase was shown not to have a significant additive beneficial effect on endothelium-dependent vasomotion beyond that observed with superoxide dismutase. 21,26

In conclusion, the present study confirms previous observations of endothelial dysfunction in hypercholes- terolemic patients and expands those findings by demon- strating that acute administration of CuZn SOD does not improve acetylcholine-induced vasodilation. These re- sults, therefore, do not support the hypothesis of in- creased extracellular destruction of nitric oxide by super- oxide anions as the mechanism responsible for this endothelial abnormality.

1. Jayakody L, Senaratne M, Thomson A, Kappagoda T. Endothelium-dependent relaxation in experimental atherosclerosis in the rabbit. Circ Res 1987;60:251-264. 2. Freiman PC, Mitchell GG, Heistad DD, Armstrong ML, Harrison DG. Athero- sclerosis impairs endothelium-dependent vascular relaxation to acetylcholine and thrombin in primates. Circ Res 1986;58:783-789. 3. Andrew HE, Bruckdorfer KR, Dunn RC, Jacobs M. Low-density lipoproteins inhibit endothelium-dependent relaxation in rabbit aorta. Nature 1987;327:237-239. 4. Cohen RA, Zitnay KM, Haudenschild CC, Cunningham LD. Loss of selective endothelial cell vasoactive functions caused by hypercholesterolemia in pig coro- nary arteries. Circ Res 1988;63:903-910. 5. Creager MA, Cooke JP, Mend&on ME, Gallagher SJ, Coleman SM, Loscalzo J, Dzau V.I. Impaired vasodilation of forearm resistance vessels in hypercholes- terolemic humans. J Clin Invest 1990;86:228-234. 6. Casino PR, Kilcoyne CM, Quyyuni AA, Hoeg JM, Panza JA. Role of nitric oxide in the endothelinm-dependent vasodilation of hypercholesterolemic patients. Circulation 1993;88:2541-2547. 7. Chowienczyk PJ, Watts GF, Cockcroft JR, Ritter JM. Impaired endothelium- dependent vasodilation of forearm resistance vessels in hypercholesterolaemia. Lmcet 199X340:1430-1432.

8. Casino PR, Kilcoyne CM, Cannon RO, Quyyumi AA, Panza JA. Impaired endothelium-dependent vascular relaxation in patients with hypercholesteroleinis extends beyond the muscarinic receptor. Am J Cardiol 1995;75:4&44. 9. Minor RL Jr, Myers PR, Guerra R Jr; Bates JN, Harrison DG. Diet-induced ath- erosclerosis increases the release of nitrogen oxides from rabbit aorta. J Clin Invest 1990;86:2109-2116. 10. Ohara Y, Peterson TE, Ham DG. Hypercholesterolemia increases endothelial superoxide anion production. .I Clin Invest 1993;91:254&255 1. 11. Whitney R-l. Measurement of changes in human limb volume by means of a mercury-in-rubber strain gauge. J Physiol 1948; 109:5P-6P. 12. Greenfield ADM, Whitney RJ, Mowbray JF. Methods for the investigation of peripheral blood flow. Br Med Bull 1963;19:101-109. 13. Hokanson DE, Sumner DS, Strandness DE. An electrically calibrated plethys- mograph for direct measurement of limb blood flow. IEEE Tram BiomedEng 1975; 22~25-29. 14. Kerslake DM. The effect of the application of an arterial occlusion cuff to the wrist on the blood flow in the human forearm. .I Physiol 1949;108:451457. 15. Bohme E, Graf H, Schultz G. Effects of sodium nitroprusside and other smooth muscle relaxants on cyclic GMP-formation in smooth muscle and platelets. Adv Cycl Nucl Res 1978;9:131-143. 16. Kukovetz WR, Holtzmann S, Wurm A, Poch G. Evidence for cyclic GMP- mediated relaxant effects of nitro-compounds in coronary smooth muscle. Naunyn Schmiedebergs Arch Pharmacol 1979;310:129-138. 17. Furchgott RF, Zawadzki JV. The obligatory role of the endothelial cells in the relaxation of arterial smooth muscle bv acetvlcholine. Nature 1980:288:373-376. 18. Furchgott RF. Role of endotheli& in &ponses of vascular smooth muscle. Circ Res 1983;53:557-573. 19. McCord JM, Fridovich I. Superoxide dismutase: an enzymatic function for ery- throcuprein (hemocuprein). J Bio2 Chem 1969;244:6049-6055. 20. Oman BA, Flares SC, McCord JM. Superoxide dismutase: pharmacological developments and applications. Adv Pharmacol 1992;23:109-161. 21. Gryglewski RJ, Palmer RMJ, Moncada S. Superoxide anion is involved in the breakdown of endothelium-derived vascular relaxing factor. Nature 1986;320: 454456. 22. Glantz SA, Slinker BK. Repeated measures. In: Primer of Applied Regression and Analysis of Variance. New York: McGraw-Hill, 1990:381463. 23. Casino PR, Kilcoyne CM, Quyyumi AA, Hoeg JM, Panza JA. Investigation of decreased availability of nitric oxide precursor as the mechanism responsible for impaired endothelium-dependent vasodilation in hypercholesterolemic patients. J Am Coil Cardiol 1994;23:84+850. 24. Gilligan DM, Guetta V, Panza JA, Garcia CE, Quyyumi AA, Cannon RO. Selective loss of microvascular endothelial function in human hypercholesterol- emia. Circulation 1994;90:3541. 25. Miigge A, Elwell JH, Peterson TE, Hofmayer TG, Heistad DD, Harrison DG. Chronic treatment with polyethylene-glycolated superoxide dismutase partially restores endothelium-dependent vascular relaxations in cholesterol-fed rabbits. Circ Res 1991;69:1293-1300. 26. Miigge A, Elwell JH, Peterson TE, Harrison DG. Release of intact endotheli- urn-dependent relaxing factor depends on endothelial superoxide dismutase activi- ty. Am .I Physiol 1991;26O:C219-c225.

PREVENTIVE CARDIOLOGY/ENDOTHELIUM-DEPENDENT VASODllATlON IN HYPERCHOLESTEROLEMIA 1161