Effects of β-blockers on the concentration and oxidizability of plasma lipids

Transcription

1 Clinical Science (1998) 94, (Printed in Great Britain) 573 Effects of β-blockers on the concentration and oxidizability of plasma lipids Simon B. DIMMITT, Peta D. WILLIAMS, Kevin D. CROFT and Lawrence J. BEILIN Western Australian Heart Research Institute, Department of Medicine, University of Western Australia, Royal Perth Hospital, Wellington Street, GPO Box X2213, Perth 6001, Western Australia 1. β-blockers improve morbidity and mortality after myocardial infarction, probably by several mechanisms. We investigated potentially relevant effects of β-blockers in vivo and in vitro on plasma lipid oxidizability. Fortytwo healthy men were randomized to receive placebo (13), metoprolol (14) or propranolol (15). 2. At 4 weeks, the effects on heart rate, blood pressure and lipids appeared similar and subjects taking a β- blocker were combined. Compared with placebo, those on a β-blocker gained 0 5 kg in weight (P 0 04), heart rate fell from 63 to 52 beats/min (P ) and blood pressure fell from 116/74 to 113/69 mmhg (P 0 005); high-density lipoprotein (HDL)-cholesterol fell from 1 26 to 1 11 mmol/l (P 0 005), there being no change in the ratio of free to esterified cholesterol in HDL, and there was an apparent rise in serum triacylglycerols from 1 18 to 1 43 mmol/l (P 0 15 when adjusted for weight gain). Low-density lipoprotein (LDL)-cholesterol and lipoprotein (a) did not change. In this study, the oxidizability of LDL was unaffected by β-blocker therapy. β-blockade was not associated with any change in LDL fatty acid profile, or β-carotene or α-tocopherol content which might account for the reduced LDL oxidizability previously reported in patients treated with β-blockers. Furthermore, neither atenolol nor propranolol, at concentrations up to 100 µmol/l, had any effect on in vitro oxidizability of LDL obtained from healthy volunteers. 3. In contrast to the favourable haemodynamic effects conferred by β-blockers, the effects on weight and serum triacylglycerols and HDL-cholesterol appear to be adverse and we did not demonstrate any changes in lipid oxidizability which might be relevant to the protective effects of β-blockers in patients with coronary disease. INTRODUCTION Treatment with β-adrenoceptor antagonists ( β- blockers) has been shown to consistently reduce cardiovascular morbidity and mortality, in particular in hypertension and after myocardial infarction [1,2]. This goes along with the considerable evidence of antiatherogenic effects of β-blockers in animal models [1,3]. However, the mechanisms by which these benefits are conferred are not fully understood [1]. β-blockers are weakly anti-arrhythmic [4], reduce myocardial ischaemia [5] and can inhibit platelet aggregation and deposition [3]. β-blockers can reduce arterial distension and strain by reducing heart rate and myocardial contractility and this may reduce trauma to the endothelium [6] and atheromatous plaque [7]. Individually or collectively, these effects may go some way to account for the up to 40% reduction in cardiovascular events seen in clinical trials in patients taking β-blockers after myocardial infarction. However, treatment with β-blockers is also associated with increased triacylglycerols and reduced high-density lipoprotein (HDL) [8], changes which might promote atherogenesis. We previously reported that β-blockers appear to reduce the susceptibility of low-density lipoprotein (LDL) to in vitro oxidation in patients with coronary disease [9]. In a case control study, LDL isolated from patients on a β-blocker proved to be 30 40% less oxidizable, as assessed by Cu +-induced conjugated diene formation [10], compared with LDL from patients not on a β-blocker. This is despite the fact that the β-blockers used in this study, metoprolol or atenolol, are relatively hydrophilic. There is considerable evidence that oxidatively modified LDL is avidly taken up by macrophages to form atherogenic foam cells in the arterial intima [11]. It seemed plausible that β-blockers may reduce clinical arterial disease events in part by rendering LDL less amenable to oxidation. We hypothesized that β-blockers may have antioxidant properties themselves or may somehow influence LDL composition in such a way as to render the LDL more resistant to oxidative change. Some evidence for the former comes from studies in which β- blockers in vitro have been shown to inhibit LDL oxidation initiated by macrophages, endothelial cells or Cu + ions [12,13]. In one study, propranolol was more effective than pindolol, which in turn was more effective than metoprolol [12]. The relative lipophilicity, and hence probable LDL uptake of the respective β-blockers, appeared to be a plausible explanation for these differences. In another study, carvedilol was more effective in reducing LDL oxidizability than Key words: antioxidants, β-blockers, coronary disease, high-density lipoprotein, low-density lipoprotein, oxidation. Abbreviations: ANOVA, analysis of variance; HDL, high-density lipoprotein; LDL, low-density lipoprotein; TBARS, thiobarbituric acid-reactive substances. Correspondence: Dr S. B. Dimmitt.

2 574 S. B. Dimmitt and others propranolol but other β-blockers (pindolol, labetalol and atenolol) proved ineffective [13]. In a further study, atenolol had no effect, while a calcium channel blocker, nifedipine, conferred antioxidant activity against in vitro Cu +-induced LDL oxidation [14]. To confirm whether β-blockers do have direct or indirect effects on LDL oxidizability, we studied the effects of β-blockers added in vitro to isolated LDL, and in addition, prospectively examined the effects of β-blockers in vivo on lipoproteins and LDL composition and oxidizability in a randomized, placebocontrolled study in healthy men. SUBJECTS AND METHODS In vitro studies Preliminary in vitro experiments were conducted to investigate the direct effect of β-blockers and on the propensity to oxidative modification of LDL. LDL was obtained from healthy volunteers for these experiments. β-blockers (pure atenolol compound supplied as 0 5 mg ml atenolol in a 0 05% citrate solution by ICI and pure propranolol HCl compound supplied by ICI) adjusted to a range of concentrations from 1 to 100 µmol l were added immediately before the initiation of the oxidation reaction with Cu + (2 µmol l). In addition, we conducted studies of citrate alone in a similar concentration range to assess whether this solute might account for the findings with atenolol. LDL oxidation LDL oxidation in vitro was examined by methodology similar to that described by Esterbauer et al. [10], with minor modifications as previously reported [9], along with a final passage of isolated LDL through a Pharmacia PD10 Sephadex column to remove the remaining EDTA just before the oxidation experiments. The cholesterol concentration of the LDL was measured using a standard enzymic method (Monotest, Boehringer Mannheim, Germany) and the LDL diluted with PBS to a standard concentration of 0 3 mmol l cholesterol. Oxidation was initiated by the addition of freshly prepared aqueous CuCl solution (final concentration 2 µmol l). Oxidation kinetics were determined by monitoring the change in absorbance at 234 nm using a DU650 UV-Vis spectrophotometer (Beckman Instruments Inc., CA, U.S.A.) with absorbance readings made automatically every 20 min for a total of 4 h. All experiments were performed at 30 C unless otherwise stated. The plot of absorbance against time was divided into three phases, i.e. a lag, a propagation and a decomposition phase, as described by Esterbauer et al. [10]. The lag time was defined as the intercept between the tangent of the absorbance curve during the propagation phase with the baseline. The rate of oxidation was calculated from the slope of the absorbance curve during the propagation phase. Total conjugated dienes produced by the oxidation procedure was calculated using maximal absorbance at 234 nm and the Beer Lambert Law with the molar absorbance of conjugated dienes assumed to be Subjects for in vivo study Male volunteers were sought for the in vivo placebocontrolled study by advertisements in the local media. Subjects were aged years and were by medical history, examination and routine investigations (glucose, creatinine, full blood count, liver function tests, electrocardiogram), free of cardiac or systemic disease. Subjects had no history of asthma nor bradyarrhythmia, were non-smokers, and were excluded if there was a history of ingestion of antioxidant vitamins or drugs which may affect lipid metabolism. After the screening visit, 35 subjects were excluded or declined consent, and 45 went on to be randomized to receive either 4 weeks of placebo (one tablet twice a day the first week, doubling to two tablets twice a day thereafter), metoprolol (50 mg twice a day the first week, doubling to 100 mg twice a day) or propranolol (40 mg twice a day the first week, doubling to 80 mg twice a day). Participants attended at baseline, and again after 1 and 4 weeks on medication. They were monitored carefully with respect to β-blocker side effects, by symptom review and by assessment of pulse and blood pressure (measured five times in succession, both supine and erect, using a Dinamap 1845SX oscillometric instrument). Blood was collected by venepuncture at baseline and at 4 weeks. Participants were encouraged to maintain their usual lifestyle and in particular, their usual diet, alcohol intake and exercise. These were monitored by food frequency questionnaire and did not appear to change during the course of the study. All medication bottles were returned and tablet counts were conducted. The study was approved by the Royal Perth Hospital Ethics Committee and the University of Western Australia Human Rights Committee. Lipids Serum cholesterol and triacylglycerols were determined by enzymic methods (Trace Scientific Pty Ltd, Baulkam Hills, NSW, Australia). HDL-cholesterol was determined after heparin manganese chloride precipitation. HDL free cholesterol was measured by a spectrophotometric cholesterol esterase method (Boehringer Mannheim) after heparin manganese chloride precipitation and was expressed as a percentage of the total cholesterol. LDL-cholesterol was derived by the method of Friedewald et al. [15]. Apolipoprotin (a) was measured by a two-site immunoradiometric assay, using two different monoclonal antibodies (Mecordia, Uppsala, Sweden). Lipoprotein isolation Blood was collected into EDTA (1 mg ml). LDL was isolated from plasma by density gradient ultra-

3 Effect of β-blockers on plasma lipids 575 centrifugation [16]. Briefly, plasma density was increased to 1 07 by addition of NaCl and then a fourstep gradient was constructed over the plasma using the following densities: 0 5 ml (NaCl), 0 5 ml 1 04 (NaCl), 0 5 ml 1 02 (NaCl) and 0 9 ml of H O. Samples were ultracentrifuged at g (average) for 4 h using a Centrikon T-1190 Ultracentrifuge (Kontron Instruments, Milano, Italy). The LDL band was collected by aspiration and passed through a Pharmacia PD10 Sephadex column to remove the excess salt and the majority of the EDTA. The LDL was stored at 4 C in the dark and studied within 24 h. Fatty acid analysis The fatty acid composition of individual LDL samples was determined, after chloroform methanol (2:1 v v) extraction, by gas chromatography of the methyl ester derivatives, using methods previously described [17]. LDL antioxidant analysis LDL β-carotene and α-tocopherol were measured simultaneously by reverse-phase HPLC using electrochemical detection, as previously described [18]. Thiobarbituric acid-reactive substances assay Thiobarbituric acid-reactive substances (TBARS) were also measured to determine the extent of lipid peroxidation. An equal volume of 25% (w v) trichloroacetic acid was added to each sample followed by an equal volume of 1% (w v) thiobarbituric acid. The mixture was heated for 30 min in a boiling water bath. After cooling on ice, the samples were centrifuged in a microfuge at rev min for 10 min and the absorbance of the clear supernatant measured at 532 nm on a DU 650 Spectrophotometer (Beckman Instruments Inc, CA, U.S.A.). The amount of TBARS was determined by comparison with a standard curve using malondialdehye generated from tetramethoxypropane immediately before assay. These results were then expressed in nmol equivalents of malondialdehyde per mg of LDL protein. LDL protein was determined by the method of Lowry et al. [19] using BSA as standard. In the present experiments, the results are expressed as a percentage of the control. Statistical methods In the in vitro study, measurements at different baseline concentrations of β-blocker were compared by one-way analysis of variance (ANOVA). Variables examined in the in vivo study were confirmed to be normally distributed by Kolmogorov Smirnov testing. Within-group changes from baseline to 4 weeks were examined by Student s t-test. Changes attributable to β-blocker treatment after 4 weeks, compared with placebo, were examined by multiple regression analysis using dummy variables. RESULTS In vitro studies Initially we examined the effect of atenolol in citrate buffer (atenolol citrate 1:100 by weight) on LDL oxidizability. Atenolol in citrate increased lag time in a dose-dependent manner over 1 to 10 µmol l (F, 54 77, P 0 001, by ANOVA, Table 1). It also appeared to decrease maximal oxidation rate, although it was difficult to obtain an accurate measure of rate at the higher doses of the drug (Table 1). TBARS formation was also progressively decreased with increasing concentrations of atenolol in citrate. However, atenolol without citrate at up to 100 µmol l had no significant effect on any of the indices of LDL oxidation (Table 1). Citrate, at doses equivalent to those in the previous experiments ( µmol l), showed a dose-dependent effect on LDL lag time (F, 24 09, P 0 001, Table 1). These results demonstrate that the effects of atenolol in citrate on LDL oxidizability were probably due to the chelating effects of citrate buffer on transitional metal ions, rather than an effect of the β-blocker. Similarly, there was no significant effect of up to 50 µmol l propranolol on lag time or rate of oxidation (Table 1). Conducting the experiment at 37 C to facilitate entry of the drug into the LDL particle (results not shown) similarly showed no effect of propranolol on LDL oxidation. Concentrations of propranolol greater than 50 µmol l generated a high background absorbance at 234 nm and were not Table 1 Effect of in vitro β-blockers on indices of LDL peroxidation Results are expressed as means (of three experiments) S.E.M. Dose (µmol/l) Lag time (min) Rate 10 4 ( abs/min) TBARS (% of control) Atenolol in citrate * ** * ** 60 Atenolol Citrate ** * ** 55 Propranolol *At least one sample did not oxidize within 4 h and an approximation of the lag time determined and thus errors are inappropriate. **Very long lag times precluded an accurate estimation of rate.

4 576 S. B. Dimmitt and others Table 2 Subject characteristics Results at baseline and after 4 weeks, expressed as means S.E.M. Statistical significance: *P 0 05, **P 0 01, ***P ( P 0 08), when difference from baseline to 4 weeks examined within groups by Student s t-test. Placebo Metoprolol Propranolol n Age (years) Body mass index (kg/m 2 ) Weight (kg) weeks * Heart rate (beats/min) Baseline weeks ** 53 2*** Blood pressure (mmhg) Baseline 115/72 2/1 124/75 3/2 122/75 3/2 4 weeks 116/74 2/1 113/68 3/2** 112/70 3/2** Excluding the four subjects who withdrew (one from placebo, three from propranolol groups). studied. Propranolol did not significantly influence LDL peroxidation as assessed by TBARS at 30 C (or 37 C), even up to concentrations of 100 µmol l. Subjects (in vivo study) The three treatment groups were satisfactorily matched at baseline by ANOVA (Table 2) with respect to age, body mass index and blood pressure, although heart rate was a little slower in the metoprolol group (P 0 03) and by Student s t-test, systolic blood pressure was higher (P 0 01). β-blockers were well tolerated overall. However, three subjects randomized to propranolol were withdrawn because of lethargy and a fourth participant from the placebo group withdrew for unrelated personal reasons. Among the 42 subjects completing the study, compliance was 85% in all but one participant (on propranolol) who returned 52% of his tablets. There was no appreciable change in diet or lifestyle based on frequency questionnaires conducted at the beginning and end of the study. Testimony to generally good compliance were the falls in heart rate and blood pressure after 4 weeks on metoprolol and propranolol (Table 2). The baseline and 4-week mean blood pressures in the metoprolol and propranolol groups were similar, suggesting the groups were matched and experienced similar β- blockade. The two groups were consequently combined for subsequent analyses, in which multiple regression analysis was used (Table 5). The small weight gain in the groups taking β-blockers (0 5 kg) reached statistical significance when the two groups were combined (P 0 04). Lipids Baseline mean cholesterol, triacylglycerols, HDLand LDL-cholesterol were not significantly different Table 3 Lipids Results at baseline and after 4 weeks, expressed as means S.E.M. Statistical significance: *P 0 05, **P 0 01, ***P , when difference from baseline to 4 weeks examined within groups by Student s t-test. expressed as median concentration (not normally distributed). Placebo Metoprolol Propranolol Cholesterol (mmol/l) Baseline weeks Triacylglycerols (mmol/l) Baseline weeks * HDL-cholesterol (mmol/l) Baseline weeks *** ** LDL-cholesterol (mmol/l) Baseline weeks Apolipoprotein (a) (units/l) Baseline weeks Table 4 LDL oxidizability and composition Results at baseline and after 4 weeks, expressed as means S.E.M. Placebo Metoprolol Propranolol Lag time (min) Baseline weeks Oxidation rate (abs 10 3 /min) Baseline weeks Diene production (µmol/mmol cholesterol) Baseline weeks β-carotene (µmol/mmol cholesterol) Baseline weeks α-tocopherol (µmol/mmol cholesterol) Baseline weeks by ANOVA (Table 3). In the two groups treated with β-blockers, HDL-cholesterol fell by 4 weeks compared with baseline (P 0 01)), and triacylglycerols rose (with metoprolol, P 0 04; with propranolol, P 0 08). Comparing the combined metoprolol and propranolol groups with placebo (Table 5), and after adjustment for the small weight gain, there was a 12% fall in HDL-cholesterol (from 1 25 to 1 11 mmol l, P 0 001) and a non-significant 18% rise in triacylglycerols (1 21 to 1 43 mmol l, P 0 15). In

5 Effect of β-blockers on plasma lipids 577 Table 5 Changes after 4 weeks in metoprolol and propranolol groups combined Results expressed as means S.E.M. P-values determined by multiple regression analysis in β-blocker-treated subjects (n 28) compared with placebo (n 14). NS, not significant. Baseline 4 weeks P-value Heart rate (beats/min) 60 1 a Systolic blood pressure (mmhg) Diastolic blood pressure (mmhg) Weight (kg) Cholesterol (mmol/l) NS Triacylglycerols (mmol/l) HDL-cholesterol (mmol/l) LDL-cholesterol (mmol/l) NS Lag time (min) NS Oxidation rate NS (abs 10 3 /min) Diene production NS (µmol/mmol cholesterol) Adjusted for weight change. addition, neither total LDL-cholesterol, apolipoprotein (a) nor the percentage of total HDL-cholesterol which was free (not shown) changed significantly in subjects taking a β-blocker. LDL composition and oxidizability with β-blocker treatment in vivo Oxidizability of LDL isolated from subjects, expressed in terms of lag time to onset of oxidation, oxidation rate and maximum diene production, did not differ between groups at baseline nor after 4 weeks of treatment with either (Table 4) or both (Table 5) β- blockers. LDL antioxidant (Table 4) and fatty acid composition similarly did not change with treatment (results not shown). DISCUSSION Contrary to the findings in some previous studies, there was no evidence in the present work of either an in vitro or in vivo effect of β-blocker therapy on LDL oxidizability or composition. The in vitro studies showed that neither atenolol nor propranolol have antioxidant activity using our LDL oxidation methodology. In previous in vitro studies, propranolol, a relatively lipophilic β-blocker, was reported to confer some protection against oxidation in Cu +-induced or cell-based oxidation systems [12,13], whereas less lipophilic agents such as metoprolol [12] or atenolol [13] proved less or not effective. The longer incubation periods used in one study [12] might account in part for the greater inhibitory effects seen with some β-blockers. Propranolol has also shown antioxidant properties at high concentration (100 µm) in a different oxidation system [20]. It has been suggested that propranolol acts as a xanthine oxidase inhibitor rather than a chain-breaking antioxidant such as α-tocopherol [21]. The exact methodology used to study LDL oxidation is likely to have an important bearing on the findings. Our finding of an inhibitory effect of citrate buffer on LDL oxidation raises the important question of whether findings by other groups may have been a consequence of similar artefact from buffers or other constituents used. The absence of any effect of β-blockers in the present in vivo randomized controlled trial contrasts with results from our previous case control study, in which the LDL isolated from coronary patients on β-blockers had a lower oxidizability than that from patients not on β-blockers [9]. Even though the magnitude of the difference was substantial (22% longer lag time and 40% slower oxidation rate in patients on a β-blocker), the difference in LDL oxidizability between patients on and not on a β-blocker may have been due to chance, or a consequence of dietary or other factors, particularly as the previous study was cross-sectional and modest in size (n 20). Most patients on a β- blocker in the coronary study were using metoprolol in similar doses to those used in the present study. In trying to reconcile the contrasting results of our two clinical studies, it is important to note that the subjects in the present study were younger (mean age 48 years, compared with 59 years in the previous coronary study) and free of clinical coronary disease. There may be qualitative attributes peculiar to the LDL in coronary patients which could have predisposed these individuals to atheroma, which are peculiarly sensitive to some antioxidant effects of β- blockers. This appears less likely given that the LDL fatty acid profile and the β-carotene and α-tocopherol content did not change in the present study, but there may have been other unmeasured attributes of LDL, such as other antioxidants, lipids or the apoproteins, which may have been relevant. Effective β-blockade was achieved in the present study, gauged by the significant reductions in heart rate and blood pressure, and in HDL. The high degree of variability of triacylglycerols [22] may well explain why the rise in triacylglycerols with β-blocker treatment did not reach statistical significance. Similar apparently adverse changes in HDL and triacylglycerols have been observed in many previous studies [8], probably related to reduced lipoprotein lipase by direct inhibition [23] or to reduced peripheral tissue perfusion. Clinical trials have consistently demonstrated reductions in cardiovacular event rate associated with short- and long-term treatment with β- blockers [1,2], presumably related to the hypotensive and negative chronotropic effects of β-blockers and perhaps other mechanisms. The weight gain associated with treatment with β- blockers is likely to be due to the inhibition of β- receptor-mediated adipose tissue lipolysis [24], possibly along with reduced whole-body metabolic rate [25]. The weight gain probably contributes at least in part to the rise in serum triacylglycerols seen with β- blocker therapy. In conclusion, we have not been able to confirm any effect of β-blockers on the propensity of LDL to

6 578 S. B. Dimmitt and others oxidative modification in vitro or in vivo, nor any compositional change in LDL which might account for the decreased LDL oxidizability previously described in coronary patients on β-blockers [9]. Unless the LDL from patients with coronary disease shows a unique specificity for effects of β-blockers it seems unlikely that the beneficial effects of β-blockers on coronary heart disease are in any way mediated by changes in plasma lipids. ACKNOWLEDGMENTS The valuable assistance of Ms Mary-Ann Powell and Ms Jackie Ritchie, clinical research nurses, Dr Val Burke, biostatistician, Mr Ken Robertson, medical scientist, and Ms Gabrielle Hall, for manuscript preparation, is gratefully acknowledged. The study was supported by the Merck, Sharpe & Dohme (Australia) Research Foundation and the National Health & Medical Research Council. Astra Pharmaceuticals kindly provided placebo and metoprolol tablets and ICI Australia provided propranolol tablets. REFERENCES 1. Kaplan JR, Manuck SB, Adams MR, Clarkson TB. (1987) The effects of beta-adrenergic blocking agents on atherosclerosis and its complications. Eur Heart J 8: Yusuf S, Wittes J, Friedman L. (1988) Overview of results of randomized clinical trials in heart disease: I. Treatments following myocardial infarction. JAMA 260: Cruickshank JM, Smith JC. (1989) The beta-receptor, atheroma and cardiovascular damage. Pharmacol Ther 42: Ahmed R, Singh N. (1993) Antiarrhythmic drugs. Curr Opin Cardiol 8: Osborne JA, Stone PH. (1994) Recent advances in the understanding and management of stable and unstable angina pectoris and asymptomatic myocardial ischaemia. Curr Opin Cardiol 9: Ablad B, Bjorkman J-A, Gustafsson D, Hansson G, Ostlund- Lindqvist A-M, Pettersson K. (1998) The role of sympathetic activity in atherogenesis: effects of β-blockade. Am Heart J 116: Fitzgerald JD. (1987) By what means might beta blockers prolong life after acute myocardial infarction? Eur Heart J 8: Roberts WC. (1989) Recent studies on the effects of beta blockers on blood lipid levels. Am Heart J 117: Croft KD, Dimmitt SB, Moulton C, Beilin LJ. (1992) Low density lipoprotein composition and oxidizability in coronary disease apparent favourable effect of beta blockers. Atherosclerosis 97: Esterbauer H, Striegl G, Puhl H, Rotheneder M. (1989) Continuous monitoring of in vitro oxidation of human low density lipoprotein. Free Rad Res Commun 6: Steinberg D, Parthasarathy S, Carew TE, Khoo JC, Witztum JL. (1989) Beyond cholesterol. Modifications of low-density lipoprotein that increase its atherogenicity. N Engl J Med 320: Maziere C, Auclair M, Maziere J-C. (1992) Lipophilic β-blockers inhibit monocyte and endothelial cell-mediated modification of low density lipoproteins. Biochim Biophys Acta 1126: Yue T-L, McKenna PJ, Lysko PG, Ruffolo Jr RR, Feuerstein GZ. (1992) Carvedilol, a new antihypertensive, prevents oxidation of human low density lipoprotein by macrophages and copper. Atherosclerosis 97: Lesnik P, Dachet C, Petit L, et al. (1997) Impact of a combination of a calcium antagonist and a β-blocker on cell- and coppermediated oxidation of LDL and on the accumulation and efflux of cholesterol in human macrophages and murine J774 cells. Arterioscler Thromb Vasc Biol 17: Friedewald WT, Levy RI, Fredrickson DS. (1972) Estimation of the concentration of low-density lipoprotein cholesterol in plasma without use of the preparative ultra-centrifuge. Clin Chem 18: Croft KD, Proudfoot J, Moulton C, Beilin LJ. (1991) The effect of lipoproteins on the release of some eicosanoids by stimulated human leukocytes. A possible role in atherogenesis. Eicosanoids 4: Croft KD, Beilin LJ, Vandongen R, Rouse I, Maserei J. (1990) Leukocyte and platelet function and eicosanoid production in subjects with hypercholesterolaemia. Atherosclerosis 83: Lang JK, Gohil K, Packer L. (1986) Simultaneous determination of tocopherols, ubiquinols and ubiquinones in blood, plasma, tissue homogenates and subcellular fractions. Anal Biochem 157: Lowry OH, Roseborough NJ, Farr AL, Randall RJ. (1951) Protein measurement with the Folin phenol reagent. J Biol Chem 193: Mak T, Weglicki WB. (1988) Protection by β-blocking agents against free radical-mediated sarcolemmal lipid peroxidation. Circ Res 63: Janero DR, Lopez R, Pittman J, burghardt B. (1989) Propranolol as xanthine oxidase inhibitor: implications for antioxidant activity. Life Sci 44: Mogadam M, Ahmed SW, Mensch AH, Godwin ID. (1990) Withinperson fluctuations of serum cholesterol and lipoproteins. Arch Intern Med 150: Kubo M, Hostetler KY. (1987) Inhibition of purified bovine milk lipoprotein lipase by propranolol and other β-adrenergic blockers in vitro. Biochem Biophys Acta 918: Arner P. (1995) The β 3 -adrenergic receptor a cause and cure of obesity? N Engl J Med 333: Lamont LS. (1995) Beta-blockers and their effects on protein metabolism and resting energy expenditure. J Cardiopulmonary Rehabil 15: Received 25 August 1997/17 December 1997; accepted 15 January 1998