EXPERIMENTAL STUDIES Jpn Circ J 1999; 63: 471–477

Combined Effects of and Bezafibrate on Lipoprotein Metabolism and Cholesteryl Ester Transfer Protein mRNA in -Fed Rabbits

Jiafu Ou, MD; Keijiro Saku, MD; Shiro Jimi, PhD*; Yuan-Lan Liao, MD; Takao Ohta, MD**; Bo Zhang, PhD; Kikuo Arakawa, MD

Probucol decreases and bezafibrate increases plasma high density lipoprotein-cholesterol (HDL-C) levels in humans. This study was performed to determine whether the HDL-C-lowering effects of probucol could be reversed by treatment with bezafibrate in hypercholesterolemic rabbits. Forty-nine normolipidemic Japanese White rabbits were divided into 5 groups [group 1: normal chow; group 2: 0.2% cholesterol (Ch) diet; group 3: 0.2% Ch and 1% probucol diet; group 4: 0.2% Ch and 1% bezafibrate diet; group 5: 0.2% Ch and 1% probucol plus 1% bezafibrate diet] and treated for 8 weeks. Plasma lipids, cholesteryl ester transfer protein (CETP) activ- ity in the lipoprotein-deficient plasma fraction, CETP mRNA in liver tissue and plasma concentrations were investigated. Serum total cholesterol (TC) increased after the rabbits in groups 2, 3, 4 and 5 were fed Ch, but overall, no significant differences were observed in serum TC and (TG) among these groups. Serum HDL-C levels increased (p<0.01) in the bezafibrate-treated group, but a significant (p<0.05) reduction in HDL-C was observed in both the Ch + probucol (group 3) and Ch + probucol plus bezafibrate (group 5) groups; no significant difference was observed between groups 3 and 5. Significant correlation (p<0.01) was found between serum low density lipoprotein cholesterol (LDL-C) levels and plasma probucol concentrations in groups 3 and 5, but no correlation was found between plasma concentrations of probucol/bezafibrate and serum HDL-C levels. CETP activity in the lipoprotein-deficient plasma fraction increased in the Ch-, Ch + probucol-, and Ch + probu- col and bezafibrate-fed groups (groups 2, 3 and 5, respectively), whereas a significant reduction in this activity was observed in the Ch + bezafibrate-fed group (group 4). An analysis of covariance showed that the CETP activity responded more sensitively to drug treatment than did the serum HDL-C level. CETP mRNA in liver tissue was assessed by Northern blotting at 8 weeks, but no changes were observed among the 5 groups. Probucol decreased and bezafibrate increased serum HDL-C levels, through CETP activity without affecting liver CETP mRNA levels, and the decrease in HDL-C levels produced by probucol could not be reversed by bezafibrate. (Jpn Circ J 1999; 63: 471–477) Key Words: ; Cholesteryl ester transfer protein; Hypercholesterolemia; Lipids; Lipoproteins; mRNA; Probucol

robucol is a potent hypolipidemic agent that has been HDL-C levels in patients with various types of dyslipopro- reported to have antioxidant effects that may retard teinemia.9–11 The large-scale, 5-year Bezafibrate Coronary P the progression or accelerate the regression of Intervention Trial (BECAIT) demonstrated atherosclerosis.1,2 Probucol is known to lower high-density that bezafibrate slowed the progression of coronary athero- lipoprotein-cholesterol (HDL-C) levels and this reduction sclerosis, and reduced coronary events in young survivors of HDL-C has been shown to be associated with increased of myocardial infarction;12,13 bezafibrate effectively cholesteryl ester transfer protein (CETP) activity3 and lowered the serum very low density lipoprotein (VLDL)- mass.4 Yamamoto et al showed that a reduction in achilles TG, VLDL-C, and fibrinogen levels, but no change was tendon xanthoma was related to a decrease in HDL-C levels observed in low density lipoprotein-cholesterol (LDL-C). after probucol treatment.5 Recent studies have shown that it Therefore, some factor (s) other than LDL-C may play reduces the rate of restenosis after balloon coronary angio- important roles in the progression/regression of coronary plasty.6,7 However, a clinical study found that probucol had lesions.12,13 Generally, direct evidence of the antiatheroscle- no effect on the regression of femoral atherosclerosis.8 rotic effects of fibrates in animal studies is rare.14 An excep- Bezafibrate is a homologue of clofibrate that has been tion is fenofibrate,15 another homologue of clofibrate, which shown to reduce plasma triglyceride (TG) and increase has been shown to inhibit an early event in the atherogene- sis of rabbits fed a cholesterol diet; however, this action has (Received November 2, 1998; revised manuscript received December been also shown to be independent of the hypolipidemic 28, 1998; accepted February 5, 1999) effects of fenofibrate.15 Departments of Internal Medicine and *Pathology, Fukuoka Univer- Combination therapy with that use different mech- sity School of Medicine, Fukuoka and **Department of Pediatrics, anisms has been evaluated in several studies of hyper- Faculty of Medicine, University of the Ryukyus, Okinawa, Japan lipoproteinemia;16–20 however, it is unclear whether probu- Mailing address: Keijiro Saku, MD, PhD, FACP, Department of Internal Medicine, Fukuoka University School of Medicine, 45-1-7 col-reduced HDL levels should be treated by drugs, and if Nanakuma Jonanku, Fukuoka 814-0180, Japan. E-mail: hh035399@ so, whether the addition of bezafibrate could possibly msat.fukuoka-u.ac.jp reverse probucol-reduced HDL levels. In addition, the

Japanese Circulation Journal Vol.63, June 1999 472 OU J et al. combination of probucol and fibrates quite often induces an ing [3H]-cholesterol oleate ([3H]-DBP) were prepared. To extreme reduction in HDL-C,16,20,21 and the mechanism of remove total lipoproteins from the rabbit plasma sample, this phenomenon is unclear. the plasma density was adjusted to d=1.21g/ml by KBr, In the present study, we investigated the effects of and then ultracentrifuged in a Beckman TL-100 Tabletop probucol, bezafibrate and their combination on lipoprotein Ultracentrifuge (Beckman Instruments, USA) for 3.5h at metabolism in hypercholesterolemic rabbits. We used 165,000 × g. The bottom fraction was collected and cholesterol-fed rabbits because in a previous HDL-apo A-I dialyzed against 10mmol/L Tris-HCl, 150mmol/L NaCl kinetic study we found that probucol had a greater effect in (pH7.4) for 24 h with 2 changes of dialysate buffer. The the hypercholesterolemic state, and that treatment with a total CET activity in lipoprotein-deficient (d>1.21 g/ml) similar dose of , another homologue of clofi- plasma was determined in terms of the percentage of [3H]- brate, increased the rate of apolipoprotein (apo) A-I synthe- cholesterol oleate transferred from discoidal bilayer parti- sis in Watanabe heritable hyperlipidemic (WHHL) cles to human LDL.26 The reaction mixture containing 110 rabbits.14 The combined effects of probucol and bezafibrate μl of [3H]-DBP (1,587.3 Bq (42.9 nCi)) as a donor of on CETP gene expression in vivo were also examined, cholesterol ester and 130μg of LDL protein as an acceptor because there has been no previous report on the effects of was incubated with 10μl of lipoprotein-deficient (d>1.21 their interaction. The underlying mechanism of the HDL g/ml) plasma for 60 min at 37°C in the presence of 1.4 deficiency was also investigated. To the best of our knowl- mmol/L 2,2'-dithio-bis-(5'-nitropyridine), an inhibitor of edge, this is the first report to demonstrate the effect of lecithin-cholesterol acyltransferase (LCAT). After incuba- probucol/bezafibrate on plasma CETP activity and liver tion, 30μl of 0.1% dextran sulphate and 30μl of 60mmol/L CETP mRNA levels in hypercholesterolemic rabbits. magnesium chloride were added to the reaction mixture. The mixture was kept on ice for 20min and centrifuged at 10,500 × g for 10 min. The supernatant was collected and Methods the pellets (LDL) were dissolved in 0.1 mol/L NaOH. Rabbits Radioactivities in both the supernatant and pellets were Male Japanese White normolipidemic rabbits (ages counted using a Beckman liquid scintillation counter. The approximately 4 months and weighing 2.1–2.6kg) were result of the assay without a sample was used as nonspe- obtained from Kyudo Co, Ltd (Fukuoka, Japan). All of the cific CETP activity. CETP activity was expressed as total animals were housed individually under a 12-h light/dark CETP activity minus nonspecific CETP activity. cycle. The project was assessed and approved by the Ethics Committee of Fukuoka University. Northern Blot Analysis The rabbits were anesthetized with sodium phenobarbi- Study Groups tal. The liver tissue was excised, immediately frozen in Rabbits were divided into 5 groups: group 1 (N), normal liquid N2, and then stored at –80°C until use. Total cellular chow diet for 8 weeks (n=10); group 2 (Ch), 0.2% choles- RNA was extracted from the liver tissue with RNAzol B terol diet for 8 weeks (n=10); group 3 (Ch+P), 1% probucol (Biotecx Laboratories, Inc). Twenty micrograms of total with 0.2% cholesterol diet for 8 weeks (n=10); group 4 RNA was fractionated on 1% agarose formaldehyde gel (Ch+B), 1% bezafibrate with 0.2% cholesterol diet for 8 and blotted onto a positively charged nylon membrane weeks (n=10); and group 5 (Ch+P+B), 1% probucol plus (Amersham). A cDNA probe (479-base fragment, 1% bezafibrate with 0.2% cholesterol diet for 8 weeks nucleotides 437–915) for rabbit CETP27 was generated by (n=9). Each dietary treatment was started on the same day. reverse transcription-polymerase chain reaction (RT-PCR) using a GeneAmp RNA PCR Kit (Perkin Elmer), and then Preparation of Probucol and Bezafibrate Diets labeled with [32P]-deoxy-cytidinetriphosphate (dCTP) Pure probucol and bezafibrate powders (supplied by using a rediprime random labeling kit (Amersham). The Daiichi Pharmaceutical Co, Tokyo, and Kissei Co, Ltd, primers used for RT-PCR of rabbit CETP were (sense) 5'- Matsumoto, respectively) were added to the basal animal CTTTCCATAAACTGCTCCTG-3', and (antisense) 5'- chow (RC-4 or RC-4 with 0.2% cholesterol, Oriental Yeast, CTGGTTGGTGTCGAAACCCT-3'. The membrane was Tokyo) at the a concentration of 1%. No organic solvent hybridized with a labeled CETP probe, as described else- was used. Crude fat and cholesterol contents of RC-4 were where,28 and then washed with 0.1% sodium dodecyl 3.0g and 2mg per 100g of dry food, respectively. A choles- sulfate (SDS) / 0.1 × SSC at 65°C. The hybridized terol diet was prepared by adding 0.2% cholesterol to RC-4. membrane was exposed to Kodak BioMax film for 2 days at –80°C. The membrane was rehybridized with human Isolation of Lipoproteins glyceraldehyde-3-phosphate dehydrogenase (G3PDH) Blood samples were taken in the morning after an (Clontech) as described earlier, after washing with 1% SDS overnight fast. Serum was adjusted to a density (d) of 1.006 for 10min at 100°C. The proportions of CETP mRNA and g/ml for VLDL, 1.063 g/ml for low-density lipoproteins G3PDH mRNA on the autoradiograph were determined by (LDL), and 1.21g/ml for HDL, using KBr solution, before densitometric scanning (300 A Computing Densitometer being ultracentrifuged.22 Lipoprotein fractions were then and Image Quant Software v3.0 Fast Scan, Molecular isolated. Serum total cholesterol (TC) and TG, and lipids Dynamics). The CETP mRNA levels were normalized to from each of the lipoprotein fractions, were measured by G3PDH mRNA and expressed as a percentage of the enzymatic methods.23,24 G3PDH mRNA level.

Determination of Cholesteryl Ester Plasma Probucol and Bezafibrate Concentrations Transfer Protein (CETP) Activity To measure the concentrations of probucol, the lipid The activity of CETP was determined according to the fraction from 1ml of plasma was extracted. Forty micro- method of Kato et al.25 Discoidal bilayer particles contain- liters of the sample was analyzed by high-performance

Japanese Circulation Journal Vol.63, June 1999 Effects of Probucol and Bezafibrate on HDL Metabolism 473 ## ## ## ## ## ## 0.14** 48.60** 22.07 134.11** 11.72* 61.25** 52.27** 179.41** 11.72* 1.68** 1.15** 1.48** 11.67* ± ± ± ± ± ± ± ± ± ± ± ± ± 0.12 2.56 1.44 52.04 9.67 42.78 6.88 5.47 10.55 359.27 5.49 19.93 14.50 187.89 6.21 149.67 15.67 422.78 6.00 11.47 5.49 19.93 3.50 17.38 9.49 22.33 ± ± ± ± ± ± ± ± ± ± ± ± ± 9.40 23.20 53.93 ## ## ## ## ## ## 0.15** 2.32 11.64 38.67 43.73** 2.80 2.64** 18.07 6.57* 11.20 109.64** 16.67 51.51** 69.11 46.00** 10.93 148.24** 42.67 6.57** 11.20 6.61 11.96** 12.13* ± ± ± ± ± ± ± ± ± ± ± ± ± Fig1. Serum HDL-C levels in the 5 groups before and after treat- ment with probucol and/or bezafibrate. Group 1 (N), normal chow diet for 8 weeks (n=10); Group 2 (Ch), 0.2% cholesterol diet for 8 weeks (n=10); Group 3 (Ch+P), 1% probucol with 0.2% cholesterol diet for 8 weeks (n=10); Group 4 (Ch+B), 1% bezafibrate with 0.2% cholesterol diet for 8 weeks (n=10); Group 5 (Ch+P+B), 1% probucol 0.11 2.62 8.04 38.80 0.92 59.88 6.24 9.36 5.41 17.04 7.38 294.72 7.57 194.90 4.99 132.72 7.83 385.80 5.41 17.04 4.14 31.20 7.08 44.28 1.93 12.40 ± ± ± ± ± ± ± ± ± ± ± ± plus 1% bezafibrate with 0.2% cholesterol diet for 8 weeks (n=9). The ± values are expressed as means±SD. *p<0.01 vs before treatment. 21.18 49.96 p<0.01 vs after treatment of group 2 by ANOVA. ## # # liquid chromatography (HPLC)27,29 using a Waters Associates Instrument (Models 510 and 481, Milford, MA, 14.17 37.30 0.15** 2.34 2.48** 19.12 21.67** 1.82 6.64* 10.14 143.20** 11.40 6.64**52.70** 10.14 61.70 53.52** 7.92 158.85** 34.40 6.35 8.08 4.37** 4.80** (Ch+B): 0.2% cholesterol plus 1% bezafibrate diet; group 5 (Ch+P+B): ± ± ± ± ± ± ± ± ± ± ± ± USA) equipped with a 4.5mm id ×250mm Nucleosil 5C18 ± column (Chemco Osaka, Japan). The column was eluted with 3% acetic acid and acetonitrile (15:85, v/v) at a flow rate of 2.0ml/min. The effluent was monitored at 254nm. For the serum bezafibrate concentration, 20μl of the 9.93 31.80 0.10 2.53 3.82 8.58 1.84 34.24 5.31 13.86 6.88 25.26 4.37 12.00 4.93 311.76 5.319.04 13.86 160.80 3.30 125.16 8.11 358.00 sample was analyzed by HPLC equipped with a 4.6mm id 2.38 9.36 ± ± ± ± ± ± ± ± ± ± ± ± ×150mm TOSOH ODS 80 TM 5μm column. The column ± was eluted with 20mmol/L phosphate K (pH3.6):methanol (60:96, v/v) at a flow rate of 0.9ml/min. The effluent was monitored at 228nm. The plasma probucol and bezafibrate concentrations were linear between 0 and 10μg/ml, and the 7.19 33.50 1.52* 18.24 0.13** 2.29 3.99* 7.62 51.97* 1.60 4.17** 50.34 4.95* 21.84 3.99* 7.62 108.90** 9.06 2.98* 7.64 50.67** 60.70 42.24** 6.24 150.16** 32.50 lowest detectable concentration was 0.1μg/ml. p<0.05 vs after treatment of group 2 by ANOVA; ± ± ± ± ± ± ± ± ± ± ± ± ± # Statistical Analysis Data were analyzed using Statview-J 4.11 and Super- ANOVA for the Macintosh. The data are presented as means ± SD. Changes in serum lipids and lipoproteins 4.93 29.10 5.71 7.68 0.12 2.56 3.63 13.98 1.46 40.02 5.72 34.26 4.09 17.87 3.63 13.98 5.42 299.76 1.86 7.44 7.56 176.50 2.98 128.22 8.87 357.60 ± ± ± ± ± ± ± ± ± ± ± ± during the study period were analyzed by the paired t-test. ± The differences between the groups during the study were examined by an analysis of variance (ANOVA), followed by Fisher’s test, and an analysis of covariance (ANCOVA). The relationships between serum lipid parameters and plasma probucol and bezafibrate concentrations were 7.50 32.60 3.13 12.78 2.74* 2.22 0.11** 2.34 4.44 3.36** 51.58 9.42 2.57 21.87 28.01 10.44 2.98* 10.40 3.36**12.33* 9.42 12.59 61.40 6.36 30.48 34.50 ± ± ± ± ± ± ± ± ± ± ± ± examined by simple regression analysis. ±

Results Table1 shows the changes in serum TC, HDL-C, LDL- C, VLDL-C, TG, HDL-TG, LDL-TG, VLDL-TG, phos- Group 1 (N) Group 2 (Ch) Group 3 (Ch+P) Group 4 (Ch+B) Group 5 (Ch+P+B) 10.33 31.70 2.53 13.44 1.45 3.46 0.12 2.79 2.86 4.57 49.14 12.30 2.64 22.60 4.61 28.68 5.00 5.96 4.575.89 12.30 3.63 70.10 16.02 6.13 54.40 ± ± ± ± ± ± ± ± ± ± ± ± pholipid (PL), HDL-PL, LDL-PL, VLDL-PL, and body ± weight (BW) in all of the rabbits throughout the entire Before After Before After Before After Before After Before After study period. Serum TC increased after rabbits in groups 2–5 were fed cholesterol (Ch), but overall, no significant differences were observed in serum TC and TG in these groups. Treatment with probucol alone (group 3) and probucol plus bezafibrate (group 5) reduced serum HDL-C (p<0.01) (21.8±4.4mg/dl at 0 weeks vs 12.0±4.8mg/dl at 8 weeks, p<0.01 and 23.2±6.0mg/dl at 0 weeks vs 11.5±1.5 TG (mg/dl) 30.80 HDL-TG (mg/dl) 12.84 VLDL-C (mg/dl) 1.78 BW (kg) 2.35 HDL-PL (mg/dl)VLDL-PL (mg/dl) 52.62 9.06 HDL-C (mg/dl) 23.04 LDL-C (mg/dl) 12.48 VLDL-TG (mg/dl) 8.90 LDL-TG (mg/dl)PL (mg/dl) 9.06 LDL-PL (mg/dl) 63.30 7.92 TC (mg/dl) 37.30 Table 1 Changes in Serum Lipids, Lipoproteins Levels and Body Weight (BW) Group 1 (N): normal chow; group 2 (Ch): 0.2% cholesterol diet; 3 (Ch+P): plus 1% probucol 4 mg/dl at 8 weeks, p<0.01, respectively), while bezafibrate plus 1% probucol and bezafibrate diet. *p<0.05 vs before treatment by paired t-test; **p<0.01

Japanese Circulation Journal Vol.63, June 1999 474 OU J et al.

Fig 3. Cholesteryl ester transfer protein (CETP) activity was assessed in the lipoprotein-deficient plasma fraction in each group. Significant increases in CETP activity were observed in groups 2, 3 and 5, whereas a significant reduction in CETP activity was observed in group 4. *p<0.01 vs before treatment, #p<0.01 between groups.

Fig 4. The relationship between the change in cholesteryl ester transfer protein (CETP) activity and the change in high density lipoprotein-cholesterol (HDL-C) in groups 1 ( ), 2 ( ), 3 ( ), 4 ( ) Fig2. The relationships between (A) the plasma probucol concen- and 5 ( ). tration and the serum LDL-C level in group 3 ( ) and group 5 ( ), (B) the plasma probucol concentration and serum HDL-C level in group 3 ( ) and group 5 ( ), and (C) the plasma bezafibrate concen- tration and serum HDL-C level in group 4 ( ) and group 5 ( ). bezafibrate concentrations. Significant positive correlation was observed between serum LDL-C levels and plasma

probucol concentrations in groups 3 and 5 (r=0.835, alone only increased serum HDL-C levels (21.2±4.1mg/dl p=0.0014; r=0.783, p=0.0098, respectively) (Fig2A), but at 0 weeks vs 31.2±12.0 mg/dl at 8 weeks, p<0.01), and no significant correlation was observed between serum these changes were significant compared to the other HDL-C levels and plasma probucol/bezafibrate concentra- groups. Fig 1 summarizes the separate and combined tions (Fig2B,C). effects of probucol and bezafibrate on serum HDL-C CETP activity was assessed in the lipoprotein-deficient levels. No significant difference was observed between plasma fraction in each group (Fig3). Significant (p<0.01) groups 3 and 5. increases in CETP activity were observed in groups 2, 3 Plasma probucol and bezafibrate concentrations were and 5 (5.65±1.09% vs 7.05±1.19%, 5.48±0.50% vs 9.17± investigated in all of the groups at 8 weeks. Mean plasma 0.54%, 5.60±0.21% vs 7.84±0.89%, before and 8 weeks probucol concentrations were 14.3±5.6μg/ml and 16.7±5.8 after treatment, respectively). A significant reduction in μg/ml in groups 3 and 5, respectively. Mean plasma CETP activity was observed in group 4 (5.38±0.26% vs bezafibrate concentrations were 2.76±0.99μg/ml and 2.43± 4.65±0.69%, before and 8 weeks after treatment). 1.21μg/ml in groups 4 and 5, respectively. No significant Significant (p<0.01) differences were also observed differences were observed within either of these 2 pairs of between groups 2 and 3, and among groups 3, 4 and 5, after groups. Fig2 shows the relationships between serum LDL- treatment. The changes in CETP activity (delta CETP C and HDL-C levels at 8 weeks and plasma probucol/ activity) and HDL-C (delta HDL-C) from the baseline were

Japanese Circulation Journal Vol.63, June 1999 Effects of Probucol and Bezafibrate on HDL Metabolism 475

apolipoprotein-cell interaction,33 and this inhibition may explain most of the decrease in HDL produced by probu- col. Finally, probucol increases the level of cholesteryl ester transfer protein,3,4,34 thus lowering serum HDL-C, and reduces HDL particle size,3,5 but contributes to the regres- sion of xanthomas.5 However, in the recent large-scale clin- ical Probucol Quantitative Regression Swedish Trial (PQRST) in humans, probucol did not produce any signifi- cant regression of femoral atherosclerosis.8 Other studies have shown that probucol reduces the rate of restenosis after balloon coronary angioplasty, independent of reduc- tions in LDL-C and HDL-C.6,7 We recently proposed a 2- step process of reverse cholesterol transport,35 where the fractional esterification rate of cholesterol in apoB-depleted plasma (FERHDL) is a functional assay of HDL; patients with high FERHDL and low HDL-C had the highest risk for coronary heart disease, and probucol might in fact cause high FERHDL and low HDL-C. Therefore, most of probu- col’s antiatherosclerotic effect is based on its antioxidant action, not through an HDL-mediated pathway. In the present study, significant reductions in serum HDL-C levels were observed after treatment with either probucol alone or probucol plus bezafibrate, whereas bezafi- brate alone increased serum HDL-C levels (Table1, Fig1). Fig5. (A) Cellular RNA was extracted from rabbit liver tissue and One of our major interests in this investigation was to deter- mRNA levels were analyzed by Northern blotting in groups 1 (lanes mine whether or not probucol-induced reductions in HDL-C 1,2), 2 (lanes 3,4), 3 (lanes 5,6), 4 (lanes 7,8), and 5 (lanes 9,10). (B) could be reversed by the addition of bezafibrate. However, Northern blots were analyzed densitometrically, and CETP mRNA the dosage of bezafibrate used in this trial did not reverse the levels were normalized to the G3PDH mRNA level. The values are probucol-induced lowering of HDL-C. Although the dose- expressed as mean±SD. response of bezafibrate (various bezafibrate concentrations) should be tested independently, CETP activity was reduced examined by simple regression analysis. A significant with the dosage of bezafibrate used, and the plasma fasting (r=–0.535, p<0.0001) negative correlation was found bezafibrate concentrations observed here correspond to the (Fig 4). Using an analysis of covariance, the changes in maximum plasma concentrations in humans after treatment CETP activity and HDL-C were compared in each group. with the usual dosage of bezafibrate (Kissei Pharmaceutical No significant difference was found in groups 1 and 2, Co, personal communication). whereas a significant change in CETP activity versus that CETP activity in the lipoprotein-deficient plasma frac- in the serum HDL-C level was noted in groups 3, 4 and 5 tion increased in groups 2, 3 and 5 after treatment (Fig3). A (p=0.0001, p=0.0001, p=0.001, respectively), indicating significant difference was observed between groups 3 and 5 that plasma CETP activity responded more sensitively to after 8 weeks of treatment, whereas the serum HDL-C drug treatment than HDL-C. levels remained lower and no difference was observed Cellular RNA was extracted from 5 rabbits that were between groups 3 and 5 (Fig1). HDL-C was increased only selected at random from each group at 8 weeks, and mRNA in group 4, where a significant reduction in CETP activity levels were analyzed by Northern blotting. No differences vs the baseline value was observed. These data suggest that were observed among the 5 groups (Fig5A). Fig5B shows bezafibrate at this dosage efficiently raised the serum HDL- a slight increase in the mean relative CETP mRNA level, C level by reducing CETP activity, as reported in a human which was normalized to the G3PDH mRNA level, in study.36 The 14% mean reduction in CETP activity in group groups 2 (211.0±8.5%), 3 (212.6±9.3%) and 5 (211.4± 5 compared with that in group 3 might be due to bezafi- 6.2%). However, these changes were not statistically brate’s effect on CETP activity. However, there was no significant. difference in serum HDL-C concentrations between groups 3 and 5, and both groups eventually showed increased CETP activities compared with the baseline value (+67% Discussion in group 3 and +40% in group 5 at 8 weeks of drug treat- Probucol decreases HDL by reducing the synthesis of ment). In group 4, a 13% mean reduction in CETP activity HDL-apoA-I in both humans30,31 and rabbits.29 We previ- was observed, and mean serum HDL-C levels increased by ously proposed that under steady-state conditions the 47% after 8 weeks. The actual mean decrease in CETP absolute synthetic and catabolic rates (ie, transport) of a activity (delta CETP activity) in group 4 was 0.192%, given plasma substance were equal, thus increased apoA-I which corresponds to an increase (delta HDL-C) in the transport can theoretically be associated with increased HDL-C level of 1.74 mg/dl based on the regression line tissue cholesterol efflux and possibly increased HDL-medi- shown in Fig 4. However, the actual mean increase in ated cholesterol removal.32 The decreased synthesis of HDL-C in group 4 was 10.0mg/dl. Thus, the direct synthe- apoA-I and HDL induced by probucol might not be desir- sis of HDL or synthesized HDL through the lipolysis of able for preventing atherosclerosis. In addition, Tsujita and TG-rich lipoproteins may also contribute to the bezafibrate- Yokoyama have shown that the reduction in HDL by induced increase in the HDL-C levels, which is supported probucol is related to the lack of HDL-generation by by our previous report that gemfibrozil, another homologue

Japanese Circulation Journal Vol.63, June 1999 476 OU J et al. of clofibrate, increased the synthesis of apoA-I and apoA- related to the HDL-C level (Fig 2B), but this relationship II, and increased activity, which resulted was not statistically significant. No correlation was in increased HDL-C levels.32 The finding that plasma CETP observed between the plasma bezafibrate concentration and activity responded more sensitively to drug treatment than the serum HDL-C level, but serum HDL-C levels were did the HDL levels, as assessed by an analysis of covari- affected by the administration of probucol and bezafibrate. ance of the changes in the CETP activity and HDL-C, also It is still unclear whether or not probucol-induced HDL- supports this proposal. All of these data suggest that C-lowering should be treated by other drugs. However, it is changes in CETP activity vs baseline values are important clear that the effect of bezafibrate at the dosage used here in setting serum HDL-C levels, and that probucol has a did not reverse the probucol-induced lowering of HDL-C. greater effect than bezafibrate at the dosage used in the As stated before, probucol reduces HDL through 3 mecha- present study, especially with regard to HDL metabolism. nisms: (1) decreased synthesis of apoA-I, (2) selective inhi- Previously, mRNA of CETP was determined in the liver bition of the free apolipoprotein-mediated cellular lipid tissue of hypercholesterolemic rabbits; increased CETP efflux as well as the binding of apolipoprotein to the cell mRNA levels as well as increased plasma CETP mass were surface, and (3) accelerated CETP systems. Apparently observed at the introduction of a high-cholesterol diet;37 bezafibrate could not counteract these 3 effects of probucol. however, there was considerable variability in CETP A drastic HDL deficiency after combination therapy with mRNA levels in a cholesterol-fed group with similar probucol and /bezafibrate has been reported by plasma CETP levels. The relationship between the mRNAs others and by us in humans;16,20,21 however, a similar of CETP and probucol has only been assessed in the reduction in HDL was not observed in this present rabbit adipose tissue of hypercholesterolemic subjects and study. Thus, the underlying mechanism remains unclear. hamsters after treatment with probucol;34 despite increased Probucol increased CETP activity and this effect was atten- CETP mass after treatment, CETP adipose tissue mRNA uated by the co-administration of bezafibrate; a lower levels were markedly decreased in the former group, CETP activity was observed with the combination of although a significant increase in plasma CETP mass and probucol and bezafibrate than with probucol alone. The adipose tissue mRNA was seen in the latter group. In our concentration of cholesterol in the diet (0.2%) and the study, no significant changes in CETP mRNA were doses of probucol and bezafibrate used in this study did not observed in the 5 groups (Fig5A). One possible explana- significantly affect liver CETP mRNA, but did change tion is that the cholesterol diet in the present study serum HDL-C levels in circulating blood by affecting the contained only 0.2% cholesterol, and the maximum serum CETP activity. TC levels were 30–40% of those in a study by Quinet et al,37 which did not affect CETP mRNA levels. Fig 5B Acknowledgments shows a slight increase in the mean CETP mRNA adjusted This work was supported by grants-in-aid from the Ministry of by G3PDH mRNA in rabbit liver in groups 2, 3 and 5; Education, Science and Culture of Japan (nos. 09670773, 10670221 and these increases were not statistically significant. In our 10670693). We wish to thank Dr Kan Liu for his interest in our project recent study of the incubation of a Chinese Hamster ovary and for his valuable comments. cell line that had been stably transfected with a human CETP gene with probucol, both cellular CETP mRNA and References CETP activities in the culture medium were increased by 1. Carew TE, Schwenke GC, Steinberg D: Antiatherogenic effect of probucol in a dose-dependent manner,38 but the probucol probucol unrelated to its hypercholesterolemic effect: evidence that concentration in both the culture medium and the cells was antioxidants in vivo can selectively inhibit low density lipoprotein very high. In the present study, the plasma probucol degradation in macrophage-rich fatty streaks and slow the progres- sion of atherosclerosis in Watanabe heritable hyperlipidemic rabbit. concentration corresponded to that in mild human hyper- Proc Natl Acad Sci USA 1987; 84: 7725–7729 cholesterolemia that is treated with the usual doses of 2. Kita T, Nagano Y, Yokode M, Ishii K, Kume N, Ooshima A, et al: probucol for several months (ie, we used 0.2% Ch-fed Probucol prevents the progression of atherosclerosis in Watanabe rabbits and the maximum serum TC was around 400mg/dl, heritable hyperlipidemic rabbits, an animal model for familial hyper- cholesterolemia. Proc Natl Acad Sci USA 1987; 84: 5928–5931 which is closer to physiological conditions than those in 3. Franceschini G, Sirtori M, Vaccarino V, Gianfranceschi G, previous studies in which rabbits were fed a 0.5% choles- Rezzonnico L, Chiesa G, et al: Mechanisms of HDL reduction after terol diet).29,37 Serum HDL-C levels did not increase after probucol: Changes in HDL subfractions and increased reverse 0.2% Ch (group 2), in contrast to our previous study using cholesteryl ester transfer. Arterioscler Thromb 1989; 9: 462–469 0.5% Ch.29 Therefore, the doses of cholesterol, probucol 4. McPherson R, Hogue M, Milne RW, Tall AR, Marcel YL: Increase in plasma cholesteryl ester transfer protein during probucol treat- and bezafibrate used in this study did not alter CETP ment: Relation to changes in high density lipoprotein composition. mRNA levels, and both probucol and bezafibrate may Arterioscler Thromb 1991; 11: 476–481 affect protein levels more effectively than CETP mRNA 5. Yamamoto A, Matsuzawa Y, Yokayama S, Funahachi T, Yamamura levels, although the effects of such drugs on both CETP T, Kishino B: Effects of probucol on xanthomata regression in famil- ial hypercholesterolemia. Am J Cardiol 1986; 57: 29H–35H mRNA translation and CETP protein degradation were not 6. Tardif JC, Cote G, Lesperance J, Bourassa M, Lambert J, Doucet S, determined in this study. Because rabbit CETP appeared to et al: Probucol and multivitamins in the prevention of restenosis after be synthesized mainly by the liver and the contribution to coronary angioplasty. N Engl J Med 1997; 337: 365–372 the total circulating CETP mass by extrahepatic tissues 7. Yokoi H, Daida H, Kuwabara Y, Nishikawa H, Takatsu F, Tomihara 39 H, et al: Effectiveness of an antioxidant in preventing restenosis after would likely be insignificant, CETP gene expression in percutaneous transluminal coronary angioplasty: the Probucol adipose tissue was not performed in this study. Angioplasty Restenosis Trial. J Am Coll Cardiol 1997; 30: 855–862 In our study, the plasma concentrations of probucol were 8. Walldius G, Erikson U, Olsson AG, Bergstrand L, Hadell K, positively correlated with the serum LDL-C levels in both Johansson J, et al: The effect of probucol on femoral atherosclerosis: the Probucol Quantitative Regression Swedish Trial (PQRST). Am J groups 3 and 5, which confirmed our previous finding that Cardiol 1994; 74: 875–883 29 probucol binds to lipids in plasma and adipose tissue. The 9. Saku K, Sasaki J, Arakawa K: Effects of slow-release bezafibrate on plasma concentration of probucol (group 3) was inversely serum lipids, lipoproteins, apolipoproteins and postheparin lipolytic

Japanese Circulation Journal Vol.63, June 1999 Effects of Probucol and Bezafibrate on HDL Metabolism 477

activities in patients with type IV and type V hypertriglyceridemia. 25. Kato H, Nakanishi T, Arai H, Nishida HI, Nishida T: Purification, Clin Ther 1989; 11: 331–340 microheterogeneity, and stability of human lipid transfer protein. J 10. Stewart JM, Packard CJ, Lorimer AR, Boag DE, Shepherd J: Effects Biol Chem 1989; 264: 4082–4087 of bezafibrate on receptor-mediated and receptor-independent low 26. Hayashibe H, Asayama K, Nakane T, Uchida N, Kawada Y, density lipoprotein catabolism in type II hyperlipoproteinaemic Nakazawa S: Increased plasma cholesteryl ester transfer activity in subjects. Atherosclerosis 1982; 44: 355–365 obese children. Atherosclerosis 1997; 129: 53–58 11. Vessby B, Lithell H, Ledermann H: Elevated lipoprotein lipase activ- 27. Nagano Y, Kita T, Yokode M, Ishii K, Kume N, Otani H, et al: ity in skeletal muscle tissue during treatment of hypertriglyceri- Probucol does not affect lipoprotein metabolism in macrophages of daemic patients with bezafibrate. Atherosclerosis 1982; 44: 113–118 Watanabe heritable hyperlipidemic rabbits. Arteriosclerosis 1989; 9: 12. de Faire U, Ericsson CG, Grip L, Nilsson J, Svane B, Hamsten A: 453–461 Retardation of coronary atherosclerosis: the Bezafibrate Coronary 28. Sambrook J, Fritsch EF, Maniatis T: Molecular cloning: A laboratory Atherosclerosis Intervention Trial (BECAIT) and other angiographic manual. New York: Cold Spring Harbor Laboratory Press, 1989 trials. Cardiovasc Drugs Ther 1997; 11(Suppl 1): 257–263 29. Ying H, Saku K, Harada R, Takami N, Sasaki N, Saito Y, et al: 13. Ericsson CG, Hamsten A, Nilsson J, Grip L, Svane B, de Faire U: Putative mechanisms of action of probucol on high-density lipopro- Angiographic assessment of effects of bezafibrate on progression of tein apolipoproteinA-I and its isoprotein kinetics in rabbits. Biochim in young male postinfarction patients. Lancet Biophys Acta 1990; 1047: 247–254 1996; 347: 849–853 30. Nestel PJ, Billington T: Effects of probucol on low density lipopro- 14. Liu R, Saku K, Jimi S, Ohta T, Zhang B, Takebayashi S, et al: tein removal and high density lipoprotein synthesis. Atherosclerosis Mechanism of action of gemfibrozil on HDL metabolism and athero- 1981; 38: 203–209 sclerosis in WHHL rabbits. Cardiovasc Drugs Ther 1997; 11: 659– 31. Atmeh RF, Stewart JM, Boag DE, Packard CJ, Lorimer AR, 668 Shepherd J: The hypolipidemic action of probucol: a study of its 15. Saitoh K, Mori T, Kasai H, Nagayama T, Tsuchiya A, Ohbayashi S: effect on high and low density lipoproteins. J Lipid Res 1983; 24: Anti-atheromatous effects of , a hypolipidemic drug. I: 588–595 Anti-atheromatous effects are independent of its hypolipidemic 32. Saku K, Gartside P, Hynd BA, Kashyap ML: Mechanism of action of effect in cholesterol-fed rabbits. Folia Pharmacol Jpn 1995; 106: gemfibrozil on lipoprotein metabolism. J Clin Invest 1985; 75: 41–50 1702–1712 16. Davignon J: Medical management of and the role of 33. Tsujita M, Yokoyama S: Selective inhibition of free apolipoprotein- probucol. Am J Cardiol 1986; 57: 22H–28H mediated cellular lipid efflux by probucol. Biochemistry 1996; 35: 17. Witstum JL, Simmons D, Steinberg D, Beltz WF, Weinreb R, Young 13011–13020 SG, et al: Intensive combination drug therapy of familial hypercho- 34. Quinet EM, Huerta P, Nancoo D, Tall AR, Marcel YL, McPherson lesterolemia with , probucol, and hydrochloride. R: Adipose tissue cholesteryl ester transfer protein mRNA in Circulation 1989; 79: 16–28 response to probucol treatment: cholesterol and species dependence. 18. Illingworth DR, Bacon S: Treatment of heterozygous familial hyper- J Lipid Res 1993; 34: 845–852 cholesterolemia with lipid-lowering drugs. Arteriosclerosis 1989; 9: 35. Saku K, Zhang B, Ohta T, Liu R, Jimi S, Shirai K, et al: Quantity and 121–134 function of high density lipoprotein as an indicator of coronay ather- 19. Saku K, Zhang B, Hirata K, Okura Y, Bai H, Liu R, et al: Combined osclerosis. J Am Coll Cardiol 1999: (in press) therapy with probucol and in hypercholesterolemia. Eur J 36. Homma Y, Ozawa H, Kobayashi T, Yamaguchi H, Sakane H, Clin Pharmacol 1993; 44: 535–539 Mikami Y, et al: Effect of bezafibrate therapy on subfractions of 20. Saku K, Zhang B, Jimi S, Bai H, Hirata K, Sasaki N, et al: High- plasma low-density lipoprotein and high-density lipoprotein, and on density lipoprotein and apolipoprotein A-I deficiency induced by activities of lecithin: cholesterol acyltransferase and cholesteryl ester combination therapy with probucol and bezafibrate. Eur J Clin transfer protein in patients with hyperlipoproteinemia. Pharmacol 1995; 48: 209–215 Atherosclerosis 1994; 106: 191–201 21. Yokoyama S, Yamamoto A, Kurasawa T: A little more information 37. Quinet EM, Agellon LB, Kroon PA, Marcel YL, Lee YC, Whitlock about aggravation of probucol-induced HDL-reduction by clofibrate. ME, et al: Atherogenic diet increases cholesteryl ester transfer Atherosclerosis 1988; 70: 179–181 protein messenger RNA levels in rabbit liver. J Clin Invest 1990; 85: 22. Havel RJ, Eder HA, Bragdon JH: The distribution and chemical 357–363 composition of ultracentrifugally separated lipoproteins in human 38. Ou J, Saku K, Jimi S, Ohta T, Zhang B, Pownall HJ, et al: serum. J Clin Invest 1955; 34: 1345–1353 Mechanism of action of probucol cholesterol ester transfer protein 23. Allain CC, Poon LS, Chan CS, Richmond W, Fu PC: Enzymatic (CETP) mRNA in a chinese hamster ovary cell line that had been determination of total serum cholesterol. Clin Chem 1974; 20: 470– stably transfected with a human CETP gene. Biochim Biophys Acta 475 1998; 1393: 153–160 24. Eggstein M, Kreutz FH: Eine neue Bestimmung der Neutralfette in 39. Nagashima M, McLean JW, Lawn RM: Cloning and mRNA tissue Blutserum und Gewebe. I. Mit Prinzip, Durchfuhrung und distribution of rabbit cholesteryl ester transfer protein. J Lipid Res Besprechung des Methiode. Klin Wochenschr 1966; 44: 262 1988; 29: 1643–1649

Japanese Circulation Journal Vol.63, June 1999