The Effects of Statin and Fibrate on Lowering Small Dense LDL- Cholesterol in Hyperlipidemic Patients with Type 2 Diabetes

128 Journal of Atherosclerosis and Thrombosis Vol.14, No.3 Lipid-Lowering Agents and Small Dense-LDL 129 Original Article

Anna Tokuno1, Tsutomu Hirano1, Toshiyuki Hayashi1, Yusaku Mori1, Takeshi Yamamoto1, Masaharu Nagashima1, Yuji Shiraishi1, Yasuki Ito2, and Mitsuru Adachi1

1First Department of Internal Medicine, Showa University School of Medicine, Tokyo, Japan. 2Denka Seiken, Co. Tokyo, Japan.

Aim: Small dense (sd)-low-density lipoprotein (LDL) is a potent atherogenic lipoprotein. The overall atherogenicity of this lipoprotein can be precisely assessed by quantifying sd-LDL rather than by measuring the LDL size. We studied the effects of representative lipid-lowering agents (statin and fibrate) on sd-LDL-cholesterol (C) in patients with type 2 diabetes. Methods: Sd-LDL-C was measured by the precipitation method established by Hirano and Ito. Large buoyant (lb)-LDL-C was calculated by subtracting sd-LDL-C from LDL-C. Type 2 diabetes patients (n =72) were administered lipid-lowering agents for three months: patients with hypercholesterolemia received 1 mg of pitavastatin and those with hypertriglyceridemia received 100 mg of micronized fenofibrate. Results: Pitavastatin reduced LDL-C by 25% and reduced TG by 8%. The statin decreased sd- LDL-C by 26%, and lb-LDL-C by 22%. Fenofibrate reduced TG by 38% and increased HDL-C by 14%. The fibrate decreased sd-LDL-C by 23% without changing LDL-C. The pitavastatin-induced reduction of sd-LDL-C was significantly correlated with the reduction of LDL-C and apo B, whereas the fenofibrate-induced reduction of sd-LDL-C was correlated with the reduction of TG. Conclusion: Both statin and fibrate reduce the potency of atherogenic sd-LDL particles, but via dif- ferent mechanisms: the former decreases total-LDL including sd-LDL, while the latter decreases sd- LDL specifically.

J Atheroscler Thromb, 2007; 14:128-132.

Key words; Small dense LDL-cholesterol, Type 2 diabetes, Pitavastatin, Fenofibrate

tative determination of sd-LDL, however, and it re- Introduction quires a time-consuming assay consisting of overnight Small dense low-density lipoprotein (sd-LDL) has electrophoresis, staining, and destaining. Analytical ul- recently been highlighted as a new risk factor for coro- tracentrifugation, the standard technique for quantify- nary heart disease (CHD). Sd-LDL is predictive of risk ing sd-LDL7), also requires a long running time, and not only in Westerners1-3), but also in Japanese in spite the need for specialized equipment renders it unsuit- of their somewhat lower levels of LDL-cholesterol (C)4, 5). able for general clinical use. Our group recently estab- LDL particle size is usually measured by gradient gel lished a simple assay for sd-LDL-C using heparin-Mg electrophoresis (GGE) using non-denaturing polyacryl- precipitation followed by direct measurement of LDL- amide according to the method of Nichols, Krauss, and C8, 9). We previously reported that the sd-LDL-C level Musliner6). This GGE procedure provides no quanti- determined by this method was substantially elevated in subjects with CHD and type 2 diabetes10). Address for correspondence: Tsutomu Hirano, First Department of Internal Medicine, Showa University School of Medicine, Statin acts potently in reducing LDL-C and is 1-5-8 Hatanodai, Shinagawa-ku, Tokyo 142-8666, Japan. widely used to prevent CHD events. It remains unclear, E-mail: [email protected] however, whether statin can reduce sd-LDL-C by mech- Received: December 12, 2006 anisms similar to LDL-C. Sd-LDL particles have weak- Accepted for publication: January 25, 2007 er affinity to the LDL receptor than large buoyant (lb)- 128 Journal of Atherosclerosis and Thrombosis Vol.14, No.3 Lipid-Lowering Agents and Small Dense-LDL 129 Original Article

LDL particles11), and the residence time of sd-LDL suring the remainder by direct homogenous LDL-C particles is much longer than that of lb-LDL particles assay (LDL-EX, Denka Seiken). The sd-LDL-C level in the blood circulation 12). If the induction of LDL re- was identical to cholesterol in the denser LDL fraction ceptors by statin treatment stimulates the uptake of with a density (d) =1.044－1.063 g/mL8). Large buoy- lb-LDL particles more potently than the uptake of sd- ant (lb)-LDL-C was calculated by subtracting sd-LDL- LDL particles, as proposed, then statin treatment might C from LDL-C. We demonstrated an excellent rela- not reduce sd-LDL as effectively as lb-LDL. Plasma tionship between estimated lb-LDL-C and cholesterol TG levels exert a strong influence on LDL size1); hence, in the lighter LDL fraction with a density (d) =1.019－ we know that a potent TG-lowering agent such as fi- 1.044 g/mL8). Apolipoprotein (apo) AⅠ and B were brate will enlarge the LDL size13). Few studies, howev- measured by immunoturbidometry (Daiichi Pure Che- er, have examined the quantitative change of LDL sub- mical Co). LDL-C and HDL-C were measured by di- species by fibrate treatment. For these reasons, we tried rect assay using commercially available kits (LDL-EX to determine the effects of statin and fibrate on the and HDL-EX; Denka Seiken Co.). concentration of LDL subspecies in patients with type The paired Student’s t-test was used to assess the 2 diabetes, one of the representative diseases with ele- significance of differences between before and after vated sd-LDL-C. lipid-lowering treatments. Correlations between two variables were calculated by Pearson’s simple linear regression analysis. Multiple linear regression analysis was Methods performed to evaluate the independent influence of Seventy-two patients (30 men and 42 women) changes of LDL-C, apoB, TG or HDL-C on the chang- with type 2 diabetes, aged 60.2±9.6, were enrolled in es of sd-LDL-C by lipid-lowering treatments. Statisti- this study. Dietary therapy was supervised by a dieti- cal significance was accepted at p＜0.05. tian and exercise therapy was prescribed by a physician according to the recommendations of the Japanese Di- Results abetes Association. Patients were treated with diet therapy alone (n =11), sulfonylureas (n =12), alpha-gluco- Table 1 (left panel) shows the serum levels of var- sidase inhibitors (n =21), metoformin (n =11), insulin ious parameters before and three months after treat- (n =2), and combinations of oral hypoglycemic agents ments with pitavastatin (left panel) and fenofibrate (n =15). All hypoglycemic agents were continued dur- (right panel). Pitavastatin treatment brought about no ing the study without any changes in dose. Hypercho- changes in glycemic control determined by fasting glu- lesterolemic subjects (LDL-C ＞140 mg/dL at baseline) cose and HbA1c. LDL-C was significantly reduced received pitavastatin (1 mg) once daily for three months. from 171 to 125 mg/dL in response to treatment, apo Hypertriglyceridemic subjects (TG ＞150 mg/dL at base- B was significantly reduced while TG was only slightly line) received micronized fenofibrate (100 mg) once dai- reduced, HDL-C was unchanged, whereas apo AⅠ was ly for three months. Patients with combined hyperlipid- slightly elevated, Sd-LDL-C was significantly reduced emia were given either pitavastatin or fenofibrate, based from 37 to 25 mg/dL and Lb-LDL-C was significantly on the severity of hypercholesterolemia or hypertriglyc- decreased from 134 to 99 mg/dL. The treatment brou- eridemia. There was no significant difference between ght about no change in the percentage of sd-LDL-C statin and fibrate groups in the method used for glyce- in total LDL-C (22 to 20%). mic control. Fasting serum samples were collected be- Fenofibrate treatment induced no changes in FPG fore and three months after the commencement of lip- or HbA1c: the levels of both were comparable to those id-lowering therapy. Informed consent was obtained measured in the statin group. Plasma TG was markedly from all subjects, and this study was approved by the reduced from 352 to 179 mg/dL by fibrate treatment, local ethics committee. whereas LDL-C was essentially unchanged, HDL-C Sd-LDL-C was measured using a commercially was significantly increased from 49 to 55 mg/dL, whe- available test kit (sd-LDL-C “Seiken”, Denka Seiken, reas apo AⅠ was unchanged, apo B was significantly Tokyo, Japan). The principle of this method has been reduced, Sd-LDL-C was significantly reduced from 45 described in detail elsewhere8, 9). When serum (0.2 to 33 mg/dL. In contrast, Lb-LDL-C was increased mL) was incubated with heparin-magnesium precipi- marginally from 90 to 99 mg/dL. The percentage of tation agents, all of the apoB-containing lipoproteins sd-LDL-C in total LDL-C was significantly reduced but sd-LDL were aggregated, and the aggregate was by fibrate treatment (33 to 25%, p＜0.001). trapped in a filter by centrifugation (5000 rpm for 1 Fig.1 shows the percent changes in TG, total- min). sd-LDL-C was obtained by subsequently mea- LDL-C, sd-LDL-C and lb-LDL-C after pitavastatin 130 Tokuno et al. Lipid-Lowering Agents and Small Dense-LDL 131

Table 1. Serum levels of various parameters before and 3 months after treatment with 1 mg of pitavastatin or 100 mg of micronized fenofibrate in patients with type 2 diabetes Pitavastatin Fenofibrate pre post p pre post p Glucose 143±52 155±66 NS Glucose 159±82 140±62 NS HbA1c 6.8±1.5 6.8±1.2 NS HbA1c 6.7±1.5 6.7±1.6 NS LDL-C 171±40 125±34 ＜0.0001 LDL-C 135±36 132±35 NS TG 206±159 160±91 ＜0.02 TG 352±230 179±84 ＜0.001 HDL-C 58±13 59±15 NS HDL-C 49±12 55±13 ＜0.005 ApoA1 142±21 149±26 ＜0.01 ApoA1 140±26 147±28 NS ApoB 132±25 101±23 ＜0.0001 ApoB 118±22 103±24 ＜0.0005 sd-LDL-C 37±20 25±15 ＜0.0001 sd-LDL-C 45±21 33±19 ＜0.01 lb-LDL-C 134±39 99±29 ＜0.0001 lb-LDL-C 90±34 99±31 NS

Data represent the mean±SD. NS＝not significant (p＞0.05), Pre＝pre-treatment, Post＝post-treatment

% % pre 140 160 pre post 120 post 140

＊＊ 120 100 ＊＊＊ 100 80 80 60 60 40 40

20 20

0 0 TG total-LDL sd-LDL lb-LDL TG total-LDL sd-LDL lb-LDL Fig.1 Fig.2 Percent changes in TG, total-LDL-C, small dense (sd)-LDL-C and Percent changes in TG, total-LDL-C, sd-LDL-C and lb-LDL-C large buoyant (lb)-LDL-C after pitavastatin treatment (1 mg/d). after micronized fenofibrate treatment (100 mg/d). ＊p＜0.01 vs ＊p＜0.01 vs pre-treatment. pre-treatment. treatment. TG was not decreased, whereas total-LDL- Table 2. Correlations between changes of sd-LDL-C and chan- C, lb-LDL-C and sd-LDL-C all decreased by similar ges of various parameters before and 3 months after treatment with 1 mg of pitavastatin or 100 mg of mi- amounts. cronized fenofibrate in patients with type 2 diabetes Fig.2 demonstrates the changes in TG, total- LDL-C, sd-LDL-C and lb-LDL-C after fenofibrate Pitavastatin Fenofibrate treatment. TG was decreased by 38%, whereas LDL-C r p r p was unchanged, Sd-LDL-C was decreased by 23%, whereas lb-LDL-C was unchanged. Glucose 0.23 NS Glucose 0.01 NS The left panel of Table 2 shows the correlation HbA1c 0.07 NS HbA1c 0.15 NS coefficients between the change of sd-LDL-C and those TG 0.24 NS LDL-C 0.04 NS of serum parameters during pitavastatin treatment. The HDL-C －0.20 NS TG 0.60 ＜0.002 reduction of sd-LDL-C by pitavastatin treatment was LDL-C 0.36 ＜0.02 HDL-C －0.46 ＜0.05 significantly associated with the reductions of LDL-C lb-LDL-C 0.02 NS lb-LDL-C 0.15 NS and apo B. Multiple regression analysis revealed that ApoA1 0.04 NS ApoA1 0.38 NS the reduction of sd-LDL-C was significantly correlated ApoB 0.48 ＜0.001 ApoB 0.26 NS 130 Tokuno et al. Lipid-Lowering Agents and Small Dense-LDL 131

with the reduction of LDL-C and apo B, respectively. result, fenofibrate substantially lowered TG and ele- The right panel of Table 2 shows the correlations vated HDL-C even at the low dose administered. Lb- between the change of sd-LDL-C and those of serum LDL was unchanged by fibrate treatment, whereas sd- parameters during fenofibrate treatment. The reduc- LDL-C was markedly decreased. When TG-rich lipo- tion of sd-LDL-C by fenofibrate treatment was sub- proteins are increased, LDL particles become TG-en- stantially associated with the reduction of TG and the riched and cholesterol in the blood is depleted by the increase of HDL-C, but was not associated with the action of cholesteryl ester transfer protein1, 12). Sd- reduction of LDL-C. Multiple regression analysis re- LDL particles are generated from TG-rich and choles- vealed that the reduction of sd-LDL-C was significant- terol-poor LDL particles by the action of hepatic TG ly correlated with the reduction of TG, but not with lipase1, 7). Fibrate likely suppresses lipid exchange be- the increase of HDL-C. tween TG-rich lipoprotein and LDL via the marked Creatinine kinase and laboratory examinations reduction in TG-rich lipoproteins. TG-rich large related to liver function were essentially unchanged by VLDL (VLDL1) is proposed to be a precursor of sd- pitavastatin and fenofibrate treatments (data not shown). LDL, while smaller VLDL (VLDL2) is proposed to be a precursor of lb-LDL12). Fibrate may reduce sd-LDL particles by suppressing the production of VLDL1 Discussion and stimulating the transfer from VLDL1 to VLDL2 Pitavastatin is a newly developed statin now wide- via the activation of lipoprotein lipase17). In an earlier ly used in Japan and many other countries. The usual study, we reported the enlargement of LDL by anoth- dose is 2 mg/day, and the maximal dose is 4 mg/day 14). er fibrate, bezafibrate, and proposed that the change Preliminary studies have revealed that 2 mg of pitavas- was wholly attributable to the TG-lowering effect of tatin confers about the same cholesterol-lowering ef- the agent18). Similarly, the present study confirmed an fect as 10 mg of atorvastatin. In the present study, we association between the decrease of sd-LDL-C and the examined the effect of a relatively low dose of pitavas- decrease of TG during fenofibrate treatment. Though tatin (1 mg) on LDL subspecies. Even at the low dose we did not measure the actual changes of the LDL size administered, pitavastatin decreased LDL-C by 25% in this study, the specific decrease in sd-LDL-C with- (significant). The low dose also reduced sd-LDL-C and out affecting lb-LDL-C implied that the average LDL lb-LDL by comparable levels, and the former reduc- size was increased. tion was significantly correlated with the pitavastatin- We finally determined that low doses of statin induced reduction of LDL-C. These results suggest and fibrate were both effective in decreasing sd-LDL- that the induction of LDL receptors by statin stimu- C concentration, and by comparable rates of reduc- lates the uptake of all LDL particles, irrespective of tion. We should bear in mind, however, that the base- their size. Another study found that 2 mg of pitavas- line sd-LDL-C level and the phenotype of hyperlipid- tatin significantly increased the LDL size15). It may be emia differed significantly between the statin- and fi- that the potent induction of LDL receptors by a high- brate-treated groups. Though we cannot readily con- dose of pitavastatin preferentially facilitates the remov- clude which drug reduces sd-LDL more potently, we al of sd-LDL particles. Pitavastatin also appears to re- can expect substantial reductions in sd-LDL-C when duce the TG level with a 2 mg dose 14, 15), and its hypo- statin is administered to hypercholesterolemic subjects triglyceridemic action was confirmed even at the low- and when fibrate is administered to hypertriglyceride- er dose of 1 mg administered in the present study. The mic subjects. Large-scale clinical trials will help to elu- ratio of sd-LDL in LDL-C was unchanged by pitavas- cidate whether the effects of statin and fibrate in re- tatin treatment, implying no change in LDL size. While ducing sd-LDL particles are associated with the pre- a lower TG is generally expected to increase the size of vention of CHD events. LDL particles, the mild reduction in TG by low-dose pitavastatin might be insufficient to bring about this effect. References There have been a number of reports that fibrate, 1) Berneis KK and Krauss RM: Metabolic origins and clini- a potent TG-lowering agent, enlarges LDL in patients cal significance of LDL heterogeneity. J Lipid Res, 2002; 13, 16) 43:1363-1379 with type 2 diabetes . Few reports, however, have 2) Lamarche B, Lemieux I, and Després JP: The small dense demonstrated the effects of fibrate on the concentra- LDL phenotype and the risk of coronary heart disease: tions of LDL subspecies. In the present study we used epidemiology, patho-physiology and therapeutic aspects. a low dose of fenofibrate (100 mg) as a suitable com- Diabetes Metabolism, 1999; 25:199-211 parison with the low dose of pitavastatin (1 mg). As a 3) Gardner CD, Fortmann SP, and Krauss RM: Association 132 Tokuno et al.

of small dense low-density lipoprotein particles with the density lipoproteins. J Biol Chem Dec, 1984; 259:14728- incidence of coronary artery disease in men and women. 14735 JAMA, 1996; 276:875-881 12) Packard CJ and Shepherd J: Lipoprotein heterogeneity and 4) Koba S, Hirano T, Yoshino G, Sakai K, Sakaue T, Adachi apolipoprotein B metabolism. Arterioscler Thromb Vasc M, and Katagiri T: Remarkably high prevalence of small Biol, 1997; 17:3542-3556 dense low-density lipoprotein in Japanese men with coro- 13) Yoshino G, Hirano T, and Kazumi T: Treatment of small nary artery disease, irrespective of the presence of diabe- dense LDL. J Atheroscler Thromb, 2002; 9:266-275 tes. Atherosclerosis, 2002; 160:249-256 14) Saito Y, Teramoto T, Yamada N, Itakura H, Hata Y, Naka- 5) Koba S, Hirano T, Kondo T, Shibata M, Suzuki H, Mu- ya N, Mabuchi H, Tsushima M, Sasaki J, Ogawa N, and rakami M, Geshi E, and Katagiri T: Significance of small Goto Y: Clinical efficacy of NK-104 (Pitavastatin), a new dense low-density lipoproteins and other risk factors in synthetic HMG-CoA reductase inhibitor, in the doswe find- patients with various types of coronary heart diseases. Am ing, double blind, three-group comparative study. Rishou Heart J, 2002; 144:1026-1035 Iyaku, 2001; 17:829-855 (Summary in English) 6) Nichols AV, Krauss RM, and Musliner TA: Nondenatur- 15) Sone H, Takahashi A, Shimano H, Ishibashi S, Yoshino G, ing polyacrylamide gradient gel electrophoresis. Methods Morisaki N, Saito Y, Kawazu S, Teramoto T, Fujita T, Shi- Enzymol, 1986; 128:417-433 ba T, Iwamoto Y, Kuzuya N, Akanuma Y, and Yamada N: 7) Zambon A, Hokanson JE, Brown BG, and Brunzell JD: HMG-CoA reductase inhibitor decreases small dense low- Evidence for a new pathophysiological mechanism for cor- density lipoprotein and remnant-like particle cholesterol onary artery disease regression: hepatic lipase-mediated in patients with type-2 diabetes. Life Sci, 2002; 71:2403- changes in LDL density. Circulation, 1999; 99:1959-1964 2412 8) Hirano T, Ito Y, Saegusa H, and Yoshino G: A novel and 16) Vakkilainen J, Steiner G, Ansquer JC, Aubin F, Rattier S, simple method for quantification of small dense LDL. J Foucher C, Hamsten A, and Taskinen MR, on behalf of Lipid Res, 2003; 44:2193-2201 the DAIS group: Relationships between low-density lipo- 9) Hirano T, Ito Y, and Yoshino G: Measurement of small protein particle size, plasma lipoproteins, and progression dense low-density lipoprotein particles. J Atheroscler Thr- of coronary artery disease. The Diabetes Atherosclerosis omb, 2005; 12:67-72 Intervention Study (DAIS). Circulation, 2003; 107:1733- 10) Hirano T, Ito Y, Koba S, Toyoda M, Ikejiri A, Saegusa H, 1737 Yamazaki J, and Yoshin G: Clinical significance of small 17) Milosavljevic D, Griglio S, Le Naour G, and Chapman dense low-density lipoprotein cholesterol levels determined MJ: Preferential reduction of very low density lipoprotein-1 by the simple precipitation method. Arterioscler Thromb particle number by fenofibrate in type ⅡB hyperlipide- Vasc Biol, 2004; 24:558-563 mia: consequences for lipid accumulation in human mono- 11) Bradley WA, Hwang SL, Karlin JB, Lin AH, Prasad SC, cyte-derived macrophages. Atherosclerosis, 2001; 155:251- Gotto AM Jr, and Gianturco SH: Low-density lipopro- 260 tein receptor binding determinants switch from apolipo- 18) Hirano T, Kazumi T, and Yoshino G: Long-term efficacy protein E to apolipoprotein B during conversion of hy- of bezafibrate on reducing small dense LDL by its hypo- pertriglyceridemic very-low-density lipoprotein to low- triglyceridemic action. Curr Ther Res, 2000; 61:127-136