Antioxidant Effects of Calcium Antagonists Rat Myocardial

223

Antioxidant Effects of Calcium Antagonists ['Ii] Rat Myocardial Membrane Lipid Peroxidation

Hitoshi Sugawara, Katsuyuki Tobise, and Kenjiro Kikuchi

We studied the antioxidant effects of nine calcium antagonists (nisoldipine, benidipine, nilvadipine, felodipine, nicardipine; nitrendipine, nifedipine, verapamil, and diltiazem) by means of rat myocardial membrane lipid peroxidation with a nonenzymatic active oxygen-generating system (DHF/FeC13-ADP). The order of antioxidant potency of these agents was nilvadipine > nisoldipine > felodipine > nicardipine > verapamil > benidipine. Their IC50 values (,uM) were 25.1, 28.2, 42.0, 150.0, 266.1, and 420.0, respectively. In contrast, nitrendipine, nifedipine, and diltiazem had little inhibitory effect on lipid peroxidation.These six calcium antagonists could be divided into four types on the basis of their antioxidant mechanisms. Nilvadipine, nisoldipine, and verapamil, which showed antioxidant effects both before and after the addition of active oxygen, and reduced the dihydroxyfumarate (DHF) auto-oxidation rate, were chain-breaking and preventive antioxidants. Felodipine, which showed antioxidant effects both before and after exposure to active oxygen and increased the DHF auto-oxidation rate, was only a chain-breaking antioxidant. Nicardipine, which showed an antioxidant effect only before exposure to active oxygen and reduced the DHF auto-oxidation rate, was mainly a preventive antioxidant. Benidipine, which showed an antioxidant effect only before exposure to active oxygen and had no appreciable effect on the DHF auto-oxidation rate, could interrupt the chain reaction of lipid peroxidation at the initial step alone. Although these results suggest that the antioxidant properties of some calcium antagonists may be beneficial clinically in protecting against cellular damage caused by lipid peroxidation, further studies are required to establish the antioxidant effects of these agents in vivo. (Hypertens Res 1996; 19: 223-228) Key Words: antioxidant effect, calcium antagonist, dihydropyridine calcium antagonists, myocardial membrane lipid peroxidation

Cellular damage due to lipid peroxidation associ- there are conflicting reports with regard to nifedi- ated with the production of active oxygen species pine, one of the most widely used dihydropyridine plays an important role in various pathologic states, calcium antagonists. Janero et al. found that nisoldi- such as ischemia-reperfusion injury (1), coronary pine was the most potent dihydropyridine antiox- arteriosclerosis (2), glomerular sclerosis (3), di- idant, whereas nifedipine had little effect as com- abetes mellitus (1), and the process of "normal" ag- pared with other dihydropyridine calcium antagon- ing (4). Calcium antagonists have been established ists in studies using a xanthine oxidase-dependent to be safe and effective in the management of active oxygen-generating system and liposomes de- essential hypertension and angina pectoris. Since rived from rat cardiac membrane phospholipids (6). treatment with anti-hypertensive and anti-anginal In contrast, Mak and Weglicki found that nifedipine agents is usually long term, additional protective had the highest antioxidant activity among all cal- effects against lipid peroxidation by oxygen radicals cium antagonists tested with an active oxygen- may be beneficial in preventing organ damage. generating system using dihydroxyfumarate (DHF) However, it is unclear whether the calcium antago- auto-oxidation and isolated sarcolemma of the nists have such antioxidant effects and, if so, what canine heart (7). The results of our previous report mechanisms are involved. (8) confirmed those of Janero et al. (6) and indi- Janero et al. first reported a direct effect of cal- cated that the membrane peroxidation reaction mix- cium antagonists on the lipid peroxidation of car- ture used by Mak and Weglicki (7) may be unsuit- diac membranes (5). They found that bepridil and able due to the generation of sucrose-derived prenylamine were the most potent antioxidants thiobarbituric acid-reactive substances (TBARS) among the calcium antagonists tested. However, rather than lipid-derived TBARS.

From the First Department of Internal Medicine, Asahikawa Medical College, Asahikawa, Hokkaido, Japan. Address for Reprints: Hitoshi Sugawara, M.D., Ph.D., Omiya Medical Center, Jichi Medical School, 1-847 Amanuma- cho, Omiya, Saitama 330, Japan. Received September 8, 1995; accepted in revised form July 15, 1996. 224 Hypertens Res Vol. 19, No. 4 (1996)

To determine the antioxidant effects and mechan- pellet was resuspended in 10 mM Tris-HC1 buffer ( isms of antioxidant action of nine calcium antagon- pH 7.4) and adjusted to a protein concentration of ists (nisoldipine, benidipine, nilvadipine, felodipine, 1-2 mg/ml and stored at - 80°C. Protein concentra- nicardipine, nitrendipine, nifedipine, verapamil and tions were determined according to the method of diltiazem) we studied rat myocardial membrane Lowry et al. (10). lipid peroxidation by means of a nonenzymatic active oxygen-generating system. We were able to de- Active Oxygen-Generating System monstrate differences in antioxidant potency among A nonenzymatic active oxygen-generating system in- calcium antagonists and to classify the mechanisms volving the auto-oxidation of DHF/Fe3+ -ADP was of the antioxidant effects of these agents into four used as described by Mak et al. (11). After the types. auto-oxidation of DHF, the resultant superoxide an- ions produced hydroxyl radicals in a reaction cataly- zed by Fe3+-ADP and initiated a lipid peroxidation Methods chain reaction (12). Animals Seven-week-old male Sprague-Dawley rats were Experimental Protocols purchased from Charles River Laboratories, Inc. Experiment 1: Under amber lighting provided by a (Tokyo, Japan) and housed for one week at the sodium lamp, 100 g of myocardial membrane pro- animal laboratory of Asahikawa Medical College tein was incubated and shaken in 10 mM Tris-HC1 before use in this study. buffer (pH 7.4) for 30 min at 37°C in a warm water bath with each agent at a final concentration of 1 Materials mM. After adding 6.66 mM DHF, 1 mM ADP, and Dihydroxyfumarate (DHF) was purchased from 0.1 mM FeC13 (in a total volume of 0.5 ml), the Aldrich Chemical Co . (Milwaukee, WI). Adenosine mixtures were incubated and shaken in a warm wa- diphosphate (ADP), nifedipine, nicardipine, and ter bath at 37°C for a further 120 min, and the ex- verapamil were purchased from Sigma Chemical tent of lipid peroxidation was measured. As shown Co. (St. Louis, MO). Nisoldipine was provided by previously (8), the IC50 value was determined for Bayer Yakuhin Ltd. (Osaka, Japan), benidipine by all agents that inhibited lipid peroxide production Kyowa Hakko Kogyo Co., Ltd. (Tokyo, Japan), significantly as compared with control. nilvadipine by Fujisawa Pharmaceutical Co., Ltd. Experiment 2: Subsequent experiments were per- (Osaka, Japan), felodipine by Hoechst Japan Ltd. ( formed with the agents that inhibited lipid peroxida- Tokyo, Japan), nitrendipine by Yoshitomi Phar- tion significantly as compared with control. Under maceutical Industry Ltd. (Osaka, Japan), and dil- similar conditions as experiment 1, 100 g of mem- tiazem by Tanabe Seiyaku Co., Ltd. (Osaka, brane protein in 10 mM Tris-HC1 buffer (pH 7.4) Japan). The other agents used were purchased from was preincubated and shaken in a warm water bath Nacali Tesque Inc. (Tokyo, Japan). Concentrated at 37°C for 30 min, and 6.66 mM DHF and 0.1 mM stock solutions of nisoldipine, benidipine, nilvadi- FeCl3-1 mM ADP were added. After 30 min, the pine, felodipine, nicardipine, nitrendipine, nifedi- test agent (1 mM) was added in a total volume of pine, verapamil, or diltiazem were diluted with 0.5 ml, the mixture was incubated and shaken in a ethanol before use, so that the final solvent concen- warm water bath at 37°C for a further 90 min, and tration in the reaction system did not affect lipid the extent of lipid peroxidation was measured. peroxidation. Water was purified with a Milli-Q Plus filter (Millipore, Tokyo, Japan) before use. Determination o f Lipid Peroxidation Lipid peroxide production was measured as thiobar- Preparation of Crude Myocardial Membranes bituric acid-reactive substances (TBARS) using the Eight-week-old rats were sacrificed by decapitation. improved thiobarbituric acid (TBA) test of Ogura et Their hearts were isolated and washed with phy- al. (13). Although the results of this test may not be siological saline. The aorta, atria, and fatty tissue related directly to actual lipid peroxide production were removed, and the ventricles were frozen in li- (14), we defined TBARS (nmol) as the malondial- quid nitrogen and stored at - 80°C. All subsequent dehyde (MDA) equivalent of lipid peroxide produc- procedures were conducted at 4°C. With the use of tion based on data obtained with a standard solu- the method of Wagner et al., as modified by Kane- tion of 1,1,3,3-tetraethoxypropane (9.487 nmol/ml). ko et al. (9), the crude myocardial membrane frac- After the last incubation in a warm water bath, tion was prepared from the heart tissue without uti- 10x1 of 2% (W/V) butylated hydroxytoluene (BHT) lizing chelating agents such as EDTA or a sucrose in methanol was immediately added to stop the gradient. Briefly, the myocardium was minced in 50 reaction. Then, 200pl of 10% (W/V) sodium mM Tris-HC1 buffer (pH 7.4), homogenized for 20 s dodecyl sulfate, and 3 ml of the TBA reagent were with a Polytron PTA 10S homogenizer (Brinkman added. The TBA reagent was a mixture of equal Inst., Inc.) on setting No.S, and then centrifuged at volumes of 0.67% (W/V) TBA aqueous solution 1,000 X g for 10 min. The supernatant obtained was and glacial acetic acid. Next, the mixture was heat- centrifuged at 48,000 X g for 25 min. The resultant ed for 30 min at 90°C and subsequently chilled in pellet was then resuspended in the same buffer and ice water for 10 min, after which 3 ml of chloroform was centrifuged again at 48,000 X g for 25 min. This was added while stirring. After centrifugation at procedure was repeated twice, after which the final 1650 X g, the absorbance of the supernatant was de- Sugawara et al: Antioxidant Effects of Calcium Antagonists 225 termined at 532 nm using a UV-160 Shimadzu and felodipine but not for benidipine and nicardi- ultraviolet-visible light spectrophotometer (Kyoto, pine. Japan). Effect of Calcium Antagonists on DHF Auto-Oxida- Determination of DHF Auto-Oxidation tion (Fig. 3) The DHF auto-oxidation rate was determined by The DHF auto-oxidation rate (nmol/min) tested in the nitroblue tetrazolium (NBT) method (12) for all the presence of each agent was as follows: control, agents that inhibited lipid peroxidation production 3.37; nicardipine, 1.09; nilvadipine, 1.73; verapamil, significantly as compared with control. NBT (50 1.73; nisoldipine, 1.76; benidipine, 2.51; and felodi- NM), the test agent, and 3.33 mM DHF were added pine, 6.15. The DHF auto-oxidation rate was re-d to 10 mM Tris-HC1 buffer (pH 7.4) in a total reac- uced by nicardipine, nilvadipine, verapamil, and tion volume of 2 ml, and the change in absorbance nisoldipine. This rate was virtually unchanged by at 560 nm was determined at 37°C using a Hitachi benidipine and increased by felodipine. 557 double-wavelength spectrophotometer (Tokyo, Japan). The DHF auto-oxidation rate was then calculated from the NBT reduction rate using the Discussion Lambert-Beer formula and a molecular absorbance The present study, clearly demonstrated con- coefficientof 17.9 X 103M-1 cm-1 (12). centration-dependent anti-lipid peroxidant effects in rat myocardial membrane for six calcium Statistical Differences antagonists: nisoldipine, verapamil, benidipine, nil- Data were expressed as mean ± SD. For statistical vadipine, felodipine, and nicardipine. No such analysis, the outlier test, analysis of variance, and effects were obtained with nitredipine, nif edipine, multiple range tests were used. A p value of 0.05 or or diltiazem (Fig. 1). The order of antioxidant less was considered to indicate statistical signi- potency of the six agents with antioxidant activity ficance. was nilvadipine > nisoldipine > felodipine > nicardipine > verapamil > benidipine. The antioxidant potency of these calcium antagonists was unrelated Results to calcium channel blocking intensity. Antioxidant Effects of Nine Calcium Antagonists on Nisoldipine, verapamil, nilvadipine, and felodi- Rat Myocardial Membrane Lipid Peroxidation (Fig. pine were found to inhibit lipid peroxidation 30 min 1) after the addition of DHF/Fe3+-ADP (Fig. 2). The The extent of lipid peroxidation (nmol/mg protein) DHF auto-oxidation rate was decreased by nicardi- in the presence of each agent (1 mM) was as fol- pine, nilvadipine, verapamil, nisoldipine, but in- lows: control, 53.4 ± 3.21; nisoldipine, 1.57 ± creased by felodipine. Benidipine had virtually no 0.62; verapamil, 1.82 ± 0.55; benidipine, 2.36 ± effect on the DHF auto-oxidation rate (Fig. 3). We 0.58; nilvadipine, 3.64 ± 1.90; felodipine, 4.18 ± previously reported that lipid peroxidation in the 1.28; nicardipine, 17.1 ± 15.7; nitrendipine, 38.5 ± same experimental system was inhibited by SOD, 2.39; nifedipine, 46.3 ± 8.98; and diltiazem, 54.9 catalase and a-tocopherol and that the DHF auto- ± 4.98. Nisoldipine, verapamil, benidipine, nilvadi- oxidation rate was decreased by SOD and catalase, pine, felodipine, and nicardipine significantly inhi- but not by a-tocopherol (8). The mechanisms of the bited lipid peroxidation in a concentration-depen- antioxidant effects of these calcium antagonists dent manner, and their IC50 values (pM) were 28.2, could be classified into four types on the basis of 266. 420.0, 25 42.0, and 150.0, respectively. these findings. Nisoldipine, verapamil, and nilvadi- The order of antioxidant potency of these agents pine, which showed antioxidant effects both before was nilvadipine > nisoldipine > felodipine > nicardi- and after exposure to active oxygen and reduced pine > verapamil > benidipine. In contrast, no signi- the DHF auto-oxidation rate, had both a preventive ficant inhibitory effect on lipid peroxidation was antioxidant effect (6, 15) similar to SOD or cata- shown for nitrendipine, nifedipine, and diltiazem. lase, and a chain-breaking effect (6, 15) simllar to a-tocopherol. Felodipine, which showed antioxidant Effect of Calcium Antagonists on Lipid Peroxidation effects both before and after exposure to active ox- 30 Min After the Addition of DHF/Fe3+-ADP. (Fig. ygen and increased the DHF auto-oxidation rate, 2) was only a chain-breaking antioxidant. Nccardipine, When each agent was applied to the crude myocar- which showed an antioxidant effect only during dial membrane fraction 30 min after the addition of preincubation with myocardial membrane before DHF/Fe3+-ADP, the extent of lipid peroxidation the addition DHF/Fe3+-ADP exposure, and re- (nmol/mg protein) was as follows: control A (total duced the DHF auto-oxidation rate, was mainly a incubation time, 30 min), 11.1 ± 5.40; control B preventive antioxidant. Benidipine, which showed (total incubation time, 120 min), 49.0 ± 14.0; nisol- antioxidant activity only before exposure to active dipine, 5.20 ± 1.06; verapamil, 5.95 ± 1.44; nilva- oxygen, and had no appreciable effect on the DHF dipine, 7.96 ± 4.70; felodipine, 9.43 ± 1.78; be- auto-oxidation rate, could interrupt lipid peroxida- nidipine, 40.7 ± 6.13; and nicardipine, 56.8 ± tion at the initial step alone. Electron spin reso- 13.1. The inhibitory effects of the agents 30 min af- nance demonstrate that the scavenged free radical ter the addition of DHF/Fe3+-ADP were clearly de- species may differ among the four types of calcium monstrated for nisoldipine, verapamil, nilvadipine, antagonists. The distribution of these calcium anta- 226 Hypertens Res Vol. 19, No. 4 (1996)

Fig. 1. Comparative antioxidant effects of calcium antagonists on lipid peroxidation in rat myocardial membranes. Values are the mean ± SD of 5-11 samples. Closed bars Fig. 3. Effects of calcium antagonists on DHF auto-ox- show p<0.0.5 as compared with control. idization measured by NBT reduction rate. Values show the means of two experiments.

than those in normal subjects. We considered structural requirements for the antioxidant activity of dihydropyridine calcium antagonists. Our data and those of Janero et al. (6) are summarized in Table 1. Five of the most potent dihydropyridine antioxidants (nilvadipine, nisoldipine, felodipine, nicardipine, and benidipine) had both a methyl ester and a hydrophobic ester at C-3 or C-5 of the dihydropyridine ring. The structural difference between felodipine and nitrendipine sug- gested that substitutions on the phenyl ring may determine antioxidant potency (6). Janero et al. reported that there was no absolute correlation between the antioxidant potency of calcium antagonists and their lipophilicity as expressed by the partition coefficient. Therefore, they men- tioned that lipophilicity may be only one factor con- Fig. 2. Antioxidant effects of calcium antagonists 30 min tributing to the antioxidant effect of these agents (5, after addition of DHF/ FeCl3-ADP. Control A, incubation 6). Similarly, our experimental data showed that be- for 30 min after the addition of DHF/FeCl3-ADP. Control nidipine had the highest partition coefficient in the B, incubation for 120 min after the addition of octanol/buffer system and lipid membrane (18), DHF/FeCl3-ADP, as previously shown in Fig. 1. Values although the potency of the antioxidant activity of are the mean ± SD of 5-10 samples. * p < 0.05 vs. Con- this agent was not high. We reconfirmed that lipid trol B. solubility dose not essentially contribute to the antioxidant effect of calcium antagonists (5, 6, 8). Although our present results suggest that the antioxidant properties of some calcium antagonists gonists in the myocardial membrane fraction may may be beneficial clinically in protecting against differ, because these antioxidants interact with free cellular damage caused by the lipid peroxidation, radicals at the same distributive phase. Since these further studies are necessary to establish the antiox- results were obtained under amber lighting provided idant effects of these agents in vivo. by a sodium lamp, our present experiment meets the requirement for darkness as recommended by Rojstaczer et al. (16). Conclusions The antioxidant effects of these agents were The present results show that the order of antiox- observed for concentrations at least 10-fold higher idant potency of the calcium antagonists tested was than therapeutic serum concentrations. These nilvadipine > nisoldipine > felodipine > nicardipine > agents may inhibit lipid peroxidation if they verapamil > benidipine. Nitrendipine, nifedipine, accumulate in cardiac membranes in sufficiently and diltiazem had little inhibitory effect on lipid higher concentrations than in serum in vivo (17). In peroxidation. The mechanisms of the antioxidant some pathological states, concertrations of calcium effects of these calcium antagonists were classified antagonists in cardiac membranes may be higher into four types. Nllvadipine, nisoldipine, and vera- Sugawara et al: Antioxidant Effects of Calcium Antagonists 227

Table 1. Comparative Antioxidant Potency and Chemical Structure of Dihydropyridine Calcium Antagonists

pamil were chain-breaking and preventive antiox- The authors wish to express their thanks to Honorary idants. Felodipine was only a chain-breaking antiox- Professor Sokichi Onodera, Asahikawa Medical College, idant. Nicardipine was mainly a preventive antiox- for his support. We also thank Dr. Tomohiro Kurosaki idant. Benidipine interrupted the initial chain reac- and Mr. Harry A. Friedman for critical reading of this tion of lipid peroxidation. manuscript. Although these results suggest that the antioxidant properties of some calcium antagonists may be References beneficial clinically, further studies are required to establish if calcium antagonists have antioxidant 1 Sugawara H, Tobise K, Minami H, Uekita K, Yoshie effects in vivo. H, Onodera S: Diabetes mellitus and reperfusion injury increase the level of active oxygen-induced lipid peroxidation in rat cardiac membranes. J Clin Exp Acknowledgements Med (Igaku no Ayumi) 1992; 163: 237-238. 2. Kok FJ, van Poppel G, Melse J, et al: Do antiox- Nisoldipine, benidipine, nilvadipine, felodipine, nitrendi- idants and polyunsaturated fatty acids have a com- pine, and diltiazem were kindly supplied by Bayer Yaku- bined association with coronary atherosclerosis? hin Ltd. (Osaka, Japan), Kyowa Hakko Kogyo Co., Ltd. Atherosclerosis 1991; 31: 85-90. (Tokyo, Japan), Fujisawa Pharmaceutical Co., Ltd. (Osa- 3 Alfrey AC: Toxicity of tuble fluid iron in the nephro- ka, Japan), Hoechst Japan Ltd. (Tokyo, Japan), Yoshito- tic syndrome. Am J Physiol 1992; 263: F637-F641. mi Pharmaceutical Industry Ltd. (Osaka, Japan), and 4 Ji LL, Dillon D, Wu E: Myocardial aging: antiox- Tanabe Seiyaku Co., Ltd. (Osaka, Japan), respectively. idant enzyme systems and related biochemical prop- 228 Hypertens Res Vol. 19, No. 4 (1996)

erties. Am J Physiol 1991; 261: R386-R392. 12. Kukreja RC, Okabe E, Schrier GM, Hess ML: Ox- 5 Janero DR, Burghardt B, Lopez R: Protection of ygen radical-mediated lipid peroxidation and inhibi- cardiac membrane phospholipid against oxidative in- tion of Cat+-ATPase activity of cardiac sarcoplasmic jury by calcium antagonists. Biochem Pharmacol reticulum. Arch Biochem Biophys 1988; 261: 447- 1988; 37: 4197-4203. 457. 6 Janero DR, Burghardt B: Antiperoxidant effects of 13. Ogura R, Sakanashi T, Nagata 0, et al: Assay for dihydropyridine calcium antagonists. Biochem Phar- lipid peroxide content in mitochondria by the macol 1989; 38: 4344-4348. thiobarbituric acid reaction. Kurume Med J 1987; 34: 7 Mak IT, Weglicki WB: Comparative antioxidant acti- 53-58. vities of propranolol, nifedipine, verapamil, and dil- 14. Janero DR: Malondialdehyde and thiobarbituric acid- tiazem against sarcolemmal membrane lipid peroxida- reactivity as diagnostic indices of lipid peroxidation tion. Circ Res 1990; 66: 1449-1452. and peroxidative tissue injury. Free Radical Biol Med 8 Sugawara H, Tobise K, Onodera S: Absence of anti- 1990; 9: 515-540. oxidant effects of nifedipine and diltiazem on 15. Niki E: Defense system against oxygen toxicity in myocardial membrane lipid peroxidation in contrast vivo. Kikan Kagaku Sosetsu 1990; 7: 177-190 (in with those of nisoldipine and propranolol. Biochem Japanese). Pharmacol 1994; 47: 887-892. 16. Rojstaczer N, Triggle DJ: Calcium channel antagon- 9 Kaneko M, Lee S-L, Wolf CM, Dhalla NS: Reduc- ists as antioxidants. Cardiovasc Drug Rev 1994; 12: tion of calcium channel antagonist binding sites by 70-84. oxygen free radicals in rat heart. J Mol Cell Cardiol 17. Aruoma 01, Smith C, Cecchini R, Evans PJ, Hal- 1989; 21, 935-943. liwell B .: Free radical scavenging and inhibition of 10. Lowry OH, Rosebrough NJ, Farr AL, Randall RJ: lipid peroxidation by /9-blockers and by agents that Protein measurement with the folin phenol reagent. J interfere with calcium metabolism. A physiologically- Biol Chem 1951; 193: 265-275. significant process? Biochem Pharmacol 1991; 42: 11. Mak IT, Misra HP, Weglicki WB : Temporal relation- 735-743. ship of free radical-induced lipid peroxidation and 18. Nosaka C, Ishii A: Partition of [3H] benidipine hyd- loss of latent enzyme activity in highly enriched hepa- rochloride into the lipid membrane. The Clin Rep tic lysosomes. J Biol Chem 1983; 258: 13733-13737. 1991; 25: 3887-3893.