European Journal of Obstetrzcs & Gynecology and Reproductrve Biology. 37 (1990) 155-162 Elsevler
EUROBS
155
00980
Low-dose oral contraceptives lower plasma levels of apolipoprotein E Armin Steinmetz
‘, Katharina Bauer I, Ralph Jtirgensen and Hans Kaffarnik ’
’
’ Zenirum Innere Medirm, Endokrinologie and Stoffwechsel and ’ Insiitut ftir medrrrnrsche Sratrstlh, Philipps Umversitiir, D-3550 Marburg, F. R. G. Accepted
for publication
10 November
1989
Summary Three different oral contraceptive preparations were studied before and after a 3 month treatment period with respect to their effects on plasma lipoprotein parameters. A total of 58 healthy women requesting oral contraception were randomly assigned to three groups. Each woman received either monophasic preparations containing ethinylestradiol and desogestrel (M-DG); ethinylestradiol and gestodene (M-GD); or a triphasic preparation of ethinylestradiol and levonorgestrel (T-LN). As has been reported in other studies, the concentrations of total plasma cholesterol and apolipoproteins B and A-IV did not change significantly in any group. HDL cholesterol, triglycerides, apolipoproteins A-I and A-II increased or tended to increase. Despite the effects of the three hormone preparations on these lipoprotein parameters, however, each led to a highly significant decrease in apolipoprotein E plasma levels. Considering the recently reported observations that oral contraceptives increase the hepatic uptake of cholesterol-rich remnants, this decrease in apo-E plasma levels may in women that take oral contraceptives be directly correlated with increased hepatic lipoprotein metabolism. Oral contraceptive,
low dose; Plasma
levels; Apohpoprotein
E
Abbreviations: Apo, apolipoprotein; VLDL, very-low-density lipoprotein; LDL, low-density lipoprotem; IDL, intermediate density lipoprotein; HDL, high-density lipoprotein; PEG, poly(ethylene glycol); EE, ethinylestradiol; M-DG, ethinylestradiol and desogestrel; M-GD, ethinylestradiol and gestodene; T-LN, ethinylestradiol and levonorgestrel. Correspondence: A. Steinmetz, Zentrum Innere Medizin, Endokrinologie und Stoffwechsel, Philipps UniversltBt, Baldingerstrasse, D-3550 Marburg, F.R.G.
0028-2243/90/$03.50
Q 1990 Elsevier Science Publishers
B.V. (Biomedical
Division)
156
Introduction
Numerous studies have established relationships between the use of oral contraceptives and the changes in plasma lipoprotein levels (reviewed in Refs. 1 and 2). The high levels of the synthetic estrogen ethinylestradiol prescribed in the 1970s was later reduced to about 30 ~18 per day. This led to a decrease in the incidence of venous thrombembolism but did not seem to affect the incidence of myocardial infarction and stroke [3]. Although estrogens were originally thought to be responsible for the arterial pathology, attention was later focused on progestogens [4,5]. Their intrinsic androgenic activities were shown to increase LDL cholesterol, to lower HDL cholesterol levels [6,7] and thus to induce lipoprotein changes believed to be atherogenic [8,9]. Most studies conducted to establish the influence of oral contraceptives on lipoprotein metabolism have been concerned with the measurement of total plasma and lipoprotein lipids and apoproteins A-I, A-II and B. As yet no information is available on the influence of oral contraceptives on apolipoprotein E. We have therefore studied three different formulations of oral contraceptives currently used in Europe with respect to their influence on conventional lipid parameters and additionally apolipoprotein E. Besides ethinylestradiol, the preparations contained other selected progestogen components. Subjects and Methods
Subjects Healthy young women requesting oral contraceptives at a gynecological outpatient facility were examined for eligibility and then randomly allocated to three different treatment groups. Informed consent was obtained from all subjects. Criteria for entering the study were: age range 16-35 years, regular menstrual cycle of 25-36 days, absence of contra-indications to oral contraceptives, absence of significant endocrinologic, hematologic or metabolic disorders. No oral contracep-
TABLE
I
Dosage
of the different
contraceptive
preparations
used in the study
Abbreviation
Composition
M-DG
30 /.tg ethinylestradiol 150 pg desogestrel
l-21
30 pg ethinylestradiol 75 pg gestodene
l-21
M-GD
T-LN
30 50 40 75 30 125
pg gg pg pg pg pg
ethinylestradlol levonorgestrel ethinylestradiol levonorgestrel ethinylestradiol levonorgestrel
Day
l7-11 11-21
6
157
tives had been used or hormones taken for 3 months prior to the study. The women were requested not to change their eating habits. They received either monophasic formulations containing desogestrel (M-DG) or gestodene (M-GD) or a triphasic preparation containing levonorgestrel (T-LN). The dosages of the preparations given are summarized in Table I. The three groups did not differ significantly in age (M-DG, n = 23, 21.5 & 3.9 years, range 16-30; M-GD, n = 21, 20.1 + 4.6 years, range 16-35; T-LN, n = 15, 22.5 _t 4.2 years, range 16-35) or mean body mass index (21.1, 20.3 and 22.1, respectively). Levels of physical activity and alcohol intake were comparable between the three groups. The distribution of apolipoprotein E genetic phenotypes did not differ among the three treatment groups. Baseline measurements were performed in the luteal phase between days 18 and 22 of the cycle; blood samples were likewise taken between days 18 and 22 in the treatment period after three cycles. Lipoprotein measurements and apoprotein analysis Blood samples were drawn after an least 12 h of fasting between 8 and 10 a.m. and serum was obtained by low-speed centrifugation (4000 X g, 30 min) at 4” C. Aliquots for apoprotein determinations were stored at - 20 ’ C and analyzed within 4 months. For lipoprotein analysis, samples were kept at + 4” C and processed within 2 days. Plasma cholesterol and triglycerides were measured enzymatically, using the commercially available kits of Boehringer (Mannheim, F.R.G.). VLDL-, LDL- and HDL-cholesterol levels were determined by an electrophoretic procedure provided by Immuno Chemie (Heidelberg, F.R.G.). HDL cholesterol was in parallel measured enzymatically after precipitation of apo-B-containing lipoproteins, using the commercially available kit of Boehringer (Mannheim, F.R.G.). Apolipoproteins A-I, A-II and B were quantitatively measured by the use of radial immunodiffusion with kits from Immuno Chemie (Heidelberg, F.R.G.). Apolipoproteins E and A-IV were quantified by electroimmunoassay. Antibodies to human apo-E were raised as outlined before [lo]. The assay was performed in 1% agarose gels on 12 x 9 cm plates. The Verona1 electrophoresis buffer contained, in addition, 1% sodium deoxycholate, 0.5% Triton X-100 and 2.5% PEG 4000. Samples were first diluted 1:2 in a 2-fold concentrated preparation of this buffer and then further diluted an additional 3- and 6-fold in electrophoresis buffer. Electrophoresis was then carried out at 25 mA for 4 h. Purified apo-E was used as a primary standard and pooled frozen plasma as a secondary standard. The interassay coefficient of variation was 6.8% and the intraassay coefficient of variation was 3.7%. The assay for Apo A-IV has been previously described [ll]. Statistical analysis Concentrations of lipids, lipoproteins Mann-Whitney rank sum test.
and apolipoproteins
were compared
by the
Results The mean plasma lipid and lipoprotein values before and after three cycles of oral contraceptive treatment are shown in Table II. Whereas total plasma cholesterol
158 TABLE II Plasma lipid and lipoprotein parameters at baseline and after three cycles of oral contraceptive administration. The combinations contained ethinylestradiol and alternatively desogestrel (M-DG), gestodene (M-GD) or levonorgestrel (T-LN). (Values in mg/l, mean f SD), NS = not significant at the 5% niveau Parameter
Drug
Baseline
Treatment
P
Total cholesterol
M-DG M-GD T-LN
184.3k 33.7 180.7 f 32.2 176.1 If:26.6
186.4& 32.4 175.4k 36.9 185.3 f26.9
NS
M-DG M-GD T-LN
86.4 k 43.2 77.6 f 44.1 73.6 f 25.6
104.4 * 44.1 98.5 k51.1 90.1+25.1
NS < 0.01
HDL-cholesterol
M-DG M-GD T-LN
53.7 + 12.5 55.8 k 12.1 59.1 f 17.7
59.6 f 13.6 56.2 5 12.7 61.1+ 14.7
-=z0.05 NS NS
LDL-cholesterol
M-DG M-GD T-LN
124.8 + 34.2 118.1* 30.5 115.9* 33.1
122.1+ 30.9 112.4* 33.4 122.1 k28.5
VLDL-cholesterol
M-DG M-GD T-LN
5.8* 6.9 4.2+ 3.4 3.6+ 4.2
Triglycerides
4.9* 5.2* 4.1+
NS
NS
< 0.05
NS NS NS NS NS NS
3.2 3.6 3.3
TABLE III Changes in plasma apoprotein parameters during the use of oral contraceptives. Baseline levels and concentrations after three cycles of the combined use of ethinylestradiol and either desogestrel (M-DG), gestodene (M-GD) or levonorgestrel (T-LN) are displayed. Mean levels ( f SD) in mg/dl, NS = not significant at the 5% niveau Parameter
Drug
Baseline
Treatment
P
Ape-AI
M-DG M-GD T-LN
137.8 k28.3 144.9 *19.3 139.1 f 22.1
159.8 k 30.4 161.5 f 26.7 150.6 + 29.8
< 0.05 < 0.05 NS
Apo-AI1
M-DG M-GD T-LN
45.5 f 48.9 f 47.4 f
6.3 4.4 5.1
55.2 k 9.7 55.5 + 6.9 53.4 f 8.8
< 0.01 < 0.01 < 0.05
Apo-AIV
M-DG M-GD T-LN
15.0 f 4.8 15.3 + 5.2 15.3 + 4.2
14.4 f 3.8 14.3 * 3.1 14.0 + 5.2
NS NS NS
APO-B
M-DG M-GD T-LN
79.1 f23.2 78.5 k30.2 74.7 +21.7
81.1 +20.4 78.3 k21.4 75.2 k26.2
NS NS NS
APO-E
M-DG M-GD T-LN
5.59+ 1.80 6.08f 1.61 6.54* 2.09
4.81 i 5.14* 5.29*
1.57 1.41 1.46
< 0.001 < 0.01 < 0.01
159
and LDL cholesterol did not change significantly in any group, plasma triglycerides rose. The rise in plasma triglyceride levels in the M-DG group was not statistically significant, this was due, at least in part, to the high variation among values. VLDL-cholesterol also did not change significantly in any group and likewise showed high variations among values. Only in the M-DG group, HDL cholesterol rose significantly. Measurements of HDL-cholesterol concentrations, determined by the precipitation procedure, yielded similar results. The only significant increase occurred in the M-DG group (data not shown). Apo A-I increased in the M-DG and also in the M-GD group. The increase, however, was not significant in the T-LN group. Apo A-II increased in all three groups. A consistent finding during the use of the three different oral contraceptive regimens was the significant reduction of apo-E concentrations in all groups. The difference between the higher mean values of apo A-IV before and the lower levels found in women taking oral contraceptives was not statistically significant. The results of the apoprotein changes are summarized in Table III. Discussion Up to now, most studies evaluating the effects of oral contraceptive use on lipoprotein metabolism have focused on the measurement of total plasma and lipoprotein lipids and of apolipoproteins A and B. Bertolini et al. [l] and Knopp [2] recently reviewed the mechanisms by which estrogens and progestogens affect plasma lipoproteins. They found that estrogens increased both hepatic synthesis and secretion of HDL and VLDL. Progestins inhibit (probably by their antiestrogenic properties) VLDL synthesis. Furthermore, whereas estrogens lower hepatic triglyceride lipase (HTGL) activity and subsequently raise HDL, levels, androgenic progestins increase HTGL activity and lower HDL, levels. This effect was not seen with weakly androgenic progestins and antiandrogenic progestins [1,2]. Thus the lipoprotein changes monitored in women taking oral contraceptives are critically dependent upon the dose and type of estrogen and progestogen (androgenic and antiestrogenic power) used in the formulation. With respect to conventionally measured lipoprotein parameters, our results are in general agreement with data published recently by Bertolini and colleagues [l]. They used virtually the same formulations as those used in this study and monitored the changes in lipid metabolism after 2, 4 and 6 cycles [12]. Although in their study, the oral contraceptive-induced changes in lipid metabolism were dependent on the duration of treatment, the most pronounced changes had already occurred after only two cycles of treatment [12]. M&z and colleagues [13] and Bergink and co-workers [14] studied preparations comparable to ours also over a three cycle period. Our results are consistent with those of Mlrz et al. [13] especially in that no significant changes were seen in total cholesterol, LDL-cholesterol or apo-B. Our results also agree with those of Bergink et al. [14] in which comparable desogestrel + EE preparations were used. In a recent study, Harvengt et al. [15] investigated preparations corresponding to M-DG and T-LN in our study. These workers failed to see significant changes in LDL-cholesterol and apo-B after either 3 or 6 months of treatment.
160
In our study, we monitored in addition apo-E and apo-AIV levels. The lowering of apolipoprotein E plasma levels seen was not unexpected, as we had already recently shown that ethinylestradiol application to post-menopausal women (1 pg/kg per day) led to a 36% drop in plasma apo-E levels within 4 days of treatment [16]. Somewhat surprising, however, was the fact that there was still a significant lowering of plasma apo-E during treatment with all formulations studied when the preparations were administered at a dose of about (0.5 pg/kg per day), and additionally combined with different progestogens. There may be several explanations for this phenomenon. It was shown earlier that liver membranes of estrogentreated rats bound increased amounts of LDL [17], probably by the induction of an LDL receptor [18]. In liver-membrane binding and perfusion experiments with livers from estrogen treated rats, increased binding of lipoproteins containing apo-B and apo-E was shown [19]; furthermore, estrogen induced an increase in the uptake of cholesterol-rich very-low-density lipoproteins as shown in rabbit liver-perfusion studies [20]. In indirect immunofluorescence studies with perfused livers of normal and estrogen treated rabbits, we could also show that estrogen treatment led to an internalization into hepatocytes of apo-E which normally surrounds these cells [21]. In addition, observations in humans argue for the induction of a similar mechanism by estrogens. Estrogen treatment leads to a pronounced lipid-, remnant- and apo-E lowering effect in most patients with type III hyperlipidemia. This disease is characterized by the accumulation of remnant particles including abnormal pmigrating VLDL (/3-VLDL) [22-251. Also, in the Lipid Research Clinics Program Prevalence Study, Knopp and co-workers [26] found P-VLDL less common among women using hormones than among those not using them. This might furthermore argue for a stimulation of remnant removal by estrogens. Along the same line, Berr and colleagues [27], studied retinylpalmitate-labeled chylomicron remnant metabolism in young women off and on contraceptive steroids, and found significantly increased plasma clearance rates of chylomicron remnants. They interpret their results as indicating an increased hepatic uptake of these remnants [27]. Similar results were reported by Gleeson and colleagues [28]. The effects of progestogens per se on apolipoprotein E are not yet clear, but the known estrogen effect [16] is sufficient to account for the changes seen in apo-E of all three drugs used in the present study. We could not directly evaluate the fraction of apo-E-containing lipoproteins which decreased as apo-E in plasma diminished. The use of ultracentrifugation to isolate VLDL or IDL and to determine apo-E content did not seem appropriate as this procedure is known to dissociate apo-E from lipoproteins. On the other hand, more gentle dissociating methods, such as gel filtration, are not practical for such a large number of samples as we processed. Also, since only a particular subset of particles is diminished, the change might not be detectable by measuring VLDL- or IDL-cholesterol concentrations. VLDLcholesterol levels determined in the present study failed to indicate significant changes. Nevertheless data obtained by Berr et al. [27] and Gleeson et al. [28] in women on oral contraceptives together with the bulk of clinical [22-251 and epidemiological [26] evidence in humans and experimental animals [17-211 strongly suggest that estrogen exerts a major action on the lowering of remnants of triglyceride rich lipoproteins thought to be specifically atherogenic. Thus a lowering
161
of apo-E plasma levels should be an indicator of a more beneficial effect through decreasing remnant levels, which predominantly contain apo-B and apo-E. APO-E concentration can therefore, be a useful additional parameter to estimate the change in the overall risk of atherosclerosis induced by the use of oral contraceptives, a matter which is still an object of controversy [2]. Acknowledgement
The authors are indebted to Dr. Rodden for linguistic advice. We thank Sabine Motzny and Uhike Otte for excellent technical and Caroline Branlant for skillful secretarial assistance. A.S. was supported by a grant from the Deutsche Forschungsgemeinschaft. References 1 Bertolini S, Capitanio GL, Valice S, Dago A, Grace S, Degl’Innocenti L, De Cecco L. Lipoprotein metabolism in women using oral contraceptives. In: Genazzani AR, ed. Gynecological Endocrinology, pp. 27-35, Gamforth: Parthenon Publishing, 1986. 2 Knopp RH. Cardiovascular effects of endogenous and exogenous sex hormones over a woman’s lifetime. Am J Obstet Gynecol 1988;158:1630-1643. 3 Bottmger LE, Boman G, Eklund G, Westerholm B. Oral contraceptives and thrombembolic disease: effects of lowering - estrogen - content. Lancet 1980$:1097-1101. and arterial disease. Evidence from the Royal College of General Prac4 Kay CR. Progestogens titioneers’ Study. Am J Obstet Gynecol 1982;142:762-765. and cardiovascular reactions associated with 5 Meade TW, Greenberg G, Thompson SG. Progestogens oral contraceptives and a comparison of the safety of 50 and 30 pg oestrogen preparations. Br Med J 1980;1:1157-1161. A, Samsioe G, Svanborg A. Lipid metabohc studies m oophorectomised 6 Silfverstolpe G, Gustafson women: effects on serum lipids and lipoproteins of three synthettc progestogens. Maturitas 1982;4:103-111. U, Wallentin L. L-Norgestrel and progesterone have dtfferent Influences 7 Fahraeus L, Larsson-Cohn on plasma lipoproteins. Eur J Clin Invest 1983;13:447-453. lipoprotems, and risk of 8 Kannel WB, Castelli WP, Gordon T, MC Namara PM. Serum cholesterol, coronary heart diseases. The Framingham study. Am Intern Med 1971;74:1-12. MC, Kannel WB, Dawber TR. High-density lipoprotein as a 9 Gordon T, Castelli WP, Hjortland protective factor against coronary heart disease. Am J Med 1977;62:707-714. of apolipoprotein E from whole blood plasma by immunoblotting. J Liptd 10 Stemmetz A. Phenotyping Res 1987;28:1364-1370. K. Changes of apolipoprotein A-IV m 11 Stemmetz A. Czekehus P, Thiemann E, Motzny S, Kaffarnik the human neonate: evidence for different inductions of apolipoprotems A-IV and A-I in the postpartum period. Atherosclerosis 1988:69:21-27. G, Montagna S, Grace M, Saturnino M, Balestrert Z, 12 Bertolim S, Elicio R, Cordera GL, Gapitanio De Cecco L. Effects of three low-dose oral contraceptive formulations on lipid metabolism. Acta Obstet Gynecol Stand 1987;66:327-32. crossover comparison of two 13 M&z W, Gross W, Romberg G, Taubert HD, Kuhl H. A randomized low-dose contraceptives: effects on serum lipids and lipoproteins. Am J Obstret Gynecol 1985;153:287-293. HJ, Lund J, Nummi S. Effects of levonorgenestrel ‘and desogestrel in 14 Bergink EW, Kloosterboer low-dose oral contraceptive combinations on serum lipids, apolipoprotems A-I and B and glycosylated proteins. Contraception 1984;30:61-72. C, Desager JP, Gaspard U, Lepot M. Changes in lipoprotein composition m women 15 Harvengt receiving two low-dose oral contraceptives containing ethinylestradiol and gonane progestms. Contraception 1988;37:565-575.
162 16 Hazzard WR, Applebaum-Bowden D, Steinmetz A, MC Lean P, Mathis C, Fontana D, Warmick GR, Albers JJ. Apolipoprotein E: Response to estrogen administration. 7th International Symposium on Atherosclerosis. Proceedings of Poster Communications, Melbourne 1985;379 (abstract). 17 Kovanen PT, Brown MS, Goldstein JL. Increased binding of low-density lipoprotein to liver membranes from rats treated with l’la-ethinyl estradiol. J Biol Chem 1979;254:11367-11373. 18 Cooper AD, Nutik R, Chen J. Characterization of the estrogen-induced lipoprotein receptor of rat liver. J Lipid Res 1987;28:59-68. 19 Windler EET, Kovanen PI, Chao YS, Brown MS, Have1 RJ, Goldstein JL. The estradiol-stimulated lipoprotein receptor of rat liver. A binding site that mediates the uptake of rat lipoproteins containing apolipoproteins B and E. J Biol Chem 1980;225:10464-10471. 20 Floren CH, Kuswaha RS, Hazzard WR, Albers JJ. Estrogen-induced increase in uptake of cholesterol rich very-low-density lipoproteins in perfused rabbit liver. Metabolism 1981;30:367-375. 21 Iozzo RV, Steinmetz A, Kushwaha RS. Estrogen treatment changes the cellular distribution of apolipoprotein E in the liver. Cell Biol Int Rep 1982;6:875-881. 22 Chait A, Brunzell J, Albers JJ, Hazzard WR. Type III hyperlipoproteinemia (‘remnant removal disease’). Insight into the pathogenic mechanism. Lancet 1977;i:1176-1178. 23 Kushwaha RS, Hazzard WR, Cagne C, Chait A, Albers JJ. Type III hyperlipoproteinemia: paradoxical hypolipidemic response to estrogen. Ann Int Med 1977;87:517-525. 24 Falko JM, Schonfeld G, Witztum JL, Kolar J, Weidman SW. Effects of estrogen therapy on apolipoprotein E in type III hyperlipoproteinemia. Metabolism 1979;28:1171-1177. 25 Stuyt PMJ, Demacker PNM, Van? Laar A. A study of the hypolipidemic effect of estrogen in type III hyperlipoproteinemia. Horm Metabol Res 1986;18:607-610. 26 Knopp RH, Walden LE, Heiss G, Johnson JL, Wahl PW. Prevalence and clinical correlates of beta-migrating very-low density lipoprotein. Lipid Research Clinics Program Prevalence Study. Am J Med 1986;81:493-502. 27 Berr F, Eckel RH, Kern F. Contraceptive steroids increase hepatic uptake of chylomicron remnants in healthy young women. J Lipid Res 1986;27:645-651. 28 Gleeson JM, Dukes CS, Elstad NL, Cham IF, Wilson DE. Effects of estrogen/progestin agents on plasma retinoids and chylomicron remnant metabolism. Contraception 1987;35:69-78.
EUROBS
155
00980
Low-dose oral contraceptives lower plasma levels of apolipoprotein E Armin Steinmetz
‘, Katharina Bauer I, Ralph Jtirgensen and Hans Kaffarnik ’
’
’ Zenirum Innere Medirm, Endokrinologie and Stoffwechsel and ’ Insiitut ftir medrrrnrsche Sratrstlh, Philipps Umversitiir, D-3550 Marburg, F. R. G. Accepted
for publication
10 November
1989
Summary Three different oral contraceptive preparations were studied before and after a 3 month treatment period with respect to their effects on plasma lipoprotein parameters. A total of 58 healthy women requesting oral contraception were randomly assigned to three groups. Each woman received either monophasic preparations containing ethinylestradiol and desogestrel (M-DG); ethinylestradiol and gestodene (M-GD); or a triphasic preparation of ethinylestradiol and levonorgestrel (T-LN). As has been reported in other studies, the concentrations of total plasma cholesterol and apolipoproteins B and A-IV did not change significantly in any group. HDL cholesterol, triglycerides, apolipoproteins A-I and A-II increased or tended to increase. Despite the effects of the three hormone preparations on these lipoprotein parameters, however, each led to a highly significant decrease in apolipoprotein E plasma levels. Considering the recently reported observations that oral contraceptives increase the hepatic uptake of cholesterol-rich remnants, this decrease in apo-E plasma levels may in women that take oral contraceptives be directly correlated with increased hepatic lipoprotein metabolism. Oral contraceptive,
low dose; Plasma
levels; Apohpoprotein
E
Abbreviations: Apo, apolipoprotein; VLDL, very-low-density lipoprotein; LDL, low-density lipoprotem; IDL, intermediate density lipoprotein; HDL, high-density lipoprotein; PEG, poly(ethylene glycol); EE, ethinylestradiol; M-DG, ethinylestradiol and desogestrel; M-GD, ethinylestradiol and gestodene; T-LN, ethinylestradiol and levonorgestrel. Correspondence: A. Steinmetz, Zentrum Innere Medizin, Endokrinologie und Stoffwechsel, Philipps UniversltBt, Baldingerstrasse, D-3550 Marburg, F.R.G.
0028-2243/90/$03.50
Q 1990 Elsevier Science Publishers
B.V. (Biomedical
Division)
156
Introduction
Numerous studies have established relationships between the use of oral contraceptives and the changes in plasma lipoprotein levels (reviewed in Refs. 1 and 2). The high levels of the synthetic estrogen ethinylestradiol prescribed in the 1970s was later reduced to about 30 ~18 per day. This led to a decrease in the incidence of venous thrombembolism but did not seem to affect the incidence of myocardial infarction and stroke [3]. Although estrogens were originally thought to be responsible for the arterial pathology, attention was later focused on progestogens [4,5]. Their intrinsic androgenic activities were shown to increase LDL cholesterol, to lower HDL cholesterol levels [6,7] and thus to induce lipoprotein changes believed to be atherogenic [8,9]. Most studies conducted to establish the influence of oral contraceptives on lipoprotein metabolism have been concerned with the measurement of total plasma and lipoprotein lipids and apoproteins A-I, A-II and B. As yet no information is available on the influence of oral contraceptives on apolipoprotein E. We have therefore studied three different formulations of oral contraceptives currently used in Europe with respect to their influence on conventional lipid parameters and additionally apolipoprotein E. Besides ethinylestradiol, the preparations contained other selected progestogen components. Subjects and Methods
Subjects Healthy young women requesting oral contraceptives at a gynecological outpatient facility were examined for eligibility and then randomly allocated to three different treatment groups. Informed consent was obtained from all subjects. Criteria for entering the study were: age range 16-35 years, regular menstrual cycle of 25-36 days, absence of contra-indications to oral contraceptives, absence of significant endocrinologic, hematologic or metabolic disorders. No oral contracep-
TABLE
I
Dosage
of the different
contraceptive
preparations
used in the study
Abbreviation
Composition
M-DG
30 /.tg ethinylestradiol 150 pg desogestrel
l-21
30 pg ethinylestradiol 75 pg gestodene
l-21
M-GD
T-LN
30 50 40 75 30 125
pg gg pg pg pg pg
ethinylestradlol levonorgestrel ethinylestradiol levonorgestrel ethinylestradiol levonorgestrel
Day
l7-11 11-21
6
157
tives had been used or hormones taken for 3 months prior to the study. The women were requested not to change their eating habits. They received either monophasic formulations containing desogestrel (M-DG) or gestodene (M-GD) or a triphasic preparation containing levonorgestrel (T-LN). The dosages of the preparations given are summarized in Table I. The three groups did not differ significantly in age (M-DG, n = 23, 21.5 & 3.9 years, range 16-30; M-GD, n = 21, 20.1 + 4.6 years, range 16-35; T-LN, n = 15, 22.5 _t 4.2 years, range 16-35) or mean body mass index (21.1, 20.3 and 22.1, respectively). Levels of physical activity and alcohol intake were comparable between the three groups. The distribution of apolipoprotein E genetic phenotypes did not differ among the three treatment groups. Baseline measurements were performed in the luteal phase between days 18 and 22 of the cycle; blood samples were likewise taken between days 18 and 22 in the treatment period after three cycles. Lipoprotein measurements and apoprotein analysis Blood samples were drawn after an least 12 h of fasting between 8 and 10 a.m. and serum was obtained by low-speed centrifugation (4000 X g, 30 min) at 4” C. Aliquots for apoprotein determinations were stored at - 20 ’ C and analyzed within 4 months. For lipoprotein analysis, samples were kept at + 4” C and processed within 2 days. Plasma cholesterol and triglycerides were measured enzymatically, using the commercially available kits of Boehringer (Mannheim, F.R.G.). VLDL-, LDL- and HDL-cholesterol levels were determined by an electrophoretic procedure provided by Immuno Chemie (Heidelberg, F.R.G.). HDL cholesterol was in parallel measured enzymatically after precipitation of apo-B-containing lipoproteins, using the commercially available kit of Boehringer (Mannheim, F.R.G.). Apolipoproteins A-I, A-II and B were quantitatively measured by the use of radial immunodiffusion with kits from Immuno Chemie (Heidelberg, F.R.G.). Apolipoproteins E and A-IV were quantified by electroimmunoassay. Antibodies to human apo-E were raised as outlined before [lo]. The assay was performed in 1% agarose gels on 12 x 9 cm plates. The Verona1 electrophoresis buffer contained, in addition, 1% sodium deoxycholate, 0.5% Triton X-100 and 2.5% PEG 4000. Samples were first diluted 1:2 in a 2-fold concentrated preparation of this buffer and then further diluted an additional 3- and 6-fold in electrophoresis buffer. Electrophoresis was then carried out at 25 mA for 4 h. Purified apo-E was used as a primary standard and pooled frozen plasma as a secondary standard. The interassay coefficient of variation was 6.8% and the intraassay coefficient of variation was 3.7%. The assay for Apo A-IV has been previously described [ll]. Statistical analysis Concentrations of lipids, lipoproteins Mann-Whitney rank sum test.
and apolipoproteins
were compared
by the
Results The mean plasma lipid and lipoprotein values before and after three cycles of oral contraceptive treatment are shown in Table II. Whereas total plasma cholesterol
158 TABLE II Plasma lipid and lipoprotein parameters at baseline and after three cycles of oral contraceptive administration. The combinations contained ethinylestradiol and alternatively desogestrel (M-DG), gestodene (M-GD) or levonorgestrel (T-LN). (Values in mg/l, mean f SD), NS = not significant at the 5% niveau Parameter
Drug
Baseline
Treatment
P
Total cholesterol
M-DG M-GD T-LN
184.3k 33.7 180.7 f 32.2 176.1 If:26.6
186.4& 32.4 175.4k 36.9 185.3 f26.9
NS
M-DG M-GD T-LN
86.4 k 43.2 77.6 f 44.1 73.6 f 25.6
104.4 * 44.1 98.5 k51.1 90.1+25.1
NS < 0.01
HDL-cholesterol
M-DG M-GD T-LN
53.7 + 12.5 55.8 k 12.1 59.1 f 17.7
59.6 f 13.6 56.2 5 12.7 61.1+ 14.7
-=z0.05 NS NS
LDL-cholesterol
M-DG M-GD T-LN
124.8 + 34.2 118.1* 30.5 115.9* 33.1
122.1+ 30.9 112.4* 33.4 122.1 k28.5
VLDL-cholesterol
M-DG M-GD T-LN
5.8* 6.9 4.2+ 3.4 3.6+ 4.2
Triglycerides
4.9* 5.2* 4.1+
NS
NS
< 0.05
NS NS NS NS NS NS
3.2 3.6 3.3
TABLE III Changes in plasma apoprotein parameters during the use of oral contraceptives. Baseline levels and concentrations after three cycles of the combined use of ethinylestradiol and either desogestrel (M-DG), gestodene (M-GD) or levonorgestrel (T-LN) are displayed. Mean levels ( f SD) in mg/dl, NS = not significant at the 5% niveau Parameter
Drug
Baseline
Treatment
P
Ape-AI
M-DG M-GD T-LN
137.8 k28.3 144.9 *19.3 139.1 f 22.1
159.8 k 30.4 161.5 f 26.7 150.6 + 29.8
< 0.05 < 0.05 NS
Apo-AI1
M-DG M-GD T-LN
45.5 f 48.9 f 47.4 f
6.3 4.4 5.1
55.2 k 9.7 55.5 + 6.9 53.4 f 8.8
< 0.01 < 0.01 < 0.05
Apo-AIV
M-DG M-GD T-LN
15.0 f 4.8 15.3 + 5.2 15.3 + 4.2
14.4 f 3.8 14.3 * 3.1 14.0 + 5.2
NS NS NS
APO-B
M-DG M-GD T-LN
79.1 f23.2 78.5 k30.2 74.7 +21.7
81.1 +20.4 78.3 k21.4 75.2 k26.2
NS NS NS
APO-E
M-DG M-GD T-LN
5.59+ 1.80 6.08f 1.61 6.54* 2.09
4.81 i 5.14* 5.29*
1.57 1.41 1.46
< 0.001 < 0.01 < 0.01
159
and LDL cholesterol did not change significantly in any group, plasma triglycerides rose. The rise in plasma triglyceride levels in the M-DG group was not statistically significant, this was due, at least in part, to the high variation among values. VLDL-cholesterol also did not change significantly in any group and likewise showed high variations among values. Only in the M-DG group, HDL cholesterol rose significantly. Measurements of HDL-cholesterol concentrations, determined by the precipitation procedure, yielded similar results. The only significant increase occurred in the M-DG group (data not shown). Apo A-I increased in the M-DG and also in the M-GD group. The increase, however, was not significant in the T-LN group. Apo A-II increased in all three groups. A consistent finding during the use of the three different oral contraceptive regimens was the significant reduction of apo-E concentrations in all groups. The difference between the higher mean values of apo A-IV before and the lower levels found in women taking oral contraceptives was not statistically significant. The results of the apoprotein changes are summarized in Table III. Discussion Up to now, most studies evaluating the effects of oral contraceptive use on lipoprotein metabolism have focused on the measurement of total plasma and lipoprotein lipids and of apolipoproteins A and B. Bertolini et al. [l] and Knopp [2] recently reviewed the mechanisms by which estrogens and progestogens affect plasma lipoproteins. They found that estrogens increased both hepatic synthesis and secretion of HDL and VLDL. Progestins inhibit (probably by their antiestrogenic properties) VLDL synthesis. Furthermore, whereas estrogens lower hepatic triglyceride lipase (HTGL) activity and subsequently raise HDL, levels, androgenic progestins increase HTGL activity and lower HDL, levels. This effect was not seen with weakly androgenic progestins and antiandrogenic progestins [1,2]. Thus the lipoprotein changes monitored in women taking oral contraceptives are critically dependent upon the dose and type of estrogen and progestogen (androgenic and antiestrogenic power) used in the formulation. With respect to conventionally measured lipoprotein parameters, our results are in general agreement with data published recently by Bertolini and colleagues [l]. They used virtually the same formulations as those used in this study and monitored the changes in lipid metabolism after 2, 4 and 6 cycles [12]. Although in their study, the oral contraceptive-induced changes in lipid metabolism were dependent on the duration of treatment, the most pronounced changes had already occurred after only two cycles of treatment [12]. M&z and colleagues [13] and Bergink and co-workers [14] studied preparations comparable to ours also over a three cycle period. Our results are consistent with those of Mlrz et al. [13] especially in that no significant changes were seen in total cholesterol, LDL-cholesterol or apo-B. Our results also agree with those of Bergink et al. [14] in which comparable desogestrel + EE preparations were used. In a recent study, Harvengt et al. [15] investigated preparations corresponding to M-DG and T-LN in our study. These workers failed to see significant changes in LDL-cholesterol and apo-B after either 3 or 6 months of treatment.
160
In our study, we monitored in addition apo-E and apo-AIV levels. The lowering of apolipoprotein E plasma levels seen was not unexpected, as we had already recently shown that ethinylestradiol application to post-menopausal women (1 pg/kg per day) led to a 36% drop in plasma apo-E levels within 4 days of treatment [16]. Somewhat surprising, however, was the fact that there was still a significant lowering of plasma apo-E during treatment with all formulations studied when the preparations were administered at a dose of about (0.5 pg/kg per day), and additionally combined with different progestogens. There may be several explanations for this phenomenon. It was shown earlier that liver membranes of estrogentreated rats bound increased amounts of LDL [17], probably by the induction of an LDL receptor [18]. In liver-membrane binding and perfusion experiments with livers from estrogen treated rats, increased binding of lipoproteins containing apo-B and apo-E was shown [19]; furthermore, estrogen induced an increase in the uptake of cholesterol-rich very-low-density lipoproteins as shown in rabbit liver-perfusion studies [20]. In indirect immunofluorescence studies with perfused livers of normal and estrogen treated rabbits, we could also show that estrogen treatment led to an internalization into hepatocytes of apo-E which normally surrounds these cells [21]. In addition, observations in humans argue for the induction of a similar mechanism by estrogens. Estrogen treatment leads to a pronounced lipid-, remnant- and apo-E lowering effect in most patients with type III hyperlipidemia. This disease is characterized by the accumulation of remnant particles including abnormal pmigrating VLDL (/3-VLDL) [22-251. Also, in the Lipid Research Clinics Program Prevalence Study, Knopp and co-workers [26] found P-VLDL less common among women using hormones than among those not using them. This might furthermore argue for a stimulation of remnant removal by estrogens. Along the same line, Berr and colleagues [27], studied retinylpalmitate-labeled chylomicron remnant metabolism in young women off and on contraceptive steroids, and found significantly increased plasma clearance rates of chylomicron remnants. They interpret their results as indicating an increased hepatic uptake of these remnants [27]. Similar results were reported by Gleeson and colleagues [28]. The effects of progestogens per se on apolipoprotein E are not yet clear, but the known estrogen effect [16] is sufficient to account for the changes seen in apo-E of all three drugs used in the present study. We could not directly evaluate the fraction of apo-E-containing lipoproteins which decreased as apo-E in plasma diminished. The use of ultracentrifugation to isolate VLDL or IDL and to determine apo-E content did not seem appropriate as this procedure is known to dissociate apo-E from lipoproteins. On the other hand, more gentle dissociating methods, such as gel filtration, are not practical for such a large number of samples as we processed. Also, since only a particular subset of particles is diminished, the change might not be detectable by measuring VLDL- or IDL-cholesterol concentrations. VLDLcholesterol levels determined in the present study failed to indicate significant changes. Nevertheless data obtained by Berr et al. [27] and Gleeson et al. [28] in women on oral contraceptives together with the bulk of clinical [22-251 and epidemiological [26] evidence in humans and experimental animals [17-211 strongly suggest that estrogen exerts a major action on the lowering of remnants of triglyceride rich lipoproteins thought to be specifically atherogenic. Thus a lowering
161
of apo-E plasma levels should be an indicator of a more beneficial effect through decreasing remnant levels, which predominantly contain apo-B and apo-E. APO-E concentration can therefore, be a useful additional parameter to estimate the change in the overall risk of atherosclerosis induced by the use of oral contraceptives, a matter which is still an object of controversy [2]. Acknowledgement
The authors are indebted to Dr. Rodden for linguistic advice. We thank Sabine Motzny and Uhike Otte for excellent technical and Caroline Branlant for skillful secretarial assistance. A.S. was supported by a grant from the Deutsche Forschungsgemeinschaft. References 1 Bertolini S, Capitanio GL, Valice S, Dago A, Grace S, Degl’Innocenti L, De Cecco L. Lipoprotein metabolism in women using oral contraceptives. In: Genazzani AR, ed. Gynecological Endocrinology, pp. 27-35, Gamforth: Parthenon Publishing, 1986. 2 Knopp RH. Cardiovascular effects of endogenous and exogenous sex hormones over a woman’s lifetime. Am J Obstet Gynecol 1988;158:1630-1643. 3 Bottmger LE, Boman G, Eklund G, Westerholm B. Oral contraceptives and thrombembolic disease: effects of lowering - estrogen - content. Lancet 1980$:1097-1101. and arterial disease. Evidence from the Royal College of General Prac4 Kay CR. Progestogens titioneers’ Study. Am J Obstet Gynecol 1982;142:762-765. and cardiovascular reactions associated with 5 Meade TW, Greenberg G, Thompson SG. Progestogens oral contraceptives and a comparison of the safety of 50 and 30 pg oestrogen preparations. Br Med J 1980;1:1157-1161. A, Samsioe G, Svanborg A. Lipid metabohc studies m oophorectomised 6 Silfverstolpe G, Gustafson women: effects on serum lipids and lipoproteins of three synthettc progestogens. Maturitas 1982;4:103-111. U, Wallentin L. L-Norgestrel and progesterone have dtfferent Influences 7 Fahraeus L, Larsson-Cohn on plasma lipoproteins. Eur J Clin Invest 1983;13:447-453. lipoprotems, and risk of 8 Kannel WB, Castelli WP, Gordon T, MC Namara PM. Serum cholesterol, coronary heart diseases. The Framingham study. Am Intern Med 1971;74:1-12. MC, Kannel WB, Dawber TR. High-density lipoprotein as a 9 Gordon T, Castelli WP, Hjortland protective factor against coronary heart disease. Am J Med 1977;62:707-714. of apolipoprotein E from whole blood plasma by immunoblotting. J Liptd 10 Stemmetz A. Phenotyping Res 1987;28:1364-1370. K. Changes of apolipoprotein A-IV m 11 Stemmetz A. Czekehus P, Thiemann E, Motzny S, Kaffarnik the human neonate: evidence for different inductions of apolipoprotems A-IV and A-I in the postpartum period. Atherosclerosis 1988:69:21-27. G, Montagna S, Grace M, Saturnino M, Balestrert Z, 12 Bertolim S, Elicio R, Cordera GL, Gapitanio De Cecco L. Effects of three low-dose oral contraceptive formulations on lipid metabolism. Acta Obstet Gynecol Stand 1987;66:327-32. crossover comparison of two 13 M&z W, Gross W, Romberg G, Taubert HD, Kuhl H. A randomized low-dose contraceptives: effects on serum lipids and lipoproteins. Am J Obstret Gynecol 1985;153:287-293. HJ, Lund J, Nummi S. Effects of levonorgenestrel ‘and desogestrel in 14 Bergink EW, Kloosterboer low-dose oral contraceptive combinations on serum lipids, apolipoprotems A-I and B and glycosylated proteins. Contraception 1984;30:61-72. C, Desager JP, Gaspard U, Lepot M. Changes in lipoprotein composition m women 15 Harvengt receiving two low-dose oral contraceptives containing ethinylestradiol and gonane progestms. Contraception 1988;37:565-575.
162 16 Hazzard WR, Applebaum-Bowden D, Steinmetz A, MC Lean P, Mathis C, Fontana D, Warmick GR, Albers JJ. Apolipoprotein E: Response to estrogen administration. 7th International Symposium on Atherosclerosis. Proceedings of Poster Communications, Melbourne 1985;379 (abstract). 17 Kovanen PT, Brown MS, Goldstein JL. Increased binding of low-density lipoprotein to liver membranes from rats treated with l’la-ethinyl estradiol. J Biol Chem 1979;254:11367-11373. 18 Cooper AD, Nutik R, Chen J. Characterization of the estrogen-induced lipoprotein receptor of rat liver. J Lipid Res 1987;28:59-68. 19 Windler EET, Kovanen PI, Chao YS, Brown MS, Have1 RJ, Goldstein JL. The estradiol-stimulated lipoprotein receptor of rat liver. A binding site that mediates the uptake of rat lipoproteins containing apolipoproteins B and E. J Biol Chem 1980;225:10464-10471. 20 Floren CH, Kuswaha RS, Hazzard WR, Albers JJ. Estrogen-induced increase in uptake of cholesterol rich very-low-density lipoproteins in perfused rabbit liver. Metabolism 1981;30:367-375. 21 Iozzo RV, Steinmetz A, Kushwaha RS. Estrogen treatment changes the cellular distribution of apolipoprotein E in the liver. Cell Biol Int Rep 1982;6:875-881. 22 Chait A, Brunzell J, Albers JJ, Hazzard WR. Type III hyperlipoproteinemia (‘remnant removal disease’). Insight into the pathogenic mechanism. Lancet 1977;i:1176-1178. 23 Kushwaha RS, Hazzard WR, Cagne C, Chait A, Albers JJ. Type III hyperlipoproteinemia: paradoxical hypolipidemic response to estrogen. Ann Int Med 1977;87:517-525. 24 Falko JM, Schonfeld G, Witztum JL, Kolar J, Weidman SW. Effects of estrogen therapy on apolipoprotein E in type III hyperlipoproteinemia. Metabolism 1979;28:1171-1177. 25 Stuyt PMJ, Demacker PNM, Van? Laar A. A study of the hypolipidemic effect of estrogen in type III hyperlipoproteinemia. Horm Metabol Res 1986;18:607-610. 26 Knopp RH, Walden LE, Heiss G, Johnson JL, Wahl PW. Prevalence and clinical correlates of beta-migrating very-low density lipoprotein. Lipid Research Clinics Program Prevalence Study. Am J Med 1986;81:493-502. 27 Berr F, Eckel RH, Kern F. Contraceptive steroids increase hepatic uptake of chylomicron remnants in healthy young women. J Lipid Res 1986;27:645-651. 28 Gleeson JM, Dukes CS, Elstad NL, Cham IF, Wilson DE. Effects of estrogen/progestin agents on plasma retinoids and chylomicron remnant metabolism. Contraception 1987;35:69-78.
