Sex Differences in Lipid and Lipoprotein Metabolism: It's Not Just about Sex Hormones

Studies of lipid kinetics in both sexes have elucidated many of the mechanisms responsible for the sex differences in the plasma lipid profile; however, the underlying physiologic modulators remain unclear.

Abstract

It is commonly thought that sex hormones are important regulators of plasma lipid kinetics and are responsible for sexual dimorphism in the plasma lipid profile. Here we discuss the findings from studies evaluating lipid and lipoprotein kinetics in men and women in the context of what we know about the effects of exogenous sex hormone administration, and we conclude that it is more complicated than that. It has become clear that normal physiological alterations in the hormonal milieu (i.e. due to menopause or throughout the menstrual cycle) do not significantly affect plasma lipid homeostasis. Furthermore, parenterally administered estrogens have either no effect or only very small beneficial effects, whereas orally administered estrogens raise plasma triglyceride concentrations—a phenomenon that is not consistent with the observed sex differences and likely results from the hepatic “first-pass effect.” The effects of progestogens and androgens mimic only in part the differences in plasma lipids between men and women. Thus, the underlying physiological modulators of plasma lipid metabolism responsible for the differences between men and women remain to be elucidated.

Premenopausal women have a less proatherogenic plasma lipid profile than men. Specifically, they have greater high-density lipoprotein (HDL) cholesterol concentration and lower low-density lipoprotein (LDL) cholesterol, very low-density lipoprotein (VLDL) cholesterol, total plasma triglyceride, and VLDL triglyceride concentrations (both during fasted and fed conditions) than age-matched men (1–8). Furthermore, both the concentration and average size of circulating VLDL are smaller (due to lower concentrations of large and medium VLDL) (9–15), whereas the concentration of LDL is smaller but the average size of LDL is larger (due to a shift toward large at the expense of small LDL) (9–15) in women than in men. The concentration of circulating HDL is not different in men and women, but women have larger HDL particles than men (due to a shift toward large and cholesterol-rich HDL) (10–12, 16, 17). These differences in plasma lipid concentrations and lipoprotein particle concentrations, subclass distributions, and sizes likely account for at least part of the cardioprotective effect of female sex (9, 10).

Sex Differences in Lipid Metabolism and Lipoprotein Kinetics

The kinetic mechanisms underlying the differences in the plasma lipid profile between men and premenopausal women have only partially been elucidated and are schematically summarized in Table 1. Curiously, the lower concentrations of VLDL triglyceride and LDL particles in plasma in women than in men are associated with accelerated (rather than reduced) VLDL triglyceride and LDL production; enhanced plasma VLDL triglyceride and LDL particle clearance compensates for this and actually results in overall lower steady-state concentrations (18–22). In contrast, the secretion rate of VLDL apolipoprotein B-100 (VLDL particles) is lower in women than in men. Consequently, women produce fewer but (on average) triglyceride-richer VLDL than men (19). The larger average size of nascent VLDL particles likely facilitates the removal of VLDL triglyceride from the circulation by enhancing their susceptibility to hydrolysis by lipoprotein lipase (23). The sex difference in HDL concentration is associated with a greater HDL apolipoprotein A-I synthesis rate (without differences in removal rate) in women than in men and both greater HDL apolipoprotein A-II synthesis and plasma clearance rates (17). There is no information regarding possible sex differences in cholesterol kinetics in the various lipoprotein fractions in human subjects. However, animal studies support an important role of endogenous sex hormones in mediating cholesterol metabolism in a sexually dimorphic manner (24).

Table 1.

Phenotypic differences in lipid and lipoprotein metabolism among premenopausal and postmenopausal women and men

	Sex differences		Effect of menopause in postmenopausal vs. premenopausal women
	Premenopausal women vs. men	Postmenopausal women vs. men
Estradiol	↑↑	↔	↓↓
Progesterone	↑	↔	↓
Testosterone	↓↓	↓↓	↔
Total and VLDL triglyceride concentrations	↓↓	↓	↔
VLDL triglyceride production	↑	↔	↑
VLDL triglyceride removal	↑↑	↑	↑
VLDL apolipoprotein B-100 production	↓	↓	↔
VLDL apolipoprotein B-100 removal	↔	↔	↔
HDL cholesterol concentration	↑	↑	↔
HDL apolipoprotein A-I production	↑	↔	↔
HDL apolipoprotein A-I removal	↔	↓	↔
HDL apolipoprotein A-II production	↑	↔	↔
HDL apolipoprotein A-II removal	↑	↔	↔
LDL cholesterol concentration	↓	↔	↑
LDL apolipoprotein B-100 production	↑	↔	↔
LDL apolipoprotein B-100 removal	↑↑	↔	↓

↑, Greater, ↓, less; ↔, no difference. Double arrows indicate more pronounced effects than single arrows.

The Case for Sex Hormones and Possible Confounders

The factors mediating the sex-specific regulation of plasma lipid kinetics and concentrations are not clear. It is thought that it is largely the influence of sex steroids that mediates sex differences in plasma lipid homeostasis—a concept that gains support from the fact that loss of ovarian function after menopause as well as hyperandrogenemia in women with polycystic ovary syndrome (PCOS) appears to increase plasma triglyceride and LDL cholesterol and decrease HDL cholesterol concentrations (1, 10, 25–27) and delay postprandial chylomicron clearance from plasma (28–30). However, these data are difficult to interpret because of potential confounding due to concomitant changes in total body fat, body fat distribution, and insulin sensitivity that accompany menopause (31), as well as PCOS (27). Healthy women tend to gain total body and intraabdominal fat and become more insulin resistant after menopause (31), and women with PCOS are typically overweight or obese (especially abdominally obese) and insulin resistant (27), and this alone could be responsible for the changes in lipid homeostasis because increased whole-body adiposity and central fat accumulation and decreased insulin sensitivity are associated with increased plasma triglyceride and LDL cholesterol and decreased HDL cholesterol concentrations (32). In fact, the results from several prospective (33–36) and cross-sectional (36–38) studies that were published during the last couple of years indicate that what has traditionally been thought to be menopause-associated dyslipidemia is, for the most part, due to chronological aging rather than the loss of ovarian function. Perhaps the only exception is a modest increase in total and LDL cholesterol and LDL apolipoprotein B-100 concentrations [due to a decrease in the LDL apolipoprotein B-100 fractional catabolic rate (30)], which seems to be specific to the menopausal transition, independent of aging or changes in body fat mass and distribution (33–36, 39). These results confirm the findings from some earlier but smaller longitudinal and cross-sectional studies (16, 30, 39–44) and are also consistent with the observation that surgically induced menopause without hormone replacement therapy has, aside from some modest transient changes, no effect on plasma lipid concentrations and lipoprotein profile when compared with a surgical control group (hysterectomy with conservation of the ovaries) (36, 42, 45). Nevertheless, we have recently found that VLDL triglyceride turnover (both hepatic VLDL triglyceride secretion and plasma VLDL triglyceride clearance) is faster in postmenopausal compared with well-matched premenopausal women despite similar plasma triglyceride and VLDL triglyceride concentrations in the two groups (46). The exact reasons for this phenomenon are unclear but are likely related to differences in hepatic fatty acid handling and changes in the architecture of nascent VLDL particles that favor more efficient removal (46), which may be further augmented by the increase in adipose tissue lipoprotein lipase activity after menopause (47). However, whether these differences are a direct consequence of the loss of ovarian function is unknown. Likewise, subjects with PCOS who are lean and normoinsulinemic (although these are rare) have a similar plasma lipid profile as healthy, age- and adiposity-matched women (48–51). Furthermore, most studies report little if any variation in plasma lipid and lipoprotein concentrations across the menstrual cycle in women (52–54), and those that do, in fact, report only small (by 5–6%) reductions in LDL cholesterol concentration during the luteal phase (52, 53) but no changes in plasma triglyceride or HDL cholesterol concentrations. Normal physiological alterations in the hormonal milieu therefore do not appear to significantly affect plasma lipid homeostasis. On the other hand, studies that evaluated sex steroid hormone action by administering them in the form of hormone replacement or supplementation therapy indicate that they do affect plasma lipid metabolism but not necessarily in ways that can help us understand the mechanisms responsible for the differences in lipoprotein kinetics and metabolism between men and women.

Insights from Exogenous Sex Hormone Administration

The majority of available data regarding the effect of sex steroid administration on plasma lipid homeostasis come from studies evaluating the effects of hormone replacement therapy in postmenopausal (both natural and surgically induced) women and the effects of oral contraceptive use in healthy premenopausal women and women with PCOS and have been reviewed in detail (55–57). These are schematically summarized in Table 2, along with the results from studies evaluating changes in lipid and lipoprotein kinetics. Estrogens do not appear to reproduce the beneficial effects on the plasma lipid profile one might expect from the differences in plasma lipid concentrations between men and women. Briefly, oral estrogen preparations given to postmenopausal women (55), premenopausal women (56), women with PCOS (57), as well as men with prostatic carcinoma (58), and male-to-female transsexuals (not on antiandrogen therapy) (59) increase total plasma and VLDL triglyceride concentrations, increase HDL cholesterol concentration, and reduce total and LDL cholesterol concentrations. Although these effects are somewhat dose-dependent, it is unlikely that they are solely due to supraphysiological availability of estrogens in these studies; although in premenopausal women and men sex steroids were given in addition to their normal physiological availability (56–59), the doses used in postmenopausal women were considered replacement doses (55). The hypertriglyceridemic effect of oral estrogen preparations (55–59) is largely due to increased hepatic secretion of large, triglyceride-rich VLDL particles (60) and therefore associated with both increased hepatic VLDL apolipoprotein B-100 and VLDL triglyceride secretion rates (61–65). Although the production of small VLDL particles is also increased (mainly from large VLDL rather than direct hepatic secretion), this is counterbalanced by equivalent increases in the direct removal rate and the conversion to intermediate-density lipoproteins (IDLs) (60). Overall VLDL triglyceride and VLDL particle clearance rates are not affected by oral estrogen (61–65) because of accelerated conversion of large to small VLDL particles with a simultaneous slowing of the direct conversion of large VLDL to IDL (60). The rise in plasma HDL cholesterol concentration in response to oral estrogen treatment (55–59) is predominantly due to an increase in the cholesterol content in the large, cholesterol-rich HDL₂ subfraction, likely attributed to augmented production rates of HDL apolipoprotein A-I (with little if any changes in HDL apolipoprotein A-II kinetics) combined with no changes or mild increases in HDL apolipoprotein A-I removal rates (62, 66–68), although a few studies have reported increases in HDL apolipoprotein A-I production rate coupled with reductions in clearance (69, 70). The decrease in plasma LDL cholesterol concentration in response to oral estrogen treatment (55–59) is associated with an increase in the LDL apolipoprotein B-100 fractional catabolic rate in large LDL particles that overcomes a smaller increase in LDL apolipoprotein B-100 production rate (the production of small LDL particles is also increased, but this is counterbalanced by equivalent increases in removal rates) (58, 60, 64, 71) that occurs as a result of increased rate of IDL-to-LDL conversion rather than from direct hepatic LDL production (60). Reduced dietary cholesterol absorption may also contribute to the hypocholesterolemic effect of estrogen (71). Although this effect of estrogen therapy is comparatively small, it has been observed after not only oral but also transdermal administration (71) and may therefore account for the modest increases in LDL cholesterol after menopause (33–36, 39) and during the follicular phase of the menstrual cycle (52, 53) when endogenous estrogen availability is low.

Table 2.

Effects of exogenous sex hormones on lipid and lipoprotein metabolism

	Estrogens		Progestogens	Androgens (oral and parenteral)
				Aromatizable		Nonaromatizable
	Oral	Parenteral	Oral and parenteral	Men	Women	Men	Women
Total and VLDL triglyceride concentrations	↑↑	↔ or ↓	↓	↔	↔/↑a	↔	↔
VLDL triglyceride production	↑↑	?	↔	?	↑↑	?	?
VLDL triglyceride removal	↔	?	↑	?	↑	?	?
VLDL apolipoprotein B-100 production	↑↑	↔	↔	↔	?	?	?
VLDL apolipoprotein B-100 removal	↔	↔	↑	↔	?	?	?
HDL cholesterol concentration	↑↑	↔ or ↑	↓	↓	↓	↓↓	↓↓
HDL apolipoprotein A-I production	↑↑	↔	↓↓	↓	↓	↓↓	↓↓
HDL apolipoprotein A-I removal	↔	↔	↓	↑	↑	↑↑	↑↑
HDL apolipoprotein A-II production	↔	?	?	?	?	?	?
HDL apolipoprotein A-II removal	↔	?	?	?	?	?	?
LDL cholesterol concentration	↓↓	↔ or ↓	↔ or ↓	↔	↔	↔ or ↑	↔ or ↑
LDL apolipoprotein B-100 production	↑	↔	↔ or ↑	?	?	?	?
LDL apolipoprotein B-100 removal	↑↑	↔	↔ or ↑	?	?	?	?

↑, Increase; ↓, decrease; ↔, no effect; ?, unknown. Double arrows indicate more pronounced effects than single arrows.

^aOral preparations increase, whereas parenteral preparations do not increase plasma triglyceride concentrations in women (see text).

Transdermal administration of 17β-estradiol, which better mimics normal endogenous estrogen delivery, has little or no effect on HDL cholesterol concentration (0–5% increase), only modestly decreases (by < 10%) LDL cholesterol concentration, and does not affect or only modestly decreases (by < 10%) plasma triglyceride concentration in premenopausal (56) and postmenopausal women (55), women with PCOS (57), as well as healthy men (72) and men with prostate cancer (73–75). Similarly, transdermal administration of estradiol, in contrast to oral administration, does not alter plasma VLDL triglyceride and VLDL apolipoprotein B-100 (64), HDL apolipoprotein A-I (67), and LDL apolipoprotein B-100 (64, 71) concentrations and kinetics. The amounts of estradiol administered in these studies ranged from supraphysiological [e.g. plasma estradiol concentration increased to more than 1000 pmol/liter (73, 74)] to doses that resulted in plasma estradiol concentrations that fall within the range of those in healthy premenopausal women during the normal menstrual cycle (76, 77) [e.g. <500 pmol/liter (64, 67, 71)]. It is therefore unlikely that estrogens play a major role in regulating plasma lipid kinetics under normal physiological conditions.

Progestogens (given alone or in combination with estrogens) reduce plasma triglyceride concentration but also decrease plasma HDL cholesterol concentration, irrespective of the route of delivery, dose, and administration pattern (cyclic or continuous), in both premenopausal and postmenopausal women (55, 56, 61, 78–86) but appear to have little effect on plasma LDL cholesterol and LDL apolipoprotein B-100 concentrations, although they increase both the production and the removal rates of LDL apolipoprotein B-100 but do not affect the net balance (82, 86). The hypotriglyceridemic effect of progestogens is mediated by enhanced VLDL triglyceride and VLDL apolipoprotein B-100 plasma clearance rates (61, 78), whereas the HDL cholesterol-lowering effect is possibly associated with decreased production of HDL apolipoprotein A-I, which overcomes a simultaneous reduction in clearance (66). Nevertheless, normal physiological fluctuations of progesterone secretion throughout the menstrual cycle do not affect plasma triglyceride and VLDL triglyceride and VLDL apolipoprotein B-100 concentrations and kinetics (54); there are no studies of LDL and HDL kinetics during the menstrual cycle.

In contrast to the vast amount of information regarding the effect of female sex hormones on lipid metabolism, there are only limited data available with regard to androgens. Evidence so far indicates that androgen receptor activation is most likely responsible for at least part of the sex differences in HDL cholesterol concentration. Testosterone (given orally, transdermally, or im) (87–92) and both aromatizable (e.g. androstenediol and androstenedione, given orally) (93–95) and nonaromatizable (e.g. stanozolol, given orally) (88, 96–98) androgen derivatives reduce (in a dose-dependent manner) HDL cholesterol and apolipoprotein A-I concentrations [due to a combination of both reduced HDL apolipoprotein A-I production rate and increased fractional catabolic rate (96)] in both men and women (premenopausal and postmenopausal). Stanozolol-induced reductions in HDL cholesterol and apolipoprotein A-I concentrations (30–50%) are considerably greater than those induced by testosterone (8–20%) and aromatizable androgens (5–12%), suggesting that aromatization blunts the unfavorable effects of androgens on HDL metabolism (90). On the other hand, testosterone administration is associated with only modest (<5%) reductions in LDL cholesterol concentration when given to hypogonadal men in replacement doses (90, 91); has no effect on LDL cholesterol concentration in eugonadal men (regardless of dose) (90) and previously untreated female-to-male transsexuals receiving moderate to high doses (87, 89); and may increase LDL cholesterol concentration in peri- and postmenopausal women when given orally but not transdermally (92). Therefore, androgen action is most likely not responsible for the sex differences in plasma LDL cholesterol concentration.

The effects of androgen receptor agonists on plasma triglyceride concentration and metabolism are not entirely clear but may be sex-specific. In healthy young and elderly men, administration of neither testosterone (73, 88, 99–101) nor aromatizable androgen derivatives (95) affects plasma triglyceride concentrations [or VLDL triglyceride concentration and VLDL apolipoprotein B-100 concentration and kinetics (101)], irrespective of the route of delivery (whether oral, transdermal, or im), dose (100), and treatment duration (102). It is possible, but unlikely, that the lack of effect is due to testosterone in these studies being available in excess of needs because testosterone therapy has also no effect on plasma triglyceride concentrations in hypogonadal men (89, 91, 102, 103). In women, aromatizable androgens or testosterone given orally raises plasma triglyceride concentrations (92, 104) due to a robust increase in hepatic triglyceride secretion, which overcomes a simultaneous but less intense increase in plasma triglyceride clearance (104), whereas nonaromatizable androgens given orally or testosterone applied transdermally has no effect on plasma triglyceride concentrations (92, 97). Gains in body weight with androgen treatment were small and similar between studies (<5%) (97, 104) and thus not likely to explain the different outcomes. Hence, although the oral route of administration may be responsible for much of the hypertriglyceridemic effect of androgens in women, these observations suggest that aromatization is also likely to be involved in the apparent sex-specific effect of androgens (hypertriglyceridemia in women not in men), because aromatization takes place predominantly in adipose tissue and women have typically more body fat than men. These results may also provide a possible mechanism responsible for the hypertriglyceridemia in obese women with PCOS, who have more efficient peripheral aromatization of androgens to estrogens (105).

Summary and Conclusions

Understanding sex differences in lipid metabolism and the factors involved in the regulation of lipid kinetics and the plasma lipid profile, a major risk factor for cardiovascular disease, is important because there are significant differences in the ways men and women experience cardiovascular disease. Clearly, the traditional view that the female “advantage” (i.e. the cardioprotective lipid profile in premenopausal women) is due to differences in the sex hormone milieu—in particular, the availability of estrogens and androgens (106, 107)—between men and women and premenopausal and postmenopausal women is not entirely supported by the results from studies examining the effects of normal physiological alterations in the hormonal milieu (i.e. due to menopause or throughout the menstrual cycle) and those evaluating the effects of exogenous sex steroid administration on plasma lipid kinetics and concentrations (Tables 1 and 2). Parenterally administered estrogens have either no effect or only very small beneficial effects, whereas orally administered estrogens raise plasma triglyceride concentrations, a phenomenon that is not consistent with the observed sex differences and likely results from the hepatic “first-pass effect.” Similarly, the effects of progestogens and androgens mimic only in part the differences in plasma lipids between men and women and, at least in the case of androgens, may depend on the sex of the subject. The lack, in many cases, of randomized, placebo-controlled, dose-response studies and the obvious difficulties in reproducing differences between men and women in sex hormonal milieu experimentally (e.g. the normal fluctuations of estradiol and progesterone during the menstrual cycle or the circadian rhythm of testosterone) limits interpretation of the available data.

Sexual dimorphism in lipid metabolism therefore seems to be the result of a complex network of hormone action in combination with other, possibly sex-specific, direct or indirect modulators of lipid metabolism. The exact nature of these factors is unclear, but there are seemingly many [e.g. insulin and adipokines (107), gene expression and imprinting (108), etc.] that need to be explored in the future. An obvious candidate would be differences in insulin action between men and women. Insulin is an important regulator of lipid metabolism, and there are differences between the sexes in insulin sensitivity of glucose metabolism in the liver and muscle (109); however, little is known regarding possible sex differences in the regulation of lipid metabolism by insulin (109). It is unlikely that differences in body composition between men and women (110–115), even if these may be causally related to lipid metabolism [e.g. by affecting fatty acid availability for VLDL triglyceride synthesis (116–118)], are responsible for this phenomenon because sex differences in the plasma lipid profile persist even when men and women are matched for percentage of body fat (9). Nevertheless, accumulation of excess body fat appears to affect lipid kinetics differently in men and women (2, 119). The underlying physiological modulators of plasma lipid metabolism responsible for the differences between men and women remain to be elucidated.