Visceral obesity, the metabolic syndrome, androgens and estrogens

Introduction

Adult men and women differ in their fat distribution; the regional distribution of body fat is a characteristic of masculinity and femininity. In premenopausal women a larger proportion of fat is stored in peripheral fat depots such as breasts, hips and thighs. Men tend to deposit excess fat in the abdominal regions (both subcutaneous and intra-abdominal or visceral fat depots) and generally have a larger visceral fat depot than (premenopausal) women. As regional localization of body fat is considered a secondary sex characteristic, it is likely that sex steroids are involved in the male and female patterns of fat deposition. This view is strengthened by the observation that variations in sex steroid levels in different phases of (reproductive) life parallel regional differences in fat storage and fat mobilization. Until puberty, boys and girls do not differ very much in the amount of body fat and its regional distribution, although girls may have somewhat more body fat than boys. From puberty on, differences become manifest. The ovarian production of estrogens and progesterone induce an increase in total body fat as well as selective fat deposition in the breast and gluteofemoral region. Pubertal boys show a strong increase in fat-free mass, while the amount of total body fat does not change very much. Adolescent boys lose subcutaneous fat but accumulate fat in the abdominal region, which, in most boys, is not very visible in that stage of development but is clearly demonstrable with imaging techniques [1]. The sex steroid-induced regional distribution is not an all-or-none mechanism; it is a preferential accumulation of excess fat. Obese men and women still show their sex-specific fat accumulation, but store their fat also in the ‘fat depots of the other sex’. Not only do the fat distributions differ between the sexes from puberty on; the dynamics of fat cell size and fat metabolism are also different. The amount of fat in a certain depot is dependent on the number and size of the fat cells. Fat cells in the gluteal and femoral region are larger than in the abdominal region. The activity of lipoprotein lipase, the enzyme response for accumulation of triglycerides in the fat cell, is higher in the gluteofemoral region than in the abdominal area. Conversely, lipolysis is regulated by hormone-sensitive lipase, which in turn is regulated by several hormones and by the sympathetic nervous system. Catecholamines stimulate lipolysis via the β-adrenergic receptor, while α2-adrenoreceptors inhibit lipolysis. Hormones affect the catecholamine of receptors. Testosterone stimulates the β-adrenergic receptor, while estrogens/progesterone favor α2-adrenoreceptors. Insulin stimulates fat accumulation. It is not an unreasonable speculation that the sex steroid-dependent fat distribution serves (or may be better has served!) the different roles of men and women in reproduction. The visceral fat depot has a high metabolic activity with a high turnover of triglycerides, which sets rather acutely large amounts of free fatty acids (FFA) free. The visceral fat depot drains on the portal vein, which drains its blood in the liver, providing FFA as fuel for quick and high degrees of physical activity. So, the reserve energy supplies of men can be mobilized fast and are readily available to fuel metabolism. Women's fat stores lend themselves less to quick mobilization. Pregnancy and lactation are situations that release energy from female stores at the buttocks and thighs, but do so at a slow pace. Again, these sex differences are not absolute but a matter of predominance.

As indicated above, a large number of studies have documented that visceral obesity is associated with low plasma total testosterone levels [2-5]. Part of the explanation for the low total testosterone levels might be the lower plasma levels of sex hormone-binding globulin (SHBG) encountered in obese men; these men are often hyperinsulinemic which, in turn, might be part of the explanation for the lowered levels of SHBG [5]. Normally, an increase in SHBG levels is observed in aging men, in all likelihood due to a decrease of growth hormone and IGF-I levels with aging [6]. Studies have found that low total and free testosterone levels are associated with coronary artery disease in men and of total testosterone with cardiovascular risk factors in healthy men [7], and that these are not independent variables [8]. From another study it appeared that rather than age per se, it is the value of plasma testosterone that is related to features of the metabolic syndrome [9]. A prospective study found that men with low testosterone and low SHBG levels are more likely to develop insulin resistance and subsequent diabetes mellitus type II [10]. It could be established that sleep apnea and daytime sleepiness is much more related to visceral obesity than to general obesity, and its occurrence related to plasma insulin and leptin levels [11]. Inflammatory cytokines released from adipose tissue produce fatigue and sleepiness in these men.

Relationships, sex steroids and fat distribution in adulthood and aging

While the evidence that pubertal sex steroids induce a sex-specific fat distribution with preferential abdominal/visceral fat accumulation in males and preferential gluteofemoral fat accumulation in females is quite solid, later in life a number of paradoxes occur in the relationship between sex steroids and fat distribution.

Acquired adult onset hypogonadism in men is associated with a higher amount of subcutaneous fat than in eugonadal men, but the amount of visceral fat appears to be not less than in a comparison group of eugonadal men [12]. So, apparently while androgens induce visceral fat accumulation, once fat has been stored in the visceral depot it does not need continued androgen stimulation as opposed to maintenance of bone and muscle mass, which are lower in the men with adult onset hypogonadism than in eugonadal controls [6]. Induction of androgen deficiency in healthy men by administration of a luteinizing-hormone releasing-hormone (LHRH) agonist leads to an increase of fat mass [13]. Correlation studies in large groups of subjects have shown that visceral fat increases with aging. There is an inverse correlation between the amount of visceral fat and plasma insulin on the one hand and levels of testosterone and SHBG on the other. Therefore, the effects of overfeeding and starvation on sex steroid levels are interesting. In an experiment, 12 pairs of identical twins were overfed for 120 days, resulting in an average weight gain of 8 kilograms. The excess fat was accumulated both subcutaneously and abdominally/viscerally, with a clear preference for the latter localization. Plasma testosterone levels did not decline but SHBG levels declined. Even though plasma testosterone did not decline, there was a significant negative correlation between changes of visceral fat and of plasma testosterone. In other words: a gain in visceral fat led to a decrease in plasma testosterone. A further correlation that appeared was between fasting insulin and changes in abdominal fat and in testosterone [14]. The latter is interesting in view of the fact that plasma levels of insulin are negatively correlated with SHBG [6],[7]. In vitro studies have shown that insulin inhibits hepatic production of SHBG. Lower SHBG levels lead to a fall of total plasma testosterone, since this leads to an increase in free testosterone in the first instance and subsequently to a higher metabolic breakdown of testosterone. It could be shown that a higher degree of visceral obesity was correlated with lower SHBG levels and with higher levels of 3α-diol-G, a metabolite of testosterone, indicating that a lowering of SHBG induces testosterone metabolism. Plasma insulin was positively associated with 3α-diol-G [15].

Conversely, weight loss either produced an increase in testosterone itself [14],[16] and/or an increase in SHBG [16],[17]. The study of Pritchard et al. [14] found that the rise in testosterone levels correlated inversely with loss of abdominal fat and with plasma insulin. Collectively, these studies suggest that a high degree of visceral adiposity is associated with high insulin levels and low SHBG levels, low total plasma testosterone, and with an increase of metabolites of testosterone, and vice versa.

Correlation studies cannot unravel the cause-and-effect relationships between the correlates, whether low testosterone induces visceral fat deposition or whether a large visceral fat depot leads to low testosterone levels. Prospective studies have confirmed that lower endogenous androgens predict central adiposity in men [3], and that these low testosterone levels are significantly inversely associated with levels of blood pressure, fasting plasma glucose and triglyceride, body mass index and HDL-cholesterol [18]. A recent study in Japanese–American men found that low testosterone levels predicted an increase in visceral fat 7.5 years later [19]. Unfortunately, SHBG levels were not measured, but baseline testosterone levels were significantly correlated with fasting C-peptide and fasting insulin levels, while adjustment for baseline visceral fat diminished the association between baseline testosterone and plasma insulin, evidencing a role of the amount of visceral fat in their interrelationship. The amount of visceral fat at baseline also predicted an increase in visceral fat over the follow-up period of 7.5 years. The latter is important since in all follow-up studies the hypotestosteronemia associated with visceral obesity may already have been presented when the subjects were initially included in the study. It does not come as a surprise that subjects with a degree of visceral obesity at a younger age show an increase thereof in their later life, as was the case in the study of Tsai et al. [19]. A five-year follow-up study of Swedish men indicated that elevated plasma cortisol and low testosterone were prospectively associated with an increased incidence of cardiovascular-related events and diabetes mellitus type II [20]. A cautionary note is that middle-aged and elderly men with risk factors for cardiovascular disease usually have lower plasma testosterone levels than healthy men, and these epidemiological findings may be self-serving in their predictions. The already existing excess of visceral fat in the men in the study may have been the mechanism of low testosterone in these men. It is of note, that with aging, plasma testosterone levels show also a decline independently of the amount of visceral fat.

Elevated androgen levels in women increase the amount of visceral fat, women with polycystic ovarian syndrome being the classic example [21]. In the medical literature, this is often presented as a paradox, in the sense that high levels of androgens in women and low levels of androgens in men are associated with visceral obesity. The paradox is partially semantic: a testosterone level of 3–5 nmol/l is very high for an adult women but would mean hypogonadism in men. So the use of the terms high and low testosterone must be related to the sex. A further element is the relation to age. Women with polycystic ovarian syndrome are relatively young (of reproductive age) when they come to the attention of the medical profession, while the relationship between visceral obesity and low androgen levels in men is typically an epidemiological finding in elderly men. Apparently, similar to the situation in androgen naive teenage boys, androgens in women with polycystic ovarian syndrome are capable of visceral fat accumulation when these women are exposed to androgens postpubertally, when their polycystic ovaries start to produce androgens.

In postmenopausal women there is a larger degree of (male type) upper body fat accumulation in comparison to the former gluteofemoral fat storage. Following female menopause, the androgen levels drop considerably. So, obviously premenopausal estrogen/progesterone levels are required to maintain a premenopausal female type of fat distribution, as also evidenced by the fact that postmenopausal hormone replacement (partially) restores the premenopausal fat distribution [22].

Increasing testosterone and the effects on adipose tissue

It is a widely held belief that androgens have an atherogenic effect and thus lead to cardiovascular disease. Two review papers have examined this relationship and conclude that it is not tenable to link testosterone to the etiology of cardiovascular disease [23],[24].

Rather, it appears that lower-than-normal testosterone levels in men are associated with cardiovascular risk factors, of which visceral obesity with its association with cardiovascular risk factors, might be the intermediate. Visceral adiposity is associated with low testosterone levels in cross-sectional studies, and from this type of study, the cause-and-effect relationship is not immediately evident. Does visceral adiposity induce low levels of testosterone, or do low levels of testosterone induce visceral adiposity? Some prospective studies argue that low testosterone predicts visceral adiposity [3],[12],[18],[19],[25-28]. Several studies have examined the effects of testosterone administration on adipose tissues in men. Some have found a decrease in fat mass with testosterone replacement therapy [12],[25-28], while others found no change [29],[30]. But longer-term studies of testosterone supplementation of older men have consistently demonstrated a decrease in fat mass [31-34]. In a recent study wherein young men were made testosterone deficient with an LHRH agonist, testosterone deficiency was associated with gains in subcutaneous, intermuscular and intra-abdominal fat. Graded restoration of plasma testosterone with exogenous testosterone led to a dose-dependent reduction of adipose tissue with no quantitative differences between trunk and peripheral non-trunk fat mass [35].

Testosterone inhibits the expression of the activity of lipoprotein lipase, the main enzymatic regulator of triglyceride uptake in the fat cell, preferentially in abdominal fat and less so in femoral fat and, maybe, mobilizes lipids from the visceral fat depot. A study of testosterone administration restoring testosterone levels to mid-normal values with a duration of 8–9 months [34] found a decrease of the visceral fat mass, a decrease of fasting glucose and lipid levels and an improvement of insulin sensitivity; in addition, a decrease in diastolic blood pressure was observed [34]. It is not well established whether androgen treatment improves insulin sensitivity, a strong predictor of diabetes mellitus and cardiovascular disease, of which the improvement is a meaningful intervention to reduce the risk of disease [36]. Some studies have found indications of a beneficial effect on insulin sensitivity [28]; others do not confirm this result [37]. More studies are needed to examine whether restoring testosterone levels to normal in viscerally obese men will lead to a reduction of visceral fat and an improvement of the cardiovascular and diabetogenic risk factors associated with it, but there is certainly an increased interest [38],[39].

Estrogens and the metabolic syndrome

Estrogens in men, which are largely a product of peripheral aromatization of androgens, receive increasing attention. Traditionally conceptualized as ‘female hormones’, estrogens appear to have unexpected but important effects on the male reproductive/metabolic system (for review [40],[41]). The discovery of one man with deficient estrogen action (on the basis of a receptor mutation) and four others with deficient estrogen synthesis, and further studies in knockout mouse models, strongly suggest that in the virtual absence of estrogen action, symptoms of the metabolic syndrome develop, such as atherosclerosis at an early age, insulin resistance, visceral obesity, and a high-risk lipid profile are present, which improved upon administration of estrogens in the men with deficient estrogen synthesis [41],[42]. It would seem that low-to-absent estrogen action in men is a factor in the metabolic syndrome. In support of a role of estrogens is the increased risk of cardiovascular disease in men with a variation in the estrogen receptor alpha gene [43]. The circulating testosterone levels in these men were high rather than low and symptoms deteriorated upon administration of testosterone. So, this observation distinguishes these cases from the men studied in population studies which predict an increased cardiovascular risk in men with low testosterone levels [3],[12],[18],[19],[25-28] and improvement upon testosterone administration [12],[25-28]. In studies which found that low testosterone levels predict an increased cardiovascular risk [3],[12],[18],[19],[25-28] plasma estradiol levels were certainly not low. Plasma estradiol levels in men show no tendency to decline with aging [40]. So, a number of paradoxes are presented here, for which presently no plausible explanation exists. Speculatively, it might be that the protective effects that androgens seem to have on the occurrence of the metabolic syndrome are only manifest in an endocrine milieu where there is a normal degree of biological actions of estrogens [42], potentially through interaction of estrogens with androgen receptors [44].