| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Editorial |
College of Physicians and Surgeons Columbia University New York, New York 10032
Address all correspondence and requests for reprints to: Dr. John P. Bilezikian, College of Physicians and Surgeons, Columbia University, New York, New York 10032.
The issue is estrogens in men. The new paradigm is that estrogens are important for optimal accrual and maintenance of the male skeleton (1, 2). It is an interesting twist from previous suppositions that placed androgens in a position of primacy for male skeletal health and estrogens in a similar position for female skeletal health. Experiments of nature in human subjects have shattered this convenient classification, instructing us that the matter is neither so simple nor necessarily so intuitive.
Most estrogens in the male skeleton are derived from peripheral aromatization of circulating testosterone and other C19 steroids by the actions of the cytochrome P450 enzyme system. Aromatase, the responsible gene product, catalyzes hydroxylation reactions, leading to the production of C18 estrogenic steroids. Estrogens are active in the skeleton by virtue of their effects through estrogen 
and ß receptors and perhaps also through nongenomic pathways (2). The cellular bases for estrogen’s effects are still obscure, but clearly receptors exist in osteoblasts and bone marrow stromal cells, in addition to which estrogens have indirect actions on skeletal metabolism by influencing the production of bone-resorbing cytokines (3). They have proapoptotic effects on osteoclasts and antiapoptotic effects on osteoblasts and osteocytes (4).
We know what we know about this subject in part because of inactivating mutations in human male subjects that have involved the estrogen gene (5) or the aromatase gene (6, 7, 8, 9). In both kinds of inactivating mutations a distinctive phenotype includes tall stature, open epiphyses, continued longitudinal growth, delayed bone age, lack of pubertal growth spurt, eunuchoid proportions, genu valgum, elevated bone turnover markers, and markedly reduced bone density. These observations are all the more noteworthy because androgen levels are typically normal or elevated in these men. When men with aromatase deficiency are treated with estrogen, epiphyses close quickly, longitudinal growth ceases, and bone mineral density (BMD) increases dramatically (7, 8). A recent report of an aromatase deficient 16-yr-old male (10) provides evidence that estrogens may also be important for periosteal bone growth. We previously thought that periosteal bone growth was the province of androgens. The observations of Bouillon et al. (10), however, suggest that periosteal bone growth is also influenced by estrogen in the growing male skeleton. Using peripheral quantitative tomography, the gain in areal bone density with estrogen replacement, as determined by dual-energy x-ray absorptiometry, was shown to be due to gains in cross-sectional area and cortical thickness, not to changes in volumetric density. The increase in areal BMD was mainly driven by an increase in bone size. As proposed by Bouillon et al. (10), it is attractive to consider biphasic actions of estrogens on periosteal bone growth, with lower amounts in the prepubertal skeleton aiding periosteal apposition, whereas at higher concentrations in late pubertal boys and adults they act as an inhibitor of periosteal bone growth.
In addition to the importance of estrogens in accruing bone mass in the male skeleton, it is also now widely accepted that estrogens are important in maintaining the male skeleton (11, 12, 13). This has been a difficult proposition to document, because in the aging male, both estrogen and androgen levels decline, particularly their free bioavailable fraction. In part, this is due to increases in SHBG levels. Cross-sectional and longitudinal studies in middle-aged and elderly men have attributed the decline in BMD more to declining estrogen levels than to declining androgen levels (11, 12, 13, 14). Moreover, men who sustain fractures have lower levels of total and bioavailable estradiol than men who do not fracture (15). Dynamic short-term studies in which male subjects were temporarily rendered biochemically hypogonadal by pharmacological means have permitted a specific sex steroid assignment to several skeletal dynamics (16, 17, 18). Estradiol, but not testosterone, completely prevents the increase in bone turnover after the hypogonadal state is induced, whereas both estrogen and testosterone each seem to have independent effects on bone formation (16, 17, 18).
The natural decline of estrogen levels in the aging male is clearly quite variable. Among a host of factors that may be responsible for this, such as declining androgen levels and changes in SHBG, the aromatase enzyme itself may play a role. The human CYP19 gene that harbors the cytochrome P450 system has great polymorphic variability (2). An association between one particular polymorphism, namely a tetranucleotide repeat sequence (TTTA) has been related to circulating estradiol levels in the male (19, 20, 21). The shorter repeat sequences have been shown to be more likely to be associated with lower BMD, greater bone loss, higher bone turnover, and fractures than longer repeat sequences (20, 21). Greater in vitro aromatase efficiency and higher estrogen production have been reported when fibroblasts obtained from high TTTA repeat sequence genotypes are compared with fibroblasts obtained from low TTTA repeat sequence genotypes (21). Curiously, the clinical expression of these associations appears to be influenced by body fat. Because fat is a major source of aromatase activity, an abundance of fat could produce sufficient amounts of circulating estradiol to prevent bone loss even if the subject is genetically endowed with the shorter TTTA repeat sequence.
There may be a critical level of bioavailable estrogen below which men show higher rates of bone loss and increased bone turnover (22). The median value of 40–55 pmol/liter has been demonstrated in a number of studies to correlate with this threshold effect (12, 22, 23, 24). About 50% of men fall below this estradiol threshold, helping to account for the important role of declining estrogen levels in many men as they age. Furthermore, in limited trials with raloxifene, a selective estrogen receptor modulator, men whose levels fell below this threshold had an agonist response to the selective estrogen receptor modulator, whereas men whose estrogen levels were higher did not respond (25, 26, 27).
Now we come to the article by Orwoll et al. (28) in this issue of the JCEM that provides more information about the subject at hand. The cohort reported by Orwoll et al. (29) is the osteoporotic fractures in men (MrOS) study, an ongoing, large, epidemiological study of aging men. In the United States, approximately 6000 men are being studied at six medical centers. Using a stratified sample design, 2643 participants for whom sex steroids were measured are reported in this paper. The medium age of the sample was 73 yr, with 15% of men 80 yr or older. This is the largest cohort of men in whom sex steroid measurements are available. MrOS will be the male equivalent of the Study of Osteoporotic Fractures that has provided over the past 10 yr a wealth of data on the aging female skeleton (30, 31). The data from MrOS are only now becoming available. It is understandable, therefore, that the early reports are preliminary. In fact, the paper in this issue of the JCEM reports only on changing sex steroid and SHBG levels and associated interrelationships. There are no data referable to the skeleton or to other organ systems that might be affected by estrogens or androgens. This is disappointing, because we can do no more than comment on the levels and relationships themselves as there are no surrogate or true end points with which to relate the data.
In MrOS, levels of testosterone and estradiol decline with age. The declines are modest, at best, for total sex steroid levels and somewhat greater when free steroid levels are measured. Similar modest declines in bioavailable testosterone and estradiol for a cohort over 65 yr have also been reported by Ferrini et al. (32). At any of the 5-yr cutoff points, there was considerable variability. With age, a greater proportion of men were shown to have levels of testosterone less than 300 ng/dl, with the group as a whole showing a percentage of 17% with levels below this value. Of additional interest were the levels of circulating estradiol in this population. Forty-three percent of the entire MrOS population had levels that were greater than 40 pmol/liter, the level noted above to be a threshold value. A majority of men (64%) over 80 yr had levels below this putative threshold. Not surprisingly, higher body mass index values were related to lower testosterone and higher estradiol levels, perhaps a function of adipocyte aromatase activity. The multivariate analyses reveal SHBG as an important component of the total testosterone level. However, for free testosterone levels, only 18% of the variance could be explained by SHBG, age, body mass index, or free estradiol levels. A larger proportion of free estradiol levels was related to free testosterone and SHBG levels. The negative relationship between estradiol and SHBG raises questions about whether SHBG has an effect, itself, on estradiol levels. It is consistent with the report by de Ronde et al. (33) but contrary to other studies in which SHBG has been viewed as an estrogen enhancer (34).
This paper points out, but does not help to elucidate, the basis for the age-related variability in testosterone and estrogen levels. To a certain extent, estrogen levels are related to testosterone. The possibilities include variability in aromatase gene activity by virtue of either the genetic polymorphism of the TTTA repeat sequence or other polymorphisms. Those with higher TTTA repeat sequences would be expected to have higher estrogen levels and, therefore, be relatively protected from age-related declines. It is also possible that the various promoters of the aromatase gene, of which there are many (35, 36), might be regulated by various tissue-specific local factors. Moreover, the circulating levels only begin to tell the story at the tissue level, where aromatase activity might be more determinant. It is also important to consider the clearance rate of estradiol, as pointed out by the authors, as another source of variability of the data. A concern, also duly noted by the researchers, relates to the measurements of sex steroids. Because this is a controversial area (37, 38), the authors have taken great effort to ensure that the data will not be confounded by this source of variability.
One would have liked to see relationships established between sex steroid measurements and the skeleton. Another important consideration is the relationship between these levels and metabolic parameters. Returning to the human experiment of nature, the aromatase-defective men, several studies have documented metabolic abnormalities (6, 39). In the study by Maffei et al. (39), insulin resistance, diabetes mellitus type 2, acanthosis nigricans, liver steatohepatitis, and signs of precocious atherogenesis were documented in the 29-yr-old affected subject. When he was treated with estrogen, these measures of metabolic disarray as well as two carotid plaques improved.
Hopefully, in reports to come, Orwoll and his group will be able to provide more definitive data with regard to the longitudinal course of the sex steroids and their interrelationships in this large cohort of aging men. In addition, we need information about key end points of these hormones, namely, the skeleton and metabolism. MrOS is ideally suited to enlighten us. In these respects, perhaps MrOS will confirm that what is good for the goose’s skeleton is also good for the gander’s skeleton!
Footnotes
Abbreviation: BMD, Bone mineral density.
Received January 17, 2006.
Accepted January 27, 2006.
References
gene polymorphisms. Bone 32:412–420[Medline]
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
