0013-7227/02/$15.00/0
Printed in U.S.A.
The Journal of Clinical Endocrinology & Metabolism 87(4):1443–1450
Copyright © 2002 by The Endocrine Society
CLINICAL REVIEW 144
Estrogen and the Male Skeleton
SUNDEEP KHOSLA, L. JOSEPH MELTON III,
AND
B. LAWRENCE RIGGS
Endocrine Research Unit (S.K., B.L.R.) and Department of Health Sciences Research (L.J.M.), Mayo Clinic and Mayo
Foundation, Rochester, Minnesota 55905
Because estrogen (E) and T are the major sex steroids in
women and men, respectively, the traditional view had been
that E primarily regulated bone turnover in women and T
played the analogous role in men. The description of ERdeficient and aromatase-deficient males, however, initiated a
major shift in our thinking on the relative roles of T and E in
regulating the male skeleton, because these individuals all
had unfused epiphyses, high bone turnover, and osteopenia.
Similar, albeit less striking, findings were noted in mouse
models with knock-out of either the ER-␣ or the aromatase
genes. Although these human experiments of nature and
mouse knock-out models clearly demonstrated an important
role for E in the growth and maturation of the male skeleton,
they did not define the role of E vs. T in regulating the adult
male skeleton. The past several years have witnessed an ac-
E
STROGEN (E) DEFICIENCY has a dramatic, and sometimes devastating, effect on the female skeleton. Indeed, loss of E after natural or surgical menopause is the
single most important factor in the development of postmenopausal bone loss and osteoporosis (1). Consequently,
there has been an enormous amount of research performed
over the past decade on E regulation of the female skeleton.
By contrast, other than under the circumstances of medical
or surgical castration, men do not have the equivalent of a
menopause and thus do not manifest a rapid phase of bone
loss, although they clearly do undergo age-related bone loss
(2–5). Because serum total T levels decline only marginally
with age in men (6), it was assumed that men remain gonadally competent throughout their lives. Moreover, because
T is the major circulating sex steroid in men, it was also
assumed that the loss of T caused the increased bone resorption and bone loss in castrated men. Although T had long
been known to be aromatized to E in men by both the testes
and peripheral tissues (7), the skeletal effects of E in men
were largely ignored.
This traditional view of the role of T and E in bone metabolism in men has undergone a fairly dramatic paradigm
shift since 1994, when the sentinel case of the ER-negative
male was described (8). This led to a radical change in our
thinking about sex steroid regulation of the male skeleton,
which was fueled further by the description of two
Abbreviations: ARKO, Aromatase knock-out; BMD, Bone mineral
density; E, estrogen; ␣ERKO, ER-␣ knock-out; ␤ERKO, ER-␤ knock-out;
DERKO, ER-␣/␤ double knock-out; Dpd, deoxypyridinoline; NTx, Ntelopeptide of type I collagen; PINP, N-terminal extension peptide of
type I collagen.
cumulation of evidence from observational as well as direct
interventional studies that now clearly indicates that E plays
a major, and likely dominant, role in bone metabolism in men.
These data also suggest that a threshold level of bioavailable
(or non-SHBG bound) E is needed for skeletal E sufficiency in
the male, and that with aging, an increasing percentage of
elderly men begin to fall below this level. It is this subset of
men who may be at greatest risk for the development of agerelated bone loss and osteoporosis. Moreover, these men may
also be the ones most likely to respond favorably to treatment
with selective E receptor modulators, or perhaps even to T
replacement, because the skeletal effects of the latter may be
mediated largely via aromatization to E. (J Clin Endocrinol
Metab 87: 1443–1450, 2002)
aromatase-deficient males (9 –11), mouse knock-out models
and studies in rats using an aromatase inhibitor (12–22),
observational and interventional human studies (6, 23–31),
and a provocative new idea about gender-neutral effects of
sex steroids on bone (32). Thus, several independent lines
of investigation have significantly altered our understanding
of the role of sex steroids in regulating bone metabolism in
males (33, 34). This review summarizes the key studies that
led to this paradigm shift in our thinking, as well as some
recent evidence suggesting that the male skeleton may require a threshold level of E for preventing increased bone
resorption and bone loss.
Human experiments of nature
During the course of evaluating a tall man aged 28 yr for
genu valgum, Smith et al. (8) noted that the patient had
unfused epiphyses and, on further evaluation, E levels that
were well into the premenopausal female range despite no
evidence of feminization. This constellation of findings correctly suggested a heretofore undescribed syndrome of E
resistance, which was confirmed by direct sequence analysis
of his ER gene [which we now know as ER-␣, as opposed to
the more recently described ER-␤ (35)]. Thus, he had homozygous cytosine to thymine substitutions at codon 157,
resulting in a premature stop codon. Although this sentinel
case has provided important insights into a number of aspects of E action in males, perhaps none has been as dramatic
as the impact this patient had on our understanding of E
regulation of bone metabolism in men.
Despite having elevated E levels (for a male) and normal
T levels, this individual had unfused epiphyses, elevated
1443
1444
J Clin Endocrinol Metab, April 2002, 87(4):1443–1450
Khosla et al. • Clinical Review
markers of bone remodeling, and marked skeletal osteopenia
(Table 1) (8). As would be expected with complete E resistance, E treatment had no effect on any of these parameters.
Soon after the description of this case, two cases of complete
aromatase deficiency were described (9 –11), with virtually
identical skeletal phenotypes. These men had normal or elevated T levels but low E levels, and E treatment resulted in
epiphyseal closure, suppression of bone resorption, and
marked increases in bone mass (Fig. 1) (10, 11). These human
experiments of nature taught us, therefore, that inability
either to respond to or to produce E resulted in dramatic
skeletal consequences in males, despite the presence of normal, or even supranormal, T levels. E was clearly important,
indeed critical, for development of the male skeleton. Even
in retrospect, this was not what most of us would have
predicted.
Yet, the counterpart of the ER-mutant male, the testicular
feminization males, with varying (and sometimes complete)
androgen insensitivity, were also telling us something important about the role of T in the male skeleton. Failure to
respond to T or other androgens did not delay epiphyseal
closure, but testicular feminization males also appeared to
have reduced bone mineral density (BMD), with the patients
with complete androgen insensitivity having lower BMD
than those with partial defects (36). Although the deficits in
BMD in the testicular feminization males may have been
related, in part, to inadequate treatment or compliance with
E replacement therapy after orchidectomy, it is likely that the
androgen insensitivity itself also played a role.
Mouse knock-out models and studies in rats using an
aromatase inhibitor
Studies using knock-out mice, along with data from rats
treated with an aromatase inhibitor and the testicular feminization rat, have also provided useful information on the
role of E and T, at least in the male rodent skeleton. However,
for reasons that remain unclear, the mouse knock-out models
have generally shown much less dramatic phenotypes than
the corresponding human mutations. Thus, ER-␣ knock-out
(␣ERKO) and ER-␣/␤ double knock-out (DERKO) mice have
shortened femoral length (12, 13) and decreased appendicular bone growth that is greater in females than in males (12,
TABLE 1. Bone turnover markers and BMD in the ER-negative
male
Patient
Serum
Osteocalcin (ng/ml)
Bone-specific alkaline phosphatase
(ng/ml)
Urine
Pyridinium (nmol/mmol Cr)
D-Pyridinium (nmol/mmol Cr)
N-telopeptide (nmol BCE/mmol Cr)
Spine BMD (g/cm2)
Normal range
18.7
34.2
3–13
4.3–19.0
110
32
248
0.745
20– 61
4–9
23–110
a
BCE, Bone collagen equivalent; Cr, creatinine.
a
Spine BMD was 3.1 SD below the mean for age-matched normal
women and more than 2 SD below the mean for 15-yr-old boys (the
patient’s bone age).
[Adapted with permission from E. P. Smith et al.: N Engl J Med
331:1056 –1061, 1994 (8). Massachusetts Medical Society.]
14). However, ␣ERKO males have cortical osteopenia and
increased bone turnover that is greater in the male than in the
female (12). In contrast, the ER-␤ knock-out (␤ERKO) males
have a skeletal phenotype that is indistinguishable from that
of the wild-type (13), whereas the ␤ERKO females have an
increase in cortical bone that is associated with increased
periosteal apposition (15, 16). ␤ERKO females also do not
appear to have the age-related reductions in cancellous bone
that are present in wild-type mice (15, 16), suggesting that
either ER-␤ may be permissive for age-related bone loss in
female mice or deletion of ER-␤ may lead to enhanced sensitivity of bone to ER-␣, and hence to an increase in E action
despite age-related decreases in serum E levels. Finally,
ovariectomy results in the same degree of bone loss in adult
DERKO mice as in wild-type females, and the bone loss can
be prevented by E treatment, although 5-fold higher doses of
E are needed in the DERKO compared with the wild-type
mice (17). Whether this is due to sex-nonspecific, nongenomic mechanisms involving E action via the androgen
receptor (32) or to the presence of ER-␣ splice variants in the
DERKO mice that allow partial sensitivity to E is at present
unclear. Indeed, a recent study has found that osteoblasts do
express a shorter form of ER-␣ (of ⬃46 kDa, in contrast to the
full-length form of 66 kDa) that arises through alternate
splicing (37). This short form does possess at least partial
transcriptional activity and is, in fact, present in the bones of
␣ERKO mice (37). These caveats notwithstanding, the mouse
knock-out models do indicate that E plays an important role
in the male skeleton and that ER-␣ is the primary receptor
mediating E actions on bone, with ER-␤ playing a more
limited role and perhaps, in fact, repressing E action.
The aromatase gene has also been knocked out (ARKO)
and, similar to the humans with aromatase deficiency, ARKO
male mice have osteopenia (18, 19). However, whereas one
report found that bone turnover was reduced in these mice
(18), a second report noted increased bone turnover in the
male ARKO mice (19), similar to the human males with
aromatase deficiency. The findings in the ARKO mice essentially confirm previous studies in which treatment with
an aromatase inhibitor decreased BMD and bone size in
growing rats (21). In addition, aged male rats treated with an
aromatase inhibitor were found to have reductions in BMD
and increased indices of bone resorption (20, 22) Finally,
testicular feminized rats lacking the androgen receptor have
a female skeletal phenotype, but are not osteopenic (38).
Indeed, the major defect in these animals appears to be a
reduction in bone size, rather than deficits in bone density.
Thus, deletion of ER-␣ or aromatase in the male mouse
results in osteopenia, although the findings in the mouse
knock-out models have generally not been as striking as in
the human ER-mutant or aromatase-deficient males. In addition, studies using an aromatase inhibitor clearly demonstrate that E regulates bone metabolism in both the growing
and mature male rat.
Studies in adult men
As important as the human ER-␣ mutant, the aromatasedeficient, and the testicular feminization males were in enhancing our understanding of E and T regulation of bone
Khosla et al. • Clinical Review
J Clin Endocrinol Metab, April 2002, 87(4):1443–1450 1445
FIG. 1. Changes in BMD in an aromatasedeficient male treated with E (0.3 mg/d of conjugated estrogens initially, with a gradual
increase to 0.75 mg/d). [Adapted with permission from J. P. Bilezikian et al.: N Engl J Med
339:599 – 603, 1998 (11). Massachusetts Medical Society.]
TABLE 2. Spearman correlation coefficients relating rates of change in BMD at the radius and ulna to serum sex steroid levels among a
sample of Rochester, Minnesota, men stratified by age
Young
T
E2
Estrone
Bioavailable T
Bioavailable E2
Middle-aged
Elderly
Radius
Ulna
Radius
Ulna
⫺0.02
0.33b
0.35c
0.13
0.30b
⫺0.19
0.22a
0.34b
⫺0.04
0.20
⫺0.18
0.03
0.17
0.07
0.14
⫺0.25
0.07
0.23a
0.01
0.21a
a
Radius
Ulna
0.13
0.21a
0.16
0.23b
0.29b
0.14
0.18a
0.14
0.27b
0.33c
a
P ⬍ 0.05; b P ⬍ 0.01; c P ⬍ 0.001.
[Reproduced with permission from S. Khosla et al.: J Clin Endocrinol Metab 86:3555–3561, 2001 (23). The Endocrine Society.]
metabolism in males, what was being observed were E and
T effects on the growing, immature skeleton. Thus, the severe
deficits in bone mass in the ER-␣ mutant and aromatasedeficient males clearly reflect the importance of E in the
acquisition of peak bone mass. These effects of E on the
growing skeleton cannot, however, be directly extrapolated
to men with mature, adult skeletons. As such, the human
experiments of nature left unresolved the question of what
role E played in regulating bone metabolism in adult men,
as well as whether E (or E deficiency) had any role in mediating age-related bone loss in men. Answers to these questions required extensive clinical studies in adult and aging
men.
Although several studies had measured T and E levels in
men and related these to BMD with somewhat inconclusive
results (39, 40), more recent studies using newer, more sensitive assays, particularly assays for E2 capable of accurately
assessing the relatively low circulating E2 levels present in
men, have found fairly consistent results. Thus, virtually all
of the recent cross-sectional observational studies done in
men have found that E2, and particularly the non-SHBG
bound (or bioavailable) E2 levels, correlated better with BMD
at multiple sites than either total or bioavailable T levels (6,
23–30). Moreover, even in a group of T-deficient men, selected specifically to have T levels below 10.4 nmol/liter (300
ng/dl), Amin et al. (29) found that serum E2 levels were much
more robust predictors of BMD than serum T levels. Consequently, there is now little doubt that E is a better predictor
of BMD in men than T, at least as judged from cross-sectional
data. Moreover, although total T and E2 levels change little
over life in men, the bioavailable fractions of both T and E2
decline markedly (by ⬃70 and 50%, respectively), due prin-
cipally to a marked age-related increase in serum SHBG
levels in men (6, 23). This further raises the possibility that
these declining sex steroid (and particularly declining bioavailable E2) levels may contribute to bone loss in aging men.
Cross-sectional observational studies, however, clearly
have limitations, because they cannot fully dissociate possible E effects on bone mass acquisition early in life (which
we know is E-dependent from the human experiments of
nature) from its effects on bone loss in senescence. To address
this issue, we recently reported results from a prospective
study in which the gain in BMD in young (age, 20 – 40 yr) men
vs. the loss in BMD in elderly (age, 60 –90 yr) men was related
to sex steroid levels (23). The findings in this study were also
fairly unequivocal: both the increase in BMD in the young
men and the decline in BMD in the elderly men were most
closely associated with serum E2 levels, particularly the bioavailable E2 levels in the elderly men (Table 2). These longitudinal data thus indicate not only that E is important for
optimal bone mass acquisition in young adulthood, but also
that declining levels of bioavailable E2 may contribute substantially to age-related bone loss in men.
The weakness of observational data (even if they are longitudinal) is that correlation does not prove causality. For
that, one has to intervene by perturbing one variable and
then directly assessing the effect on the variable of interest.
Thus, to establish a causal role for E in regulating bone
turnover and to quantify the relative contributions of E vs. T
in determining bone resorption and formation in men, we
pharmacologically altered T and E levels in elderly men and
assessed the impact of these perturbations on markers of
bone turnover (31). Endogenous T and E production was
completely eliminated in 59 elderly men (mean age, 68 yr)
1446
J Clin Endocrinol Metab, April 2002, 87(4):1443–1450
FIG. 2. Percentage changes in bone resorption markers (urinary Dpd
and NTx) (A) and bone formation markers (serum osteocalcin and
PINP) (B) in a group of elderly men (mean age, 68 yr) made acutely
hypogonadal and treated with an aromatase inhibitor (group A), E
alone (group B), T alone (group C), or both E and T (group D). See text
for details. Asterisks indicate significance for change from baseline: *,
P ⬍ 0.05; **, P ⬍ 0.01; ***, P ⬍ 0.001. [Adapted with permission from
A. Falahati-Nini et al.: J Clin Invest 106:1553–1560, 2000 (31). American Society for Clinical Investigation.]
using a combination of a GnRH agonist and an aromatase
inhibitor, and the men were initially studied under conditions of full sex steroid replacement at physiologic doses
using T and E patches. After collection of baseline urine and
blood samples for assessment of markers of bone turnover,
the men were randomized to one of four groups, with all four
groups continuing treatment with the GnRH agonist and the
aromatase inhibitor: group A (⫺T, ⫺E) had both T and E
patches withdrawn; group B (⫺T, ⫹E) stopped the T patch
but continued the E patch; group C (⫹T, ⫺E) continued the
T patch but stopped the E patch; and group D (⫹T, ⫹E)
continued both patches. After 3 wk of the respective treatments, the baseline studies were repeated in all subjects.
Figure 2A shows the impact of these interventions of the
bone resorption markers, which increased significantly in
group A (⫺T, ⫺E), but remained unchanged in group D (⫹T,
⫹E). E alone (group B) was almost completely able to prevent
the increase in bone resorption markers, whereas T alone
(group C) was much less effective. In fact, using a two-factor
ANOVA model, the E effect on urine deoxypyridinoline
(Dpd) and N-telopeptide of type I collagen (NTx) excretion
was highly significant (P ⫽ 0.005 and 0.0002, respectively)
with at best a borderline T effect on NTx excretion (P ⫽ 0.232
and 0.085 for Dpd and NTx, respectively). On the basis of
these data, we estimated that, in normal elderly men, E
accounted for approximately 70% of the total effect of sex
steroids on bone resorption, with T contributing (in the absence of aromatization to E) at most 30% of the effect. In a
subsequent study, Leder et al. (41) have been able to demonstrate an independent effect of T on bone resorption, but
because that study lacked an E-alone group, the relative
contributions of T vs. E toward the regulation of bone resorption could not be quantified.
Khosla et al. • Clinical Review
FIG. 3. BMD in 596 men, aged 51– 85 yr, based on quartiles of serum
bioavailable E2 levels, after adjustment for age and body weight. A,
Total hip BMD (F ⫽ 5.14; P ⬍ 0.002); B, distal forearm BMD (F ⫽ 4.99;
P ⫽ 0.002); C, whole body bone mineral content (F ⫽ 3.15; P ⬍ 0.03).
[Reproduced with permission from P. Szulc et al.: J Clin Endocrinol
Metab 86:192–199, 2001 (30). The Endocrine Society.]
Figure 2B shows the changes in the bone formation markers in the four groups. Both serum osteocalcin and N-terminal extension peptide of type I collagen (PINP) levels decreased significantly in group A (⫺T, ⫺E), indicating an
important effect of sex steroids in maintaining bone formation. These findings are consistent with the induction of
osteoblast apoptosis that has been observed in the setting of
sex steroid withdrawal both in vivo and in vitro (42, 43).
Interestingly, the decrease in serum osteocalcin, which is a
marker of function of mature osteoblasts and osteocytes (44),
was prevented equally well by either T or E (ANOVA P
values of 0.013 and 0.002, respectively). This may represent
the in vivo correlate of the recent provocative in vitro demonstration by Kousteni et al. (32) of sex nonspecific, nongenomic actions of T and E on preventing apoptosis. By
contrast, PINP, which is produced not only by mature osteoblasts, but also by cells in the entire osteoblast lineage
(from the immature marrow stromal cell to the osteocyte)
(44) was modulated primarily by E and not T (ANOVA P
values, 0.0001 and 0.452, respectively), suggesting that it is
principally E that regulates osteoblast differentiation in vivo.
This direct interventional study thus provided unequivocal evidence for an important, and likely dominant, role for
E in bone metabolism in normal elderly men. Subsequently,
similar findings were reported by Taxel et al. (45), who demonstrated that treatment of elderly men with an aromatase
inhibitor for 9 wk resulted in significant increases in bone
resorption and decreases in bone formation markers. Combined with the now overwhelming observational data, these
direct interventional studies leave little doubt that E is critical
not only for the growing, but also for the adult and aging
Khosla et al. • Clinical Review
J Clin Endocrinol Metab, April 2002, 87(4):1443–1450 1447
FIG. 4. Urinary NTx levels (A) and annualized rates of
change in mid-radius BMD (B) in a cohort of elderly men
(aged 60 –90 yr) as a function of serum bioavailable E2
levels. F, Subjects with bioavailable E2 levels below 40
pmol/liter (11 pg/ml); E, those with values above 40 pmol/
liter. r and P values indicate separate correlation coefficients for those individuals either below or above bioavailable E2 levels of 40 pmol/liter. [Adapted with permission
from S. Khosla et al.: J Clin Endocrinol Metab 86:3555–
3561, 2001 (23). The Endocrine Society.]
skeleton. The major question left unresolved at this point is
not whether E regulates bone metabolism in the adult male,
but how much E is enough?
Is there a threshold for skeletal E sufficiency in men?
Although this question cannot be answered directly at
present, considerable indirect evidence suggests that the
male skeleton does indeed require a threshold level to suppress bone resorption and bone loss. This evidence comes
again from both observational and interventional studies.
In their cross-sectional analysis of 596 men aged 51– 85 yr,
Szulc et al. (30) found that men in the lowest quartile for
bioavailable E2 levels [⬍53 pmol/liter (14 pg/ml)] had significantly lower BMD at multiple sites compared with men
in the upper three quartiles (Fig. 3). Moreover, it was this
group that also had the most significant increases in bone
turnover markers. Similarly, we have found that bone resorption markers were unrelated to circulating E2 levels in
elderly men with bioavailable E2 levels above 40 pmol/liter
(11 pg/ml), whereas there was a clear inverse association
between these markers and serum bioavailable E2 levels in
the men with bioavailable E2 levels below this value (Fig. 4A)
(23). The value of 40 pmol/liter for bioavailable E2 is also the
median bioavailable E2 level in this population of elderly
men, and more than 90% of postmenopausal women have
bioavailable E2 levels below this value (23). In addition, the
rate of bone loss at the radius in these men was unrelated to
serum bioavialable E2 levels when the latter were above 40
pmol/liter, but clearly related to bioavailable E2 levels when
these levels were below this value (Fig. 4B). Furthermore,
even young men with bioavailable E2 levels below this value
were either losing, or at least not gaining, bone mass (23). On
the basis of these data, we proposed that men with bioavialable E2 levels less than 40 pmol/liter [which correspond in
elderly men to total E2 levels of 114 pmol/liter (31 pg/ml)
and in young men, because of their lower SHBG levels, to
total E2 levels of 73 pmol/liter (20 pg/ml)] were at greatest
risk for increases in bone resorption and bone loss (23).
These observational results have now been independently
confirmed by two interventional studies. In the first study,
Rochira et al. (46) sequentially treated the aromatase-deficient male they had described with three different doses of
the E patch (50, 25, and 12.5 ␮g/d), achieving serum E2 levels
of 356 pmol/liter (97 pg/ml), 88 pmol/liter (24 pg/ml), and
FIG. 5. Changes in lumbar spine BMD as a function of time in an
aromatase-deficient male treated with progressively lower doses of
transdermal E2. [Adapted with permission from V. Rochira et al.:
J Clin Endocrinol Metab 85:1841–1845, 2000 (46). The Endocrine
Society.]
55 pmol/liter (15 pg/ml), respectively. They found that although BMD at the spine continued to increase in the patient
at the intermediate E dose, BMD decreased when the lowest
dose was used (Fig. 5), suggesting that there was a potential
threshold level (at least in young men) for skeletal E sufficiency somewhere between 55 and 88 pmol/liter (15–24 pg/
ml). This is remarkably similar to our independent estimate
from observational data for approximately 73 pmol/liter (20
pg/ml) being necessary for skeletal E sufficiency in young
men (23).
Comparable confirmation of a threshold E level in elderly
men comes from a recently completed study in which 50 men
with a mean age of 69 yr were randomized to treatment with
either placebo or raloxifene, 60 mg/d for 6 months (47).
Overall, raloxifene treatment had no effect on the bone resorption marker, urine NTx, in these men. However, on the
basis of regression analysis, the baseline E2 level was clearly
related to the change in NTx excretion in the raloxifene
group, but not the placebo group (Fig. 6). This analysis indicated that men with serum E2 levels below 96 pmol/liter
(26 pg/ml) responded to raloxifene with a decrease in NTx
excretion, whereas men with serum E2 levels above this
value actually had an increase in NTx excretion after raloxifene therapy. Consistent with this, the men in whom urinary
1448
J Clin Endocrinol Metab, April 2002, 87(4):1443–1450
Khosla et al. • Clinical Review
FIG. 6. Baseline E2 levels vs. the change in urinary NTx
excretion (6 months, baseline) in raloxifene-treated (A) and
placebo-treated (B) men. A, The x-intercept (corresponding
to a ⌬ NTx ⫽ 0) is at approximately 26 pg/ml. [Reproduced
with permission from P. M. Doran et al.: J Bone Miner Res
16:2118 –2125, 2001 (47). American Society for Bone and
Mineral Research.]
NTx excretion decreased after raloxifene therapy had significantly lower baseline serum E2 levels than the men in
whom urinary NTx excretion didn’t change or increased
after raloxifene therapy (22 ⫾ 2 pg/ml vs. 30 ⫾ 3 pg/ml; P ⫽
0.03). By contrast, no such difference was present in the
placebo-treated men (28 ⫾ 3 pg/ml vs. 28 ⫾ 3 pg/ml; P ⫽
0.832, for the corresponding analysis in the placebo group).
One interpretation of these findings is that below 96 pmol/
liter (26 pg/ml), the male skeleton is relatively E deficient,
and raloxifene reduces bone resorption. By contrast, above
this value the skeleton is relatively E sufficient, and raloxifene competes with the more potent endogenous E, E2,
causing a net increase in bone resorption. Indeed, similar to
these findings, the selective ER modulator, tamoxifen, prevents bone loss in E-deficient postmenopausal women (48 –
51), but actually causes bone loss in E-sufficient premenopausal women (49, 50).
Collectively, therefore, the data from the observational
and interventional studies are consistent with the hypothesis
that there is, indeed, a threshold E level for skeletal E sufficiency in men. The best estimate is that this may be at a
bioavailable E2 level of approximately 40 pmol/liter (11 pg/
ml), which would correspond to a total E2 level in elderly
men of 114 pmol/liter (31 pg/ml) and in younger men [because of their lower SHBG levels (6, 23)] to a total E2 level of
73 pmol/liter (20 pg/ml). Clearly, formal dose response
studies are needed to further test this hypothesis. These
findings also indicate that, although local aromatization of
androgens to estrogens may contribute significantly to bone
metabolism (52), a minimum circulating level of E2 (derived
from nonskeletal sources such as the testes and adipose tissue) may still be necessary; below this level, local E2 production simply cannot compensate for the decline in circulating E2 levels. Again, additional work attempting to
quantify the contribution of local E production in bone vs. E
derived from the circulation toward regulating bone metabolism is needed to address this issue.
Finally, Fig. 7 places these findings in perhaps a more
global framework that includes pre- and postmenopausal
women. As evident, virtually all postmenopausal women
have bioavailable E2 levels below 40 pmol/liter and are thus
at risk for bone loss. By contrast, premenopausal women and
young men are generally above this level and are consequently protected against bone loss. A subset of middle-aged
men and an even larger proportion (perhaps as much as 50%)
of elderly men fall below this level, and these may well be the
FIG. 7. Bioavailable E2 levels in three groups of Rochester, Minnesota, men (young, ages 22–39 yr; middle aged, 40 –59 yr; and elderly,
60 –90 yr) and in post- and premenopausal Rochester women. Shown
are the medians and interquartile ranges. The dashed line indicates
a bioavailable E2 of 40 pmol/liter (11 pg/ml). [Reproduced with permission from S. Khosla et al.: J Clin Endocrinol Metab 86:3555–3561,
2001 (23). The Endocrine Society.]
men most likely to develop age-related increases in bone
resorption and bone loss.
Summary and implications for the pathogenesis and
treatment of bone loss in men
The paradigm shift in our thinking that began with the
description of the human experiments of nature has now
brought us to more clearly appreciate the role of E in the male
skeleton. The evidence from multiple lines of investigation is
now overwhelming that E plays a major, and likely dominant, role in regulating bone metabolism in men. T is not
unimportant, however, and contributes in several ways: it
likely has some effect on bone resorption, clearly helps maintain bone formation, and perhaps most importantly, provides the necessary substrate for aromatization to E in the
testes and in peripheral tissues, including locally in bone (52).
Moreover, T also has independent effects on bone size,
largely by enhancing periosteal bone apposition (53).
The evidence that elderly men with low bioavailable E2
levels [below ⬃40 pmol/liter (11 pg/ml)] are the ones that
have the greatest increases in bone resorption markers and in
rates of bone loss also suggests that age-related bone loss
in men may, at least in part, be due to relative E deficiency
in these men. Also, there likely is a sex steroid-independent,
age-related decrease in osteoblast function (54), although
given the evidence that both E and T are important for
Khosla et al. • Clinical Review
maintaining bone formation (31, 32), it is possible that even
this defect is, in part, due to E and T deficiency in elderly men.
Finally, these findings suggest that not all men are likely to
respond to treatment with selective ER modulators or possibly even T replacement, because the effects of the latter on
the skeleton are likely mediated in large part by aromatization to E. Rather, future studies using these agents should
target men with low bioavailable E2 levels, because these are
the individuals who are E deficient and likely to have a
favorable skeletal response to hormonal therapy. As clinicians, we already knew this to be true in the case of pre- vs.
postmenopausal women, who are either E sufficient or deficient. Although this distinction may be more subtle in men,
it does now appear that the same may apply to men with
relative E sufficiency or deficiency.
Acknowledgements
Received December 13, 2001. Accepted January 17, 2002.
Address all correspondence and requests for reprints to: Sundeep
Khosla, M.D., Mayo Clinic, 200 First Street SW, 5-194 Joseph, Rochester,
Minnesota 55905. E-mail: [email protected].
This work was supported by Grants PO1 AG-04875 from the National
Institute on Aging and R01 AR-27065 from the National Institute of
Arthritis, Musculoskeletal and Skin Diseases.
References
1. Riggs BL, Khosla S, Melton III LJ 1998 A unitary model for involutional
osteoporosis: estrogen deficiency causes both Type I and Type II osteoporosis
in postmenopausal women and contributes to bone loss in aging men. J Bone
Miner Res 13:763–773
2. Meier DE, Orwoll ES, Jones JM 1984 Marked disparity between trabecular
and cortical bone loss with age in healthy men. Measurement by vertebral
computed tomography and radial photon absorptiometry. Ann Intern Med
101:605– 612
3. Jones GT, Nguyen V, Sambrook PN, Kelly PJ, Eisman JA 1994 Progressive
loss of bone in the femoral neck in elderly people: longitudinal findings from
the Dubbo Osteoporosis Epidemiology Study. Br Med J 309:691– 695
4. Fatayerji D, Cooper AM, Eastell R 1999 Total and regional bone mineral
density in men: effect of age. Osteoporos Int 10:59 – 65
5. Melton III LJ, Khosla S, Atkinson EJ, O’Connor MK, O’Fallon WM, Riggs
BL 2000 Site-specific bone density over life in men compared to women.
Osteoporos Int 11:592–599
6. Khosla S, Melton III LJ, Atkinson EJ, O’Fallon WM, Klee GG, Riggs BL 1998
Relationship of serum sex steroid levels and bone turnover markers with bone
mineral density in men and women: a key role for bioavailable estrogen. J Clin
Endocrinol Metab 83:2266 –2274
7. Hemsell DL, Grodin JM, Brenner PF, Siiteri PK, MacDonald PC 1974 Plasma
precursors of estrogen. II. Correlation of the extent of conversion of plasma
androstenedione to estrone with age. J Clin Endocrinol Metab 38:476 – 479
8. Smith EP, Boyd J, Frank GR, Takahashi H, Cohen RM, Specker B, Williams
TC, Lubahn DB, Korach KS 1994 Estrogen resistance caused by a mutation
in the estrogen-receptor gene in a man. N Engl J Med 331:1056 –1061
9. Morishima A, Grumbach MM, Simpson ER, Fisher C, Qin K 1995 Aromatase
deficiency in male and female siblings caused by a novel mutation and the
physiological role of estrogens. J Clin Endocrinol Metab 80:3689 –3698
10. Carani C, Qin K, Simoni M, Faustini-Fustini M, Serpente S, Boyd J, Korach
KS, Simpson ER 1997 Effect of testosterone and estradiol in a man with
aromatase deficiency. N Engl J Med 337:91–95
11. Bilezikian JP, Morishima A, Bell J, Grumbach MM 1998 Increased bone mass
as a result of estrogen therapy in a man with aromatase deficiency. N Engl
J Med 339:599 – 603
12. Couse JF, Korach KS 1999 Estrogen receptor null mice: what have we learned
and where will they lead us? Endocr Rev 20:358 – 417
13. Vidal O, Lindberg MK, Hollberg K, Baylink DJ, Andersson G, Lubahn DB,
Mohan S, Gustafsson J, Ohlsson C 2000 Estrogen receptor specificity in the
regulation of skeletal growth and maturation in male mice. Proc Natl Acad Sci
USA 97:5474 –5479
14. Vidal O, Lindberg M, Savendahl L, Lubahn DB, Ritzen EM, Gustafsson JA,
Ohlsson C 1999 Disproportional body growth in female estrogen receptor␣-inactive mice. Biochem Biophys Res Commun 265:569 –571
15. Windahl SH, Hollberg K, Vidal O, Gustafsson J-A, Ohlsson C, Andersson
J Clin Endocrinol Metab, April 2002, 87(4):1443–1450 1449
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
G 2001 Female estrogen receptor ␤⫺/⫺ mice are partially protected against
age-related trabecular bone loss. J Bone Miner Res 16:1388 –1398
Ke HZ, Chidsey-Frink KL, Qi H, Crawford DT, Simmons HA, Brown TA,
Petersen DN, Allen MR, McNeish JD, Thompson DD 2001 The role of
estrogen receptor-␤ (ER-␤) in the early age-related bone gain and later agerelated bone loss. J Bone Miner Res 16(Suppl 1):S160
Gentile MA, Zhang H, Harada S, Rodan GA, Kimmel DB 2001 Bone response
to estrogen replacement in OVX double estrogen receptor-(␣ and ␤) knockout
mice. J Bone Miner Res 16(Suppl 1):S146
Oz OK, Zerwekh JE, Fisher C, Graves K, Nanu L, Millsaps R, Simpson ER
2000 Bone has a sexually dimorphic response to aromatase deficiency. J Bone
Miner Res 15:507–514
Miyaura C, Toda K, Inada M, Ohshiba T, Matsumoto C, Okada T, Ito M,
Shizuta Y, Ito A 2001 Sex- and age-related response to aromatase deficiency
in bone. Biochem Biophys Res Comm 280:1062–1068
Vanderschueren D, Van Herck E, De Coster R, Bouillon R 1996 Aromatization of androgens is important for skeletal maintenance of male rats. Calcif
Tissue Int 59:179 –183
Vanderschueren D, Van Herck E, Nijs J, Ederveen AGH, De Coster R,
Bouillon R 1997 Aromatase inhibition impairs skeletal modeling and decreases bone mineral density in growing male rats. Endocrinology 138:2301–
2307
Vanderschueren D, Boonen S, Ederveen AGH, De Coster R, Van Herck E,
Moermans K, Vandenput L, Verstuyf A, Bouillon R 2000 Skeletal effects of
estrogen deficiency as induced by an aromatase inhibitor in an aged male rat
model. Bone 27:611– 617
Khosla S, Melton III LJ, Atkinson EJ, O’Fallon WM 2001 Relationship of
serum sex steroid levels to longitudinal changes in bone density in young
versus elderly men. J Clin Endocrinol Metab 86:3555–3561
Slemenda CW, Longcope C, Zhou L, Hui SL, Peacock M, Johnston C 1997 Sex
steroids and bone mass in older men: positive associations with serum estrogens and negative associations with androgens. J Clin Invest 100:1755–1759
Greendale GA, Edelstein S, Barrett-Connor E 1997 Endogenous sex steroids
and bone mineral density in older women and men: the Rancho Bernardo
study. J Bone Miner Res 12:1833–1843
Center JR, Nguyen TV, Sambrook PN, Eisman JA 1999 Hormonal and biochemical parameters in the determination of osteoporosis in elderly men. J Clin
Endocrinol Metab 84:3626 –3635
Ongphiphadhanakul B, Rajatanavin R, Chanprasertyothin S, Piaseu N,
Chailurkit L 1998 Serum oestradiol and oestrogen-receptor gene polymorphism are associated with bone mineral density independently of serum testosterone in normal males. Clin Endocrinol (Oxf) 49:803– 809
Van den Beld AW, de Jong FH, Grobbee DE, Pols HAP, Lamberts SWJ 2000
Measures of bioavailable serum testosterone and estradiol and their relationships with muscle strength, bone density, and body composition in elderly
men. J Clin Endocrinol Metab 85:3276 –3282
Amin S, Zhang Y, Sawin CT, Evans SR, Hannan MT, Kiel DP, Wilson PWF,
Felson DT 2000 Association of hypogonadism and estradiol levels with bone
mineral density in elderly men from the Framingham study. Ann Intern Med
133:951–963
Szulc P, Munoz F, Claustrat B, Garnero P, Marchand F 2001 Bioavailable
estradiol may be an important determinant of osteoporosis in men: the MINOS
study. J Clin Endocrinol Metab 86:192–199
Falahati-Nini A, Riggs BL, Atkinson EJ, O’Fallon WM, Eastell R, Khosla S
2000 Relative contributions of testosterone and estrogen in regulating bone
resorption and formation in normal elderly men. J Clin Invest 106:1553–1560
Kousteni S, Bellido T, Plotkin LI, O’Brien CA, Bodenner DL, Han L, Han K,
DiGregorio GB, Katzenellenbogen JA, Katzenellenbogen BS, Roberson PK,
Weinstein RS, Jilka RL, Manolagas SC 2001 Nongenotropic, sex-nonspecific
signaling through the estrogen or androgen receptors: dissociation from transcriptional activity. Cell 104:1–20
Vanderschueren D, Boonen S, Bouillom R 1998 Action of androgens versus
estrogens in male skeletal homeostasis. Bone 23:391–394
Grumbach MM, Auchus RJ 1999 Estrogen: consequences and implications of
human mutations in synthesis and action. J Clin Endocrinol Metab 84:4677–
4694
Kuiper GG, Enmark E, Pelto-Huikko M, Nilsson S, Gustafsson JA 1996
Cloning of a novel estrogen receptor expressed in rat prostate and ovary. Proc
Natl Acad Sci USA 93:5925–5930
Marcus R, Leary D, Schneider DL, Shane E, Favus M, Quigley CA 2000 The
contribution of testosterone to skeletal development and maintenance: lessons
from the androgen insensitivity syndrome. J Clin Endocrinol Metab 85:1032–
1037
Denger S, Reid G, Kos M, Flouriot G, Parsch D, Brand H, Korach KS,
Sonntag-Buck V, Gannon F 2001 ER␣ gene expression in human primary
osteoblasts: evidence for the expression of two receptor proteins. Mol Endocrinol 15:2064 –2077
Vanderschueren D, Van Herck E, Suiker AMH, Visser WJ, Schot LPC, Chung
K, Lucas RS, Einhorn TA, Bouillon R 1993 Bone and mineral metabolism in
the androgen-resistant (testicular feminized) male rat. J Bone Miner Res 8:
801– 809
Meier DE, Orwoll ES, Keenan EJ, Fagerstrom RM 1987 Marked decline in
1450
40.
41.
42.
43.
44.
45.
46.
J Clin Endocrinol Metab, April 2002, 87(4):1443–1450
trabecular bone mineral content in healthy men with age: lack of association
with sex steroid levels. J Am Geriatr Soc 35:189 –197
Kelly PJ, Pocock NA, Sambrook PN, Eisman JA 1990 Dietary calcium, sex
hormones, and bone mineral density in men. Br Med J 300:1361–1364
Leder BZ, Schoenfeld DA, LeBlanc K, Finkelstein JS 2001 Both estradiol and
testosterone play fundamental roles in the maintenance of normal bone turnover in men. J Bone Miner Res 16(Suppl 1):S171
Tomkinson A, Reeve J, Shaw RW, Noble BS 1997 The death of osteocytes via
apoptosis accompanies estrogen withdrawal in human bone. J Clin Endocrinol
Metab 82:3128 –3135
Manolagas SC 2000 Birth and death of bone cells: basic regulatory mechanisms
and implications for the pathogenesis and treatment of osteoporosis. Endocr
Rev 21:115–137
Lian JB, Stein GS, Canalis E, Gehron Robey P, Boskey AL 1999 Bone formation: osteoblast lineage cells, growth factors, matrix proteins, and the mineralization process. In: Favus MF, ed. Primer on the metabolic bone diseases
and disorders of mineral metabolism, 4th Ed. Philadelphia: Lippincott Williams & Wilkins; 14 –38
Taxel P, Kennedy DG, Fall PM, Willard AK, Clive JM, Raisz LG 2001 The
effect of aromatase inhibition on sex steroids, gonadotropins, and markers of
bone turnover in older men. J Clin Endocrinol Metab 86:2869 –2874
Rochira V, Faustini-Fustini M, Balestrieri A, Carani C 2000 Estrogen replacement therapy in a man with congenital aromatase deficiency: effects of
different doses of transdermal estradiol on bone mineral density and hormonal
parameters. J Clin Endocrinol Metab 85:1841–1845
Khosla et al. • Clinical Review
47. Doran PM, Riggs BL, Atkinson EJ, Khosla S 2001 Effects of raloxifene, a
selective estrogen receptor modulator, on bone turnover markers and
serum sex steroid and lipid levels in elderly men. J Bone Miner Res 16:
2118 –2125
48. Love RR, Mazess RB, Barden HS, Epstein S, Newcomb PA, Jordan VC,
Carbone PP, DeMets DL 1992 Effects of tamoxifen on bone mineral density
in postmenopausal women with breast cancer. N Engl J Med 326:852– 856
49. Powles TJ, Hickish T, Kanis JA, Tidy A, Ashley S 1996 Effect of tamoxifen
on bone mineral density measured by dual-energy x-ray absorptiometry in
healthy premenopausal and postmenopausal women. J Clin Oncol 14:78 – 84
50. Burshell AL, Anderson SJ, Leib ES, Johnston CC, Costantino JP 1999 The
effect of tamoxifen on pre and postmenopausal bone. J Bone Miner Res
14(Suppl 1):S158
51. Grey AB, Stapleton JP, Evans MC, Tatnell MA, Ames RW, Reid IR 1995 The
effect of the antiestrogen tamoxifen on bone mineral density in normal late
postmenopausal women. Am J Med 99:636 – 641
52. Simpson E, Rubin G, Clyne C, Robertson K, O’Donnell L, Jones M, Davis
S 2000 The role of local estrogen biosynthesis in males and females. Trends
Endocrinol Metab 11:184 –188
53. Turner RT, Wakley GK, Hannon KS 1990 Differential effects of androgens on
cortical bone histomorphometry in gonadectomized male and female rats.
J Orthop Res 8:612– 617
54. Lips P, Coupron P, Meunier PJ 1978 Mean wall thickness of trabecular bone
packets in the human iliac crest: changes with age. Calcif Tissue Res 26:13–17