Laboratory for Experimental Medicine and Endocrinology (Legendo) Catholic University Leuven 3000 Leuven, Belgium
Address correspondence and requests for reprints to: Roger Bouillon, Laboratory for Experimental Medicine and Endocrinology (Legendo), Katholieke Universiteit Leuven, Herestraat 49, B-3000 Leuven, Belgium. E-mail: Roger.Bouillon{at}med.kuleuven.ac.bc
![]() |
Introduction |
---|
![]() ![]() ![]() |
---|
In this issue of the journal, Hayes et al. (4) applied an intelligent approach, described earlier by the same group, to circumvent this problem. The effects of selective estrogen deficiency (by aromatase inhibition) on GnRH secretion was evaluated in a tandem study of either normal men or men with idiopathic hypogonadotrophic hypogonadism (IHH) treated with fixed GnRH pulses. Their results confirmed the already well established inhibitory effects of estrogens on LH and FSH secretion in men. However, the interest of the present study lies in the comparison of the effect of estrogen deficiency on gonadotropin secretion in normal men with free oscillating GnRH vs. IHH men with a hypothalamic GnRH clamp. Indeed, despite similar sex steroid hormone concentrations, the response of LH both in terms of frequency and amplitude of pulses was significantly greater in normal men compared with IHH men. Therefore, at least part of the difference in LH output is due to differences in GnRH secretion. Hayes et al. (4) further dissected out, again indirectly, whether this increase of GnRH output was due to an increase of GnRH amplitude or differences in pituitary sensitivity.
The acute administration of a submaximal dose of a specific GnRH antagonist resulted in a similar decrease of LH output in normal men whether estrogen replete (without aromatase inhibitor) or estrogen deplete (with aromatase inhibitor) conditions, implying that the hypothalamic GnRH pulses are of equal amplitude in both groups. The increased pulse frequency in estrogen-deficient males is, thus, due to a direct hypothalamic effect on GnRH pulse frequency, whereas the increased amplitude of LH would be due to decreased negative feedback at the pituitary level. This increased sensitivity of the pituitary gland in response to sex steroids was, indeed, already documented by earlier experiments in IHH men during a similar hypothalamic clamp (5, 6). Therefore, the hypothalamus is now another target tissue where the effects of androgens are mediated via the ER after prior aromatization.
The study by Hayes et al. (4) is a clear example of a well-designed clinical study to answer an at first sight unapproachable question in men. The major end point, the frequency of LH pulses as marker for GnRH pulse frequency has been evaluated by an established technique but could have been improved by deconvolution analysis.
Many questions, however, remain unanswered.
First, what are the target cells for estrogen action and their mode of
action on the GnRH network: the GnRH neuron itself (which seems to lack
functional ER during adult life) or the surrounding glia cells or
afferent neurons? (7). Moreover, which ER ( or ß) is
involved and are the effects generated via the classical genomic mode
of action or mediated via rapid nongenomic mechanisms? Indeed, in
animal experiments a rapid (<2 h) inhibitory effect of estrogens on
GnRH release suggested a nongenomic mode of action.
Second, where are the estrogens responsible for the action on the
hypothalamic GnRH secretion produced? Are these estrogens produced
locally by the aromatase activity within the GnRH network or are
systemic estrogens the key messengers? Indeed, the aromatase activity
is found in many tissues and cells in higher primates, whereas in most
other vertebrates aromatase activity is rather limited to the gonades
and the brain. In humans, CYP19 is, however, also expressed in the
placenta, adipose tissue, hair follicle, muscle, bone, and liver
(8). In the brain CYP19 activity is found in various
places, including the anterior and mediobasal hypothalamus. It is
presently unclear whether GnRH neurons or other cells of the GnRH
network express high levels of aromatase activity (7). If
produced locally then the regulation of the CYP19 gene should be taken
into account. Indeed, many cytokines including those found in
life-threatening stress (e.g. tumor necrosis factor and
interleukin-1) or prostaglandins have potent effects on aromatase
expression in other cells and, thus, may potentially mediate stress
effects on the hypothalamic-pituitary gonadotrophic axis
(9).
Third, is there still a place for direct action of androgens (with or
without prior 5-reduction) via the androgen receptor (AR) or
all are actions on the GnRH network mediated via the ER? Because the
GnRH network has a gender-specific pattern, it would at first sign seem
to be directly androgen responsive (7). The
masculinization of the GnRH neurons originates, indeed, from their
exposure to an androgen pulse during the early neonatal period and
eliminates forever the estrogen-mediated positive feedback on the GnRH
network, which is so typical for the female ovulatory LH surge.
However, exposure to an estrogen pulse during the first 5 days of life
also masculinizes the GnRH neurons, suggesting that aromatization of
testosterone and the ER are the pathways involved. In contrast,
nonaromatizable androgens are unable to inhibit GnRH secretion in
animals and men (10). Other arguments seem to indicate
that the AR activation nevertheless has a direct role on GnRH/LH/FSH
secretion. Indeed, gonadotropin secretion in both animals and men with
an inactivating AR mutation is frankly increased despite high
circulating estrogen levels pointing toward a separate direct negative
effect of androgens (10, 11, 12). These effects, however,
could be due to direct androgen action at the pituitary level (see
following paragraph).
Fourth, is the inhibitory action and mode of action of estrogens at the
pituitary level different from their effects on the hypothalamus?
Again, the same questions can be raised regarding the origin (local or
systemic) of estrogens, their use of the ER and ß, and the choice
between genomic/nongenomic activation pathways. Moreover, the negative
effects could be located at different levels because estrogens can
decrease the number of GnRH receptors on gonadotrophs
(13). The GnRH peptidase activity also fluctuates.
Therefore, genomic effects at the transcriptional level are the most
likely mechanism. In contrast to the GnRH network, several arguments
point toward direct effects of testosterone on LH secretion via the AR.
Indeed, prior conversion into dihydrotestosterone by the abundant
pituitary 5
-reduction was found to be essential for testosterone
action on GnRH-induced calcium signaling in the rat gonadotroph
(14).
It is now increasingly clear that androgens and estrogens are not
always hormones with counteracting effects. In fact, in many situations
both hormones use the same final pathway, after local transformation
and activating the same receptor and postreceptor mechanisms. The
hypothalamic regulation of GnRH secretion is now a new example. This
was already previously well documented for the gonadotroph in the
pituitary (5, 6), the growth plate chondrocyte, bone
cells, and testicular efferent ductuli (2, 3, 15, 16, 17).
Indeed, the pubertal growth plate does not close in men with either an
ER or aromatase mutation. Therefore, such men show continuous linear
growth and develop eunuchoid characteristics. Administration of
estrogens to aromatase-deficient men, however, result in mineralization
and closure of the growth plate (2). Moreover, bone cells
and especially osteoblast-mediated osteoclastogenesis are sensitive to
both estrogens and androgens, and aromatase inhibition causes similar
degree of osteoporosis in male animals as orchidectomy
(18). The role of estrogens for male reproduction below
the level of GnRH and LH/FSH regulation is not yet solved. Male
-ERKO mice show an apparently normal development of the reproductive
tract but are nevertheless infertile, with decreased sperm quality
resulting in the inability to fertilize wild-type oocytes (3, 17). This failure of spermatogenesis in
-ERKO mice is caused
by severely impaired fluid absorption, luminal distension and atrophy
of the seminiferous tubuli. No such abnormalities are found in ß-ERKO
mice. Aromatase mutations in male rats and men also have variable
effects on fertility. However, failure of spermatogenesis in aromatase
knockout mice is not caused by a primary defect in seminiferous
tubuli fluid reabsorption (such as described in
-ERKO mice), but
seems secondary to direct failure of germ cell differentiation
(15).
The role of prior aromatization of androgens on the male brain (beside
GnRH secretion) and especially on male behavior is also not settled
(19). Indeed, in ER and aromatase mutant men, sexual
behavior seemed to be grossly normal. -ERKO male mice nevertheless
show decreased mounting attempts with an especially low rate of
intromissions and ejaculations.
-ERKO male mice also have a marked
deficit in all male aggressive behavioral indices, which may,
therefore, be estrogen dependent (3). AR-deficient XY mice
and men (Tfm), in contrast, do not display male-like gender behavior.
Therefore, most observations point toward a dominant role for the
androgen rather than ER in male gender role behavior.
Of course, not all androgen actions are mediated via prior
aromatization and the ER (see Fig. 1).
Sometimes, the AR and ER also regulate genes independently in the same
(e.g., osteoblast-derived cytokines) or opposite direction
(e.g. lipoproteins). Moreover, a number of crucial
gender-specific characteristics, such as growth, differentiation of
prostate or hair follicles, and muscle development, directly depend on
the activation of the AR, with or without prior 5
-reduction of
testosterone.
|
![]() |
Acknowledgments |
---|
![]() |
Footnotes |
---|
Received July 17, 2000.
Accepted July 17, 2000.
![]() |
References |
---|
![]() ![]() ![]() |
---|