| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ARTICLES |
Medical Research Council Reproductive Biology Unit, Center for Reproductive Biology (N.A., C.M., M.W., K.J.T., J.S.F., M.M., M.R.M., R.M.S.), Edinburgh, Scotland EH3 9ET; the Institute of Experimental Morphology and Anthropology, Bulgarian Academy of Science (N.A.), 1113 Sofia, Bulgaria; the School of Biological and Molecular Sciences, Oxford Brookes University (N.P.G.), Headington, Oxford, United Kingdom OX3 0BP
Address all correspondence and requests for reprints to: Dr. R. M. Sharpe, Medical Research Council Reproductive Biology Unit, Center for Reproductive Biology, 37 Chalmers Street, Edinburgh, Scotland EH3 9ET. E-mail: r.sharpe{at}ed-rbu.mrc.ac.uk
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
GnRHa-treated rats exhibited a significant increase in germ cell volume per Sertoli cell and a reduction in germ cell apoptosis, probably because of the raised FSH levels. Despite similar raised FSH levels, rats treated with DES (10 or 1 µg) or EE (10 µg) had reduced germ cell volume/Sertoli cell and increased germ cell apoptosis, especially when compared with GnRHa-treated animals. The latter changes were associated with an increase in lumen size per testis, indicative of impaired fluid resorption from the efferent ducts, resulting in fluid accumulation in the testis. Rats treated neonatally with 0.1 µg DES showed reduced germ cell apoptosis comparable to that in GnRHa-treated animals. The changes in apoptotic rate among treatment groups occurred across all stages of the spermatogenic cycle. It is concluded that 1) neonatal estrogen treatment results in dose-dependent alterations in Sertoli cell numbers, germ cell volume, efficiency of spermatogenesis, and germ cell apoptosis in adulthood; 2) the relatively poor spermatogenesis in estrogen-treated animals is most likely due to altered testis fluid dynamics and/or altered Sertoli cell function; 3) as indicated by FSH (LH) and testosterone levels, the hypothalamic-pituitary axis and Leydig cells are probably more sensitive than the Sertoli cells to reprogramming by estrogens neonatally; and 4) elevated FSH levels in adulthood may improve the efficiency of spermatogenesis.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
(6, 7), although
recent evidence suggests that male (BERKO) mice with targeted
inactivation of estrogen receptor-ß do not show fertility problems
(8). Conversely, it is clear that exposure of the developing male to
exogenous estrogens either in utero or neonatally can result
in a range of abnormalities of reproductive development and function
(9, 10, 11, 12, 13). The latter include effects on the developing Sertoli cells
(12), which express estrogen receptor-ß (3, 4), and suggest that
inappropriately high estrogen exposure during neonatal life can reduce
final Sertoli cell number and impair Sertoli cell functional maturation
(12). In the past, impairment of adult spermatogenesis in neonatally
estrogenized rats has been attributed to reduced gonadotropin secretion
(primarily reduced FSH levels) induced during the neonatal treatment
(14, 15), a conclusion that is difficult to reconcile with recent data
(12). A further issue is that the majority of early studies involving
neonatal estrogen administration (9) have used extremely high doses of
potent estrogens (e.g. 200–500 µg injected on day 1) and
even our own recent studies have used total doses of 60 µg
diethylstilbestrol (DES) spread over a 12-day period (12). Results from
such studies are therefore difficult to relate straightforwardly to the
prevailing concern about whether weakly estrogenic environmental
estrogens can induce impairment of spermatogenesis and whether human
exposure to such compounds could have contributed to the reported fall
in sperm counts or to other disorders of male reproductive health (16, 17). This has become a major issue but progress toward obtaining a
definitive answer has been slow. One of the obstacles to addressing the concerns expressed above is our lack of understanding of the mechanistic basis via which neonatal estrogen treatment is able to impair adult testicular function. A number of fundamental questions remain unanswered. For example, are the estrogen-induced changes in spermatogenesis in adulthood simply a consequence of reduced Sertoli cell number (12) or do they result from the reported abnormal changes to the structure and function of the efferent ducts (11, 13, 18) as is the case in ERKO mice? (19). Alternatively, is Sertoli cell function impaired permanently in some fundamental way (12), or are there changes in Leydig cell function and testosterone production (9) that result in suboptimal hormonal support for spermatogenesis? Finally, in the context of evaluating human risk from exposure to environmental estrogens, an important question is what level of estrogen exposure will induce adverse changes to spermatogenesis and what end points most clearly reflect inappropriate estrogen exposure? The present studies were undertaken to answer these questions and, in so doing, to establish which aspects of the hypothalamic-pituitary-testis axis were most sensitive to lifelong disruption by neonatal estrogen exposure.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Wistar rats bred in our own animal house were maintained under standard conditions. Beginning on postnatal day 2 (day of birth = day 1), rats were subjected to one of the following treatments: 1) sc injection of either DES (Sigma Chemical Co., Poole, UK) at a dose of 10, 1, or 0.1 µg in 20 µl corn oil or ethinyl estradiol (EE) at a dose of 10 µg in 20 µl corn oil on days 2, 4, 6, 8, 10, and 12; 2) sc injection of 10 mg/kg of a long acting GnRH antagonist (Antarelix, Europeptides, Argenteuil, France) in 20 µl 5% mannitol on days 2 and 5 [this treatment regimen has been shown previously (12) to suppress FSH secretion to undetectable levels up to days 18–25]; and 3) injection with the appropriate vehicle (controls). As oil-injected and 5% mannitol-injected control rats did not differ significantly for any of the parameters measured in initial studies, data for these animals was pooled and in all subsequent experiments all controls were injected with corn oil as above. Final group sizes of 7–22 rats from the different treatment groups, representing animals from 2 separate experiments, were subsequently killed on days 75–90 to study testis weight and blood hormone levels and, in a representative proportion of animals, gross morphology of the testis and cell volume quantification. Additional groups of 5–9 animals in all but the EE-treated group were evaluated at 18 days of age for determination of Sertoli cell numbers as described below. For all animals, body and testis weights were recorded at death.
Collection of samples and processing of tissues
On the day of sampling, animals were anesthetized with halothane
or flurothane, and blood was collected from the heart into a
heparinized syringe; animals were then killed by cervical dislocation.
Plasma samples were stored at -20 C until used for hormone analysis.
In animals sampled at 18 days of age for determination of Sertoli cell
number, the testes were removed and immersion fixed in Bouin’s fluid
for 24 h at room temperature before being processed as described
below. Representative adult animals of those sampled at 75–90 days of
age were perfusion fixed with Bouin’s fixative as described previously
(20). Before processing, Bouin’s-fixed testicular tissue was cut
transversely into two to six slices with a razor blade. Fixed tissue
was then processed for 17.5 h in an automated Shandon processor
and embedded in paraffin wax. Sections of 5-µm thickness were cut and
floated onto slides coated with 2% 3-aminopropyltriethoxy-silane
(Sigma Chemical Co., St. Louis, MO) and dried at 50 C
overnight before being used for cell quantification studies and
visualization of apoptotic germ cells as described below.
Measurement of Sertoli cell numbers
The time chosen for determination of Sertoli cell numbers was 18
days of age, because final Sertoli cell number is established before
this age (21, 22), and preliminary experiments using labeling with
bromodeoxyuridine (12) had confirmed that Sertoli cell division was
negligible by day 18 in control and DES-treated rats (unpublished
data). Testis tissue from 18-day-old rats was immersion fixed in
Bouin’s fixative for 24 h. Testes were then sampled in a random
systematic manner by slicing them transversely into four pieces using a
razor blade and then processing either slices 1 and 3 or slices 2 and 4
through graded ethanols before infiltrating with JB4 resin (TAAB,
Berkshire, UK). After polymerization, 20-µm sections were cut on a
Reichart 2050 microtome using a Diatome Histoknife, mounted onto glass
slides, and stained with Harris’ hematoxylin. Sertoli cells were then
counted using the optical disector method as described by Wreford
(23).
Measurement of Leydig [3ß-hydroxysteroid dehydrogenase (3ßHSD)
positive] cell volume per testis
To evaluate whether neonatal treatment with EE (10 µg) or GnRH
antagonist (GnRHa) altered Leydig cell numbers in adulthood, sections
from Bouin’s perfusion-fixed testes were immunostained for 3ßHSD, as
described previously (24), and the volume of 3ßHSD-positive cells per
testis was determined using point-counting methods similar to those
outlined below for enumeration of apoptotic cells. The differences were
that one section from each of three blocks for each of four animals per
group were examined, points falling over 3ßHSD-positive cytoplasm or
over the nuclei of cells with 3ßHSD-positive cytoplasm were scored
separately, and both were then independently expressed as relative
volumes per testis. These data were converted to absolute volumes per
testis by multiplication by testis weight (=volume), as shrinkage was
minimal (see below).
Quantification of apoptotic and nonapoptotic (=viable) germ cells,
and determination of germ cell volume per Sertoli cell and seminiferous
tubule lumen volume per testis
The purpose of these investigations was to determine the
efficiency of spermatogenesis by 1) enumerating total germ cell volume
supported by each Sertoli cell, and 2) enumerating apoptotic germ cells
as a proportion of total germ cell mass. DNA fragmentation was detected
by in situ DNA 3'-end labeling in histological sections
using a nonradioactive labeling method that resulted in a high
degree of specificity and low background staining, as described and
validated in detail previously (12).
Apoptotic cells in cross-sections of testes from three to six rats in each treatment group were evaluated and examined under oil immersion using a Leitz (Rockleigh, NJ) x63 plan apo objective fitted to a Leitz Laborlux microscope with a 121-point eyepiece graticule. Using a systematic clockface sampling pattern from a random starting point, 16 fields were counted. Points falling over Sertoli cell or germ cell nuclei (apoptotic or nonapoptotic) or over seminiferous tubule lumens were scored and expressed as a percentage of the 121 points possible; nonapoptotic spermatogonia, spermatocytes, and round and elongate spermatids were distinguished, but in the present studies, their total volumes were combined for individual animals before analysis. For each animal, the values for percent volume were converted to absolute volumes per testis by reference to testis volume (=weight), as shrinkage was minimal, i.e. testis weights before and after fixation were comparable in each treatment group. For each animal, the total germ cell volume per testis (apoptotic plus nonapoptotic) was expressed relative to Sertoli cell volume per testis in the same animal; correction for Sertoli cell volume was applied, as Sertoli cell numbers were reduced in treated rats compared with controls (see Results). The data for round spermatid volume per testis based on nuclear point counting can be equated to cell number per testis, as preliminary studies showed no major differences in round spermatid nuclear diameters between values for control, DES (10 µg)-treated, and GnRHa-treated animals (our unpublished data).
To provide some indication of the qualitative support of germ cells by Sertoli cells, the proportion of apoptotic/total germ cells was calculated for each animal. Finally, to evaluate the stage dependence of germ cell apoptosis, three complete testis transverse sections per animal for each of three animals per treatment group were scanned systematically, and seminiferous tubule cross-sections at stages I–V, VII–VIII, and IX–XIV of the spermatogenic cycle were scored for the number of apoptotic germ cells per cross-section. A total of 46–400 round cross-sections (in most instances at least 120) of tubules at each stage grouping were evaluated per animal, and a mean value for the number of apoptotic cells/cross-section in the 3 stage groupings was derived for each animal.
Measurement of plasma hormone levels
Plasma levels of FSH and LH were measured by RIA using materials
supplied by the NIDDK. Results have been expressed in terms of rat FSH
RP-2 and rat LH RP-3 standards, respectively. In the FSH assay, plasma
levels measured in hypohysectomized rats range from 1.2–2.3 ng/ml, and
values in this range are therefore considered to be baseline. Plasma
levels of inhibin B were measured using a two-site enzyme-linked
immunoassay that uses a capture antibody directed against the
C-terminal portion of the human ßB-subunit and the F(ab)
fraction of a mouse monoclonal antibody (R1) to the N-terminal portion
of the inhibin -subunit conjugated to alkaline phosphatase (25, 26, 27).
The assay has been previously validated for measurement of inhibin B in
the rat (27, 28), and in the present studies it was confirmed that rat
plasma diluted in parallel with the inhibin B standard and that levels
of inhibin B in castrate adult male rat plasma were reduced to
undetectable levels (<60 pg/ml).
Plasma levels of testosterone were measured using an enzyme-linked immunosorbent assay adapted from an earlier RIA method (29). Plasma, to which was added trace amounts of [3H]testosterone (Amersham International, Aylesbury, UK), was extracted twice with 10 vol hexane-ether (4:1, vol/vol), and the organic phase was dried down under N2 at 55 C. The efficiency of extraction averaged 75%. The second antibody was immobilized to an enzyme-linked immunosorbent assay plate by addition of 100 µl acid purified donkey antigoat/sheep IgG (250–350 mg/ml) diluted in 0.1 M sodium carbonate buffer, pH 9.6. The plate was sealed and incubated overnight at 4 C. The wells were then washed twice with 0.1% Tween-20 and incubated for 10 min at room temperature with 0.2 ml of the same solution to block nonspecific binding sites. Samples in duplicate (50 µl) were assayed after dilution in 0.1 M PBS, pH 7.4, containing 0.1% gelatin (Sigma Chemical Co.) and incubated overnight at 4 C with 50 µl sheep antitestosterone-3-carboxymethyloxide-BSA diluted 1:100,000 plus 50 µl testosterone-3-carboxymethyloxide labeled with 1:20,000 diluted horseradish peroxidase (Amdex, Amersham Pharmacia Biotech, Uppsala, Sweden). The plate was then washed several times with 0.1% Tween-20 before addition of 0.2 ml substrate (5 mM O-phenylenediamine; Sigma Chemical Co.) and 0.03% hydrogen peroxide diluted in 0.1 M citrate-phosphate, pH 5.0, to each well. The plate was then incubated in the dark for 10–30 min until the color reaction was optimal. The reaction was stopped by the addition of 50 µl 2 M sulfuric acid to each well, and the OD was read at 492 nm in a plate reader. The limit of detection was 12 pg/ml. All samples were assayed together in one run.
Statistics
Comparison of testis weights, hormone and Sertoli cell number,
and germ cell and Leydig cell volume data for the various treatment
groups was made using ANOVA, as all data were normally distributed,
whereas data for stage-specific counts of apoptotic germ cells were
transformed logarithmically before analysis. Where significant
differences between groups were indicated, subgroup comparisons also
used ANOVA, but used the variance from the experiment as a whole as the
measure of error.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
). Neonatal suppression of gonadotropin
levels via administration of GnRHa reduced testis weight to a similar
extent as the 10-µg doses of DES and EE, but caused a relatively
smaller reduction in germ cell volume per testis than did these
estrogens (Fig. 1). Gross morphology of the testes was comparable to
that in controls in animals treated with GnRHa or 0.1 µg DES (Fig. 2). In animals treated with 10 µg of
either DES or EE, there was distension of the rete testis and
associated dilatation of seminiferous tubule lumens, which was most
severe adjacent to the rete testis (Fig. 2). Tubules in the latter area
exhibited variable disruption of spermatogenesis, including Sertoli
cell-only tubules, and elsewhere in the testis, tubules were reduced in
diameter and exhibited increased numbers of degenerating/apoptotic germ
cells (see below), although tubules with grossly normal spermatogenesis
were frequent (Fig. 2). Animals treated with 1 µg DES exhibited
largely normal spermatogenesis and tubule diameter, with only
occasional tubules in which variable loss of germ cells/increased
apoptosis was evident (Fig. 2).
|
|
) were matched by changes in both Sertoli
cell number (Fig. 3) and Sertoli cell
volume per testis (Fig. 4), measured by
the two methods. Thus, treatment with DES caused a dose-dependent,
significant decrease in final Sertoli cell number per testis, whereas
treatment with GnRHa reduced Sertoli cell numbers by nearly
50%. Similar changes were observed in these treatment groups for
Sertoli cell volume per testis, determined by point counting in adult
testes (Fig. 4). Only Sertoli cell volume was determined for animals
treated neonatally with EE, and this revealed a significant decrease
(Fig. 4).
|
|
). It was anticipated that inhibin B levels would
parallel the change in Sertoli cell number/volume, with FSH levels
showing a mirror image of inhibin B levels. Surprisingly, FSH levels
were elevated significantly, and to a similar level in all animals
treated with GnRHa, DES, or EE, independent of the dose administered or
the degree of reduction in Sertoli cell number (Fig. 4); the data
illustrated was derived from animals from two separate experiments, and
comparable results were obtained in both experiments. There was a
similar, although not identical, disparity between plasma levels of FSH
and inhibin B (Fig. 4). Thus, animals treated with EE or 10 or 1 µg
DES showed a decrease in inhibin B levels that more or less paralleled
the decrease in Sertoli cell volume, whereas animals treated with GnRHa
showed a disproportionately small decrease, and rats treated with the
lowest dose (0.1 µg) of DES showed no decrease at all in their
inhibin B levels despite their very clear elevation of FSH levels (Fig. 4).
Changes in plasma levels of LH and testosterone
Plasma levels of testosterone showed a major dose-dependent
decrease in animals treated neonatally with estrogens (Fig. 5). Thus, average testosterone levels
were reduced by 83%, 71%, and 38% respectively, in animals treated
with 10, 1, or 0.1 µg DES, whereas EE-treated rats showed a 70%
reduction. In contrast, testosterone levels in GnRHa-treated rats were
comparable to control levels (Fig. 5). Plasma levels of LH did not
differ significantly from control values in any of the treatment groups
(Fig. 5).
|
).
Compared with GnRHa-treated animals, EE-treated rats showed a
significant reduction in Leydig cell volume per testis (Table 1).
|
). This apparent difference in
effect of EE and DES at the same dose was not significant
(P > 0.2) and was probably attributable to one
unusually high value in one animal in the DES (10 µg)-treated group.
Subcategorization of germ cells indicated that the volume of pachytene
spermatocytes or round spermatids per Sertoli cell showed changes that
paralleled those shown for total germ cell volume in the various
treatment groups (data not shown).
|
), which
is most likely explained by the elevation of FSH levels (see
Discussion). As DES- and EE-treated rats all had plasma FSH
levels as high as those in the GnRHa-treated group (Fig. 4), germ cell
volume/Sertoli cell in these groups was compared with that in the
GnRHa-treated group. This indicated a significant reduction in germ
cell volume/Sertoli cell in EE-treated and DES (10 or 1 µg)-treated
animals compared with those in the GnRHa-treated group (Fig. 6). The
validity of this comparison was reinforced by data for the proportion
of germ cells that were apoptotic (Fig. 6, middle panel).
This showed that GnRHa-treated animals exhibited a significant
reduction in the relative proportion of apoptotic germ cells compared
with controls, whereas DES-treated rats showed a stepwise pattern, with
grossly elevated rates of apoptosis in animals treated neonatally with
10 µg DES, similar rates to controls in the 1 µg DES group and a
significantly reduced rate of apoptosis in the 0.1 µg DES group.
EE-treated rats showed an approximate doubling in the relative
proportion of apoptotic germ cells, although this was only
significantly different from the GnRHa-treated group.
Two possible causes for the increased apoptotic rate in higher dose
estrogen-treated rats were evaluated: 1) the possibility that it
resulted from inadequate testosterone support (30, 31), and 2) the
possibility that it resulted from a build-up of seminiferous tubule
fluid (STF) in tubule lumens (13, 19). Enumeration of apoptotic germ
cells in relation to stage of the spermatogenic cycle in each of the
treatment groups indicated that where increases or decreases in
apoptosis occurred they did so in all stage groupings, with no evidence
for preferential changes at the androgen-dependent stages VII–VIII
(data not shown). In contrast, build-up of STF in tubule lumens was
evident in DES (and EE)-treated animals, with a dose-dependent increase
in the percentage of testis volume occupied by lumen, a change that
paralleled the increase in the proportion of apoptotic germ cells (Fig. 6). In contrast, GnRHa-treated rats and animals treated with the lowest
dose of DES did not differ from controls in this parameter (Fig. 6).
Relative sensitivity to neonatal estrogen exposure
The relative sensitivities of the various end points measured in
these studies to alteration by neonatal DES exposure are listed in
Table 2. From this it is evident that
relatively larger changes (compared with the control mean) are induced
in FSH and testosterone levels by all doses of DES, although testis
weight is also a sensitive indicator of inappropriate estrogen exposure
neonatally. Although the change in FSH levels is noticeable for its
apparent reverse dose response, it is emphasized that there was no
significant (P > 0.05) difference in FSH levels among
the different DES dose groups.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
There is a substantial literature showing adverse effects of fetal/neonatal estrogen exposure on spermatogenesis and sperm output in adulthood in rodents (9, 10, 11, 12, 13, 14, 15) and limited evidence to suggest that the same may be true for man (16, 17, 35). Although earlier studies in rodents largely concluded that these effects resulted from suppression of FSH secretion during treatment (14, 15), comparison with animals in which suppression of FSH levels to the limit of detection is induced by neonatal administration of a GnRHa does not support this conclusion (12). Furthermore, detailed comparison of FSH levels during and after neonatal treatment with DES shows that only the highest dose (10 µg) suppresses FSH levels consistently, although still not as effectively as with GnRHa treatment (12) (our unpublished data). The considerably different phenotypes exhibited by DES (10 µg)- and GnRHa-treated animals makes it unlikely that gonadotropin suppression can account for most of the changes observed in the former group.
The most important factor that determines the ceiling of sperm production and output in all mammals is the number of Sertoli cells per testis (30). The present findings confirm and extend our previous observations by showing that neonatal treatment of rats with potent estrogens such as DES or EE results in a permanent and dose-dependent reduction in Sertoli cell number, either measured directly at 18 days of age or inferred by reduction in volume per testis in adulthood. Based on our previous evidence (12), this effect may not be explicable by a reduction in FSH levels during the treatment period (when Sertoli cells are replicating), but may also reflect a direct effect of estrogen on the rate of division of the developing Sertoli cells. As a dose of DES as low as 0.1 µg (administered six times over 12 days) reduced Sertoli cell number, it will be important in future studies to explore whether even lower doses can still induce such an effect. It will also be important to test whether 17ß-estradiol can alter Sertoli cell numbers using our treatment regimen, as a recent study in which rats were fed 17ß-estradiol in their diets at 2.5 parts/million failed to demonstrate any decrease in Sertoli cell numbers (36). From the present data, it is unquestionable that the numerically most important factor that limits sperm output in adult rats exposed neonatally to estrogens is the decrease in Sertoli cell number, but it is also evident that other factors, both positive and negative, may affect final sperm output in these animals.
The hormones FSH and testosterone play important roles in spermatogenesis, although in nonseasonal animals testosterone appears to be the most important regulatory factor (30). It is well established in rats that testosterone exerts its main effects on spermatogenesis at stages VII–VIII of the spermatogenic cycle (30, 37, 38), and inadequate levels of testosterone within the testis result in the appearance of apoptotic germ cells at these stages, whereas in controls apoptosis at these stages is very rare (31, 39, 40). In rats treated neonatally with DES/EE or GnRHa, in which increases or decreases in the apoptotic rate of germ cells occurred relative to control values, they did so at all stages and without a preferential change at stages VII–VIII, which suggests that the major reduction (70–80%) in plasma levels of testosterone in DES/EE-treated groups may not be matched by a deficiency in intratesticular testosterone. The fact that the increase in germ cell apoptosis in DES/EE-treated groups occurred at all stages of the spermatogenic cycle suggests that it may have a more general cause, as is discussed below.
Rats treated neonatally with GnRHa had reduced Sertoli cell numbers and testis size, but exhibited significantly improved efficiency of spermatogenesis, as indicated by increased germ cell volume supported per Sertoli cell and decreased germ cell apoptosis compared with control animals. It is our contention that the increase in efficiency of spermatogenesis in GnRHa-treated rats results from the supranormal FSH levels. This is based on two arguments. First, in rats in which Sertoli cell number is increased experimentally by induction of neonatal hypothyroidism, the resulting elevation of inhibin B secretion leads to a permanent lowering of FSH levels (28, 41), which is associated with a substantial reduction in the number of round spermatids supported by each Sertoli cell (42). This is exactly the opposite change from that shown for GnRHa-treated rats in the present studies. Second, there is considerable (mainly indirect) evidence in the literature suggesting that in adult animals, FSH may increase the number of germ cells entering meiosis and thereafter progressing through spermatogenesis (reviewed in Refs. 30, 32). Assuming that an increase in FSH levels is responsible for the increase in efficiency of spermatogenesis in GnRHa-treated rats, it would be predicted that rats treated neonatally with estrogens would also exhibit increased efficiency of spermatogenesis as they exhibited an elevation of FSH levels that matched that in the GnRHa-treated rats. However, except in animals treated with the lowest dose (0.1 µg) of DES, the present evidence points toward reduced, rather than increased, efficiency of spermatogenesis in DES/EE-exposed rats, especially when they were compared with GnRHa-treated animals.
The reduced efficiency of spermatogeneis in DES/EE-treated rats stemmed in part from increased germ cell apoptosis, which was not stage specific and seemed to go hand in hand with increased seminiferous tubule lumen size. The latter change is thought to result from maldevelopment of the rete testis and efferent ducts, previously shown in neonatally estrogen-treated rodents (11, 13, 18, 43), with consequent impaired resorption of STF by the efferent duct epithelium (13, 18). Similar, although more extreme, changes have been shown in ERKO mice (6, 19), and it is thought that a similar underlying mechanism may account for the change both in ERKO males and, at the other extreme, in animals overexposed to estrogens neonatally (1). It may therefore be that the positive stimulus to the Sertoli cell via increased FSH levels in DES/EE-treated rats is counteracted by the negative influence on spermatogenesis of STF accumulation in the tubule lumens. The finding that animals treated neonatally with only 0.1 µg DES exhibited no lumen distension and also had significantly reduced germ cell apoptosis (comparable to that in GnRHa-treated rats) supports this interpretation. However, as low rates of germ cell apoptosis occurred in treatment groups with elevated FSH levels and normal or only moderately reduced testosterone levels, it is also possible that interactive effects between these two hormones may positively support spermatogenesis.
Aside from effects on spermatogenesis, the present studies also provide clear evidence for permanent reprogramming of the hypothalamic-pituitary-testis axis at various levels by neonatal estrogen exposure. The most surprising of these changes was the complete discordance between the changes in blood levels of FSH and inhibin B. In other studies (28), we have shown that final Sertoli cell number per testis determines the blood levels of inhibin B, and this, in turn, determines the blood levels of FSH, i.e. the lower the levels of inhibin B, the higher the levels of FSH and vice versa, a finding matched by several studies in adult men (44, 45). The present findings indicate that this relationship does not apply straightforwardly to adult rats in which Sertoli cell numbers have been altered by neonatal estrogen administration, as these showed a relatively uniform elevation of FSH levels despite considerable differences in final Sertoli cell number/volume per testis. Moreover, although inhibin B levels grossly reflected Sertoli cell number in the different treatment groups, animals treated with higher estrogen doses exhibited a disproportionately large reduction in inhibin B levels. This change, which is also evident earlier in development (our unpublished data), could indicate permanent impairment of Sertoli cell function, consistent with previous suggestive evidence (12). Regardless of changes in Sertoli cell function, the finding that plasma FSH levels were elevated independently of inhibin B or testosterone (see below) levels, argues strongly for permanent alterations in the wiring of the hypothalamus/pituitary axis induced by neonatal estrogen exposure (9, 46). The primary defect in FSH regulation appears to be insensitivity to normal feedback signals from the gonads, and the finding that LH levels were normal in estrogen-treated rats despite gross lowering of blood testosterone levels suggests that inappropriate feedback sensitivity may also exist with this axis. However, dynamic stimulation tests of the hypothalamic-pituitary axis will be necessary to establish whether this interpretation is correct.
The present finding that neonatal estrogen exposure dramatically and dose dependently lowers testosterone levels in adulthood was unexpected. This change was not a consequence of suppressed gonadotropin secretion neonatally due to the estrogen treatment, as rats treated neonatally with GnRHa had unaltered testosterone levels and, for unknown reasons, elevated Leydig cell volume per testis in adulthood. The reduction in testosterone levels in neonatally estrogen-treated rats did not stem from a deficiency in LH in adulthood, and preliminary analysis indicated that it did not result from a deficiency in Leydig cell volume per testis. The most likely explanation remaining is reduced responsiveness of the Leydig cells, but this remains to be defined. Nevertheless, it is clear that estrogen exposure neonatally is somehow able to permanently reprogram adult Leydig cell function, and identification of the pathways involved could provide important insight into the fundamental regulation of Leydig cells.
Finally, the present findings have potential implications regarding the issue of environmental estrogens and the application of methods for their detection (33, 34). Based on the DES dose-response relationship for the various parameters measured in the present studies, it is clear that the Leydig cells (as indicated by testosterone levels) and the gonadotropes (as indicated by FSH levels) appear to be more sensitive to lower doses of DES developmentally than are the various testicular end points, except for testis weight. Although such changes are likely to reflect a complex set of circumstances (i.e. altered feedback signals, altered sensitivity to these signals, etc.), they have potential utility for the detection of subtle effects of environmental estrogens in vivo. Studies to evaluate such possibilities are in progress.
In summary, the present findings add to the evidence that neonatal exposure of the developing Sertoli cells to estrogens is able to exert long lasting effects on the numbers and function of these cells and on the efficiency of spermatogenesis, changes that may involve both direct and indirect effects on spermatogenesis. The present data also provide interesting new evidence to suggest that hypothalamic/pituitary programming of FSH (and LH) secretion neonatally appears to be extremely sensitive to estrogens and that Leydig cell function in adulthood may also be susceptible to reprogramming by neonatal estrogen exposure. Finally, the finding that elevation of FSH levels in adulthood is associated with significant improvement in the efficiency of spermatogenesis, at least in neonatally GnRHa-treated rats, could provide new insights into the regulation of spermatogenesis.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Received March 11, 1999.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
(ER) within the testis and excurrent ducts of the rat and marmoset
monkey from perinatal life to adulthood. J Endocrinol 153:485–495[Abstract]
-hydroxylase/C17–20
lyase (P450c17) in the fetal rat testis is reduced by maternal exposure
to exogenous estrogens. Endocrinology 137:1063–1070[Abstract]
This article has been cited by other articles:
 |
|
 |
|
  H. M. Scott, G. R. Hutchison, M. S. Jobling, C. McKinnell, A. J. Drake, and R. M. Sharpe Relationship between Androgen Action in the "Male Programming Window," Fetal Sertoli Cell Number, and Adult Testis Size in the Rat Endocrinology, October 1, 2008; 149(10): 5280 - 5287. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. R Hutchison, H. M Scott, M. Walker, C. McKinnell, D. Ferrara, I. K. Mahood, and R. M Sharpe Sertoli Cell Development and Function in an Animal Model of Testicular Dysgenesis Syndrome Biol Reprod, February 1, 2008; 78(2): 352 - 360. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Warita, K. Okamoto, K.-i. Mutoh, Y. Hasegawa, Z.-P. Yue, T. Yokoyama, Y. Matsumoto, T. Miki, Y. Takeuchi, H. Kitagawa, et al. Activin A and Equine Chorionic Gonadotropin Recover Reproductive Dysfunction Induced by Neonatal Exposure to an Estrogenic Endocrine Disruptor in Adult Male Mice Biol Reprod, January 1, 2008; 78(1): 59 - 67. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M L Gould, P R Hurst, and H D Nicholson The effects of oestrogen receptors {alpha} and {beta} on testicular cell number and steroidogenesis in mice Reproduction, August 1, 2007; 134(2): 271 - 279. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. M. Scott, G. R. Hutchison, I. K. Mahood, N. Hallmark, M. Welsh, K. De Gendt, G. Verhoeven, P. O'Shaughnessy, and R. M. Sharpe Role of Androgens in Fetal Testis Development and Dysgenesis Endocrinology, May 1, 2007; 148(5): 2027 - 2036. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Schauwaers, K. De Gendt, P. T. K. Saunders, N. Atanassova, A. Haelens, L. Callewaert, U. Moehren, J. V. Swinnen, G. Verhoeven, G. Verrijdt, et al. Loss of androgen receptor binding to selective androgen response elements causes a reproductive phenotype in a knockin mouse model PNAS, March 20, 2007; 104(12): 4961 - 4966. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Assinder, R. Davis, M. Fenwick, and A. Glover Adult-only exposure of male rats to a diet of high phytoestrogen content increases apoptosis of meiotic and post-meiotic germ cells Reproduction, January 1, 2007; 133(1): 11 - 19. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. Ferrara, N. Hallmark, H. Scott, R. Brown, C. McKinnell, I. K. Mahood, and R. M. Sharpe Acute and Long-Term Effects of in Utero Exposure of Rats to Di(n-Butyl) Phthalate on Testicular Germ Cell Development and Proliferation Endocrinology, November 1, 2006; 147(11): 5352 - 5362. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. Delbes, C. Levacher, and R. Habert Estrogen effects on fetal and neonatal testicular development. Reproduction, October 1, 2006; 132(4): 527 - 538. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P R Dalsenter, G M Santana, S W Grande, A J. Andrade, and S L Araujo Phthalate affect the reproductive function and sexual behavior of male Wistar rats Human and Experimental Toxicology, June 1, 2006; 25(6): 297 - 303. [Abstract] [PDF]
|
|
