Background
Traditionally, the regulation of mammalian spermatogenesis has been viewed as being dependent upon the interactions of gonadotrophins and androgens, however, recent observations in experimental rodents and man have indicated additional roles for estrogen in this process [[
1,
2] for reviews]. Human studies have demonstrated the presence of aromatase in the testis together with the expression of two splice variants of ERβ in somatic and germ cell compartments, though expression of ERα appears to be restricted to the efferent ductules [
3]. The importance of estrogen in male reproduction is illustrated by the infertility displayed by men with inactivating mutations in the gene encoding aromatase or ERα [
4]. Mice have been widely used to investigate the function of estrogens in male reproduction, indeed the targeted deletion of estrogen receptors [
5] and aromatase [
6] provided the first evidence for a role for estrogen in sperm production. In contrast to primates, ERα is expressed abundantly in the Leydig cells of mice, though it seems likely that the infertility in male ERα knockout mice results from inadequate fluid resorption in the efferent ductules rather than a primary deficit in spematogenesis [
7]. The function of ERα in the interstitial cells of the rodent testis is not understood.
We and other research groups have exploited the hypogonadal (
hpg) mouse [
8], which cannot produce mature GnRH decapeptide due to a truncation in the GnRH gene, as a model to study the role of estrogen in male reproductive function [
9,
10].
Hpg mice are infertile because they do not produce gonadotropin and, in consequence, testis development does not proceed beyond the neonatal stage [
11]. It is well established that treatment of
hpg mice with LH can stimulate steroidogenesis [
12] and treatment with FSH and androgens can induce qualitatively normal spermatogenesis [
13,
14]. However, we have also observed that chronic estradiol treatment alone can induce spermatogenesis in
hpg mice [
9,
15,
16], though this is associated with paradoxical increases in pituitary and serum FSH concentrations. The importance of this estrogen-induced increase in FSH secretion is not known: it has not been established whether estradiol acting directly at the level of the testis is sufficient to produce qualitative spermatogenesis in
hpg mice or whether the effects of estradiol in inducing spermatogenesis are more indirect, occurring via the rise in FSH production.
This aim of the study was to investigate the relative contribution of direct (testis) and indirect (pituitary) actions of estradiol in stimulating spermatogenesis. The initial objective of the study was to use immunohistochemistry and Western blotting to localise ERα in the pituitary gland and testis of hpg mice, in order to identify putative sites of estrogenic action in promoting spermatogenesis in this mutant strain. The main objective was to measure the qualitative and quantitative effects of estradiol on testis development and spermatogenesis in hpg mice in the presence and absence of rhFSH. If combined estradiol and rhFSH treatment proved to be more effective than either of these treatments alone (i.e. there was a synergistic effect), we would conclude that these hormones exert separate actions. If, however, combined estradiol and rhFSH treatment were to be only as efficacious as would be expected from a higher rhFSH dose (i.e. an additive effect) then we would conclude that estradiol most likely acts via stimulation of FSH secretion.
Methods
General
All animal procedures were approved by the University of Nottingham Local Ethical Review Committee and carried out in accordance with the Animals Scientific Procedures Act (UK) 1986 (project licence PPL 40/2372). Laboratory Animal Science Association (LASA) guidelines were followed for administration of substances [
17]. Adult
hpg male mice were identified by their micropenis, reduced anogenital distance and small scrotal sac.
Western blot procedure
Pituitary glands and testes from hpg and age-matched littermates mice (approximately 135 days old) were placed in phosphate-buffered saline at pH7.4 (PBS) containing a protease inhibitor. Tissues were sonicated on ice for 3 × 10 sec, centrifuged at 1300 RPM for 5 min and the supernatant extracted and stored at -20°C until used. Testis and pituitary protein were incubated at 100°C for 5 mins in the presence of 5% mercaptoethanol. Samples were then immediately cooled on ice and loaded onto the gel. Protein samples were separated on a 10% SDS gel and transferred to a nitrocellulose membrane (Amersham). Membranes were blocked with 5% milk (Marvel) in PBS at 25°C for 1 hour before incubation overnight at 4°C with a polyclonal antisera raised in rabbit recognising ERα (1:50 dilution of MC-20, Santa Cruz, USA). After overnight incubation membranes were washed in PBS, then PBS/Tween 20, and subsequently incubated with 1:2000 anti-rabbit IgG peroxidase conjugate (Sigma, Poole, UK). Visualization of the protein was achieved using the supersignal system (Pierce, USA). The molecular weight of the protein was estimated by comparing its mobility with standard marker proteins of known molecular weights. The relative amounts of the protein indicated by the intensity of the bands was evaluated by analysing scans of the film and performing densitometry using the NIH Image computer package.
Immunohistochemistry
Testis and pituitary tissue was fixed by immersion in 4% paraformaldehyde, then embedded in paraffin wax so that 5 um sections could be cut on a microtome. Mounted sections were then placed in xylene to remove the paraffin wax, and placed in decreasing concentrations of ethanol and then tap water to rehydrate the tissue. The pituitary and testis sections were subject to antigen retrieval in 0.01 M citrate buffer pH6 by microwaving at full power (650 W) for 2 × 5 mins. For the studies on estrogen receptor distribution, sections were washed in phosphate buffered saline (PBS) buffer and blocked for endogenous peroxidase with 3% hydrogen peroxide. Following rinsing in PBS, samples were blocked with normal goat serum (Vector Labs, Peterborough, UK) then incubated overnight at 4°C with the previously validated rabbit polyclonal ERα antiserum (MC-20, Santa Cruz) at a dilution of 1:50. Sections were washed in PBS and incubated with a biotinylated anti-rabbit IgG at 25°C for 1 hour followed by an avidin biotin horseradish peroxidase complex (Elite ABC: Vector Labs, Peterborough, UK). Immunoreactivity was revealed by incubation with diaminobenzidine (DAB) and the reaction stopped by immersing sections in tap water. Sections were counterstained with eosin. After dehydration through increasing concentration of ethanol, sections were mounted using DePeX mounting medium (Sigma, Poole, UK).
For the studies on where the expression of the steroidogenic enzymes 3β hydroxysteroid dehydrogenase (3βHSD) was used to identify Leydig cells, a similar protocol was followed but omitting the pre-treatment with 3% hydrogen peroxide. Sections were incubated overnight in a polyclonal rabbit anti-3βHSD antiserum (dilution of 1:1000) donated by Prof Ian Mason, University of Edinburgh, and previously validated for use in rodents [
18,
19]. Sections were washed in PBS and incubated with a biotinylated anti-rabbit IgG at 25°C for 1 hour followed by a fluorescein-avidin complex (Vector Labs). After washes in PBS then distilled water, sections were coverslipped using Vectashield (Vector Labs), sealed with nail polish and viewed under a Leica DMRB Fluorecence microscope.
Experimental protocol: effects of estradiol and rhFSH
Under general anesthesia, male
hpg mice ranging in age from 3–7 months were implanted with either a subcutaneous capsule containing 2% estradiol (n = 11) or cholesterol (n = 15) as previously described [
9]. Subgroups of
hpg mice bearing estradiol implants were then treated daily with 1 IU rhFSH (n = 6) or vehicle (n = 5) for 50 days. Correspondingly, subgroups of
hpg mice bearing cholesterol implants were treated daily with 1 IU rhFSH (n = 5) or 5 IU rhFSH (n = 6) or vehicle (n = 5) for 50 days. The doses of rhFSH were based on those used in a previous dose-response study in
hpg mice and were designed to produce a submaximal (1 IU/day) and a maximal (5 IU/day) effect [
14]. On experimental day 50, blood samples were collected by cardiac puncture under terminal anesthesia (sodium pentobarbitone, i.p.) 2 h after the last FSH treatment. Blood was allowed to clot at room temperature before being spun at 13000 rpm and then serum samples were stored at -20°C until assayed. After blood collection, the pituitary, testes, epididymides and seminal vesicles were excised, trimmed of fat and connective tissue and weighed. One testis was placed in Bouin's fixative, and subsequently embedded in paraffin wax so that 5 um sections could be cut on a microtome and stained with haematoxylin-eosin for light microscope analysis.
Morphometric (stereological) analyses
The primary morphometric aims were to estimate volumes of seminiferous tubules and seminiferous epithelium within testes. To this end, stereological estimates [
20,
21] were made on tissue images randomised for location and orientation by a multistage systematic uniform random sampling design. Approximately four sections per testis, and six microscopical fields per section, were randomly chosen for analysis. Fields were sampled as images captured on a Leica DM4000B brightfield microscope via OpenLab software (Improvision, Coventry, UK) and printed to a final magnification of ×1000 using a stage micrometer scale as an external calibration standard. Volume densities of testicular ingredients were determined by randomly superimposing a transparent grid comprising 35 test points arranged in a quadratic array. Test points falling on a given testis and its ingredients were summed over all fields from all sections. The total number of points landing on a given ingredient (tubule, lumen, epithelium), divided by the total number of points landing on the testis sections, provided an unbiased estimate of its volume density. The absolute volume of each ingredient per testis was estimated subsequently from the product of the volume density and the processed testicular volume. In order to assess whether changes in tubule volume were attributable to changes in calibre or total length, mean diameters of seminiferous tubules were measured after selecting tubules using an unbiased counting frame [
20]. In the case of elliptical profiles, the short axis was measured. To give an indication of the degree of germ cell development, each of the 24 fields per testis collected for stereological analysis was also scored by an observer blind to the experimental treatment to indicate the most mature germ cell type present.
Quantitative assessment of Leydig and Sertoli cell numbers was made in order to obtain relative measures of net cell proliferation. Estimates were obtained by randomly superimposing a forbidden line unbiased counting frame [
20] on enlarged photomicrographs of tissues immunostained for 3β-HSD (Leydig cell marker) or on tissue sections stained with haematoxylin and eosin where the nuclei of Sertoli cells could be unequivocally identified by their characteristic location and appearance (eg tripartite nucleolus, [
11]). Using this counting frame, the numbers of nuclear profiles per unit area of testis were used to obtain model-based estimates of the numbers of nuclei per unit volume of tissue (N
v). To this end, we obtained estimates of section thickness (T, always 5 μm) and the mean diameters (D, in μm) of nuclei for each cell type. To estimate the mean nuclear diameter, the diameters of 10 randomly selected nuclei were measured. On the explicit assumption that each cell contains a single nucleus, the total number of specific cell types per testis (N) was then calculated from
where V
testis represents testis volume (cm
3) was calculated by dividing testis mass (g) by tissue density. The latter was taken to be 1.05 g/cm
3 [
22].
Hormone assays
Serum FSH concentrations were measured using both a commercially available human FSH ELISA kit (IDS Ltd UK) to determine the levels of recombinant hFSH circulating, and a rat FSH immunoradiometric kit (IDS Ltd UK) in order to estimate the total circulating FSH concentrations (i.e. endogenous and exogenous). FSH concentrations in extracts of mouse pituitary gland, or in untreated wild-type mouse serum, did not dilute in parallel with the rat standards, and there was substantial cross reactivity of rhFSH in the rat IRMA (1 mIU rhFSH reading as approximately 1.5 ng/ml of the rat FSH standard) so all experimental samples were compared at the same dilution in a single assay and must be considered as relative concentrations rather than absolute values. The minimum detection limit in the rat FSH IRMA was 0.2 ng/ml and the intra-assay CV was 2.0%. The minimum detection limit in the human FSH ELISA was 1 mIU/ml and the intra-assay CV was 7.5%.
Statistical analysis
Data for organ weights were analysed using one-factor ANOVA followed by Tukey's multiple comparison test (Prism v4.0, GraphPad Software, San Diego, CA). FSH values were analysed using Kruskal-Wallis non-parametric tests (also Prism) because when values were below the limit of detection of the assay they were assigned that value for the purposes of statistical analysis.
Discussion
The first objective of these studies was to confirm the expression of ERα receptors in the pituitary gland and testis of
hpg mice. Complementary immunohistochemical and Western blot studies revealed that ERα immunoreactivity is present in both the pituitary and testis of
hpg mice at levels comparable to that in wild-type litter mates. The immunohistochemical studies demonstrated that the distribution of ERα in the testis of the
hpg mouse was restricted to putative Leydig cells in the interstitium and peritubular myoid cells, consistent with previous studies in the mouse [
23]. Interestingly, we also observed ERα immunoreactivity in angular nuclei adjacent to the basal lamina in testes from wild-type littermates, these are very likely to be Sertoli cells. Previous studies in mice have not reported ERα expression in this cell type [
23]. Our observations suggest that estradiol could potentially act via ERα in both the pituitary and testis in the
hpg mouse. We were unable to confirm using immunohistochemistry whether ERβ is also expressed in the testis of the
hpg mice due to technical limitations of the antiserum, but Western blots did reveal protein bands of the appropriate molecular weight of ERβ in tissues from both
hpg and wild-type mice (data not shown). Previous studies have provided rather conflicting evidence for ERβ expression in the mouse testis, some studies demonstrating ERβ expression in the Leydig cells and elongating spermatids of adult mice [
24], others reporting a developmental decline in ERβ expression in the mouse testis, with immunoreactivity being undetectable in the adult [
25]. As ERβ expression does not appear to be required for normal testis function in the mouse [
26], our studies have focussed on the role of the ERα receptor.
We again observed robust stimulatory effects of estradiol treatment on the testis in
hpg mice, consistent with our previous studies [
9,
15,
16]. The current study revealed that these effects of estradiol were both qualitatively and quantitatively very similar to the effects of treatment with FSH. It is well established that treatment with FSH alone will induce testicular development in
hpg mice [
14], though we observed a greater degree of development. In our study, both doses of rhFSH induced germ cell development up to the elongating spermatid stage and the also the formation of small lumina in the tubules, whereas Singh and Handelsman [
14] reported that the same doses of rhFSH applied for the same duration (50 days) mainly stimulated premeiotic germ cell proliferation, and that relatively few cells progressed to the round spermatid stage. We observed a significant increase in numbers of Sertoli cells after estradiol and high dose rhFSH treatment, which Singh and Handelsman [
14] did not detect. Given that the estradiol treatment clearly elevated circulating FSH concentrations, this observation is consistent with the key role of FSH in Sertoli cell proliferation in mice [
27], and with more recent studies where genetic induction of FSH production in
hpg mice has increased Sertoli cell production [
28]. We also observed an increased number of Leydig cells in
hpg mice treated with the high dose of FSH. Stimulatory actions of FSH on Leydig cell activity and morphology have been reported previously in
hpg mice and are likely to reflect indirect actions mediated by paracrine output from Sertoli cells [
29]. Unfortunately insufficient serum was available to measure circulating testosterone concentrations after the various experimental treatments as we considered it to be much more important to measure endogenous and recombinant FSH in the serum that we collected in order to verify the efficacy of our treatment protocols. In our previous studies we have been unable to measure detectable increases in either serum testosterone [
9,
15] or in intratesticular testosterone [
15] after estradiol treatment, but on the basis of previous studies in
hpg mice it seems likely that the FSH treatments would have increased testicular androgen production [
29].
Though model-based, our estimates of cell numbers are consistent with those obtained by ourselves and others using more robust, design-based stereological procedures [
9,
27,
30,
31]. An important difference between the current study and that of Singh and Handelsman [
14] is that in the latter study the 50-day rhFSH treatment was initiated at a juvenile age (day 21), whereas we treated
hpg mice at an "adult" age. We have previously observed that there is a degree of spontaneous testicular development in
hpg mice probably reflecting a low level of GnRH-independent production of gonadotropins [
1]. Thus, our experimental rhFSH treatments may well have been superimposed upon a marginally higher endogenous level of testicular development or androgen production, and given the key role for androgens in murine spermatogenesis [
13,
15] this seems a highly likely explanation for the differences between the current and the previous study.
The key outcome of the current studies is that the combined estradiol and rhFSH did
not produce a greater degree of stimulation of the seminiferous epithelium that treatment with either hormone alone. This is contrary to our initial prediction, and to studies in the neonatal rat where combined treatment with estradiol benzoate and purified human FSH produce effects on development of the seminiferous epithelium not seen after treatment with either hormone alone, including increased tubule diameters, induction of a lumen and increased numbers of pachytene spermatocytes [
32]. The observation that treatment with a higher dose of rhFSH (5 IU daily) did increase the volume of the seminiferous epithelium and tubules further demonstrates that the lack of additive or synergistic effect is not a consequence of maximal proliferation rates already being achieved. Given the qualitative and quantitative similarity of the effects of estradiol and rhFSH that we observed, we conclude that the primary action of estradiol on testis function is indirect, via the stimulation of FSH release, in the
hpg mouse experimental model. Although this stimulatory effect of estradiol may seem like a paradoxical response in a male mammal, it was demonstrably the case that circulating FSH levels were increased in this and previous studies in the
hpg mouse. This paradoxical effect in the male is not due to incomplete masculinization of the
hpg mouse [
16], rather it appears to be a reflection of the fact that development of the reproductive axis in the
hpg mouse is arrested at a neonatal stage [
1,
11] when steroid negative feedback pathways have not yet developed. It is worth noting that certain actions of estradiol on the male reproductive tract in mice
are highly likely to be direct actions. In the current study we observed that estradiol treatment produced a substantial increase in the wet weight of the seminal vesicles, whereas FSH had no effect on this tissue nor did it augment the actions of estradiol. This is consistent with a previous study in the
hpg mouse [
9], which also provided evidence for direct effects of estradiol on the prostate gland [
9].
Recent evidence points to an autocrine or paracrine role for estradiol in the testis during fetal and early neonatal development whereby it inhibits production of androgens by Leydig cells [
33]. Perhaps the endocrine actions of estradiol revealed in the
hpg mouse represent a physiological role in co-ordinating early development. Whilst estradiol production acts within the testis to retard early androgen production, it also acts in the pituitary gland to stimulate FSH production and thereby promote early Sertoli cell proliferation. Thus, when pubertal increases in LH production eventually occur and thereby stimulate androgen production, the somatic cell complement within the tubules has already been established and will support the androgen-driven spermatogenesis.
Competing interests
The author(s) declare that they have no competing interests.
Authors' contributions
HB carried out the tissue processing and stereological analyses on the seminiferous tubules, MOW carried out the in vivo studies and the Western Blot and ER immunohistochemistry, GRH and RAW carried out the immunostaining for 3β-HSD and the quantitative assessment of Leydig and Sertoli cell numbers, TMM developed the stereological procedures and oversaw their application, FJPE managed the project and wrote and revised the manuscript.