Mouse models of altered gonadotrophin action: insight into male reproductive disorders

Enhanced LH–LHR activity

To examine the effects of promiscuous LHR activation, our laboratory generated two transgenic mouse models with enhanced LH/human chorionic gonadotrophin (hCG) action. The first model generated expressed the hCGβ (Tg(UBC‐CGB)1Hht) subunit under the human ubiquitin C promoter, allowing ubiquitous, persistent and low-level expression of hCGβ from late gestation onwards (Rulli et al. 2003). The rationale behind this was that when the transgene was coexpressed in pituitary gonadotroph and thyrotroph cells with the glycoprotein hormone common α-subunit (CGA (αGSU)), bioactive heterodimers hCGαβ would be produced. We termed this model hCGβ, and in males, it attained moderately three- to fourfold elevated levels of bioactive hCG compared with endogenous LH (Rulli et al. 2003). hCGβ+ males were fertile with full spermatogenesis and normal sperm quality despite reduced testis size and serum FSH (Rulli et al. 2003), echoing the phenotype observed in activating LHR mutations in humans. However, the onset of puberty was normal, with no evidence of precocious puberty, which is the hallmark of human males with enhanced LHR activation (Themmen & Huhtaniemi 2000). As modest elevation in LH/hCG action had no effect on fertility or the timing of puberty, we went on to test the effect of grossly elevated LH/hCG on these factors. To achieve this, we generated another mouse model expressing the αGSU, also under the human ubiquitin C promoter, and crossed with the hCGβ+ mice, creating a double transgenic line (hCGαβ+), with a 1000-fold higher circulating concentration of bioactive LH/hCG observed when compared with WT mice (Rulli et al. 2003). hCGαβ+ males were infertile, despite exhibiting comparable spermatogenesis as evidenced by histological analysis of testis and caudal epididymal sperm motility and morphology to hCGβ+ and WT littermates. Infertility appeared to be mechanical and/or behavioural in origin, with hCGαβ+ males displaying extremely aggressive behaviour, often resulting in severe injury or death of WT females housed with the males, and mating ability impaired as evidenced by the lack of vaginal plugs during breeding studies. Testes size was smaller with enlarged seminal vesicles and prostate, dilated vasa deferentia and bladder, as well as kidney defects in adulthood (Rulli et al. 2003). Testicular steroidogenesis was also enhanced, despite a near-total down-regulation of cell surface LHR expression, echoing studies showing that <0.1% occupation of LHR is required for full testicular steroidogenesis (Mendelson et al. 1975). As with hCGβ+ males, precocious puberty was not detected in hCGαβ+ males, despite highly elevated serum testosterone with the timing of the balano-preputial separation and onset of spermatogenesis indistinguishable from WT males (Ahtiainen et al. 2005). Interestingly, juvenile hCGαβ+ males developed Leydig cell adenomas, reaching their maximum size at 10 days post partum but disappearing by puberty, coinciding with the normal regression pattern of foetal Leydig cells. The gene expression of foetal and adult Leydig cell markers suggested that the adenomas originated from the foetal Leydig cell population, providing evidence that the adult Leydig cells may be resistant to developing gonadotrophin-induced adenomas (Ahtiainen et al. 2005).

Recent studies of the hCGαβ+ animals have revealed that the hypothalamic function of prepubertal males was altered, displaying accelerated GnRH pulse frequency and increased GnRH content of GnRH neurons, coupled to decreased pituitary expression of GNRHR (Gonzalez et al. 2011). A profound and persistent malfunction of the neuroendocrine feedback control of the gonadotrophin axis was evidenced, with FSH levels persistently low throughout life and unresponsive to castration or the anti-androgen flutamide both pre- and postpubertally, but with re-establishment by blockade of perinatal androgen action (Gonzalez et al. 2011). These findings suggest that androgen excess, during a critical window between gestational day 18 and postnatal day 14, is able to disrupt the developmental programming of the male hypothalamic–pituitary–gonadal axis. A direct testosterone-dependent regulation of hypothalamic aromatase expression was also demonstrated, indicating that locally produced oestrogens might play a key role in the hypothalamic–pituitary phenotype of hCGαβ+ mice.

Additional GM models to test the effects of elevated hCG or LH have also been utilised by others. A transgenic model overexpressing hCGβ expressed under the metallothionein (MT1) promoter did not show elevated circulating dimeric hCG nor obvious changes in testicular phenotype, yet MT1-hCGβ males were infertile, speculated to be due to free circulating hCGβ subunit binding to LHR and competing with endogenous LH for receptor occupancy (Matzuk et al. 2003). Coexpression of MT1-hCGα and hCGβ subunits was conducted, to form the active hCG heterodimer. Male mice with low expression of MT1-hCGαβ were initially fertile and indistinguishable from WT littermates. However, by 6–7 months, these mice were progressively infertile but no histological abnormalities were observed and no obvious phenotypic explanation was available to indicate the cause of infertility. Male mice with high expression of the MT1-hCGαβ transgenes, as with ubiquitin C-expressed hCGαβ male mice, were infertile, the origin of which appearing to be through disrupted mating behaviour as evidenced by the lack of vaginal plugs when housed with either super-ovulated or naturally cycling female mice. Male mice were also noted to be aggressive when caged with other male or female mice, and displayed altered sexual behaviour. Serum testosterone was highly elevated, and circulating gonadotrophins decreased. Testis size was reduced, and histological analyses indicated Leydig cell hyperplasia and in some tubules Sertoli cell-only-like syndrome, with germ cell loss, echoing observation of LHR over-activity in humans. A transgenic model for elevated LH consisting of a fusion protein of the bovine LHβ subunit and the hCGβ C-terminal peptide (bLHβ-CTP) under the common α-subunit promoter has also been studied. However, this model failed to produce sufficiently elevated LH/hCG bioactivity in male animals, as they presented with no apparent phenotype (Risma et al. 1995).

To constitutively activate LHR, a novel transgenic approach using covalently linked hCGβ and αGSU to reconstitute heterodimeric hCG, fused to rat LHR expressed under inhibin α subunit promoter, termed ‘yoked’ LHR (YHR), was utilised (Meehan et al. 2005, Meehan & Narayan 2007). In pre-pubertal males, enhanced LH/LHR action was observed, with increased circulating testosterone and seminal vesicle weights, probably due to the early expression of the transgene driven by the inhibin α promoter. However, despite this elevation in testosterone, as with the hCGβ+ and hCGαβ+ animals, the timing of puberty was normal. After puberty, there was a trend for enhanced LHR action, with decreased seminal vesicle weights and reduced testis size due to a decrease in seminiferous tubule volume. However, normal spermatogenesis was noted. As with hCGβ+ animals, serum FSH was suppressed in both pre- and postpubertal animals; however, LH was only suppressed in pre-pubertal animals. This defect may be the consequence of dysregulation in hypothalamic–pituitary communication and may reflect differences in the regulation of LH and FSH secretion.

To date, a single CAM LHR mouse model has been described, the result of a knockin D582G LHR mutation, the most commonly observed CAM in human boys with familial male-limited precocious puberty (McGee & Narayan 2013). As with the human mutation, D582G LHR resulted in precocious puberty, with decreased testis weight and increased seminal vesicle weight at 3 weeks post partum. Serum and intra-testicular testosterone were increased from day 7 post-partum; however, serum FSH and LH remained below the limit of detection throughout the tested lifespan of the animals, due to steroid hormone feedback. Sertoli cell development was unaltered; however, Leydig cell hyperplasia was observed, with enhanced expression of steroidogenic genes in most age groups tested. Although precocius puberty was observed, spermatogenesis was not altered in these male mutants. Although initially fertile, progressive infertility was detected, but normal levels of epididymal sperm were noted, indicating a potential abnormality in seminal vesicle and prostate function and/or lower urinary tract; however, detailed analysis of accessory gland function was not carried out.

Enhanced FSH–FSHR activity

GM mouse models with elevated FSH have been generated to explore enhanced ligand-dependent activation of FSHR. As with MT-hCGβ and MT-hCGαβ, Kumar et al. took the approach of overexpressing human αGSU and the human FSHβ subunit under the MT1 promoter. The MT1-αGSU and MT1-FSHβ transgenic mice were fertile. Inter-crossing of these transgenic mouse strains generated mice overexpressing dimeric FSH (MT1-FSHαβ), with high levels of circulating FSH. Male mice were largely infertile, with just one in ten animals producing one litter of pups in a 6 month period. Mating studies suggested a lack of mating activity in these animals. Testicular size and morphology were indistinguishable from WT; similarly, epididymal weights were comparable. However, serum testosterone was elevated and seminal vesicles enlarged, due to increased androgenic action. Histological analysis of the testes showed little difference from WT; moreover, analysis of epididymal sperm numbers showed MT1-FSHαβ animals to have increased sperm number, with no difference in motility or viability. These findings suggest that the infertility observed in MT1-FSHαβ animals appears to result from behavioural changes rather than a direct impact on spermatogenesis. It is possible that the increase in testosterone resulted in altered and/or aberrant seminal vesicle secretions, or functional incompetence of the sperm.

Mouse models of enhanced FSHR activity have primarily utilised the hpg mouse model as a background in which to generate the mutations. The hpg mouse, resulting from a naturally occurring deletion mutation in Gnrh1 (Cattanach et al. 1977), with a phenotype of hypogonadotrophic hypogonadism, provides the advantage of testing the effects and direct contribution of FSH/FSHR-dependent testicular function, in the absence of circulating LH and activation of LHR. Using the rat androgen-binding protein promoter for specific integration into Sertoli cells, Haywood et al. (2002) created a transgenic line expressing the human Asp567Gly FSHR CAM (TG-FSHR+). Testicular expression was confirmed, and enhanced ligand-independent cAMP production was detected in cultured TG-FSH+ Sertoli cells. In a WT background, testis weights and fertility were comparable between TG-FSH+ animals and WT littermates. However, in the hpg background, testis weights were significantly increased in comparison to hpg littermates; moreover, treatment with testosterone at equivalent levels to the maximum observed in hpg mouse testis vastly increased testis size in hpg TG-FSH+ animals in comparison to hpg littermates. Histological analysis of the testes showed the presence of both round and elongated spermatids, and examination of Sertoli cell structure showed the maturation of this cell type. Although intra-testicular testosterone was increased in hpg TG-FSHR+ animals, serum testosterone was not different from hpg littermates. A similar phenotype was also observed in a transgenic model overexpressing complete FSH (αGSU and FSHβ subunits) in a WT or hpg mouse background (Allan et al. 2001), demonstrating that without LH-induced testosterone production, FSH/FSHR activity is sufficient for Sertoli cell maturation and can promote spermatogenesis to some extent. However, LH/LHR activity, and consequential testosterone production, is required for the completion of spermiogenesis.

In our laboratory, a knockin constitutively activating mFshrD580H mouse model has been generated (Oduwole/Peltoketo personal communications). Interestingly, despite this mutation having deleterious effects on female reproduction (Peltoketo et al. 2010), male animals did not present with any obviously altered phenotype during embryogenesis, puberty or adulthood. The gross morphology and histology of the reproductive tract and testis appeared no different to WT littermates, indicating that enhanced FSHR activity alone in the WT background, as opposed to hpg mice, had neither positive nor deleterious effects on male reproductive function.

Diminished LH–LHR activity

The first GM approach exploring the effects of loss of function of gonadotrophins utilised deletion of αGSU. Deletion of the αGSU gene in male mice showed normal pre-natal and pre-pubertal sexual differentiation and gonadal development, confirming that pre-pubertal gonadal development in mice is independent of gonadotrophin action (Kendal et al. 1995). However, male animals, being also hypothyroid, failed to undergo puberty and exhibited a lack of sex steroid production. Post-pubertal animals lacked gonadal development and function, with diminished testis size and smaller seminiferous tubules, and spermatogenesis blocked at the first meiotic division. The presence of vas deferens and epididymis showed that the αGSU KO mice were able to produce sufficient testosterone in utero. As the αGSU gene is an integral part of both heterodimeric thyroid-stimulating hormone (TSH) and FSH, it should be noted that phenotypic effects observed from deletion of αGSU are not the result of lacking just LH action, but also TSH and FSH action. The mouse model demonstrated that mice devoid of glycoprotein hormone production are viable, which is perhaps not unsurprising given that mice do not express or secrete placental CG, and rather rely upon placental lactogens and alternative hormonal support for maintenance of pregnancy, in contrast to humans in whom hCG is vital.

To decipher the effects of deleting Lhr, our laboratory took the approach of generating an Lhr knockout (LuRKO) mouse. As with the αGSU knockout mice, Lhr deletion resulted in alterations of the reproductive tract from the pubertal period onwards, exhibiting normal pre-pubertal development (Zhang et al. 2004). Elevated FSH and LH were observed, with a decrease in sex steroid concentrations, due to lack of steroid feedback to the hypothalamic–pituitary axis (Pakarainen et al. 2007). Adult LuRKO males were infertile with underdeveloped testes and hypoplastic accessory sex organs. Testes were cryptorchid and significantly reduced in size, with narrow seminiferous tubules, decreased number and size of Leydig cells and arrested spermatogenesis at the round spermatid stage. The expression of Leydig cell-specific genes, while similar at birth, became gradually low or undetectable in adulthood. Accessory sex organs, including the prostate and seminal vesicles, were undetectable (Lei et al. 2001, Zhang et al. 2001). A similar phenotype to the LuRKO mice was also observed with the deletion of LHB, mimicking the reproductive phenotypes displayed in αGSU null male mice (Ma et al. 2004); however, LHB knockout males exhibited unaltered serum FSH, contrasting from the hypogonadotrophic and hypergonadotrophic phenotypes of αGSU and LuRKO male mice respectively.

An interesting difference that exists between human and mouse inactivating LHR mutations, is that normal pre-pubertal development is observed in male mice; however, in human counterparts, complete inactivation of LHR results in pseudohermaphroditism (Themmen & Huhtaniemi 2000). This indicates that LH action in utero is not a prerequisite for foetal Leydig cell androgen and insulin-like growth factor 3 production required for intrauterine testicular development and descent and masculinisation in male mice, highlighting the presence of additional safety mechanisms present for maintaining foetal Leydig cell function by a network of paracrine factors (El-Gehani et al. 1998, Themmen & Huhtaniemi 2000, Peltoketo et al. 2011).

Testosterone replacement therapy in LuRKO animals leads to partial reversal of the hypogonadal phenotype, with achievement of full spermatogenesis; however, male mice remained sub-fertile due to poor accessory gland development and poor sexual behaviour (Pakarainen et al. 2005). Abnormalities such as vigorous inflammation of the epididymis and the prostate were conspicuous in a proportion of the testosterone-treated mice. The incidence of low ejaculatory frequency and low sperm count in cauda epididymides were also observed. Whether or not testosterone replacement, or lack of sufficient androgen priming prepubertally before testosterone replacement, is responsible for these abnormalities is not, however, clear. A striking physiological finding in the LuRKO mice is a late-onset recovery of qualitatively full spermatogenesis at ∼12 months of age, when the passage of round spermatids to elongated spermatids can be found. This suggests that spermatogenesis can proceed qualitatively to completion with support of the basal LH-independent low intra-testicular testosterone present in the LuRKO testis (Zhang et al. 2003), though a much higher threshold of testosterone may be required to induce qualitatively and quantitatively full spermatogenesis (Huhtaniemi et al. 2006). This finding was confirmed and extended in our recent study (Oduwole et al. 2014), observing that a narrow margin separated the testosterone doses that activated peripheral male sexual androgen action and spermatogenesis. When extrapolated to humans, this may jeopardise the current approach to hormonal male contraception, as it will be practically impossible to define a single dose of testosterone that can suppress gonadotrophins and attain azoospermia. Therefore, it is only a total abolition of intra-testicular testosterone action that can bring about total and complete suppression of spermatogenesis.

Diminished FSH–FSHR activity

Targeted ablation of bioactive FSH was achieved through deletion of exons 1 and 2 and partial deletion of exon 3 of FSHβ (Kumar et al. 1997). Phenotypic examination of Fshb (FSHβ) KO males showed reduced testis size, with decreased seminiferous tubule diameter and volume. However, Leydig cell populations were unaffected and speculated qualitatively to be enhanced in number, however, net Leydig cell number probably did not differ from WT littermates due to the reduced testis size. Accessory sex glands were of comparable size to age-matched littermates, consistent with comparable circulating serum testosterone and adequate Leydig cell number and function. Epididymal sperm were decreased by 75% in comparison to heterozygous and WT littermates, with motility decreased by 40%; however, no difference in viability was observed. Despite this, Fshb KO animals were fertile, with normal serum LH, probably reflecting negative feedback from circulating testosterone. The maintenance of spermatogenesis and Sertoli cell function in the absence of FSH-activated FSHR is suggestive of potential testicular or extra-testicular paracrine factors that can compensate for FSH function in the testis, or that basal constituitive ligand-independent FSHR activity is sufficient to maintain tonic testis function and spermatogenesis in male mice.

The generation of FSHR knockout mice (FORKO) provided additional insight into the dependence of spermatogenesis on FSH (Dierich et al. 1998, Abel et al. 2000). As with Fshb null males, FORKO males were fertile, with reduced testis size and decreased spermatogenesis. To examine key differences in these models, a study was conducted to directly compare the phenotypes observed of FORKO and FSHβ GM mice (Baker et al. 2003). Comparison of serum and intra-testicular testosterone showed a reduced level of circulating testosterone in FORKO animals, which was not observed in the FSHβ mouse model; yet both models exhibited diminished intra-testicular testosterone, indicating that local production of testosterone was impaired in both FORKO and FSHβ mice. Serum LH was elevated in FORKO animals, but not FSHβ animals. Interestingly, Leydig cell-specific steroidogenic genes such as Cyp11a1 were diminished in the FORKO model, with decreased Leydig cell number to ∼60% of control, suggesting a potential failure of Leydig cell proliferation and/or differentiation at puberty in FORKO animals, which was not observed in Fshb KO animals. This effect is likely to be reflective of the decreased Leydig cell number observed in these animals and represents a key difference between these animal models. As both models were fertile, these studies revealed that FSH action is not critical for the maintenance of foetal Leydig cells, as shortly after birth, when the maintenance of these cells is critically dependent on gonadotrophin action. As FSHR is expressed solely in Sertoli cells, the action of FSHR on Leydig cell development must be via Sertoli cell-secreted paracrine factors. Previously studies have implicated factors such as desert hedgehog and PDGF; however, to date nothing has conclusively been described to be the key factor(s) mediating these paracrine effects. It is likely that FSHR action mediates and ensures sufficient Sertoli cell activity for output of such trophic factors, and why spermatogenesis is impaired when either FSHβ or FSHR action is abrogated. Whereas Fshb and Fshr KO male mice are fertile, there is some discrepancy in humans on the phenotype of men with inactivated FSH function. The three men described with inactivating FSHB mutations are all azoospermic (Lindstedt et al. 1998, Phillip et al. 1998, Layman et al. 2002), whereas the five men with inactivating FSHR mutations have oligozoospermia of variable severity (Tapanainen et al. 1997). This discrepancy can be resolved only through detection of new cases of these extremely rare mutations.