Role of FSH in the regulation of spermatogenesis: clinical aspects

Under normal conditions male reproduction in mammals requires four distinct components which have to interact in a concerted fashion with female reproductive functions to guarantee successful procreation: •sperm as the actual carrier of genetic information and as the fertilizing agent; •semen as the transport medium for sperm; •organs to transfer semen into the female genital tract; • male sexual behaviour. In addition to genetic, neuronal and vascular integrity hormones are essential for the regular function of the four components. In particular, the endocrine feedback-mechanism involving hypothalamus, pituitary and testis is responsible for each component and their coordinated interplay. One hypothalamic hormone, GnRH, is responsible for the release of two gonadotrophins, FSH and LH, and the exact role of the two gonadotrophins in male reproduction remains an enigma of andrological research. Through mediation by specific receptors both gonadotrophins act on the testis: LH on Leydig cells and FSH on Sertoli cells, and it is unclear why the single hypothalamic signal is translated into a dual pituitary signal controlling male reproductive functions. The early concept of regulation of testicular function surmised that LH stimulates testosterone production and FSH controls spermatogenesis (Greep, 1937). However, so far only one mammalian animal model could be discovered where such a clear separation of the two functions may occur, i.e. the Djungarian hamster. In this seasonal breeder FSH can induce full spermatogenesis and viable sperm (function 1 above) while testosterone appears to be required only for functions 2–4 (Lerchl et al., 1993; Niklowitz et al., 1997). In most other mammals, in particular, in nonhuman and human primates the roles played by the two gonadotrophins in the regulation of male reproductive functions appear to be much more closely interwoven. However, recent findings in genetically manipulated mice have suggested that FSH may not be required for spermatogenesis (e.g. Kumar et al., 1997; Dierich et al., 1998). Although extrapolating data from small laboratory rodents to the humans has its limitations, these findings have left many clinicians in a quandary concerning the role of FSH in human spermatogenesis and its implications for diagnosis and treatment of male hypogonadism and infertility. The purpose of this brief review is to summarize the current knowledge about the role of FSH as distinct from that of LH and testosterone in human male reproduction. Experimental evidence from nonhuman primates as the most relevant animal model is used to substantiate findings in humans. Since initiation of spermatogenesis may have different hormonal requirements than maintenance of the process, the two situations are considered separately. Moreover, a distinction between qualitatively and quantitatively normal spermatogenesis will be drawn. The evidence cited for a role of FSH alone or LH/testosterone alone or for synergism of the two gonadotrophins is summarized in Tables 1 and 2. Patients with hypogonadotrophic hypogonadism represent a valuable experimental model to study the role of gonadotrophins in spermatogenesis. Patients with secondary hypogonadism due either to idiopathic hypogonadotrophic hypogonadism (IHH) and Kallmann's syndrome or to hypopituitarism can be effectively treated with pulsatile GnRH or a combination of both gonadotrophins in order to induce spermatogenesis and achieve fertility (Whitcomb & Crowley, 1990; Burgues et al., 1997; Büchter et al., 1998). Under these treatments sperm appear in the ejaculate and sperm concentrations increase progressively. However, sperm concentrations reach normal values in only few patients, although this is not necessary for the induction of pregnancy, and the time required to achieve a pregnancy can be quite variable (Büchter et al., 1998). Under clinical conditions it appears that both gonadotrophins, either stimulated by GnRH or given directly, are required to achieve fertility in such patients while FSH alone or in combination with low-dose testosterone is not able to induce fertility (Schaison et al., 1993). This failure of FSH treatment to induce spermatogenesis in hypogonadotrophic patients contrasts with the observation of the fertile eunuch syndrome, i.e. hypogonadal patients with isolated lack of LH/testosterone action, but with active spermatogenesis and sperm in the ejaculate (Pasqualini & Burr, 1950). In one of these patients (Axelrod et al., 1979) a gene defect for the LH-β subunit could be detected (Weiss et al., 1992). It has, however, been argued that such patients may have enough residual LH activity to produce intratesticular testosterone sufficient for spermatogenesis, thus preventing these patients from being unanimously accepted as the model example that FSH alone would be sufficient to initiate spermatogenesis (Smals et al., 1978). It could also be that FSH itself may stimulate testosterone production in Leydig cells, mediated by Sertoli cell factors, thus providing enough testosterone necessary for spermatogenesis by its own action (Levalle et al., 1998). The concept that both gonadotrophins are required for fully normal spermatogenesis was recently confirmed by the description of men with inactivating FSH-receptor mutations. The five patients had reduced testes volumes, but sperm were present in the ejaculate and two had fathered children (Tapanainen et al., 1997). Assuming that the receptor mutations were able to abolish FSH action totally, these cases would demonstrate that LH/testosterone action alone would be sufficient to induce spermatogenesis and sperm in the ejaculate. However, the question arises how completely the FSH-receptor mutation had inactivated FSH action, as FSH circulated in high concentrations in all patients. Similarly, the recent description of a single azoospermic patient with a mutation in the FSH β subunit gene (Phillip et al., 1998) does not convincingly demonstrate that the lack of FSH alone would cause azoospermia, because this 18 year-old patient had an additional defect in Leydig cell function evidenced by high LH and low testosterone. On the other hand, another man with a FSH β gene mutation was azoospermic and normally virilized, suggesting that FSH action might indeed be absolutely necessary for the initiation of spermatogenesis (Lindstedt et al., 1998). Earlier, patients with isolated FSH deficiency as diagnosed by serum FSH determinations were described (Rabinowitz et al., 1975; Mozaffarian et al., 1983; Al-Ansari et al., 1984). But again, these patients had additional defects including cryptorchidism and 47XXY karyotype, which makes it difficult to attribute the effects on spermatogenesis solely to a lack of FSH. Although the experimental treatment of prepubertal monkeys with FSH alone for four weeks was able to induce cell proliferation in the seminiferous epithelium, only the combined administration of both hCG and FSH was capable of increasing testicular volume and the number of Sertoli cells and spermatogonia (Arslan et al., 1993; Schlatt et al., 1995). Thus it is evident that FSH alone is capable of inducing proliferation of Sertoli cells and spermatogonia in the prepubertal primate, but this does not result in qualitatively and quantitatively normal spermatogenesis unless testosterone is simultaneously present. On the other hand, exogenous testosterone alone given in very high concentrations is capable of inducing qualitatively complete spermatogenesis in immature cynomolgus monkeys (Marshall et al., 1984). However, despite prolonged treatment for almost a year, the very low sperm concentrations in the ejaculate showed that testosterone alone was not capable of inducing quantitatively normal spermatogenesis. To achieve this both LH/testosterone and FSH action appear to be required—as clinical experience in patients with secondary hypogonadism has clearly demonstrated (see above). The role of FSH in the maintenance of spermatogenesis in the primate has been studied by suppressing FSH action in adult males. One possible way to reach this goal is immunization against FSH or the FSH receptor. Several experiments have shown that long-term immunization against FSH induces severe testicular regression, oligozoospermia or even azoospermia in monkeys (Srinath et al., 1983; Moudgal et al., 1992). Even if the monkeys are only oligozoospermic and not azoospermic they remain infertile in mating tests, suggesting that the few sperm left are functionally impaired and that the severe oligozoospermia achieved by selective FSH suppression may be sufficient to induce infertility (Moudgal & Sairam, 1998). Consequently, in a pilot study conducted recently in India, human volunteers received an ovine FSH vaccine at a dose sufficient to achieve FSH deprivation throughout the period of four spermatogenic cycles. Due to the short duration of FSH deprivation, the experiment showed only marginal suppression of sperm counts (around 30–65%) but some parameters of sperm quality such as chromatin condensation and acrosome content were altered. Interestingly, spermatogenesis was reduced in the presence of normal levels of LH and testosterone. These data would suggest that the sole elimination of FSH action results not only in quantitative, but also qualitative damage of spermatogenesis (Moudgal et al., 1997). Long-term administration of hCG to normal volunteers induced suppression of spermatogenesis, via suppression of endogenous FSH (Matsumoto et al., 1986). This treatment resulted in oligozoospermia, which was reversed by simultaneous administration of FSH but not testosterone, demonstrating that FSH in connection with testosterone is fundamental for the maintenance of quantitatively normal spermatogenesis. In the monkey, the selective effect of FSH on the maintenance of spermatogenesis was studied by treating male adult cynomolgus monkeys with a GnRH antagonist which completely suppresses pituitary gonadotrophin production, and with simultaneous administration of pure FSH. It was shown that FSH could prevent and delay the reduction of testicular volume induced by the GnRH antagonist. Moreover, the combined administration of FSH and GnRH antagonist was able to maintain the number of type A (pale) spermatogonia and, in part, of round spermatids during the 8 weeks of treatment, while in animals treated with GnRH antagonist only, the number of these germ cells decreased much faster and more drastically (Weinbauer et al., 1991). These data show that FSH alone is capable of maintaining spermatogenesis in the primate, at least in part. This finding has very important implications for the pharmacological suppression of spermatogenesis for contraceptive purposes. In contraceptive trials in which either the long-acting testosterone ester testosterone buciclate (Behre et al., 1995) or transdermal testosterone together with oral levonorgestrel (Büchter et al., 1999) were administered to normal volunteers in order to suppress spermatogenesis, azoospermia was reached only in those subjects whose serum FSH levels were suppressed below the lower limit of detection, thus supporting the essential role of FSH for the maintenance of spermatogenesis. A particularly impressive example of the role of FSH in the maintenance of spermatogenesis in the human is the case of a hypophysectomized, fertile man with an activating mutation of the FSH receptor. The patient had been hypophysectomized because of a chromophobic adenoma and had normal fertility despite undetectable serum gonadotrophins while he was treated with testosterone enanthate. In this patient a heterozygous mutation exchanging codon 567 of the FSH receptor from Ala to Gly was discovered. This mutation was able to induce constitutive receptor activity in vitro, resulting in cAMP production in the absence of FSH stimulation (Gromoll et al., 1996). Serum inhibin B was at the upper limit of the normal range (Simoni et al., 1997), corroborating the conclusion that sperm production in this man was sustained by an FSH-like activity, such as that conferred by an activating mutation of the FSH receptor. Although the residual intratesticular testosterone might have contributed to the patient's fertility, low androgen levels such as those observed in this patient are usually not compatible with normal sperm production in hypophysectomized men. This case highlights the crucial importance of FSH for the maintenance of spermatogenesis and corroborates an earlier observation that a patient with an FSH-secreting pituitary adenoma had qualitatively normal spermatogenesis up to elongated spermatids despite low LH as well as testosterone and Leydig cell atrophy (Cunningham & Huckins, 1977). In summary, although FSH appears to play a more prominent role in the maintenance of primate spermatogenesis than in the initiation, qualitatively and quantitatively normal spermatogenesis is best maintained by the combined effects of FSH and LH/testosterone. As FSH is an essential therapeutic component for stimulation of spermatogenesis in hypogonadotrophic patients it would be of interest to investigate whether different FSH preparations vary in their potency to induce sperm production and paternity. Three preparations are available: the oldest contains FSH activity in the form of human menopausal gonadotrophin (hMG) extracted from urine. Thereafter, highly purified preparations of urinary FSH became available and most recently recombinant human FSH was introduced into clinical practice (for review see Simoni & Nieschlag, 1995). Unfortunately, no direct comparison of the three preparations in the treatment of male hypogonadism is available and a historical comparison is complicated by the fact that different therapeutic regimens were used. The available albeit scanty information reveals no convincing advantage of the highly purified or recombinant preparations over the original urinary preparations (Kliesch et al., 1995; Burgues et al., 1997; Büchter et al., 1998; unpublished observation). With the latter preparations appearance of sperm in the ejaculate can be induced in 95% of hypogonadotrophic patients and paternity is achieved in about 78% (Büchter et al., 1998). As FSH is essential for normal spermatogenesis and fertility, it has been speculated that abnormalities in FSH action could, in certain cases, cause idiopathic male infertility. However, aside from secondary hypogonadism, abnormalities in FSH action could be detected only in isolated cases, which fails to account for a substantial number of men with infertility so far classified as idiopathic. For example, despite an extensive search, abnormalities in the bioactivity of FSH could not be identified as a cause of male infertility (Jockenhövel et al., 1989). FSH receptor mutations have been found in women with primary and early secondary amenorrhoea (Aittomäki et al., 1995; Beau et al., 1998). In the course of a large screening attempt to find FSH receptor mutations in infertile male patients we found that the receptor is polymorphic at two sites in exon 10, corresponding to the extracellular domain close to the transmembrane domain and to the intracellular domain, respectively (Simoni et al., 1999), extending a previous observation in women (Liu et al., 1998). In particular we found two receptor isoforms carrying Ala307-Ser680 and Thr307-Asn680, respectively. These two polymorphic variants show similar affinity for the hormone in vitro in transiently transfected COS-7 cells. The two receptor variants are equally effective in mediating FSH action in terms of cAMP concentrations in response to FSH but the Ala-Ser variant does so in the presence of a receptor concentration three times lower than the Thr-Asn variant. This indicates that the Ala-Ser receptor variant has a higher activity for the hormone. This different activity in vitro, however, is not reflected in vivo by differences in parameters which might be FSH-dependent, such as serum inhibin B and FSH levels and testicular volume, both in proven fathers and in infertile men subdivided according to their FSH receptor genotype. Finally, the two receptor variants show a Mendelian distribution and frequency of the two different alleles similar in proven fathers and in infertile men. This suggests that the FSH receptor genotype is not a major determinant of fertility status in man. The above mentioned inactivating mutations of the FSH receptor remain a rare isolated cause of infertility restricted to five Finnish cases (Tapanainen et al., 1997), as screening for these mutations in other parts of the world was unrewarding (Jiang et al., 1998; Tuerlings et al., 1998; Simoni et al., 1999). Despite the scarcity of pathological findings in FSH action in male infertility, numerous attempts have been made to find a role for FSH in the treatment of male infertility. Because of the high success rate of hCG/hMG in inducing pregnancies in hypogonadotrophic men, these therapies were also applied to patients with normo- or hypergonadotrophic infertility. Numerous uncontrolled open studies tried to demonstrate their effectiveness and this kind of treatment was performed for more than two decades. However, a placebo-controlled, prospective, double-blind, randomized study of hCG/hMG treatment in normogonadotrophic men with oligoasthenoteratozoospermia could not demonstrate any beneficial effect of the verum on sperm parameters and pregnancy rates in comparison with placebo (Knuth et al., 1987). Although FSH was given in combination with hCG the study demonstrated that there was no role for FSH in the treatment of idiopathic male infertility. With the arrival of purified hMG and later, of recombinant human FSH, investigators again felt compelled to use these preparations empirically in the treatment of idiopathic male infertility. The first to use purified hMG were Acosta et al. (1992), who reported an increase in in vitro fertilization and pregnancy rates seen in wives of men treated with highly purified hMG who had failed to fertilize oocytes prior to treatment. In view of the fact that this was an uncontrolled study, the results have to be considered inconclusive. Although other investigators cautioned against premature conclusions from uncontrolled studies and requested controlled trials (Baker et al., 1992; Simoni & Nieschlag, 1995), the publication by Acosta et al. (1992) precipitated a host of investigations applying purified hMG in male infertility. We found several meeting abstracts describing such studies—all uncontrolled—and hardly any has been published since in a peer-reviewed journal. Studies which have appeared in journals are either similarly uncontrolled (Glander & Kratzsch, 1997) or, in addition, suffer from insufficient subject numbers (Merino et al., 1996; or from et al., et al., 1997). The only study of the for failed to placebo treatment to purified hMG et al., 1997). these not find a higher pregnancy rate in the We recently a double-blind, randomized study recombinant human FSH placebo for weeks in patients et al., 1998). It not any in parameters of semen or pregnancy However, in the verum we a increase in testicular volume as detected by and an chromatin condensation of sperm as by For the time there is no for FSH treatment for idiopathic male infertility. studies to are either in or the results not the effectiveness of FSH in sperm parameters or pregnancy testicular volumes, more sperm and in the appearance of some sperm to investigate the role of FSH in human spermatogenesis and to a concept for treatment of idiopathic male infertility that will the of Moreover, recombinant the production of particular FSH isoforms or in which could find a therapeutic The recent finding that not only FSH but also the FSH receptor is polymorphic (Simoni et al., 1999) the of FSH action, but it be considered in the of FSH and As a result of the between hypothalamus, pituitary and testis to a certain FSH the status of spermatogenesis. In the diagnosis of male infertility FSH serum determinations had in fact testicular (Behre et al., 1997). FSH levels in the presence of testes and azoospermia syndrome and low FSH levels to hypothalamic or pituitary testicular volume and azoospermia or severe oligozoospermia is simultaneously FSH indicates primary of spermatogenesis, while normal FSH in the presence of azoospermia indicates the of FSH is with the number of seminiferous germ cells et al., et al., The of serum FSH by in vitro even the most recombinant et al., does not the of the and the of mutations of the FSH β only the bioactivity but not the of the hormone is the only for this kind of With the of and the of single sperm extracted from testicular for the question whether in azoospermic patients sperm are present in the testicular or not became much more important than the of the Serum FSH not sufficient to this question et al., 1997; et al., 1998). inhibin B was demonstrated to spermatogenesis more than FSH et al., 1998; et al., 1999), but the of inhibin B also appears to be low to the presence or absence of low numbers of sperm in the testis et al., 1999). FSH and LH/testosterone each and alone are able to initiate and maintain spermatogenesis to a certain However, only to a certain be and this remains far below qualitatively normal spermatogenesis. For qualitatively and quantitatively normal spermatogenesis the of both gonadotrophins is that both gonadotrophins are controlled by the hypothalamic they are both responsible for spermatogenesis. their function is not combined into one hormone remains of the endocrine enigma the of male reproductive The for two hormones for the purpose may reproductive success and may the more with to It may also for of in the of the four major components of male reproduction mentioned at the of this Finally, it is possible that of action of both gonadotrophins is mediated by that The of will be in the to investigate the role of in testis The are by the The in the by is