Copyright © 2001 by The Endocrine Society

Abbreviations:

CV,

Coefficient of variation;

FAS,

free α-subunit;;

Inh B,

inhibin B;;

IPI,

interpulse interval;;

NL,

normal.

IN THE UNITED States, 10–15% of couples are infertile, with a male factor implicated in up to 50% of cases ^[1] . To date, the etiology of most cases of male infertility remains poorly understood, with inconsistent elevation of FSH levels. Indeed, the role of FSH in both the initiation and the maintenance of spermatogenesis is controversial and exhibits important species differences ^{[2] [3] [4] [5] [6]} . In male mice rendered FSH deficient due to targeted disruption of the FSH β-subunit ^[2] or FSH receptor gene ^[3] , fertility is preserved despite reduced testes size and partial spermatogenic failure. In the human there is discordance between the phenotype of men with mutations in the FSH receptor gene who are oligospermic ^[4] and those with mutations of the FSH β-subunit who are azoospermic ^{[5] [6]} . This discrepancy in the degree of spermatogenic failure suggests that ligand deficiency may well have a more severe expression than that of the FSH receptor and raises question about the completeness of the receptor knockout models. Given these human data indicating that FSH is required for quantitatively normal spermatogenesis, delineating the feedback regulation of FSH is key to defining the pathophysiology of male infertility as well as determining the feasibility of hormonal approaches to male contraception.

It has long been known that regulation of gonadotropin secretion in the human male involves a complex interplay between stimulation by GnRH secretion and inhibitory feedback by sex steroids comprising T and E2 from Leydig cells. However, in the last decade it has also been increasingly appreciated that for FSH, there is an additional level of complexity mediated by nonsteroidal factors. This nonsteroidal regulation of FSH comprises an endocrine negative feedback loop mediated by inhibin B (Inh B) secretion from Sertoli cells ^{[7] [8] [9] [10] [11] [12] [13]} in addition to autocrine/paracrine modulation within the pituitary mediated by activin and follistatin ^[14] . Thus, an integrated approach to the study of FSH regulation requires utilization of models that permit isolation of the effects of sex steroids from those of nonsteroidal factors.

To date, most clinical investigation in the human has addressed only the sex steroid component of gonadal feedback ^{[15] [16] [17] [18] [19] [20] [21] [22] [23] [24]} . However, the nonsteroidal component of gonadal feedback is particularly important given that most cases of male infertility represent a defect in the seminiferous tubule compartment of the testis in the absence of any obvious perturbation in Leydig cell function. It has only recently been possible to quantitate the biologically active nonsteroidal gonadal regulators of FSH reliably with the development of sensitive and specific assays for Inh B ^[25] . To date, inhibin has not been available in sufficient purity or quantity for human administration; thus, for the present, studies on inhibin

5542

Thus, in the present experiments we set out to elucidate the Inh B contribution to FSH regulation by comparing the FSH levels of normal men after acute sex steroid withdrawal with those in age-matched functional castrates who lack both sex steroids and Inh B. We hypothesized that suppression of sex steroid negative feedback to castrate levels would fail to recreate castrate levels of FSH in an Inh B-replete model, implying that Inh B is the major regulator of FSH secretion in the human male. We also hypothesized that Inh B would have no effect on GnRH pulse frequency, such that despite markedly different Inh B levels, similar pulse frequencies would be observed in the normal men after sex steroid withdrawal and in the functional castrates.

Subjects and Methods

Normal (NL) men

Ten NL men (age range, 25–45 yr) participated in the study. All NL study subjects met the following criteria: 1) normal pubertal development, sexual function, and general health; 2) normal physical examination including a testicular volume 20 ml or more; 3) normal serum levels of T, E2, LH, FSH, TSH, and PRL; and 4) normal semen analysis according to WHO criteria ^[29] . Normal subjects represent the integrated model of both sex steroid and nonsteroidal components of gonadotropin negative feedback. Suppression of sex steroid synthesis in these subjects thus reveals the nonsteroidal input to FSH control.

Men with castrate T and Inh B levels

Three age-matched men (age range, 23–47 yr) with castrate levels of T and Inh B were also studied. Two patients were agonadal and thus true castrates; the first after a severe burn injury to the groin at the age of 10 yr, and the second after congenital unilateral torsion and contralateral reflex sympathetic orchiopathy. The third patient, who had Klinefelter syndrome, was not a true castrate, in that his testes were still present. However, we felt that it was appropriate to classify him as a functional castrate given a testes size of only 1 ml in association with a T level of 59 ng/dl (2.1 nmol/liter) and an undetectable Inh B level. All three patients had been off T replacement for at least 8 wk before participating in the study. Castrate men represent a model characterized by the complete absence of both gonadal sex steroids and nonsteroidal factors. Accordingly, they define the FSH levels that occur when the hypothalamic-pituitary axis is unrestrained by either component of gonadal feedback.

The study was approved by the human research committee at the Massachusetts General Hospital, and all subjects provided written informed consent.

Study protocol

Subjects were admitted to the General Clinical Research Center of Massachusetts General Hospital at 1700 h and had an iv cannula inserted in a forearm vein. Commencing at 1800 h, blood sampling was performed every 10 min for 12 h, after which subjects were discharged home. On completion of the baseline frequent sampling study, an interval of at least 14 d was allowed to elapse before commencing sex steroid suppression due to blood volume considerations. Acute sex steroid withdrawal was induced in 10 NL men using high dose ketoconazole, an antifungal imidazole that blocks steroidogenesis predominantly by inhibiting C_17–20 lyase ^{[30] [31]} . The ketoconazole regimen employed comprised a loading dose of 1 g at midnight, followed by a maintenance dose of 400 mg, four times daily, as previously described ^[24] . Because this dose of ketoconazole may impair cortisol biosynthesis, glucocorticoid replacement was provided in the form of dexamethasone (0.5 mg twice daily). We ^[19] and others ^[32] have demonstrated previously that this dose of dexamethasone does not suppress gonadotropin secretion in men. In view of the potential hepatotoxicity of ketoconazole ^[31] , liver function tests were monitored daily. Participants were withdrawn from the study if hepatic transaminases increased to greater than 3 times the upper limit of normal. The same 12-h frequent blood-sampling paradigm performed at baseline was repeated during h 66–78 (d 3) and 138–150 (d 6) of sex steroid ablation to define the extent and time course of changes in gonadotropin secretion during acute sex steroid ablation.

Exp 2: neuroendocrine characterization of men with castrate T and Inh B levels

In the three patients with castrate levels of T and Inh B, a more intensive sampling paradigm (every 5 min for 24 h) was employed given the importance of sampling interval and duration on analysis of pulsatile hormone secretion ^{[33] [34]} and the absence of blood volume considerations, as frequent sampling was performed on just one occasion.

In both Exp 1 and Exp 2, all samples were assayed for LH and free α-subunit (FAS), whereas FSH was measured in hourly samples. T and E2 levels were determined at baseline and at 6-h intervals during the study. Inh B was measured in a pool comprising equal aliquots of each sample obtained during the study. Mean LH and FSH levels were calculated for each frequent blood-sampling study. Pulsatile LH and FAS secretion were analyzed using a modification of the Santen and Bardin method ^[35] as previously validated by our group ^[36] .

Hormone assays

Serum LH and FSH concentrations were determined by microparticle enzyme immunoassay using the automated Abbott AxSYM system (Abbot Laboratories, Inc., Chicago, IL). The Second International Reference Preparation was used as the reference standard. The assay sensitivity for both LH and FSH was 1.6 mIU/ml. The intraassay coefficients of variation (CVs) for LH and FSH were less than 7% and less than 6%, respectively, with interassay CVs for both hormones of less than 7.4%. The reference range for gonadotropins in this assay was based on the 95% confidence limits of 36 NL men, aged 18–45 yr, and was 4.9–14.9 IU/liter for LH and 1.3–10.1 IU/liter for FSH. Serum FAS concentrations were determined by a monoclonal antibody RIA using highly purified α-subunit of hCG as the assay calibrator ^[37] . Serum T concentrations were measured using the Coat-A-Count RIA kit (Diagnostic Products, Los Angeles, CA), which has intra- and interassay CVs of less than 10% and a detection limit of 4 ng/dl. E2 was measured by the Abbott AxSYM system, which has an analytical sensitivity of 10 pg/ml, an intraassay CV of less than 6.4%, and an interassay CV of less than 10.6%. Inh B was measured using a commercially available, double antibody ELISA (Serotec, Oxford, UK) as previously described ^[25] . In our use, the clinical detection limit of this assay is 50 pg/ml, with a CV of 4–6% within plate and 15–18% between plates.

Statistical methods

In the NL men, mean hormone levels at baseline were compared with those during d 3 and 6 of sex steroid suppression using ANOVA for repeated measures followed by post-hoc Newman-Keuls testing for individual differences. The gonadotropin responses (mean LH and FSH levels, mean pulse frequency, and amplitude of both LH and FAS) of the NL men on d 6 of ketoconazole administration were compared with those in the castrate subjects using an unpaired t test. To correct for the different sampling paradigms employed, only data from samples drawn every 10 min for 12 h were included when comparing gonadotropin secretion in the functional castrates with that in the NL men after sex steroid ablation. P < 0.05 was taken to be statistically significant.

Results

Exp 1: neuroendocrine characterization of NL men before and after acute sex steroid suppression

One individual was withdrawn from the study on d 2 of ketoconazole administration because of a greater than 3-fold

5543

Exp 2: neuroendocrine characterization of men with castrate T and Inh B levels

The castrate subjects had a mean T level of 66 ± 8 ng/dl (2 ± 0.4 nmol/liter), an E2 level of 20 ± 1 pg/ml (73 ± 4pmol/liter),

Figure 1. Hormonal profile of 9 normal men at baseline and in response to gonadal steroid withdrawal induced by the administration of ketoconazole for 150 h. , Mean ± SEM data for the group; ■, individual data. Asterisks represent a significant change from mean baseline values: *, P < 0.05; **, P < 0.0005. The shaded area represents the 95% confidence limits for 36 normal men, aged 18–45 yr. To convert the values for T to nanomoles per liter, multiply by 0.03467. To convert the values for E2 to picomoles per liter, multiply by 3.671.

Discussion

Using an in vivo human model of short-term gonadal steroid withdrawal, this study demonstrates that selective ablation of sex steroids results in only a modest increase in FSH levels, which remain within the normal adult male range, implying that a nonsteroidal factor, presumably Inh B, is likely to be the major regulator of FSH secretion in the human male.

Most of the literature on FSH regulation in the human has focused on the sex steroid component of gonadal feedback.

5544

Figure 2. Comparison of the gonadotropin concentrations in normal men at baseline () and in response to sex steroid withdrawal () with those in functional castrates (■). Asterisks represent a significant difference from the baseline values of normal men (*, P < 0.05; **, P < 0.0005). The dashed line represents the upper 95th confidence limit for 36 normal men, aged 18–45 yr.

The concept of dual endocrine activity of the testis was first proposed more than 65 yr ago with the demonstration that administration of an aqueous testicular extract to gonadectomized male rats could suppress the formation of castration cells in the anterior pituitary, suggesting a physiological role for a nonsteroidal gonadal secretory product ^[40] . Over the last several decades, efforts to explore inhibin physiology in the human have been hampered by numerous methodological difficulties, including the lack of inhibin for human administration ^[13] . For this reason studies on inhibin physiology in the human have, of necessity, been indirect and inferential. There is now an abundance of literature attesting to the importance of Inh B as an important endocrine feedback regulator of FSH secretion in the human male ^{[8] [9] [10] [11] [41] [42]} . However, the relative importance of Inh B vs. sex steroids to FSH control in the human has not been established.

In the present study short-term, reversible sex steroid withdrawal with ketoconazole was employed to isolate the sex steroid and nonsteroidal components of FSH feedback. Clearly, the duration of castration is an important limitation of this study, in that the impact of acute sex steroid ablation is being compared with that of longer term castration. Although there are few data on the time course of the gonadotropin rise after gonadectomy in the human male, studies in the monkey suggest that gonadotropins reach stable castrate levels approximately 20 d after orchidectomy ^[43] . Although our sequential frequent sampling studies demonstrated that gonadotropin levels had actually reached a plateau during the 1-wk period of ketoconazole administration, it is possible that a further secondary rise in gonadotropins would have been observed had the duration of castration been extended. However, in a study of nine men with prostate cancer treated with a similar ketoconazole regimen for 12 months, no significant change was observed in FSH levels, whereas LH levels continued to rise until approximately 60 d ^[44] . Although neither LH nor FSH reached castrate levels in the short window of sex steroid ablation employed in this study, a differential gonadotropin response was already clearly apparent, such that LH had increased to approximately 40% of castrate levels, whereas FSH remained at only approximately 15% of castrate levels. The fact that removal of sex steroid negative feedback caused LH levels to increase well above the normal adult male range compared with a modest rise in FSH that did not exceed normal levels thus points to an important role of Inh B in FSH regulation. The modest increase in FSH is all the more significant given the 25% decrease in Inh B levels observed in this study, which is also likely to have contributed to the FSH rise given the known correlation between Inh B and FSH in normal men ^{[8] [45] [46]} .

The impact of inhibin administration ^{[47] [48] [49] [50]} and immunoneutralization ^{[51] [52] [53]} on FSH levels has been examined in a variety of animal models. However, one must be cautious about extrapolating data from nonprimates to the human given the species differences that have been observed in FSH regulation. Thus, although administration of inhibin antiserum results in an increase in FSH levels in the adult male rhesus monkey ^[52] , no effect on FSH is observed in the adult male rat ^[53] . Support for the hypothesis that Inh B is likely to be the most important regulator of FSH secretion in the human male is provided by a series of elegant studies in the monkey demonstrating that only the institution of combined treatment with T and gonadal peptides derived from charcoal-extracted porcine follicular fluid prevents the postcastration rise in FSH ^[48] . Selective withdrawal of follicular fluid treatment then results in a dramatic rise in serum FSH levels, which returns to normal once therapy with gonadal peptides is restored. Subsequent studies performed after inhibin had

5545

It is interesting that ablation of sex steroids caused an approximately 25% reduction in Inh B levels despite a 2-fold increase in FSH levels as we have previously reported ^[24] . These data are in keeping with the hypothesis that the negative feedback effect of Inh B on FSH secretion is more robust than the positive stimulating effect of FSH on Inh B secretion ^[55] . The mechanism for the decline in Inh B after biochemical castration is unclear. Potential explanations are that lack of T or E2 either directly affects Sertoli cell production of Inh B or, alternatively, interferes with spermatogenesis and thus the paracrine regulation of inhibin by germ cells ^{[56] [57] [58]} . It is possible, therefore, that transient disruption of spermatogenesis at the spermatid stage resulting from castrate levels of T during the period of ketoconazole administration led to the decrease in Inh B levels observed in this study. It also possible that ketoconazole may have had a direct suppressive effect on Inh B secretion independent of its effect to lower T levels. However, given the rapid fall in T after administration of the loading dose of ketoconazole, it is impossible to distinguish between these two effects.

Concerning the GnRH pulse frequency estimations in these studies, two important conclusions can be made. First, when a comparable sampling paradigm is used, a similar GnRH pulse frequency is observed in normal men after sex steroid withdrawal and in functional castrates, indicating that GnRH pulse frequency is regulated by sex steroids rather than by Inh B. Second, using FAS as a marker of GnRH pulse frequency, pulses of GnRH occur approximately every 35 min in castrate men, indicating that the true castrate frequency is significantly faster than what we and others proposed based on analysis of LH pulsatility ^{[18] [59]} . In future experiments it will be interesting to examine the impact on FSH secretion of increasing GnRH pulse frequency to castrate levels using the model of GnRH-deficient men in whom sex steroid secretion has been blocked with ketoconazole.

Using the acute model of ketoconazole-induced gonadal steroid withdrawal, we demonstrated that suppression of sex steroids in normal men results in increases in LH, but not FSH, to levels above the normal range. From these experiments we infer that a nonsteroidal gonadal factor, most likely Inh B, is a major regulator of FSH in the human male. Confirmation of this hypothesis awaits the availability of inhibin for human administration.

Acknowledgments

We gratefully acknowledge the nurses of the General Clinical Research Center for their excellent clinical care, and the technicians of the Reproductive Endocrine Sciences Center RIA Core for their superb technical contributions to this study.

References

5546