Differential regulation of gonadotropin secretion by testosterone in the human male: Absence of a negative feedback effect of testosterone on follicle-stimulating hormone secretion

Studies of sex steroid regulation of gonadotropin secretion in the human ma le have focused primarily on the respective site(s) of negative feedback of testosterone (T) and estradiol (E-2). The use of pharmacological doses of sex steroids in these studies has precluded conclusions about the relative roles of T and E-2 in gonadotropin feedback. Thus, the aims of the present study were to 1) determine the relative contributions of T vs. E-2 to the s ex steroid component of gonadotropin regulation, and 2) distinguish the fee dback effects of T that that are direct (i.e. mediated by the androgen rece ptor) vs. indirect (mediated by aromatization to E-2). Two experimental interventions were used: 1) inhibition of aromatization by a selective aromatase inhibitor to examine the impact of selective E-2 wit hdrawal; and 2) acute medical castration to examine the effect of ablating both T and E-2. Sixteen normal (NL) men (mean age, 30.5 +/- 2.2 yr) were st udied. Nine NL subjects were treated with the aromatase inhibitor, anastroz ole (10 mg, orally, daily, for 5 days). Twelve NL men underwent medical cas tration with ketoconazole (1-g loading dose followed by 400 mg, orally, fou r times a day for 5 days). Ketoconazole-treated subjects received concomita nt treatment with dexamethasone (0.5 mg twice daily) to prevent the develop ment of adrenal insufficiency. Single blood samples were drawn daily betwee n 0800-1000 h. To ensure that dexamethasone was not altering the gonadotrop in response to sex steroid ablation by a direct pituitary effect, live GnRH -deficient men (mean age, 37.6 +/- 3.9 yr) underwent GnRH dose-response stu dies at baseline and after treatment with dexamethasone (0.5 mg twice daily ). Aromatase blockade caused significant lowering of E-2 (33 +/- 3 to 14 +/- 1 pg/mL; P < 0.0005) with a corresponding increase in T levels (563 +/- 42 t o 817 +/- 81 ng/dL; P < 0.05). Treatment with ketoconazole resulted in equi valent suppression of E-2 (41 +/- 4 to 14 +/- 1 pg/mL; P < 0.0005), but als o induced castrate levels of T (491 +/- 28 to 40 +/- 3 ng/dL; P < 0.0005). Both treatment regimens were associated with a significant increase in gona dotropin levels. For LH, the percent increase in serum levels after castrat ion was almost 3-fold greater than that seen after selective E-2 withdrawal (275 +/- 23% with ketoconazole us. 95.6 +/- 21% with anastrozole; P < 0.00 5). Despite the divergent changes in T levels with these two maneuvers (a m arked decrease after ketoconazole and a significant increase with anastrozo le), the percent rise in FSH levels was similar in the two protocols (91 +/ - 6% vs. 71 +/- 7%, respectively; P = NS). Inhibin B levels were unchanged after selective E-2 withdrawal (156 +/- 23 vs. 176 +/- 19 pg/mL), but decre ased slightly with ketoconazole (156 +/- 15 to 131 +/- 11 pg/mL; P < 0.05). In contrast to the effects of glucocorticoid administration on gonadotropi n secretion in women, no significant changes were observed in the GnRH-defi cient men treated with dexamethasone in terms of mean LH levels (19.8 +/- 3 .2 vs. 23.3 +/- 5.4 IU/L), mean LH pulse amplitude after GnRH (16.0 +/- 2.5 vs. 19.0 +/- 5.1 IU/L), or mean FSH levels (8.0 +/- 1.9 vs. 9.2 +/- 2.4 IU /L, pre vs. post). These studies provide evidence of differential regulation of gonadotropin s ecretion by T in the human male. T exerts both direct and indirect feedback on LH secretion, whereas its effects on FSH appear to be mediated largely by aromatization to E-2. From these data we conclude that in terms of sex s teroid feedback, E-2 is the predominant regulator of FSH secretion in the h uman male.