Aromatase Inhibition in the Human Male Reveals a Hypothalamic Site of Estrogen Feedback*

The preponderance of evidence states that, in adult men, estradiol (E2) inhibits LH secretion by decreasing pulse amplitude and responsiveness to GnRH consistent with a pituitary site of action. However, this conclusion is based on studies that employed pharmacologic doses of sex steroids, used nonselective aromatase inhibitors, and/or were performed in normal (NL) men, a model in which endogenous counterregulatory adaptations to physiologic perturbations confound interpretation of the results. In addition, studies in which estrogen antagonists were administered to NL men demonstrated an increase in LH pulse frequency, suggesting a potential additional hypothalamic site of E2 feedback. To reconcile these conflicting data, we used a selective aromatase inhibitor, anastrozole, to examine the impact of E2 suppression on the hypothalamic-pituitary axis in the male. Parallel studies of NL men and men with idiopathic hypogonadotropic hypogonadism (IHH), whose pituitary-gonadal axis had been normalized with long-term GnRH therapy, were performed to permit precise localization of the site of E2 feedback. In this so-called tandem model, a hypothalamic site of action of sex steroids can thus be inferred whenever there is a difference in the gonadotropin responses of NL and IHH men to alterations in their sex steroid milieu. A selective GnRH antagonist was also used to provide a semiquantitative estimate of endogenous GnRH secretion before and after E2 suppression. Fourteen NL men and seven IHH men were studied. In Exp 1, nine NL and seven IHH men received anastrozole (10 mg/day po x 7 days). Blood samples were drawn daily between 0800 and 1000 h in the NL men and immediately before a GnRH bolus dose in the IHH men. In Exp 2, blood was drawn (every 10 min x 12 h) from nine NL men at baseline and on day 7 of anastrozole. In a subset of five NL men, 5 microg/kg of the Nal-Glu GnRH antagonist was administered on completion of frequent blood sampling, then sampling continued every 20 min for a further 8 h. Anastrozole suppressed E2 equivalently in the NL (136 +/- 10 to 52 +/-2 pmol/L, P < 0.005) and IHH men (118 +/- 23 to 60 +/- 5 pmol/L, P < 0.005). Testosterone levels rose significantly (P < 0.005), with a mean increase of 53 +/- 6% in NL vs. 56 +/- 7% in IHH men. Despite these similar changes in sex steroids, the increase in gonadotropins was greater in NL than in IHH men (100 +/- 9 vs. 58 +/- 6% for LH, P = 0.07; and 85 +/- 6 vs. 41 +/- 4% for FSH, P < 0.002). Frequent sampling studies in the NL men demonstrated that this rise in mean LH levels, after aromatase blockade, reflected an increase in both LH pulse frequency (10.2 +/- 0.9 to 14.0 +/- 1.0 pulses/24 h, P < 0.05) and pulse amplitude (5.7 +/- 0.7 to 8.4 +/- 0.7 IU/L, P < 0.001). Percent LH inhibition after acute GnRH receptor blockade was similar at baseline and after E2 suppression (69.2 +/- 2.4 vs. 70 +/- 1.9%), suggesting that there was no change in the quantity of endogenous GnRH secreted. From these data, we conclude that in the human male, estrogen has dual sites of negative feedback, acting at the hypothalamus to decrease GnRH pulse frequency and at the pituitary to decrease responsiveness to GnRH.