T he specific roles of luteinizing hormone (LH) and follicle-stimulating hormone (FSH) in controlling human spermatogenesis are poorly understood. We studied the effect of an experimentally induced, selective LH deficiency on sperm production in normal men. After a 3-mo control period, five men received 200 mg testosterone enanthate (T) i.m./wk to suppress LH, FSH, and sperm counts. Then, while continuing T at the same dosage, human FSH (hFSH) was administered simultaneously to replace FSH activity, leaving LH activity suppressed. Four men received 100 IU hFSH s.c. daily plus T (high dosage hFSH) for 13-14 wk, while one man received 50 IU hFSH s.c. daily plus T (low dosage hFSH) for 5 mo. The effect on sperm production of the selective LH deficiency produced by hFSH plus T administration was assessed.In the four men who received the high dosage hFSH regimen, sperm counts were markedly suppressed during T administration alone (0.3±0.2 million/cm³, mean±SE, compared with 94±12 million/cm³ during the control period). Serum LH bioactivity (determined by in vitro mouse Leydig cell assay) was suppressd (140±7 ng/ml compared with 375±65 ng/ml during control period) and FSH levels (by radioimmunoassay) were reduced to undetectable levels (<25 ng/ml, compared with 98±21 ng/ml during control period) during T alone. With the addition of 100 IU hFSH s.c. daily to T, sperm counts increased significantly in all subjects (33±7 million/cm³, P < 0.02 compared with T alone). However, no subject consistently achieved sperm counts within his control range. Sperm morphology and motility were normal in all four men and in vitro sperm penetration of hamster ova was normal in the two men tested during the hFSH-plus-T period. During high-dosage hFSH administration, serum FSH levels increased to 273±44 ng/ml (just above the normal range for FSH, 30-230 ng/ml). Serum LH bioactivity was not significantly changed compared with the T-alone period (147±9 ng/ml). After the hFSH-plus-T period, all four men continued to receive T alone after hFSH was stopped. Sperm counts were again severely suppressed (0.2±0.1 million/cm³), demonstrating the dependence of sperm production on hFSH administration.Serum T and estradiol (E₂) levels increased two- to threefold during T administration alone compared with the control period. Both T and E₂ levels remained unchanged with the addition of hFSH to T, confirming the lack of significant LH activity in the hFSH preparation.In the one man who received low dosage hFSH treatment, sperm counts were reduced to severely oligospermic levels, serum FSH was suppressed to undetectable levels, and serum LH bioactivity was markedly lowered during the T-alone period. With the addition of 50 IU hFSH s.c. daily to T, sperm counts increased, to a mean of 11±3 million/cm³. During this period, serum FSH levels increased to a mean of 105±11 ng/ml (slightly above this man's control range and within the normal adult range), while LH bioactivity remain suppressed. After hFSH was stopped and T alone was continued, sperm counts were again severely reduced to azoospermic levels.We conclude that FSH alone is sufficient to reinitiate sperm production in man during gonadotropin suppression induced by exogenous T administration. FSH may stimulate sperm production in this setting by increasing intratesticular T through androgenbinding protein production or by increasing the sensitivity of the spermatogenic response to the intratesticular T present during exogenous T administration.