Sex steroid contraceptive regimes result in incomplete suppression of spermatogenesis in 30-45% of Caucasian men. The basis for this is unclear, but differences in the activity of 5 alpha-reductase (5 alpha R) have been demonstrated. Two isoforms of 5 alpha R have been described: 5 alpha R1 is found in skin, whereas the predominant from in reproductive tissues is 5 alpha R2. To investigate possible contributions of these isoenzymes, we have investigated androgen-dependent changes in seminal plasma androgens (5 alpha R2) and sebum production (5 alpha 1) during administration of a supraphysiological dose (200 mg IM weekly) of testosterone enanthate (TE) to 33 normal men. Eighteen men rapidly (< 20 weeks treatment) became azoospermic, the remainder having a mean sperm density of 2.0 +/- 0.6 x 10(6) at that time. The concentrations of testosterone and 3 alpha, 17 beta-androstanediol glucuronide (AdiolG) were lower in seminal plasma than in blood but rose by a similar degree (100%) after 16 weeks TE treatment in both groups. There were no differences in seminal-plasma concentration of testosterone or AdiolG between azoospermic and oligozoospermic responders, either pretreatment or after 16 weeks TE treatment. Although the concentrations of dihydrotestosterone (DHT) were similar in seminal plasma and blood pre- and posttreatment, there was a selective increase in seminal plasma DHT concentration in the oligozoospermic responders from 2.12 +/- 0.29 to 2.94 +/- 0.33 nmol/L (P < 0.05), while there was no significant change in the azoospermic responders (2.18 +/- 0.31-2.54 +/- 0.27 nmol/L) after 16 weeks of TE treatment. Dihydrotestosterone in seminal plasma is primarily derived from 5 alpha R activity in the epididymis. The concentration of prostaglandin E2 (PGE2) in seminal plasma was unchanged during TE treatment. Sebum excretion was increased during TE treatment, but there were no difference between azoospermic and oligozoospermic responders pretreatment or after 16 weeks TE treatment. These results are consistent with the hypothesis that incomplete suppression of spermatogenesis during TE treatment is associated with a relatively higher 5 alpha R activity in the reproductive tract (epididymis and/or testis) during TE treatment. As the predominant form of 5 alpha R in the reproductive tract is 5 alpha R2 (type 2), we conclude that the increase in activity derives from this form of the enzyme, rather than the type 1 form (5 alpha R1) predominantly found in nongenital skin.