Continuous administration of the antiprogesterone RU486 inhibits ovulation in women and in monkeys; in this regard RU486 may act as a progestin agonist rather than as an antagonist. We compared the site(s) and mechanism(s) of RU486-induced ovulation inhibition with those of levonorgestrel (LNG). Six regularly menstruating cynomolgus monkeys each received placebo, RU486 (1 mg/kg/d) or LNG (2 g/kg/d) i.m. between days (cd) 2-22 of three separate menstrual cycles. Serum levels of estradiol (E2), progesterone (P4), androstenedione, LH and FSH were analyzed by RIAs in daily blood samples. Basal and GnRH-stimulated (1 and 50 g of GnRH i.v. 2 h apart) secretion of LH and FSH was assessed using serial blood samples collected for 12 h on cd 10. Mean cycle length was prolonged by RU486 and LNG treatments from 32 d to 70 d and 52 d, respectively (p < 0.02). Ovulation was inhibited in five of the six primates during RU486, and in all six during LNG treatment. During RU486 treatment, serum E2 levels were similar to those of the control cycle; despite peaks of E2 secretion, no LH peaks were seen. In contrast, E2 concentrations were profoundly suppressed during LNG treatment (p < 0.005). The reduction in serum E2 was accompanied by lower levels of androstenedione, and suppressed ratio of E2/androstenedione (p < 0.02) suggesting both reduced synthesis and aromatization of androgen precursors during administration of LNG. Consequently, LNG treatment was associated with higher levels of serum FSH and LH (p < 0.001; 1-way ANOVA). Similarly, as during the luteal phase of the menstrual cycle, the amplitude of basal LH-pulses was increased during LNG treatment (p < 0.05), whereas RU486 treatment did not affect basal LH secretion. The GnRH-stimulated release of LH was similar during the placebo, RU486 and LNG cycles; enhanced release of FSH was seen during administration of LNG. Thus, in the present model system, RU486 seems to inhibit ovulation mainly at the level of hypothalamus, possibly by interfering with the steroidal positive feedback signals from the ovary. However, LNG inhibits ovulation differently, most likely via direct progesterone-like effects on folliculogenesis and the hypothalamus. The pituitary does not appear to be the major site of action(s) of RU486 or LNG. Thus, the differential mechanisms of ovulation inhibition by RU486 and LNG seem to result from lesser intraovarian impact of RU486 as well as dissimilar influences on tonic gonadotropin secretory levels. We conclude that when inhibiting ovulation, RU486 does not act as a progestin agonist, but rather, functions through a hypothalamic mechanism(s), which might be unique to RU486 as a progesterone antagonist.