The perifusion technique was used to investigate the action of diazepam (DZ), a benzodiazepine molecule known to compete for TRH receptor binding in rat pituitary, on TRH-induced TSH and GH release. The release kinetics for the two hormones from quartered pituitaries were measured in response to a 6-min pulse of TRH (10 nM), without or with DZ addition for a period of 30 min before and during the TRH pulse, plus an additional 15-min period. The dynamic patterns of TSH and GH release in response to TRH were characterized by a rapid increase in hormone release, declining slowly over the next 20 min. The rate of release represented 2.98 +/- 0.02 (+/- SE) and 1.75 +/- 0.06 times the corresponding basal level for TSH and GH, respectively, when evaluated over the first 15 min of the response to TRH. Addition of increasing doses of DZ suppressed the stimulatory effect of TRH in a dose-related manner, with an ID50 of 3 nM for both TSH and GH. The maximal effect of DZ was obtained with a concentration of 10 nM for both hormones. Ro 15-1788 (100 nM), a selective antagonist of the central type of benzodiazepine-binding sites (added to the perifusion system 30 min before DZ and then during the whole period of DZ perifusion), completely abolished (P less than 0.01) the inhibitory effect of DZ (10 nM) on the TRH-induced TSH and GH responses. When added alone before the TRH pulse, Ro 15-1788 had no effect on the TSH response to TRH. In contrast, PK 11,195 (100 nM), a selective antagonist of the nonneuronal benzodiazepine-binding sites, was unable to abolish the inhibitory action of DZ on TRH-stimulated TSH release. In addition, the effects of four other types of benzodiazepine (flurazepam, chlordiazepoxide, midazolam, and medazepam), all tested at a 10-nM concentration, corroborated these findings. Furthermore, DZ inhibition of the TSH response was nullified by picrotoxin (1 microM), but not by bicuculline (1 microM), two gamma-aminobutyric acid antagonists that had no effect, by themselves, on this response. For comparison, the effect of DZ (10 nM) was also tested on the release of GH in response to human GH-releasing factor-(1-44)-NH2 (10 nM) and was found to be ineffective.(ABSTRACT TRUNCATED AT 400 WORDS)