Apparent pA2 values of benzodiazepine antagonists and partial agonists in monkeys.

Drugs that bind to benzodiazepine recognition sites of gamma-aminobutyric acid type A receptor complexes may function as agonists in some behavioral assays and as antagonists in other behavioral assays. The present studies compared the effects of the benzodiazepines midazolam, flumazenil, bretazenil, Ro 41-7812, and Ro 42-8773 and the beta-carboline, beta-carboline-3-carboxylate-t-butyl ester (beta-CCt) under two different types of schedule-controlled responding in squirrel monkeys. One group of monkeys responded under a fixed-ratio schedule of stimulus-shock termination, and a second group of monkeys responded under a multiple fixed-ratio schedule of food presentation involving suppressed and nonsuppressed behavior. Under the schedule of stimulus-shock termination, midazolam produced dose-related decreases in response rate, and these effects were surmountably antagonized by flumazenil, bretazenil, Ro 41-7812, Ro 42-8773, and beta-CCt. Schild plot analysis of these data revealed the following mean pA(2) values: flumazenil, 7.18; bretazenil, 7.62; Ro 41-7812, 7. 06; Ro 42-8773, 6.95. Apparent pA(2) values were not calculated for beta-CCt because the CL of the slope of the Schild plot included positive values. Under the multiple schedule, midazolam, bretazenil, and Ro 42-8773 dose-dependently increased rates of suppressed responding, whereas flumazenil, Ro 41-7812, and beta-CCt had no significant rate-altering effects. Flumazenil antagonized the antisuppressant effects of midazolam and bretazenil; however, individual variability in these effects prohibited the determination of apparent pA(2) values. These results indicate that in vivo pA(2) values may be determined for benzodiazepine-site ligands. These results further demonstrate that some benzodiazepine-site ligands, e. g., bretazenil and Ro 42-8773, may function as both agonists and as competitive antagonists in vivo.