1. Previous studies in our laboratory have shown that the synthetic xanthine analogue denbufylline, a selective type 4 phosphodiesterase (PDE-4) inhibitor, is a potent activator of the hypothalamo-pituitary-adrenal (HPA) axis when given orally or intraperitoneally (i.p.) to adult male rats. This paper describes the results of experiments in which well established in vivo and in vitro methods were used to compare the effects of denbufylline on HPA function with those of two other selective PDE-4 inhibitors, rolipram and BRL 61063 (1,3-dicyclopropylmethyl-8-amino-xanthine). For comparison, parallel measurements of the immunoreactive- (ir-) luteinising hormone (LH) were made where appropriate. 2. When injected intraperitoneally, rolipram (40 and 200 micrograms kg-1, P < 0.005), denbufylline (0.07-0.6 microgram kg-1, P < 0.05) and BRL 61063 (30 micrograms kg-1, P < 0.005) each produced marked rises in the serum ir-corticosterone concentrations. However, lower doses of rolipram (1.6 and 8 micrograms kg-1) and BRL 61063 (0.25-6 micrograms kg-1) were without effect (P > 0.05). By contrast, intracerebroventricular (i.c.v.) injection of rolipram (8 ng-1 micrograms kg-1) or denbufylline (50 ng-1 microgram kg-1) failed to influence the serum ir-corticosterone concentration. BRL 61063 (8-120 ng kg-1, i.c.v.) was also ineffective in this regard although at a higher dose (1 microgram kg-1, i.c.v.) it produced a small but significant (P < 0.05) increase in ir-corticosterone release. Denbufylline also increased the serum ir-LH concentration when given peripherally (0.2-0.6 microgram kg-1, i.p., P < 0.05) or centrally (100 ng kg-1, i.c.v., P < 0.05) but rolipram (1.6-200 micrograms kg-1, i.p. or 8 ng-1 microgram kg-1, i.c.v.) and BRL 61063 (0.25-30 micrograms kg-1, i.p. or 1 ng-1 microgram kg-1, i.c.v.) did not (P > 0.05). 3. In vitro rolipram (10 microM, P < 0.01), denbufylline (1 mM, P < 0.001) and BRL 61063 (1 and 10 microM, P < 0.05) stimulated the release of corticotrophin releasing hormone (ir-CRH-41) but lower concentrations of the drugs were without effect as also was BRL 61063 at 100 microM (P > 0.05); the rank order of potency was thus BRL 61063 > rolipram > denbufylline. The adenylyl cyclase activator forskolin (100 microM, P < 0.01) also stimulated the release of ir-CRH-41, producing effects which were additive with those of rolipram and denbufylline but not with those of BRL 61063. The secretory responses to forskolin (100 microM) were accompanied by a highly significant increase in the cyclic AMP content of the hypothalamic tissue (P < 0.005). Rolipram (10 microM) also significantly (P < 0.05) elevated the hypothalamic cyclic AMP but denbufylline (10 mM) and BRL 61063 (10 microM) did not. However, all three PDE-inhibitors potentiated the rise in cyclic AMP induced by forskolin (P < 0.05). None of the drugs tested, alone or in combination, modified the release of arginine vasopressin (ir-AVP) from the hypothalamus. 4. Rolipram (100 microM), denbufylline (100 microM) and BRL 61063 (100 microM) stimulated the release of corticotrophin (ir-ACTH) from pituitary tissue in vitro (P < 0.05) but in lower concentrations they were without significant effect. In addition, rolipram (10 microM, P < 0.05), denbufylline (0.1 microM, P < 0.05) and BRL 61063 (10 microM, P < 0.05) potentiated the significant (P < 0.05) rises in ir-ACTH secretion induced by forskolin (100 microM). Forskolin (100 microM) also produced a highly significant increase (P < 0.01) in the tissue cyclic AMP content which was further potentiated by rolipram (10 microM), denbufylline (10 microM) and BRL 61063 (10 microM) which, alone did not affect the cyclic AMP content of the tissue. 5. Since both denbufylline and BRL 61063 possess significant adenosine A1 receptor blocking activity, further studies examined the potential influence of these receptors on the secretion in vitro of CRH-41, AVP and ACTH. The release of ir-CRH-41 was increased significantly by adenosine deaminase (ADA, 5microml-1, P<0.05) and the A1-receptor antagonist, 1,3-dicyclopropyl-8-cyclopentylxanthine (DPCPX, 0.1-10nM, P<0.05). The responses to ADA were abolished by the A1 receptor agonist N6-cyclo-hexyladenosine (CHA, 100nM, P<0.05) which alone had no significant effect on ir-CRH-41 release. ADA (0.1-10microml-1) and DPCPX (1nM) had weak stimulant and inhibitory effects, respectively, on the release of ir-ACTH from the pituitary gland while CHA (0.1-10nM) was without effect. Ligand binding studies with [3H]-DPCPX as a probe demonstrated the presence of specific high affinity A1 binding sites in the hypothalamus (Kd=0.7nM; Bmax=367+/-32fmolmg-1 protein) and in the hippocampus (Kd=1nM; Bmax=1165 +/-145fmolmg-1 protein). In both tissues binding of the ligand was displaced by CHA (IC50=1nM (hypothalamus) and 2nM (hippocampus)), BRL 61063 (IC50=80nM (hypothalamus) and 100nM (hippocampus)) and denbufylline (IC50=5microM (hypothalamus) and 9microM(hippocampus)) but not by rolipram. 6.The results suggest that rolipram, denblufylline and BRL 61063 stimulate the HPA axis in the rat, acting at the levels of both the hypothalamus and the pituitary gland. Their actions may be explained, at least in part, by inhibition of PDE-4 but additional actions including blockade of hypothalamic adenosine A1 receptors by denbufylline and BRL 61063 cannot be excluded.