The pulsatile but not the continuous application of parathyroid hormone (PTH) increase bone mass in vivo. To study the effects of intermittent hormonal administration on bone-derived cells in vitro, we established a perifusion system using the human osteosarcoma cell line SaOS-2. Cells were grown in suspension culture attached to collagen beads and were then loaded into a 3 ml syringe for perifusion experiments. The application of PTH(1-34) resulted in a dose-dependent increase of cAMP release by SaOS-2 cells into the effluent medium. Cyclic AMP accumulation was rapidly desensitized by approx. 80% after 30 min of continuous exposure to PTH(1-34) (10(-7) M), while cells remained responsive to forskolin. The recovery of PTH responsiveness required at least 2 h of hormone-free perifusion. Desensitization in the experimental setting was dose-dependent (EC50 = 1 x 10(-10) M PTH(1-34)). Neither 8Br-cAMP (2 x 10(-4) M) nor PMA(1 x 10(-7) M) had an effect on the PTH(1-34)-induced desensitization of the adenylate cyclase. Radioreceptor assays showed that [125I]-[Tyr36]hPTHrP(1-36)amide binding to SaOS-2 cells was decreased by 60-70% by PTH(1-34) (1 x 10(-6) M), bPTH(1-84) (1.8 x 10(-6) M) and bPTH(3-34) (2 x 10(-6) M), whereas 8Br-cAMP (2 x 10(-4) M) had no effect on radioligand binding. PMA (1 x 10(-7) M) appeared to slightly increase [125I]PTHrP binding. This observation is consistent with a small (3-fold) increase in PTH-induced cAMP release as a result of PMA pre-treatment. Receptor internalization was dose-dependent EC50 = 3 x 10(-7) M PTH(1-34)). The maximal effect occurred after 10-30 min and was largely reversible within 2 h. Monensin (3 x 10(-5) M) inhibited the recovery from receptor internalization. We conclude that a perifusion system using SaOS-2 cells is a suitable model to study the effect of discontinuous application of PTH on cAMP release. A rapid, homologous desensitization of PTH(1-34) stimulated cAMP accumulation has been observed that does not appear to involve protein kinase A or C.