The purpose of this study was to characterize the trophic effect of pancreatic duct cells on the islets of Langerhans. Ductal epithelium and islets were isolated from hamster pancreata. In addition, duct-conditioned medium (DCM) was prepared from primary duct cultures that had been passaged twice to remove other cellular elements. Three experimental groups were then established: Group 1, 100 islets alone; Group 2, 100 islets+80 duct fragments; and Group 3, 100 islets in 25% DCM. All tissues were embedded in rat tail collagen for up to 12 days and the influence of pancreatic ductal epithelium on islet cell survival was examined. By day 12, 20.6+/-3. 0% (S.E.M.) of the islets cultured alone developed central necrosis, compared with 6.7+/-2.0% of the islets co-cultured with ducts and 5.6+/-1.5% of the islets cultured in DCM (P<0.05). The presence of apoptotic cell death was determined by a TdT-mediated dUTP-biotin nick end labelling (TUNEL) assay and by a specific cell death ELISA. DNA fragmentation in islets cultured alone was significantly increased compared with islets cultured either in the presence of duct epithelium or in DCM (P<0.05). More than 80% of TUNEL-positive cells were situated in the inner 80% of the islet area, suggesting that most were beta-cells. DCM was analysed for known growth factors. The presence of a large amount of IGF-II (34 ng/ml) and a much smaller quantity of nerve growth factor (4 ng/ml) was identified. When the apoptosis studies were repeated to compare islets alone, islets+DCM and islets+IGF-II, the cell death ELISA indicated that IGF-II produced the same beneficial result as DCM when compared with islets cultured alone. We conclude that pancreatic ductal epithelium promotes islet cell survival. This effect appears to be mediated in a paracrine manner by the release of IGF-II from cells in the ductal epithelium.