Limitations in success of clinical islet transplantation may be coupled to a long-term decline in the secretory capacity of the grafted human islet tissue. To address this issue human or mouse islets were transplanted to the subcapsular space of the kidney of nude mice. After 4 or 12 weeks, the grafts were removed and tested for insulin secretory dynamics in a perifusion system. Insulin secretion of non-transplanted human islets was examined as well. Insulin extracted from 12-week human islet grafts was significantly lower than that from 4-week grafts. Quite in contrast, 12-week mouse islet grafts contained as much insulin as the 4-week grafts. When stimulated with high glucose, insulin secretion was increased by about 6-fold in non-transplanted human islets and 3-fold in 4-week-grafts. The 12-week-grafts were just marginally stimulated by the high glucose stimulation. The mouse islets maintained a 2- to 3-fold insulin response at both time points when challenged high glucose. In non-transplanted human islets glucose-induced insulin secretion was inhibited by noradrenaline, while there was no such effect in the human islet grafts. Addition of acetylcholine potentiated glucose-induced insulin secretion 1-4 fold in both non-transplanted and grafted human islets. When human islet grafts were stimulated by both glucose and caffeine or arginine, insulin secretion was increased severalfold in comparison to glucose stimulation alone. The present results indicate that human islets, in contrast to mouse islets, progressively diminish their insulin content, as well as the capacity to secrete insulin in response to glucose after transplantation into nude mice. Moreover, grafted human islets also lose their responsiveness to the neurotransmittor noradrenaline. These findings may partly explain why clinical islet transplantation so far has met with limited success.