Cross-talk between muscarinic- and adenosine-receptor signalling in the regulation of cytosolic free Ca2+ and insulin secretion.

The effects of A1-adenosine-receptor occupation on Ca2+ handling in the insulin-secreting RINm5F cell line were investigated. The selective A1-agonist N6-cyclopentyladenosine (CPA) had no effect itself on the cytosolic free Ca2+ concentration in cells loaded with Fura 2. However, CPA (1) attenuated the rise due to activation of voltage-gated Ca2+ channels with Bay K 8644, and (2) caused a secondary increase (EC50 approx. 300 nM) if added after the primary Ca(2+)-mobilizing agonists vasopressin or carbamoylcholine (carbachol). Prior addition of CPA (10 microM) also potentiated (by approx. 20%) the subsequent Ca2+ peak due to maximal (100 microM) carbachol, but did not alter the EC50 of the carbachol response. Detailed analysis of the secondary rise in Ca2+ revealed further features. First, it was due to mobilization from intracellular stores, since it persisted in the absence of extracellular Ca2+. Second, it was associated with a rapid (5-15 s) increase in phospholipase C (PLC) activity, as measured by h.p.l.c. analysis of Ins(1,4,5)P3. This increase was only apparent after prior stimulation with carbachol. Third, and unlike the response to carbachol, it was mediated by a pertussis-toxin-sensitive G-protein. Fourth, it was not secondary to a decrease in cyclic AMP. Fifth, it was absolutely dependent on continued occupation of the primary receptor, since it was abolished if carbachol was displaced with the antagonist atropine. This implies a dynamic cross-talk between the two receptor coupling systems, rather than covalent modification as a result of the prior activation of PLC. Sixth, it was not associated with any desensitization of the ability of CPA to inhibit forskolin-stimulated adenylate cyclase activity. Glyceraldehyde (10 mM)-induced insulin secretion was also potently inhibited by CPA > 10 nM, but the secretory response to 100 microM carbachol was unaffected up to 10 microM. The results suggest that, in vivo, adenosine would inhibit secretion due to carbohydrate nutrients much more effectively than that due to stimuli which activate PLC.