The b-cells of the pancreatic islets secrete insulin. Insulin secretion is regulated by changes in b-cell membrane potential that culminate in the opening or closure of voltage-gated calcium channels. Work on rodent b-cells have resulted in a fairly complete picture of the cellular control of insulin secretion. A major discovery was the discovery of the ATP-regulated potassium channels (K_ATP-channels) that constitute the resting membrane conductance of the b-cells. These channels close in response to glucose stimulation and this results in membrane depolarization. The K_ATP-channels are the target of hypoglycaemic sulphonylureas and they stimulate insulin secretion by direct interaction with the channel protein. Glucose- or tolbutamide-induced inhibition of the K_ATP-channels (via acceleration of glucose metabolism and increased ATP production) leads to the initiation of action potential firing in the beta-cell. During these action potentials, voltage-dependent Ca²⁺-channels activate leading to an increased intracellular calcium concentration that triggers the fusion of insulin-containing secretory vesicles.

The ion channels involved in beta-cell action potential firing have been characterized in some detail. Thus, in mouse b-cells the upstroke of the action potential depends principally on activation of L-type Ca²⁺-channels of the a1C subclass. In addition, R-type and P/Q-type Ca²⁺-channels mediate some limited calcium influx. Action potential repolarization is mediated by activation of voltage-dependent delayed rectifying K⁺-channels (Kv2.1).

The regulation of insulin secretion from human b-cells has widely been assumed to be very similar (if not identical) to that of rodent cells. Increased availability of human islet cells (as a by-product of islet transplantation programmes) has challenged this view. Like their rodent counterparts, human b-cells are equipped with K_ATP-channels and their closure triggers electrical activity. However, the complement of the voltage-gated Ca²⁺-channels involved in action potential firing exhibits marked differences. Thus, action potential firing in human b-cell involves both voltage-gated Na⁺-channels and low-threshold T-type Ca²⁺-channels; two types of ion channels not playing any role at all in mouse b-cells. The L-type Ca²⁺-channels that are essential for insulin secretion in rodent b-cells are not all involved in insulin secretion in human beta-cells but are rather required for action potential firing. Insulin secretion instead depends on Ca²⁺-influx through P/Q-type Ca²⁺-channels. Finally, action potential repolarization in human beta-cells results from opening of large-conductance Ca²⁺-activated K⁺-channels (channels that rarely activate in mouse beta-cells) whereas delayed rectifying Kv2.1/2.2 channels play a relative minor role. These differences between human and mouse b-cells are summarized in Table 1.

It is unlikely that the differences are limited to those already described and confined to ion channels. Our work so far has demonstrated the potential pitfalls of extrapolating rodent data to the situation in man. More work is required to understand the molecular and cellular causes of human diabetes (rodent diabetes is a very rare clinical condition!). Detailed knowledge about the genes expressed in human beta-cells is also essential for the interpretation of the genetic data that are now emerging. Clearly, understanding the functional consequences of a given gene polymorphism linked to diabetes depends critically on whether the gene is expressed in the beta-cell or not.

Table 1