Background and Aims:

Stimulus-secretion coupling in pancreatic a-cells is poorly understood. Changes in the free cytosolic Ca²⁺ concentration ([Ca²⁺]_c) are implicated in the control of glucagon release, but the mechanisms by which secretagogues modify a-cell [Ca²⁺]_c remain highly controversial. These were studied here in single a-cells isolated from a mouse model expressing a fluorescent protein (FP) under the control of the glucagon promoter.

Materials and Methods:

Single cells, obtained by dispersion of isolated islets, were cultured for 1–3 days on glass coverslips in RPMI medium containing 7 mM glucose (G). For [Ca²⁺]_c measurements, the cells were loaded with 1 mM fura-PE3 in the culture medium. The coverslips were then transferred on the stage of an inverted microscope. Only cells expressing the FP (easily recognized upon excitation at 490 nm) were studied. [Ca²⁺]_c was measured with an imaging system by exciting the cells alternatively at 340 and 380 nm and collecting the emitted fluorescence at 510 nm . The fluorescence of the FP did not interfere with that of fura-PE3. Patch-clamp (perforated mode) was used to measure the K⁺_ATP current.

Results:

All islet cells expressing the FP were a-cells, as shown by immunodetection of glucagon. As expected, a-cells responded by a large [Ca²⁺]_c rise upon stimulation by 10mM adrenaline or 10 mM arginine. During perifusion with 0.5 mM G, [Ca²⁺]_c oscillated in 55% (47/86) of a-cells, but was low in the others. Application of 15 mM G to oscillating cells did not significantly change [Ca²⁺]_c, whereas inhibition of ATP synthesis by 2 mM azide abolished [Ca²⁺]_c oscillations and decreased [Ca²⁺]_c to nearly basal levels. Patch-clamp recordings showed that a-cells possess a K⁺_ATP current with similar characteristics to that of ?-cells: a large activation by 250 mM diazoxide (Dz) + azide, and a robust inhibition by 250 mM tolbutamide (Tb). Closing K⁺_ATP channels by addition of 10–500 mM Tb to a medium containing 0.5 mM G increased [Ca²⁺]_c in all a-cells. By contrast, increasing [G] from 0.5 to 15 mM in the continuous presence of 10 mM Tb decreased [Ca²⁺]_c by 40%. Opening K⁺_ATP channels with 100 mM Dz abolished [Ca²⁺]_c oscillations occurring in 0.5 mM G, and lowered [Ca²⁺]_c to basal levels. Dz also fully reversed the effects of 10 mM Tb, and completely prevented 10 mM arginine from increasing [Ca²⁺]_c. Inhibition of L-type Ca²⁺ channels with 1mM nimodipine virtually abolished [Ca²⁺]_c oscillations in 0.5 mM G, and largely reversed the effects of 10mM Tb (77%) or 10 mM arginine (89%).

Conclusions:

In contrast with previous suggestions, our results show that opening and closure of K⁺_ATP channels, or inhibiting mitochondrial metabolism with azide in a-cells change [Ca²⁺]_c in a similar way as in b-cells. However, contrary to b-cells, isolated a-cells are poorly responsive to glucose alone, which only slightly decreases [Ca²⁺]_c. This suggests that glucose does not induce major changes in the ATP/ADP ratio in single a-cells and supports the hypothesis that inhibition of glucagon secretion by a-cells in situ mainly results from an indirect effect of glucose involving non a-cells. The rise in [Ca²⁺]_c elicited by agents such as Tb or arginine mainly depends on Ca²⁺ influx through L-type Ca²⁺ channels.