Two microelectrodes have been used to measure membrane potentials simultaneously in pairs of mouse pancreatic islet cells. In the presence of glucose at concentrations between 5.6 and 22.2 mM, injection of current i into cell 1 caused a membrane potential change in this cell, V1, and, provided the second microelectrode was less than 35 micron away, in a second impaled cell 2, V2. This result establishes that there is electrical coupling between islet cells and suggests that the space constant of the coupling ratio within the islet tissue is of the order of a few beta-cell diameters. The current-membrane potential curves i-V1 and i-V2 are very similar. By exchange of the roles of the microelectrodes, no evidence of rectification of the current through the intercellular pathways was found. Removal of glucose caused a rapid decrease in the coupling ratio V2/V1. In steady-state conditions, the coupling ratio increases with the concentration of glucose in the range from 0 up to 22 mM. Values of the equivalent resistance of the junctional and nonjunctional membranes have been estimated and found to change with the concentration of glucose. Externally applied mitochondrial blockers induced a moderate increase in the junctional resistance possibly mediated by an increase in intracellular Ca2+.

When the electrode tips were separated by less than about 20 ,tm and current injection showed the cells to be ionically coupled the two signals were in phase and had almost identical shape. The phase relations between cells further apart were variable, the leading cell usually being located deeper within the islet than the other impaled cell. Increasing the glucose concentration increased electrical activity, reduced any phase lags and made the shape of the bursts more similar. There was less lag between the responses from two cells when the glucose concentration was suddenly reduced, than when it was suddenly increased. Qualitatively similar observations were made in glibenclamide-treated mice, a treatment previously shown to increase dye coupling between islet cells. However, the response to increasing glucose concentrations showed less phase lag; likewise the phase lag between bursts was reduced. Furthermore the response to current injected into one cell could be detected at much larger distances (up to 80,tm) than in control islets. This suggests that electrical coupling offl-cells was improved in sulphonylureatreated mice. Electron microscopy of both control and glibenclamide-treated mouse islets fixed at the end of each electrophysiological experiment showed the region impaled by the electrodes to be well preserved and, whenever the electrodes penetrated at least 20 ,um into the islet, to contain a large proportion of f-cells. The data support the view that, within an islet, most but not necessarily all cells are electrically synchronized, and that the coupling can be modulated by natural and pharmacological secretagogues.

The effects of charybdotoxin (CTX) on single [Ca2+]-activated potassium channel (K(Ca)) activity and whole-cell K+ currents were examined in rat and mouse pancreatic beta-cells in culture using the patch-clamp method. The effects of CTX on glucose-induced electrical activity from both cultured beta-cells and beta-cells in intact islets were compared. K(Ca) activity was very infrequent at negative patch potentials (-70 less than Vm less than 0 mV), channel activity appearing at highly depolarized Vm. K(Ca) open probability at these depolarized Vm values was insensitive to glucose (10 and 20 mM) and the metabolic uncoupler 2,4 dinitrophenol (DNP). However, DNP blocked glucose-evoked action potential firing and reversed glucose-induced inhibition of the activity of K+ channels of smaller conductance. The venom from Leiurus quinquestriatus hebreus (LQV) and highly purified CTX inhibited K(Ca) channel activity when applied to the outer aspect of the excised membrane patch. CTX (5.8 and 18 nM) inhibited channel activity by 50 and 100%, respectively. Whole-cell outward K+ currents exhibited an early transient component which was blocked by CTX, and a delayed component which was insensitive to the toxin. The individual spikes evoked by glucose, recorded in the perforated-patch modality, were not affected by CTX (20 nM). Moreover, the frequency of slow oscillations in membrane potential, the frequency of action potentials and the rate of repolarization of the action potentials recorded from pancreatic islet beta-cells in the presence of glucose were not affected by CTX. We conclude that the K(Ca) does not participate in the steady-state glucose-induced electrical activity in rodent pancreatic islets.

SUMMARY1. Insulin release and fl-cell membrane potentials in response to glucose at 37 and 27 TC have been measured simultaneously in single, micro-dissected, perifused islets of Langerhans from normal mice.2. Insulin release and 45Ca outflow in response to glucose at 37 and 27 'C have been measured simultaneously from perifused islets isolated by collagenase digestion from normal rats.3. The effect of cooling on fl-cell membrane potassium permeability was assessed by changes in measured membrane potential and input resistance (in the mouse) and by changes in 86Rb outflow (in the rat).4. Resting and active fl-cell membrane parameters (i.e. membrane potential, spike frequency, input resistance, 45Ca outflow and 86Rb outflow), in both mouse and rat islets, were affected only slightly by cooling to 27 'C, with temperature coefficients of 2 or lower.5. At 27 'C glucose-stimulated insulin release was inhibited completely in mouse islets and almost completely in rat islets. The temperature coefficients in both preparations were greater than 5.6. It is concluded that #-cell electrical activity and changes in membrane permeability induced by glucose are not consequences of insulin release.