Voltage-dependent Intracellular Calcium Release from Mouse Islets Stimulated by Glucose*

insulin secretion depends upon elevation of intracellular calcium concentration, which is thought to arise from Ca2+ influx through voltage-dependent calcium channels. Using mouse islets, we in component of

An increase in @-cell intracellular calcium concentration ( [Ca'+Ii)' is required for glucose-stimulated insulin secretion (1)(2)(3)(4). Electrophysiological recordings from islets of Langerhans have demonstrated that glucose-stimulated insulin secretion is associated with the appearance of phasic depolarizations which, in the physiological range of glucose concentrations, consist of bursts of calcium-dependent action potentials ( 5 ) . Other observations have indicated that glucose also induces phasic increases in [ Ca2+Ii in islets of Langerhans, primary P-cells, and cultured /I-cell lines (6)(7)(8)(9)(10). In the current model, this glucose-evoked [Ca2+Ii rise is thought to originate from membrane depolarization, secondary to the closure of ATP-dependent K+ channels ( K A~~) .
Voltage-dependent Ca2+ channels then open, leading to a further depolarization, an increase in [Ca2+], and exocytosis of insulin-containing granules (5, 11). In studies characterizing the depolarizationactivated calcium transients in @-cells and glucose-activated [Ca2+Ii increases in islets, the phasic increases in [Ca'+]; were * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. attributed solely to Ca2+ channel activation (6)(7)(8)(9)(10). However, other P-cell insulin secretagogues including acetylcholine induce [Ca2+Ji transients and insulin secretion, not by depolarization but instead by mobilizing intracellular Ca2+ stores via stimulating inositol 1,4,5-trisphosphate (IPS) production (12). Such an intracellular Ca2+ pool could likewise be mobilized secondary to depolarization via a charge-, Na+ current-, or Ca2+ current-coupled mechanism, as has been described in other cell types (13)(14)(15). In this report, utilizing primary cultures of intact mouse islets of Langerhans, we provide evidence that glucose in addition to activating Ca2+ influx also induces voltage-dependent release of Ca2+ from intracellular stores.

MATERIALS AND METHODS
Intact islets of Langerhans were obtained from the pancreata of 3-7-month-old C57BL/KsJ mice by collagenase digestion. After separation from acinar tissue on a discontinuous Ficoll gradient, the islets were individually hand-picked and cultured on uncoated coverslips in RPMI-1640 medium supplemented with fetal bovine serum, 11.6 mM glucose, 100 microunits/ml penicillin, and 100 pg/ml streptomycin for periods of 6-12 days before use. Medium was changed every 2-3 days. No studies on acutely isolated islets were possible, since the adhering process to the coverslips took a number of days. However, microelectrode recordings of electrical bursting activity from islets exposed to glucose were identical in islets maintained in culture and in acutely microdissected islet preparations.
Islets were loaded with fura-2 by a 25-min incubation at 37 "C in Krebs-Ringer buffer ((in mM) 119 NaC1, 4.7 KC1, 2.5 CaC12, 1.2 MgSO,, 1.2 KHZPO,, 25 NaHC03, 2 glucose) containing 5 PM acetoxymethyl ester of fura-2 (Molecular Probes Inc.). The specimen chamber (volume, 1 ml) was mounted on a temperature-controlled stage (Medical Systems Inc.) of an inverted microscope (Diaphot, Nikon Inc.) equipped with a X 40 Fluor objective (Nikon Inc.), and was continuously perfused with Krebs-Ringer buffer at 37 "C and pH 7.4 at a rate of 2.5 ml/min. Fura-2 dual excitation (340 and 380 nm) and fluorescence detection (510 nm) were accomplished using a Photoscan-2 ratio fluorescence photometry system (Nikon Inc.); the 340, 380, and 340/380 nm ratio signals were acquired continuously at 2 Hz. Estimations of free [Ca*+], were made using the methodology described elsewhere (16), where the maximum and minimum ratio calibrations were achieved by exposing a fura-2-loaded islet to 30 p~ ionomycin and 20 mM EGTA, pH 8.5, respectively. A value of 224 nM for the K d of Ca'+ for the dye was assumed. Glucose-activated Intracellular Calcium Release from Islets (phase l), peak values being achieved within 30-50 s of onset, and remained at these levels for periods ranging from 2 to 5 min before declining. After this period, phasic sinusoidal oscillations of [Ca'+Ii developed (phase 2) with a frequency of 0.5-3/min. The time course of these oscillations mirrored the phasic bursts of Ca2+-dependent action potentials recorded from islets exposed to glucose (8)(9)(10)(11)(12), supporting the idea that the phase 2 Ca2+ oscillations at least arise secondary to depolarization-driven Ca2+ influx. We treated islets with caffeine in order to investigate the role of mobilization of intracellular Ca2+ stores in mediating glucose-induced alterations of [Ca2+Ii. Caffeine has complex actions on intracellular Ca2+ stores in a variety of cell types; rapid exposure induces a transient release from intracellular storage sites, whereas prolonged incubation disables the store from releasing further ca'+ (17-19). In the presence of 2 mM glucose, rapid (<2 s) exposure of an islet to 5 mM caffeine caused a phasic increase in [Ca'+Ii ( Fig. 2 A ) . This increase was not blocked either by short incubations (1 min) with the Ca2+ channel blocker CoC12 (1 mM) or by low external Ca'+ (e100 KM), indicating that caffeine releases Ca'+ from intracellular stores in mouse islets. Rapid exposure to caffeine during 12 mM glucose-induced phase 2 [Ca'+]i oscillations suppressed the Ca'+ transients for a period of 30-60 s (Fig.  2B), thereafter being replaced by lower amplitude, higher frequency oscillations. Continuous superfusion with 5 mM caffeine caused a smaller and more slowly developing rise in [Ca2+Ii in the 2 mM glucose-containing solution (Fig. 2C). Elevating glucose to 12 mM in the continued presence of caffeine resulted in a blunted or completely absent phase 1 [Ca'+]i transient. Furthermore, the phase 2 [Ca2+]i oscillations were greatly attenuated and of higher frequency (Fig. 2C, a). Upon removal of caffeine in the presence of 12 mM glucose, there was a large and sustained phase 1 [Ca'+li transient (Fig.  2C, b ) consistent with the notion that this event represents augmented intracellular Ca2+ release, since caffeine promotes the refilling of intracellular Ca2' reservoirs (20). Re-exposure to 12 mM glucose following a further period in 2 mM glucose produced responses identical to the initial control (Fig. 2C,  c). The related xanthine compound, theophylline, which also mobilizes [Ca2+Ii stores (21,22), produced qualitatively similar changes in [Ca'+]i in the mouse islet to caffeine ( n = 3). On the other hand, ryanodine, which usually blocks intracellular Ca'+ release in the same systems where xanthines are effective (23, 24), failed to block the glucose-activated Ca2+ transient, either when applied during phase 2 [Caz+]i oscillations (Fig.  2 0 ) or following pretreatment prior to glucose exposure (n = 3). These data indicate that xanthine-specific intracellular Ca2+ mobilization and sequestration play a critical role in glucose-activated phase 1 [ Ca2+Ii transients and an important contributory role in phase 2 [ Ca2+]i oscillations.
The trigger mechanism for glucose-induced Ca2+ release was investigated in the following experiments. Exposure of the islet to 500 PM diazoxide during the glucose-induced phase 2 [Ca2+Ii oscillations totally suppressed the Ca'+ transients (Fig. 3A), an effect consistent with the phase 2 oscillations arising from phasic depolarization-induced Ca'+ influx, since diazoxide, by opening KATp channels, will suppress glucoseactivated depolarization (25). However, pretreatment with diazoxide before glucose exposure prevented the phase 1 Ca2+ transient as well (Fig. 3B). Since our results suggest that the phase 1 Ca'+ transient primarily represents Ca2+ release, this indicated that the trigger mechanism for release was voltagedependent. Whether the depolarization itself or the secondary influx of Ca2+ was the trigger for Ca2+ release was examined by lowering external Ca'+ below 50-100 WM. A variety of  (Fig. 3C). A and B, suppressive effect of 0.5 mM diazoxide during glucose-induced [Ca2+], oscillations ( A ) or prior to addition of glucose ( B ) . In both cases diazoxide suppressed glucose-dependent increases in [Ca2+];. Note that the initial phase 0 glucose-induced lowering of [Caz+]; was unaffected. C, effect of lowering extracellular CaZ+ concentration, [Ca2+l0, on glucose-stimulated alterations in [Ca2+],. EGTA (2.5 mM) was added 1 min prior to glucose exposure and for a further 20 min during exposure to 12 mM glucose. Note that on removal of EGTA a complete (phases 0, 1, and 2) pattern of responses to glucose was generated. A subsequent exposure to EGTA was made during the glucose-induced phase 2 [Ca2+]; oscillations. D, effect of Li+ substitution. Similar protocol to C, except that 119 mM NaCl was substituted with LiC1. Note that glucose has qualitatively the same effect on [Ca2+];. E, same protocol as in C, except that 1 PM TTX was applied to the perifusion buffer. Note that the inhibitory effect of TTX was completely reversible. F, KClinduced Ca2+ mobilization is only dependent on membrane depolarization. In the absence of extracellular calcium, the KC1-induced mobilization of [Ca"], was unaffected by 5 PM TTX. Representative records from a total of 13 experiments.  (Fig. 3C). Phase 2 [Ca2+Ii oscillations were not observed, indicating that external Ca2+ was too low to support Ca2+ current-driven Ca2+ transients. In accordance with this idea, EGTA addition during the glucose-activated phase 2 [Ca"+]i oscillations arrested the Ca2+ transients (Fig. 3C). Substitution of external Na+ with Li', which can also permeate through Na' channels, had no effect on the glucose-induced Ca2+ transient in the low Ca2+-containing solution, ruling out Na+/Ca2+ exchange as being responsible for the increase in [Ca2+Ji (26) (Fig. 3 0 ) . The glucose-induced phase 1 [Ca2+]i transient was, however, totally suppressed by application of 1 PM TTX, a blocker of voltage-dependent Na+ channels (Fig.  3E). Furthermore, T T X was without effect on the glucoseactivated [Ca2+Ii changes in normal Ca2+-containing solutions. The most likely explanation for these results was that both Na+ and CaZ+ influx triggered release of the Ca2+ store by a common mechanism by virtue of causing membrane depolarization; the failure of TTX to block release in normal CaZ+containing solutions would then relate to the greater importance of Ca2+ influx in bringing about P-cell depolarization (27). We tested this notion further by depolarizing the islet independently of glucose using 40 mM KCI. In the presence of 0.5 mM EGTA (0 mM Ca2+ added), KC1 produced a similar rise in [Caz+li to glucose, due to intracellular Ca2+ release (Fig.  3F). In contrast to the effects of glucose, addition of up to 5 PM TTX had no effect on the KCI-induced intracellular Ca2+ mobilization (Fig. 3F). These observations support the notion that the glucose-induced release of intracellular Ca2+ is dependent only upon membrane depolarization, the combined block of voltage-dependent Ca2+ and Na+ influx being sufficient to inhibit complete glucose-dependent depolarization and thereby prevent intracellular Ca2+ release.

1200s
Our data therefore indicate that glucose induces release of intracellular Ca2+ from a xanthine-sensitive but ryanodineinsensitive store via a depolarization-dependent mechanism. Since caffeine has been reported to block IP3 synthesis (28) and IP3 production itself may be a depolarization-activated process (29, 30), it is conceivable that glucose might mobilize Ca2+ stores by this mechanism. This effect would clearly differ from the voltage-independent Ca2+-mobilizing effect of muscarinic agonists on the P-cell (25) but would be consistent with reported elevations of IP3 by glucose (31) and the observations that IP3 receptors can mediate Ca2+-induced Ca2+ release (32). Furthermore, the putative involvement of IP, in synchronizing electrical oscillations in the islet (33) may explain the faster, low amplitude Ca2+ oscillations in the presence of caffeine (Fig. 2, B and C). Obviously, measurements of changes in islet IP3 production upon depolarization are warranted to fully test this hypothesis. A recent report has challenged the notion that IPS is a significant intracellular Ca2+ mobilizer in the islet (34), their findings suggesting that cyclic adenosine diphosphate-ribose, and not IP3, induces Ca2+ release from islet-derived microsomes. The applicability of their data to glucose-dependent [Ca2+]; mobilization is unclear since high concentrations of ryanodine (100 p~) were able to inhibit completely their microsomal Ca2+ release, whereas we were unable to detect any effect of ryanodine on the glucoseactivated Ca2+ release process (Fig. 2 0 ) . Whatever second messenger is involved, glucose mobilization of intracellular Ca2+ stores plays a vital role in determining both the nature and magnitude of the glucose-induced Ca2+ increases in the islet, in particular as it relates to the phase 1 [Ca2+]i transient. One of the earliest detectable defects in patients with type I1 (maturity onset) non-insulin-depend-

Glucose-activated Intracellular Calcium
Release from Islets ent diabetes mellitus is a loss of early (first phase) insulin secretion in response to a glucose challenge (35). In perifused mouse islets, there seems to be a close temporal relationship between the first phase and later second phase of glucosedependent insulin secretion, and the phase 1 and phase 2 ea2+ responses that we have observed (36, 37). This suggests that defects in intracellular Ca2+ release mechanisms may play an important role in the pathogenesis of this disease.