Insulinotropic Glucagon-like Peptide-1-mediated Activation of Non-selective Cation Currents in Insulinoma Cells Is Mimicked by Maitotoxin*

Maitotoxin (MTX) activates a Ca2+-dependent non-selective cation current (ICa-NS) in insulinoma cells whose time course is identical to non-selective cation currents activated by incretin hormones such as glucagon-like peptide-1 (GLP-1), which stimulate glucose-dependent insulin secretion by activating cAMP signaling pathways. We investigated the mechanism of activation of ICa-NS in insulinoma cells using specific pharmacological reagents, and these studies further support an identity between MTX- and GLP-1-activated currents. ICa-NS is inhibited by extracellular application of genistein, econazole, and SKF 96365. This inhibition by genistein suggests that tyrosine phophorylation may play a role in the activation of ICa-NS. ICa-NS is not inhibited by incubation of cells in glucose-free solution, by extracellular tetrodotoxin, nimodipine, or tetraethylammonium, or by intracellular dialysis with 4-aminopyridine, ATP, ryanodine, or heparin. ICa-NS is also not significantly inhibited by staurosporine, which does, however, partially inhibit the MTX-induced rise of intracellular Ca2+ concentration. These effects of staurosporine suggest that protein kinase C may not be involved in the activation of ICa-NS but that it may regulate intracellular Ca2+ release. Alternatively, ICa-NS may have a small component that is carried through separate divalent cation-selective channels that are inhibited by staurosporine. ICa-NS is neither activated nor inhibited by dialysis with KF, KF + AlF3 or GTPγS (guanosine 5′-O-(3-thiotriphosphate)), suggesting that GTP-binding proteins do not play a major role in the activation of this current.

␤-cells, however, are often unresponsive to glucose alone, but can become responsive by combined stimulation with glucose and hormones, such as insulinotropic hormone glucagon-like peptide-1 (GLP-1; Ref. 2), that elevate intracellular cAMP levels (3)(4)(5). One mechanism underlying this increased responsiveness is the enhanced closure of K ϩ ATP channels (2). A second mechanism by which GLP-1, pituitary adenylate cyclase-activating polypeptide (PACAP), and cAMP can induce ␤-cell depolarization is through the activation of voltageindependent, non-selective cation currents (6,7). Similar cation currents are also activated by MTX, a polyether toxin isolated from dinoflagellates that in ␤-cells has been shown to stimulate insulin secretion and inositol trisphosphate (Ins(1,4,5)P 3 ) production (8) and to enhance the influx of monovalent cations (9).
MTX-sensitive currents are activated by depletion of intracellular Ca 2ϩ stores (10), and GLP-1 enhances intracellular Ca 2ϩ mobilization through the potentiation of ryanodine-sensitive Ca 2ϩ -induced Ca 2ϩ release in ␤TC3 cells (11,12). Increased cAMP levels stimulate Ca 2ϩ release from secretory granules and reduce mitochondrial Ca 2ϩ uptake in ␤-cells (13,14). These observations raise the possibility that the activation of non-selective cation currents by PACAP and GLP-1 may be a secondary consequence of glucose-and cAMP-dependent intracellular Ca 2ϩ release. The physiological role of Ca 2ϩ releaseactivated currents in ␤-cells remains controversial, but such currents have been suggested to play a role in the cholinergic modulation of electrical bursting activity (15), and may control the membrane potential and intracellular Ca 2ϩ ([Ca 2ϩ ] i ) oscillations in response to nutrient stimulation (10).
Both PACAP (16) and GLP-1 (17) are potent insulin secretagogues in the presence of slightly elevated glucose levels. The activation of a voltage-independent, non-selective cation current by these hormones under conditions that stimulate insulin secretion suggests that this current may play an important role in depolarizing ␤-cells to initiate insulin secretion. The aim of this study is to examine the mechanism of activation of the MTX-sensitive current and to compare the properties of I Ca-NS with the current activated by GLP-1, PACAP, and cAMP to determine whether these currents are likely to be carried through the same channels.
Test Solutions-Cells were bathed in a standard extracellular solution (SES) containing: 138 mM NaCl, 5.6 mM KCl, 2.6 mM CaCl 2 , 1.2 mM MgCl 2 , 10 mM HEPES (295 mosM; pH adjusted to 7.4 with NaOH, approximately 4 mM), and 0.8 mM D-glucose unless indicated as being different in the text. Na ϩ -free, N-methyl-D-glucamine (NMG) solutions were prepared using 138 mM NMG substituted for NaCl and adjusted to pH 7.4 with HCl. A Na ϩ -free, 100 Ca solution was also prepared containing: 100 mM CaCl 2 , 1.2 mM MgCl 2 , 10 mM HEPES (282 mosM; pH adjusted to 7.4 with KOH, approximately 6 mM) Other test solutions were prepared by substitution of NaCl with 140 mM KCl, CsCl, LiCl, or choline chloride. Ca 2ϩ -free solutions were prepared by substituting MgCl 2 or MnCl 2 for CaCl 2 . Low [Cl Ϫ ] o solution was prepared by substituting NaCl with Na-aspartate.
Test solutions containing MTX were applied to individual cells by focal application from micropipettes using a PicoSpritzer II pressure ejection system (General Valve, Fairfield, NJ). A gravity-fed bath superfusion system was used to exchange and refresh bath solutions. Maitotoxin, econazole, staurosporine, nimodipine, and glyburide were obtained from Sigma. Tetrodotoxin, SKF 96365, genistein, GTP␥S, ryanodine, and heparin were obtained from Calbiochem. Tetraethylammonium chloride (TEA) and 4-aminopyridine (4AP) were obtained from Aldrich.
Measurement of Intracellular Calcium-Cells were prepared for measurement of [Ca 2ϩ ] i by incubation in fura-2 acetoxymethyl ester (fura-2 AM; Molecular Probes, Inc., Eugene, OR). Cells were loaded in SES supplemented with 2% fetal bovine serum, 0.03% pluronic F-127, and 1 M fura-2 AM for 15 min at room temperature (20 -22°C). Coverslips with fura-loaded adherent cells formed the base of a recording chamber mounted on a temperature-controlled stage (Micro Devices, Jenkintown, PA). Cells were visualized using a Zeiss IM35 microscope equipped with a Nikon UVF100 100ϫ objective. Measurements of [Ca 2ϩ ] i were performed at 1-s intervals from the average of 10 video frames using a dual excitation wavelength video imaging system (IonOptix Corp., Milton, MA). Experiments were conducted at 32°C. [Ca 2ϩ ] i was estimated from the ratio of 510 nm emission fluorescences due to excitation by 350 nm and 380 nm wavelength light from the following equation (18). Values of ␤, R min , and R max were determined using fura-2 pentapotassium salt and calibration solutions from Molecular Probes, Inc. Patch Clamp Recording Techniques-Cell resting potentials and membrane currents were measured under current clamp or voltage clamp using either whole-cell or perforated patch configurations (19,20). Patch pipettes pulled from borosilicate glass (Kimax-51, tip resistance 2-4 megohms) were fire polished and tip-dipped in K-or Cspipette solution containing: 95 mM K 2 SO 4 (or Cs 2 SO 4 ), 7 mM MgCl 2 , 5 mM HEPES (pH adjusted to 7.4 with NaOH; final concentration of Na ϩ approximately 2 mM). Pipettes were then back-filled with the same solution, to which nystatin (240 g/ml) was added for perforated patch recording.
The patch pipette was connected to an Heka Electronik EPC-9 patch clamp amplifier (Instrutech Corp., Mineola, NY) interfaced with a Macintosh Quadra 840AV computer running Pulse version 8.0 software (Instrutech Corp.). The series resistance (R s ) and cell capacitance (C m ) were monitored following seal formation, and experiments were conducted when R s declined to Ͻ35 megohms. In voltage clamp experiments, R s was compensated for by 60%.
Values are given as mean Ϯ S.E. Statistical significance was determined using Student's t test computed using Lotus 1-2-3 spreadsheets.

RESULTS
MTX activates a non-selective cation current in ␤TC6 cells ( Fig. 1) that has a reversal potential of -7.8 Ϯ 1.0 mV (n ϭ 72) in SES (142 mM [Na ϩ ] o ). This current was observed in all cells tested, and its reversal potential becomes more negative as [Na ϩ ] o is reduced (Fig. 1, Table I), although the shift is less than predicted from the Nernst potential for Na ϩ . I Ca-NS is also observed in solutions where extracellular Na ϩ is replaced by Cs ϩ , K ϩ , Li ϩ , choline ϩ , or NMG ϩ and in cells dialyzed with K ϩ , Cs ϩ , Na ϩ , choline ϩ , or NMG ϩ with Na ϩ in the bathing solution (Table II), confirming the non-selective nature of this current. Changing from normal [Cl Ϫ ] o to low [Cl Ϫ ] o extracellular solution did not affect the amplitude or reversal potential of the current (data not shown), confirming the cation selectivity of I Ca-NS .
It is reported that Ca 2ϩ is impermeant through MTX-activated channels in mouse ␤-cells (10) in contrast with a report that Ca 2ϩ is permeant through such channels in mouse L-cells (21). We therefore decided to re-examine the Ca 2ϩ permeability of I Ca-NS . Ca 2ϩ influx through MTX-sensitive channels in HIT-T15 cells is suggested by an increase of [Ca 2ϩ ] i when hyperpolarizing voltage steps are applied following activation of the MTX-sensitive current and by the rapid and reversible fall of [Ca 2ϩ ] i when extracellular Ca 2ϩ is removed (Figs. 2 and 3A). Fig. 2 shows that a hyperpolarizing voltage clamp step from Ϫ70 mV to Ϫ100 mV had no effect on [Ca 2ϩ ] i in a HIT-T15 cell before stimulation with MTX, but that similar voltage steps applied after activation of I Ca-NS resulted in a pronounced increase of [Ca 2ϩ ] i that was reversibly abolished by removal of extracellular Ca 2ϩ . Hyperpolarizing pulses applied following activation of non-selective cation currents by GLP-1 (22) or 8-bromo-cAMP (7) also increased [Ca 2ϩ ] i .
We further tested the effects of removal of extracellular Ca 2ϩ to confirm the activation of MTX-sensitive currents by Ca 2ϩfree solutions (10) and to determine the effects of activation of the current on [Ca 2ϩ ] i in the absence of extracellular Ca 2ϩ . In 15/18 cells tested, bath perfusion with a Ca 2ϩ -free solution containing 50 M EGTA (the Ca 2ϩ concentration in this solution was measured (using K 5 fura-2) to be 150 nM) before FIG. 1. MTX activates a non-selective cation current mainly carried by Na ؉ . Panel A shows the whole cell current from a perforated patch (Cs-pipette solution) voltage-clamped ␤TC6 cell (C m 11.3 pF) initially bathed in SES (142 mM Na ϩ ). All bath solutions contained 5 mM TEA, 2 M TTX, and 100 M Cd 2ϩ (to inhibit Ca 2ϩ -activated K ϩ channels, voltage-dependent Na ϩ channels, and VDCCs respectively). Breaks in the trace (labeled 1-6) mark where series of four voltage ramps from Ϫ70 mV to ϩ30 mV (1 V/s) were applied. Currents were filtered at 1 kHz. A 2-s pulse of 10 pM MTX was applied (arrow) and the bath solution subsequently changed to 14 mM Na ϩ (Tris-HCl-substituted), returned to SES, exchanged with 14 mM Na ϩ (NMG-substituted), and again returned to SES as indicated. Panel B shows the averaged currents from series of voltage ramps (as indicated in A) before MTX (1), after MTX stimulation (2), and after changing the bath to 14 mM Na ϩ (3). Panel C shows the currents from the ramp series subtracted as indicated. The reversal potential of the MTX-activated current in this cell was 0 mV in 142 mM Na ϩ and reversibly shifted to Ϫ31 mV in 14 mM Na ϩ ( Table I). The currents during ramp series (5) were too small to allow accurate determination of the reversal potential. stimulation with MTX caused a small, reversible increase of the membrane current and a fall of [Ca 2ϩ ] i (Fig. 3A). Holding currents at Ϫ70 mV increased from Ϫ0.96 Ϯ 0.16 pA/pF in SES to Ϫ1.64 Ϯ 0.30 pA/pF (p ϭ 0.005) and [Ca 2ϩ ] i decreased from 117 Ϯ 15 nM to 99 Ϯ 15 nM on removal of extracellular Ca 2ϩ (not statistically significant, p ϭ 0.4). In 3/18 cells the holding current showed a much larger increase on transfer to Ca 2ϩ -free solution (Fig. 3B). The holding current in these cells increased by Ϫ21.8 Ϯ 4.3 pA/pF and [Ca 2ϩ ] i reduced from 75 Ϯ 15 nM to 37 Ϯ 4 nM (p ϭ 0.1). This current inactivates rapidly, and [Ca 2ϩ ] i is transiently elevated when SES (2.6 mM Ca 2ϩ ) is reintroduced to the bath (Fig. 3B). Elevation of [Ca 2ϩ ] i and activation of a non-selective cation current were also observed in a similar subset of cells following stimulation with thapsigargin (2-10 M, data not shown), as reported previously (10).
Bath perfusion with Ca 2ϩ -free solution after the activation of I Ca-NS results in a rapid, reversible fall of [Ca 2ϩ ] i but has only a small effect on the current amplitude (Figs. 2 and 3A). The reversal potential of the MTX-activated current in Ca 2ϩ -free solution is Ϫ10.1 Ϯ 2.8 mV (n ϭ 12), not significantly different (p ϭ 0.43) from the reversal potential measured in SES. These data suggest that Ca 2ϩ -influx carries only a small proportion of the current through MTX-sensitive channels in the presence of Na ϩ . However, the rapid and reversible effects of removing extracellular Ca 2ϩ on [Ca 2ϩ ] i suggests that I Ca-NS does permit Ca 2ϩ entry.
Further evidence for the influx of divalent cations through MTX-sensitive channels is suggested by increased Mn 2ϩ quenching of intracellular fura-2 fluorescence following MTXstimulation (Fig. 3C). Fig. 3C (i and ii) shows the raw fura-2 fluorescence emission values, application of a 60-s pulse of Ca 2ϩ -free, Mn 2ϩ -substituted solution before stimulation with MTX produces a gradual quenching of the fluorescence signals, a small increase in the holding current (Fig. 3C, iii), and little or no change in the 350 nm/380 nm fluorescence ratio (Fig. 3C, iv). Following activation of the inward current and rise in [Ca 2ϩ ] i , a pulse of Ca 2ϩ -free, Mn 2ϩ -substituted solution produces rapid quenching of fura-2 fluorescence (Fig. 3C, i and ii) with little or no effect on membrane current (Fig. 3C, iii) and reduces the fluorescence ratio (Fig. 3C, iv), consistent with a decrease in [Ca 2ϩ ] i . Similar activation of Mn 2ϩ quenching has been observed following stimulation of insulinoma cells with PACAP (6), cAMP (22), and thapsigargin (23,24).
A role for Ca 2ϩ influx through MTX-sensitive channels is further supported by the effects of applying high [Ca 2ϩ ] o solutions. A Na ϩ -free test solution containing 100 mM Ca 2ϩ was applied to a ␤TC6 cell prior to stimulation with MTX and produces a small increase of [Ca 2ϩ ] i (Fig. 4), similar to the effects of applying elevated [Ca 2ϩ ] o solutions to mouse islets (25). Application of 100 mM [Ca 2ϩ ] o solution following stimulation with MTX produces a pronounced inhibition of I Ca-NS (Fig.  4) and a negative shift in the reversal potential of the current (Table I). This inhibition of I Ca-NS by the 100 mM Ca 2ϩ solution is accompanied by a rise of [Ca 2ϩ ] i (Fig. 4). These data could be Ϫ25.0 Ϯ 1.4

TABLE II Ionic dependence of reversal potentials (mV) for MTX-induced currents
Table shows the effect of extracellular (bath) and intracellular (pipette) ion substitutions on the reversal potential of I Ca-NS . The reversal potential with SES (Na ϩ ) bath solution and Cs ϩ -pipette solution is taken to be the reference value and ion substitutions that produced a significant change in the reversal potential are indicated. These data confirm the non-selective nature of the MTX-sensitive channels. *, p Ͻ 0.001; **, p ϭ 0.005; ***, p ϭ 0.014. explained by Ca 2ϩ influx through a single class of non-selective cation channel with a lower permeability to Ca 2ϩ than to Na ϩ under these experimental conditions. Alternatively, I Ca-NS may have two (or more) components, one component being monovalent cation selective, and the other, smaller, component being selective for divalent cations.
The non-selective cation current activated by 8-bromo-cAMP in HIT-T15 cells is inhibited by whole cell dialysis with Ca 2ϩfree, EGTA-buffered solutions or by loading the cells with BAPTA (1,2-bis(2-aminophenoxy)ethane N,N,NЈ,NЈ-tetraacetic acid), a Ca 2ϩ chelator (7) leading us to test the dependence of I Ca-NS activation on [Ca 2ϩ ] i (Fig. 5). Whole cell recordings from HIT-T15 cells were performed with either normal K-pipette solution (nominally Ca 2ϩ -free) or the same solution with 5 mM EGTA added (Ca 2ϩ -free). Whole cell dialysis with Ca 2ϩ -free intracellular solution inhibited activation of I Ca-NS compared with control cells from the same platings dialyzed with nomi-nally Ca 2ϩ -free solution or compared with cells in perforated patch voltage clamp (Fig. 5). These observations suggest that physiological [Ca 2ϩ ] i levels are required for activation of the current. It is notable that dialysis of the cells with Ca 2ϩ -free solution does not induce activation of the current alone, whereas dialysis with this solution might be expected to deplete intracellular Ca 2ϩ stores and thus activate store-operated currents.
Ca 2ϩ -activated non-selective cation (Ca-NS) channels are expressed in ␤-cells that are activated at cytosolic [Ca 2ϩ ] Ͼ 10 Ϫ4 M and are blocked by 1 mM ATP and also by 10 mM 4AP when applied to the cytosolic face of isolated, inside-out patches (26,27). Activation of I Ca-NS is observed with physiological [Ca 2ϩ ] i levels but inhibition of the current by dialysis of cells with Ca 2ϩ -free solutions suggests that it may be carried through Ca-NS channels. To test this possibility, ␤TC6 cells were bathed in glucose-free SES with 5 mM TEA and 10 nM glyburide added (to block Ca 2ϩ -activated K ϩ channels and K ϩ ATP channels) and dialyzed in the whole cell recording mode with K-pipette solution and 10 mM 4AP either with or without 2 mM ATP. MTX-induced currents in cells dialyzed without ATP had a mean peak amplitude of Ϫ23.6 Ϯ 7.9 pA/pF (n ϭ 6),  (Table I).

FIG. 5. Activation of the MTX-sensitive current is dependent upon intracellular Ca 2؉ . Bar graph of MTX-induced current amplitudes (inverted scale) in HIT cells held at
Ϫ70 mV in perforated patch (perf., n ϭ 5) or whole cell recording dialyzed with nominally Ca 2ϩ -free (WCR, n ϭ 5) and Ca 2ϩ -free (ϩ 5 mM EGTA, n ϭ 6) K-pipette solutions. Whole cell recording currents are not significantly different from perforated patch current amplitudes (p ϭ 0.39), whereas calcium-free currents (EGTA) are significantly inhibited compared with whole cell recording currents (p ϭ 0.003). and cells from the same platings dialyzed with 2 mM ATP had peak amplitudes of Ϫ22.5 Ϯ 7.5 pA/pF (n ϭ 6, not significantly different, p ϭ 0.9). These data indicate that activation of I Ca-NS is not glucose-dependent, consistent with previous reports of MTX-stimulated, glucose-independent insulin secretion (28). These differences in the sensitivity of I Ca-NS to [Ca 2ϩ ] i and to block by ATP and 4AP in whole cell recordings compared with inside-out patches may indicate that I Ca-NS is not carried through the Ca-NS channels reported previously (26,27) or may reflect the different recording configurations.
The non-selective cation current activated by PACAP in ␤TC6 cells is inhibited by SKF 96365 (22), a blocker of depletion-activated currents (29) that inhibits MTX-induced Ca 2ϩ influx and insulin secretion (30). Fig. 6 shows that application of 50 M SKF 96365 reversibly inhibits I Ca-NS in ␤TC6 cells, similar to its effect on the PACAP-induced current in these cells (22), further supporting the suggestion that these currents are carried by the same channel type.
The activation of Mn 2ϩ quenching of intracellular fura-2 fluorescence by MTX (Fig. 3C) is similar to that observed following thapsigargin treatment of insulinoma cells (23,24). Thapsigargin-sensitive Ca 2ϩ pools can also be depleted by econazole, which inhibits Ca 2ϩ -ATPases (31) and thereby elevates [Ca 2ϩ ] i and also inhibits Mn 2ϩ quenching in HIT cells (23). We therefore tested the effect of econazole on the activation of I Ca-NS . Fig. 7A shows that 10 M econazole significantly reduces the amplitude of I Ca-NS compared with control cells from the same platings. The basal (pre-MTX) [Ca 2ϩ ] i rose from 52 Ϯ 7 nM to 182 Ϯ 41 nM (n ϭ 5, p ϭ 0.01) in the presence of econazole and the peak amplitude of I Ca-NS decreased from Ϫ12.6 Ϯ 2.3 pA/pF to Ϫ3.6 Ϯ 1.1 pA/pF (n ϭ 5, p ϭ 0.01). The rise of [Ca 2ϩ ] i above basal during the MTX response was reduced from 449 Ϯ 206 nM to 49 Ϯ 14 nM (p ϭ 0.09) by econazole.
The effect of genistein, a tyrosine kinase inhibitor, on the activation of I Ca-NS was tested as capacitative Ca 2ϩ influx activated by thapsigargin can also be inhibited by genistein (Ref. 32, Fig. 7B Protein kinase C (PKC) activates capacitative Ca 2ϩ entry in rat insulinoma (RINm5F) cells, and this activation of Ca 2ϩ entry can be blocked by 1 M staurosporine (24). We therefore tested the effects of 1 M staurosporine on membrane currents and on the rise of [Ca 2ϩ ] i following MTX stimulation of ␤TC6 cells (Fig. 7C). Staurosporine had no significant effect on I Ca-NS ; the mean peak amplitude of control currents was Ϫ54.2 Ϯ 19.5 pA/pF (n ϭ 5) compared with Ϫ51.1 Ϯ 7.7 pA/pF (n ϭ 6, p ϭ 0.9) in cells from the same platings after addition of 1 M staurosporine to the bathing solution. The pre-MTX basal [Ca 2ϩ ] i in control cells was 61 Ϯ 11 nM and 96 Ϯ 15 nM in cells bathed in staurosporine (p ϭ 0.1), whereas the peak rise (increase above basal levels) of [Ca 2ϩ ] i was reduced from 1653 Ϯ 252 nM to 142 Ϯ 28 nM (p Ͻ 0.001) by staurosporine.
Activation of some types of non-selective cation current has been shown to be mediated by GTP-binding proteins (33). We therefore examined the potential role of GTP-binding proteins in the activation of I Ca-NS by dialysis of ␤TC6 cells with Cspipette solution supplemented with 10 mM 4AP, 2 mM Na 2 ATP, and either 10 mM KF, 10 mM KF ϩ 100 M AlF 3 , or 100 M GTP␥S. The bath contained SES with 5 mM TEA and 1 M TTX. Whole cell dialysis of cells for 15-20 min with KF (n ϭ 4), KF ϩ AlF 3 (n ϭ 5), or GTP␥S (n ϭ 5) failed to activate inward currents in cells that all subsequently responded to stimulation with MTX (data not shown).
MTX stimulates an increase in Ins(1,4,5)P 3 levels in ␤-cell lines (8), and this increase might activate I Ca-NS through an intracellular Ca 2ϩ release mechanism. Heparin is a specific blocker of Ins(1,4,5)P 3 receptors that should inhibit activation of I Ca-NS if Ins(1,4,5)P 3 -gated Ca 2ϩ stores play an important role. ␤TC6 cells were dialyzed in whole cell recordings with Cs-pipette solution plus 10 mM 4AP, 2 mM Na 2 ATP, and 0.5 mg/ml heparin for 3-4 min before stimulation with MTX. Activation of I Ca-NS was not inhibited in cells dialyzed with heparin. The amplitude of the currents was not significantly different from that in control cells from the same platings, and the reversal potential of the currents was Ϫ7.0 Ϯ 2.0 mV (n ϭ 9), not significantly different from control values (p ϭ 0.96).
GLP-1 enhances intracellular Ca 2ϩ mobilization from ryanodine-sensitive stores in ␤TC3 cells, and ryanodine reduces the amplitude of [Ca 2ϩ ] i spikes produced by depolarizing voltage clamp steps within 1 or 2 min (11). We introduced 100 M ryanodine into ␤TC6 cells by whole cell dialysis in the Cs/4AP/ ATP-pipette solution (as above) and allowed 3-4 min dialysis before stimulation with MTX. Ryanodine failed to prevent activation of I Ca-NS under these conditions, and the current amplitude and reversal potential (-8.0 Ϯ 1.9 mV, n ϭ 5) are not significantly different from control cells (p ϭ 0.97). DISCUSSION We propose that MTX activates the same non-selective cation current as stimulation with the peptide hormones GLP-1 and PACAP (6,7,22). These hormones couple through GTPbinding proteins (G-proteins) to activate adenylyl cyclase and elevate intracellular cAMP in ␤-cells (3)(4)(5), and cAMP analogs can also activate these non-selective cation currents. However, activation of G-proteins, by dialysis of cells with KF, KF ϩ AlF 3 , or GTP␥S (compounds that stimulate G-protein mediated activation of non-selective cation currents in epithelial cells (33) and activate K ϩ ATP channels in RINm5F and HIT-T15 insulinoma cells (34,35)), neither activated nor inhibited the MTX-sensitive current.
Depletion of intracellular Ca 2ϩ stores activates a MTX-sensitive current in mouse ␤-cells (10), and the stimulation of Ins(1,4,5)P 3 production by MTX (8) raises the possibility that Ins(1,4,5)P 3 -gated stores may play a role in the activation of this current. Parasympathetic, cholinergic stimulation of ␤-cells stimulates Ins(1,4,5)P 3 production, potentiates glucoseinduced insulin secretion (36), and also activates a TTX-insensitive Na ϩ -dependent depolarizing current (37) that may also be carried through Ca 2ϩ release-activated non-selective cation channels (15). The role of intracellular Ca 2ϩ release in the activation of this current has, however, been disputed, and cholinergic activation of this Na ϩ current is reported to be mediated by M 3 -type muscarinic receptors being coupled to Na ϩ channels (38). We observed that dialysis with heparin, a blocker of Ins(1,4,5)P 3 receptors, failed to inhibit activation of I Ca-NS , suggesting that Ca 2ϩ release from these stores may not be critical.
The presence of a Ca 2ϩ store depletion-activated current in pancreatic ␤-cells was proposed from studies showing: 1) that the state of filling of endoplasmic reticulum stores could regulate the membrane potential in mouse ␤-cells (39), 2) that thapsigargin can activate Mn 2ϩ quenching of intracellular fura-2 in RINm5F cells (24), and 3) that the Mn 2ϩ quenching pathway is inhibited by econazole in HIT-T15 cells (23). Activation of I Ca-NS by thapsigargin and its inhibition by both econazole and SKF 96365 are consistent with the suggestion that I Ca-NS may represent a Ca 2ϩ release-activated current (10,31).
A role for PKC in the activation of store-operated Ca 2ϩ entry in RINm5F cells was suggested from observations that the sustained [Ca 2ϩ ] i rise in response to combined stimulation with PKC-activating phorbol esters and thapsigargin is inhibited by staurosporine (24). We observed that staurosporine reduced the MTX-induced [Ca 2ϩ ] i rise but did not significantly inhibit the amplitude of I Ca-NS , suggesting that PKC may not play a direct role in the activation of I Ca-NS but does regulate [Ca 2ϩ ] i responses. Such effects could be mediated through inhibition of Ca 2ϩ release from intracellular Ca 2ϩ stores, or could reflect the inhibition of a small, divalent cation selective component of I Ca-NS carried through a distinct set of channels other than the non-selective cation channels. The high concentration of staurosporine (1 M) used in these experiments would also be expected to inhibit cAMP-dependent protein kinase, and further studies are required to elucidate the role(s) of PKC and cAMPdependent protein kinase in the pathway(s) leading to the activation of I Ca-NS and regulation of [Ca 2ϩ ] i changes.
Inhibition of both I Ca-NS and the MTX-induced [Ca 2ϩ ] i rise by genistein, an inhibitor of tyrosine kinases (40) that blocks thapsigargin-and carbachol-induced Ca 2ϩ entry (32,41), suggests a role for tyrosine phosphorylation in the activation pathway of I Ca-NS . However, the role of tyrosine phosphorylation remains ambiguous as only certain tyrosine kinase inhibitors (including genistein) effectively block capacitative Ca 2ϩ entry (42). Thapsigargin-induced Ca 2ϩ entry can also occur in the absence of detectable tyrosine phosphorylation but is still inhibited by tyrosine kinase inhibitors (43), and, therefore, the role of tyrosine kinases in activation of I Ca-NS remains to be clarified.
Ca-NS channels are expressed in ␤-cells that are activated at cytosolic [Ca 2ϩ ] Ͼ 10 Ϫ4 M (26,27). Activation of I Ca-NS is observed at physiological [Ca 2ϩ ] i (approximately 100 nM) and is inhibited by dialysis of cells with Ca 2ϩ -free solution, similar to the inhibition of cAMP-activated currents (7) and suggesting a Ca 2ϩ -dependent step in the activation pathway of these channels. Ca-NS channels are also expressed in pancreatic acinar cells, and these channels are activated at much lower cytosolic Ca 2ϩ concentrations in whole cell records than in isolated patches (44); a similar difference in Ca 2ϩ sensitivity seems likely to occur for Ca-NS channels in ␤-cells. Ca-NS channels are blocked by 1 mM ATP and by 10 mM 4AP when applied to the intracellular face of isolated membrane patches (27); however, these two compounds did not inhibit I Ca-NS when introduced into the cytosol by whole cell dialysis (at 2 mM and 10 mM, respectively). It remains to be determined whether these differences in sensitivity to [Ca 2ϩ ] i , 4AP, and ATP are a consequence of the different recording conditions or suggest that Ca-NS channels do not carry I Ca-NS .
Reducing extracellular [Cl Ϫ ] has no effect on I Ca-NS amplitude or on its reversal potential, confirming the cation selectivity of the channel. Depletion of intracellular ATP levels through bathing cells in glucose-free media and dialyzing with ATP-free solutions or dialysis of cells with 2 mM ATP has no effect on I Ca-NS amplitudes. This further distinguishes the MTX-induced current from the non-selective anion current described in insulin-secreting cells that increases in amplitude upon dialysis with 2 mM ATP (45).
Elevation of [Ca 2ϩ ] i is observed following the activation of non-selective cation currents by MTX, GLP-1, or PACAP in voltage clamped cells where activation of voltage-dependent Ca 2ϩ channels is prevented (6,7,22). This rise of [Ca 2ϩ ] i is reversed by removal of extracellular Ca 2ϩ , suggesting that Ca 2ϩ influx is associated with the non-selective current, although it remains to be determined whether a single class of channel is permeant to both monovalent and divalent cations, or if two (or more) distinct conductances are involved. The physiological role of Ca 2ϩ influx associated with I Ca-NS in the stimulation of insulin secretion remains to be determined. It has been reported that sustained Ca 2ϩ influx through L-type VDCCs is strongly coupled to insulin secretion from HIT-T15 cells, whereas more transient Ca 2ϩ influx through N-type Ca 2ϩ channels is only weakly coupled (46). The Ca 2ϩ influx associated with I Ca-NS is prolonged and may, therefore, be able to contribute to the sustained [Ca 2ϩ ] i elevation that triggers secretion. However, the magnitude of this Ca 2ϩ influx is likely to be small compared with that through L-type channels as, under physiological conditions, the cells will depolarize to a value close to that for the reversal potential of I Ca-NS , and also influx through L-type VDCCs raises [Ca 2ϩ ] i very rapidly, whereas the [Ca 2ϩ ] i increase associated with I Ca-NS develops much more slowly. This slow time course of the rise in [Ca 2ϩ ] i is consistent with a small amplitude Ca 2ϩ influx and would explain why it is difficult to resolve a decrease in I Ca-NS amplitude on changing to Ca 2ϩ -free solution with normal extracellular Na ϩ concentrations. It therefore seems that the main physiological role for these non-selective cation currents in the control of insulin secretion will be to depolarize the membrane potential and activate VDCCs.
The currents activated by GLP-1, PACAP, cAMP analogs, and MTX are Ca 2ϩ -dependent non-selective cation currents that activate over tens of seconds and persist for extended periods following removal of the stimulus. These currents are all insensitive to TTX, L-type Ca 2ϩ channel blockers, TEA, and ryanodine but are inhibited by NMG, SKF96365, and La 3ϩ . Based upon these similarities between the temporal properties of the currents, and their associated [Ca 2ϩ ] i changes, and the pharmacology of the currents, we propose that these agents activate the same non-selective cation channels. The precise mechanism(s) controlling the activation of these non-selective cation channels remains to be determined, but a role for tyrosine kinase-induced phosphorylation is suggested by the effects of genistein. Activation of I Ca-NS may also be controlled by the state of filling of intracellular Ca 2ϩ stores, or may be partly due to Ca 2ϩ release from intracellular stores raising cytosolic Ca 2ϩ levels. We propose that activation of MTX-sensitive non-selective cation channels may play an important role in depolarizing ␤-cells in response to stimulation by GLP-1 and PACAP during feeding to initiate insulin secretion without large elevations of blood glucose. We also suggest that Ca 2ϩ is permeant through the MTX-sensitive channels and suggest that spontaneous activity of these channels may form the depolarizing, non-selective background conductance that permits Ca 2ϩ influx (25) and opposes the activity of ATP-sensitive K ϩ channels in regulating the resting potential of ␤-cells under both basal conditions and in response to hormonal stimulation.