Glucose-mediated inhibition of calcium-activated potassium channels limits α-cell calcium influx and glucagon secretion

Abstract

Pancreatic α-cells exhibit oscillations in cytosolic Ca²⁺ (Ca²⁺_c), which control pulsatile glucagon (GCG) secretion. However, the mechanisms that modulate α-cell Ca²⁺_c oscillations have not been elucidated. As β-cell Ca²⁺_c oscillations are regulated in part by Ca²⁺-activated K⁺ (K_slow) currents, this work investigated the role of K_slow in α-cell Ca²⁺ handling and GCG secretion. α-Cells displayed K_slow currents that were dependent on Ca²⁺ influx through L- and P/Q-type voltage-dependent Ca²⁺ channels (VDCCs) as well as Ca²⁺ released from endoplasmic reticulum stores. α-Cell K_slow was decreased by small-conductance Ca²⁺-activated K⁺ (SK) channel inhibitors apamin and UCL 1684, large-conductance Ca²⁺-activated K⁺ (BK) channel inhibitor iberiotoxin (IbTx), and intermediate-conductance Ca²⁺-activated K⁺ (IK) channel inhibitor TRAM 34. Moreover, partial inhibition of α-cell K_slow with apamin depolarized membrane potential (V_m) (3.8 ± 0.7 mV) and reduced action potential (AP) amplitude (10.4 ± 1.9 mV). Although apamin transiently increased Ca²⁺ influx into α-cells at low glucose (42.9 ± 10.6%), sustained SK (38.5 ± 10.4%) or BK channel inhibition (31.0 ± 11.7%) decreased α-cell Ca²⁺ influx. Total α-cell Ca²⁺_c was similarly reduced (28.3 ± 11.1%) following prolonged treatment with high glucose, but it was not decreased further by SK or BK channel inhibition. Consistent with reduced α-cell Ca²⁺_c following prolonged K_slow inhibition, apamin decreased GCG secretion from mouse (20.4 ± 4.2%) and human (27.7 ± 13.1%) islets at low glucose. These data demonstrate that K_slow activation provides a hyperpolarizing influence on α-cell V_m that sustains Ca²⁺ entry during hypoglycemic conditions, presumably by preventing voltage-dependent inactivation of P/Q-type VDCCs. Thus, when α-cell Ca²⁺_c is elevated during secretagogue stimulation, K_slow activation helps to preserve GCG secretion.

INTRODUCTION

Under conditions of low blood glucose, pancreatic α-cells secrete glucagon (GCG), which stimulates hepatic glucose output (35, 46, 49). Thus, GCG secretion serves a central role in preventing hypoglycemia and maintaining glucose homeostasis. Some of the first studies examining α-cell function revealed that extracellular Ca²⁺ is required for GCG secretion (18, 41). It was subsequently shown that Ca²⁺ entry through voltage-dependent Ca²⁺ channels (VDCCs) stimulates GCG secretion (43, 47, 50). As VDCC activation is mediated by membrane potential (V_m) depolarization (7, 56), changes in α-cell V_m controlled by ion channel activity modulates GCG secretion. Furthermore, in diabetic states, changes in gene expression of transcripts coding for Ca²⁺ handling proteins as well as ion channel function have been linked to perturbations in GCG secretion (5, 53, 69). Although these observations indicate an important role for Ca²⁺ in controlling GCG secretion, our understanding of α-cell Ca²⁺ handling is incomplete.

It has become clear that α-cells display glucose-regulated electrical excitability that controls Ca²⁺ entry through VDCCs (4, 10, 36). In turn, the electrical excitability of α-cells is controlled by ion channel control of V_m (10, 20, 21, 29, 32, 62, 68, 69). For example, K⁺ channels such as ATP-sensitive K⁺ (K_ATP) channels, voltage-gated K⁺ (Kv) channels, G protein-coupled inwardly rectifying K⁺ (GIRK) channels, and two-pore domain K⁺ (K2P) channels tune α-cell V_m and thus VDCC activity (10, 30, 43, 57, 64, 69). Inhibition of most K⁺ channels reduce GCG secretion by depolarization-induced voltage-dependent inactivation (VDI) of VDCCs (50, 57, 69). Following VDI of VDCCs, V_m hyperpolarization is required to regenerate channel activity (57). One physiological way this can be accomplished is through V_m hyperpolarization between waves of electrical excitability. Indeed, α-cells undergo transient spontaneous V_m hyperpolarization in part because of somatostatin (SST)-mediated GIRK channel activation (4). However, spontaneous hyperpolarization events persist under low glucose conditions, whereas SST secretion is greatly reduced, suggesting the existence of other mechanisms (4). Recently, it has been shown that a subset of α-cells exhibit oscillations in cytosolic Ca²⁺ (Ca²⁺_c) similar to β-cells, which may allow for recovery from VDCC VDI (36, 37, 39, 64). As α-cell GCG secretion depends on Ca²⁺ entry through VDCCs, control of Ca²⁺ oscillations would be predicted to influence GCG secretion. In pancreatic β-cells, a Ca²⁺-activated K⁺ (K_slow) current plays a key role in the termination of Ca²⁺ oscillations (12, 20, 21, 29, 32, 68). However, the nature and importance of K_slow in modulating α-cell Ca²⁺ oscillations and GCG secretion has not been investigated.

In β-cells, small-conductance Ca²⁺-activated K⁺ (SK) channels and intermediate-conductance Ca²⁺-activated K⁺ (IK) channels contribute to K_slow currents (12, 15, 32, 68). Importantly, transcriptome analyses of highly pure mouse and human islet cell populations show that α-cells also express some of these channels. For example, α-cells express a substantial amount of KCNN3 transcript (gene encoding SK3 channels), as well as lower levels of KCNN1 (gene encoding SK1 channels) and KCNN2 (gene encoding SK2 channels) transcripts (3, 14). The same transcriptome analyses detected minimal KCNN4 (gene encoding IK channels) in α-cells (3, 14); however, KCNN4 expression was also low in β-cells despite the importance of IK channels to β-cell K_slow (15). Therefore, low levels of KCNN4 transcript can produce functional ion channels that regulate islet cell electrical excitability. Thus, it is important to determine how SK and IK channels influence α-cell Ca²⁺ handling and GCG secretion.

Although a functional role for K_slow has not been established in α-cells, large-conductance Ca²⁺-activated K⁺ (BK) channels (encoded by KCNMA1) are highly expressed in α-cells in which they facilitate V_m hyperpolarization (3, 14, 57). BK channels are sensitive to changes in both Ca²⁺ and voltage; Ca²⁺-binding left shifts the activation voltage of BK channels to more negative V_m values, which results in increased channel activity under physiological conditions (65). BK channels account for a significant portion of K⁺ current in human and mouse α-cells, and selective inhibition with iberiotoxin (IbTx) reduces GCG secretion (57), which is presumably due to VDI of VDCCs. However, the exact impact of BK channels on α-cell Ca²⁺ handling and K_slow current propagation have not been examined.

This manuscript shows for the first time, to the best of our knowledge, that α-cells exhibit K_slow currents that modulate α-cell Ca²⁺ handling and GCG secretion. SK, BK, and IK channels contribute to α-cell K_slow, which is modulated in part by Ca²⁺ influx through L- and P/Q-type VDCCs, as well as Ca²⁺ released from endoplasmic reticulum (ER) Ca²⁺ (Ca²⁺_ER) stores. Selective inhibition of SK channels depolarized α-cell V_m and transiently enhanced Ca²⁺ entry; however, prolonged inhibition of SK or BK channels reduced Ca²⁺ entry. Moreover, apamin-mediated SK channel inhibition reduced GCG secretion, which was likely due to VDI of VDCCs. These findings demonstrate the importance of α-cell V_m hyperpolarization, facilitated by K_Ca channel activation, in sustaining VDCC function and preserving GCG secretion during secretagogue stimulation.

METHODS

Ethical approval.

Animals were handled in compliance with guidelines approved by the Vanderbilt University Animal Care and Use Committee (protocol no. M1600063–00). All mice used in these studies were 8- to 18-wk-old, age-matched males on a C57Bl6/J background. Healthy human islets were provided through the Integrated Islet Distribution Program from multiple isolation centers. The Integrated Islet Distribution Program obtained informed consent for deceased donors in accordance with National Institutes of Health guidelines before reception of human islets for our studies. Work detailed here was approved by the Vanderbilt University Health Sciences Committee Institutional Review Board (IRB no. 110164). Islet donor information is provided in Table 1.

Table 1.

Summary of human islet donor information

Age, yr	Sex	Race	BMI	Assay Performed
32	Male	African American	27.8	GCG secretion
49	Female	Caucasian	31.6	GCG secretion
56	Female	Hispanic	22.7	GCG secretion
54	Male	Caucasian	21.7	GCG secretion
59	Female	Caucasian	32.3	GCG secretion
50	Male	Caucasian	22.4	GCG secretion
43	Female	Caucasian	30.9	GCG secretion
57	Female	Caucasian	25.8	Ca2+c imaging
32	Male	Caucasian	26.2	Ca2+c imaging
45	Female	Caucasian	32.1	Ca2+c imaging
32	Male	Caucasian	32.3	Ca2+c imaging

BMI, body mass index; Ca²⁺_c, cytosolic Ca²⁺; GCG, glucagon.

Chemicals and reagents.

All research materials were purchased from Sigma-Aldrich (St. Louis, MO) or Thermo-Fisher (Waltham, MA), unless otherwise specified.

Transgenic mouse models.

Transgenic mice expressing a tandem-dimer red fluorescent protein (tdRFP) fluorescent reporter, specifically in α-cells, were generated by crossing mice expressing GCG-IRES-Cre with mice expressing a tdRFP fluorescent reporter preceded by a loxP-flanked STOP cassette in the Rosa26 locus (α-RFP) (25, 37, 40). Similarly, transgenic mice expressing a GCaMP3 fluorescent Ca²⁺ indicator, specifically in α-cells, were generated by crossing mice expressing GCG-IRES-Cre with mice expressing GCaMP3 preceded by a loxP-flanked STOP cassette in the Rosa26 locus (α-GCaMP3; Jackson Laboratories; cat. no. 014538) (10, 64).

Islet isolation and cell culture.

Mouse pancreata were digested with collagenase P (Roche; Basel, Switzerland), and islets were isolated using density gradient centrifugation as previously described (12, 13, 62). After isolation, mouse islets were gently triturated in 0.005% trypsin to disperse them into single cells or islet cell clusters (or left as whole islets). Following reception, human islets were washed twice with Versene, incubated in 0.005% trypsin for 5 min in a 37°C water bath, and gently triturated to disperse into single cells. Whole islets, partially dispersed islet cell clusters, and completely dispersed islet cells were plated on poly-d-lysine-coated glass bottom dishes and cultured at 37°C, 5% CO₂ in RPMI-1640 medium (11 mM glucose) supplemented with 15% FBS, 100 IU/ml penicillin, and 100 mg/ml streptomycin.

Perforated-patch voltage-clamp α-cell K_slow recording.

Single α-cells were identified by tdRFP fluorescence and patched in Krebs-Ringer-HEPES buffer (KRHB) with (in mM) 119 NaCl, 2.0 CaCl₂, 4.7 KCl, 25 HEPES, 1.2 MgSO₄, and 1.2 KH₂PO₄ [adjusted to pH 7.4 with NaOH, an osmotic concentration of 310 Osm with sucrose, and supplemented with 1 (KRHB-1mM) or 11 (KRHB-11mM) mM glucose]. Borosilicate patch pipettes (6–12 MΩ) were loaded with intracellular solution containing (in mM) 11.8 NaCl, 63.7 KCl, 1 MgCl_2, 28.4 K₂SO₄, 20.8 HEPES, and 0.5 EGTA (adjusted to pH 7.25 with KOH and supplemented with 0.05 mg/ml of the pore-forming agent amphotericin B) (63). To preserve the integrity of intracellular metabolism and signaling, a perforated-patch voltage-clamp technique was utilized to record α-cell K_slow currents (48). Using an Axopatch 200B amplifier with pCLAMP10 software (Molecular Devices; Sunnyvale, CA), K_slow currents were generated with a previously published (12, 20) β-cell K_slow protocol (K_slow; Fig. 1A) or a protocol developed to more closely resemble an α-cell action potential (AP; K_{slow, α}; Fig. 2A). The K_slow protocol is as follows: the command voltage was held at −40 mV for 7 s to establish baseline current density, ramped between −40 and 20 mV for 5.2 s (wave form: triangle; train rate: 5 Hz; pulse width: 0.2 s) to activate K_slow and returned to −40 mV to inactivate VDCCs and Kv channels (12, 20). The K_{slow, α} protocol was as follows: the command voltage was held at −45 mV for 300 ms to establish baseline current density, stepped to −70 mV for 5 s to alleviate VDI of P/Q-type VDCCs, ramped from −45 mV to 20 mV over 50 ms to activate K_slow, and returned to −45 mV to inactivate VDCCs and Kv channels. K_slow currents from 2 to 3 untreated α-cells were measured in a single dish at room temperature, and then buffer containing the indicated treatment (Figs. 1 and 2) was added. The cells were equilibrated for 15 min, and K_slow currents from 2 to 3 treated α-cells were recorded at room temperature. Agatoxin, apamin, IbTx, isradipine, thapsigargin (Tg), and TRAM 34 were purchased from Alomone Laboratories (Jerusalem, Israel). UCL 1684 was purchased from Tocris Bioscience (Minneapolis, MN). K_slow currents obtained using the K_slow protocol were fitted to a model of two-phase exponential decay (Eqs. 1A–1C) using GraphPad Prism software, and the associated kinetic parameters were determined. K_{slow, max} indicates the peak K_slow current density (pA/pF), plateau indicates steady state K_slow current as time approaches infinity (pA/pF), % fast indicates the percentage of K_slow occurring in the fast-phase, K_f indicates fast-phase rate constant (s⁻¹), and K_s indicates slow-phase rate constant (s⁻¹).

Fig. 1.

α-Cell Ca²⁺-activated K⁺ (K_slow) currents are activated by Ca²⁺ waves generated by repeated membrane potential depolarization. A: overview of pulse train K_slow protocol (top) and a representative K_slow current recorded from a single red fluorescent protein-expressing (α-RFP) α-cell (bottom). Inset shows a magnification of the K_slow tail current. B: average α-cell K_slow currents (n ≥ 12 cells from 3 mice) with (red) and without (blue) extracellular Ca²⁺ (2 mM). C: area under the curve (AUC) of α-cell K_slow currents from B. Fast phase (0–2 τ_f s), slow phase (2–3 τ_s s), and total (0–3 s). D: average α-cell K_slow currents (n ≥ 15 cells from 3 mice) with vehicle (red) or agatoxin (100 nM; blue). E: AUC of α-cell K_slow currents from D. F: average α-cell K_slow currents (n ≥ 16 cells from 3 mice) with vehicle (red) or nifedipine (50 μM; blue). G: AUC of α-cell K_slow currents from F. H: average α-cell K_slow currents (n ≥ 13 cells from 3 mice) with vehicle (red) or thapsigargin (Tg; 2 μM; blue) at 1 mM glucose. I: AUC of α-cell K_slow currents from H. J: average α-cell K_slow currents (n ≥ 10 cells from 3 mice) with vehicle (red) or Tg (blue) at 11 mM glucose. K: AUC of α-cell K_slow currents from J. L: average α-cell K_slow currents (n ≥ 17 cells from 3 mice) with vehicle (red) or apamin (100 nM; blue). M: AUC of α-cell K_slow currents from L. N: average α-cell K_slow currents (n ≥ 18 cells from 3 mice) with vehicle (red) or iberiotoxin (IbTx; 100 nM; blue). O: AUC of α-cell K_slow currents from N: average K_slow tail currents were fit to a model of two-phase decay. Statistical analysis was conducted using unpaired two-tailed t-tests, and uncertainty is expressed as SE (*P < 0.05, **P < 0.01, and ***P < 0.001). n.s., not significant.

Fig. 2.

α-Cell Ca²⁺-activated K⁺ (K_slow) currents are also activated by Ca²⁺ influx resulting from a single membrane potential depolarization. A: overview of single action potential K_slow, α-protocol (top), and a representative K_slow current recorded from a single red fluorescent protein-expressing (α-RFP) α-cell (bottom). Inset shows a magnification of the K_slow tail current. B: average K_slow currents (n ≥ 7 cells from 4 mice) from α-cells treated with a vehicle control (black), and α-cells treated with agatoxin (green), thapsigargin (Tg; red), or isradipine (light blue; 10 μM). C: peak α-cell K_slow from B. D: average K_slow currents (n ≥ 15 cells from 4 mice) from α-cells treated with a vehicle control (black) and α-cells treated with apamin (green), IbTx (red), or apamin+IbTx (light blue). E: peak α-cell K_slow from D. F: average K_slow currents (n ≥ 11 cells from 4 mice) from α-cells treated with a vehicle control (black) and α-cells treated with UCL 1684 (green; 100 nM) or TRAM 34 (red; 100 nM). G: peak α-cell K_slow from F. Statistical analysis was conducted using unpaired two-tailed t-tests, and uncertainty is expressed as SE (*P < 0.05, **P < 0.01, and ***P < 0.001).

$$A=\left({\text{K}}_{\text{slow,\hspace{0.17em}max}}-\text{Plateau}\right)\times \text{\%\hspace{0.17em}fast}\times 0.01$$ (1A)

$$B=\left({\text{K}}_{\text{slow,\hspace{0.17em}max}}-\text{Plateau}\right)\times \left(100-\text{\%\hspace{0.17em}fast}\right)\times 0.01$$ (1B)

$${\text{K}}_{\text{slow}}=\text{Plateau}+A\times exp\left(-{\text{K}}_{\text{f}}\times t\right)+B\times exp\left(-{\text{K}}_{\text{s}}\times t\right)$$ (1C)

Fast- (τ_f; s) and slow-phase time constants (τ_s; s) were derived from the inverse of the fast- and slow-phase rate constants. Select kinetic parameters are presented in Tables 2–7. Area under the curve (AUC) was calculated for K_slow fast-phase [from t = 0 − (2 × τ_f) s], K_slow slow-phase (from t = (2 × τ_f) − 3 s), and for total K_slow (from t = 0 − 3 s). K_slow currents obtained using the K_{slow, α} inactivated more rapidly and were monophasic, thus K_{slow, max} was employed as a measure of the magnitude of α-cell K_slow. Negative K_slow AUC values were set to zero, as K_slow is an outward current.

Table 2.

α-Cell K_slow is activated by extracellular Ca²⁺

Parameter	2 mM Ca2+ (n = 12 cells)	0 mM Ca2+ (n = 13 cells)	P Value
Kslow, max, pA/pF	3.480 ± 0.116	2.661 ± 0.057	<0.001
Percent fast, %	64.82 ± 1.43	82.39 ± 2.07	<0.001
τf, s	0.044 ± 0.004	0.071 ± 0.005	<0.001
τs, s	1.614 ± 0.122	0.580 ± 0.079	<0.001

K_slow currents were recorded from tdRFP-positive α-cells in KRHB with 1 mM glucose (n ≥ 12 cells from 3 mice). Cells were incubated for 15 min before recording in KRHB without Ca²⁺. Statistical analysis was conducted using an unpaired two-tailed t-test, and uncertainty is expressed as SE. KRHB, Krebs-Ringer-HEPES buffer; K_{slow, max}, peak Ca²⁺-activated K⁺; tdRFP, tandem-dimer red fluorescent protein; τ_f, fast-phase time constant; τ_s, slow-phase time constant.

Table 7.

α-Cell K_slow is modulated by SK and BK channels

Parameter	Nontreated (n = 18 cells)	Apamin (n = 17 cells)	IbTx (n = 18 cells)	Nontreated vs. Apamin P Value	Nontreated vs. IbTx P Value	Apamin vs. IbTx P Value
Kslow, max, pA/pF	4.582 ± 0.073	3.714 ± 0.070	3.276 ± 0.076	<0.001	<0.001	<0.001
Percent fast, %	53.82 ± 0.91	73.93 ± 0.67	64.93 ± 1.10	<0.001	<0.001	<0.001
τf, s	0.117 ± 0.007	0.049 ± 0.002	0.065 ± 0.005	<0.001	<0.001	ns
τs, s	2.130 ± 0.131	1.809 ± 0.119	1.540 ± 0.101	ns	<0.01	ns

K_slow currents were recorded from tdRFP-positive α-cells in KRHB with 1 mM glucose (n ≥ 17 cells from 3 mice). Cells were incubated for 15 min before recording in the same KRHB supplemented with 100 nM apamin or 100 nM IbTx. Statistical analysis was conducted using a one-way ANOVA, and uncertainty is expressed as SE. BK, large-conductance Ca²⁺-activated K⁺; IbTx, iberiotoxin; KRHB, Krebs-Ringer-HEPES buffer; K_{slow, max}, peak Ca²⁺-activated K⁺; ns, not significant; SK, small-conductance Ca²⁺-activated K⁺; tdRFP, tandem-dimer red fluorescent protein; τ_f, fast-phase time constant; τ_s, slow-phase time constant.

Perforated-patch current-clamp α-cell V_m recording.

α-Cells within whole α-RFP islets were identified by tdRFP fluorescence and patched in KRHB-11mM at room temperature. Changes in α-cell V_m were monitored using a perforated-patch current-clamp technique as previously described (12, 13, 62, 64). Electrically active α-cells were allowed to stabilize for 5–10 min, perifused with KRHB-11mM for 5 min, then perifused with KRHB-11mM with 100 nM apamin for 10 min.

Cytosolic Ca²⁺ imaging.

Dispersed human or α-RFP mouse islet cells were cultured in islet media supplemented with 2 µM Fura-2-acetoxymethyl ester (AM) for 25 min at 37°C, 5% CO₂. The cells were then incubated in KRHB-1mM for 20 min at 37°C. Mouse α-cells were identified by tdRFP fluorescence; human islet cells were fixed on ice with 4% paraformaldehyde (Electron Microscopy Sciences; Hatfield, PA) for 20 min after imaging, and α-cells were identified by immunofluorescence [1:1,000 mouse anti-GCG (MABN238; Millipore; Burlington, MA) and 1:500 Cy3-conjugated donkey anti-mouse (cat. no. 715–166–150; Jackson Immunoresearch; West Grove, PA)]. Cells were perifused with KRHB at a flow of 2 ml/min at 37°C with the glucose concentrations and treatments indicated (Fig. 4). α-Cell cytosolic Ca²⁺ (Ca²⁺_c) was measured every 5 s as a ratio of Fura-2 AM emission at 340 nm and 380 nm (F₃₄₀/F₃₈₀) using a Nikon Eclipse TE2000-U microscope equipped with an epifluorescence illuminator (Sutter Instrument; Novato, CA), an HQ2 CCD camera (Photometrics Scientific; Tucson, AZ), and Nikon Elements software (Nikon; Melville, NY). Changes in α-cell Ca²⁺_c were quantified by calculating the AUC of Fura-2 AM responses.

Fig. 4.

A: average (n ≥ 60 cells from 3 mice) Fura-2 acetoxymethyl ester (AM) responses (F₃₄₀/F₃₈₀) of dispersed red fluorescent protein-expressing (α-RFP) α-cells to apamin (100 nM) at 1 mM (top) and 11 mM (bottom) glucose. The bars above the traces denote when stimuli are present. B: area under the curve (AUC) of mouse α-cell Fura-2 AM responses before (black) and after (white) the addition of apamin. C: average (n ≥ 99 cells from 3 mice) Fura-2 AM responses (F₃₄₀/F₃₈₀) of dispersed α-RFP α-cells to iberiotoxin (IbTx; 100 nM) at 1 mM (top) and 11 mM (bottom) glucose. D: AUC of mouse α-cell Fura-2 AM responses before (black) and after (white) the addition of IbTx. E: representative Fura-2 AM response (F₃₄₀/F₃₈₀) of dispersed human α-cells to apamin at 1 mM glucose. F: AUC (left) and average Fura-2 AM response (right) of human α-cells before (black) and after (white) the addition of apamin (n ≥ 56 cells from 4 donors). Human islet cells were fixed with 4% paraformaldehyde, and α-cells were identified by glucagon staining. Statistical analysis was conducted using paired two-tailed t-tests, and uncertainty is expressed as SE (*P < 0.05, ***P < 0.001).

Whole α-GCaMP3 islets were cultured in RPMI-1640 supplemented with 1 mM or 11 mM glucose for 20 min at 37°C, 5% CO₂ then perifused with KRHB with the indicated glucose concentrations and treatments (see figure legends) at a flow of 2 ml/min at 37°C during imaging. Alternatively, whole α-GCaMP3 islets were cultured for 30 min at 37°C, 5% CO₂ in KRHB with the indicated glucose concentrations and treatments (see figure legends) then imaged at 37°C under static conditions. Fluorescence emission at 488 nm was measured every 3 s as an indicator of α-cell Ca²⁺_c using a Nikon spinning disk confocal microscope equipped with a Yokogawa CSU-X1 spinning disk head and an Andor DU-897 EMCCD camera or a Zeiss LSM780 confocal microscope. As GCaMP3 is a single-wavelength Ca²⁺ probe, all data was normalized to minimum fluorescence intensity at 488 nm (F/F_min).

Hormone secretion assays.

Following isolation, mouse islets were allowed to recover overnight in islet media supplemented with 0.5 mg/ml BSA at 37°C, 5% CO₂. Human islets were allowed to recover for 4 h after reception in islet media supplemented with 0.5 mg/ml BSA at 37°C, 5% CO₂. The islets were next incubated for 1 h at 37°C, 5% CO₂ in DMEM with 0.5 mg/ml BSA, 0.5 mM CaCl₂, and 10 mM HEPES (DMEM*) supplemented with 10% FBS and 5.6 mM glucose, and then handpicked on ice into 24-well plates. For GCG secretion assays, 20 size-matched islets were placed in wells containing 450 µl of DMEM* supplemented with glucose concentrations and treatments as indicated in figure legends. For SST secretion assays, 40 size-matched islets were placed in wells containing 380 µl of DMEM* supplemented with the glucose concentrations and treatments indicated in Fig. 7. GCG secretion was measured over a 60-min period, and SST secretion was measured over a 90-min period at 37°C, 5% CO₂, after which the supernatants were collected, supplemented with protease inhibitor, and stored at −20°C until analyzed. GCG concentrations of islet secretion supernatants were analyzed by the Vanderbilt Hormone Assay and Analytical Services Core using Sigma-Aldrich GCG radioimmunoassay kits (GL-32K; guinea pig anti-GCG antibody 1032-K), and SST concentrations of islet secretion supernatants were determined using a Phoenix SST ELISA kit (CEK-060–03; Burlingame, CA), according to manufacturer instructions.

Fig. 7.

Small-conductance Ca²⁺-activated K⁺ channel inhibition reduces glucagon (GCG) secretion from mouse islets. A: GCG secretion from mouse islets treated with vehicle (H₂O; black) or apamin (200 nM; white) at 1 (n = 5 mice) and 11 mM glucose (n = 8 mice). B: somatostatin (SST) secretion from mouse islets treated with vehicle (black) or apamin (white) at 1 and 11 mM glucose (n = 3 mice). Statistical analysis was conducted using paired two-tailed t-tests, and uncertainty is expressed as SE (*P < 0.05 and **P < 0.01).

Data analysis.

All data are presented as mean values ± SE for the specified number of samples (n). Statistical significance between sample groups was determined using unpaired two-tailed t-tests, paired two-tailed t-tests, or one-way ANOVA as appropriate. Microsoft Excel was utilized for two-tailed t-tests, whereas one-way ANOVAs were completed using GraphPad Prism software. P < 0.05 was considered statistically significant.

RESULTS

α-Cell K_slow currents are activated by Ca²⁺ influx through VDCCs and Ca²⁺_ER release.

To determine if α-cells display K_slow currents, we measured tail currents from single α-cells following Ca²⁺ influx in response to a train of voltage ramps each resembling an AP (Fig. 1A) (12, 20, 21, 68) or to one voltage ramp replicating a single AP (Fig. 2A). Utilizing these voltage-clamp protocols, it was determined that mouse α-cells display a significant K_slow current (Figs. 1A and 2A) and that this current is greatly reduced when extracellular Ca²⁺ is removed (Fig. 1, B and C; Table 2). Removal of extracellular Ca²⁺ also reduced peak K_slow current (K_{slow, max}), accelerated fast- and slow-phase K_slow decay, and decreased slow-phase K_slow AUC (Table 2). The selective P/Q-type VDCC inhibitor agatoxin (200 nM) decreased K_{slow, max}, accelerated slow-phase K_slow decay, and decreased slow-phase K_slow AUC (Figs. 1, D and E and 2, B and C; Table 3). Interestingly, Ca²⁺ entry through P/Q-type VDCCs had a more profound effect on the slow-phase of K_slow, indicating that intracellular Ca²⁺ stores or Kv channels may contribute to the fast-phase of K_slow. The L-type VDCC blocker nifedipine was also employed to examine the role of L-type VDCCs in α-cell K_slow activation (Fig. 1, F and G); however, application of this compound totally abrogated outward tail currents. As nifedipine has been reported to inhibit K⁺ channels (39, 60), isradipine, a more selective inhibitor of L-type VDCCs, was employed (58), which also almost completely eliminated α-cell K_slow currents (Fig. 2, B and C). These results demonstrate that Ca²⁺ influx through both P/Q- and L-type VDCCs activates α-cell K_slow, and during a Ca²⁺ wave generated by repeated AP firing, this current can persist for up to several seconds.

Table 3.

α-Cell K_slow is activated by Ca²⁺ influx through P/Q-type VDCCs

Parameter	Nontreated (n = 16 cells)	Agatoxin (n = 15 cells)	P Value
Kslow, max, pA/pF	3.291 ± 0.085	1.964 ± 0.049	<0.001
Percent fast, %	67.11 ± 1.082	80.93 ± 1.017	<0.001
τf, s	0.056 ± 0.004	0.061 ± 0.004	ns
τs, s	1.857 ± 0.148	1.327 ± 0.165	<0.05

K_slow currents were recorded from tdRFP-positive α-cells in KRHB with 1 mM glucose (n ≥ 15 cells from 3 mice). Cells were incubated for 15 min before recording in KRHB supplemented with 100 nM agatoxin. Statistical analysis was conducted using an unpaired two-tailed t-test, and uncertainty is expressed as SE. KRHB, Krebs-Ringer-HEPES buffer; K_{slow, max}, peak Ca²⁺-activated K⁺; ns, not significant; tdRFP, tandem-dimer red fluorescent protein; τ_f, fast-phase time constant; τ_s, slow-phase time constant; VDCCs, voltage-dependent Ca²⁺ channels.

It has been previously reported that β-cell K_slow is activated by release of Ca²⁺ from Ca²⁺_ER stores (20). Thus, we next examined the contributions of Ca²⁺_ER release to α-cell K_slow currents. This was accomplished by comparing α-cell K_slow currents (DMSO vehicle-treated) to K_slow currents following Ca²⁺_ER depletion with the sarco/ER Ca²⁺-ATPase (SERCA) inhibitor Tg (2 µM) (19, 42). As Ca²⁺_ER storage is dependent on ATP energization of SERCAs and β-cell K_slow is glucose-sensitive, these studies were conducted at both 1 and 11 mM glucose (20, 32). When α-cell Ca²⁺_ER was depleted with Tg at 1 mM glucose, K_{slow, max} decreased along with fast- and slow-phase K_slow AUC, whereas slow-phase K_slow decayed more rapidly (Figs. 1, H and I and 2, B and C; Table 4). Tg-mediated depletion Ca²⁺_ER at 11 mM glucose had an analogous effect on fast-phase α-cell K_slow AUC but did not significantly affect slow-phase K_slow AUC (Fig. 1, J and K; Table 5). Interestingly, total α-cell K_slow currents were only modestly decreased at high (11 mM) glucose; K_{slow, max} was decreased, and slow-phase K_slow decay was accelerated when compared with K_slow currents recorded in 1 mM glucose (Table 6). These data suggest that low glucose conditions enhance α-cell Ca²⁺_ER release, leading to increased K_slow activity.

Table 4.

α-Cell K_slow is activated by release of Ca²⁺ from endoplasmic reticulum stores at 1 mM glucose

Parameter	Vehicle (n = 14 cells)	Tg (n = 15 cells)	P Value
Kslow, max, pA/pF	3.711 ± 0.112	0.915 ± 0.083	<0.001
Percent fast, %	46.29 ± 1.88	45.67 ± 5.03	ns
τf, s	0.047 ± 0.006	0.063 ± 0.031	ns
τs, s	1.360 ± 0.058	6.211 ± 1.086	<0.001

K_slow currents were recorded from tdRFP-positive α-cells in KRHB with 1 mM glucose (n ≥ 14 cells from 3 mice). Cells were incubated for 15 min before recording in KRHB supplemented with 2 µM Tg. Statistical analysis was conducted using an unpaired two-tailed t-test, and uncertainty is expressed as SE. KRHB, Krebs-Ringer-HEPES buffer; K_{slow, max}, peak Ca²⁺-activated K⁺; ns, not significant; tdRFP, tandem-dimer red fluorescent protein; Tg, thapsigargin; τ_f, fast-phase time constant; τ_s, slow-phase time constant.

Table 5.

α-cell K_slow is activated by release of Ca²⁺ from endoplasmic reticulum stores at 11 mM glucose

Parameter	Vehicle (n = 13 cells)	Tg (n = 10 cells)	P Value
Kslow, max, pA/pF	3.386 ± 0.097	1.329 ± 0.084	<0.001
Percent fast, %	54.01 ± 1.58	55.97 ± 2.84	ns
τf, s	0.034 ± 0.004	0.035 ± 0.007	ns
τs, s	0.975 ± 0.035	5.495 ± 0.335	<0.001

K_slow currents were recorded from tdRFP-positive α-cells in KRHB with 11 mM glucose (n ≥ 10 cells from 3 mice). Cells were incubated for 15 min before recording in KRHB supplemented with 2 µM Tg. Statistical analysis was conducted using an unpaired two-tailed t-test, and uncertainty is expressed as SE. KRHB, Krebs-Ringer-HEPES buffer; K_{slow, max}, peak Ca²⁺-activated K⁺; ns, not significant; tdRFP, tandem-dimer red fluorescent protein; Tg, thapsigargin; τ_f, fast-phase time constant; τ_s, slow-phase time constant.

Table 6.

α-Cell K_slow activity is decreased under high glucose conditions

Parameter	1 mM glucose (N = 14 cells)	11 mM glucose (N = 13 cells)	P Value
Kslow, max, pA/pF	3.711 ± 0.112	3.386 ± 0.097	<0.05
Percent fast, %	46.29 ± 1.88	54.01 ± 1.58	<0.01
τf, s	0.047 ± 0.006	0.034 ± 0.004	ns
τs, s	1.360 ± 0.058	0.975 ± 0.035	<0.001

K_slow currents were recorded from tdRFP-positive α-cells in KRHB with 1 mM or 11 mM glucose (n ≥ 13 cells from 3 mice). Cells were incubated for 15 min before recording. Statistical analysis was conducted using an unpaired two-tailed t-test, and uncertainty is expressed as SE. KRHB, Krebs-Ringer-HEPES buffer; K_{slow, max}, peak Ca²⁺-activated K⁺; ns, not significant; tdRFP, tandem-dimer red fluorescent protein; τ_f, fast-phase time constant; τ_s, slow-phase time constant.

SK, BK, and IK channels are molecular determinants of α-cell K_slow.

We next explored the identity of the K_Ca channels that contribute to α-cell K_slow currents. Inhibition of SK channels with 100 nM apamin decreased α-cell K_{slow, max}, accelerated fast-phase K_slow decay, and decreased fast- and slow-phase K_slow AUC (Figs. 1, L and M and 2, D and E; Table 7), whereas SK channel inhibition with 100 nM UCL 1684 decreased α-cell K_{slow, max} (Fig. 2, F and G). Similarly, BK channel inhibition with 100 nM IbTx decreased α-cell K_{slow, max}, accelerated fast- and slow-phase K_slow decay, and decreased fast- and slow-phase K_slow AUC (Figs. 1, N and O and 2, D and E; Table 7). Inhibition of IK channels with 100 nM TRAM 34 also decreased α-cell K_{slow, max} (Fig. 2, F and G). The effect of BK channel inhibition (33.1 ± 3.3% decrease) on α-cell K_slow was slightly larger than SK (21.1 ± 1.4% decrease) or IK (26.0 ± 1.5% decrease) channel inhibition. This may be due to the fact that in α-cells KCNMA1 (gene encoding BK channels) is more highly expressed than KCNN3 (gene encoding SK3 channels) or KCNN4 (gene encoding IK channels) (3, 14) and that BK channel unitary conductance (100–220 pS) is significantly greater than that of SK channels (2–20 pS) or IK channels (20–85 pS) (27, 38, 45). Interestingly, the combined effect of IbTx and apamin on α-cell K_{slow, max} was indistinguishable from IbTx alone (34.9 ± 3.1% decrease). This is consistent with previous reports that found that the effects of IbTx and apamin are not additive (44, 55). Thus, other K⁺ channels including IK channels are predicted to contribute significantly to α-cell K_slow currents.

SK channels modulate α-cell electrical excitability and Ca²⁺ handling.

As K_slow currents hyperpolarize V_m, partial inhibition of α-cell K_slow with apamin would be predicted to depolarize V_m. Indeed, apamin-mediated SK channel inhibition decreased α-cell AP amplitude (10.4 ± 1.9 mV decrease; Fig. 3, A and B; P < 0.01) and depolarized V_m (3.8 ± 0.7 mV; Fig. 3, A and C; P < 0.01) in whole α-RFP islets. Interestingly, SK channel inhibition caused a similar change in α-cell electrical excitability as inhibition of α-cell K_ATP activity (i.e., reduction of AP amplitude and V_m depolarization) (1). However, unlike glucose-mediated inhibition of α-cell K_ATP activity, selective inhibition of SK channels did not influence the frequency of AP firing or the time between periods of electrical activity. This is likely because α-cells express SK channels at a lower level than K_ATP channels and also because the unitary conductance of K_ATP channels (~50 pS) is larger than SK channels (2–20 pS) (3, 34, 45).

Fig. 3.

Small-conductance Ca²⁺-activated K⁺ (SK) channels increase α-cell action potential (AP) amplitude and hyperpolarize plateau membrane potential (V_m). A: representative V_m recording from a red fluorescent protein-expressing (α-RFP) α-cell within an islet cluster (10–20 cells) with 11 mM glucose. The bar over the recording denotes when apamin was present (100 nM). B: individual and average (n = 5 islet clusters) AP amplitudes recorded from α-RFP α-cells within islet clusters before (black) and after (white) the addition of apamin. C: individual and average (n = 5 islet clusters) plateau V_m recorded from α-RFP α-cells within islet clusters before (black) and after (white) the addition of apamin. Statistical analysis was conducted using paired two-tailed t-tests, and uncertainty is expressed as SE (**P < 0.01). tdRFP, tandem-dimer red fluorescent protein.

Electrical excitability triggers Ca²⁺ entry into α-cells through V_m depolarization-mediated activation of VDCCs. Therefore, the contributions of SK and BK channels to Ca²⁺ handling were examined in dispersed single α-cells. Mouse α-cell Ca²⁺_c was significantly increased at 1 mM glucose in response to inhibition of SK channels with apamin (11.2 ± 2.1 to 27.4 ± 2.9 AUC; Fig. 4, A and B; P < 0.001) or inhibition of BK channels with IbTx (9.1 ± 1.9 to 28.1 ± 3.4 AUC; Fig. 4, C and D; P < 0.001). At 11 mM glucose, SK channels did not regulate Ca²⁺ influx into mouse α-cells, whereas inhibition of BK channels with IbTx modestly increased Ca²⁺_c in mouse α-cells at 11 mM glucose (15.1 ± 2.1 to 21.7 ± 2.8 AUC; Fig. 4, C and D, P < 0.05). As apamin-mediated SK channel inhibition increased mouse α-cell Ca²⁺_c under low glucose conditions, its effect on dispersed single human α-cells was also examined at 1 mM glucose. Apamin-mediated SK channel inhibition significantly increased human α-cell Ca²⁺_c at 1 mM glucose (15.8 ± 2.8 to 28.8 ± 4.4; Fig. 4, E and F; P < 0.01). Thus, these data show that K_Ca channels play an important role in regulating α-cell Ca²⁺_c and suggests that K_slow may be more active under low glucose conditions.

As α-cell Ca²⁺ handling and GCG secretion are tightly regulated by paracrine signaling from neighboring β- and δ-cells, we assessed SK channel modulation of α-cell Ca²⁺ handling within intact islets (4, 17, 33). This was accomplished by utilizing islets isolated from transgenic mice with α-cell-specific expression of the genetically encoded GCaMP3 Ca²⁺ indicator (10). At 1 mM glucose, SK channel inhibition increased total α-cell Ca²⁺_c (42.9 ± 10.6% increase; Fig. 5, A and C; P < 0.01) and maximum Ca²⁺ oscillation amplitude (28.0 ± 5.4% increase; Fig. 5, A and D; P < 0.001) relative to vehicle-treated islets, whereas at 11 mM glucose, SK channel inhibition did not affect total α-cell Ca²⁺ _c but slightly increased maximum Ca²⁺ oscillation amplitude (14.2 ± 4.4% increase; Fig. 5, B and D; P < 0.05) relative to vehicle-treated islets. Short-term (20 min) exposure to elevated glucose also modestly increased total α-cell Ca²⁺_c compared with islets under low glucose conditions (35.3 ± 13.9% increase; Fig. 5, A–C; P < 0.05). This indicates that SK channel inhibition increases α-cell Ca²⁺_c in a glucose-dependent manner, which may modulate secretagogue-stimulated GCG secretion.

Fig. 5.

Small-conductance Ca²⁺-activated K⁺ (SK) channel inhibition transiently increases α-cell cytosolic Ca²⁺ (Ca²⁺_c) within whole islets. A: representative α-GCaMP3 islet α-cell responses to apamin (100 nM; F/F_min) at 1 mM glucose. The bars above the traces denote when apamin was present. B: representative α-GCaMP3 islet α-cell responses to apamin at 11 mM. C: area under the curve (AUC) of α-GCaMP3 islet α-cell responses relative to nontreated α-cells at 1 mM glucose (n ≥ 147 cells from 5 mice) before (black; from 0 s to 600 s) and after (white; from 1,000 s to 1,600 s) the addition of apamin. D: maximum amplitude of α-GCaMP3 islet α-cell responses relative to nontreated α-cells at 1 mM glucose (n ≥ 147 cells from 5 mice) before (black) and after (white) the addition of apamin. Statistical analysis was conducted using one-way ANOVA, and uncertainty is expressed as SE (*P < 0.05, **P < 0.01, and ***P < 0.001). n.s., not significant.

In dispersed single α-cells, SK and BK channel inhibition only transiently increased Ca²⁺_c; therefore, we next examined the consequences of sustained channel inhibition on α-cell Ca²⁺_c within intact islets. When islets were pretreated for 30 min with apamin at 1 mM glucose before the start of imaging and compared with vehicle-treated islets (Fig. 6, A–D), total α-cell Ca²⁺_c decreased (38.5 ± 10.4% decrease; Fig. 6E; P < 0.001) along with maximum Ca²⁺ oscillation amplitude (15.7 ± 5.0% decrease; Fig. 6F; P < 0.01). This reduction in Ca²⁺_c was similar to the decrease in total α-cell Ca²⁺_c (28.3 ± 11.1% decrease; Fig. 6E; P < 0.05) and maximum Ca²⁺_c oscillation amplitude (17.7 ± 4.6%; Fig. 6F; P < 0.01) observed when islets were imaged following a 30-min incubation at 11 mM glucose (compared with those recorded under the same conditions but cultured at 1 mM glucose). Moreover, apamin did not further reduce α-cell Ca²⁺_c under high glucose conditions following a 30-min pretreatment at 11 mM glucose. Similarly, a 30-min pretreatment of islets with IbTx at 1 mM glucose before imaging (Fig. 6, G–J) decreased total Ca²⁺_c compared with vehicle-treated islets (31.0 ± 11.7% decrease; Fig. 6K; P < 0.05) but did not alter maximum Ca²⁺ oscillation amplitude. These results suggest that sustained SK or BK channel inhibition eventually decreases α-cell Ca²⁺_c. As SK or BK channel inhibition at 11 mM glucose caused no further reduction in α-cell Ca²⁺_c, it is likely that channel activity is reduced along with α-cell Ca²⁺_c under hyperglycemic conditions.

Fig. 6.

Sustained small-conductance Ca²⁺-activated K⁺ or large-conductance Ca²⁺-activated K⁺ channel inhibition decreases α-cell cytosolic Ca²⁺ (Ca²⁺_c) within whole islets. Representative α-GCaMP3 islet α-cell responses following a 30-min pretreatment with vehicle (H₂O; A) or apamin (B) at 1 mM glucose. Representative α-GCaMP3 islet α-cell responses following a 30-min pretreatment with vehicle (C) or apamin (D) at 11 mM glucose. Area under the curve (AUC; E) and (F) maximum amplitude of α-GCaMP3 islet α-cell apamin responses relative to vehicle-treated controls at 1 mM glucose (n ≥ 39 cells from 3 mice). Representative α-GCaMP3 islet α-cell responses following a 30-min pretreatment with vehicle (H₂O; G) or iberiotoxin (IbTx; H) at 1 mM glucose. Representative α-GCaMP3 islet α-cell responses following a 30-min pretreatment with vehicle (I) or IbTx (J) at 11 mM glucose. AUC (K) and maximum amplitude (L) of α-GCaMP3 islet α-cell IbTx responses relative to vehicle-treated controls at 1 mM glucose (n ≥ 59 cells from 6 mice). Statistical analysis was conducted using one-way ANOVA, and uncertainty is expressed as SE (*P < 0.05, **P < 0.01, and ***P < 0.001). n.s., not significant.

SK channels regulate GCG secretion.

It has been demonstrated that Ca²⁺ signaling in α-cells modulates GCG secretion (4, 10, 59, 67). Furthermore, a previous report determined that IbTx-mediated BK channel inhibition decreases islet GCG secretion (57). Thus, we went on to assess the impact of SK channel inhibition on GCG secretion from mouse islets. SK channel inhibition significantly reduced GCG secretion from mouse islets at both 1 mM glucose (4.0 ± 0.5 pg·islet⁻¹·h⁻¹ to 3.1 ± 0.3 pg·islet⁻¹·h⁻¹; Fig. 7A; P < 0.05) and 11 mM glucose (0.54 ± 0.07 pg·islet⁻¹·h⁻¹ to 0.30 ± 0.05 pg·islet⁻¹·h⁻¹; Fig. 7A; P < 0.05). Although apamin treatment reduced GCG secretion at 11 mM glucose, SK channel inhibition had no significant impact on α-cell Ca²⁺_c. This suggests that SK channel modulation of GCG secretion from mouse islets under high glucose conditions could be due to either paracrine effects or a switch from P/Q-type to L-type VDCC activity (11).

SST secreted from pancreatic δ-cells is a potent inhibitor of GCG secretion (4, 30, 59, 64), and δ-cells express SK channel transcripts (KCNN1, KCNN2, and KCNN3) (3, 14). Thus, we investigated whether the inhibitory effect of apamin on GCG secretion is due in part to a δ-cell SK channel-mediated increase in SST secretion. Apamin had no effect on SST secretion at 1 mM glucose, but it significantly decreased SST secretion at 11 mM glucose (0.95 ± 0.09 pg·islet⁻¹·h⁻¹ to 0.51 ± 0.03 pg·islet⁻¹·h⁻¹; Fig. 7B; P < 0.01). It is unclear why apamin decreased SST secretion at 11 mM glucose; however, as reduced SST would amplify GCG secretion, these results suggest that apamin reduces GCG secretion through its action on α-cell SK channels.

As apamin mediated a transient rise in human α-cell Ca²⁺_c under low glucose conditions (Fig. 4, E and F), SK channel modulation of human islet GCG secretion was also investigated. At 1 mM glucose, apamin treatment led to a significant decrease in GCG secretion (20.5 ± 5.6 pg·islet⁻¹·h⁻¹ to 14.7 ± 4.5 pg·islet⁻¹·h⁻¹; Fig. 8, A and C; P < 0.05); however, at 11 mM glucose, apamin had no effect on GCG secretion (Fig. 8, B and D). These data indicate an important role for SK channels in human α-cells in which they promote GCG secretion during sustained secretagogue stimulation, presumably by reducing VDI of P/Q-type VDCCs.

Fig. 8.

Small-conductance Ca²⁺-activated K⁺ channel inhibition reduces glucagon (GCG) secretion from human islets. A: GCG secretion from individual human islet preparations treated with vehicle (H₂O; black) or apamin (200 nM; white) at 1 mM glucose (n = 7 islet donors). B: GCG secretion from individual human islet preparations treated with vehicle or apamin at 11 mM glucose (n = 7 islet donors). C: average GCG secretion from human islet preparations at 1 mM glucose (from A). D: average GCG secretion from human islet preparations at 11 mM glucose (from B). Statistical analysis was conducted using paired two-tailed t-tests, and uncertainty is expressed as SE (*P < 0.05).

DISCUSSION

GCG secretion is critical for the maintenance of glucose homeostasis; however, our understanding of α-cell function remains incomplete. It has been known for several decades that Ca²⁺ entry into α-cells is required for GCG secretion (18), but the factors that orchestrate α-cell Ca²⁺ handling have not been fully elucidated. Here, we show for the first time, to our knowledge, that K_Ca channels contribute to α-cell K_slow currents, which are key regulators of α-cell Ca²⁺ handling and GCG secretion. We found that α-cell K_slow currents are mediated by a combination of SK, BK, and IK channel activity. Furthermore, SK channel inhibition depolarized α-cell V_m transiently elevating Ca²⁺_c, which was followed by a sustained decrease in Ca²⁺_c and GCG secretion. These data suggest that K_slow currents provide a hyperpolarizing influence on α-cell V_m that sustains Ca²⁺ entry during extended periods of hypoglycemia, presumably by limiting VDI of P/Q-type VDCCs. Thus, when α-cell Ca²⁺_c is elevated during secretagogue stimulation, activation of K_slow enhances GCG secretion.

Our findings reveal that K_Ca channels contribute to α-cell K_slow and, like other α-cell K⁺ channels, enhance GCG secretion by facilitating Ca²⁺ influx through P/Q-type VDCCs (9, 10, 57). As anticipated, α-cell K_slow currents augmented V_m hyperpolarization by K_ATP channels (1), as partial inhibition of K_slow currents depolarized α-cell V_m. This hyperpolarizing influence should decrease VDI of P/Q-type VDCCs, which would promote Ca²⁺ entry (50). Indeed, although α-cell Ca²⁺_c initially increased when SK channels were inhibited at 1 mM glucose, consistent with VDI of P/Q-type VDCCs, over time Ca²⁺_c decreased. Sustained BK channel inhibition with IbTx also led to a decrease in α-cell Ca²⁺_c. In both cases, this decrease in Ca²⁺_c was glucose dependent, as it was only observed under low glucose conditions. It is probable that under high glucose conditions, K_ATP channel inhibition significantly increases VDI of P/Q-type VDCCs and thus decreases Ca²⁺ influx independently of SK or BK channel activity. This is further supported by the observation that prolonged (30 min) exposure of islets to 11 mM glucose resulted in a reduction in Ca²⁺_c similar to SK or BK channel inhibition at 1 mM glucose. Although SK channel inhibition did not further reduce α-cell Ca²⁺_c at 11 mM glucose, it significantly decreased GCG secretion from mouse islets. In mouse α-cells, P/Q-type VDCC currents account for a small portion of total α-cell transmembrane Ca²⁺ currents (~20%), and the majority of these channels would be expected to be inactive under high glucose conditions because of VDI (22). Thus, under hyperglycemic conditions, SK channel inhibition does not significantly alter bulk Ca²⁺_c, because the change in overall Ca²⁺ entry because of VDI of P/Q-type VDCCs is minimal compared with Ca²⁺ influx through L-type VDCCs (11). However, it is possible that at high glucose SK channel inhibition does decrease P/Q-type VDCC activity further, and this small change in local Ca²⁺_c near GCG granules reduces GCG secretion (50). These results demonstrate the importance of K_Ca channel hyperpolarization of α-cell V_m in maintaining P/Q-type VDCC activity, which enhances GCG secretion.

In α-cells, Ca²⁺_c is modulated by Ca²⁺ release from intracellular Ca²⁺ stores, which could also tune K_slow currents (53). Our findings show that α-cell K_slow currents are stimulated by Ca²⁺_ER release, as depletion of Ca²⁺_ER significantly reduced α-cell K_slow currents. Moreover, glucose metabolism modulates Ca²⁺_ER handling and thus would presumably regulate α-cell K_slow activity. For example, energization of α-cell SERCAs under high glucose conditions increases the ratio of Ca²⁺_ER uptake to Ca²⁺_ER release and thus reduces the influence of Ca²⁺_ER leak on α-cell K_slow activity, whereas reduced SERCA activity under hypoglycemic conditions would decrease the ratio of Ca²⁺_ER uptake to Ca²⁺_ER release, leading to a more prominent role for Ca²⁺_ER in elevating Ca²⁺_c and activating α-cell K_slow. Depletion of Ca²⁺_ER stores under low glucose conditions also facilitates translocation of stromal interaction molecule 1 to the plasma membrane activating store-operated Ca²⁺ entry (SOCE) through Orai1 channels (61). Interestingly, although the fast component of K_slow was Ca²⁺_ER-dependent, it was not affected by glucose concentration. This is likely because the fast component of α-cell K_slow is activated predominantly by Ca²⁺-induced Ca²⁺ release from Ca²⁺_ER stores. Ca²⁺-induced Ca²⁺ release is a rapid phenomenon mediated by the opening of Ca²⁺_ER release channels (e.g., ryanodine receptors and IP3 receptors) and is supported by a large ionic driving force (Ca²⁺_ER is ~10,000× greater than Ca²⁺_c); therefore, it is unlikely that slow changes in Ca²⁺_c mediated by SERCA activity influence the process (2, 6, 16, 28). However, Ca²⁺_ER leak and SOCE occur on a timescale that is compatible with the slow component of α-cell K_slow (51, 66). Thus, the data suggest that glucose-modulated mobilization of intracellular Ca²⁺ stores plays a significant role in tuning α-cell K_slow currents, which would be expected to influence glucose regulation of GCG secretion.

In addition to changes in bulk Ca²⁺_c, K_Ca channel activity is regulated by Ca²⁺ microdomains generated near Ca²⁺ permeable ion channels (70). The site of Ca²⁺ entry is also of great consequence to α-cell GCG secretion. For example, P/Q-type VDCCs concentrate to areas near exocytosis machinery and thus facilitate a localized rise in Ca²⁺_c near GCG granules, which promotes granule exocytosis (50). Interestingly, P/Q-type VDCCs colocalize with and activate both BK and SK channels in cerebellar Purkinje cells, which play a role in controlling the Ca²⁺ microdomains created around these VDCCs (26). Therefore, α-cell P/Q-type VDCCs and K_Ca (i.e., BK and SK) channels may work in concert to fine tune Ca²⁺_c in the vicinity of GCG granules. Furthermore, SK channels are located near and activated by Ca²⁺_ER release through ER-localized ryanodine receptors (31). Similar cellular architecture could exist in α-cells in which K_Ca channels are regulated by P/Q-type VDCCs on the plasma membrane as well as Ca²⁺ release channels on the ER. SK channels also interact with and are activated by SOCE through Orai1 channels (8). Therefore, it is also possible that activation of Orai1 by stromal interaction molecule 1 under hypoglycemic conditions further stimulates K_Ca channel activity. Taken together, this suggests that α-cell K_Ca channels reside near Ca²⁺ permeable ion channels, in which they would be predicted to modulate local V_m around P/Q-type VDCCs. Indeed, this may contribute to the decreased GCG secretion observed under high glucose conditions during sustained SK channel inhibition that occurs without a significant change in bulk Ca²⁺_c. In the future, it will be important to identify the intracellular localization of α-cell K_Ca channels in relation to plasmalemmal and ER-localized Ca²⁺ permeable channels as well as explore the impact of α-cell K_slow on Ca²⁺ microdomains in close proximity to GCG granules.

Although α-cell Ca²⁺_c is regulated in part by intrinsic mechanisms, it is well established that α-cell Ca²⁺ handling is also modulated by paracrine signals from β- and δ-cells (4, 23, 24, 53, 54, 64). Therefore, paracrine regulation of α-cell Ca²⁺_c would also modulate K_slow currents. Glucose-stimulated secretion of SST from δ-cells diminishes α-cell Ca²⁺_c by activating hyperpolarizing GIRK channels (30, 64). Furthermore, GIRK activation can lead to spontaneous V_m hyperpolarization, which would influence α-cell Ca²⁺ oscillations (4). Decreased SST secretion under hypoglycemic conditions elevates Ca²⁺_c by reducing α-cell GIRK channel activity, which would activate K_slow currents, whereas under high glucose conditions paracrine signaling limits Ca²⁺ influx into α-cells, which would decrease K_slow activity and reduce GCG secretion (17, 33, 52). This is supported by the finding that SK channels predominantly regulate α-cell Ca²⁺_c and GCG secretion under hypoglycemic conditions in both mouse and human islets. Interestingly, in the presence of 1 mM glucose, both intact and dispersed mouse islet preparations displayed a transient increase in α-cell Ca²⁺_c following SK channel inhibition. Although this suggests that paracrine signaling is not the sole cause for the glucose dependence of SK channel function in α-cells, paracrine factors would still be predicted to influence total α-cell SK channel conductance. Furthermore, SK channel expression in other islet cell types (e.g., δ-cells) could affect hormone secretion and thus α-cell paracrine signaling (14). Indeed, apamin reduced SST secretion in the presence of 11 mM glucose; however, SK channel inhibition also inhibited GCG secretion at 1 mM glucose, whereas apamin did not impact SST secretion. Further studies will investigate how islet paracrine signals that modulate α-cell V_m and/or Ca²⁺_c influence K_slow activity and GCG secretion as well as how K_slow and GIRK currents work together to orchestrate α-cell electrical excitability.

In conclusion, our data indicate that α-cell K_slow currents are enhanced by activation of SK, BK, and IK channels, which augment α-cell Ca²⁺_c and GCG secretion, presumably by preventing VDI of P/Q-type VDCCs. The α-cell K_slow currents are activated by Ca²⁺ influx through P/Q- and L-type VDCCs as well as Ca²⁺ released from Ca²⁺_ER stores. These results suggest that K_slow currents play a prominent role in maintaining α-cell P/Q-type VDCC activity under low glucose conditions in which elevated α-cell Ca²⁺_c sustains K_Ca channel activity, whereas under high glucose conditions, α-cell Ca²⁺_c decreases because of a combination of intrinsic and paracrine-mediated effects, which diminishes α-cell K_slow currents. These findings show for the first time, to our knowledge, that K_slow plays an important role in tuning α-cell V_m and VDCC activity during hypoglycemic conditions, which amplifies GCG secretion.

GRANTS

M. T. Dickerson was supported by the Vanderbilt Integrated Training in Engineering and Diabetes Grant T32DK101003 and NIH Grant DK-097392. In addition, research in the laboratory of D. A. Jacobson was supported by NIH Grant DK-081666, NIH Grant DK115620, ADA Grant 1-17-IBS-024, and a Vanderbilt University Diabetes Research Training Center Pilot and Feasibility Grant P60-DK-20593. Hormone immunoassays were carried out by the Vanderbilt Hormone Assay Core (supported by NIH Grants DK-059637 and DK-020593). Confocal microscopy was performed using the Vanderbilt Cell Imaging Shared Resource (supported by NIH Grant DK-020593).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

M.T.D. and D.A.J. conceived and designed research; M.T.D., P.K.D., M.K.A., K.R.V., A.S.T., K.L.J., N.C.V., G.A., and D.A.J. performed experiments; M.T.D., P.K.D., M.K.A., K.R.V., A.S.T., K.L.J., N.C.V., G.A., and D.A.J. analyzed data; M.T.D. and D.A.J. interpreted results of experiments; M.T.D. and D.A.J. prepared figures; M.T.D. and D.A.J. drafted manuscript; M.T.D. and D.A.J. edited and revised manuscript; M.T.D., P.K.D., M.K.A., K.R.V., A.S.T., K.L.J., N.C.V., G.A., and D.A.J. approved final version of manuscript.