Role of KATP channels in glucose-regulated glucagon secretion and impaired counterregulation in type 2 diabetes

Abstract

Glucagon, secreted by pancreatic islet α cells, is the principal hyperglycemic hormone. In diabetes, glucagon secretion is not suppressed at high glucose, exacerbating the consequences of insufficient insulin secretion, and is inadequate at low glucose, potentially leading to fatal hypoglycemia. The causal mechanisms remain unknown. Here we show that α cell KATP-channel activity is very low under hypoglycemic conditions and that hyperglycemia, via elevated intracellular ATP/ADP, leads to complete inhibition. This produces membrane depolarization and voltage-dependent inactivation of the Na(+) channels involved in action potential firing that, via reduced action potential height and Ca(2+) entry, suppresses glucagon secretion. Maneuvers that increase KATP channel activity, such as metabolic inhibition, mimic the glucagon secretory defects associated with diabetes. Low concentrations of the KATP channel blocker tolbutamide partially restore glucose-regulated glucagon secretion in islets from type 2 diabetic organ donors. These data suggest that impaired metabolic control of the KATP channels underlies the defective glucose regulation of glucagon secretion in type 2 diabetes.

Copyright © 2013 The Authors. Published by Elsevier Inc. All rights reserved.

Figures

Graphical abstract

Figure 1

Regulation of Glucagon Secretion by Glucose and Tolbutamide (A) Glucagon secretion from mouse islets at 1 mM and 6 mM glucose in the absence and presence of 10 mM mannoheptulose (MH, n = 6). ∗p < 0.05 versus 1 mM glucose; †p < 0.05 versus 6 mM glucose without mannoheptulose. (B) (Left) Representative traces of the increase in [ATP/ADP]cyt measured as Δ(F/F0) recorded from individual α cells within the islet, in response to 6 mM glucose and 2 μM FCCP. The gray line represents parallel measurements of tdRFP fluorescence in the α cell. The data are representative of 44 α cells in 9 intact islets from 3 different animals. (Right) Confocal images of Perceval (green), tdRFP (red), and the combination of the two (yellow). Scale bar, 200 μm. (C) Glucagon secretion from muse islets in the absence or presence of CYN154806 (which blocks SSTR2, the main somatostatin receptor subtype in α cells; Yue et al., 2012) at 1 mM or 6 mM glucose (left) or with and without tolbutamide (0.1 mM, right; n = 6–7). ∗p < 0.05 or ∗∗p < 0.01 versus 1 mM glucose; ††p < 0.01 versus 6 mM glucose without CYN154806. (D) Insulin and glucagon secretion were measured using the perfused pancreas preparation. Data are normalized to the maximal insulin or glucagon secretion in each mouse (n = 5–7). For clarity, significances are not indicated, but effects of glucose and tolbutamide on insulin and glucagon secretion were statistically significant (∗p < 0.05 or better). Error bars indicate SEM.

Figure 2

Effects of Glucose, Tolbutamide, and Diazoxide on Electrical Activity in α Cells (A–C) (Left) Electrical activity in mouse α cells (identified by action potential firing at 1 mM glucose) at 1 mM and 6 mM glucose (A, n = 9), 200 μM tolbutamide (B, n = 5), or 1 or 100 μM diazoxide (C, n = 5). (Right) Averaged action potentials recorded around the indicated time points, shown on an expanded timescale (averages of 308 and 181 spikes at 1 and 6 mM glucose in A; 54 and 223 in the absence and presence of tolbutamide in B; and 277, 560, and 190 at 1 mM and 6 mM glucose, and 6mM glucose + 1 μM diazoxide in C). See also Figures S1 and S2 and Table S1.

Figure 3

Modulation of KATP Channel Activity by Glucose in Mouse Pancreatic α Cells (A) (Left) Perforated patch whole-cell currents in mouse α cells evoked by ±10 mV voltage pulses (indicated schematically above the current traces) at 1 mM glucose (black) and 6 mM glucose (red, +glc.) or 1 mM glucose plus 100 μM tolbutamide (red, +tolb.). Histograms (right) show mean whole-cell KATP-currents (expressed as membrane conductance) in glucose and tolbutamide, as indicated. ∗p < 0.05 versus 1 mM glucose. (B) Dose-response curve for the effects of increasing concentrations of diazoxide (1–300 μM) on the whole-cell KATP-channel activity (expressed as membrane conductance) in α cells within intact islets. ∗p < 0.05 versus no diazoxide for concentrations >1 μM (not indicated for clarity). (Inset) Perforated patch whole-cell currents in mouse α cells evoked by ±10 mV voltage pulses at 6 mM glucose (black) and 6 mM glucose + 3 μM diazoxide (red, +diaz.). (C) Relationship between glucagon secretion and mean whole-cell conductance (G, perforated patch) in mouse α cells at 6 mM glucose and the indicated diazoxide concentrations (black). Red square, glucagon secretion and G at 1 mM glucose. ∗∗∗p < 0.001 versus no diazoxide. (D) Effects of glucose and diazoxide on glucagon secretion in the absence and presence of 2–6 mM of amino acid mixtures (see Table S1). Glucose and diazoxide were also added to the medium, as indicated. n = 8–9 experiments. ∗∗p < 0.01 for effects of 1 mM glucose versus 6 mM glucose (at 0, 2, or 6 mM AAM); ‡p < 0.05 for effect of diazoxide versus no diazoxide at 6 mM glucose in the presence of 2 mM AAM; ¶p < 0.05 for effect of diazoxide at 1 mM glucose in the presence of 6 mM AAM. Error bars indicate SEM. See also Figure S2 and Table S1.

Figure 4

Impact of Membrane Potential on Action Potential Amplitude (A) Electrical activity in mouse α cells hyperpolarized to ∼−80 mV by injection of −5 pA and stimulated by 5 s current pulses to different membrane potentials applied at a frequency of 0.2 Hz. (B) Individual spikes from (A), as indicated by asterisks, shown on an expanded timescale. (C) Relationship between interspike membrane potential and action potential peak voltage (Vp). The amplitude of each action potential and the associated interspike voltage were measured. For display, data were binned according to the most negative interspike potential (bin width, 5 mV) and averaged (n = 7–12 action potentials for each data point). The line is a Boltzmann fit to the mean data with a midpoint of −52 mV. The red arrow indicates the decrease in peak voltage predicted from the 9 mV glucose-induced depolarization. (D) Voltage-dependent inactivation of Na+-current in identified α cells in intact islets. Peak currents during a 1 ms test pulses to 0 mV are displayed against the membrane potential during the preceding 200 ms conditioning pulses (frequency, 1 Hz). Current responses are presented as h∞ (= I/Imax). The response following a conditioning pulse to −100 mV was taken as unity (n = 6). The midpoint of inactivation was −47 mV. Arrow indicates the decrease in h∞ predicted from the glucose-induced depolarization. (Inset) Na+-currents during depolarizations to 0 mV following conditioning pulses to −100 mV, −55 mV, and −45 mV. See also Table S1. (E) Examples of action potentials recorded from an α cell in an intact islet exposed to 1 mM glucose under control conditions, after addition of TTX (0.1 μg/ml) and following washout of the blocker (ten individual spikes averaged for each experimental condition). (F) Effects of TTX on glucagon secretion measured from isolated NMRI mouse islets at 1 mM glucose in the absence and presence of TTX (as indicated). ∗p < 0.05 versus 1 mM glucose without TTX. (G) Absolute (left) and relative (right) glucagon secretion from mouse islets at 1 mM glucose and different [K+]o (n = 6). Relative release is normalized to that at 3.6 mM [K+]o). The red arrow indicates the equivalent decrease (47% ± 4%) in glucagon secretion produced by increasing glucose from 1 mM to 6 mM in control experiments (at 3.6 mM [K]o; seven different experimental series performed over several years). (Inset) Stimulation of glucagon secretion by 70 mM [K+]o. ∗p < 0.01 and ∗∗∗p < 0.001 versus 3.6 mM [K+]o. Error bars indicate SEM.

Figure 5

Ca2+ Channel Activation and Exocytosis in α Cells (A) Voltage dependence of α cell exocytosis (n = 5) was measured as changes in membrane capacitance (ΔCm) in response to 500 ms voltage pulses (holding potential −70 mV) using the standard whole-cell configuration. ΔCm was normalized to the largest response in each cell (ΔCmax). The red arrow indicates the 75% suppression of exocytosis predicted to result from a 9 mV reduction of the action potential. (B) Glucagon secretion from islets exposed to glucose and ω-agatoxin, as indicated (n = 6). ∗∗p < 0.01 versus 1 mM glucose under control conditions; †††p < 0.001 versus 1 mM glucose under control conditions. (C) ΔCm recorded during 500 ms depolarizations to 0 mV in α cells with/without 200 nM ω-agatoxin (n = 4). ∗p < 0.05 versus no ω-agatoxin. (D) ω-agatoxin-sensitive Ca2+-currents (lower) evoked by voltage-clamp commands based on the averages of 20 action potentials recorded in the presence of 1 and 6 mM glucose (top). (E and F) ΔCm recorded during 50 and 300 ms depolarizations to 0 mV with intracellular EGTA included in the pipette medium at 50 μM (E) or 1 mM (F). (G) ΔCm recorded in the perforated patch whole-cell configuration at 1 and 6 mM glucose in response to 500 ms voltage pulses from −70 mV to 0 mV (n = 10). Error bars indicate SEM. See also Figure S3 and Table S1.

Figure 6

Effects of the Kir6.2-V59M Mutation on Glucagon Secretion (A and B) Whole-cell resting conductance (Gr) measured with the perforated patch method in α cells within control (A; n = 6–7) and α-V59M islets (B; n = 6) at the indicated concentrations (mM) of glucose and tolbutamide. ∗p < 0.05 versus 1 mM glucose. †p < 0.05 versus 1 mM glucose in control α cells. Inset in (A) shows conductance changes on an expanded vertical axis. (C) Electrical activity in α cells in α-V59M islets at the indicated glucose (black bar) and tolbutamide (red bar) concentrations. Examples showing α cells (i) refractory to glucose but responding to tolbutamide and (ii) stimulated by glucose and where tolbutamide reduced spike height. (D) Glucagon secretion from control (left and black; n = 15–21 from 8 mice) and α-V59M islets (right and red; n = 26 from 12 mice) at the indicated glucose and tolbutamide concentrations (in mM). ∗∗∗p < 0.001 versus 1 mM glucose (control islets), ††p < 0.01 versus 1 mM glucose (α-V59M islets), and ‡p < 0.05 versus 6 mM glucose in control islets. (E) Effects of adrenaline (adr.; 5 μM) on glucagon secretion at 1 mM glucose in wild-type type (left; n = 6) and α-V59M islets (right; n = 9). ∗p < 0.05 and ∗∗p < 0.01 versus 1 mM glucose (for comparisons within strain). (F) As in (E), but testing the effects of membrane depolarization produced by increasing extracellular K+ to 70 mM. ∗p < 0.05 and ∗∗p < 0.01 versus 3.6 mM [K+]o (for comparisons within strain) and ‡p < 0.05 versus 70 mM [K+]o in control islets. Error bars indicate SEM. See also Figure S4 and Table S1.

Figure 7

Glucagon Secretion and Type 2 Diabetes (A) Glucagon secretion at 1 and 6 mM glucose in islets from human organ donors without (ND) and with known T2D normalized to glucagon content. White bars show glucagon secretion in the five islet preparations where glucose stimulated glucagon release. ∗∗∗p < 0.001 versus 6 mM glucose; †p < 0.02 versus 1 mM glucose in five T2D preparations with no glucose-induced inhibition of glucagon secretion. (B) Glucagon secretion in the absence (black) and presence (red) of 2 μM diazoxide in ND human islets (n = 7 experiments, using islets from two donors). ∗p < 0.05 versus 1 mM glucose. †p < 0.05 versus 1 mM glucose in the presence of diazoxide. (C) Glucagon secretion from islets of nondiabetic human donors homozygous for the high-risk (TT; n = 6) and low-risk (CC; n = 8) Kir6.2 (KCNJ11) alleles. ∗p < 0.05 and ∗∗p < 0.01 versus 1 mM glucose. (D) Glucagon secretion from mouse islets at 1 mM and 6 mM glucose in the absence (black) and presence (red) of 0.5 μM oligomycin (n = 8–16). Tolbutamide 10 μM was added as indicated. ∗∗∗p < 0.001 versus 1 mM glucose; †††p < 0.001 versus 1 mM glucose plus oligomycin; ‡‡p < 0.01 versus 1 mM glucose without oligomycin; §p < 0.05 versus 1 mM glucose with oligomycin (no tolbutamide); ¶p < 0.05 versus 1 mM glucose plus oligomycin (with tolbutamide). (E) Effects of glucose (1 or 6 mM) on glucagon secretion normalized to glucagon content in islets from T2D donors in the absence (left) and presence (right; n = 4) of tolbutamide (10 μM). ∗p < 0.05 versus 1 mM glucose. (F) Relative effect (%) of elevating glucose from 1 to 6 mM on glucagon secretion in ND (black) and T2D islets (color coded as in A) and in four T2D preparations in the presence of tolbutamide (far right). Significance between 1 mM and 6 mM glucose is indicated. ∗∗∗p < 0.001, ∗∗p < 0.02, ∗p < 0.05. Error bars indicate SEM. See also Figures S5 and S6 and Table S1.