Gamma-aminobutyric acid (GABA) is an autocrine excitatory transmitter in human pancreatic beta-cells

Abstract

Objective: Paracrine signaling via gamma-aminobutyric acid (GABA) and GABA(A) receptors (GABA(A)Rs) has been documented in rodent islets. Here we have studied the importance of GABAergic signaling in human pancreatic islets.

Research design and methods: Expression of GABA(A)Rs in islet cells was investigated by quantitative PCR, immunohistochemistry, and patch-clamp experiments. Hormone release was measured from intact islets. GABA release was monitored by whole-cell patch-clamp measurements after adenoviral expression of alpha(1)beta(1) GABA(A)R subunits. The subcellular localization of GABA was explored by electron microscopy. The effects of GABA on electrical activity were determined by perforated patch whole-cell recordings.

Results: PCR analysis detected relatively high levels of the mRNAs encoding GABA(A)R alpha(2), beta(3,) gamma(2), and pi subunits in human islets. Patch-clamp experiments revealed expression of GABA(A)R Cl(-) channels in 52% of beta-cells (current density 9 pA/pF), 91% of delta-cells (current density 148 pA/pF), and 6% of alpha-cells (current density 2 pA/pF). Expression of GABA(A)R subunits in islet cells was confirmed by immunohistochemistry. beta-Cells secreted GABA both by glucose-dependent exocytosis of insulin-containing granules and by a glucose-independent mechanism. The GABA(A)R antagonist SR95531 inhibited insulin secretion elicited by 6 mmol/l glucose. Application of GABA depolarized beta-cells and stimulated action potential firing in beta-cells exposed to glucose.

Conclusions: Signaling via GABA and GABA(A)R constitutes an autocrine positive feedback loop in human beta-cells. The presence of GABA(A)R in non-beta-cells suggests that GABA may also be involved in the regulation of somatostatin and glucagon secretion.

FIG. 1.

Functional detection of endogenous GABA_AR Cl⁻ channels in human islet cells. A: Patch-clamp recording of currents evoked by puffer application of GABA (1 mmol/l, as indicated by the bars) to an identified β-cell in the absence (black trace) and presence (gray trace) of 50 μmol/l SR-95531 in the same cell. The cell was held at −70 mV throughout the experiment. B: As in A, showing a δ-cell (note the difference in scale bars). C: As in A, showing an α-cell. The cell had been incubated in the presence of 0.5 μmol/l insulin for 1 h before the experiment.

FIG. 2.

Expression of GABA_AR subunits in human islets. A: Expression profiling of GABA_AR subunits in islets by quantitative RT-PCR (n = 3 preparations from three donors). B–D: Co-labeling of pancreatic tissue sections with anti-insulin and anti-GABA_AR β_2/3 (B), anti-GABA_AR α_1–6 (C), or anti-GABA_AR γ₂ (D). Insulin is shown in red and GABA_AR subunits in green; co-localization results in yellow labeling (scale bars = 10 μm). a.u., arbitrary units.

FIG. 3.

Quantal release of GABA from human β-cells. A: A β-cell overexpressing α₁/β₁ GABA_AR was held at −70 mV and infused with intracellular solution containing 2 μmol/l free Ca²⁺ (at 5 mmol/l extracellular glucose). SR-95531 (10 μmol/l) was applied as indicated by the bar. The inset shows a part of the trace (indicated by *) on an expanded time scale. B: A train of 10 500-ms voltage-clamp depolarizations from −70 to 0 mV was applied to a cell overexpressing α₁/β₁ GABA_AR (with intracellular solution containing 50 μmol/l EGTA). GABA-induced transient currents are indicated by arrows. The inset shows a part of the trace (marked by the dotted rectangle) on an expanded time base. Note that the direction of transient currents is outward at 0 mV. The initial downward component represents the opening of the voltage-gated Na⁺ and Ca²⁺ currents triggered by the depolarization. The activation of the GABA_AR accounts for the outward current. The recording shown is representative of seven experiments.

FIG. 4.

Glucose- and tolbutamide-induced GABA release from human β-cells. Experiments were performed in small clusters of islet cells overexpressing α₁/β₁ GABA_AR. The patch-clamped cell was held at −70 mV and infused with pipette solution containing 10 mmol/l EGTA. A: The glucose concentration in the bath was increased from 1 to 6 mmol/l as indicated. B (upper): The extracellular glucose concentration was increased from 1 to 20 mmol/l. SR-95531 (10 μmol/l) was included as indicated. B (lower): Sections of top trace (as indicated by letters i and ii) shown on an expanded time base. C: Summary of observed frequencies of GABA release (TICs/min) at the indicated glucose concentrations (*P < 0.05). The measurements were made at steady state (1–5 min after addition of glucose). D: Tolbutamide (100 μmol/l) was applied as indicated (extracellular glucose concentration 4 mmol/l). E: SR-95531 (10 μmol/l) was applied as indicated at 1 mmol/l extracellular glucose.

FIG. 5.

Storage and secretion of GABA by insulin-containing LDCVs in human β-cells. A: GABA release was detected by patch-clamping in an identified β-cell overexpressing α₁/β₁ GABA_AR (upper trace). The cell was held at −70 mV and infused with intracellular solution containing 2 μmol/l free Ca²⁺. Serotonin release was measured simultaneously in the same cell by carbon fiber amperometry (lower trace). The dashed lines indicate simultaneous occurrence of GABA-induced TICs and amperometric currents. B: Immunogold labeling of GABA in a human β-cell. Scale bar: 250 nm.

FIG. 6.

Effect of GABA_AR blockade on insulin secretion from isolated human islets. Insulin secretion was measured at 1, 3, 6, 10, or 20 mmol/l extracellular glucose in the absence (○) or presence (●) of SR-95531 (10 μmol/l) as indicated (n = 9–12 from four donors; ‡P < 0.001 for the effect of SR-95531). Under control conditions insulin release at all glucose concentrations was significantly different from the previous lower glucose concentration (not indicated). The results at 1, 6, and 20 mmol/l glucose were repeated with islets from seven different donors with the same results.

FIG. 7.

Effects of GABA on the membrane potential of human β-cells. Membrane potential was recorded in noninfected cells in the perforated-patch configuration, using the Cl⁻-impermeable antibiotic gramicidin as the perforating agent. Recordings were made at 6 mmol/l extracellular glucose. A: Application of GABA (100 μmol/l, indicated by bars) to a β-cell in the absence (left) or presence (right) of 50 μmol/l SR-95531. B: Application of 10 μmol/l GABA (bar) to an electrically active β-cell.