Expression of genes encoding the GABA signalling system components in islets
A comprehensive expression profile of genes encoding proteins participating in the GABA signalling cascade in excitable cells was created using cDNA microarray from islets from 54 non-diabetic donors (ESM Fig. 1a, b). Of the genes encoding the GABAA channel subunits, the β3, γ2, δ, ε, π and ρ2 subunits were most prominently expressed in the islets (ESM Fig. 1a). High expression signals were also detected for GABARAP (GABA receptor-associated protein), radixin (RDX), the NKCC1 transporter (SLC12A2) and GAD65 (also known as GAD2) (ESM Fig. 1b).
We further examined and compared gene expression in islets from individuals with and without type 2 diabetes using RT-qPCR analysis (Fig. 1). The highest level of expression was obtained for the genes for α1, α2, β3, γ2, π and ρ2 GABAA channel subunits in islets from normoglycaemic individuals (n = 14). In islets from type 2 diabetic donors, the α1, α2, β2 and β3 GABAA subunits were significantly downregulated (Fig. 1a). The other genes in the GABA signalling cascade (Fig. 1b) were all similarly expressed and did not differ significantly between islets from individuals with type 2 diabetes and those from normoglycaemic donors. We further examined whether we could detect downregulation of genes in the islets from hyperglycaemic donors (not diagnosed with type 2 diabetes; n = 6). Only the α2 GABAA channel subunit was significantly downregulated in islets from hyperglycaemic individuals compared with normoglycaemic individuals (Fig. 1c, α1, p = 0.127; β2, p = 0.536; β3, p = 0.386).
Localisation of GABAA α1 and α2 channel subunit proteins in human pancreatic islets
To determine which specific GABAA channel subunits are present in pancreatic islet beta cells, we analysed expression of the subunits in sorted beta cells from one normoglycaemic donor using RT-qPCR. The α1, α2, α5, β3, γ2, δ, π and ρ2 GABAA channel subunits were present in the cells (Fig. 2a). To identify the cellular and subcellular location of the GABAA channel subunits, we immunostained for the α1 or α2 subunits together with insulin and glucagon in pancreatic islets (Fig. 2b–e). The results show that, in human pancreatic islets, the α1 and the α2 GABAA channel subunit proteins are present in the plasma membrane and cytoplasm of alpha and beta cells. The α2 subunit appears particularly prominent in the cells, whereas α1 expression is more limited.
Interstitial GABA activates GABAA channel currents in intact islets
The exact interstitial GABA concentration in the islets is not known, but can be assumed to be in the submicromolar range or similar to that in the extracellular fluid in the brain . The highest GABA concentration is expected to be around the beta cell release sites. Using the patch-clamp technique and recording from cells in intact islets, we examined currents from GABAA channels activated by the interstitial GABA originating within the islets, as no GABA was added experimentally. Figure 3a shows whole-cell currents, which were inhibited by the GABAA channel competitive antagonist, SR95531 (100 μmol/l). The upward shift in the baseline current when SR95531 was applied shows the level of the GABA-activated GABAA current. The levels of lines 1 and 2 in Fig. 3a correspond to the peak values (Fig. 3b) of Gaussian fits to histograms of 30 s current records from before and after SR95531 application. The difference between the peak values was 4.4 pA and is the level of the GABA-generated current in the cell. This type of current is termed ‘tonic’, as it is long lasting and can significantly affect cell excitability . In two cells, synaptic-like transient currents were recorded (data not shown) similar to that reported by Braun et al [1, 8]. We examined whether larger tonic currents were generated if we applied no glucose to the cells, but applied 10 mmol/l glutamine. Under these conditions, no tonic currents were recorded (n = 5), but when we applied 20 mmol/l glucose to islets from the same donor, tonic currents were evoked (n = 3). We then examined whether a positive modulator of GABAA channels, pentobarbital, enhanced the GABA-generated tonic currents (Fig. 3c, d). The level of line 3 (Fig. 3c) corresponds to the peak value (Fig. 3d) of the Gaussian fit to a histogram of 30 s current record after 100 μmol/l pentobarbital application to the islet. The level of the GABA-generated tonic current in the cell was 3.6 pA and was enhanced to 7.1 pA by 100 μmol/l pentobarbital (Fig. 3d). Interestingly, GABA-generated tonic currents were minimal or not detected in islets from type 2 diabetic donors until we applied pentobarbital (n = 5, Fig. 3e, f). The results are consistent with pentobarbital enhancing GABAA currents by increasing both the open probability and the channel conductance of GABAA channels, resulting in higher apparent affinity of the channels for GABA. The enhanced current was inhibited by SR95531 (Fig. 3e, f). We recorded single-channel currents from three different cells (Fig. 4a, b, c) in islets from normoglycaemic donors (Fig. 4a, slow time scale (s); Fig. 4a expanded, Fig. 4b and Fig. 4c, fast time scale (ms); the glucose concentration was 20 mmol/l and the holding potential: Fig. 4a, −90 mV; Fig. 4b, −70 mV; Fig. 4c, −70 mV). In all three cells, when the islets were perfused with 100 μmol/l SR95531, the single-channel currents were inhibited. Figure 4a (top current trace, time scale s) shows SR95531 inhibition of the channels in one of the cells. The most prominent single-channel current amplitude recorded in each cell is indicated in Fig. 4 by the dotted lines and gave channel conductance of 51 pS, 36 pS and 71 pS for the cells in Fig. 4a, b and c, respectively.
Effects of GABAA channel and GABAB receptor antagonists on insulin and glucagon secretion in islets from individuals with or without type 2 diabetes
In islets from both normoglycaemic and type 2 diabetic donors, 16.7 mmol/l glucose stimulated insulin secretion and reduced glucagon secretion relative to the secretion observed in response to 1 mmol/l glucose (ESM Fig. 2). To study the effects of GABA on insulin and glucagon release, we examined the effects of the GABAA and GABAB antagonists, SR95531 (10 μmol/l, Fig. 5a, b) and CGP55845 (10 μmol/l, Fig. 5c, d), on hormone secretion in islets from normoglycaemic and type 2 diabetic individuals incubated with basal (1 mmol/l) or high-concentration (16.7 mmol/l) glucose.
Figure 5a and c show that insulin secretion was increased at 1.0 mmol/l but significantly decreased at 16.7 mmol/l glucose in islets from individuals with type 2 diabetes compared with islets from normoglycaemic donors. At 1 mmol/l glucose, insulin secretion was increased by a factor of 1.6 in the islets from type 2 diabetic donors and was unaffected by either SR95531 (10 μmol/l, Fig. 5a) or CGP55845 (Fig. 5c). At 16.7 mmol/l glucose, insulin secretion in islets from the type 2 diabetic individuals was decreased by 60% (Fig. 5a), but could be increased about 1.5-fold with the GABAB receptor antagonist CPG55845 (Fig. 5c) in islets from both type 2 diabetic and normoglycaemic donors. As 10 μmol/l SR95531 may not completely inhibit very high-affinity GABAA channels, we examined in islets from two normoglycaemic individuals whether 100 μmol/l SR95531 had any effect on insulin secretion. In low (1 mmol/l) glucose, 100 μmol/l SR95531 significantly increased insulin secretion (1G + SR100 μmol/l: 0.20 ± 0.03 ng islet−1 h−1, n = 12 from two individuals; 1G: 0.14 ± 0.01 ng islet−1 h−1, n = 12 from four individuals, p < 0.05) to the level observed for type 2 diabetic individuals (0.23 ± 0.02 ng islet−1 h−1, n = 12 from three individuals), but failed to modulate insulin secretion in 16.7 mmol/l glucose (n = 12 from two individuals, data not shown).
Figure 5b and d show that glucagon secretion was only significantly increased in islets from type 2 diabetic individuals compared with normoglycaemic donors when the islets were exposed to 16.7 mmol/l glucose. Glucagon secretion could be enhanced with 10 μmol/l SR95331 (Fig. 5b) in islets exposed to either 1 or 16.7 mmol/l glucose, whereas CGP555845 (Fig. 5d) was without effect. SR95531 increased glucagon secretion at 1.0 and 16.7 mmol/l glucose by a factor of 1.3 and 1.5, respectively, in islets from type 2 diabetic donors, and a factor of 1.4 and 1.9, respectively, in islets from normoglycaemic donors. The somewhat reduced effect of 10 μmol/l SR95531 on islets from type 2 diabetic individuals may be due to the fact that these islets exhibited increased basal glucagon secretion compared with the normoglycaemic donors. The GABAB receptor antagonist, CGP55845, did not modulate glucagon secretion at either low or high glucose concentrations (Fig. 5d). Together these results demonstrate regulation of hormone release by the GABAA channels and GABAB receptors.
We further examined whether the mRNA expression of GABAA channel subunits α1 and α2 in islets correlated with the HbA1c or insulin secretion of donors. Interestingly, expression of the GABAA channel subunits, α1 and α2, in islets correlated negatively with the HbA1c level of donors (Fig. 5e, f), and α2, but not α1, correlated positively with insulin secretion stimulated with 16.7 mmol/l glucose (Fig. 5e, f).