Jorge Tamarit-Rodriguez

Our previous work has demonstrated that islet depolarization with KCl opens connexin36 hemichannels in β-cells of mouse pancreatic islets allowing the exchange of small metabolites with the extracellular medium. In this study, the opening of these hemichannels has been further characterized in rat islets and INS–1 cells. Taking advantage of hemicannels’opening, the uptake of extracellular ATP and its effect on insulin release were investigated. 70 mM KCl stimulated light emission by luciferin in dispersed rat islets cells transduced with the fire-fly luciferase gene: it was suppressed by 20 mM glucose and 50 μM mefloquine, a specific connexin36 inhibitor. Extracellular ATP was taken up or released by islets depolarized with 70 mM KCl at 5 mM glucose, depending on the external ATP concentration. 1 mM ATP restored the loss of ATP induced by the depolarization itself. ATP concentrations above 5 mM increased islet ATP content and the ATP/ADP ratio. No ATP uptake occurred in non-depolarized or KCl-depolarized islets simultaneously incubated with 50 μM mefloquine or 20 mM glucose. Extracellular ATP potentiated the secretory response induced by 70 mM KCl at 5 mM glucose in perifused rat islets: 5 mM ATP triggered a second phase of insulin release after the initial peak triggered by KCl-depolarization itself; at 10 mM, it increased both the initial, KCl-dependent, peak and stimulated a greater second phase of secretion than at 5 mM. These stimulatory effects of extracellular ATP were almost completely suppressed by 50 μM mefloquine. The magnitude of the second phase of insulin release due to 5 mM extracellular ATP was decreased by addition of 5 mM ADP (extracellular ATP/ADP ratio = 1). ATP acts independently of K_ATP channels closure and its intracellular concentration and its ATP/ADP ratio seems to regulate the magnitude of both the first (triggering) and second (amplifying) phases of glucose-induced insulin secretion.

The existence of functional connexin36 ( Cx36) hemichannels in β-cells was investigated in pancreatic islets of rat and wild-type ( Cx36^+/+), monoallelic ( Cx36^+/−), and biallelic ( Cx36^−/−) knockout mice. Hemichannel opening by KCl depolarization was studied by measuring ATP release and changes of intracellular ATP (ADP). Cx36^+/+ islets lost ATP after depolarization with 70 mM KCl at 5 mM glucose; ATP loss was prevented by 8 and 20 mM glucose or 50 μM mefloquine (connexin inhibitor). ATP content was higher in Cx36^−/− than Cx36^+/+ islets and was not decreased by KCl depolarization; Cx36^+/− islets showed values between that of control and homozygous islets. Five minimolar extracellular ATP increased ATP content and ATP/ADP ratio and induced a biphasic insulin secretion in depolarized Cx36^+/+ and Cx36^+/− but not Cx36^−/− islets. Cx36 hemichannels expressed in oocytes opened upon depolarization of membrane potential, and their activation was inhibited by mefloquine and glucose (IC₅₀ ∼8 mM). It is postulated that glucose-induced inhibition of Cx36 hemichannels in islet β-cells might avoid depolarization-induced ATP loss, allowing an optimum increase of the ATP/ADP ratio by sugar metabolism and a biphasic stimulation of insulin secretion. Gradual suppression of glucose-induced insulin release in Cx36^+/− and Cx36^−/− islets confirms that Cx36 gap junction channels are necessary for a full secretory stimulation and might account for the glucose intolerance observed in mice with defective Cx36 expression. Mefloquine targeting of Cx36 on both gap junctions and hemichannels also suppresses glucose-stimulated secretion. By contrast, glucose stimulation of insulin secretion requires Cx36 hemichannels' closure but keeping gap junction channels opened.

We have demonstrated recently that branched-chain α-keto acid stimulation of insulin secretion is dependent on islet GABA (γ-aminobutyric acid) metabolism: GABA transamination to succinic semialdehyde is increased by 2-oxoglutarate, generated in α-keto acid transamination to its corresponding α-amino acid. The present work was aimed at investigating whether glucose also promotes islet GABA metabolism and whether the latter contributes to the stimulation of insulin secretion. Glucose (20 mM) decreased both the content and release of islet GABA. Gabaculine (1 mM), a GABA transaminase inhibitor, partially suppressed the secretory response of rat perifused islets to 20 mM glucose at different L-glutamine concentrations (0, 1 and 10 mM), as well as the glucose-induced decrease in islet GABA. The drug also reduced islet ATP content and the ATP/ADP ratio at 20 mM glucose. Exogenous succinic semialdehyde induced a dose-dependent increase in islet GABA content by reversal of GABA transamination and a biphasic insulin secretion in the absence of glucose. It depolarized isolated β-cells and triggered action potential firing, accompanied by a reduction of membrane currents through ATP-sensitive K+ channels. The gene expression and enzyme activity of GABA transaminase were severalfold higher than that of 2-oxoglutarate dehydrogenase in islet homogenates. We conclude that, at high glucose concentrations, there is an increased diversion of glucose metabolism from the citric acid cycle into the ‘GABA shunt’. Semialdehyde succinic acid is a cell-permeant ‘GABA-shunt’ metabolite that increases ATP and the ATP/ADP ratio, depolarizes β-cells and stimulates insulin secretion. In summary, an increased islet GABA metabolism may trigger insulin secretion.

We have previously shown that oxo-4-methylpentanoate promotes islet GABA (γ-aminobutyric acid) metabolism and stimulates insulin secretion. The main aim of this work was to explore the participation of the transamination of branched-chain 2-oxoacids in these processes with the aid of several inhibitors of this enzyme activity. No correlation was found between the transamination of branched-chain 2-oxoacids in islet homogenates and insulin secretion. However, in vivo transamination rates correlated better with the secretion capacity of the different branched-chain 2-oxoacids. Gabapentin, a specific inhibitor of the cytosolic isoenzyme, showed greater potential to decrease the in vitro transamination rates of oxo-3-methylbutyrate and oxo-3-methylpentanoate than those of oxo-4-methylpentanoate and oxohexanoate; this correlated with its capacity to decrease insulin secretion. 4-Methylvaleric acid very strongly inhibited the transamination of all the branched-chain 2-oxoacids and blocked their capacity to decrease islet GABA and to stimulate insulin secretion. KCl at 70 mM at stimulated islet GABA release, subsequently decreasing its tissue concentration. This ‘non-metabolic’ decrease of GABA suppressed the second phase of insulin secretion triggered by oxo-4-methylpentanoate and oxohexanoate. Oxo-4-methylpentanoate and oxo-3-methylpentanoate suppressed dose-dependent 2-oxoglutarate dehydrogenase activity in islet homogenates. In conclusion, the transamination of branchedchain 2-oxoacids is more important to the stimulation of insulin secretion than their catabolism, and transamination decreases islet GABA concentrations by promoting GABA metabolism. Also, inhibition of 2-oxoglutarate dehydrogenase by branched-chain 2-oxoacids may increase metabolic flux in the ‘GABA-shunt’ at the expense of reduced tricarboxylic acid cycle flux.

We have carried out a detailed examination of l-glutamine metabolism in rat islets in order to elucidate the paradoxical failure of l-glutamine to stimulate insulin secretion. l-Glutamine was converted by isolated islets into GABA (γ-aminobutyric acid), l-aspartate and l-glutamate. Saturation of the intracellular concentrations of all of these amino acids occurred at approx. 10 mmol/l l-glutamine, and their half-maximal values were attained at progressively increasing concentrations of l-glutamine (0.3 mmol/l for GABA; 0.5 and 1.0 mmol/l for Asp and Glu respectively). GABA accumulation accounted for most of the 14CO2 produced at various l-[U-14C]glutamine concentrations. Potentiation by l-glutamine of l-leucine-induced insulin secretion in perifused islets was suppressed by malonic acid dimethyl ester, was accompanied by a significant decrease in islet GABA accumulation, and was not modified in the presence of GABA receptor antagonists [50 µmol/l saclofen or 10 µmol/l (+)-bicuculline]. l-Leucine activated islet glutamate dehydrogenase activity, but had no effect on either glutamate decarboxylase or GABA transaminase activity, in islet homogenates. We conclude that (i) l-glutamine is metabolized preferentially to GABA and l-aspartate, which accumulate in islets, thus preventing its complete oxidation in the Krebs cycle, which accounts for its failure to stimulate insulin secretion; (ii) potentiation by l-glutamine of l-leucine-induced insulin secretion involves increased metabolism of l-glutamate and GABA via the Krebs cycle (glutamate dehydrogenase activation) and the GABA shunt (2-oxoglutarate availability for GABA transaminase) respectively, and (iii) islet release of GABA does not seem to play an important role in the modulation of the islet secretory response to the combination of l-leucine and l-glutamine.