High glucose activates hypoxia inducible factors 1 and 2 in cultured rat islets and INS1-E cell monolayers

Bensellam, Mohammed [UCL] Jonas, Jean-Christophe [UCL]

(eng) Background and Aim: We previously tested the effects of a 18h culture in 2, 5, 10 and 30 mmol/l glucose (G2, G5, G10 and G30) on the transcriptome of cultured rat islets and identified 18 clusters with distinct glucose-dependent mRNA profiles. Of these, genes that were up-regulated between G10 and G30 are particularly interesting, as they may contribute to beta-cell glucotoxicity or protection against it. Besides genes involved in the unfolded protein response, this cluster contains most glycolytic enzymes and other genes typically induced by hypoxia, such as Adrenomedullin (Adm). Hypothesis: High glucose, which increases O2 consumption in β-cells, may induce hypoxia and thereby cause glucotoxicity. In this study, we tested whether overnight culture in high glucose activates the hypoxia-inducible transcription factor HIF in cultured rat beta-cells. Material and Methods: Male Wistar rat islets were precultured for one week in serum-free RPMI medium containing 10 mmol/l glucose (G10) and 5 g/l BSA, during which islets with signs of central necrosis were systematically discarded. INS1-E cells were cultured as monolayers (passage 72 to 80) in the presence of 10% FBS. Rat islets and INS1-E cells (70% confluence) were then cultured 18h in the presence of increasing glucose concentrations (G2, G5, G10 and G30) with various test substances and at different O2 concentrations (1%, 20% or 60% corresponding to a pO2 of 7.6, 152 and 456 mmHg). Results: Compared with G2, culture in G30 significantly increased 4-fold the protein levels of HIF1α and HIF2α, and 2-fold their dimerization partner HIF1β in INS1-E cell nuclear extracts while significantly decreasing HIF1β protein levels by ~50% in cytosolic extracts. High glucose also significantly up-regulated the mRNA levels of several HIF target genes (Adm, Glyceraldehyde phosphate dehydrogenase, Aldolase A, Lactate dehydrogenase A). These glucose effects, which were confirmed in whole rat islets, were mimicked by 6h exposure to a low pO2 (1% O2) or by 18h treatment with 10-30 µmol/l CoCl2, a known activator of HIF. In contrast, they were almost completely inhibited by 60% O2 or by various agents that inhibit Ca2+ influx and insulin secretion (250 µmol/l diazoxide, 1 µmol/l nimodipine and 1 µmol/l clonidine) and thereby likely reduce without suppressing the glucose stimulation of β-cell O2 consumption. In comparison with HIF target genes, Thioredoxin interacting protein (Txnip), one of the most glucose-responsive genes in cultured rat islets, was regulated in a completely different manner. Thus, Txnip mRNA levels were not increased by CoCl2 and 1% O2, and their increase by G30 was not reduced by 60% O2 and was markedly enhanced by diazoxide, nimodipine and clonidine. Conclusion: High glucose induces HIF1α and HIF2α protein stabilization, HIF1/2α-HIF1β dimer translocation to the nucleus, and increased expression of glycolytic enzymes and other HIF target genes in cultured rat beta-cells. These effects likely result from the glucose stimulation of ATP utilization and O2 consumption which may induce beta-cell hypoxia. These effects could contribute to in vitro β-cell glucotoxicity not only in whole islets but also in insulin-secreting cells cultured as monolayers.