Eukaryotic translation initiation factor 2A protects pancreatic beta cells during endoplasmic reticulum stress while rescuing global translation inhibition

Abstract

Aims/hypothesis

The endoplasmic reticulum (ER) stress-induced unfolded protein response helps determine beta cell survival rate in diabetes. The alternative eukaryotic translation initiation factor 2A (EIF2A) has been proposed to mediate translation initiation independent of the α subunit of EIF2 (EIF2S1) during cellular stress, but its role in beta cells has not been comprehensively examined.

Methods

For in vitro experiments, we used MIN6 cells, primary mouse pancreatic islets, and human islets obtained under informed consent. Thapsigargin (1 µmol/l) or palmitate complexed with BSA (0.5 mmol/l) was used to induce ER stress. Transient transfection and lentiviral infection were used for transgene delivery. For in vivo experiments, adeno-associated viral particles expressing EIF2A or GFP under the control of a rat insulin promoter were delivered via intraductal injection to 6-week-old female Akita mice randomised into three groups (two cohorts, n=10–11). Tail blood was collected for blood glucose measurements for single time points as well as during glucose and insulin tolerance tests.

Results

EIF2A protein abundance and specificity was high in human and mouse islets relative to other tissues. We used STRING and AlphaFold pulldown to predict interacting proteins and binding partners, verifying EIF1AX with co-immunoprecipitation. Both thapsigargin and palmitate significantly increased EIF2A mRNA and EIF2A protein levels in MIN6 cells, mouse islets and human islets. Knockdowns of EIF2A, the related factor EIF2D or both EIF2A and EIF2D were not sufficient to cause apoptosis. On the other hand, transient or stable EIF2A overexpression protected MIN6 cells, primary mouse islets and human islets from ER stress-induced, caspase-3-dependent apoptosis. Mechanistically, EIF2A overexpression decreased endoplasmic reticulum to nucleus signalling 1 (ERN1, also known as inositol-requiring enzyme 1 α or IRE1α) expression in thapsigargin-treated MIN6 cells or human islets. In vivo, beta cell-specific EIF2A viral overexpression reduced ER stress and improved insulin secretion and glucose tolerance in Ins2^Akita/WT mice. EIF2A overexpression significantly increased expression of genes involved in mRNA translation and reduced expression of pro-apoptotic genes (e.g. Aldh1a3). Proteomic analysis of EIF2A-overexpressing human islets revealed significant changes in pathways associated with ribosomes and protein processing in ER. Remarkably, the decrease in global protein synthesis during unfolded protein response was prevented by EIF2A, despite ER stress-induced EIF2S1 phosphorylation. The protective effects of EIF2A were additive to those of ISRIB, a drug that counteracts the effects of EIF2S1 phosphorylation. Cells overexpressing EIF2A showed higher expression of translation factor EIF2B5, which may contribute to the lack of translational inhibition in these cells.

Conclusions/interpretation

We conclude that EIF2A is a novel target for beta cell protection and the circumvention of EIF2S1-mediated translational repression.

Graphical Abstract