Adipocyte-derived exosomal ANGPTL2 promotes beta cell dysfunction via the EIF2α/ATF4/CHOP stress pathway in mice.

aims/hypothesisObesity-associated lipotoxicity drives beta cell failure in type 2 diabetes, but the adipose-to-islet signals that promote beta cell dysfunction remain incompletely defined. We hypothesised that angiopoietin-like protein 2 (ANGPTL2) carried by lipotoxic adipocyte-derived exosomes contributes to adipose-to-islet signalling that impairs beta cell function.

methodsExosomes from palmitate- or PBS-treated 3T3-L1 adipocytes were applied to MIN6 cells and primary mouse islets to assess beta cell function and identity. In vivo, mice underwent adipose-specific Angptl2 knockdown under high-fat diet (HFD) feeding, followed by metabolic phenotyping and assessment of beta cell function. Adipose-derived exosomes from HFD-fed mice with or without adipose Angptl2 knockdown were injected into normal-chow-fed recipient mice, and endocrine marker composition and islet secretory function were evaluated. Mechanistic studies in MIN6 cells used co-immunoprecipitation, pharmacological rescue experiments and Eif2α (also known as Eif2s1) knockdown to interrogate stress signalling.

resultsExosomes derived from palmitate-treated adipocytes were efficiently internalised by beta cells and resulted in impaired insulin content and secretion, accompanied by reduced beta cell identity markers and induction of dedifferentiation signatures. ANGPTL2 was enriched in lipotoxic adipocyte-derived exosomes and accumulated in beta cells in an exosome-dependent manner. Angptl2 silencing in adipocytes lowered exosomal ANGPTL2 and mitigated islet dysfunction in vitro. In vivo, adipose-specific Angptl2 knockdown reduced islet ANGPTL2 without changing islet Angptl2 mRNA and improved glucose tolerance, with the AUC of the IPGTT decreasing from 2300.2 ± 165.5 to 1583.7 ± 88.8 mmol/l × min (mean ± SD). In adipose-exosome transfer experiments, recipients of HFD-short hairpin (sh)Angptl2-exosome (Exo) showed lower islet ANGPTL2 abundance and improved ex vivo islet glucose-stimulated insulin secretion (GSIS), with stimulated insulin secretion increasing from 6.0 ± 0.3% (mean ± SD) in the HFD-sh-negative control (NC)-Exo group to 10.9 ± 0.6% in the HFD-shAngptl2-Exo group. Mechanistically, ANGPTL2 associated with eukaryotic translation initiation factor 2α (EIF2α) and increased protein kinase R-like endoplasmic reticulum kinase (PERK)-associated EIF2α by approximately 2.8-fold, whereas this difference was lost after further normalisation to input EIF2α, supporting a substrate-availability model. PERK inhibition and integrated stress response inhibitor both attenuated ANGPTL2-induced activating transcription factor-4 (ATF4)/C/EBP homologous protein (CHOP) signalling and improved GSIS. CONCLUSIONS/

interpretationLipotoxic adipocyte-derived exosomes increase ANGPTL2 abundance in recipient beta cells and contribute to beta cell dysfunction and identity disruption by amplifying PERK-dependent EIF2α/ATF4/CHOP signalling. Adipose ANGPTL2 and circulating exosome-associated ANGPTL2 represent candidate targets and biomarkers for preserving beta cell function in obesity-associated dysglycaemia.