An epigenetic switch in β-cells links mitochondrial stress to inflammatory fate via HMGB1 acetylation

Pancreatic β-cell failure in diabetes is driven by chronic inflammation, yet how metabolic stress determines pro-inflammatory cell fate remains unclear. Here, we report that sublethal oxidative stress activates a β-cell-enriched epigenetic switch that licenses intrinsic inflammation. We identify a β-cell-enriched vulnerability wherein oxidative stress disrupts mitochondrial NAD⁺/acetyl-CoA flux, skewing the nuclear equilibrium between the deacetylase SIRT1 and acetyltransferases p300/CBP. This metabolic-epigenetic imbalance induces hyperacetylation of the alarmin HMGB1 at K96/K128-a modification remarkably prominent in β-cells compared to macrophages or hepatocytes. This site-specific acetylation acts as a molecular gate for HMGB1 nucleocytoplasmic translocation, triggering TLR/RAGE-mediated inflammation. Simultaneously, we discover that oxidative stress co-opts the mechanosensitive Hippo pathway, which sequesters YAP to transcriptionally repress SIRT1, thereby forming a functionally integrated signalling axis that exacerbates HMGB1 acetylation. Therapeutically, reconstruction of mitochondrial retrograde signalling via NAD⁺ supplementation (e.g., NMN) or dual inhibition of mitochondrial ROS and Hippo signalling restored acetylation homeostasis and suppressed HMGB1-driven inflammation. Notably, this combinatorial targeting demonstrates greater efficacy than either intervention alone (~73% reduction in inflammatory markers) in mitigating β-cell failure across murine and porcine models. Further validation in non-human primates was specifically implemented to address the unique translational gap of rodent models and available human single-cell datasets: by leveraging human-like islet anatomy and systemic physiological microenvironment, we verified the pharmacodynamic robustness and in vivo feasibility of this strategy in a clinically recapitulative large-animal setting, rather than merely confirming cross-species molecular concordance. Our work unveils a stress-responsive signalling network in which metabolic and mechanical cues are integrated at the epigenetic level to control the inflammatory fate of β-cells, providing a new mechanistic framework for diabetic pathogenesis and a rationale for combinatorial therapeutic intervention. Oxidative stress triggers mitochondrial dysfunction in pancreatic β-cells, depleting NAD⁺ and accumulating acetyl-CoA. This metabolic crisis skews the SIRT1/p300 balance, inducing β-cell-enriched hyperacetylation of HMGB1 at K96/K128-a molecular switch for its nucleocytoplasmic translocation, which activates TLR/RAGE-mediated intrinsic inflammation. Concurrent Hippo pathway activation further exacerbates this process by repressing SIRT1, forming an integrated signaling axis. To ensure reproducibility, we define a minimal validation workflow using β-TC-6 cells or primary islets under standardized sublethal stress. The central mechanism can be verified by monitoring HMGB1 K96/K128 acetylation and nucleocytoplasmic translocation, while the inflammatory cascade can be effectively blunted through modular rescue approaches, including NAD+ supplementation (NMN), mitochondrial ROS inhibition or hippo pathway inhibition.