Extended Data Figure 10: Model of the interplay between UPR-induced autophagy, NF-κB signalling and inflammation.

a, Deletion of Xbp1 in the intestinal epithelium, specifically in Paneth cells, leads to ER stress and activation of the PERK–eIF2α branch of the UPR. ATF4, a transcriptional mediator of this pathway, transactivates genes essential for autophagosome formation, such as Map1lc3b (LC3b) and Atg7, which catalyses the creation of the ATG12–ATG5 conjugate that stabilizes ATG16L1 through complex formation²¹. UPR-induced autophagy in the intestinal epithelium is essential for restoration of homeostasis and restraint of ER stress-induced intestinal inflammation due to XBP1 deficiency. Activation of the UPR in the setting of XBP1 deficiency results in activation of IRE1α, resulting in the recruitment of TRAF2 and activation of IKK2 leading to IκBα degradation^4,27,45,46. As shown here, UPR-mediated autophagy serves an important role in restraining NF-κB activation, conceivably by removing hyperinflammatory ER membranes containing activated IRE1α. Pharmacological augmentation of this compensatory autophagy-dependent mechanism via inhibition of eIF2α dephosphorylation through salubrinal, or via the mTOR inhibitor rapamycin, results in amelioration of UPR-induced enteritis, which is driven by the commensal microbiota, NF-κB, and TNF-R1 signalling in intestinal epithelial cells and myeloid cells, whereby the ligand TNF may originate from XBP1-deficient intestinal epithelial cells⁴. b, ATG16L1 deficiency in intestinal epithelial cells leads to ER stress as revealed through upregulation of the chaperone GRP78 in intestinal epithelial cells, increased expression of GRP78 protein in Paneth cells, increased IRE1α expression and increased splicing of Xbp1 mRNA in intestinal crypts as well as increased intestinal epithelial cell death. This leads to increased sensitivity of the epithelium to environmental triggers (for example, dextran sodium sulphate) that further challenge the UPR and its compensatory pathways. c, Deficiency of ATG16L1 or ATG7 in the intestinal epithelium results in abrogation of the compensatory autophagic mechanism that restrains IRE1α activity, conceivably via removal of hyperinflammatory ER membranes, and further fosters intestinal epithelial cell death in the context of ER stress due to Xbp1 deficiency, resulting in spontaneous transmural small intestinal inflammation that is associated with further increases in NF-κB activation and cell death via the mechanisms described in a. The UPR allows for responses to a variety of signals that have an impact on protein folding, including genetic (for example, rare XBP1 variants, ORMDL3 as risk factor of inflammatory bowel disease^4,47), environmental (for example, low O₂ tension in the intestinal tract) and microbial factors (for example, microbial toxins such as trierixin⁴⁸), which determines the level of ER stress in the intestinal epithelium. UPR-induced autophagy function provides a buffer to cope with different levels of ER stress and vice versa. However, in the presence of genetic risk variants, such as ATG16L1 (refs 1, 17, 49) or IRGM (ref. 50), which are relatively prevalent in the general population, this compensatory mechanism is impaired, resulting in development and/or exacerbation of intestinal inflammation in the setting of unabated ER stress.