Extended Data Figure 5: Atg16l1 intestinal-epithelial-cell-specific deletion abrogates UPR-induced autophagy and exacerbates ER stress-induced spontaneous intestinal inflammation due to Xbp1 deficiency.

a, Immunoblot of primary intestinal epithelial cell scrapings. Note identical GAPDH loading control as in Fig. 3b (n = 3). b, Representative high-magnification TEM images (n = 2). Note intact ER in wild-type mice; severe distortion of ER in Atg16l1^ΔIEC mice; autophagic vacuoles with the characteristic double membrane (arrow) in Xbp1^ΔIEC mice; and absence of ER and autophagic vacuoles in Atg16l1/Xbp1^ΔIEC mice. Scale bar, 200 nm. ER, endoplasmic reticulum. c, Quantification of autophagic vacuoles (n = 10 unpaired Student’s t-test; mean ± s.e.m.). d, Representative images of p62 immunostaining on small intestinal sections (n = 3). Brown staining indicates p62 with maximal signal intensity in the region of Paneth cells. Scale bars, 20 μm. e, High-resolution haematoxylin and eosin images. Note the extension of the inflammation into muscularis propria (arrow) in Atg16l1/Xbp1^ΔIEC mice. Scale bars, 20 μm. f, Deep inflammation score (as described in Methods) in mice assessed at 18 weeks of age (n = 11; median shown; Kruskal–Wallis with post-hoc Dunn’s correction). g, Enteritis histology score (n = 14; 7–10-week-old mice housed in Innsbruck; median shown; Kruskal–Wallis with post-hoc Holm’s-corrected Mann–Whitney U-test). h, qRT–PCR of grp78 in primary intestinal epithelial cell scrapings (n = 7; unpaired Student’s t-test; mean ± s.e.m.). i, Disease activity index during DSS colitis (n = 7/4, two-way ANOVA with post-hoc Bonferroni; mean ± s.e.m.). j, Representative endoscopic images from the colon at day 5 of DSS treatment (n = 3). k, Representative haematoxylin and eosin staining of colonic sections from 5-day DSS-treated and untreated mice (n = 7/4). Scale bars, 50 μm. l, Colitis score after 5 days of DSS administration (n = 7/4, Mann–Whitney U-test). *P < 0.05, **P < 0.01, ***P < 0.001.