Fig 2: Levels of Paneth cells markers in CD patients and hospitalized controls. (a) mRNA expression of LYZ, DEFA5 and DEFA6 in ileal samples from CD patients and hospitalized controls. (b) Microscopic images of IHC staining of LYZ in ileal samples from CD patients and healthy controls. Scale bar = 500, 20 µm (20x, 63x). (c) Intensity of the staining per crypt was measured using ImageJ software. (d) Lysozyme level was measured in fecal samples from CD patients treated or not treated with AZA by ELISA. (e) qPCR was performed to measure mRNA levels of KI67 in ileal biopsies CD patients and controls. (f) Representative images of ileal KI67 staining in CD patients and controls. Scale bar = 20 µm (63x). (g) Numbers of Ki67 + cells were counted using ImageJ software. (a,c,e,g) Kruskal–Wallis test; (d) Unpaired t test with Welch’s correction; results are shown as truncated violin plots with median represented as dashed line. *p < 0.05, **p < 0.01. HC: hospitalized controls; CD: non-AZA treated CD patients; CD-A: azathioprine-treated CD patients; DEFA5, 6: alpha-defensin 5, 6; LYZ: lysozyme.

Fig 3: Effect of AZA on intestinal organoids with mitochondrial impairment. (a) Atp8 mutant SI organoids were maintained on standard medium for four days. Consequently, DMSO or AZA was introduced on day 4 at a concentration of 1 µM for three days. (b) Microscopic appearance of SI organoids treated with DMSO or AZA. Scale bar = 50 µm (40x). (c) Diameter of DMSO and AZA-treated organoids. (d) Glucose and lactate levels were measured in the supernatants of SI organoids. (e) LDH level was measured in the supernatants of organoids. mRNA expression of Ki67 (f), Defa3 (g) and Lyz1 (h) in DMSO and AZA-treated organoids. (i) Representative images of organoid sections stained with Lyz antibody. Scale bar = 50 µm (20x), 10 µm (63x). (j) Number of Lyz + cells was counted in each organoid. mRNA levels of differentiation markers, Hes1 (k) and Atoh1 (l) were assessed by qPCR. Statistical differences were detected by paired t test. All results are shown as truncated violin plots with median represented as dashed line except (d) mean with SEM. **p < 0.01. Atoh1: Atonal BHLH Transcription Factor 1; AZA: azathioprine; Defa3: alpha-defensin 3; Hes1: Hairy and enhancer of split-1; LDH: lactate dehydrogenase.

Fig 4: Effect of AZA on the intestinal cell line, MODE-K cells. (a) Schematic summary of azathioprine metabolism. (b) mRNA expression of HPRT1, TPMT and XDH was measured in MODE-K via qPCR. (c) Overnight-incubated MODE-K cells were treated with either AZA or DMSO for three days. (d) Cytotoxic effect of AZA was determined in MODE-K treated with various concentrations of AZA (0.1–1000 µM) for three days by neutral red assay. (e) Viable cell numbers of DMSO or 1 µM AZA-treated cells were counted by trypan blue exclusion. Red dashed line represents starting seeding density. (f) LDH level was measured in the supernatants of AZA and DMSO-treated cells. (g) Western blot was performed from whole protein extracts with respective antibodies in cells treated with AZA or DMSO. (h) Relative protein expression of cleaved PARP-1 in cells treated with AZA or DMSO. (i) Mitochondrial bioenergetics and area under the curve (AUC) (j) were measured in MODE-K using MitoXpress Xtra oxygen consumption assay. (k) mRNA levels of Lyz1 were measured in DMSO or AZA-treated cells via qPCR. All experiments were repeated three times. (b) one-way ANOVA with Tukey’s multiple comparisons test; (d,i) one-way ANOVA with Dunnett's multiple comparisons test; (e,f,h,j,k) paired t test; results are shown as truncated violin plots with median represented as dashed line except (f,i) mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. 6-MeMP: 6-methyl mercaptopurine; 6-MP: 6-mercaptopurine; 6-TGNs: 6-thioguanine nucleotides; 6-TIMP: 6-thioinosine monophosphate; 6-TUA: 6-thiouric acid; AZA: azathioprine; HPRT: hypoxanthine phosphoribosyltransferase; LDH: lactate dehydrogenase; LYZ: lysozyme; RFU: relative fluorescence units; TPMT: thiopurine S-methyltransferase; XDH: xanthine dehydrogenase.

Fig 5: Model performance in the validation cohort.a Comparison of AUROC for different proteomics-based models in predicting SI; b, c Comparison of balanced accuracy and F1-score across models for SI prediction; d Comparison of AUROC for different proteomics-based models in predicting 28-day mortality; e, f Comparison of balanced accuracy and F1-score across models for 28-day mortality prediction; g Confusion matrix for SI prediction; h Confusion matrix for 28-day mortality prediction; i Comparison of the expression levels of four selected proteins between patients with (n = 20) and without secondary infection (n = 40) in the validation cohort 1. For all statistical analyses, n denotes the number of biological replicates, defined as independent human subjects. Statistical analyses were performed by a two-tailed unpaired Student’s t test, and data were presented as mean ± SD. Statistical significance was observed for LYZ (P = 0.0005), SERPIND1 (P = 0.0030), CALM1 (P = 0.0295), and DPT (P = 0.0080). The center line denotes the median; box limits indicate the upper and lower quartiles; whiskers extend to 1.5× the interquartile range; points represent individual samples. Source data are provided as a Source Data file.