JAK inhibitors alleviate EGFR-inhibitor-induced diarrhea by protecting intestinal stem cells from adaptive-immune-exacerbated injury

Yuan Cheng; Chenyue Xu; Dazhao Lv; Meng Tian; Haoyu Wang; Bingxue Yang; Leying Chen; Qing You; Shiyi Zhang

doi:10.1038/s41467-026-71739-8

. 2026 Apr 13;17:5151. doi: 10.1038/s41467-026-71739-8

JAK inhibitors alleviate EGFR-inhibitor-induced diarrhea by protecting intestinal stem cells from adaptive-immune-exacerbated injury

Yuan Cheng ¹, Chenyue Xu ¹, Dazhao Lv ¹, Meng Tian ¹, Haoyu Wang ¹, Bingxue Yang ¹, Leying Chen ¹, Qing You ^1,^✉, Shiyi Zhang ^1,^✉

PMCID: PMC13249821 PMID: 41974707

Abstract

While intestinal stem cells (ISC) are essential for epithelial homeostasis, their dynamic regulation during immune-mediated injury remains undefined. Here we show that suppression of jejunal ISC proliferation contributes to pathology arising from oral EGFR (epidermal growth factor receptor) tyrosine kinase inhibitor (TKI) treatment. Suppression of adaptive immunity via genetic intervention reverses ISC suppression and accelerates mucosal repair via inhibiting the TKI-induced, chemokine-directed migration of T and B lymphocytes from Peyer’s patches. Spatial transcriptomics reveals enhanced crosstalk between adaptive immune cells and ISCs in the jejunum. Ex vivo modelling demonstrates that activated T cells directly impair ISC survival through IFN-γ and TNF, with IFN-γ-induced JAK (Janus kinase) /STAT signaling serving as a critical downstream effector. Accordingly, targeted JAK inhibition mitigates EGFRi (epidermal growth factor receptor inhibitor)–induced diarrhea without substantially compromising antitumor efficacy. This work thus redefines TKI-induced enteropathy as an immune-driven pathology and identifies JAK inhibition as a mechanism-based supportive management of targeted therapy toxicities.

Subject terms: Diarrhoea, Drug development, Mucosal immunology, Targeted therapies

Immune-related adverse effects often hinder targeted therapies, such as severe diarrhoea caused by EGFR tyrosine kinase inhibitor treatment. Here authors show that EGFR inhibition induces diarrhoea via changes in the jejunal chemokine milieu, facilitating migration of T and B lymphocytes from Peyer’s patches to the jejunum, which results in inhibited intestinal stem cell proliferation via IFN-γ-induced JAK/STAT signalling.

Introduction

Intestinal stem cells (ISC) are pivotal regulators of physiological self-renewal and post-injury regeneration in the gastrointestinal tract^1,2. The Lgr5+ stem cell at the crypt base remains the most archetypal population for its capacity to regenerate all epithelial lineages in the small intestine^3,4. While the ISCs compartment is essential for maintaining epithelial integrity and facilitating regeneration following injury, emerging evidence highlights the role of immunological influences in tissue repair^1,5,6.

As key adaptive immune components residing in the lamina propria, T and B lymphocytes have both been reported to actively participate in tissue homeostasis, microbial containment, and early pathogen defense^7,8. Paradoxically, during tissue damage, T cells secrete inflammatory cytokines that mediate substantial gastrointestinal injury^9,10. The prototypical proinflammatory cytokines IFN-γ and TNF-α associate with Paneth cell niche disruption^11–13. IFN-γ reduces epithelial proliferation in murine colitis models^14,15, while supraphysiological TNF-α (>20 ng/mL) markedly suppresses organoid-forming capacity. In discrete reports, expanded B cell populations in inflammatory bowel disease patients impair ISC-epithelial interactions critical for mucosal healing compared to healthy controls^16,17. Current research on T and B lymphocytes has documented their dualistic regulatory effects on intestinal homeostasis, manifesting both protective and pathogenic potentials. However, the specific roles and migration mechanisms of these adaptive immune cells remain poorly delineated. Furthermore, a comprehensive analysis integrating both T/B lymphocyte subsets in GI pathophysiology remains conspicuously absent. To address these knowledge gaps, we aim to systematically investigate the crosstalk between ISCs mentioned above and adaptive immunity, along with its underlying molecular circuitry, during epidermal growth factor receptor inhibitor (EGFRi)-mediated gastrointestinal injury.

The epidermal growth factor receptor (EGFR) represents the first validated growth factor receptor target for cancer therapeutics¹⁸. Therapies targeting the EGFR signaling pathway have been extensively employed in treating a variety of cancers, including non-small cell lung cancer (NSCLC), colorectal cancer, pancreatic cancer, and head and neck squamous cell carcinoma^18–20. However, patients receiving EGFRi, such as afatinib, face a high incidence—up to 75%—of treatment-related diarrhea^21,22. In severe cases (≥grade 3, as defined by the NCI Common Terminology Criteria for Adverse Events, NCI CTCAE), intestinal toxicity may compromise treatment efficacy by necessitating dose delays, reductions, or even permanent discontinuation. The clinical imperative to address EGFRi-induced diarrhea extends beyond symptomatic management. Current guidelines lack a mechanistic rationale, relying on reactive measures such as loperamide, adsorbents, and electrolyte replacement^23,24. Therefore, elucidating the toxicological mechanisms underlying EGFR inhibitor-induced diarrhea is critically important. While damage to the intestinal crypts and subsequent intestinal inflammation have been implicated as potential causes^25,26, the precise crypt compartment affected and the relationships among these cellular populations during intestinal injury remain poorly understood.

Through integrated transcriptomic and proteomic analysis of an EGFRi-induced rat diarrhea model, we identify ISC proliferation suppression and adaptive immune hyperactivation as key pathogenic mechanisms across diarrhea progression stages. Oral EGFRi administration inhibits ISC proliferation, inducing villus atrophy while triggering massive release of immune chemokines that drive excessive recruitment of T/B cells. These activated lymphocytes exacerbate intestinal damage through JAK (Janus kinase)-STAT-mediated inflammatory cytokine production. Crucially, pharmacological blockade of JAK-STAT signaling prevents EGFRi-induced diarrhea in rats. Our findings thus reveal therapeutic targets for mitigating this dose-limiting adverse effect.

Results

Impaired jejunum contributes to EGFRi-induced diarrhea

To elucidate the mechanism underlying EGFR inhibitor-induced diarrhea, we first established a grading system based on rat and mice fecal morphology and perianal condition (Supplementary Fig. 1A). We then constructed rat diarrhea models using different types of EGFR inhibitors (osimertinib, gefitinib, afatinib) (Fig. 1A and Supplementary Fig. 1B, C). Among these agents, afatinib-induced diarrhea showed the most consistent and stable features and was selected as the representative drug for subsequent investigations. To recapitulate the clinical features of EGFRi-induced diarrhea within a shortened experimental timeframe, we developed a short-term EGFRi-induced diarrhea model. Based on linear extrapolation, each 1 mg/kg oral dose corresponds to an AUC of 325 nmol h L⁻¹ in rats²⁷. The doses of 20, 50, and 100 mg/kg correspond to rat AUC₀–₂₄ values of approximately 6500, 16250, and 32500 nmol h L⁻¹ (Table S1), respectively, which are approximately 1.8×, 4.5×, and 9.0× the clinical exposure of 60 mg (AUC ≈ 3622 nmol h L⁻¹)²⁸ Our direct observations from the in vivo dose exploration showed that 20 mg/kg induced Grade I diarrhea (Supplementary Fig. 1D); 50 mg/kg induced Grade I–III diarrhea (achieving the clinically relevant diarrhea severity profile we aimed to model); while 100 mg/kg resulted in significant lethality (Supplementary Fig. 1E). Therefore, we ultimately selected 50 mg/kg as the modeling dose (Fig. 1B, C). Based on the modeling data, diarrhea occurred within three days of oral afatinib administration, accompanied by a gradual increase in intestinal permeability over time (Fig. 1D). Anatomical examination showed that lesions were primarily localized in the rat jejunum (intestinal wall thinning, edema, hyperemia), with mild damage in the ileum (mild hyperemia), while the duodenum and colon showed no significant injury (Fig. 1E and Supplementary Fig. 1F). H&E staining of the duodenum, ileum, and colon corroborated the anatomical findings, showing no significant damage in the duodenum or colon, while mild damage was observed in the ileum in the later stages of diarrhea (Supplementary Fig. 1G). Hematoxylin and eosin (H&E) staining revealed progressively severe villous atrophy in the jejunum of afatinib-treated rats (Fig. 1F), with typical features of mucosal thinning, exposure of the lamina propria, and a decrease in goblet cell numbers (Fig. 1G). Immunofluorescence analysis (EdU and Ki67) showed a gradual reduction in cell proliferation within the intestinal crypts in afatinib-treated jejunum tissues (Fig. 1H; Supplementary Fig. 1H, I).

Fig. 1 — A Schematic representation of the treatment schedule for rats. B Diarrhea severity scores in rats after mid-dose afatinib treatment (n = 6 rats per group, 50 mg/kg). This was repeated n = 3 independent times with similar results. C Changes in body weight of rats following different dose afatinib administration (n = 6 rats per group), low-dose 20 mg/kg; high-dose 100 mg/kg. Rats that lost more than 30 percent of their body weight were euthanized. This was repeated n = 3 independent times with similar results. D Alterations in intestinal permeability, as indicated by changes in the levels of FITC, D-LA, and LPS in the blood, consistent results were observed across five biological replicates (n = 5 rats per group)⁶¹. E Representative images of the jejunum in rats on day 10 post-treatment. MC: methylated cellulose (**F–H**) Temporal changes in small intestinal villi morphology, assessed by H&E staining, EdU immunostaining, and PAS staining (goblet cells), consistent results were observed across five biological replicates (n = 5 rats per group). Scale, 500 μm (F, G). Scale, 100 μm (H). F Statistical analysis of villus length over time. G Statistical analysis of goblet cell numbers over time. H Statistical analysis of crypt proliferating cell numbers (EdU) over time. Data are shown as individual values with group means ± SEM. P-values were determined by unpaired two-tailed Student’s t tests (D, F, G, H). n.s. not significant. Source data are provided as a Source Data file.

Similarly, tumor-bearing mice administered continuous oral afatinib also developed diarrhea symptoms and villous atrophy comparable to those observed in rats (Fig. 2A–D and Supplementary Fig. 1A). However, due to the relatively small size of C57BL/6 mouse feces and the subtle nature of perianal soiling, it was difficult to confirm and observe the grading of diarrhea. Therefore, we chose rats, which are easier to observe, as the primary model for diarrhea studies.

Fig. 2 — A Schematic representation of the treatment protocol for tumor-bearing mice. C57BL/6 mice were subcutaneously injected with 1 × 10⁶ Lewis lung carcinoma (LLC1) cells to establish the syngeneic model. After tumors reached approximately 50–70 mm³, mice were randomly assigned to two groups (Ctrl and Afa) as indicated. B Diarrhea severity scores in tumor-bearing mice after afatinib treatment (n = 5 mice per group). This was repeated n = 3 independent times with similar results. C Representative images of the jejunum of tumor-bearing mice on day 10, showing both the full length and localized changes (n = 5 mice per group). D Quantitative analysis of the morphological changes of intestinal villi, consistent results were observed across five biological replicates (n = 5 rats per group). E Schematic depiction of small intestinal organoid isolation and culture. F Representative images of organoids treated with EGFRi for 12 and 24 h. Measurement of intestinal organoid proliferation through EdU staining. This was repeated n = 3 independent times with similar results. Scale, 200 μm. G Quantitative analysis of EdU incorporation in intestinal organoids to evaluate proliferative activity. n = 3 biological replicates. This was repeated n = 2 independent times with similar results. Data are shown as individual values with group means ± SEM. P-values were determined by unpaired two-tailed Student’s t tests (D), and one-way ANOVA with Tukey’s post hoc test (G). n.s. not significant. Source data are provided as a Source Data file.

In both rat and mouse diarrhea models, we observed significant suppression of crypt cell proliferation in the small intestine. To determine whether villus atrophy resulted from increased apoptosis, we examined jejunal tissues for apoptotic markers, including TUNEL staining and cleaved caspase-3 expression. Notably, during the first three days of treatment, neither TUNEL-positive cells nor cleaved caspase-3 levels were elevated; instead, both showed a modest decrease. These findings suggest that villus atrophy was not caused by enhanced epithelial apoptosis (Supplementary Fig. 1J). To further investigate the impact of EGFR inhibitors on crypt proliferation, we established an in vitro small intestine organoid model (Fig. 2E). Prolonged afatinib treatment markedly reduced cell proliferation in the crypts, as assessed by EdU incorporation, indicating significant inhibitory effects in this region (Fig. 2F, G). The staining results of the viability of the organoids further indicate that the Edu results are not caused by the differences in the viability of the organoids (Supplementary Fig. 1K, L).

These findings indicate that EGFR inhibitors impair proliferation in the jejunal crypts and that damage to the jejunum may serve as the primary cause underlying the development of diarrhea.

Multi-omics reveals the critical role of mucosal immunity at different stages of diarrhea

To better understand the pathophysiological characteristics of EGFR inhibitor-induced diarrhea, we conducted high-throughput RNA sequencing (RNA-Seq) analysis on jejunum samples from rats treated with afatinib or solvent at various stages of disease progression. In the early phase (Day 1, Day 2, and Day 3), visible pathological changes were first observed in the jejunum of afatinib-treated rats (Fig. 1F). Given that the jejunum was the primary site of early pathology, it was chosen for further omics investigation. Using a threshold of |log2FC| >. 1 and p < 0.05, we identified 145 genes that were downregulated and 493 genes that were upregulated in the afatinib group (Fig. 3A). Functional enrichment analysis revealed that early-stage diarrhea was primarily associated with inflammatory responses, with innate and adaptive immune cells, playing significant roles (Fig. 3B). The JAK-STAT signaling pathway, a major downstream pathway involved in cytokine-cytokine receptor interactions, was also identified as a critical regulator in this process (Fig. 3B)²⁹. Further analysis of differentially expressed genes (DEG) within the JAK-STAT module showed that most related genes were upregulated in the jejunum of afatinib-treated rats, including a range of inflammatory cytokines, immune cell chemokines, and key molecules such as Pattern Recognition Receptors (PRR) (Supplementary Fig. 2A). These findings suggest that the JAK-STAT pathway is activated at an early stage of the disease progression, with inflammation representing an early hallmark of afatinib-induced diarrhea.

Subsequently, we compared the DEGs across different stages of diarrhea (early Day 1–Day 3, mid Day 4–Day 6, and late stages Day 7–Day 10) (Supplementary Fig. 2B–D). Overlapping DEGs across these stages were selected for functional enrichment analysis. Since Gene Ontology biological process (GO:BP) terms can encompass both positive and negative regulators, we separately analyzed the upregulated and downregulated sets of overlapping DEGs. Specifically, we first identified upregulated DEGs in the afatinib-treated (Afa) groups across the early, mid, and advanced stages and visualized their overlap using a Venn diagram (Fig. 3C). To determine biological processes that are shared across multiple stages, we performed functional (pathway) enrichment analysis on DEGs derived from the overlapping regions of the Venn diagram, encompassing both pairwise-shared and commonly shared genes across the three stages (3 + 26 + 163 + 25 genes; Fig. 3D). The results showed that these differentially expressed genes (DEGs) were enriched in immune-related pathways, particularly those regulating B and T lymphocyte activation, which is consistent with the results in Fig. 2B. Similarly, downregulated DEGs across the three stages were analyzed using a Venn diagram (Fig. 3E), followed by functional (pathway) enrichment analysis of downregulated DEGs derived from all overlapping regions of the Venn diagram, including genes shared between any two stages as well as those shared across all three stages (1 + 2 + 591 + 1 genes; Fig. 3F). In contrast to the upregulated genes, these shared downregulated DEGs were mainly enriched in pathways related to metabolic processes, with the strongest suppression observed in the mid and advanced stages of diarrhea. To further delineate the temporal characteristics of mucosal immune activation, we examined the evolution of innate and adaptive immune signatures across these stages. Pathways associated with innate immunity were transiently elevated during the early phase, reflecting an acute epithelial stress response. However, these innate immune pathways subsided thereafter, while genes related to adaptive immunity, particularly those linked to T- and B-cell activation, remained persistently upregulated through the mid and late stages of disease progression. This pattern suggests that innate immune activation serves primarily as an initial trigger, whereas sustained epithelial injury is largely maintained by prolonged adaptive immune responses.

We also performed enrichment analysis on differentially expressed proteins (DEP) from the early and mid-stages. As shown in Fig. 3G, proteomic data revealed that the early-stage Afa group was primarily associated with metabolic pathways, while highly expressed proteins in the mid-stage Afa group were enriched in the JAK-STAT signaling pathway, demonstrating the robustness of our findings. B cells played pivotal roles at all stages, with their numbers significantly increased (Fig. 3H). Notably, changes in CD4⁺ T cells were observed throughout all stages, while alterations in CD8⁺ T cells became evident only in the advanced-stage of diarrhea (Fig. 3I).

To further characterize the immune landscape, we performed immune cell deconvolution analyses based on integrated transcriptomic data. The results demonstrated that adaptive immune cells were the most transcriptionally active populations following EGFR inhibitor treatment. Specifically, multiple T-cell subsets, including regulatory and helper T cells, showed significant expansion and activation, while B-cell populations—particularly follicular and germinal-center B cells—also displayed marked upregulation of activation markers (Supplementary Fig. 2G-I). In contrast, macrophage- and dendritic cell–related gene signatures remained at relatively low levels throughout the disease course, indicating that innate immune activation was limited to early phases.

Given that the correlation between mRNA and protein expression is approximately 0.3–0.5, examining mRNA expression alone may not fully capture the molecular dynamics involved³⁰. The combined analysis of both transcriptome and proteome further reinforced that T and B cells are crucial players (Fig. 3J, K and Supplementary Fig. 2E, F) in both the early and mid-stage of diarrhea, with the JAK-STAT pathway serving as a central mediator of these immune responses. While the enrichment of the JAK-STAT pathway observed in Supplementary Fig. 2F did not reach the conventional statistical significance threshold, it displayed a trend toward significance. Furthermore, this pathway was identified among overlapping DEGs/DEPs, exhibiting consistent upregulation trends at both the transcriptomic and proteomic levels during mid-stages. These findings indicate that the sustained immune activation observed during the middle and late stages of EGFRi-induced diarrhea is primarily driven by adaptive immune responses, rather than by prolonged innate immune dysregulation.

T and B cells exacerbate EGFRi-Induced intestinal injury

Multi-omics integrated analyses suggest a strong association between T and B cell activation and the progression of diarrhea. Considering the availability of relevant immunodeficient mice, we chose mice for the experiments. To further investigate the role of these immune cells in EGFRi-induced diarrhea, we employed three distinct immunodeficient mouse models—T cell-deficient, B cell-deficient, and T, B cell-deficient mice. Mice of all three genotypes were treated and subsequently assessed for small intestinal villus damage and the severity of diarrhea. Histological analysis (H&E staining) and Ki67 immunofluorescence revealed that, in the absence of either T or B cells, the degree of small intestinal villus atrophy in mice treated with oral afatinib was significantly alleviated, and intestinal stem cell proliferation was only mildly suppressed. Notably, T, B cell-deficient mice exhibited minor villus atrophy following afatinib treatment, with almost no effect on intestinal stem cell proliferation (Fig. 4A–C). The incidence of diarrhea in these immunodeficient mice closely mirrored the extent of villus damage (Supplementary Fig. 3A).

Fig. 4 — **A–C** Intestinal changes in three immunodeficient mouse models following 10 days of oral afatinib treatment, consistent results were observed across five biological replicates (n = 5 mice per group). H&E staining. Scale, 500 μm. A Ki67 immunostaining. Scale, 100 μm. B, and statistical analysis of villus length and crypt proliferating cell numbers **(C)** are shown. D Experimental schematic: Immunodeficient mice were euthanized at the 6-day time point post-afatinib treatment, and the same initial number of crypts were isolated for organoid culture. E, G Status of intestinal organoids from normal and two types of immunodeficient mice after oral afatinib treatment. This was repeated n = 2 independent times with similar results. F Representative images of organoid growth status. H Quantification of the number of organoids per well following co-culture with CD4⁺ and CD8⁺ T cells at different time points (biological replicates, n = 3). This was repeated n = 2 independent times with similar results. Data are shown as individual values with group means ± SEM. P-values were determined by unpaired two-tailed Student’s t tests (C), and two-way ANOVA with Tukey’s post hoc test (H). Source data are provided as a Source Data file.

In addition, treatment of normal mice with Bruton tyrosine kinase inhibitor (BTKi) in conjunction with oral EGFRi similarly resulted in partial alleviation of small intestinal villus damage (Supplementary Fig. 3B). These findings collectively suggest that the activation of T and B cells plays a vital role in EGFRi-induced diarrhea, exacerbating villus damage and promoting the onset of diarrhea in the small intestine.

To further explore the impact of immune cells on intestinal stem cell proliferation, we conducted additional experiments as illustrated in the Fig. 4D. After six days of oral afatinib treatment, the two immunodeficient mouse models and one control group were euthanized, and their small intestines were dissected. Crypts from the jejunal segments were isolated and cultured as organoids, and their growth was subsequently monitored and quantified. The results revealed that, following afatinib treatment, crypt growth was significantly impaired in normal mice compared to T cell-deficient and B cell-deficient mice. Moreover, the proportion of collapsed organoids in the crypts of normal mice was notably higher than in the other two groups (Fig. 4E-G).

In parallel, we developed an in vitro co-culture system to investigate the interaction between immune cells and the intestinal stem cell niche. To further delineate the contribution of different T-cell subsets, CD4⁺ and CD8⁺ T cells were separately isolated and co-cultured with small intestinal organoids. Both activated CD4⁺ and CD8⁺ T-cell subsets significantly suppressed organoid growth (Fig. 4H and Supplementary Fig. 3C–D).

Investigating the crosstalk between T, B cell and stem cells using spatial transcriptomics

To systematically investigate cellular remodeling and spatial heterogeneity in the early phase of drug-induced diarrhea, we employed the 10× Visium HD platform for high-resolution spatial transcriptomic analysis of jejunal tissues from afatinib-treated mice. As outlined in Fig. 5A, B, two biological samples were analyzed. To strengthen the spatial analysis, an unbiased clustering approach was applied prior to deconvolution (Supplementary Fig 4A). Then, spatial mapping of cell types was performed using the Robust Cell Type Decomposition (RCTD) algorithm, which leveraged annotated scRNA-seq data to deconvolve spatial transcriptomics profiles^31,32. The correlation analysis between spatial transcriptomics and single-cell data revealed a strong association among intestinal epithelial cells, goblet cells, and intestinal stem cells (Supplementary Fig. 4B), thereby cross-validating and confirming the robustness of our cell type annotation approach. By integrating spatial features of intestinal architecture and histological information from H&E staining, we identified 12 major cellular populations, including B cells, T cells, myeloid cells, goblet cells, enterocytes (mature and immature), and intestinal stem cells (Fig. 5B, Supplementary Fig 4B). To further validate the identity of these populations, we conducted gene set enrichment scoring for marker genes associated with cell types of interest, which corroborated the classification of the 12 cell clusters (Supplementary Fig. 4C).

Fig. 5 — A Schematic workflow of spatial transcriptomics (ST) analysis. B Unsupervised clustering of all spatial spots across two tissue sections identifies 12 distinct clusters with spatial distribution patterns (left). Magnified view of the boxed region alongside the corresponding H&E field (middle). Intestinal regions were anatomically segmented based on structural landmarks (right). C Differential intercellular interaction networks. Red and blue lines denote increased and decreased interaction frequency/strength in the experimental group relative to controls, respectively, with line thickness proportional to the magnitude of change. D Comparative heatmap of T/B cell–intestinal stem cell interactions. E Bubble plot illustrating ligand-receptor pairs governing T/B cell–stem cell crosstalk. Rows: Ligand-receptor gene pairs; columns: interacting cell type pairs. F Gene Set Enrichment Analysis (GSEA) comparing KEGG pathways and GO terms between groups from the spatial transcriptomics. G Gene Ontology enrichment profiles of T/B cells and intestinal stem cells. H The spatial localization information of T and B cells with Jak pathway activity. I Violin plots depicting chemokine expression levels across cell types from the spatial transcriptomics. x-axis: cell types; y-axis: normalized gene expression values. The center bar represents the median value; box limits indicate the 25th and 75th percentiles (interquartile range, IQR); whiskers extend to the minimum and maximum values within 1.5 times the IQR; data points beyond this range are shown as outliers. Statistical significance was assessed using a two-sided Wilcoxon rank-sum test. Source data are provided as a Source Data file.

Those clusters of interest, including T and B cells, are predominantly enriched in the mucosal layer, whereas intestinal stem cells reside at the base of the crypts. (Fig. 5B and Supplementary Fig 4D). Subsequent analyses focused on these two specific compartments. Analysis using the widely adopted CellChat V1 framework demonstrated significant rewiring of jejunal cellular communication networks between control and afa-treated groups (Supplementary Fig. 4E). Differential interactions were observed among 11 cell subsets (Fig. 5C). Consistent with the integrated transcriptomic and proteomic (“multi-omics”) analysis, the spatial transcriptomic data also revealed robust T/B cell–stem cell interactions. (Fig. 5D and Supplementary Fig. 4F). These dynamics were governed by five key ligand-receptor pairs: Lgals9-Lgals1 and App-Cd74 mediated immune cell communication to coordinate phased immune responses^33,34; Guca2b-Gucy2c modulated T cell-stem cell interactions to maintain intestinal homeostasis^35,36; while Col4a1-Sdc1 and Cdh1 homotypic binding regulated stem cell behavior through adhesion, migration, and tissue structural organization (Fig. 5E)^37,38. Validation with the updated CellChat V2 algorithm confirmed conserved alterations in T/B cell-stem cell communication during early afatinib-induced diarrhea, reinforcing the robustness of these findings (Supplementary Fig. 4G, H). While such approaches can provide many insights into tissue dynamics and cell-cell interactions, some clarifications should be noted. Spatial transcriptomics do not provide accurate single-cell resolution, and spatial spots are likely to encompass a mixture of cells, especially in a complex tissue architecture like the small intestine. It should be noted, however, that these inferred ligand–receptor interactions are computational predictions based on transcriptomic correlations rather than experimentally validated mechanisms. Therefore, while such integrative analyses provide valuable insights into tissue dynamics and cell–cell communication, their functional relevance requires further experimental confirmation.

For the three major cell populations of interest—T cells, B cells, and intestinal stem cells, we performed pathway enrichment analyses based on GSVA and GSEA to infer transcriptional-level changes in pathway activity. GSVA reflects sample-level variation in pathway activity, whereas GSEA evaluates coordinated gene set enrichment across the entire expression profile; both methods provided complementary evidence for transcriptional-level inferred pathway alterations rather than direct functional validation. Both analyses suggested that, during early jejunal injury, gene programs associated with T/B cell chemotaxis appeared to be enhanced, with T and B cells showing transcriptional enrichment of JAK–STAT signaling pathways, accompanied by downregulation of gene programs related to intestinal stem cell growth (Fig. 5G, F). GSVA and GSEA offer transcriptional-level pathway activity inferences from spatial transcriptomic data, which do not represent direct functional evidence, while complementary multi-omics datasets (transcriptomic and proteomic) provide contextual information that guided the interpretation of these findings. Using the JAK-STAT signaling pathway gene set (KEGG: map04630; link: https://www.genome.jp/entry/map04630), we performed AddModuleScore analysis separately in both groups. The top 40% of bin8s with the highest scores were selected and intersected with T cell and B cell populations, respectively, to define the T cell–JAK-STAT and B cell–JAK-STAT subsets. Leveraging spatial information, we further demonstrated that T cell–JAK-STAT and B cell–JAK-STAT subsets are predominantly enriched along the villi and tend to localize in closer proximity to the submucosal layer (Fig. 5H). GO enrichment and chemokine profiling highlighted CCL19, CCL20, CXCL12, and CXCL13 as key regulators of immune cell recruitment. These chemokines are primarily produced by subsets of intestinal epithelial cells, enteroendocrine cells, and fibroblasts (Fig. 5I). These findings provide multidimensional insights into the molecular pathogenesis of drug-induced diarrhea and highlight the potential involvement of T/B cell–intestinal stem cell crosstalk, while functional validation remains necessary to confirm these transcriptionally inferred mechanisms, whose biological relevance is further supported by experiments presented in preceding and following figures.

JAK inhibitors attenuate intestinal inflammatory signaling

The gut-associated lymphoid tissue (GALT) plays a pivotal role in maintaining intestinal homeostasis, comprising primarily of Peyer’s patches, mesenteric lymph nodes, and lymphocytes within the mucosal lamina propria^39,40. During intestinal inflammation, immune cells in the mucosal Peyer’s patches are influenced by chemokines, leading to their migration out of the patches, through the mesenteric lymph nodes, and eventually returning to the intestinal mucosal lamina propria via the bloodstream^41,42. To clearly delineate the changes in T and B cell populations within the various GALT compartments following drug administration, we performed flow cytometry experiments (Supplementary Fig. 5A, B). As shown in the figure, oral administration of afatinib resulted in a marked increase in CD4, CD8 T cells, and B cells within the intestinal lamina propria, while no significant changes were observed in the immune cell populations of the mesenteric lymph nodes (Fig. 6A, B). In contrast, oral administration of JAKi significantly attenuated the expansion of CD4, CD8 T cells, and B cells in the intestinal lamina propria. Moreover, JAKi also inhibited the migration of CD4 T cells and B cells from Peyer’s patches (Fig. 6A, B). Furthermore, to clarify the origin of these infiltrating immune cells, we examined α4β7 expression—a key integrin mediating gut-homing of lymphocytes derived from Peyer’s patches. afatinib treatment markedly increased the proportion of α4β7⁺ immune cells in the jejunal lamina propria, whereas this increase was substantially reduced by JAK inhibition (Fig. 6C, D and Supplementary Fig. 6A). These findings suggest that a significant fraction of the infiltrating lymphocytes in the lamina propria originates from Peyer’s patches.

Fig. 6 — Changes in immune cell populations in the Lamina Propria, Peyer’s patches, and mesenteric lymph nodes of mice following oral drug treatment. T cell alterations (A), B cell alterations (B). n = 6 biological replicates. C, D The changes in the expression of α4β7 in immune cell populations in the Lamina Propria and Peyer’s patch lymph nodes of mice after oral administration. Changes in T cell expression of α4β7 (C) and changes in B cell expression of α4β7 (D). n = 6 biological replicates. E Changes in chemokine expression in the jejunum of rats during the first three days post-drug administration (qPCR). n = 3 biological replicates. This was repeated n = 2 independent times with similar results. F Changes in chemokine expression in small intestinal organoids at three time points following drug treatment (ELISA). n = 3 biological replicates. This was repeated n = 3 independent times with similar results. G Changes in inflammatory cytokine levels in the jejunum of mice during the first three days post-treatment (qPCR). n = 3 biological replicates. This was repeated n = 2 independent times with similar results. H Representative images and quantitative statistics of the viability of organoid and immune cell co-culture after three days of treatment with different drug inhibitors. n = 3 biological replicates. This was repeated n = 2 independent times with similar results. Data are shown as individual values with group means ± SEM. P-values were determined by one-way ANOVA with Tukey’s post hoc test (A–D), unpaired two-tailed Student’s t tests (E–G) and two-way ANOVA with Sidak’s post hoc test (H). Source data are provided as a Source Data file.

In the context of intestinal inflammation, inflammatory cytokines and chemokines act as key mediators and bridges for immune cell activation, and are considered central to the pathogenesis of the disease, contributing to its initiation, progression, and eventual resolution^43,44. To investigate the effects of drug treatment on chemokines and chemokine receptors, we first assessed their expression in the intestine and organoids. Our results demonstrate that JAKi significantly downregulates the expression of common T and B cell chemokines, including CXCL13, CCL19, CCL20, and CCL21, both in the intestine and in organoid cultures (Fig. 6E, F and Supplementary Fig. 6B, C). This reduction in chemokine levels suggests a possible association with decreased chemotaxis and aggregation of T and B cells in the intestine. Furthermore, the decreased expression of chemokine receptors (CXCR5, CCR6, CXCR4, and CCR7) on B cells further substantiates this observation (Supplementary Fig. 6D). TNFα and IFNγ are well-established inflammatory cytokines that have been reported to induce intestinal damage⁴⁵. Consistent with this, our data show that oral afatinib significantly increases the expression of both TNFα and IFNγ in the intestine, whereas JAKi markedly reduces their expression (Fig. 6G). This conclusion was also validated in an in vitro co-culture model of intestinal organoids and immune cells, where both JAKi and TNF-α blockade effectively alleviated the growth suppression of intestinal organoids (Fig. 6H). In addition to IFN-γ and TNF-α, which emerged as the predominant mediators of intestinal injury, our subsequent analyses also revealed a mild induction of IL-15 expression following EGFR inhibitor treatment. Although cytokines such as IL-2, IL-6, and IL-9 remained largely unchanged, the suppression of IL-15 by tofacitinib suggests that JAK inhibition may also modulate secondary pro-inflammatory signals beyond IFN-γ and TNF-α, contributing to the overall protective effect on intestinal homeostasis (Supplementary Fig. 7A–E). Collectively, these findings suggest that JAK inhibition can attenuate immune cell-mediated intestinal damage by reducing the expression of key inflammatory cytokines.

Oral administration of JAK inhibitors ameliorates diarrhea in a rat model

We further assessed the effects of tofacitinib on the rat diarrhea model that we established. We selected a pan-JAK inhibitor: a broad JAK inhibitor was expected to provide more comprehensive suppression, and tofacitinib, as an orally available and clinically characterized agent, served as a suitable proof-of-concept drug in our model. We selected 50 mg/kg, oral gavage in rats to ensure robust pharmacodynamic coverage within the time frame of our study while remaining within toxicology-derived safety margins (Table S1). Combined oral administration of tofacitinib effectively prevented the onset of diarrhea, reduced its severity, and improved both small intestinal structure and body weight (Fig. 7A–D). ELISA-based analysis of intestinal barrier integrity revealed substantial damage to the barrier in afatinib-treated rats. In contrast, rats co-treated with afatinib and tofacitinib exhibited stable barrier function comparable to the control group (Fig. 7E). Hematoxylin and eosin (H&E) staining revealed progressively severe structural damage in the jejunum of afatinib-treated rats over time. In the lamina propria of the small intestine, T cells (CD3, CD4, and CD8 antibodies) and B cells (CD20 antibody) infiltration were observed in the afatinib group. Oral administration of tofacitinib significantly alleviated the structural damage, crypt proliferation, and immune cell infiltration (Fig. 7F–I and Supplementary Fig. 8A, B). Additionally, we examined the activation of STATs in the rat diarrhea model. After afatinib treatment, levels of phosphorylated STAT1 (P-STAT1) and STAT3 (P-STAT3) were significantly elevated in villus regions. Notably, tofacitinib treatment effectively abolished this activation. The phosphorylation reduction mainly occurred in the immune cells within the lamina propria, with only weak P-STAT signals detected in the tofacitinib-treated group (Fig. 7J). These results also provide opportunities to gain more insights into the impacts of afatinib and tofacitinib on the immune rather than epithelial cells.

Fig. 7 — **(A)** Schematic of the treatment protocol. Male rats were randomly assigned to three groups (n = 5 rats per group): Control group (vehicle and placebo), Afa group (Afatinib, 50 mg/kg, plus placebo), and Afa + Tofa group (Afatinib, 50 mg/kg, plus tofacitinib, 50 mg/kg). This was repeated n = 3 independent times with similar results. **(B)** Changes in body weight of rats across the three groups (n = 5 rats per group). This was repeated n = 3 independent times with similar results. **(C)** Diarrhea severity scores in rats from the three groups (n = 5 rats per group). This was repeated n = 3 independent times with similar results. **(D)** Representative images of the jejunum from rats in all three groups on day 10 post-treatment. **(E)** Changes in intestinal permeability. n = 5 biological replicates. **(F to I)** Temporal changes in small intestinal villi morphology in the three groups, assessed by H&E staining (n = 5 biological replicates) (F) and Edu immunostaining (n = 5 biological replicates) (G and H). At the end of the experiment, tissue samples from the control, Afa, and Afa + Tofa groups were analyzed for CD19 (B cells), CD4 and CD8 (T cells) (I), and p-STAT1 and p-STAT3 immunostaining **(J)**. This was repeated n = 3 independent times with similar results. In a separate experiment, male rats were randomly assigned to five groups (n = 5 rats per group) as indicated in the schematic: Afa + placebo, Afa + Tofacitinib, Afa + Levofloxacin, Afa + Loperamide, and Afa + Octreotide (afatinib plus octreotide injection). Data are shown as individual values with group means ± SEM. P-values were determined by one-way ANOVA with Tukey’s post hoc test (E, F, H). n.s. not significant. Source data are provided as a Source Data file.

Alterations in gut microbiota composition have been shown to influence the development of side effects associated with anti-tumor therapies⁴⁶. We conducted 16S rRNA sequencing to analyze the bacterial composition in rats treated with afatinib, afatinib plus tofacitinib, or PBS. After afatinib treatment, a significant reduction in bacterial community abundance was observed, while tofacitinib treatment restored the microbial balance (Supplementary Fig. 8C). Principal coordinate analysis (PCoA) of microbiome data revealed significant clustering in the microbiota of afatinib-treated rats, which was effectively normalized by tofacitinib (Supplementary Fig. 8D). Furthermore, fecal microbiota composition analysis indicated an increased proportion of harmful bacteria in the afatinib group, while tofacitinib treatment significantly reduced the abundance of harmful bacteria and restored the proportion of beneficial gut microbiota (Supplementary Fig. 8E–G).

Currently, there are no specific interventions targeting EGFRi-induced GI side effects in clinical practice. Loperamide, antibiotics, and octreotide are commonly used to manage targeted therapy-induced diarrhea^47,48. Therefore, we compared the therapeutic effects of tofacitinib, loperamide, antibiotics, and octreotide in the afatinib-induced rat diarrhea model. The results demonstrated that oral administration of tofacitinib and levofloxacin significantly improved both body weight and diarrhea severity in the rats. Loperamide improved diarrhea but only moderately enhanced body weight. Rats treated with octreotide showed no significant improvement in either body weight or diarrhea (Fig. 8A–D and Supplementary Fig. 8H). Histological analysis revealed that only the tofacitinib group showed marked improvements in small intestine structural integrity, crypt proliferation, and inflammation. Loperamide showed moderate improvement in villous atrophy and crypt proliferation but had no effect on inflammation. Levofloxacin and octreotide showed no significant histological improvements (Fig. 8D). In summary, tofacitinib demonstrated superior efficacy compared to current supportive treatments for EGFRi-induced diarrhea in preclinical models. Our conclusion is based on experimental evidence in animal models, and further clinical studies will be required to establish translational relevance.

Fig. 8 — A Changes in body weight across the five groups. This was repeated n = 3 independent times with similar results. B Diarrhea severity scores in rats from the five groups. This was repeated n = 3 independent times with similar results. C Diarrhea incidence across the five groups. This was repeated n = 3 independent times with similar results. D Representative images of the jejunum from rats in all five groups, showing H&E staining, Ki67, CD3 (T cells), and CD19 (B cells). Scale, 100 μm. This was repeated n = 3 independent times with similar results. Data are shown as individual values with group means ± SEM. Source data are provided as a Source Data file.

Oral administration of tofacitinib did not affect the anti-tumor effect of afatinib

Motivated by the promising ability of tofacitinib to prevent afatinib-induced diarrhea, we investigated whether tofacitinib impacts the antitumor efficacy of afatinib.

We first assessed the effects of co-administration of a JAK inhibitor (JAKi) and an EGFR inhibitor (EGFRi) in vitro using two representative lung cancer cell lines. The results demonstrated that afatinib exerted a pronounced inhibitory effect on both PC9 and LLC1 cells, with PC9 exhibiting significantly higher sensitivity to afatinib compared to LLC1 (Fig. 9A, B). This difference is likely attributable to the fact that LLC1 is a ras-mutant lung cancer cell line, rather than an EGFR-mutant one. Moreover, the toxicity profile of afatinib was not altered by its combination with tofacitinib in either cell line (Fig. 9A, B).

Fig. 9 — A Effect of Tofacitinib on the afatinib-induced inhibition of human lung cancer cell proliferation. This was repeated n = 2 independent times with similar results. B Effect of Tofacitinib on the afatinib-induced inhibition of mouse lung cancer cell proliferation. This was repeated n = 3 independent times with similar results. C Schematic of the treatment protocol. Nude or C57BL/6J mice were randomly assigned to three groups (n = 7 mice per group). Control group (vehicle plus placebo), Afa group (afatinib plus placebo), and Afa+Tofa group (afatinib plus tofacitinib). D Representative images of xenograft tumors from nude mice. Scale bar: 1 mm. n = 7 biological replicates. E Tumor volume measurements in nude mice across the four groups. n = 7 biological replicates. F Tumor weights in nude mice across the four groups. n = 7 biological replicates. G Representative images of LLC1 tumors excised from C57BL/6J mice at the end of the experiment. n = 5 biological replicates. H Tumor volume changes in LLC1 tumors in C57BL/6J mice from different groups. n = 5 biological replicates. I Subcutaneous tumor weights in C57BL/6J mice at the conclusion of the experiment. n = 5 biological replicates. J Representative images of tumor-infiltrating CD8 + T lymphocytes in LLC1 tumors. n = 5 biological replicates. Scale, 100 μm. K Quantification of tumor-infiltrating CD8 + T lymphocytes in LLC1 tumors. n = 5 biological replicates. Data are shown as individual values with group means ± SEM. P-values were determined by one-way ANOVA with Tukey’s post hoc test (F, I, K), and two-way ANOVA with Tukey’s post hoc test (E, H). n.s. not significant. Source data are provided as a Source Data file.

We employed the EGFR inhibitor-sensitive non-small cell lung cancer (NSCLC) cell line PC9 as a model. PC9 xenograft-bearing nude mice were treated with a combination of tofacitinib and afatinib (Fig. 9C). The addition of tofacitinib did not modify the antitumor efficacy of afatinib compared to afatinib monotherapy (Fig. 9D). Afatinib treatment, either alone or in combination with tofacitinib, resulted in a significant reduction in tumor volume and weight relative to placebo (Fig. 9E, F), with no significant differences in tumor growth observed between the afatinib and afatinib plus tofacitinib groups (Fig. 9E, F and Supplementary Fig. 9A). Interestingly, an unexpected finding was that tofacitinib monotherapy also exhibited a noticeable antitumor effect in PC9 xenograft–bearing nude mice. This is not an entirely novel observation, as previous studies have reported antitumor activity associated with JAK inhibition^49,50.

In most autoimmune diseases, JAK inhibition induces immunosuppression^51,52. To evaluate whether oral JAK inhibitors suppress immune cells involved in antitumor responses, we established a syngeneic lung cancer model (LLC1) in C57 mice, representing a fully immunocompetent tumor model (Fig. 9C). Afatinib treatment exhibited antitumor activity, although the effect was less pronounced compared to the PC9 model (Fig. 9G). Co-administration of tofacitinib did not impact the antitumor efficacy of afatinib, and tofacitinib monotherapy itself did not promote tumor growth. (Fig. 9H, I).

Investigations of the anti-tumor immune response are also provided in the setting of afatinib and tofacitinib treatment. Cytotoxic CD8 + T cells of the adaptive immune system are the most powerful effectors in the anticancer immune response. Analysis of tumor-infiltrating CD8 + T cells revealed that afatinib treatment alone resulted in a reduction in CD8 + T cells. The combination of afatinib and tofacitinib also led to a similar reduction. This suggests that the addition of tofacitinib did not alter the extent of CD8 + T cell reduction, as shown in Fig. 9J. Additionally, five-part differential blood analysis demonstrated no significant differences in total white blood cell (WBC) counts across the groups (Supplementary Fig. 9B). Likewise, there were no significant differences in total cell counts between the afatinib monotherapy group and the afatinib plus tofacitinib combination group (Supplementary Fig. 9C–G).

In summary, these results suggest that JAK inhibitors may mitigate EGFR inhibitor–induced diarrhea without substantially compromising antitumor efficacy, although further evaluation in additional tumor models and clinical settings is warranted. Moreover, while our in-vivo studies using the NSCLC cell lines PC9 and LLC provide mechanistic insight, they cannot fully predict the long-term impact of tofacitinib on tumor biology. Although tofacitinib did not affect afatinib’s antitumor efficacy in the models tested, these results may not apply to all types of tumors, as different cancers may respond differently to JAK inhibition depending on factors like JAK–STAT pathway activation, immune interactions, and EGFR signaling. Therefore, additional studies in other tumor models with varied characteristics are needed to assess the potential for tumor-type-specific effects. Taken together, these data support the cautious, short-term use of tofacitinib to manage life-threatening afatinib-induced diarrhea, accompanied by vigilant follow-up to minimize any potential risk.

Discussion

Here, we report that the inhibition of EGFR signaling predominantly affects the jejunal segment (Fig. 10). After the suppression of intestinal stem cell proliferation in the jejunum, villous atrophy occurs concurrently with the release of a significant amount of immune cell chemokines (CXCL12, CXCL13, CCL19, CCL20, and CCL21), promoting the migration of T and B cells from the Peyer’s patches to the intestinal lamina propria. Our study systematically delineates the dynamics (migratory trajectories) of intestinal T and B lymphocytes during disease pathogenesis, employing a multidimensional analytical framework that extends beyond mere interaction networks between these lymphocyte populations.

Adaptive immune cells accumulate in the jejunum in an EGFRi-induced diarrhea model, with a notable enrichment and activation of the JAK-STAT pathway. Here, we employed three kinds of distinct immunodeficient mice (T cell-deficient, B cell-deficient, and T/B cell-deficient) to establish an EGFRi-induced diarrhea model. Additionally, we developed an ex vivo co-culture system of intestinal organoids and T cells to delineate the impact of T cells during immune-mediated intestinal stem cell (ISC) injury. Our findings demonstrate that the presence of T and B cells prolongs recovery from ISC proliferation inhibition. These results provide a causal link between the accumulation of adaptive immune cells in EGFRi-induced diarrhea and support the potential therapeutic benefit of JAK inhibitors. This result is consistent with previous findings from the A. M. Hanash laboratory¹⁰. Previous studies have reported that EGFRi suppresses DNA replication and proliferation of Lgr5⁺ ISCs via the MEK pathway; however, this inhibition is rapidly reversible⁵³. These findings also provide indirect evidence suggesting that adaptive immune cells may be a key contributor to EGFRi-induced diarrhea. Furthermore, studies from the Eduardo J. Villablanca laboratory demonstrated that co-culture of B cells with fibroblasts and intestinal organoids similarly showed that B cells exert a negative influence on intestinal injury repair¹⁶. Notably, while our study primarily focused on αβ T cells, we recognize that γδ T cells may also influence epithelial integrity during EGFRi-induced intestinal injury. Future studies employing TCRδ-deficient mice will help clarify their specific contribution.

Intestinal damage mediated by adaptive immune cells, particularly in EGFRi-induced diarrhea, represents a highly localized response involving cellular injury, migration, and effector functions. Given the complexity of this process and the involvement of multiple cell types across specific temporal windows, elucidating the distinct, precise interactions between T/B cells and ISCs remains a formidable challenge. To gain deeper insight, we leveraged the 10× Visium HD platform to generate a spatial transcriptomic atlas of the murine jejunum—the first such atlas of this region using this technology. CellChatV1, a widely used analytical tool, revealed increased communication frequency between T/B cells and ISCs when ISC proliferation was suppressed. Comparative analysis using the latest CellChatV2 confirmed these findings with high concordance. Further, interactions between T/B cells and ISCs were modulated by key ligand-receptor pairs, including Lgals9-Lghm, App-Cd74, Guca2b-Gucy2c, Col4a1-Sdc1, and Cdh1-Cdh1. Moreover, enteroendocrine cells, intestinal epithelial cells, and fibroblasts released large quantities of immune chemokines during this process, facilitating the migration of T and B cells from Peyer’s patches to the jejunal lamina propria and triggering JAK-STAT pathway activation in T and B cells.

The GI effects of IFNγ and TNF-α have been extensively studied in various experimental models, where they have been shown to induce epithelial toxicity through both cell-autonomous and non-autonomous negative regulatory feedback loops^14,54. These key inflammatory cytokines can also directly program ISC death^10,55. Similarly, we observed a marked increase in IFNγ and TNF-α expression during EGFRi-mediated GI injury. In our ex vivo co-culture system of intestinal organoids and T cells, neutralizing antibodies against these cytokines alleviated the growth inhibition of organoids. These findings underscore the critical role of IFNγ and TNF-α from crypt-adjacent donor T cells in mediating direct cytotoxicity to intestinal epithelial cells.

Mounting evidence suggests a close interplay between immune cells and intestinal injury repair. B cells have been implicated in disrupting the interactions between intestinal epithelial cells and ISCs during colonic mucosal healing¹⁶, while T cell-derived IFNγ directly targets ISCs and induces apoptosis via a JAK/STAT-dependent mechanism¹⁰. The intestinal injury process involves a coordinated response of multiple inflammatory cytokines and chemokines, engaging a diverse array of recruited immune cells. Notably, B-cell inhibitors exhibited either negligible or only modest alleviation in our EGFRi-induced diarrhea model, suggesting that targeting individual immune cell populations may be insufficient to ameliorate the pathological intestinal phenotype. Importantly, JAK inhibitors demonstrated remarkable therapeutic efficacy across multiple models, including rat diarrhea models, tumor-bearing mouse models, and ex vivo organoid-immune cell co-culture systems. This study offers a framework for understanding inflammation in EGFRi-induced diarrhea and provides promising therapeutic insights. Given their ability to simultaneously target multiple cytokine signaling pathways and disrupt inflammatory circuits, JAK inhibitors represent a compelling therapeutic option for EGFRi-induced diarrhea.

In summary, through integrative multi-omics analysis and cutting-edge 10× HD spatial transcriptomics, we establish the pivotal role of adaptive immune cells in immune-mediated GI injury. By leveraging evidence from cellular, organoid, and murine models, we comprehensively characterize the specific interactions between T/B cells and ISCs. Our study advances the understanding of EGFRi-induced diarrhea pathogenesis and offers promising clinical intervention strategies, although further validation in patient-based studies is required.

Notably, translating our findings to real-world clinical practice requires careful consideration, particularly with respect to the long-term safety of JAK inhibition in patients with cancer. We do not exclude the theoretical concern that prolonged use of JAK inhibitors, such as tofacitinib, may increase malignancy risk in certain populations. Observational data in rheumatoid arthritis (RA) cohorts have shown a higher incidence of malignancy with tofacitinib compared with TNF inhibitors⁵⁶, and a network meta-analysis similarly reported a modest increase versus TNFi but not versus placebo or methotrexate⁵⁷. However, these findings derive from patients receiving chronic immunosuppression for autoimmune disease and cannot directly demonstrate that tofacitinib increases tumor incidence in individuals who already have cancer. The U.S. FDA’s integrated review further concluded that genetic toxicology and two-year rodent carcinogenicity studies indicated a low risk of direct, drug-induced carcinogenicity because of large exposure margins. However, the review also cautioned that immunosuppression-associated malignancies were observed in long-term monkey toxicology studies and in clinical trials. Our own experiments in NSCLC cell lines PC9 and LLC1 offer mechanistic insight but cannot predict the long-term consequences of tofacitinib exposure on tumor biology. Importantly, the clinical context we address is fundamentally different from chronic autoimmune therapy: tofacitinib was administered as a short-term intervention to control severe, afatinib-induced diarrhea that jeopardized continuation of life-prolonging EGFR-targeted treatment. This time-limited use substantially reduces cumulative immunosuppressive burden, yet still demands vigilant follow-up. While tofacitinib did not reduce afatinib’s antitumor efficacy in the two preclinical tumor models tested (PC9 and LLC), we acknowledge that these results may not be representative of all tumor types. Different malignancies may exhibit distinct JAK–STAT pathway activation, immune microenvironment interactions, and EGFR signaling dependencies, all of which could affect the response to JAK inhibition. Therefore, further studies in additional tumor models, with diverse molecular characteristics and immune environments, are essential to explore potential tumor-type-specific effects. We therefore emphasize that, while our data suggest potential benefit of JAK inhibition for managing EGFR-inhibitor–related toxicity, its application in oncology should remain cautious and carefully monitored until supported by prospective long-term clinical studies.

Nevertheless, this study has several limitations. First, we were unable to establish a direct B cell-intestinal organoid co-culture system, preventing direct assessment of B cell-mediated effects on ISCs. However, existing literature supports the negative impact of B cells on intestinal injury repair¹⁶. Second, while our multi-omics and spatial transcriptomic analyses indicate significant gene expression changes in innate and adaptive immune cells during the early phase of diarrhea, innate immune cells involvement diminishes in later stages, leading us to focus primarily on adaptive immune responses. Third, we did not differentiate between quiescent and actively cycling ISCs. Prior studies employing EGFR flox x Lgr5-CreERT2 models have shown that EGFR inhibition alone does not produce overt macroscopic phenotypes following Lgr5+ cell ablation⁵⁸, raising important questions for further investigation. Lastly, the lack of clinical trial data remains a limitation, necessitating future patient-based studies to validate the therapeutic efficacy of JAK inhibitors in EGFRi-induced diarrhea. While we acknowledge this absence of clinical data, we also recognize the potential limitations regarding the interpretation of gut microbiota changes. Specifically, the observed shifts in gut microbiota may not be primary drivers of EGFR inhibitor (EGFRI)-induced diarrhea and intestinal damage, but rather secondary effects of immune modulation. Indeed, our antibiotic treatment experiment (Fig. 8A–D) showed that, although antibiotics partially alleviated diarrhea, the underlying intestinal villus damage remained severe. This suggests that gut microbiota alterations may be a consequence of immune changes rather than a primary cause. Further studies, including immune biomarkers and histological analysis, will be required to clarify the causality of these shifts.

Methods

Ethics approval

All animal procedures were conducted in strict accordance with the guidelines of the Institutional Animal Care and Use Committee (IACUC) of Shanghai Jiao Tong University (A2023105-002).

Animal

Male Sprague-Dawley rats, aged 6–7 weeks, with body weights ranging from 230 to 260 g at the start of treatment. Male BALB/c nude mice and male C57BL/6J mice, aged 6–7 weeks, with body weights ranging from 18 to 21 g at the start of treatment. Animals that lost more than 30 percent of their body weight were euthanized by carbon dioxide asphyxiation. Only male rats and mice were used in this study to eliminate potential confounding effects of estrogen on the experimental outcomes. Animals were obtained from GemPharmatech Co., Ltd, Shanghai, China, and were group-housed under a 12 h/12 h light/dark cycle (n = 3 and 5), with access to autoclaved tap water and sterile rodent chow.

Cell lines and culture conditions

The human lung cancer cell line PC9 (RRID: CVCL_B260) and Lewis lung carcinoma (LLC1) cells (RRID: CVCL_4358) were purchased from the National Collection of Authenticated Cell Culture (NCACC). None of the above cell lines was contaminated. Both PC-9 and LLC1 were cultured in RPMI Medium 1640 medium (BioEngine) supplemented with 10% fetal bovine serum (FBS), 100 units/ml penicillin, and 100 mg/ml streptomycin. All cells were maintained in 5.0% CO₂ at 37 °C. Both cell lines were authenticated by STR profiling and tested negative for mycoplasma.

Reagents

Afatinib, tofacitinib, ozanimod, and BTKi (Acalabrutinib) were purchased from Goyic, China. The antibodies used for immunohistochemistry and immunofluorescence are as follows: phospho-STAT1 (7649) and phospho-STAT3 (9145) antibodies were from Cell Signaling Technology. Anti-Ki67 antibody (ab279653), anti-CD8 alpha antibody (ab237709), and anti-CD8 alpha antibody (ab217344) were from Abcam. CD19 (6OMP31) and CD3 (eBioG4.18 (G4.18)) were from eBioscience. The antibodies used in the flow cytometry analysis are as follows: CD45 (103132, BioLegend), CD3 (553064, BD Pharmingen), CD4 (100510, BioLegend), CD8 (100737, BioLegend), and CD19 (552854, BD Pharmingen). The secondary antibodies used were from Beyotime. The primary antibody was diluted at a ratio of 1:500, and the secondary antibody was diluted at a ratio of 1:5000.

Diarrhea grade evaluation for rats

Diarrhea grade was quantified by two independent assessors using a comprehensive grading system where 0 = normal, normal stool or absent; 1 = slight, slightly wet and soft stool with mild perianal soiling; 2 = moderate, wet and unformed stool with moderate perianal soiling; 3 = severe, watery stool with severe perianal soiling.

PAS staining

Paraffin sections were sequentially dewaxed in xylene I (20 min) and xylene II (20 min), followed by rehydration through graded ethanol: 100% ethanol I (10 min), 100% ethanol II (10 min), 95% ethanol (5 min), 90% ethanol (5 min), 80% ethanol (5 min), 70% ethanol (5 min), and finally rinsed in distilled water. Sections were oxidized in 0.5% periodic acid for 10 min, washed in running tap water for several minutes, and then rinsed twice with distilled water. Schiff reagent staining was performed in the dark for 15–30 min, followed by a running water wash for 10 min. Nuclei were counterstained with hematoxylin for 1–2 min, differentiated briefly in 1% acid alcohol, and returned to blue under running tap water. Dehydration and clearing were carried out through 95% ethanol I (5 min), 95% ethanol II (5 min), 100% ethanol I (5 min), 100% ethanol II (5 min), xylene I (5 min), and xylene II (5 min). Sections were air-dried briefly and mounted with neutral balsam. Microscopic observation and image acquisition were then performed. Staining results: glycogen and neutral mucins appeared red; nuclei stained blue.

Assessment of Intestinal Permeability

Rats were fasted for 6 h prior to the experiment. Baseline blood samples were collected via retro-orbital bleeding, followed by oral gavage of FITC-dextran (5 kDa, 600 mg/kg). Four hours post-gavage, a second blood sample was collected via the same route. All blood samples were collected into K3-EDTA–coated tubes (Sarstedt, Germany), centrifuged at 4 °C for 15 min at 3000 rpm, and plasma was transferred to clear Eppendorf tubes. Plasma from PBS-treated animals was used to generate a standard curve. Fluorescence intensity of FITC-dextran in plasma (excitation: 488 nm; emission: 520 nm) was measured using a microplate reader. The initial blood sample was analyzed using the corresponding commercial assay kit (D-LA, LPS).

Quantitative measurement of the concentration of D-Lactate (D-LA) and lipopolysaccharides (LPS) in plasma, according to the manufacturer’s recommendations for kits. All ELISA Kits were purchased from Cusabio (CSB-E12634r and CSB-E14247r).

Immunofluorescent staining

Paraffin sections were deparaffinized in xylene I (10 min), xylene II (10 min), and xylene III (10 min), followed by rehydration through absolute ethanol I–III (5 min each) and rinsing in distilled water. Antigen retrieval was performed by heating slides in retrieval buffer using a microwave at high power for 6 min, then at low power for 15 min. Buffer evaporation was minimized to avoid drying of sections. Slides were cooled to room temperature and washed in PBS (pH 7.4) on a decolorization shaker three times for 5 min each. Endogenous peroxidase activity was blocked by incubating the sections with 5% hydrogen peroxide for 25 min at room temperature in the dark. Slides were washed in PBS (pH 7.4) three times (5 min each). After removing excess PBS, nonspecific binding was blocked for 30 min using either 10% normal rabbit serum (for goat primary antibodies) or 3% BSA (for other primary antibodies). Sections were then incubated with the primary antibody at 4 °C overnight in a humidified chamber.

After three PBS washes (5 min each), slides were incubated with HRP-conjugated secondary antibody for 50 min at room temperature, followed by another set of PBS washes (3 × 5 min). Tyramide signal amplification (TSA) reagent was applied and incubated in the dark at room temperature for 10 min, followed by three washes in TBST (5 min each). Nuclei were stained with DAPI for 10 min in the dark and washed again in PBS (3 × 5 min). Autofluorescence was quenched using Reagent B for 5 min, followed by a 10-min rinse under running water. Sections were mounted with an anti-fade mounting medium. Fluorescence images were acquired using a fluorescence microscope.

RNA isolation and library preparation

Rats were treated with or without afatinib for 3, 6, and 10 days, and jejunal tissues were analyzed to examine gene expression profiles. RNA-Seq was used to assess gene expression.

Total RNA was extracted using the TRIzol reagent (Invitrogen, CA, USA) according to the manufacturer’s protocol. RNA purity and quantification were evaluated using the NanoDrop 2000 spectrophotometer (Thermo Scientific, USA). RNA integrity was assessed using the Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA). Then the libraries were constructed using the VAHTS Universal V6 RNA-seq Library Prep Kit according to the manufacturer’s instructions. The transcriptome sequencing and analysis were conducted by Novogene Co., Ltd (Beijing, China) and GENEWIZ Co., Ltd (Suzhou, China).

Quality control of RNA-Seq raw data and differential expression gene analysis

Raw sequencing data (raw reads) from the Illumina 550AR sequencer were processed to filter out low-quality reads. Clean reads from each sample were obtained and used in the following analysis. The RNA reads were aligned against the reference hg19 and gencodev27lift37 database (downloaded from https://www.gencodegenes.org) by STAR (v2.7.8a). Based on the aligned reads, raw count and transcripts per million (TPM) values were calculated in RSEM (v1.3.0). Then, gene expression level was summarized from the transcript level. Differential expression genes (DEGs) were identified by the DESeq2 package. The genes with |log₂FoldChange| > 1.5 and p-value < 0.05 were considered as significantly DEGs. Volcano and heatmap diagrams were conducted in R with ggpubr and the Complexheatmap package. KEGG pathway enrichment was analyzed on the KOBAS-i webtool.

Total protein extraction

A total of 12 rat jejunal segments were utilized for proteomic analysis, distributed as follows: the control group (n = 4) and the Afa group (n = 8). Within the Afa group, samples were further categorized into early-stage (n = 3, corresponding to 3 days of Afa treatment) and mid-stage (n = 5, corresponding to 6 days of Afa treatment). Jejunal tissues were analyzed for protein expression profiles. TMT quantitative proteomics was employed to evaluate protein expression.

The sample was ground individually in liquid nitrogen and lysed with SDT (containing 100 mM NaCl) and 1/100 volume of DTT, followed by 5 min of ultrasonication on ice. Centrifuge at 12,000 g for 15 min at 4 °C, collect the supernatant, heat at 95 °C for 8–15 min, followed by an ice bath for 2 min, then add an adequate amount of IAM solution and incubate in the dark for 1 h. Then samples were completely mixed with 4 times volume of precooled acetone by vortexing and incubated at −20 °C for at least 30 min. Samples were then centrifuged at 12,000 g for 15 min at 4 °C, and the precipitation was collected. After washing with 1 mL cold acetone, the pellet was dissolved completely by Dissolution Buffer (DB buffer).

Protein quality test

BSA standard protein solution was prepared according to the instructions of the Bradford protein quantitative kit, with a gradient concentration ranging from 0 to 0.5 μg/μL. BSA standard protein solutions and sample solutions with different dilution multiples were added into a 96-well plate to fill up the volume to 20 µL, respectively. Each gradient was repeated three times. The plate was added 180 µL G250 dye solution quickly and placed at room temperature for 5 minutes, and the absorbance at 595 nm was detected. The standard curve was drawn with the absorbance of the standard protein solution, and the protein concentration of the sample was calculated. Twenty micrograms of the protein sample was loaded to 12% SDS-PAGE gel electrophoresis, wherein the concentrated gel was performed at 120 V for 20 min, and the separation gel was performed at 150 V for 50 min. The gel was stained with Coomassie Brilliant Blue R-250 and decolored until the bands were visualized clearly.

TMT labeling of peptides

Hundred microliters of 0.1 M TEAB buffer was added to reconstitute, and 20 μL of acetonitrile-dissolved TMT labeling reagent was added, sample was mixed with shaking for 2 h at room temperature. Then, the reaction was stopped by adding 5% ammonia. All labeling samples were mixed with an equal volume, desalted, and lyophilized. For multiple labeled groups, a common reference was created by pooling an equal quantity of each sample.

LC-MS analysis

UHPLC-MS/MS analyses were performed using an EASY-nLCTM 1200 UHPLC system (Thermo Fisher, Germany) coupled with a Q ExactiveTM HF-X (Thermo Fisher, Germany) or Orbitrap Exploris 480 mass spectrometer (Thermo Fisher, Germany) in Novogene Co., Ltd. (Beijing, China).

Differential expression proteins analysis

Differential expression proteins (DEPs) were identified by the DESeq2 package. The proteins with |log₂FoldChange| > 1.5 and p-value < 0.05 were considered as significantly DEPs. KEGG pathway enrichment was analyzed on the KOBAS-i webtool.

Score of infiltration immune cells

Based on the gene expression matrix, the ImmuCellAI package was used to estimate the score of infiltration immune cells. In brief, the SSGSEA algorithm and its associated gene sets are used to estimate cell type scores, which are then corrected using a compensation matrix. This process yields the immune cell scores for each sample.

Visium HD spatial transcriptome sequencing

Freshly collected jejunum tissue was divided into blocks of appropriate size and immediately fixed in 4% FPA and paraffin-embedding. The sections were attached to the slides (Sigma-Aldrich, P0425). Deparaffinization, H&E staining, imaging, and decrosslinking of the sections were performed according to the 10x Genomics Visium HD FFPE Tissue Preparation Handbook (CG000684). Probe hybridization, probe release, and library construction were performed using the Visium HD Spatial Gene Expression Reagent kit from 10x Genomics (PN-1000676 for Mouse, 6.5 mm) according to the User Guide (CG000685). The transcriptome libraries were subjected to high-throughput sequencing with the PE-150 mode. The sequencing and bioinformatics analysis were provided by OE Biotech Co., Ltd. (Shanghai, China).

Spatial transcriptomic profiling using the 10× Genomics Visium HD platform was performed on two samples comprising two tissue sections in total. The transcriptome was captured at single-bin resolution with each bin defined as an 8 × 8 μm grid. Each sample yielded 175,868–194,486 high-quality spatial bins, with average UMI counts ranging from 557 to 629 per bin and median gene detection of 413–453 genes, confirming robust data quality for downstream analyses. Using the Seurat pipeline, we identified the top 3000 highly variable genes (HVG). Following dimensionality reduction via principal component analysis (PCA), UMAP visualization resolved 12 distinct spatial clusters.

Single-cell RNA-seq datasets are available through the Broad Institute Single Cell Portal under accession numbers SCP2760 (RNA-seq). Clusters were annotated for major cell lineages, including enterocytes, GCs, enteroendocrine cells, tuft cells, T cells, and innate lymphoid cells, B cells, myeloid cells and granulocytes, fibroblast/mural and endothelial cells on the basis of known markers such as Muc2 for GCs, Cd3e for T cells, and Igha for IgA B cells³¹.

Leveraging morphological image segmentation and gene expression information, the Bin2cell pseudo-single-cell segmentation approach was also applied to reconstruct individual cells from high-resolution Visium HD data^59,60. However, the results indicate that single-cell analysis based on this dataset is not suitable for our samples (Supplementary Fig. 10).

Intestinal organoid extraction and culture

C57BL/6J mice, aged 6–8 weeks, were euthanized by CO₂ asphyxiation. The intestines were carefully removed using forceps, and the mesentery, blood vessels, and adipose tissue were dissected away. Intestinal segments were placed in a 10 cm dish on ice. The segments were longitudinally opened with small scissors and rinsed multiple times with ice-cold PBS to remove intestinal contents. The intestinal segments were placed mucosal-side up on a glass slide, and the mucus and villi were scraped off until the tissue became semi-transparent. The segments were then rinsed again with ice-cold PBS, and this process was repeated until the supernatant became clear. The supernatant was discarded, and the tissue was minced into small pieces and resuspended in 25 mL of 2.5 mM (2–5 mM) ice-cold EDTA-PBS. The mixture was incubated on a shaker at 50 rpm for 30 min. Afterward, the solution was centrifuged at low speed (1 min), and the EDTA-PBS was removed. The tissue fragments were then resuspended in 10 mL of cold PBS containing 10% FBS, gently pipetted up and down three times, and allowed to settle until the majority of the tissue fragments precipitated. The suspension was passed through a 70 µm cell strainer, and the filtrate was collected. The filtrate was centrifuged at 300 × g for 5 min at 4 °C, and the supernatant was discarded. The pellet was resuspended in cold complete medium (Abs) for crypt counting. The same initial number of crypts or cells was used across different conditions.

Crypts were mixed with a 1:1 ratio of 25 µL basic medium and 25 µL Matrigel, using pre-cooled tips to pipette and mix on ice. Pre-warmed 24-well plates were prepared, and 50 µL of the crypt-Matrigel mixture was added to the center of each well, avoiding bubble formation. After the addition, the plates were incubated at 37 °C for 15 min to allow the Matrigel to solidify into a dome shape. Once polymerized, 300 µL of complete medium was added to each well, and the plates were returned to the incubator. The medium was changed every 2 days using pre-warmed medium at room temperature.

Organoid and immune cell co-culture

A suitable number of organoids were collected and washed several times with ice-cold PBS to remove the majority of the Matrigel. The organoids were then mixed with immune cells at a ratio of 1:1000, based on the number of intestinal cells and immune cells. The mixture was prepared by combining serum-free medium for immune cells and complete medium for organoids in a 1:1 ratio. The organoid-immune cell mixture was added to a 96-well plate, with a final volume of 200 µL per well. The plate was cultured at 37 °C with 5% CO₂, and the status of the organoids and immune cells was continuously monitored daily.

Chemokine receptor profiling

Organoids were treated with afatinib (1 μM final concentration) for 24 h. At the endpoints, culture supernatants were collected as conditioned media. B cells were then cultured in either (1) a 1:1 mixture of organoid medium and B cell medium, or (2) a 1:1 mixture of conditioned medium and B cell medium. Expression of chemokine receptors on B cells was analyzed at the indicated time points using qPCR.

Quantitative real-time PCR

Total RNA was extracted from the indicated cells or rat intestinal tissue via RNAsimple Total RNA Kit (TIANGEN) with Buffer RZ (TIANGEN) according to the manufacturer’s instructions. One microgram of total RNA was reverse transcribed to cDNA synthesis using a ReverTra Ace qPCR RT Master Mix (FSQ-201, TOYOBO). The cDNA was analyzed by the Quantitative real-time PCR on ABI ViiA7 or ABI 7500 Real Time-PCR System using Hieff qPCR SYBR Green Master Mix (YEASEN) and the primers listed in the Supplementary table 2. Relative expression of the target genes was normalized to that of GAPDH with the 2-ΔΔCt method.

Flow cytometry analysis

Single-cell suspensions were prepared from the mouse jejunal lamina propria, Peyer’s patches, and mesenteric lymph nodes. The cells were resuspended in PBS. Viable and non-viable cells were separated using the Zombie NIR™ Fixable Viability Kit (77184, BioLegend) according to the manufacturer’s instructions. The cells were then incubated with fluorescent dyes for surface staining, thoroughly mixed, and prepared for analysis. Flow cytometry was performed using a Fortessa X20 (BD Biosciences), and data were analyzed using FlowJo software (TreeStar).

After isolation and labeling of jejunal tissue, the number of immune cells detectable by flow cytometry was relatively low. To more accurately assess changes in T and B cell populations, we therefore quantified their relative proportions—calculating CD4⁺, CD8⁺, and CD19⁺ cells as percentages of the CD45⁺ leukocyte pool.

Xenograft tumor model

PC9 cells (5 × 10⁶ cells per mouse) were subcutaneously injected into the flanks of nude mice. When tumor size reached approximately 200–400 mm³, mice were randomly assigned to three groups (n = 7/group). Group 1 was the control group, receiving daily oral gavage of a placebo for 10 consecutive days. Group 2 received daily oral gavage of 30 mg/kg afatinib with placebo, and Group 3 received daily oral gavage of 30 mg/kg afatinib and 30 mg/kg tofacitinib. Tumor volume (V = 0.52 × a × b², a and b denote the length and width, respectively) and body weight were measured every 2 days. After 10 days, all mice were euthanized by carbon dioxide asphyxiation, and tumor tissues were collected, photographed, and weighed.

For the LLC1 tumor model, LLC1 cells (1.0 × 10⁶) were subcutaneously injected into the dorsal region of C57BL/6 mice. When tumor size reached approximately 50–90 mm³, mice were randomly assigned to three groups (n = 5/group). Group 1 was the control group, receiving daily oral gavage of a placebo for 10 consecutive days. Group 2 received daily oral gavage of 60 mg/kg afatinib with placebo, and Group 3 received daily oral gavage of 60 mg/kg afatinib and 60 mg/kg tofacitinib. Tumor volume was measured every 2 days with calipers, and body weight was monitored daily. After 10 days of treatment, blood was collected through the orbital sinus for complete blood count analysis. All mice were then euthanized by carbon dioxide asphyxiation, and tumor tissues were collected, photographed, and weighed.

Quantification and statistical analysis

Kaplan-Meier analysis was used to determine the time to first onset of diarrhea and the median time to onset, with comparisons made using the log-rank test. Unless otherwise stated in the figure legends, all data are presented as mean ± SEM. Statistical analyses were performed using unpaired Student’s t test and one-way or two-way analysis of variance (ANOVA). Statistical analysis was conducted with GraphPad Prism 8.0 (GraphPad Software). Details of the statistical tests used and summary values for biological replicates are provided in the respective figure legends. For all experiments, the investigators were blinded to group allocation during data collection and analysis.

Patient and public involvement

Patients and/or the public were not involved in the design, conduct, reporting, or dissemination plans of this research.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Supplementary information

Supplementary Information^{(6MB, pdf)}

Reporting Summary^{(110.1KB, pdf)}

Transparent Peer Review File^{(2.5MB, pdf)}

Source data

Source Data^{(1.5MB, xlsx)}

Acknowledgements

This work was supported by the National Natural Science Foundation of China (grant no. T2550061), Shanghai Municipal Science and Technology Major Project, and the Young Leading Scientists Cultivation Plan supported by Shanghai Municipal Education Commission (ZXWH1082101). We are thankful to the Shanghai Cancer Institute for providing the technical platform. We would like to thank Shanghai Jiao Tong University Laboratory Animal Center for Animal Feeding Services. The authors would like to thank Mengqiu Hao from Shanghai OE Biotech for their assistance in Visium spatial transcriptome sequencing.

Author contributions

Conceptualization: Yuan Cheng, Qing You, and Shiyi Zhang. Methodology: Yuan Cheng, Qing You, Chenyue Xu, and Leying Chen. Resources: Shiyi Zhang. Investigation: Yuan Cheng, Qing You, Chenyue Xu, and Haoyu Wang. Visualization: Yuan Cheng, Qing You, and Chenyue Xu. Validation: Yuan Cheng, Qing You, Chenyue Xu, Meng Tian, Haoyu Wang, Dazhao Lv, and Bingxue Yang. Software: Yuan Cheng and Chenyue Xu. Funding acquisition: Shiyi Zhang. Project administration: Shiyi Zhang. Writing—original draft: Yuan Cheng. Writing—review and editing: Yuan Cheng, Qing You, Chenyue Xu, Meng Tian, Haoyu Wang, Bingxue Yang, and Shiyi Zhang.

Peer review

Peer review information

Nature Communications thanks Carlo Perricone and the other anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Data availability

The mass spectrometry proteomics data have been deposited to ProteomeXchange Consortium (https://www.iprox.cn//page/project.html?id=IPX0012903000) via the iProX partner repository with the dataset identifier PXD066986. Raw data have been deposited to the National Center for Biotechnology Information (NCBI) under the BioProject number PRJNA1427952, that are publicly accessible at https://www.ncbi.nlm.nih.gov/bioproject/PRJNA1427952. All data relevant to the study are included in the article or uploaded as supplementary information. Source data are provided with this paper.

Competing interests

The authors declare no competing interests.

Footnotes

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Contributor Information

Qing You, Email: youqing@sjtu.edu.cn.

Shiyi Zhang, Email: zhangshiyi@sjtu.edu.cn.

Supplementary information

The online version contains supplementary material available at 10.1038/s41467-026-71739-8.

References

1.Metcalfe, C., Kljavin, N. M., Ybarra, R. & de Sauvage, F. J. Lgr5+ stem cells are indispensable for radiation-induced intestinal regeneration. Cell Stem Cell14, 149–159 (2014). [DOI] [PubMed] [Google Scholar]
2.Choi, J. & Augenlicht, L. H. Intestinal stem cells: guardians of homeostasis in health and aging amid environmental challenges. Exp. Mol. Med.56, 495–500 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
3.Barker, N. et al. Identification of stem cells in small intestine and colon by marker gene Lgr5. Nature449, 1003–1007 (2007). [DOI] [PubMed] [Google Scholar]
4.Sato, T. et al. Single Lgr5 stem cells build crypt-villus structures in vitro without a mesenchymal niche. Nature459, 262–265 (2009). [DOI] [PubMed] [Google Scholar]
5.Barker, N. et al. Lgr5(+ve) stem cells drive self-renewal in the stomach and build long-lived gastric units in vitro. Cell Stem Cell6, 25–36 (2010). [DOI] [PubMed] [Google Scholar]
6.Lindemans, C. A. et al. Interleukin-22 promotes intestinal-stem-cell-mediated epithelial regeneration. Nature528, 560–564 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
7.Spencer, J. & Sollid, L. M. The human intestinal B-cell response. Mucosal Immunol.9, 1113–1124 (2016). [DOI] [PubMed] [Google Scholar]
8.Dikiy, S. & Rudensky, A. Y. Principles of regulatory T cell function. Immunity56, 240–255 (2023). [DOI] [PubMed] [Google Scholar]
9.Grivennikov, S. I., Greten, F. R. & Karin, M. Immunity, inflammation, and cancer. Cell140, 883–899 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
10.Takashima, S. et al. T cell-derived interferon-gamma programs stem cell death in immune-mediated intestinal damage. Sci. Immunol.4, 10.1126/sciimmunol.aay8556 (2019). [DOI] [PMC free article] [PubMed]
11.Farin, H. F. et al. Paneth cell extrusion and release of antimicrobial products is directly controlled by immune cell-derived IFN-gamma. J. Exp. Med.211, 1393–1405 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
12.Raetz, M. et al. Parasite-induced TH1 cells and intestinal dysbiosis cooperate in IFN-gamma-dependent elimination of Paneth cells. Nat. Immunol.14, 136–142 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
13.Gunther, C. et al. Caspase-8 regulates TNF-alpha-induced epithelial necroptosis and terminal ileitis. Nature477, 335–339 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
14.Nava, P. et al. Interferon-gamma regulates intestinal epithelial homeostasis through converging beta-catenin signaling pathways. Immunity32, 392–402 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
15.Lebrusant-Fernandez, M. et al. IFN-gamma-dependent regulation of intestinal epithelial homeostasis by NKT cells. Cell Rep.43, 114948 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
16.Frede, A. et al. B cell expansion hinders the stroma-epithelium regenerative cross talk during mucosal healing. Immunity55, 2336–2351 e2312 (2022). [DOI] [PubMed] [Google Scholar]
17.Dorn, I., Schlenke, P., Mascher, B., Stange, E. F. & Seyfarth, M. Lamina propria plasma cells in inflammatory bowel disease: intracellular detection of immunoglobulins using flow cytometry. Immunobiology206, 546–557 (2002). [DOI] [PubMed] [Google Scholar]
18.Ciardiello, F. & Tortora, G. EGFR antagonists in cancer treatment. N. Engl. J. Med.358, 1160–1174 (2008). [DOI] [PubMed] [Google Scholar]
19.Chong, C. R. & Janne, P. A. The quest to overcome resistance to EGFR-targeted therapies in cancer. Nat. Med.19, 1389–1400 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
20.Koch, A. T., Love-Homan, L., Espinosa-Cotton, M., Stanam, A. & Simons, A. L. MyD88-dependent signaling decreases the antitumor efficacy of epidermal growth factor receptor inhibition in head and neck cancer cells. Cancer Res.75, 1657–1667 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
21.Sung, C. Y. W. et al. Macrophage depletion protects against cisplatin-induced ototoxicity and nephrotoxicity. Sci. Adv.10, eadk9878 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
22.Cui, H. et al. The support of genetic evidence for cardiovascular risk induced by antineoplastic drugs. Sci. Adv.6, 10.1126/sciadv.abb8543 (2020). [DOI] [PMC free article] [PubMed]
23.Sequist, L. V. et al. Phase III study of afatinib or cisplatin plus pemetrexed in patients with metastatic lung adenocarcinoma with EGFR mutations. J. Clin. Oncol.31, 3327–3334 (2013). [DOI] [PubMed] [Google Scholar]
24.Califano, R. et al. Expert Consensus on the Management of Adverse Events from EGFR Tyrosine Kinase Inhibitors in the UK. Drugs75, 1335–1348 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]
25.Yang, J. C. et al. Diarrhea associated with afatinib: an oral ErbB family blocker. Expert Rev. Anticancer Ther.13, 729–736 (2013). [DOI] [PubMed] [Google Scholar]
26.Sun, K., Wang, X., Zhang, H., Lin, G. & Jiang, R. Management and mechanisms of diarrhea induced by tyrosine kinase inhibitors in human epidermal growth factor receptor-2-positive breast cancer. Cancer Control31, 10732748241278039 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
27.Lee, D. Y. & Kang, H. E. Pharmacokinetics of DA-6886, a new 5-HT(4) receptor agonist, in rats. Pharmaceutics14, 10.3390/pharmaceutics14040702 (2022). [DOI] [PMC free article] [PubMed]
28.Sommerfelt, H. et al. Cloned polynucleotide and synthetic oligonucleotide probes used in colony hybridization are equally efficient in the identification of Escherichia coli. J. Clin. Microbiol.28, 642 (1990). [DOI] [PMC free article] [PubMed] [Google Scholar]
29.Hu, X., Li, J., Fu, M., Zhao, X. & Wang, W. The JAK/STAT signaling pathway: from bench to clinic. Signal Transduct. Target. Ther.6, 402 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
30.Ponomarenko, E. A. et al. Workability of mRNA sequencing for predicting protein abundance. Genes14, 10.3390/genes14112065 (2023). [DOI] [PMC free article] [PubMed]
31.Mayassi, T. et al. Spatially restricted immune and microbiota-driven adaptation of the gut. Nature636, 447–456 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
32.Cable, D. M. et al. Robust decomposition of cell type mixtures in spatial transcriptomics. Nat. Biotechnol.40, 517–526 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
33.Anderson, A. C. et al. Promotion of tissue inflammation by the immune receptor Tim-3 expressed on innate immune cells. Science318, 1141–1143 (2007). [DOI] [PubMed] [Google Scholar]
34.Ye, C. et al. Single-cell and spatial transcriptomics reveal the fibrosis-related immune landscape of biliary atresia. Clin. Transl. Med.12, e1070 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
35.Fruhbeck, G. et al. High plasma and lingual uroguanylin as potential contributors to changes in food preference after sleeve gastrectomy. Metabolism128, 155119 (2022). [DOI] [PubMed] [Google Scholar]
36.Yin, L. et al. Single-cell analysis of cellular heterogeneity and interactions in the ischemia-reperfusion injured mouse intestine. J. Pharm. Anal.13, 760–775 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
37.Iozzo, R. V., Zoeller, J. J. & Nystrom, A. Basement membrane proteoglycans: modulators Par Excellence of cancer growth and angiogenesis. Mol. Cells27, 503–513 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]
38.Berx, G. & van Roy, F. Involvement of members of the cadherin superfamily in cancer. Cold Spring Harb. Perspect. Biol.1, a003129 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]
39.Montorsi, L. et al. Double-negative B cells and DNASE1L3 colocalise with microbiota in gut-associated lymphoid tissue. Nat. Commun.15, 4051 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
40.Kunisawa, J., Kurashima, Y. & Kiyono, H. Gut-associated lymphoid tissues for the development of oral vaccines. Adv. Drug Deliv. Rev.64, 523–530 (2012). [DOI] [PubMed] [Google Scholar]
41.Rothkotter, H. J., Pabst, R. & Bailey, M. Lymphocyte migration in the intestinal mucosa: entry, transit and emigration of lymphoid cells and the influence of antigen. Vet. Immunol. Immunopathol.72, 157–165 (1999). [DOI] [PubMed] [Google Scholar]
42.Canales-Herrerias, P. et al. Gut-associated lymphoid tissue attrition associates with response to anti-alpha4beta7 therapy in ulcerative colitis. Sci. Immunol.9, eadg7549 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
43.Neurath, M. F. Targeting immune cell circuits and trafficking in inflammatory bowel disease. Nat. Immunol.20, 970–979 (2019). [DOI] [PubMed] [Google Scholar]
44.Graham, D. B. & Xavier, R. J. Pathway paradigms revealed from the genetics of inflammatory bowel disease. Nature578, 527–539 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
45.Li, Q. et al. Interferon-gamma and tumor necrosis factor-alpha disrupt epithelial barrier function by altering lipid composition in membrane microdomains of tight junction. Clin. Immunol.126, 67–80 (2008). [DOI] [PubMed] [Google Scholar]
46.He, Z. hen et al. Chemotherapy-induced microbiota exacerbates the toxicity of chemotherapy through the suppression of interleukin-10 from macrophages. Gut Microbes16, 2319511 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]
47.Benson, A. B. et al. Recommended guidelines for the treatment of cancer treatment-induced diarrhea. J. Clin. Oncol.22, 2918–2926 (2004). [DOI] [PubMed] [Google Scholar]
48.Andreyev, J. et al. Guidance on the management of diarrhoea during cancer chemotherapy. Lancet Oncol.15, e447–e460 (2014). [DOI] [PubMed] [Google Scholar]
49.Shi, X. J. et al. Discovery of tofacitinib derivatives as orally active antitumor agents based on the scaffold hybridization strategy. Eur. J. Med. Chem.203, 112601 (2020). [DOI] [PubMed] [Google Scholar]
50.Okita, R. et al. Effect of platinum‑based chemotherapy on the expression of natural killer group 2 member D ligands, programmed cell death‑1 ligand 1 and HLA class I in non‑small cell lung cancer. Oncol. Rep.42, 839–848 (2019). [DOI] [PubMed] [Google Scholar]
51.Schwartz, D. M., Bonelli, M., Gadina, M. & O’Shea, J. J. Type I/II cytokines, JAKs, and new strategies for treating autoimmune diseases. Nat. Rev. Rheumatol.12, 25–36 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]
52.Simon, N. et al. Characterisation of the antinociceptive effect of baricitinib in the collagen antibody-induced arthritis mouse model. Ann. Rheum. Dis.84, 421–434 (2025). [DOI] [PubMed] [Google Scholar]
53.Basak, O. et al. Induced quiescence of Lgr5+ stem cells in intestinal organoids enables differentiation of hormone-producing enteroendocrine cells. Cell Stem Cell20, 177–190.e174 (2017). [DOI] [PubMed] [Google Scholar]
54.Capaldo, C. T. et al. IFN-gamma and TNF-alpha-induced GBP-1 inhibits epithelial cell proliferation through suppression of beta-catenin/TCF signaling. Mucosal Immunol.5, 681–690 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
55.Zhao, X. et al. TNF‑alpha promotes the malignant transformation of intestinal stem cells through the NF‑kappaB and Wnt/beta‑catenin signaling pathways. Oncol. Rep.44, 577–588 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
56.Curtis, J. R. et al. Malignancy risk with tofacitinib versus TNF inhibitors in rheumatoid arthritis: results from the open-label, randomised controlled ORAL Surveillance trial. Ann. Rheum. Dis.82, 331–343 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
57.Russell, M. D. et al. JAK inhibitors and the risk of malignancy: a meta-analysis across disease indications. Ann. Rheum. Dis.82, 1059–1067 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]
58.Riehl, T. E., Alvarado, D., Ee, X., Ciorba, M. A. & Stenson, W. F. Hyaluronic acid promotes Lgr5(+) stem cell proliferation and crypt fission through TLR4 and PGE(2) transactivation of EGFR. Am. J. Physiol. Gastrointest. Liver Physiol.319, G63–G73 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
59.Polanski, K. et al. Bin2cell reconstructs cells from high resolution Visium HD data. Bioinformatics40, 10.1093/bioinformatics/btae546 (2024). [DOI] [PMC free article] [PubMed]
60.Chen, T. et al. The genome sequence archive family: toward explosive data growth and diverse data types. Genom. Proteom. Bioinform.19, 578–583 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
61.Luo, Y., Luo, L., Xia, M., Liu, Q. & Zhang, G. Studies on the changes in rectal permeability and intestinal microbiota with developmental age in young rats. Front. Microbiol.16, 1551693 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Information^{(6MB, pdf)}

Reporting Summary^{(110.1KB, pdf)}

Transparent Peer Review File^{(2.5MB, pdf)}

Source Data^{(1.5MB, xlsx)}

Data Availability Statement

[CR1] 1.Metcalfe, C., Kljavin, N. M., Ybarra, R. & de Sauvage, F. J. Lgr5+ stem cells are indispensable for radiation-induced intestinal regeneration. Cell Stem Cell14, 149–159 (2014). [DOI] [PubMed] [Google Scholar]

[CR2] 2.Choi, J. & Augenlicht, L. H. Intestinal stem cells: guardians of homeostasis in health and aging amid environmental challenges. Exp. Mol. Med.56, 495–500 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR3] 3.Barker, N. et al. Identification of stem cells in small intestine and colon by marker gene Lgr5. Nature449, 1003–1007 (2007). [DOI] [PubMed] [Google Scholar]

[CR4] 4.Sato, T. et al. Single Lgr5 stem cells build crypt-villus structures in vitro without a mesenchymal niche. Nature459, 262–265 (2009). [DOI] [PubMed] [Google Scholar]

[CR5] 5.Barker, N. et al. Lgr5(+ve) stem cells drive self-renewal in the stomach and build long-lived gastric units in vitro. Cell Stem Cell6, 25–36 (2010). [DOI] [PubMed] [Google Scholar]

[CR6] 6.Lindemans, C. A. et al. Interleukin-22 promotes intestinal-stem-cell-mediated epithelial regeneration. Nature528, 560–564 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR7] 7.Spencer, J. & Sollid, L. M. The human intestinal B-cell response. Mucosal Immunol.9, 1113–1124 (2016). [DOI] [PubMed] [Google Scholar]

[CR8] 8.Dikiy, S. & Rudensky, A. Y. Principles of regulatory T cell function. Immunity56, 240–255 (2023). [DOI] [PubMed] [Google Scholar]

[CR9] 9.Grivennikov, S. I., Greten, F. R. & Karin, M. Immunity, inflammation, and cancer. Cell140, 883–899 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR10] 10.Takashima, S. et al. T cell-derived interferon-gamma programs stem cell death in immune-mediated intestinal damage. Sci. Immunol.4, 10.1126/sciimmunol.aay8556 (2019). [DOI] [PMC free article] [PubMed]

[CR11] 11.Farin, H. F. et al. Paneth cell extrusion and release of antimicrobial products is directly controlled by immune cell-derived IFN-gamma. J. Exp. Med.211, 1393–1405 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR12] 12.Raetz, M. et al. Parasite-induced TH1 cells and intestinal dysbiosis cooperate in IFN-gamma-dependent elimination of Paneth cells. Nat. Immunol.14, 136–142 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR13] 13.Gunther, C. et al. Caspase-8 regulates TNF-alpha-induced epithelial necroptosis and terminal ileitis. Nature477, 335–339 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR14] 14.Nava, P. et al. Interferon-gamma regulates intestinal epithelial homeostasis through converging beta-catenin signaling pathways. Immunity32, 392–402 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR15] 15.Lebrusant-Fernandez, M. et al. IFN-gamma-dependent regulation of intestinal epithelial homeostasis by NKT cells. Cell Rep.43, 114948 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR16] 16.Frede, A. et al. B cell expansion hinders the stroma-epithelium regenerative cross talk during mucosal healing. Immunity55, 2336–2351 e2312 (2022). [DOI] [PubMed] [Google Scholar]

[CR17] 17.Dorn, I., Schlenke, P., Mascher, B., Stange, E. F. & Seyfarth, M. Lamina propria plasma cells in inflammatory bowel disease: intracellular detection of immunoglobulins using flow cytometry. Immunobiology206, 546–557 (2002). [DOI] [PubMed] [Google Scholar]

[CR18] 18.Ciardiello, F. & Tortora, G. EGFR antagonists in cancer treatment. N. Engl. J. Med.358, 1160–1174 (2008). [DOI] [PubMed] [Google Scholar]

[CR19] 19.Chong, C. R. & Janne, P. A. The quest to overcome resistance to EGFR-targeted therapies in cancer. Nat. Med.19, 1389–1400 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR20] 20.Koch, A. T., Love-Homan, L., Espinosa-Cotton, M., Stanam, A. & Simons, A. L. MyD88-dependent signaling decreases the antitumor efficacy of epidermal growth factor receptor inhibition in head and neck cancer cells. Cancer Res.75, 1657–1667 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR21] 21.Sung, C. Y. W. et al. Macrophage depletion protects against cisplatin-induced ototoxicity and nephrotoxicity. Sci. Adv.10, eadk9878 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR22] 22.Cui, H. et al. The support of genetic evidence for cardiovascular risk induced by antineoplastic drugs. Sci. Adv.6, 10.1126/sciadv.abb8543 (2020). [DOI] [PMC free article] [PubMed]

[CR23] 23.Sequist, L. V. et al. Phase III study of afatinib or cisplatin plus pemetrexed in patients with metastatic lung adenocarcinoma with EGFR mutations. J. Clin. Oncol.31, 3327–3334 (2013). [DOI] [PubMed] [Google Scholar]

[CR24] 24.Califano, R. et al. Expert Consensus on the Management of Adverse Events from EGFR Tyrosine Kinase Inhibitors in the UK. Drugs75, 1335–1348 (2015). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR25] 25.Yang, J. C. et al. Diarrhea associated with afatinib: an oral ErbB family blocker. Expert Rev. Anticancer Ther.13, 729–736 (2013). [DOI] [PubMed] [Google Scholar]

[CR26] 26.Sun, K., Wang, X., Zhang, H., Lin, G. & Jiang, R. Management and mechanisms of diarrhea induced by tyrosine kinase inhibitors in human epidermal growth factor receptor-2-positive breast cancer. Cancer Control31, 10732748241278039 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR27] 27.Lee, D. Y. & Kang, H. E. Pharmacokinetics of DA-6886, a new 5-HT(4) receptor agonist, in rats. Pharmaceutics14, 10.3390/pharmaceutics14040702 (2022). [DOI] [PMC free article] [PubMed]

[CR28] 28.Sommerfelt, H. et al. Cloned polynucleotide and synthetic oligonucleotide probes used in colony hybridization are equally efficient in the identification of Escherichia coli. J. Clin. Microbiol.28, 642 (1990). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR29] 29.Hu, X., Li, J., Fu, M., Zhao, X. & Wang, W. The JAK/STAT signaling pathway: from bench to clinic. Signal Transduct. Target. Ther.6, 402 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR30] 30.Ponomarenko, E. A. et al. Workability of mRNA sequencing for predicting protein abundance. Genes14, 10.3390/genes14112065 (2023). [DOI] [PMC free article] [PubMed]

[CR31] 31.Mayassi, T. et al. Spatially restricted immune and microbiota-driven adaptation of the gut. Nature636, 447–456 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR32] 32.Cable, D. M. et al. Robust decomposition of cell type mixtures in spatial transcriptomics. Nat. Biotechnol.40, 517–526 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR33] 33.Anderson, A. C. et al. Promotion of tissue inflammation by the immune receptor Tim-3 expressed on innate immune cells. Science318, 1141–1143 (2007). [DOI] [PubMed] [Google Scholar]

[CR34] 34.Ye, C. et al. Single-cell and spatial transcriptomics reveal the fibrosis-related immune landscape of biliary atresia. Clin. Transl. Med.12, e1070 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR35] 35.Fruhbeck, G. et al. High plasma and lingual uroguanylin as potential contributors to changes in food preference after sleeve gastrectomy. Metabolism128, 155119 (2022). [DOI] [PubMed] [Google Scholar]

[CR36] 36.Yin, L. et al. Single-cell analysis of cellular heterogeneity and interactions in the ischemia-reperfusion injured mouse intestine. J. Pharm. Anal.13, 760–775 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR37] 37.Iozzo, R. V., Zoeller, J. J. & Nystrom, A. Basement membrane proteoglycans: modulators Par Excellence of cancer growth and angiogenesis. Mol. Cells27, 503–513 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR38] 38.Berx, G. & van Roy, F. Involvement of members of the cadherin superfamily in cancer. Cold Spring Harb. Perspect. Biol.1, a003129 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR39] 39.Montorsi, L. et al. Double-negative B cells and DNASE1L3 colocalise with microbiota in gut-associated lymphoid tissue. Nat. Commun.15, 4051 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR40] 40.Kunisawa, J., Kurashima, Y. & Kiyono, H. Gut-associated lymphoid tissues for the development of oral vaccines. Adv. Drug Deliv. Rev.64, 523–530 (2012). [DOI] [PubMed] [Google Scholar]

[CR41] 41.Rothkotter, H. J., Pabst, R. & Bailey, M. Lymphocyte migration in the intestinal mucosa: entry, transit and emigration of lymphoid cells and the influence of antigen. Vet. Immunol. Immunopathol.72, 157–165 (1999). [DOI] [PubMed] [Google Scholar]

[CR42] 42.Canales-Herrerias, P. et al. Gut-associated lymphoid tissue attrition associates with response to anti-alpha4beta7 therapy in ulcerative colitis. Sci. Immunol.9, eadg7549 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR43] 43.Neurath, M. F. Targeting immune cell circuits and trafficking in inflammatory bowel disease. Nat. Immunol.20, 970–979 (2019). [DOI] [PubMed] [Google Scholar]

[CR44] 44.Graham, D. B. & Xavier, R. J. Pathway paradigms revealed from the genetics of inflammatory bowel disease. Nature578, 527–539 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR45] 45.Li, Q. et al. Interferon-gamma and tumor necrosis factor-alpha disrupt epithelial barrier function by altering lipid composition in membrane microdomains of tight junction. Clin. Immunol.126, 67–80 (2008). [DOI] [PubMed] [Google Scholar]

[CR46] 46.He, Z. hen et al. Chemotherapy-induced microbiota exacerbates the toxicity of chemotherapy through the suppression of interleukin-10 from macrophages. Gut Microbes16, 2319511 (2024). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR47] 47.Benson, A. B. et al. Recommended guidelines for the treatment of cancer treatment-induced diarrhea. J. Clin. Oncol.22, 2918–2926 (2004). [DOI] [PubMed] [Google Scholar]

[CR48] 48.Andreyev, J. et al. Guidance on the management of diarrhoea during cancer chemotherapy. Lancet Oncol.15, e447–e460 (2014). [DOI] [PubMed] [Google Scholar]

[CR49] 49.Shi, X. J. et al. Discovery of tofacitinib derivatives as orally active antitumor agents based on the scaffold hybridization strategy. Eur. J. Med. Chem.203, 112601 (2020). [DOI] [PubMed] [Google Scholar]

[CR50] 50.Okita, R. et al. Effect of platinum‑based chemotherapy on the expression of natural killer group 2 member D ligands, programmed cell death‑1 ligand 1 and HLA class I in non‑small cell lung cancer. Oncol. Rep.42, 839–848 (2019). [DOI] [PubMed] [Google Scholar]

[CR51] 51.Schwartz, D. M., Bonelli, M., Gadina, M. & O’Shea, J. J. Type I/II cytokines, JAKs, and new strategies for treating autoimmune diseases. Nat. Rev. Rheumatol.12, 25–36 (2016). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR52] 52.Simon, N. et al. Characterisation of the antinociceptive effect of baricitinib in the collagen antibody-induced arthritis mouse model. Ann. Rheum. Dis.84, 421–434 (2025). [DOI] [PubMed] [Google Scholar]

[CR53] 53.Basak, O. et al. Induced quiescence of Lgr5+ stem cells in intestinal organoids enables differentiation of hormone-producing enteroendocrine cells. Cell Stem Cell20, 177–190.e174 (2017). [DOI] [PubMed] [Google Scholar]

[CR54] 54.Capaldo, C. T. et al. IFN-gamma and TNF-alpha-induced GBP-1 inhibits epithelial cell proliferation through suppression of beta-catenin/TCF signaling. Mucosal Immunol.5, 681–690 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR55] 55.Zhao, X. et al. TNF‑alpha promotes the malignant transformation of intestinal stem cells through the NF‑kappaB and Wnt/beta‑catenin signaling pathways. Oncol. Rep.44, 577–588 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR56] 56.Curtis, J. R. et al. Malignancy risk with tofacitinib versus TNF inhibitors in rheumatoid arthritis: results from the open-label, randomised controlled ORAL Surveillance trial. Ann. Rheum. Dis.82, 331–343 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR57] 57.Russell, M. D. et al. JAK inhibitors and the risk of malignancy: a meta-analysis across disease indications. Ann. Rheum. Dis.82, 1059–1067 (2023). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR58] 58.Riehl, T. E., Alvarado, D., Ee, X., Ciorba, M. A. & Stenson, W. F. Hyaluronic acid promotes Lgr5(+) stem cell proliferation and crypt fission through TLR4 and PGE(2) transactivation of EGFR. Am. J. Physiol. Gastrointest. Liver Physiol.319, G63–G73 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR59] 59.Polanski, K. et al. Bin2cell reconstructs cells from high resolution Visium HD data. Bioinformatics40, 10.1093/bioinformatics/btae546 (2024). [DOI] [PMC free article] [PubMed]

[CR60] 60.Chen, T. et al. The genome sequence archive family: toward explosive data growth and diverse data types. Genom. Proteom. Bioinform.19, 578–583 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]

[CR61] 61.Luo, Y., Luo, L., Xia, M., Liu, Q. & Zhang, G. Studies on the changes in rectal permeability and intestinal microbiota with developmental age in young rats. Front. Microbiol.16, 1551693 (2025). [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

JAK inhibitors alleviate EGFR-inhibitor-induced diarrhea by protecting intestinal stem cells from adaptive-immune-exacerbated injury

Yuan Cheng

Chenyue Xu

Dazhao Lv

Meng Tian

Haoyu Wang

Bingxue Yang

Leying Chen

Qing You

Shiyi Zhang

Abstract

Introduction

Results

Impaired jejunum contributes to EGFRi-induced diarrhea

Fig. 1. Establishment of a rat model of diarrhea induced by EGFRi.

Fig. 2. Establishment of an EGFRi-induced diarrhea model in mice and small intestinal organoids.

Multi-omics reveals the critical role of mucosal immunity at different stages of diarrhea

Fig. 3. Multi-omics reveals the critical role of mucosal immunity in diarrhea across different stages.

T and B cells exacerbate EGFRi-Induced intestinal injury

Fig. 4. T and B cells further inhibit the recovery of EGFRi-suppressed intestinal stem cell proliferation.

Investigating the crosstalk between T, B cell and stem cells using spatial transcriptomics

Fig. 5. Spatial transcriptomics unveils intercellular communication networks.

JAK inhibitors attenuate intestinal inflammatory signaling

Fig. 6. Tofacitinib inhibits the migration of T and B cells and the secretion of inflammatory cytokines.

Oral administration of JAK inhibitors ameliorates diarrhea in a rat model

Fig. 7. Oral JAK Inhibitor Treatment Improves Diarrhea.

Fig. 8. Comparison of in vivo therapeutic effects between oral JAK inhibitors and conventional drugs.

Oral administration of tofacitinib did not affect the anti-tumor effect of afatinib

Fig. 9. Oral JAK inhibitor treatment exhibits a broad safety margin.

Discussion

Fig. 10. Proposed mechanism of EGFRi-induced diarrhea and the potential therapeutic action of JAK inhibition.

Methods

Ethics approval

Animal

Cell lines and culture conditions

Reagents

Diarrhea grade evaluation for rats

PAS staining

Assessment of Intestinal Permeability

Immunofluorescent staining

RNA isolation and library preparation

Quality control of RNA-Seq raw data and differential expression gene analysis

Total protein extraction

Protein quality test

TMT labeling of peptides

LC-MS analysis

Differential expression proteins analysis

Score of infiltration immune cells

Visium HD spatial transcriptome sequencing

Intestinal organoid extraction and culture

Organoid and immune cell co-culture

Chemokine receptor profiling

Quantitative real-time PCR

Flow cytometry analysis

Xenograft tumor model

Quantification and statistical analysis

Patient and public involvement

Reporting summary

Supplementary information

Source data

Acknowledgements

Author contributions

Peer review

Peer review information

Data availability

Competing interests

Footnotes

Contributor Information

Supplementary information

References

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases