Dose-response framework for DSS-induced colitis in mice Authors: Siyan Che, Jinglong Peng, Xia Xu, Yue Wang, Mujie Huang, Lin Zhao, Bin Feng, Pan Huang, Jianjun Li, Miaomiao Wu, Yulong Yin Published: 2026 ABSTRACT SummaryDextran sulfate sodium (DSS) is widely used to model epithelial injury-driven colitis, but heterogeneous dosing complicates model selection and cross-study comparison. Here, we established a graded DSS dose-response framework by administering 0%, 1.5%, 2.0%, 2.5%, or 3.0% DSS to C57BL/6J mice for 5 days, followed by 2 days of recovery, and integrated pathological, microbial, transcriptomic, and immune profiling. Low-dose DSS (1.5%–2.0%) induced mild and largely reversible epithelial injury, early microbiota perturbation, and CD8+ T cell accumulation. Higher doses (2.5%–3.0%) caused sustained inflammation, pronounced mucosal damage, macrophage and granulocyte infiltration, and elevated pro-inflammatory cytokines. Transcriptomic analysis revealed shared activation of TNF, NF-κB, IL-17, and NOD-like receptor signaling, with greater amplitude at higher DSS concentrations. This framework supports rational DSS dose selection according to disease severity and experimental objectives.Keywords: dextran sulfate sodium, dose-dependent colitis, microbiota dysbiosis, mucosal immune activation Introduction Inflammatory bowel disease (IBD), encompassing Crohn’s disease and ulcerative colitis, is a global disorder characterized by a chronically dysregulated immune response in the gastrointestinal tract.1,2 Despite substantial advances in disease management, the global prevalence of IBD continues to increase, with more than 3.5 million affected individuals in the United States and Europe alone, imposing a growing burden on healthcare systems.3 The pathogenesis of IBD is multifactorial, involving complex interactions among host genetics, environmental exposures, intestinal microbiota, and immune regulation, which together contribute to disease initiation and progression.4 Experimental animal models have been instrumental in dissecting the interacting pathogenic components of IBD and in evaluating therapeutic strategies.5,6,7 Chemical models, particularly those based on dextran sulfate sodium (DSS), are among the most widely used owing to their technical simplicity, reproducibility, and capacity to induce epithelial injury-driven inflammation that resembles key pathological features of human ulcerative colitis.5 The DSS-induced colitis model is highly versatile, as the modulation of DSS concentration and exposure duration enables the induction of acute, chronic, or relapsing inflammatory states. Conventionally, administration of 40–50 kDa DSS in drinking water is used to generate robust colitis, with disease severity broadly correlated with DSS dose.8 However, effective DSS concentrations reported in the literature vary widely, typically ranging from 1% to 3%, and in some cases extending to 5% (w/v).9,10 This variability complicates cross-study comparisons and obscures the relationship between DSS dose and specific pathological, microbial, transcriptional, and immune outcomes.5,8 Importantly, while DSS dose is widely recognized as a determinant of colitis severity, a systematic, multi-layered characterization of dose-dependent responses across epithelial injury, gut microbiota, host transcriptional programs, and immune activation remains limited. As a result, DSS concentrations are often selected empirically, without clear alignment to specific experimental objectives, such as modeling early immune perturbation versus severe ulcerative colitis-like pathology. In this study, we address this gap by establishing a graded DSS dose-response framework in C57BL/6J mice, using 1.5%, 2.0%, 2.5%, and 3.0% DSS. By integrating phenotypic assessment, histopathology, gut microbiota profiling, transcriptomic analysis, and immune characterization, we systematically delineate how increasing DSS doses reshape intestinal pathology and host responses. This work provides a comprehensive reference for rational DSS dose selection and offers a systems-level view of how epithelial, microbial, and immune compartments are coordinately altered during DSS-induced colitis. Results Differential colitis phenotypes induced by increasing DSS concentrations To investigate the dose-dependent effects of DSS on colitis phenotypes in mice, the animals were administered different concentrations of DSS (0%, 1.5%, 2.0%, 2.5%, and 3.0%, w/v) through their drinking water for 5 days, followed by 2 days of recovery with regular water (Figure 1A). DSS treatment resulted in a significant decrease in body weight compared with the control group, with weight loss being more pronounced in the 2.5% and 3.0% groups than in the 1.5% and 2.0% groups. However, there was no significant difference between the 2.5% and 3.0% groups (Figure 1B). After the 2-day recovery period, the mice treated with 1.5% and 2.0% DSS exhibited disease activity index (DAI) scores close to those of healthy controls, whereas DAI values in the 2.5% and 3.0% groups remained significantly elevated, indicating that complete recovery may require a longer period (Figure 1C). DSS exposure also induced a dose-dependent shortening of the colon, with the 3.0% DSS group showing significantly reduced colon length compared with the 1.5% group but a comparable length to those of the 2% and 2.5% groups (Figures 1D and 1E). Consistent with these findings, spleen and liver weights also exhibited a dose-dependent trend (Figures 1F and 1G). Histopathological analysis further confirmed that colonic epithelial injury and inflammatory cell infiltration were more severe in the 2.5% and 3.0% groups than in the 1.5% and 2.0% groups (Figures 1H and 1I). Collectively, these results demonstrate that 2.5% and 3.0% DSS induce comparable degrees of colitis in mice, both reflecting severe disease phenotypes. The impact of varying concentrations of DSS on the phenotype of mice with colitis (A) Experimental design for DSS-induced colitis in C57BL/6J mice: the mice received DSS for 5 days, followed by regular water for 2 days. (B and C) Mice body weight (B) and disease activity index scores (C) were recorded daily throughout the experiment period. (D and E) Colon length; scale bars, 1 cm. (F and G) Spleen weight (F) and liver weight (G) were recorded at the experimental endpoint. (H) Histopathological changes in the colonic tissues from the experiment group were observed by H&E staining; scale bars, 50 μm. (I) Histology score. All data are presented as the mean ± SEM (n = 10–12 per group). Statistical significance was determined using one-way ANOVA, followed by the Tukey’s multiple-comparisons test; ∗p < 0.05, ∗∗p < 0.01. Impact of DSS on the composition of the gut microbiota To determine how DSS exposure reshapes the gut microbiota, we performed 16S rRNA gene sequencing of fecal samples. Principal coordinate analysis revealed a clear separation of microbial communities in the mice treated with ≥2.0% DSS from those in control mice, whereas the 1.5% DSS group largely overlapped with controls, indicating minimal perturbation at low dose (Figures S1A–S1C). At the taxonomic level, DSS induced consistent compositional shifts across doses, characterized by enrichment of Erysipelotrichales, Bacteroidaceae, Verrucomicrobiota, Akkermansia, Muribaculaceae, and Bacteroidia (Figures S1D–S1G). At the phylum level, DSS directed the enrichment of Bacteroidota and Verrucomicrobiota and inhibited Firmicutes, Proteobacteria, Actinobacteriota, and Cyanobacteria (Figures 2A and 2B). At the genus level, DSS markedly increased the enrichment of Akkermansia and Turicibacter and decreased Ruminococcus_torques group (Figures 2C and 2D). Then, we observed the unique microbiota in the colon of each group of mice, and 1.5% DSS-treated mice had no unique microbiota when compared with the other groups. In comparison, the 2.0%, 2.5%, and 3.0% DSS groups were mainly enriched in an unclassified bacterium within the order Clostridiales, Verrucomicrobiota, and Erysipelotrichaceae, respectively (Figure 2E). Firmicutes and Proteobacteria dominated the gut of all mice. Nevertheless, DSS treatment increased the abundance of potentially harmful commensals, including Helicobacter, which was enriched in the 2.0% and 2.5% DSS groups, and Clostridium sensu stricto 1, which was enriched in the 3.0% DSS group (Figure 2F). Collectively, 1.5% DSS induced a milder colitis phenotype in mice, probably because the slight changes in the gut microbiota were not sufficient to trigger gut inflammation, and the 2%, 2.5%, and 3% DSS-induced gut microbiota disorder were consistent at the levels of phylum and genus. Different concentrations of DSS-induced intestinal microbial disturbance in mice (A and C) Bacterial taxonomic profiling of the gut microbiota at the phylum level (A) and genus level (C). (B and D) Microbiota compositions at the phylum level (B) and the genus level (D). (E) Differential microbial taxa among the 0.0%, 1.5%, 2.0%, 2.5%, and 3.0% DSS groups were identified using linear discriminant analysis effect size (LEfSe) analysis with the Kruskal–Wallis test (p < 0.05) and a linear discriminant analysis (LDA) score threshold of 4.0. (F) Cluster heatmap analysis at the genus level. All data are presented as mean ± SEM (n = 8–10 per group). Statistical significance was determined using one-way ANOVA, followed by Tukey’s multiple-comparisons test. ∗p < 0.05, ∗∗p < 0.01. Dose-dependent transcriptomic remodeling reveals inflammatory activation and metabolic suppression in DSS-induced colitis To investigate the effects of DSS on the transcriptional landscape of colonic tissues, we performed bulk RNA sequencing. Principal component analysis (PCA) revealed that the core transcriptional profiles of mice treated with different DSS concentrations were largely similar, despite dose-dependent variation in disease severity (Figure 3A). Using time-course clustering analysis, we identified two major gene clusters that responded strongly to DSS exposure: cluster 1, comprising upregulated genes, and cluster 2, consisting of downregulated genes (Figure 3B). Dose-dependent transcriptomic alterations in colonic tissues of DSS-induced colitis (A) Principal component analysis (PCA) of bulk RNA sequencing data from colonic tissues of mice treated with 0.0%, 1.5%, 2.0%, 2.5%, and 3.0% DSS. Samples clustered closely across the treatment groups, indicating overall similar core transcriptional components despite differences in disease severity. (B) Time-course clustering analysis identified two major gene clusters responsive to increasing DSS concentrations. Cluster 1 contains genes showing progressive upregulation across DSS doses, whereas cluster 2 contains genes displaying a dose-dependent downregulation pattern. (C) Venn diagrams showing overlap between the upregulated and downregulated DEGs across different DSS concentrations and the two major clusters. A total of 109 shared upregulated genes and 163 shared downregulated genes were consistently altered across all DSS-treated groups. (D) KEGG pathway enrichment analysis of shared DEGs. Pathway significance is represented by dot size (gene count) and color (−log10p value). We next intersected the genes upregulated or downregulated at each DSS concentration with the genes in cluster 1 and cluster 2, respectively. This analysis identified 109 commonly upregulated genes and 163 commonly downregulated genes across all DSS-treated groups (Figure 3C). Functional enrichment analysis showed that the upregulated genes were predominantly associated with the inflammatory and immune-related pathways, including cytokine-cytokine receptor interaction, IL-17 signaling pathway, TNF signaling pathway, and NOD-like receptor signaling pathway. In contrast, the downregulated genes were mainly enriched in metabolic pathways related to xenobiotic processing, such as drug metabolism-cytochrome P450, chemical carcinogenesis-DNA adducts, metabolism of xenobiotics by cytochrome P450, and drug metabolism (Figure 3D). Escalating DSS doses drive coordinated activation of IBD signature genes and pro-inflammatory pathways Focusing on genes implicated in IBD susceptibility and pathogenesis, transcriptomic analysis revealed a graded, dose-dependent remodeling of colonic gene expression programs in response to DSS exposure. A heatmap of genome-wide association study (GWAS)-derived IBD-associated risk genes11 demonstrated clear stratification across DSS concentrations (Figure 4A). While the control and 1.5% DSS-treated mice exhibited largely preserved transcriptional profiles, the 2.5% and 3.0% DSS groups showed pronounced upregulation of genes linked to epithelial barrier disruption (Muc2 and Claudin-2), cytokine signaling (Tnf, Il1b, and Il6), chemokine-mediated leukocyte recruitment (Ccl2, Cxcl1, and Ccl20), and oxidative stress responses. The transcriptional changes in the 2.0% DSS group were intermediate, indicating a transitional stage between mild and severe inflammation. Pathway-level analysis further confirmed a graded activation pattern of inflammation-associated signaling networks (Figure 4B). The expression of genes belonging to the TNF, IL-17, NF-κB, and chemokine signaling pathways increased progressively with the increasing DSS concentrations. Low-dose DSS (1.5%) induced minimal activation, whereas high-dose DSS (2.5%–3.0%) triggered extensive transcriptional remodeling consistent with mucosal immune hyperactivation. These data indicate that the DSS dose governs not only the magnitude of colonic inflammation but also the qualitative engagement of IBD-relevant transcriptional programs, delineating a continuum from early, subtle immune perturbation to overt inflammatory remodeling. Transcriptomic profiling reveals dose-dependent activation of IBD-related genes and inflammatory pathways in DSS-induced colitis (A) Heatmap showing the expression patterns of IBD-associated genes in colonic tissues of mice treated with 0.0%, 1.5%, 2.0%, 2.5%, and 3.0% (w/v) DSS. High concentrations of DSS (2.5% and 3.0%) markedly increased the expression of genes linked to epithelial barrier damage, cytokine signaling, and immune cell recruitment, whereas low concentrations (1.5% and 2.0%) induced only modest changes. (B) Heatmap illustrating the expression of key inflammatory signaling pathways, including TNF, IL-17, NF-κB, and chemokine pathways. Expression intensity increased progressively with the DSS dose, indicating a dose-dependent activation of pro-inflammatory signaling networks consistent with disease severity. Dose-dependent activation of inflammatory pathways and immune cell recruitment in DSS-induced colitis Differential gene expression analysis between DSS-treated and control mice further revealed that DSS upregulated several apoptosis-related genes, including Gml and Gml2, as well as immune-response genes such as members of the S100 family (S100a8 and S100a9) and chemokines Cxcl2 and Cxcl3, which are known to mediate T cell recruitment to the colon (Figure 5A). We then evaluated the expression of interleukins and found that the pro-inflammatory cytokines including Il1a, Il1b, Il16, and Il17 were sharply induced in the 3.0% DSS group, while the 2.0% and 2.5% DSS groups displayed comparable transcript levels (Figure 5B). Given the involvement of S100 proteins and chemokines in T cell recruitment, we next assessed the effect of DSS on CD8+ T cell infiltration by examining the expression of the surface markers CD8α and CD8β (Figure 5C). Interestingly, even low-dose DSS (1.5%) was sufficient to increase the number of CD8+ T cells in the colon. In contrast, 3.0% DSS induced markedly higher expression of CD8α and CD8β than 1.5% DSS, whereas the 2.0% and 2.5% groups showed no substantial difference. Flow cytometry further confirmed these immune alterations (Figure 5D). DSS treatment increased the proportions of CD11b+F4/80+ macrophages and CD11b+Ly6G+ neutrophils, along with elevated TNF-α production. In parallel, CD4+ and CD8+ T cells were expanded, accompanied by increased frequencies of IFN-γ+ (Th1) and IL-17A+ (Th17) CD4+ T cells, particularly at higher DSS doses. Collectively, these findings demonstrate that intestinal immune cells are highly sensitive to DSS exposure. Even mild DSS treatment is capable of disrupting immune homeostasis, suggesting that low-dose DSS models may be useful for studying early immune alterations in mild colitis. DSS-induced colonic immune disorders in mice (A) DEGs between the DSS-treated and control mice. (B) Expression levels of interleukins. The density of CD8+ T cells detected by immunofluorescence. (C) Representative images of DSS-treated colons stained for CD8α (red) and CD8β (green) with DAPI (blue). Mean fluorescence intensities of CD8α and CD8β were normalized to the 0.0% DSS group. Scale bars, 50 μm. (D) Flow cytometry analysis of the intestinal immune cell populations. Representative contour plots showing gating on CD11b+F4/80+ macrophages, CD11b+Ly6G+ neutrophils, and CD3+CD4+ T cells, followed by intracellular cytokine staining for TNF-α, IFN-γ, and IL-17A. Quantification of immune cell subsets, including CD11b+F4/80+ macrophages (% of CD45+), TNF-α+ macrophages (% of CD11b+F4/80+), CD11b+Ly6G+ neutrophils (% of CD45+), TNF-α+ neutrophils (% of CD11b+Ly6G+), CD8+ T cells (% of CD3+), CD4+ T cells (% of CD3+), Th1 cells (IFN-γ+, % of CD4+), and Th17 cells (IL-17A+, % of CD4+). Data are presented as the mean ± SEM (n = 5–8 per group). Statistical significance was determined using one-way ANOVA, followed by Tukey’s multiple-comparisons test. ∗p < 0.05, ∗∗p < 0.01. Discussion The intestinal ecosystem is maintained through a highly coordinated network among gut microbiota, intestinal epithelial cells (IECs), and host immune cells, collectively ensuring mucosal homeostasis. Disruption of this balance generates a pro-inflammatory milieu that promotes intestinal injury.12 Extensive evidence supports the dysbiosis theory, demonstrating that, compared with healthy individuals, patients with IBD exhibit marked alterations in microbial diversity, composition, and localization,13 commonly characterized by reductions in Bacteroidetes and Firmicutes and enrichment of Proteobacteria and Actinobacteria.14,15,16 Beyond pathogenic shifts, commensal organisms such as Firmicutes, Bacteroidetes, and Bifidobacteria contribute to host protection by generating short-chain fatty acids (SCFAs), including acetate, propionate, and butyrate, which serve as essential metabolic substrates for IECs and promote epithelial barrier integrity.17,18 In our study, DSS triggered dose-dependent microbial dysbiosis. DSS exposure enriched Bacteroidota and Verrucomicrobiota while reducing Firmicutes, Proteobacteria, Actinobacteriota, and Cyanobacteria, with notable shifts already evident at 1.5% DSS. Potentially pathogenic genera such as Helicobacter and Clostridium_sensu_stricto_1 emerged in 2.0%–3.0% DSS-treated mice, consistent with reports linking these taxa to T cell dysregulation and epithelial stress during colitis progression.19,20 Importantly, our findings support a growing concept that microbial dysbiosis precedes overt epithelial injury, indicating that the microbiota acts as an upstream regulator of DSS-induced inflammation.21,22 CD8+ T cells constitute a critical interface between the epithelial stress signals and immune activation. Although their involvement in IBD has long been recognized, their functional roles remain context dependent.23 Classical cytotoxic CD8+ T lymphocytes eliminate damaged IECs and constrain pathogen dissemination, whereas regulatory CD8+ subsets contribute to immune restraint and tissue homeostasis.24,25 In addition, non-cytotoxic subsets such as Tc17 cells produce IL-17A, IL-17F, and IL-22, supporting epithelial repair and barrier function.26,27 Consistent with this functional heterogeneity, we observed a dose-dependent expansion of colonic CD8+ T cells, with distinct population dynamics between low- and high-dose DSS exposure. Even mild DSS (1.5%) was sufficient to promote CD8+ T cell accumulation, whereas higher concentrations (3.0%) were associated with a more pronounced inflammatory phenotype. These findings are in line with recent single-cell studies highlighting substantial plasticity within intestinal CD8+ T cell compartments in inflammatory settings.28,29 Beyond CD8+ T cells, flow cytometric analyses revealed coordinated remodeling of additional immune compartments, including increased macrophage and granulocyte infiltration and enhanced differentiation of CD4+ T cell subsets, particularly Th1 and Th17 cells, together with elevated TNF-α production. These observations are consistent with previous studies demonstrating that epithelial injury and microbiota perturbation drive coupled innate and adaptive immune activation in colitis. At the signaling level, the activation of multiple pattern-recognition receptor (PRR) pathways, including Toll-like receptor (TLR), NOD-like receptor (NLR), and RIG-I-like receptor signaling, underscores a central role for innate immune sensing in DSS-induced inflammation.30,31 Notably, RIG-I-like receptor signaling was responsive even to low-dose DSS, suggesting that epithelial-intrinsic antiviral sensors may also detect microbiota-derived signals or cellular stress. Such integration of epithelial stress sensing, PRR activation, and immune cell recruitment supports a model in which early mucosal immune responses are shaped by coordinated crosstalk between the epithelial and immune compartments.30,31 Downstream inflammatory pathways, including TNF, NF-κB, and IL-17 signaling, were progressively amplified with increasing DSS concentrations, consistent with the escalating immune cell infiltration and cytokine production.32 Collectively, these findings support a multi-compartment framework in which microbiota perturbation, epithelial stress, and immune activation act in concert to determine the severity and trajectory of DSS-induced colitis.33,34 Importantly, our findings underscore the value of low-dose DSS (1.5%–2.0%) models, which, despite inducing only mild and reversible acute inflammation, significantly disrupt the microbial composition and trigger early immune cell differentiation. These subtle yet biologically relevant changes make low-dose DSS models particularly useful for studying early events in mucosal immune dysregulation, microbiota-immune crosstalk, and the transition from acute to chronic inflammation. Meanwhile, high-dose DSS (≥2.5%) induces widespread epithelial destruction and immune hyperactivation, providing a more suitable model for severe ulcerative colitis-like pathology and therapeutic intervention studies. Overall, our study demonstrates that the severity and nature of DSS-induced colitis are governed by dose-dependent interactions among the microbiota, epithelium, and immune system, supporting a multi-layered framework in which dysbiosis, epithelial stress sensing, and adaptive immune activation work in concert to shape intestinal inflammation. These findings advance our understanding of the cellular and molecular mechanisms driving DSS-induced pathology and highlight the importance of dose selection in experimental colitis models. Limitations of the study Several limitations should be acknowledged. Although we established a graded dose-response framework for DSS-induced colitis by integrating pathological, microbial, transcriptomic, and immune features, the analyses were restricted to a single mouse strain, male mice, and an acute injury paradigm. Therefore, the extent to which these patterns generalize across genetic backgrounds, sexes, or chronic and relapsing models remains unresolved. Importantly, susceptibility to DSS-induced colitis is strongly influenced by the gut microbiota composition, which varies across animal facilities and even among genetically identical mice.35,36,37 Such microbiota-dependent variability can modulate immune phenotypes, disease severity, and experimental outcomes, thereby shaping the apparent dose sensitivity observed here. Accordingly, DSS concentrations may require empirical optimization across laboratories, and both the magnitude and kinetics of disease progression may differ depending on the microbial context. In addition, variability in DSS source and physicochemical properties, particularly molecular weight and batch differences, can critically influence its colitogenic activity and dose-response behavior.38 The framework defined here was established using DSS from MP Biomedicals (36–50 kDa), and extrapolation to other preparations should be approached with caution. Finally, although flow cytometry provided resolution of major immune cell subsets, transcriptomic analyses were performed at the bulk tissue level, limiting precise attribution of gene expression changes to specific cellular populations. Future studies incorporating longitudinal designs, multi-omics integration, and single-cell-resolved approaches are required to further refine and generalize this framework. Resource availability Lead contact Requests for further information and resources should be directed to and will be fulfilled by the lead contact, Miaomiao Wu (miaomiaowu0316@hunau.edu.cn). Materials availability This study did not generate new unique reagents or materials. Data and code availability •RNA-seq and 16S rRN A sequencing data generated in this study have been deposited in the NCBI SRA under BioProject accession numbers. The accession numbers are NCBI BioProject: PRJNA1392172 for RNA-seq and NCBI BioProject: PRJNA1392345 for 16S rRNA sequencing. These datasets have been released and are publicly accessible through NCBI (https://www.ncbi.nlm.nih.gov/). All other data reported in this manuscript will be shared by the lead contact upon request. RNA-seq and 16S rRN A sequencing data generated in this study have been deposited in the NCBI SRA under BioProject accession numbers. The accession numbers are NCBI BioProject: PRJNA1392172 for RNA-seq and NCBI BioProject: PRJNA1392345 for 16S rRNA sequencing. These datasets have been released and are publicly accessible through NCBI (https://www.ncbi.nlm.nih.gov/). All other data reported in this manuscript will be shared by the lead contact upon request. •This study did not generate new code. This study did not generate new code. •Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request. Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request. Acknowledgments This work was supported by Yuelushan Laboratory Breeding Program (YLS-2025-ZY03011), the National Natural Science Foundation of China (32573249 and 82470126), Natural Science Foundation of Hunan Province (2025JJ30011), Huxiang Youth Talent Program of Hunan Province (2023RC3030), Hunan Provincial Natural Science Foundation of China (no. 2024JC0007), the Open Project Program of Sichuan Provincial Key Laboratory of Animal Disease-resistant Nutrition, Sichuan Agricultural University (grant no. SZ20250102), and the Major Special Project of the Changsha Science and Technology Plan (kh2401015). We also thank Xiaolu Jin from China Agricultural University for technical support. Author contributions S.C. and J.P. performed the experiments, collected the data, and drafted the manuscript; Y.W. and M.H. assisted with methodology, data curation, and statistical analysis; L.Z. and B.F. contributed to resources, technical support, and data interpretation; P.H. assisted with bioinformatic analysis and visualization; J.L. supervised the study, guided the experimental design, and revised the manuscript; M.W. contributed to conceptualization, project administration, and critical revision of the manuscript; Y.Y. conceived the project, acquired funding, supervised the overall research, and revised and approved the final manuscript. Declaration of interests The authors declare no competing interests. Declaration of generative AI and AI-assisted technologies in the writing process During the preparation of this work, the authors used ChatGPT in order to improve language clarity and grammar. After using this tool, the authors reviewed and edited the content as needed and take full responsibility for the content of the publication. STAR★Methods Key resources table REAGENT or RESOURCESOURCEIDENTIFIERAntibodiesAPC-Cy™7 Rat Anti-Mouse CD45BD PharmingenCat# 561037;RRID: AB_396774BV650 Rat Anti-Mouse CD3BD PharmingenCat# 569683;RRID: AB_3685227PerCP Rat Anti-Mouse CD4BD PharmingenCat# 561090;RRID: AB_394587FITC Rat Anti-Mouse CD8aBD PharmingenCat# 561966;RRID: AB_10896291BB515 Rat Anti-CD11bBD PharmingenCat# 564455;RRID: AB_2665392BV510 Rat Anti-Mouse Ly-6GBD PharmingenCat# 569685;RRID: AB_3685229PE Rat Anti-Mouse F4/80BD PharmingenCat# 565410; RRID: AB_2687527PE-Cy™7 Rat Anti-Mouse IFN-γBD PharmingenCat# 561040; RRID: AB_396766Alexa Fluor® 647 Rat anti-Mouse IL-17ABD PharmingenCat# 560224; RRID: AB_1645204APC Rat Anti-Mouse TNFBD PharmingenCat# 561062; RRID: AB_398553Anti-Mouse CD8aInvitrogenCat# 14-0195-82; RRID: AB_2637159Anti-CD8 betaAbcamCat# ab228965; N/AChemicals, peptides, and recombinant proteinsDextran sulfate sodium (DSS), 36–50 kDaMP BiomedicalsCat# 9011-18-1; N/AFixable Viability Stain 575VBD PharmingenCat# 565694;Cytiva Percoll™ Centrifugation MediaCytivaCat# 45-001-747DNase IRocheCat# 10104159001DL-Dithiothreitol (DTT)Sigma-AldrichCat# D9779Ethylenedinitrilotetraacetic acid (EDTA)Sigma-AldrichCat# E9884Cytofix/Cytoperm™ Fixation/Permeabilization KitBD PharmingenCat# 554714Deposited dataRNA-seq dataNCBI SRANCBI BioProject: PRJNA139217216S rRNA sequencing dataNCBI SRANCBI BioProject: PRJNA1392345Experimental models: Organisms/strainsC57BL/6J miceHunan SJA Laboratory Animal Co., Ltd.Strain: C57BL/6J; N/ASoftware and algorithmsZEN Microscopy SoftwareCarl Zeiss Microscopy GmbHhttps://www.zeiss.com/microscopy/en/products/software/zeiss-zen.htmlGraphPad Prism version 10.0GraphPad Softwarehttps://www.graphpad.com/IBM SPSS Statistics version 26.0IBMhttps://www.ibm.com/products/spss-statisticsFlowJo version 10.1BD Bioscienceshttps://www.flowjo.com/QIIME2 version 2022.2QIIME2https://qiime2.org/ Experimental model and study participant details Wild-type male C57BL/6J mice, 6 weeks of age, were obtained from Hunan SJA Laboratory Animal Co., Ltd. (Hunan, China). After a 7-day acclimatization period, mice were housed under specific pathogen-free conditions in a controlled environment at 23 ± 1°C with a 12 h light/12 h dark cycle and ad libitum access to food and water. Mice were randomly divided into five groups. Four groups received 1.5%, 2.0%, 2.5%, or 3.0% (w/v) DSS in their drinking water for 5 days, followed by a 2-day recovery period with regular water. The remaining group served as the untreated control. Body weight was recorded daily throughout the experiment. All animal procedures were conducted in accordance with institutional guidelines and were approved by the Animal Ethics Committee of Hunan Agricultural University (approval number: 202005). Only male mice were used in this study; therefore, the potential influence of sex on DSS-induced colitis severity, microbiota remodeling, transcriptomic responses, and immune activation was not assessed. No human participants, human samples, cell lines, primary cell cultures, plants, or microbial strains were used in this study. The disease activity index (DAI) was calculated based on body weight loss, stool consistency, and the presence of occult or gross blood in the stool. At the end of the experiment, following a 12-h fast, mice were anesthetized for tissue collection. Livers were excised, rinsed in cold phosphate-buffered saline (PBS), and weighed. Colons were carefully dissected, and their lengths were measured. Fecal samples were collected aseptically and stored in sterile cryovials. Colon segments (0.5 cm) were fixed in 4% paraformaldehyde for histological and immunohistochemical analyses, while additional 2 cm sections were snap-frozen in liquid nitrogen and stored at −80°C for subsequent analyses. Method details Histological score analysis Histopathological scoring was performed in a blinded manner by a qualified pathologist according to Aggarwal et al.12 The scoring criteria included villus blunting, crypt loss and regeneration, epithelial cell apoptosis, immune cell infiltration, and mucosal ulceration of the lamina propria. Lesions were graded on a five-point scale: 0.5, focal and rare; 1, focal and mild; 2, diffuse and mild; 3, diffuse and moderate; and 4, diffuse and severe. Sequencing and processing of bacterial 16S rRNA gene in fecal DNA Fecal samples were collected individually from each mouse using sterile forceps and immediately stored at −80 °C until DNA extraction. Total microbial genomic DNA was isolated using the DNeasy PowerLyzer PowerSoil Kit (Qiagen, Germany) according to the manufacturer’s protocol. DNA concentration and purity were determined using a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, USA) and Qubit dsDNA HS Assay Kit (Invitrogen, USA). DNA integrity was confirmed by 1% agarose gel electrophoresis. The hypervariable V3-V4 region of the bacterial 16S rRNA gene was amplified using the universal primers 341F (5′-CCTACGGGNGGCWGCAG-3′) and 806R (5′-GGACTACHVGGGTATCTAAT-3′) with Ex Taq Hot Start (Takara Bio, Japan). PCR reactions were performed in 25 μL volumes containing 12.5 μL of 2× Taq Master Mix, 1 μL of each primer (10 μM), and 50 ng of template DNA. The amplification protocol consisted of an initial denaturation at 95 °C for 3 min, followed by 30 cycles of 95 °C for 30 s, 55 °C for 30 s, and 72 °C for 45 s, with a final extension at 72 °C for 10 min. PCR products were verified by 2% agarose gel electrophoresis and purified using AMPure XP beads (Beckman Coulter, USA). Purified amplicons were quantified with a Qubit 2.0 fluorometer and pooled in equimolar concentrations. Libraries were constructed using the MiSeq Reagent Kit v3 (600-cycle) and sequenced on the Illumina MiSeq platform (Illumina, USA) to generate paired-end reads (2 × 300 bp). Raw reads were processed using the QIIME2 (2022.2) pipeline. Reads were demultiplexed, and adapter, barcode, and primer sequences were trimmed using cutadapt. Quality filtering was performed with DADA2 using a Phred quality threshold of Q30, and low-quality reads or chimeras were removed. High-quality sequences were clustered into operational taxonomic units (OTUs) at 96% sequence identity using the UCLUST algorithm. Taxonomic assignment was conducted using GLSEARCH against the RDP (version 10.27, CORE update September 2012) and NCBI 16S rRNA databases. To ensure equal sequencing depth, samples were rarefied to 10,000 reads per sample based on rarefaction curves. α-diversity indices (Chao1, Shannon, and Simpson) were calculated to assess species richness and diversity. β-diversity was evaluated using weighted and unweighted UniFrac distances and visualized by principal coordinate analysis (PCoA). Statistical differences in microbial community composition were tested using PERMANOVA in QIIME2. Distinct clustering of microbial communities was observed between experimental groups, as indicated by PCoA and Shannon diversity analyses. RNA sequencing Total RNA was extracted from tissue samples using TRIzol reagent (Thermo Fisher Scientific, USA) according to the manufacturer’s instructions. RNA quantity and purity were assessed using a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, USA), and RNA integrity was evaluated with an Agilent 2100 Bioanalyzer and RNA 6000 Nano LabChip Kit (Agilent Technologies, CA, USA). Samples with an RNA integrity number (RIN) > 7.0 were used for subsequent analyses. Messenger RNA (mRNA) was enriched from total RNA using Oligo(dT)-attached magnetic beads, fragmented into short sequences by divalent cations under elevated temperature, and reverse-transcribed into complementary DNA (cDNA) using random hexamer primers. Approximately 5 μg of total RNA per sample was used for cDNA library preparation. Libraries were constructed using the Illumina TruSeq RNA Sample Preparation Kit (Illumina, San Diego, CA, USA) following the manufacturer’s protocol. The resulting cDNA libraries were quantified using a Qubit fluorometer (Invitrogen, USA), and their insert size distribution was verified on the Agilent 2100 Bioanalyzer. Sequencing was performed on an Illumina NovaSeq 6000 platform (Illumina, USA) to generate 150 bp paired-end reads. Raw sequencing reads were processed to remove adapter sequences and low-quality bases using Trimmomatic, and the quality of clean reads was assessed using FastQC. Clean reads were mapped to the Mus musculus reference genome (GRCm39) using HISAT2, and gene expression levels were quantified as fragments per kilobase of transcript per million mapped reads (FPKM) using StringTie. Time-course clustering analysis was performed using the Mfuzz package in R to identify temporal expression patterns across increasing DSS concentrations. Expression values were standardized, and genes were assigned to soft clusters based on their response trajectories, enabling the identification of dose-responsive gene modules. Differentially expressed genes (DEGs) were identified using the DESeq2 package in R with thresholds of adjusted p < 0.05 and |log2 (fold change) | > 1.5. Functional enrichment analyses of DEGs were conducted using Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) over-representation analysis implemented in the clusterProfiler R package. All raw sequencing data and processed files were uploaded to the Novogene Bioinformatics platform (https://magic.novogene.com) for further analysis. Immunofluorescence Tissue sections were rinsed with phosphate-buffered saline (PBS) and subjected to antigen retrieval as described above. After washing in PBS, sections were blocked with 5% bovine serum albumin (BSA) for 20 min at room temperature to prevent nonspecific binding. Slides were then incubated overnight at 4 °C with primary antibodies listed in the key resources table. After three washes in PBS, sections were incubated with Alexa Fluor-conjugated secondary antibodies (1:500, Invitrogen, USA) for 1 h at room temperature (25 °C) in the dark. Negative control slides were processed in parallel under identical conditions, except that the primary antibodies were omitted. Nuclei were counterstained with 4′,6-diamidino-2-phenylindole (DAPI, 1 μg/mL) for 5 min, rinsed in PBS, and mounted with antifade mounting medium. Fluorescence images were captured using a Zeiss LSM-710 confocal laser scanning microscope (Carl Zeiss, Germany), and images were processed using ZEN Microscopy Software. Flow cytometry Colon tissues were opened longitudinally, cut into 1–2 cm pieces, and incubated in HBSS/EDTA buffer (1× HBSS supplemented with 5 μM EDTA) for 2 × 20 min at 37°C with shaking (500 rpm). After each incubation, tissues were filtered through a 150-μm nylon strainer, and the flow-throughs were collected as fraction I. The remaining tissues were washed with HBSS/HEPES and further digested in HBSS containing CaCl2 (1 mM), collagenase VIII (0.3 mg/mL), and DNase I (25 μg/mL) for approximately 45 min at 37°C (300 rpm). The digested suspension was filtered through a 100-μm strainer to obtain fraction II. Fractions I and II were combined and purified by 40%/80% Percoll density gradient centrifugation to isolate intestinal immune cells. For flow cytometry analysis, cells were stained with fluorophore-conjugated antibodies as listed in the key resources table, following standard protocols. Dead cells were excluded using Fixable Viability Stain. Intracellular cytokine staining was performed using the Cytofix/Cytoperm kit (BD Pharmingen) according to the manufacturer’s instructions. Data acquisition was carried out on a FACSCelesta cytometer (BD Biosciences), and compensation was established using single-stained controls. Data were analyzed with FlowJo v10 (BD Biosciences), with low-quality events removed using FlowAI. For quantitative analysis, immune cell populations were expressed as percentages of their parent populations (e.g., CD45+ or CD3+ cells) and, where applicable, normalized to total viable cell counts to reflect relative abundance. Quantification and statistical analysis Statistical analyses were performed using GraphPad Prism and SPSS where appropriate. Data are presented as mean ± SEM unless otherwise indicated. For animal experiments, each data point represents an individual mouse as an independent biological replicate. The number of biological replicates for each experiment is indicated in the corresponding figure legends. For comparisons among multiple DSS treatment groups, statistical significance was determined using one-way analysis of variance (ANOVA) followed by Tukey’s multiple comparisons test. For microbiota community analyses, beta diversity was assessed using weighted and unweighted UniFrac distances, and group differences were evaluated by PERMANOVA in QIIME2. Differential microbial taxa were identified using LEfSe analysis with the Kruskal-Wallis test, with p < 0.05 and an LDA score threshold of 4.0. For RNA-seq analysis, differentially expressed genes were identified using DESeq2 with an adjusted p value <0.05 and |log2(fold change)| > 1.5. Functional enrichment analyses were performed using the clusterProfiler package in R. Statistical significance is indicated in the figures as follows: ∗p < 0.05 and ∗∗p < 0.01. Exact sample sizes, statistical tests, and definitions of error bars are provided in the corresponding figure legends. Footnotes Supplemental information can be found online at https://doi.org/10.1016/j.isci.2026.116892. Contributor Information Jianjun Li, Email: jianjunli@isa.ac.cn. Miaomiao Wu, Email: miaomiaowu0316@hunau.edu.cn. Yulong Yin, Email: yinyulong@isa.ac.cn. Supplemental information Document S1. Figure S1 References Associated Data Supplementary Materials Document S1. Figure S1 Data Availability Statement •RNA-seq and 16S rRN A sequencing data generated in this study have been deposited in the NCBI SRA under BioProject accession numbers. The accession numbers are NCBI BioProject: PRJNA1392172 for RNA-seq and NCBI BioProject: PRJNA1392345 for 16S rRNA sequencing. These datasets have been released and are publicly accessible through NCBI (https://www.ncbi.nlm.nih.gov/). All other data reported in this manuscript will be shared by the lead contact upon request. RNA-seq and 16S rRN A sequencing data generated in this study have been deposited in the NCBI SRA under BioProject accession numbers. The accession numbers are NCBI BioProject: PRJNA1392172 for RNA-seq and NCBI BioProject: PRJNA1392345 for 16S rRNA sequencing. These datasets have been released and are publicly accessible through NCBI (https://www.ncbi.nlm.nih.gov/). All other data reported in this manuscript will be shared by the lead contact upon request. •This study did not generate new code. This study did not generate new code. •Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request. Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.