A pathogenic CD8⁺ TRM–chemokine axis orchestrates liver fibrosis and provides circulating biomarkers during chronic Clonorchis sinensis infection Authors: Xinyue Du, Jiashun Li, Xin Wang, Wenyu Wu, Tingjun Zhu, Huibo Yan, Chenyun Wu, Yuanyuan Li, Yuwei Gao, Haibo Yi, Kokouvi Kassegne, Xiangwei Li, Xiaoyin Niu, Helena Helmby, Samuel C. Wassmer, Youqiong Ye, Banchob Sripa, Guangjie Chen, Menbao Qian, Zhaojun Wang Published: 2026 ABSTRACT Chronic liver fluke infections represent a major global health burden, driving liver fibrosis and elevating cancer risk, yet the immunological mechanisms linking persistent infection to tissue damage remain poorly defined. Using Clonorchis sinensis, a carcinogenic liver fluke infecting millions worldwide, as a representative model, we map immune responses during chronic infection in mice. We show that intrahepatic CD8⁺ tissue-resident memory T cells (TRM) progressively accumulate and drive pathology by producing chemokines that promote immune recruitment and fibroblast activation. Depletion of CD8⁺ T cells during defined disease phases markedly alleviates liver inflammation and fibrosis, demonstrating their causal role. Strikingly, C-C motif chemokine ligand 4, predominantly secreted by CD8⁺ TRM cells, correlates with fibrosis severity in both experimental models and patient cohorts, establishing it as a circulating biomarker for noninvasive disease monitoring. These findings define a pathogenic CD8⁺ TRM–chemokine axis underlying helminth-induced liver injury and highlight TRM-derived chemokines as biomarkers and therapeutic targets for fibrosis across chronic parasitic infections. Introduction Clonorchis sinensis, a Group 1 carcinogenic helminth infecting more than 15 million people worldwide, is a major cause of chronic biliary inflammation, progressive liver fibrosis, and increased risk of cholangiocarcinoma1–3. Considerable progress has been made in understanding peripheral immune responses during C. sinensis and related helminth infections, particularly in acute animal models, which have revealed important aspects of host–parasite interactions and immune modulation4–6. However, these approaches insufficiently capture the chronic phase that underlies human disease and provide limited insight into immune processes within the liver—the primary site of parasite colonization and pathology. As a result, the tissue-specific mechanisms linking persistent infection to progressive fibrosis remain poorly defined. To address this gap, we established a chronic C. sinensis mouse model that faithfully recapitulates the human disease course. Using this system, we present a comprehensive, high-resolution characterization of the spatiotemporal dynamics of host immunity across disease progression, revealing a central role for CD8⁺ tissue-resident memory T (TRM) cells and CCL chemokine signaling in orchestrating hepatic immunopathology. Our analyses demonstrated that CD8⁺ TRM cells act as pivotal drivers of liver pathology through chemokine-mediated immune recruitment and fibroblast activation. Importantly, we identified CCL4, predominantly produced by CD8⁺ TRM cells, as a circulating biomarker correlating with fibrosis severity in both mice and patients. Together, these findings establish a pathogenic CD8⁺ TRM–chemokine axis as a central mechanism of helminth-induced liver fibrosis and highlight TRM-derived chemokines as promising biomarkers and therapeutic targets. Results Chronic C. sinensis infection reshapes hepatic immunity from early Th2 polarization to persistent inflammation To establish a chronic infection model, BALB/c mice were orally gavaged with C. sinensis metacercariae, and blood and liver samples were collected at baseline and at 2, 6, and 16 weeks post-infection, corresponding to early acute, acute, and chronic phases (Supplementary Fig. 1a). Human C. sinensis infection causes hepatic inflammation and fibrosis with changes in serum liver function markers1,7. We therefore assessed these parameters in a mouse model. Histopathology revealed progressive hepatic injury and fibrosis, accompanied by biochemical evidence of liver dysfunction, including elevated ALT, AST, ALP, bile acids, increased total protein, and a reduced albumin-to-globulin ratio (Supplementary Fig. 1b–d). Infected mice also developed strong humoral responses to adult worm antigens, with marked increases in antigen-specific IgG, IgM, IgA and IgE (Supplementary Fig. 1e). Together, these findings confirm successful establishment of a chronic murine model recapitulating key features of human infection. We then analyzed hepatic immunity to reveal dynamic temporal changes. Multiplex immunohistochemistry showed increasing immune infiltration, dominated by CD11b⁺ myeloid cells (P = 0.0003) at 2 weeks, followed by recruitment of B220⁺ B cells (P = 0.0018) and CD4⁺/CD8⁺ T cells (P < 0.0001) at 6 weeks, and culminating in marked CD8⁺ T cell (P < 0.0001) accumulation at 16 weeks (Fig. 1a, b and Supplementary Fig. 2a, b). Flow cytometry further demonstrated early activation of innate populations (NK, NKT, neutrophils, mast cells, macrophages), broad immune expansion by 6 weeks, and a CD8⁺ T cell-dominated environment in the chronic phase (Fig. 1c and Supplementary Fig. 3). Th2-associated cytokines (IL-4 (P = 0.0002), IL-5 (P < 0.0001), IL-10 (P = 0.0441), IL-13 (P = 0.0373)) peaked at 6 weeks, while pro-inflammatory mediators (IL-6 (P = 0.0002), IL-17 (P = 0.0165), IL-27 (P < 0.0001), TNF (P = 0.0032), IFN-γ (P = 0.0001)) predominated by 16 weeks (Fig. 1d). Collectively, these results delineate a coordinated evolution of hepatic immunity during C. sinensis infection, transitioning from early Th2 polarization driven by innate responses to chronic inflammation characterized by CD8⁺ T cell dominance and elevated pro-inflammatory cytokines.Fig. 1Hepatic local immune responses during long-term C. sinensis infection.BALB/c mice were orally infected with C. sinensis metacercariae and analyzed at 0, 2, 6, and 16 weeks post-infection. a Representative multiplex immunohistochemistry images of liver sections stained for DAPI (blue), CD11b (green), B220 (yellow), CD4 (white), and CD8 (red) (female, n = 5 mice per time point). Scale bars: top, 500 μm; bottom, 100 μm. b Quantification of hepatic immune cell populations from immunostained liver sections in (a). c Relative abundance of liver immune cell subsets measured by flow cytometry, normalized to uninfected controls (week 0) (week 0 and 6, n = 8 mice; week 2 and 16, n = 6 mice). d Cytokine concentrations in liver homogenate supernatants across time points (week 0, 6, and 16, n = 10 mice; week 2, n = 6 mice). Data are presented as mean ± SEM from at least two independent experiments. Statistical significance was assessed by one-way ANOVA with Tukey’s multiple comparisons test or Kruskal-Wallis test with Dunn’s multiple comparisons test (b–d). P values are indicated in the corresponding panels. Source data are provided as a Source Data file. Hepatic local immune responses during long-term C. sinensis infection. BALB/c mice were orally infected with C. sinensis metacercariae and analyzed at 0, 2, 6, and 16 weeks post-infection. a Representative multiplex immunohistochemistry images of liver sections stained for DAPI (blue), CD11b (green), B220 (yellow), CD4 (white), and CD8 (red) (female, n = 5 mice per time point). Scale bars: top, 500 μm; bottom, 100 μm. b Quantification of hepatic immune cell populations from immunostained liver sections in (a). c Relative abundance of liver immune cell subsets measured by flow cytometry, normalized to uninfected controls (week 0) (week 0 and 6, n = 8 mice; week 2 and 16, n = 6 mice). d Cytokine concentrations in liver homogenate supernatants across time points (week 0, 6, and 16, n = 10 mice; week 2, n = 6 mice). Data are presented as mean ± SEM from at least two independent experiments. Statistical significance was assessed by one-way ANOVA with Tukey’s multiple comparisons test or Kruskal-Wallis test with Dunn’s multiple comparisons test (b–d). P values are indicated in the corresponding panels. Source data are provided as a Source Data file. Progressive accumulation of CD8⁺ tissue-resident memory T cells in the liver during chronic infection To further dissect the cellular landscape of the liver during C. sinensis infection, we performed single-cell RNA sequencing (scRNA-seq) of CD45⁺ NPCs from uninfected, acutely infected (6 weeks), and chronically infected (16 weeks) mice. After quality control, 53,444 cells were retained for analysis. Unsupervised clustering defined 11 major immune cell populations (Fig. 2a–c). Pseudobulk differential state testing highlighted T cells as the most transcriptionally perturbed population (Fig. 2d), and pseudobulk gene set enrichment analysis (GSEA) further revealed extensive functional reprogramming of T cells in infected livers (Fig. 2e). These findings were consistent with previous flow cytometry and imaging data, underscoring the central involvement of T cells in infection-induced hepatic immunity.Fig. 2Dynamics of hepatic immune cells and T cell subsets during long-term C. sinensis infection.BALB/c mice were orally infected with C. sinensis metacercariae. a–l Hepatic non-parenchymal cells were isolated from uninfected mice (Control) and from mice at the acute (6 weeks post-infection, Cs_acute) and chronic (16 weeks post-infection, Cs_chronic) phases, followed by single-cell RNA sequencing. a Bubble heatmap of lineage-specific marker expression across immune cells. b Uniform manifold approximation and projection (UMAP) visualization of immune cell clustering and composition across infection phases. c UMAP plot showing phase-wise distribution of immune cells; equal numbers of cells per group were randomly sampled for visualization only. d Bar plot displaying the number of differentially expressed genes (DEGs) in immune cells. e Gene set enrichment analysis (GSEA) of positive regulation of T cell activation. Differential expression was assessed using quasi-likelihood F-tests, and P values were adjusted using Benjamini-Hochberg method. NES, normalized enrichment score; p.adj, adjusted P value. f Bubble heatmap of subset-specific marker expression across T cells. g UMAP visualization of T cell subset clustering and composition across infection phases. h UMAP plot showing phase-wise distribution of T cell subsets; equal numbers of cells were randomly sampled for visualization. i Phase preference of T cell subsets estimated by Ro/e score. j Violin plot of tissue residency signature scores across T cell subsets, showing the quartiles (dashed lines) and median (solid lines). P values versus CD8⁺ TRM were calculated with two-sided Wilcoxon rank-sum tests, adjusted using Benjamini-Hochberg method and shown above each group. k Venn diagram showing the overlap and uniqueness of TCR clonotypes among CD8⁺ naive (TN), CD8⁺ effector memory (TEM) and CD8⁺ tissue-resident memory (TRM) T cells. l Feature plot of T cell subsets clonotype abundance and distribution. m Representative flow cytometry plots showing frequencies of CD8⁺CD62L⁺CD69⁻ (circulating) and CD8⁺CD69⁺CXCR6⁺ (tissue-resident) T cell subsets in different tissues of control and infected female mice. The percentages are shown. n Representative histograms of CFSE expression showing antigen-specific proliferation of hepatic CD8⁺ T cells from control (Con) and infected (Cs) mice. CsAWA, C. sinensis adult worm antigen. m,n Data are representative of two independent experiments, n = 5 mice per group. Source data are provided as a Source Data file. Dynamics of hepatic immune cells and T cell subsets during long-term C. sinensis infection. BALB/c mice were orally infected with C. sinensis metacercariae. a–l Hepatic non-parenchymal cells were isolated from uninfected mice (Control) and from mice at the acute (6 weeks post-infection, Cs_acute) and chronic (16 weeks post-infection, Cs_chronic) phases, followed by single-cell RNA sequencing. a Bubble heatmap of lineage-specific marker expression across immune cells. b Uniform manifold approximation and projection (UMAP) visualization of immune cell clustering and composition across infection phases. c UMAP plot showing phase-wise distribution of immune cells; equal numbers of cells per group were randomly sampled for visualization only. d Bar plot displaying the number of differentially expressed genes (DEGs) in immune cells. e Gene set enrichment analysis (GSEA) of positive regulation of T cell activation. Differential expression was assessed using quasi-likelihood F-tests, and P values were adjusted using Benjamini-Hochberg method. NES, normalized enrichment score; p.adj, adjusted P value. f Bubble heatmap of subset-specific marker expression across T cells. g UMAP visualization of T cell subset clustering and composition across infection phases. h UMAP plot showing phase-wise distribution of T cell subsets; equal numbers of cells were randomly sampled for visualization. i Phase preference of T cell subsets estimated by Ro/e score. j Violin plot of tissue residency signature scores across T cell subsets, showing the quartiles (dashed lines) and median (solid lines). P values versus CD8⁺ TRM were calculated with two-sided Wilcoxon rank-sum tests, adjusted using Benjamini-Hochberg method and shown above each group. k Venn diagram showing the overlap and uniqueness of TCR clonotypes among CD8⁺ naive (TN), CD8⁺ effector memory (TEM) and CD8⁺ tissue-resident memory (TRM) T cells. l Feature plot of T cell subsets clonotype abundance and distribution. m Representative flow cytometry plots showing frequencies of CD8⁺CD62L⁺CD69⁻ (circulating) and CD8⁺CD69⁺CXCR6⁺ (tissue-resident) T cell subsets in different tissues of control and infected female mice. The percentages are shown. n Representative histograms of CFSE expression showing antigen-specific proliferation of hepatic CD8⁺ T cells from control (Con) and infected (Cs) mice. CsAWA, C. sinensis adult worm antigen. m,n Data are representative of two independent experiments, n = 5 mice per group. Source data are provided as a Source Data file. Within the T cell compartment, 13 transcriptionally distinct subsets were identified (Fig. 2f–h). Th2 cells predominated during the acute phase, whereas CD8+ tissue-resident memory T cells (TRM), CD8+ effector memory T cells (TEM) and T helper type 1 cells (Th1) were more abundant in the chronic phase (Fig. 2h). Cell abundance enrichment analysis, along with correlation analysis between cell proportions and infection duration, revealed that TRM cells showed the strongest positive association with disease progression (Fig. 2i and Supplementary Fig. 4a). The AddModuleScore algorithm confirmed that CD8⁺ TRM clusters displayed the highest residency signature scores, distinguishing them from other CD8⁺ subsets (Fig. 2j, Supplementary Fig. 4b and Supplementary Table 1). TCR repertoire analysis revealed minimal clonotype overlap between CD8+ TRM and naive T cells (TN), indicating that TRM cells likely expand in situ within the liver rather than being replenished by circulating T cells (Fig. 2k). Consistently, TRM subsets were enriched for expanded clonotypes and exhibited reduced Shannon diversity relative to other T cell subsets, a pattern indicative of antigen-driven proliferation and a stable tissue-resident phenotype (Fig. 2l and Supplementary Fig. 4c). To experimentally validate these phenotypes, we assessed the expression of tissue-residency markers (CD69 and CXCR6) and the circulation-associated marker CD62L by flow cytometry. A distinct CD8⁺CD69⁺CXCR6⁺ TRM population accumulated in infected livers but was largely absent from blood and spleen (Fig. 2m and Supplementary Fig. 5). Similar phenotype was also observed in C. sinensis-infected rat models (Supplementary Fig. 6). In vitro assays further showed that hepatic CD8⁺CXCR6⁺ TRM proliferated in response to C. sinensis adult worm antigens (Fig. 2n). Together, these results demonstrate that CD8⁺ TRM progressively accumulate during chronic infection through local, antigen-driven proliferation, establishing them as tissue-resident cells with potential roles in helminth-induced liver immunopathology. Pathogenic functions of CD8⁺ TRM cells in promoting hepatic inflammation and fibrosis To elucidate the functional roles of specific immune cell subsets during chronic C. sinensis infection, we applied CellChat to systematically analyze intercellular communication networks. Among all hepatic immune populations, CD8⁺ TRM cells exhibited the strongest signaling activity, followed by CD8⁺ TEM cells (Fig. 3a and Supplementary Fig. 7a). Differential gene expression analysis identified 97 genes significantly upregulated in CD8+ TRM cells, including key chemokines (Ccl3, Ccl4, Ccl5) and granzymes (Gzmb, Gzmk) (Fig. 3b and Supplementary Fig. 7b and Supplementary Table 2). Consistent with these transcriptional profiles, hepatic NPCs from chronically infected mice secreted elevated levels of CCL3, CCL4, CCL5, and GZMB (Fig. 3c and Supplementary Fig. 7c). Flow cytometric analysis further confirmed that hepatic CD8⁺ T cells–particularly the CXCR6⁺ TRM subset–were the primary source of the increased expression of CCL4, GZMK, and GZMB (Fig. 3d, e).Fig. 3Functional features of CD8+ TRM cells and their association with liver pathology during chronic C. sinensis infection.BALB/c mice were infected by C. sinensis for 16 weeks (Cs). a–i Hepatic non-parenchymal cells (NPCs) were isolated. a Bar plot showing the total ligand–receptor interaction strengths inferred by CellChat. b Volcano plot of differentially expressed genes between CD8⁺ TRM cells and other T cells. P values were calculated using two-sided Wilcoxon rank-sum tests with Bonferroni correction. Significant upregulated genes: red, downregulated genes: blue (|log2FC | > 1, p.adj <0.05); top upregulated genes are labeled. c Concentrations of CCL3, CCL4, CCL5, and Granzyme B (GZMB) in 24-hour NPC culture supernatants (female, n = 7 mice per group). d,e Flow cytometry of NPCs from infected mice (n = 7 mice per group). Top: Representative histograms showing protein expression in total CD8⁺ T cells versus CD8⁻ NPCs (d), and in CD8⁺CXCR6⁻ versus CD8⁺CXCR6⁺ subsets (e). Bottom: Mean fluorescence intensity (MFI) of indicated markers, normalized to that in total NPCs. f Bar plot showing top Gene Ontology terms enriched in genes upregulated in CD8⁺ TRM cells versus other T cells. P values were calculated using hypergeometric tests with Benjamini-Hochberg correction. g Significantly enriched pathways across CD8⁺ TRM, TEM and TN subsets. Bar height indicates relative pathway activity. h Chord diagram of top five CD8⁺ TRM-mediated secretory signaling pathways, chord width indicates CellChat interaction strength. i Sankey diagram illustrating CD8⁺ TRM-mediated CCL signaling pathways and target populations. Ribbon width reflects CellChat interaction strength. j Representative multiplex immunohistochemistry images of liver tissues showing DAPI (blue), CD8⁺ T cells (red), Ly6G⁺ neutrophils (green), F4/80⁺ macrophages (cyan), and α-SMA⁺ fibroblasts (yellow) (female, n = 5 mice per group). Scale bars, 100μm. k Quantification of immune cell infiltration and fibrotic area (α-SMA⁺) in (j). l,m Spearman’s rank correlation between CD8⁺ T cells count and Ly6G⁺ neutrophils count, F4/80⁺ macrophages count, or α-SMA⁺ fibrotic area in (j) (n = 90 fields). Data are presented as mean ± SEM from two independent experiments. Statistical significance was assessed by two-tailed unpaired t or Mann-Whitney U test (c–e,k) and two-tailed Spearman’s rank correlation (l,m). P values are indicated in the corresponding panels. Source data are provided as a Source Data file. Functional features of CD8+ TRM cells and their association with liver pathology during chronic C. sinensis infection. BALB/c mice were infected by C. sinensis for 16 weeks (Cs). a–i Hepatic non-parenchymal cells (NPCs) were isolated. a Bar plot showing the total ligand–receptor interaction strengths inferred by CellChat. b Volcano plot of differentially expressed genes between CD8⁺ TRM cells and other T cells. P values were calculated using two-sided Wilcoxon rank-sum tests with Bonferroni correction. Significant upregulated genes: red, downregulated genes: blue (|log2FC | > 1, p.adj <0.05); top upregulated genes are labeled. c Concentrations of CCL3, CCL4, CCL5, and Granzyme B (GZMB) in 24-hour NPC culture supernatants (female, n = 7 mice per group). d,e Flow cytometry of NPCs from infected mice (n = 7 mice per group). Top: Representative histograms showing protein expression in total CD8⁺ T cells versus CD8⁻ NPCs (d), and in CD8⁺CXCR6⁻ versus CD8⁺CXCR6⁺ subsets (e). Bottom: Mean fluorescence intensity (MFI) of indicated markers, normalized to that in total NPCs. f Bar plot showing top Gene Ontology terms enriched in genes upregulated in CD8⁺ TRM cells versus other T cells. P values were calculated using hypergeometric tests with Benjamini-Hochberg correction. g Significantly enriched pathways across CD8⁺ TRM, TEM and TN subsets. Bar height indicates relative pathway activity. h Chord diagram of top five CD8⁺ TRM-mediated secretory signaling pathways, chord width indicates CellChat interaction strength. i Sankey diagram illustrating CD8⁺ TRM-mediated CCL signaling pathways and target populations. Ribbon width reflects CellChat interaction strength. j Representative multiplex immunohistochemistry images of liver tissues showing DAPI (blue), CD8⁺ T cells (red), Ly6G⁺ neutrophils (green), F4/80⁺ macrophages (cyan), and α-SMA⁺ fibroblasts (yellow) (female, n = 5 mice per group). Scale bars, 100μm. k Quantification of immune cell infiltration and fibrotic area (α-SMA⁺) in (j). l,m Spearman’s rank correlation between CD8⁺ T cells count and Ly6G⁺ neutrophils count, F4/80⁺ macrophages count, or α-SMA⁺ fibrotic area in (j) (n = 90 fields). Data are presented as mean ± SEM from two independent experiments. Statistical significance was assessed by two-tailed unpaired t or Mann-Whitney U test (c–e,k) and two-tailed Spearman’s rank correlation (l,m). P values are indicated in the corresponding panels. Source data are provided as a Source Data file. Gene Ontology enrichment analysis demonstrated that genes upregulated in TRM cells were enriched for cell chemotaxis, neutrophil migration, and cytokine activity (Fig. 3f). In contrast, CD8⁺ TEM cells were mainly associated with cytotoxicity and extravasation (Fig. 3g and Supplementary Fig. 7d). Intercellular communication analysis revealed five dominant pathways linked to TRM cells—CCL, MIF, GALECTIN, TNF, and IFN-II—with the CCL axis as the most prominent, targeting neutrophils, macrophages, Kupffer cells, and other immune subsets (Fig. 3h, i). Further analysis of crosstalk between TRM cells and stromal cells indicated that TRM cells may transmit signals to hepatic stromal cells either directly or indirectly, implying a potential role in promoting liver fibrosis (Supplementary Fig. 7e, f). Histological validation by multiplex immunohistochemistry revealed dense infiltration of CD8⁺ T cells and Ly6G⁺ neutrophils, along with marked activation of α-SMA⁺ fibroblasts within lesional areas (Fig. 3j, k and Supplementary Fig. 8a, b). Spatial mapping further demonstrated that CD8⁺ T cells were closely colocalized with neutrophils, macrophages, and activated fibroblasts (Fig. 3l, m and Supplementary Fig. 8c,d). Notably, a similar phenotypic profile was also observed in C. sinensis-infected rat models, consistent with our findings in mice (Supplementary Fig. 9a–d). These findings support a model in which CD8+ TRM cells act as local orchestrators of hepatic pathology, amplifying inflammation through chemokine-mediated immune recruitment and driving fibrosis via fibroblast activation. Phase-specific depletion of CD8⁺ T cells attenuates liver inflammation and fibrosis To directly assess the pathogenic role of CD8⁺ T cells, we performed antibody-mediated depletion using anti-CD8α antibodies during the chronic phase (16 weeks post-infection), a time point that closely models human disease progression (Fig. 4a). Intraperitoneal administration of anti-CD8α antibodies effectively depleted CD8⁺ T cells from both peripheral blood and spleen, confirming treatment efficacy without affecting fecal egg counts (Fig. 4b and Supplementary Fig. 10). However, compared to isotype IgG controls, anti-CD8α-treated mice exhibited significantly reduced liver pathology, as shown by histological assessment (Fig. 4c, d). Multiplex immunohistochemistry further demonstrated decreased neutrophil infiltration and diminished α-SMA⁺ fibroblast accumulation, indicating attenuated inflammation and fibrotic remodeling (Fig. 4e, f). Flow cytometric analysis of hepatic NPCs confirmed a substantial reduction in CD8⁺ T cells, accompanied by lower frequencies of neutrophils and eosinophils, while macrophage proportions increased (Fig. 4g).Fig. 4Phase-specific CD8⁺ T cell depletion ameliorates liver pathology in C. sinensis infection.a Schematic overview of the CD8⁺ T cell depletion protocol. BALB/c mice were orally infected with C. sinensis metacercariae, and at 16 weeks post infection (p.i.), treated intraperitoneally with either anti-CD8α antibody or isotype control on days 0, 3, 6, and 9 (n = 8 mice per group). i.p., intraperitoneal injection. Created in BioRender. Du, X. (2026) https://BioRender.com/xdjezfm. b Flow cytometric analysis showing frequencies of CD8+ T cells in peripheral blood and spleen after antibody treatment. c Representative liver sections stained with hematoxylin-eosin (H&E) and Sirius Red. Scale bars, 500 μm. d Quantification of histological activity index (HAI) scores and collagen deposition (percentage Sirius Red-positive area) from sections in (c). e Representative multiplex immunohistochemistry images of liver tissues showing DAPI (blue), CD8⁺ T cells (red), Ly6G⁺ neutrophils (green), F4/80⁺ macrophages (cyan), and α-SMA⁺ fibroblasts (yellow) (n = 6 mice per group). Scale bars, 100 μm. f Quantification of immune cell infiltration and fibrotic area (α-SMA⁺) in (e). g Relative abundance of liver immune cell subsets measured by flow cytometry, normalized to the isotype control group (n = 8 mice per group). h Transcript levels of Ccl3, Ccl4, Ccl5, Gzmb, Mmp2, Mmp7, Mmp8, Mmp12, Timp1 and Timp2 in hepatic NPCs, normalized to the isotype control group (n = 6 mice per group). i Concentrations of CCL3, CCL4, CCL5, Granzyme B (GZMB) and MMP-2 in liver homogenate supernatants (n = 8 mice per group). Data are presented as mean ± SEM from at least two independent experiments. Statistical significance was assessed by two-tailed unpaired t or Mann-Whitney U test (b,d,f–i). P values are indicated in the corresponding panels. Source data are provided as a Source Data file. Phase-specific CD8⁺ T cell depletion ameliorates liver pathology in C. sinensis infection. a Schematic overview of the CD8⁺ T cell depletion protocol. BALB/c mice were orally infected with C. sinensis metacercariae, and at 16 weeks post infection (p.i.), treated intraperitoneally with either anti-CD8α antibody or isotype control on days 0, 3, 6, and 9 (n = 8 mice per group). i.p., intraperitoneal injection. Created in BioRender. Du, X. (2026) https://BioRender.com/xdjezfm. b Flow cytometric analysis showing frequencies of CD8+ T cells in peripheral blood and spleen after antibody treatment. c Representative liver sections stained with hematoxylin-eosin (H&E) and Sirius Red. Scale bars, 500 μm. d Quantification of histological activity index (HAI) scores and collagen deposition (percentage Sirius Red-positive area) from sections in (c). e Representative multiplex immunohistochemistry images of liver tissues showing DAPI (blue), CD8⁺ T cells (red), Ly6G⁺ neutrophils (green), F4/80⁺ macrophages (cyan), and α-SMA⁺ fibroblasts (yellow) (n = 6 mice per group). Scale bars, 100 μm. f Quantification of immune cell infiltration and fibrotic area (α-SMA⁺) in (e). g Relative abundance of liver immune cell subsets measured by flow cytometry, normalized to the isotype control group (n = 8 mice per group). h Transcript levels of Ccl3, Ccl4, Ccl5, Gzmb, Mmp2, Mmp7, Mmp8, Mmp12, Timp1 and Timp2 in hepatic NPCs, normalized to the isotype control group (n = 6 mice per group). i Concentrations of CCL3, CCL4, CCL5, Granzyme B (GZMB) and MMP-2 in liver homogenate supernatants (n = 8 mice per group). Data are presented as mean ± SEM from at least two independent experiments. Statistical significance was assessed by two-tailed unpaired t or Mann-Whitney U test (b,d,f–i). P values are indicated in the corresponding panels. Source data are provided as a Source Data file. At the molecular level, hepatic NPCs from CD8+ T cell-depleted mice showed downregulation of key fibrogenic and inflammatory mediators, including the chemokines (Ccl3, Ccl4, Ccl5), the cytotoxic effector (Gzmb), and the fibrosis-associated inhibitors (Timp1, Timp2) at the transcriptional level. Conversely, expression of the matrix-degrading enzymes (Mmp2, Mmp8) was elevated (Fig. 4h, Supplementary Fig. 11). Consistent with these findings, liver tissue analysis confirmed reduced protein levels of CCL3, CCL4, and GZMB, alongside increased MMP-2 (Fig. 4i). Collectively, these findings provide direct evidence that CD8⁺ T cells—particularly the TRM subset—drive liver immunopathology in chronic C. sinensis infection by promoting immune cell recruitment and sustaining a pro-fibrotic microenvironment. Their depletion reshapes the hepatic milieu toward reduced inflammation and ameliorates the fibrotic microenvironment, highlighting the central pathogenic role of TRM cells. Circulating CCL4 as a TRM-derived biomarker correlating with hepatic complications in mice and patients Given the strong link between CD8+ TRM cell-mediated chemokine activity and progressive liver immunopathology, we explored the biomarker potential of these chemokines by mapping their spatiotemporal expression in both liver tissue and serum throughout the course of infection (Fig. 5a, b and Supplementary Fig. 12a, b). CCL3 and CCL4 levels were concurrently elevated in both the liver and peripheral blood, whereas CCL5 showed liver-specific upregulation. Jonckheere-Terpstra trend analysis confirmed a statistically significant monotonic increase in liver and serum levels of CCL3 and CCL4 over time. Notably, the temporal expression profiles of CCL4 in liver and serum were highly concordant (p = 0.1024, ρ = 0.8), suggesting that hepatic production is a major source of circulating CCL4 (Fig. 5c).Fig. 5Serum CCL4 levels correlate with hepatic complications during C. sinensis infection.a–c BALB/c mice were orally infected with C. sinensis metacercariae. Liver and serum samples were collected at 0, 2, 6, and 16 weeks post-infection (female, week 0, n = 12 mice; week 2, n = 6 mice; week 6 and 16, n = 10 mice). a Concentrations of CCL3, CCL4 and CCL5 in liver homogenate supernatants. b Corresponding chemokine concentrations in serum. c Paired analysis of CCL3, CCL4, and CCL5 levels in liver homogenate supernatants (left y-axis, pg/g) and serum (right y-axis, pg/ml) across time points. Jonckheere-Terpstra test (JT) assessed monotonic trends (JT < 0.05 regarded as significant); linear mixed-effects models tested time × tissue interaction (p < 0.05 regarded as different slopes); Spearman’s rank correlation (ρ) evaluated concordance between compartments. p, interaction P value; ρ, Spearman coefficient; L, liver; S, serum. d Serum CCL3, CCL4, and CCL5 levels in human subjects with (Cs+, n = 164 participants) or without C. sinensis infection (Cs-, n = 47 participants). e Serum CCL levels stratified by liver fibrosis stage in Cs+ individuals (F0 = 48, F1 = 31, F2 = 37, F3 = 48 participants) and compared with Cs- controls. f Correlation between liver fibrosis stage and serum chemokine concentrations in all Cs+ individuals (n = 164 participants). Data are presented as mean ± SEM from at least two independent experiments. Statistical significance was assessed by one-way ANOVA with Tukey’s multiple comparisons test or Kruskal-Wallis test with Dunn’s multiple comparisons test (a,b,e), two-tailed Mann-Whitney U test (d), and Spearman’s correlation (c,f). P values are indicated in the corresponding panels. Source data are provided as a Source Data file. Serum CCL4 levels correlate with hepatic complications during C. sinensis infection. a–c BALB/c mice were orally infected with C. sinensis metacercariae. Liver and serum samples were collected at 0, 2, 6, and 16 weeks post-infection (female, week 0, n = 12 mice; week 2, n = 6 mice; week 6 and 16, n = 10 mice). a Concentrations of CCL3, CCL4 and CCL5 in liver homogenate supernatants. b Corresponding chemokine concentrations in serum. c Paired analysis of CCL3, CCL4, and CCL5 levels in liver homogenate supernatants (left y-axis, pg/g) and serum (right y-axis, pg/ml) across time points. Jonckheere-Terpstra test (JT) assessed monotonic trends (JT < 0.05 regarded as significant); linear mixed-effects models tested time × tissue interaction (p < 0.05 regarded as different slopes); Spearman’s rank correlation (ρ) evaluated concordance between compartments. p, interaction P value; ρ, Spearman coefficient; L, liver; S, serum. d Serum CCL3, CCL4, and CCL5 levels in human subjects with (Cs+, n = 164 participants) or without C. sinensis infection (Cs-, n = 47 participants). e Serum CCL levels stratified by liver fibrosis stage in Cs+ individuals (F0 = 48, F1 = 31, F2 = 37, F3 = 48 participants) and compared with Cs- controls. f Correlation between liver fibrosis stage and serum chemokine concentrations in all Cs+ individuals (n = 164 participants). Data are presented as mean ± SEM from at least two independent experiments. Statistical significance was assessed by one-way ANOVA with Tukey’s multiple comparisons test or Kruskal-Wallis test with Dunn’s multiple comparisons test (a,b,e), two-tailed Mann-Whitney U test (d), and Spearman’s correlation (c,f). P values are indicated in the corresponding panels. Source data are provided as a Source Data file. To assess the clinical relevance of these findings, we measured serum CCL3, CCL4 and CCL5 levels in a cohort of 211 individuals from a C. sinensis-endemic region in southeastern China. C. sinensis-positive (Cs+) individuals exhibited significantly higher serum CCL4 levels compared to uninfected controls (Fig. 5d). Importantly, among Cs+ individuals, those with moderate-to-severe liver fibrosis (stages F2–F3) displayed markedly elevated serum CCL4 concentrations relative to those with mild or no fibrosis (Fig. 5e). Serum CCL4 levels positively correlated with fibrosis severity (Fig. 5f). To assess whether this association was confounded by clinical covariates, we fitted a multivariable ordinal logistic regression model with fibrosis stage as an ordered outcome. After adjustment for age, fatty liver status and bile duct dilatation, higher CCL4 remained associated with greater fibrosis severity (adjusted OR = 1.057 per 10-unit increase, 95% CI 1.018–1.097, p = 0.0041) (Supplementary Fig. 13). These findings highlight the potential of CCL4 as a noninvasive biomarker for monitoring hepatic fibrosis progression in chronic clonorchiasis. Discussion TRM cells are central mediators of localized immune responses, possessing both memory and effector functions as well as notable phenotypic and functional plasticity8–10. In this study, we identify CD8+ T cells–particularly the TRM subset–as key contributors to chronic liver injury during C. sinensis infection and propose them as potential therapeutic targets. This conclusion is supported by multiple lines of evidence: (a) during the chronic phase of infection, CD8+ TRM cells significantly expand and become the dominant lymphocyte subset among hepatic non-parenchymal cells; (b) cell-cell communication analysis reveals that CD8+ TRM cells exhibit higher levels of intercellular signaling compared to other immune subsets; (c) spatial mapping of liver tissue shows a strong correlation between CD8+ T cell localization and sites of pathological damage, including immune cell infiltration and α-SMA⁺ fibroblast activation; and (d) antibody-mediated depletion of CD8+ T cells at advanced stages of disease markedly reduces liver inflammation and remodels the fibrotic microenvironment. Together, these findings highlight CD8⁺ TRM cells as critical drivers of tissue-specific immunopathology in chronic helminth infection and underscore their therapeutic relevance in helminth-induced liver disease. TRM cells are distinguished for their transcriptional plasticity–preconditioned before tissue entry but highly adaptable to local microenvironments–allowing them to acquire specialized functions8,11,12. In our study, hepatic CD8+ TRM cells from chronically infected mice displayed a canonical tissue-resident signature. These cells expressed high levels of Itga1, Cxcr6, Zfp683, Xcl1, Cd101, Rgs1, Gpr18, Ccr5, Nr4a2 and Gzmk, while exhibiting low expression of S1pr1, Klf2, and Ccr7. This expression pattern is consistent with reduced egress capacity and retention within hepatic tissue, supporting their localized residency and functional specialization13–17. Functional profiling of hepatic CD8⁺ TRM cells in our study revealed robust activation of chemokine signaling pathways, highlighted by elevated expression of CCL3, CCL4, and CCL5. These findings suggest that CD8+ TRM cells function as key amplifiers of the local immune response, recruiting and activating other pathogenic immune subsets within the hepatic microenvironment. Notably, GZMK, a molecule closely associated with tissue residency, was also highly expressed in hepatic TRM cells18. Recent studies have linked GZMK to immune modulation via complement activation, pointing to a potential non-cytolytic role in shaping tissue-level immune dynamics19,20. In contrast to conventional TEM cells, which typically mediate cytotoxicity via the perforin–granzyme axis21,22, hepatic TRM cells in our study expressed high levels of Gzmb but relatively low levels of Prf1, indicating a potentially diminished capacity for direct target cell lysis. This suggests that TRM cells may contribute more to immune regulation and inflammatory amplification than to classical cytotoxic functions. Whether these distinct phenotypic and functional characteristics of CD8⁺ TRM cells are driven by parasite-derived factors or by cues from the altered hepatic immune microenvironment remains an important open question. In the current study, we assessed the pathogenic role of CD8⁺ T cells through systemic depletion using anti-CD8α antibodies. CD8⁺ T cell depletion resulted in a marked attenuation of fibrosis, characterized by reduced neutrophil infiltration and decreased numbers of α-SMA⁺ collagen-producing cells. Interestingly, this was accompanied by an increase in macrophage abundance. These findings suggest that depletion of CD8⁺ T cells may alter the hepatic immune microenvironment in a manner that favors collagen clearance over deposition. In particular, macrophages–known for their dual roles in fibrosis–may be functionally reprogrammed under these conditions toward a pro-resolving phenotype23,24, potentially through enhanced expression of matrix-degrading enzymes such as MMPs25,26. This shift could contribute to the observed reduction in fibrotic burden24. Further investigation is needed to clarify how CD8⁺ T cell depletion reshapes immune cell crosstalk and regulates the balance between matrix synthesis and degradation in fibrotic tissue. This study has certain limitations, primarily the lack of tools for TRM-specific depletion. Therefore, although our systemic CD8⁺ depletion data strongly implicate CD8⁺ T cells in driving pathology, the specific attribution to the TRM subset currently relies on correlative and spatial-functional evidence. The core issue is that no widely validated and exclusive surface markers are available to deplete TRM populations in vivo without affecting other CD8⁺ T cell subsets. To address this constraint as rigorously as possible, we used a timed depletion strategy administered during the period when TRM are thought to exert their most prominent pathogenic effects. Future studies to identify unique TRM markers will be essential to overcome this limitation and enable precise mechanistic studies. CCL4, a member of the CC chemokine family formerly known as macrophage inflammatory protein-1β (MIP-1β), exerts multifaceted effects on both immune and non-immune cells27–29. In this study, we identified CD8+ TRM cells as a predominant source of CCL4 within the liver during chronic C. sinensis infection. CCL4 levels were markedly elevated in both hepatic tissue and serum of infected mice, and these elevations closely correlated with the severity and progression of liver pathology. Consistent with these findings, human cohort data demonstrated significantly higher serum CCL4 concentrations in clonorchiasis patients than in healthy controls, with levels positively associated with fibrosis severity. These results highlight the potential of CCL4 as a prognostic biomarker for disease monitoring in chronic C. sinensis infection. Importantly, chronic C. sinensis infection is a well-established risk factor for hepatobiliary malignancies, particularly cholangiocarcinoma30. Increasing evidence has implicated CCL4 in tumorigenesis through its roles in immune modulation and the recruitment of immunosuppressive cell populations31,32. Our finding that hepatic CD8⁺ TRM cells represent a major source of CCL4 raises the possibility that persistent CCL4 production contributes to a pro-tumorigenic microenvironment, potentially linking chronic inflammation to malignant transformation—a mechanism that warrants further mechanistic investigation. Taken together, our findings delineate the immune landscape of chronic C. sinensis infection and highlight CD8⁺ TRM cells and chemokine signaling as key drivers of hepatic pathology. Beyond advancing our understanding of parasite-induced immune dysregulation, this study identifies TRM cells and CCL4 as promising therapeutic and diagnostic targets for mitigating liver injury and reducing the risk of helminth-associated malignancy. Methods Ethics statement Our research complies with all relevant ethical regulations. The study involving human samples was approved by the Ethics Committee of the National Institute of Parasitic Diseases, Chinese Center for Disease Control and Prevention in Shanghai, China (approval code: 2011-005), and written informed consent was obtained prior to participation. All animal care procedures were performed according to the National Institute of Health Guide for the Care and Use of Laboratory Animals. The animal study was reviewed and approved by the Animal Ethics Committee of Shanghai Jiao Tong University School of Medicine. All animal experiments were conducted in accordance with protocols approved by the Institutional Animal Care and Use Committee (IACUC) of Shanghai Jiao Tong University School of Medicine (approval code: JUMC2023-133-A). Mice and rats Wild-type BALB/c mice were purchased from Shanghai Lingchang Biotechnology Company Limited. Sprague Dawley (SD) rats were purchased from Shanghai Bikai Keyi Biotechnology Company Limited. All mice and rats were housed under specific pathogen-free conditions (SPF) at the Animal care facilities in the Shanghai Jiao Tong University School of Medicine and National Institute of Parasitic Diseases. Animals were maintained on a 12-hour light/dark cycle at a temperature of 20 °C–24 °C and relative humidity of 45–65%. Human sera Human serum samples were collected from people living in Hengxian County, Guangxi Province, China, a C. sinensis endemic area33. No compensation was provided to participants. A total of 211 participants were enrolled in this study, including 164 individuals diagnosed with C. sinensis infection (Cs+) and 47 uninfected controls (Cs-). Among the 164 Cs+ individuals, liver fibrosis staging was assessed34, yielding the following distribution: 48 cases with stage F0, 31 cases with F1, 37 cases with F2, and 48 cases with F3. Sex or gender was not considered as a variable in the study design, and no sex- or gender-based analysis was conducted. Detailed information of the participants was summarized in Supplementary Table 3. C. sinensis infection Six- to eight-week-old BALB/c mice were orally inoculated with 50 C. sinensis metacercariae via intragastric gavage35. This study used mostly female mice except for Supplementary Fig. 2, 5, 7, 8, 12. Female SD rats weighing 50 g were orally inoculated with 100 C. sinensis metacercariae via intragastric gavage36. Successful infection was confirmed by weekly quantitative fecal egg counts (expressed as eggs per gram of feces, EPG). Mice were euthanized at designated time points (0, 2, 6, and 16 weeks post-infection) for collection of blood, spleen, and liver tissues. Serum alanine aminotransferase (ALT), aspartate aminotransferase (AST), alkaline phosphatase (ALP), total bile acid (TBA), total protein (TP), and albumin/globulin (A/G) were measured using an AU5800 clinical chemistry analyzer (Beckman Coulter, USA). Rats were euthanized at 16 weeks post-infection, together with control rats, for collection of blood, spleen, and liver tissues. Histopathological analyses Liver tissues were collected at 0, 2, 6, and 16 weeks post C. sinensis infection and immediately fixed in 4% paraformaldehyde (PFA) for 48 h at 4 °C. After fixation, tissues were rinsed thoroughly with phosphate-buffered saline (PBS), dehydrated through a graded ethanol series, and embedded in paraffin. Serial sections (4 μm thick) were cut and stained with hematoxylin-eosin (H&E) and Sirius Red for histological evaluation. Microscopy was performed using a Nikon microscope (Nikon, JPN). For each sample, measurements were taken from at least three randomly selected fields, and the average of these values was used as the representative score for that sample. Liver inflammation was assessed using the histology activity index (HAI), scored from 0 to 18, based on periportal or periseptal interface hepatitis (0 to 4), confluent necrosis (0 to 6), focal lytic necrosis, apoptosis and focal inflammation (0 to 4) and portal inflammation (0 to 4)37,38. Collagen deposition was quantified from Sirius Red-stained sections using ImageJ software39. Enzyme-linked immunosorbent assay (ELISA) Antibody reactivity of mouse sera against C. sinensis adult worm antigen (AWA) was determined by ELISA. Briefly, 96-well plates were coated overnight at 4 °C with 100 μL/well of CsAWA (1 μg/mL). Plates were washed three times with PBST (0.3% Tween-20 in PBS), then blocked with 0.3% BSA in PBST for 1 h at 37 °C. After three additional washes, diluted serum samples were added and incubated for 1 h at 37 °C. Following another round of washing, horseradish peroxidase (HRP)-conjugated secondary antibodies were added and incubated for 1 h at 37 °C. The following antibodies were used at the indicated dilutions: HRP-conjugated goat anti-mouse IgG, IgM, IgA (1:5000, Invitrogen, A-10668), HRP-conjugated goat anti-mouse IgE (1:1000, Invitrogen, SA5-10263). After washing, 100 μL/well of TMB substrate (Beyotime, P0209) was added. Following color development, 50 μL/well stop solution (Beyotime, P0215) was added and absorbance was measured at 450 nm using a microplate reader (Biotek, USA). Level of granzyme B (GZMB) in cell culture supernatants and liver homogenate supernatants were quantified using a commercial ELISA kit (Invitrogen, BMS6029) according to the manufacturer’s instructions. Level of MMP-2 in liver homogenate supernatants was quantified using a commercial ELISA kit (R&D, MMP200) according to the manufacturer’s instructions. Flow cytometry Cells were first incubated with fixable viability dye eFluor 780 (1:1000, eBioscience, 65-0865-18) for 30 min at 4 °C to discriminate live/dead cells. For surface staining, cells were washed with FACS buffer (0.5% BSA in PBS), blocked with normal mouse serum for 10 min at 4 °C, and then incubated with surface staining antibodies for 30 min at 4 °C. After staining, the cells were washed and fixed in 1% paraformaldehyde (PFA) in FACS buffer. For intracellular staining, followed by surface staining as described above, cells were fixed and permeabilized with Fixation/Permeabilization solution (BD Biosciences, 554722) for 10 min at 4 °C, washed with Perm/Wash buffer (BD Biosciences, 554723), and incubated with intracellular staining antibodies for 30 min at 4 °C. Flow cytometric data were acquired using a LSRFortessa X20 flow cytometer (BD Biosciences, USA) and analyzed with FlowJo software. Antibodies were purchased from BioLegend (BL) or eBioscience (eBio) and used at 1:100 dilution unless otherwise specified: Lineage and surface markers CD11b FITC (eBio, 11-0112-85, M1/70), CD11b Brilliant Violet (BV) 711 (BL, 101242, M1/70), CD117 Pacific Blue (BL, 105819, 2B8), CD170 (Siglec-F) BV711 (BL, 155537, S17007L), CD186 (CXCR6) PerCP-Cyanine5.5 (BL, 151121, SA051D1), CD3 BV510 (BL, 100233, 17A2), CD4 FITC (BL, 201505, W3/25), CD4 APC (eBio, 17-0042-82, RM4-5), CD45 FITC (eBio, 11-0451-82, 30-F11), CD45 Pacific Blue (BL, 202225, OX-1), CD45 Super Bright 600 (eBio, 63-0451-82, 30-F11), CD45R/B220 BV421 (BL, 103239, RA3-6B2), CD49b PerCP-Cyanine5.5 (BL, 103519, HMα2), CD62L FITC (eBio, 11-0621-81, MEL-14), CD62L PerCP (BL, 202914, OX-85), CD69 BV711 (BL, 104537, H1.2F3), CD69 (abcam, ab307081, EPR25398-81) with secondary antibody goat anti-rabbit IgG Alexa Fluor 647 antibody (CST, 4414), CD8a BV605 (BL, 100744, 53-6.7), CD8a PE (BL, 201705, OX-8), FcεRIα PE (BL, 134307, MAR-1), F4/80 APC (eBio, 17-4801-82, BM8), Ly-6C PerCP-Cyanine5.5 (BL, 128011, HK1.4), Ly-6G BV510 (BL, 127633, 1A8). Intracellular markers Granzyme B PE-Cyanine7 (eBio, 25-8898-82, NGZB), Perforin BV421 (BL, 154319, S16009A). Multiplex immunohistochemistry analysis Liver tissues were collected at 0, 2, 6, and 16 weeks post C. sinensis infection and immediately fixed in 4% PFA for 48 h, washed with PBS, embedded in paraffin and sectioned at 4 μm thicknesses. Then sections were baked at 60 °C for 2 h to enhance tissue adherence. Slides were deparaffinized in xylene (2 changes) and rehydrated through a graded ethanol series (100%, 95%, 80%, and 75%), followed by rinsing in distilled water. Antigen retrieval was performed by microwave heating for 10 min in either sodium citrate buffer (pH 6.0) or Tris-EDTA buffer (pH 9.0), depending on the antigen-specific requirements. Endogenous peroxidase activity was quenched with 3% hydrogen peroxide (H2O2) for 20 min at room temperature. For intracellular antigens, membrane permeabilization was carried out using 0.2% Triton X-100 in TBS for 10 min. Sections were blocked in 5% bovine serum albumin (BSA) for 1 h at room temperature, and then sequentially incubated with primary antibodies. After washing with TBST, slides were incubated with HRP-conjugated secondary antibodies, followed by signal development using tyramide signal amplification (TSA) according to the manufacturer’s instructions for the PDOne six-color TSA-RM-82758 (Panovue, 10234100050). Microwave treatment was applied between each TSA labeling cycle. Nuclei were counterstained with DAPI Fluoromount-G™ (Yeasen, 36308ES20) after all antigens were labeled. Stained slides were scanned using an Olympus VS200 (Olympus, DEU). Quantification was performed using ImageJ software; for each sample at least three randomly selected fields were analyzed, and the mean value was used as the representative measurement for that sample. The antibodies used were as follows: CD8 alpha (1:1000, Abcam, ab217344), CD8 alpha (1:500, Abcam, ab237709), CD4 (1:500, Abcam, ab183685), CD11b (1:4000, Abcam, ab133357), B220 (1:200, ThermoFisher, 14-0452-82), alpha smooth muscle Actin (1:100, Abcam, ab5694), alpha smooth muscle Actin (1:300, CST, 19245), Ly6G (1:200, CST, 87048), Myeloperoxidase (1:100, Abcam, ab300650) with secondary antibody goat anti-rat IgG Alexa Fluor 647 (1:500, Abcam, ab150167), F4/80 (1:200, CST, 70076) and F4/80 (1:8000, ThermoFisher, 28463-1-AP). Liver non-parenchymal cell isolation Liver dissociation was carried out using the Liver Dissociation Kit (Miltenyi Biotec, 130-105-807) according to the manufacturer’s instructions. The resulting liver single-cell suspension was further enriched for non-parenchymal cells (NPCs) via density gradient centrifugation40,41. Briefly, the cells were resuspended in 30% Histodenz (Sigma-Aldrich, D2158) prepared in HBSS, and an overlay of HBSS was gently added on top. The gradient was centrifuged at 1400 × g for 20 min at 4 °C without braking. Liver NPCs were collected at the interface between the Histodenz and HBSS layers and used for downstream applications. Quantification of cytokine and chemokine concentrations Cytokine and chemokine levels in liver homogenate supernatants, cell culture supernatants, and serum samples were measured using LEGENDplex™ flow cytometry-based multiplex immunoassays (BioLegend) following the manufacturer’s protocol. Briefly, 25 μl of diluted sample or standard was combined with 25 μl of assay buffer or Matrix C, and 25 μl of mixed capture beads. The mixture was incubated for 2 h at room temperature with gentle shaking in the dark. Following a single wash step, 25 μl of detection antibodies was added and incubated for 1 h at room temperature with shaking. Subsequently, 25 μl of streptavidin-phycoerythrin (SA-PE) was added directly to each well and incubated for another 30 min under the same conditions. After a final wash, samples were resuspended in 200 μl wash buffer and acquired using a BD LSRFortessa X20 flow cytometer. Data were analyzed using LEGENDplex™ Data Analysis Software and quantified based on standard curves generated from known concentrations of each analyte. Single-cell sequencing Isolated liver non-parenchymal cells (NPCs) were processed using the Chromium Single Cell 5’ Kit v2 (10x Genomics) according to the manufacturer’s protocol. Paired-end sequencing of the resulting libraries was performed on an Illumina NovaSeq 6000 system. Raw sequencing data were processed using Cell Ranger (v6.0.1). Reads were demultiplexed, aligned to the mouse reference genome (mm10, 10x Genomics, v3.0.0) and quantified to generate a gene expression matrix. ScRNA-seq data quality control, processing, annotation, and visualization Raw gene barcode matrices generated by Cell Ranger (v6.0.1) were imported into R (v4.3.2) and converted into Seurat (v5.3.0) objects for downstream analysis42. Quality control was performed by retaining cells with 200 to 4000 detected genes (nFeatures), except for one data set (the second repetition of the 6-week group) where the upper threshold was adjusted to 3000. Cells with >10% mitochondrial gene expression were excluded. Ambient RNA contamination was corrected with SoupX (v1.6.2) with default parameters43. Putative multiplets were identified using DoubletFinder (v2.0.6). The optimal pK was determined via parameter sweep on the first 10 principal components (PCs), and the expected doublet rate was set at 6%. Barcodes predicted as doublet were excluded from further analysis44. Gene expression data from all libraries were merged using Seurat JoinLayers function. Each dataset underwent log-normalization (scaling factor = 10,000), followed by cell cycle regression (S and G2/M scores), and selection of the top 2000 highly variable genes using the vst method. Principal component analysis (PCA) was performed on the scaled data matrix, and Harmony (v1.2.3) was used to correct batch effects with sample identity as the covariate45. A shared nearest neighbor (SNN) graph was constructed on the Harmony-corrected embeddings. Clustering was performed using the Louvain algorithm at multiple resolutions (0.1, 1.5, 2 and 3). Low-quality and mixed clusters were excluded to refine the immune cell populations. Dimensionality reduction and visualization were performed using UMAP. For downstream analyses, fixed random seeds were used for computational steps involving stochastic procedures to standardize the analytical workflow. Cluster annotation was guided by canonical marker expression and generated using Seurat’s FindAllMarkers function (test.use = “wilcox”, only.pos = TRUE, min.pct = 0.25, logfc.threshold = 0.25). To generate balanced grouped UMAP visualizations, Seurat’s WhichCells function was used to randomly downsample each group to the smallest group size. This step only affected point density on plots–UMAP coordinates, and all downstream analyses were performed on the full dataset. Following the initial clustering, T cell subsets were further analyzed through iterative reclustering using the same workflow at higher resolutions (2, 3). Grouped UMAP visualizations of T cells were also downsampled for consistency, while all calculations were based on the complete dataset. Stage-preference analysis Differential state analysis was performed using the quasi-likelihood framework implemented in edgeR (v4.0.16) in combination with Muscat (v1.16.0), applied to pseudobulk expression counts46. Three pairwise contrasts were analyzed: healthy control (HC) contrasted with C. sinensis acute phase, HC contrasted with C. sinensis chronic phase, and acute contrasted with chronic phase. Genes were considered stage-responsive if the false discovery rate (FDR) < 0.05 and absolute log2fold change (log2FC) > 0.25. For each cluster, the union of all calculated stage-responsive genes was calculated and defined as the stage-dependent gene load. To quantify stage preference in T cell subsets, we calculated the ratio of observed to expected cells (Ro/e) for each cluster at each infection stage, providing a normalized measure of enrichment or depletion across time points. Gene-set enrichment and signature scoring For gene set enrichment analysis (GSEA), counts for immune cells were aggregated per sample within each condition using the AggregateExpression function to generate a pseudobulk expression matrix, where each column represented a biological replicate. Differential expression between the infected group (acute + chronic phase) and the uninfected group was assessed using a quasi-likelihood generalized linear model implemented in edgeR (v4.0.16). Genes were ranked by log2FC and analyzed using the gseGO function from clusterProfiler (v4.10.1) with annotations from org.Mm.eg.db (v3.18.0). Pathways with normalized enrichment score (NES) > 1 and adjusted P value (p.adj) <0.05 were considered significant. For Gene Ontology (GO) over-representation analysis, differentially expressed genes (DEGs) were identified using the Wilcoxon rank-sum test (FindMarkers, min.pct = 0.1). DEGs associated with red blood cell, mitochondrial, and ribosome-related genes were manually excluded. Over-representation analysis was then performed on DEGs with log2FC > 1 and FDR < 0.05 using enrichGO (biological process (BP) and molecular function (MF) categories). The top six terms per category were ranked by –log10(p.adj) and gene count. Mouse gene annotations were sourced from org.Mm.eg.db (v3.18.0). For cell-type-specific signature scoring, gene sets of interest were evaluated using the AddModuleScore function with default parameters47–50. For each sample-celltype combination, these scores were summarized by their median and compared with the CD8+ TRM using pairwise two-tailed Wilcoxon rank-sum tests, and the P values were corrected for multiple comparisons with the Benjamini-Hochberg, scores were displayed as violin plots. TCR sequence assembly TCR sequencing reads were processed using Cell Ranger (v6.0.1) provided by 10x Genomics. Reads were aligned to the mouse reference genome (vdj_GRCm38_alts_ensembl-3.1.0.gz-3.1.0) and assembled into TCR sequences. Preliminary TCR sequences were filtered to retain only high-confidence, full-length, productive sequences that were assigned valid cell barcodes and unambiguous TCR chain types. After filtering, 91.31% (9068 / 9931) of T cells were successfully assigned a paired TCR sequence. A total of 7939 clonotypes were identified, with a median of 2865 clonotypes per group (range 1708–3394). Clonotype sizes ranged from 1 to 47 cells, with 528 clonotypes found in more than one cell, and 178 clonotypes detected in more than two cells. TCR clonal analysis T cell receptor (TCR) repertoire analysis was performed using the scRepertoire (v1.12.0)51. Filtered filtered_contig_annotations.csv files were imported, and TCR libraries were merged using the combineTCR function. To ensure compatibility with Seurat object barcodes, barcode suffixes were remapped to sample-specific prefixes, following the approach recommended by the package authors. Clonotype metadata were then integrated into the Seurat T cell subsets with combineExpression. Clonotypes were defined as a unique paired CDR3α + CDR3β amino acid sequences. Clonotypes detected in 2 or more cells were classified as expanded, while those found in only one cell were designated as singletons. Clonal diversity was assessed using the Shannon index, calculated via clonalDiversity with 100 bootstrap replicates (default settings). Clone expansion distributions were computed using dplyr (v1.1.4), and shared clonotypes across T cell subsets were visualized as an area-proportional Euler diagram using the eulerr (v7.0.2). Cell-cell communication analysis Intercellular signaling was inferred using CellChat (v1.6.1), based on the package’s curated mouse ligand-receptor interaction database and default probabilistic modeling framework52,53. The total communication strength for each cell cluster was calculated by summing the probabilities of all ligand–receptor interactions, excluding autocrine loops. Visualization of signaling networks was performed using CellChat’s built-in functions. Key pathway interactions were illustrated using chord diagrams generated with netVisual_chord_gene and circle plots generated with netVisual_circle. Ligand-specific signaling flows were visualized as Sankey diagrams using the ggalluvial (v0.12.5), with CellChat-derived interaction probabilities used to scale edge widths. CD8+ T cell proliferation Liver non-parenchymal cells (NPCs) were isolated from control and C. sinensis infected mice and stained with 5 µM CFSE (Invitrogen, 65-0850-84) in PBS containing 0.1% BSA for 5 min at room temperature. After staining, cells were washed twice with 0.1% BSA in PBS and resuspended in complete cell culture medium at a concentration of 2.5 × 106 cells/mL. CFSE-stained NPCs were plated in 96-well plates (200 µl per well) and stimulated with 10 µg/mL C. sinensis adult worm antigen (AWA) for 3 days. Following stimulation, cell proliferation was assessed by flow cytometry based on CFSE dilution. CD8+ T cell function analysis Non-parenchymal cells (NPCs) were isolated from the livers of control and C. sinensis infected mice and resuspended at a concentration of 5 × 106 cells/mL in culture medium containing 5 µg/mL Brefeldin A (eBioscience, 00-4506-51). Cells were plated in 48-well plates (500 µl per well) and incubated for 4 h at 37 °C. For perforin staining, NPCs were additionally stimulated with phorbol 12-myristate 13-acetate (PMA) (Sigma-Aldrich, P1585) and ionomycin (Sigma-Aldrich, I3909) during the incubation period. For CCL4 and Granzyme K staining, cells were stained using a primary goat anti-mouse CCL4 antibody (R&D, AF-451-SP) and a rabbit anti-mouse Granzyme K antibody (Invitrogen, PA5-50980). Secondary staining was performed using donkey anti-goat IgG Alexa Fluor 488 antibody (abcam, ab150129) and goat anti-rabbit IgG Alexa Fluor 555 antibody (CST, 4413). Intracellular staining was performed as described above. CD8+ T cell depletion At 16 weeks after C. sinensis infection, mice were treated with 200 μg/mouse anti-CD8α antibody (BioXCell, BE0061) or isotype control antibody (BioXCell, BE0090), diluted in PBS. The mice were grouped based on similar levels of infection according to fecal egg counts measurement. Antibodies were administered via intraperitoneal (i.p.) injection on days 0, 3, 6 and 9. Blood, spleen and liver samples were collected on day 10 for downstream analyses. RNA isolation and real-time quantitative PCR Total RNA was extracted from mice hepatic NPCs using the High Pure RNA Isolation Kit (Roche, 11828665001) in accordance with the manufacturer’s instructions. Prior to cDNA synthesis, RNA was treated with DNase I (Thermo Scientific, EN0521) to remove trace amounts of DNA. Mouse complementary DNA (cDNA) was reverse-transcribed with RevertAid First Strand cDNA Synthesis Kit (Thermo Scientific, K1622) and real-time quantitative PCR (qPCR) was performed on a 7500 Fast Real-Time PCR System (Applied Biosystems, USA) using Hieff qPCR SYBR Green Master Mix (YEASEN, 11202ES08). The qPCR reaction was started at 95 °C for 10 minutes, followed by a two-step program with 40 cycles of 95 °C for 15 seconds and 60 °C for 60 seconds. After amplification, a melt curve analysis was performed, and specific peaks were shown. All primers’ amplification efficiency (E) were >94% with an R² value of 0.987–0.999 (Supplementary Fig. 11a). Relative gene expression was normalized to the geometric mean of genes Gapdh, Hprt1 and 18S rRNA1, which are stably expressed in the experiment (Supplementary Fig. 11b). The calculation method is using the 2−ΔΔCt method54. All PCR reactions were performed in duplicate, using their mean as the data point. Primer sequences are synthesized by Beijing Tsingke Biotechnology Company Limited and listed below (FW: Forward, RV: Reverse): Ccl3 FW: 5′-TGTACCATGACACTCTGCAAC-3′; Ccl3 RV: 5′-CAACGATGAATTGGCGTGGAA-3′; Ccl4 FW: 5′-TTCCTGCTGTTTCTCTTACACCT-3′; Ccl4 RV: 5′-CTGTCTGCCTCTTTTGGTCAG-3′; Ccl5 FW: 5′-GCTGCTTTGCCTACCTCTCC-3′; Ccl5 RV: 5′-TCGAGTGACAAACACGACTGC-3′; Gzmb FW: 5′-CTGCTACTGCTGACCTTGTCTC-3′; Gzmb RV: 5′-CTTGCTGGGTCTTCTCCTGTTC-3′; Mmp2 FW: 5′-CCAGATCACATACAGGATCATT-3′; Mmp2 RV: 5′-CCATCATGGATTCGAGAAAA-3′; Mmp7 FW: 5′-AACACTCTAGGTCATGCCTTCGCA-3′; Mmp7 RV: 5′-AGACCCAGAGAGTGGCCAAA-3′; Mmp8 FW: 5′-CCAGCACCTATTCACTACCTC-3′; Mmp8 RV: 5′-AGCATCAAATCTCAGGTGGG-3′; Mmp12 FW: 5′-CTGCCTCATCAAAATGTGCATC-3′; Mmp12 RV: 5′-ATTTGGAGCTCACGGAGACTT-3′; Timp1 FW: 5′-ACAAGTCCCAGAACCGCAGTGA-3′; Timp1 RV: 5′-CCGTCCACAAACAGTGAGTGTCA-3′; Timp2 FW: 5′-GAGCCACCGCAGTGAGCGAGAA-3′; Timp2 RV: 5′-GGGGAGGAGATGTAGCAAGGGA-3′; Gapdh FW: 5’-CTGAGCAAGAGAGGCCCTATCC-3’; Gapdh RV: 5’-CTCCCTAGGCCCCTCCTGTT-3’; Hprt1 FW: 5’-CAGACTTTGTTGGATTTGAAA-3’; Hprt1 RV: 5’-GCTCATCTTAGGCTTTGTAT-3’; 18S rRNA1 FW: 5’-GTAACCCGTTGAACCCCATT-3’; 18S rRNA1 RV: 5’-CCATCCAATCGGTAGTAGCG-3’. Statistical analyses Data are presented as mean ± standard error of the mean (SEM). Statistical analyses were performed using GraphPad Prism software. Depending on the experimental design, comparisons between groups were conducted using a two-tailed unpaired t test, two-tailed Mann-Whitney U test, one-way ANOVA with Tukey’s multiple comparisons test, and Kruskal-Wallis test with Dunn’s multiple comparisons test. Spearman’s rank correlation was used to assess associations between continuous variables. A P value < 0.05 was considered statistically significant. For correlation and trend analyses, temporal monotonic trends were evaluated using the Jonckheere-Terpstra trend test. Compartment effects and week-by-compartment interactions were analyzed using a linear mixed-effects model. Concordance between serum and liver chemokine concentrations was assessed using Spearman’s rank correlation coefficient. Fibrosis stage was modeled as an ordered outcome in a proportional-odds ordinal logistic regression model including CCL4, age, fatty liver and bile duct dilation as covariates. Adjusted odds ratios, 95% confidence intervals and two-sided Wald P values were estimated for each 10-unit increase in CCL4. Reporting summary Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article. Supplementary information Supplementary information Reporting Summary Transparent Peer Review file Supplementary information Reporting Summary Transparent Peer Review file Source data Source Data Source Data