Gut-derived memory γδ T17 cells exacerbate sepsis-induced acute lung injury in mice

Abstract

Sepsis is a critical global health concern linked to high mortality rates, often due to acute lung injury (ALI)/acute respiratory distress syndrome (ARDS). While the gut-lung axis involvement in ALI is recognized, direct migration of gut immune cells to the lung remains unclear. Our study reveals sepsis-induced migration of γδ T17 cells from the small intestine to the lung, triggering an IL-17A-dominated inflammatory response in mice. Wnt signaling activation in alveolar macrophages drives CCL1 upregulation, facilitating γδ T17 cell migration. CD44⁺ Ly6C^– IL-7R^high CD8^low cells are the primary migratory subtype exacerbating ALI. Esketamine attenuates ALI by inhibiting pulmonary Wnt/β-catenin signaling-mediated migration. This work underscores the pivotal role of direct gut-to-lung memory γδ T17 cell migration in septic ALI and clarifies the importance of localized IL-17A elevation in the lung.

Introduction

Sepsis, a grave malady imperiling global health, casts its shadow with elevated morbidity and mortality rates. The hyperactivated inflammatory response induced by sepsis can precipitate multifaceted organ dysfunction¹. Notably, the lung stands among the most susceptible organs to infection and septic insults. Approximately 45% of patients who experience sepsis will progress to develop Acute Respiratory Distress Syndrome (ARDS), characterized by disruption of the pulmonary endothelial barrier and diffuse lung injury². The excessive inflammatory response triggered by substantial immune cell infiltration in the lung during the early stages of sepsis is closely associated with the pathogenesis of ARDS³.

Recent research has increasingly highlighted the concept of “gut-origin sepsis”, underscoring the intestines’ pivotal role in driving multi-organ dysfunction during sepsis through intricate organ crosstalk mechanisms. Interactions between the gut and other organs, facilitated by critical conduits such as the mesenteric lymphatic system and portal venous system, contribute to the translocation of intestinal microbiota, metabolic products, and various mediators like pathogen-associated molecular patterns (PAMPs) and damage-associated molecular patterns (DAMPs) into the bloodstream^4,5. Despite mounting evidence linking perturbations in the intestinal microenvironment following sepsis to disorders like gut-origin liver damage, brain injury, and acute lung injury (ALI)/ARDS, mediated by intestinal microbiota and metabolism^6,7, the role of intestinal immune cells in mediating gut-origin ALI/ARDS remains poorly understood.

We previously demonstrated the migration of small intestinal γδ T cells to the lung exacerbating gut-origin lung injury following ischemic brain injury, highlighting a potential mechanism for gut-lung crosstalk⁸. However, whether sepsis induces similar migration of intestinal γδ T cells to the lung and the associated mechanisms remain unclear. γδ T cells, predominantly residing in the lamina propria of the small intestine, exhibit a dual role in inflammatory diseases. Within the inflamed milieu, resulting from disruptions in the gut epithelial barrier and bacterial infiltration, resident γδ T cells in the gut undergo differentiation, primarily manifesting as IL-17-producing inflammatory γδ T cells (referred to as γδT17 cells), which contribute to inflammation through dysregulated IL-17 expression and subsequent downstream pro-inflammatory factor production^3,9,10.

Here, we show that esketamine (S-KT), known for its immunomodulatory effects in psychiatric disorders¹¹, presents a potential therapeutic avenue for sepsis-induced ALI. In the mice model of polymicrobial sepsis induced by cecal ligation and puncture (CLP), we found a remarkable amelioration of sepsis-induced lung injury with the treatment of S-KT. Further exploration of the mechanism found that the migration of IL-7R^high CD8^low Ly6C^low memory γδ T17 cells from the small intestine to the lung following sepsis is a crucial factor contributing to the excessive inflammatory response and lung injury in sepsis. Moreover, the activation of the Wnt/β-catenin signaling following sepsis, and the subsequent transcriptional regulation of C-C motif chemokine ligand (CCL1) by the transcription factor lymphoid enhancer-binding factor 1 (Lef1) in the alveolar macrophages (AMs) mediated the migration of memory γδ T17 cells from the small intestine to the lung. Our results indicate that S-KT could mitigate sepsis-induced lung injury by suppressing the pulmonary Wnt/β-catenin signaling, thereby inhibiting the migration of small intestinal γδ T17 cells into the lung.

Results

Impacts of esketamine on septic ALI

Recently, esketamine (S-KT), an S-enantiomer of ketamine approved by the Food and Drug Administration (FDA) in the United States for treating depression, has garnered significant attention due to its anti-inflammatory action^11,12. Evidence indicates a close association between the antidepressant action of S-KT and its anti-inflammatory properties¹¹. However, its role in septic ALI remains unclear. We intend to elucidate the mechanisms and devise innovative therapeutic approaches for septic ALI through investigating the roles and mechanisms of S-KT in ALI triggered by cecal ligation and puncture (CLP)-induced polymicrobial sepsis. Surprisingly, our findings revealed a substantial enhancement in the 21-day survival rate, elevating it from 12.3% to 61.0%, following the combined prophylactic and therapeutic administration of S-KT (Fig. 1a). Furthermore, the debilitated condition of mice at both day 1 and day 3 after CLP was notably mitigated with S-KT treatment (Fig. 1b, c), and a reduction in the decline in rectal temperature at 24 h post-CLP was observed (Fig. 1d). Moreover, S-KT demonstrated its efficacy in alleviating sepsis-induced lung congestion, edema, and inflammatory cell infiltration (Fig. 1e, f). This effect was complemented by decreased lung wet-to-dry ratio (Fig. 1g), reduced vascular permeability (Fig. 1h), decreased bacterial load of the lung (Fig. 1i), and increased expression of tight junction proteins occludin and zonula occluden-1 (ZO-1) in the lung of septic mice (Supplementary Fig. 1a, b).

Fig. 1: Esketamine exerts protective effects on septic ALI.

a Kaplan–Meier analysis of survival in septic mice after prophylactic and/or therapeutic administration of esketamine (S-KT; 15 mg/kg) or 0.9% saline (n = 27 for saline, 25 for S-KT-Combined administration, 23 for S-KT-Pre-administration, 24 for S-KT-Post-administration). b Body weight changes (n = 26), c murine sepsis score (n = 10), and d core body temperature changes (n = 10) in septic mice after prophylactic and therapeutic administration of S-KT or 0.9% saline. e, f Representative H&E sections of lung and histological injury scores (n = 7), g lung W/D weight ratio (n = 19 for sham + saline at 12 h, 10 for sham + saline at 24 h, 18 for CLP + saline at 12 h, 13 for CLP + saline at 24 h, 27 for CLP + S-KT at 12 h, 16 for CLP + S-KT at 24 h), and h pulmonary capillary leakage determined with Evans blue dye (n = 7 for CLP + saline, 8 for sham + saline and CLP + S-KT) in mice after combined prophylactic and therapeutic administration of S-KT or 0.9% saline. 20×, scale bar = 100 μm. 40×, scale bar = 50 μm. i Bacterial loads in the murine lung after combined prophylactic and therapeutic administration of S-KT or 0.9% saline (n = 6 for sham + saline, 7 for CLP + saline, 5 for CLP + S-KT). Data are shown as mean ± SD. ^*P < 0.05, ^**P < 0.01, ^***P < 0.001, ^****P < 0.0001. ns not significant. P values were calculated using the Log-rank (Mantel–Cox) test (a), two-way ANOVA (b), two-sided Student’s unpaired t-test (c), two-tailed Mann–Whitney test (d), and one-way ANOVA followed by Tukey’s multiple-comparison test (f–i). Source data and exact P values are provided as a Source Data file.

Next, we aimed to gain in-depth insight into the mechanism underlying the protective mechanisms of S-KT in septic ALI, proteomic sequencing on lung tissues was performed to gain a comprehensive understanding of protein expression changes (Fig. 2a). Our proteomics analysis revealed significant differential expression of key members of the IL-17 signaling, such as matrix metallopeptidase (MMP)-3, MMP-9, prostaglandin-endoperoxide synthase 2 (Ptgs2), and TNF-alpha-induced protein 3 (TNFAIP3), between the septic mice and sham-operated mice (Fig. 2b, c). Consistently, a significant increase in inflammatory cytokines (IL-17A, IL-6, and IL-1β) was found in the lung at 24 h post-CLP, which was suppressed by S-KT treatment (Fig. 2d). This finding was further supported by immunofluorescence staining (Fig. 2e, f). Since IL-17A can stimulate the production of other pro-inflammatory mediators (IL-6, TNF-α, IL-1β, IL-8, granulocyte colony-stimulating factor, monocyte chemoattractant protein-1, macrophage inflammatory protein-2, CCL2, CXCL10 and CCL20)^3,13, and synergistically promote neutrophil recruitment and tissue infiltration¹⁴, we further assessed neutrophil infiltration in the lung by measuring myeloperoxidase (MPO) activity in the lung homogenates. Notably, a significantly higher MPO activity in the lung of the septic mice was observed, which was inhibited by S-KT treatment (Fig. 2g), indicating S-KT could alleviate substantial neutrophil infiltration in the lung during the acute phase of sepsis. Furthermore, S-KT also significantly reduced the levels of inflammatory cytokines (IL-6, TNF-α, and IL-17A) in the serum of septic mice (Fig. 2h). Our findings suggest that the IL-17 signaling in the lung is significantly activated following sepsis, and S-KT has anti-inflammatory effects associated with downregulation of IL-17A in the lung.

Fig. 2: Esketamine suppresses the initial surge of pulmonary IL-17A and mitigates the inflammatory responses in sepsis.

a The differential protein heatmap between sham-operated and CLP mice, in which each row represents a sample and each column represents a protein (n = 3). b Volcano plot of differential expression protein and c KEGG differential pathway enrichment bubble diagram in CLP mice compared to that in sham-operated mice (n = 3). d The protein levels of IL-17A, IL-6, and IL-1β in the lung (n = 6 for IL-17A and IL-1β, 9 for IL-6), e, f immunofluorescence staining and semi-quantitative analysis of IL-17A (n = 6). 40×, Scale bar = 20 μm. 80×, Scale bar = 10 μm. g pulmonary MPO activity (n = 9 for CLP + saline and CLP + S-KT, 10 for sham + saline), and h the serum levels of inflammatory cytokines IL-17A, IL-6 and TNF-α (n = 9 for sham + saline, 16 for CLP + saline, 17 for IL-6 in CLP + S-KT, 18 for IL-17A and TNF-α in CLP + S-KT) in mice after combined prophylactic and therapeutic administration of S-KT or 0.9% saline. Data are shown as mean ± SD. ^*P < 0.05, ^**P < 0.01, ^****P < 0.0001. P values were calculated using two-sided Student’s unpaired t-tests (b), one-way ANOVA followed by Tukey’s multiple-comparison test (d, f, g), and Kruskal–Wallis with Dunn’s multiple-comparison test (h). Source data and exact P values are provided as a Source Data file.

Migration of small intestinal memory γδ T17 Cells into the lung aggravates ALI

Beyond Th17 cells, several other cell types have been confirmed as sources of IL-17A, including γδ T cells, natural killer T cells, CD4⁺ T helper 17 (Th17) cells, CD8⁺ T cells, neutrophils, macrophages, and lineage^– innate lymphoid cells (ILCs)^{15,16,17,18,19,20}. Therefore, we delved deeper into the dominant cellular source of IL-17A in the lung after sepsis. Notably, through flow cytometry (FCM) analysis, we found that γδ T cells stood out as the dominant cellular source of IL-17A production and secretion in the lung 24 h post-CLP (Fig. 3a, b). To validate this finding, we further examined the proportion and absolute number of γδ T cells within the CD45⁺ immune cell population secreting IL-17A. Similarly, γδ T cells constituted the dominant cellular source of IL-17A secretion among the total immune cell population (Fig. 3c, d), but not lineage^– ILCs, CD4⁺ T cells, and CD8⁺ T cells. Additionally, we demonstrated a significant increase in the pulmonary γδ T cells in the septic mice at 24 h post-CLP (Fig. 3e, f). Thus, our findings indicate a substantial increase in the pulmonary γδ T cells 24 h post-CLP, making them the dominant source of IL-17A in the lung. Our data align with previous research demonstrating the robust ability of γδ T cells to produce IL-17A, particularly during the early stages of innate immune responses in diseases^18,21.

Fig. 3: Intestinal γδ T cells act as the dominant source of early pulmonary IL-17A in sepsis.

a, b Representative FCM plots, percentages, and quantification of lung IL-17A⁺ live γδ T cells (IL-17A⁺ CD3⁺ γδ TCR⁺) in sham-operated and CLP mice (n = 8 for sham, 6 for CLP). c, d Representative FCM plots, percentages, and quantification of lung CD45⁺ IL-17A⁺ γδ T cells (CD45⁺ IL-17A⁺ CD3⁺ γδ TCR⁺) in sham-operated and CLP mice (n = 8 for sham, 9 for CLP). e, f Representative FCM plots, percentages, and quantification of lung γδ T cells in sham-operated and CLP mice (n = 7). g Integrated UMAP of the T lymphocytes from sham-operated and CLP mice. h Feature plots of the Trdc expression (referred to γδ T cells) in the immune cell subsets. i The frequency of each cell type is depicted in the columns in sham-operated and CLP mice. j, k Representative FCM plots, percentages, and quantification of γδ T cells in the small intestine of sham-operated and CLP mice (n = 7 for sham, 8 for CLP). l, m Representative FCM plots and percentages of Ki67⁺ γδ T cells in the small intestine of sham-operated and CLP mice (n = 6 for sham, 7 for CLP). n–q FCM analysis and quantification of lung IL-17A⁺ lymphocytes, IL-17A⁺ γδ T cells, and Kaede red⁺ IL-17A⁺ γδ T cells from Kaede-tg mice after photoconversion of the small intestine (n = 10 for sham + saline, 7 for CLP + saline, 6 for CLP + S-KT). Data are shown as mean ± SD. ^*P < 0.05, ^**P < 0.01, ^****P < 0.0001. P values were calculated using two-sided Student’s unpaired t-test (a, d, f, k, m), and Kruskal–Wallis with Dunn’s multiple-comparison test (o–q). Source data and exact P values are provided as a Source Data file.

Furthermore, we observed a significant increase in intestinal permeability 24 h post-CLP, while S-KT treatment effectively ameliorated sepsis-induced damage to the intestinal barrier (Supplementary Fig. 1c–e). Consistently, we observed a significant decrease in mucin-2 (MUC-2) expression in the small intestine, and S-KT showed positive restorative effects on MUC-2 protein loss 24 h post-CLP (Supplementary Fig. 1f). The two tight junction proteins, ZO-1 and Occludin, showed a decreasing trend after sepsis, although the differences were not statistically significant (Supplementary Fig. 1g, h). To further gain a comprehensive understanding of changes in intestinal immune cells after sepsis, we conducted single-cell sequencing on CD45⁺ immune cells from the small intestine of septic mice. We found a significant increase in Trdc-positive cells (γδ T cells) in the small intestine 24 h post-CLP (Fig. 3g–i). We validated this finding through FCM analysis, which was highly consistent with our single-cell sequencing results (Fig. 3j, k). We further found that γδ T cells exhibited high expression of Ki67 24 h after CLP compared to those in the sham-operated group, indicating that γδ T cells presented an actively proliferative state under septic condition (Fig. 3l, m). These data also suggest that small intestinal γδ T cells are self-renewing, independent of replenishment from circulating γδ T cells, and display proliferative bursts during sepsis.

We and others have previously reported that in an acute ischemic brain injury model, small intestinal γδ T cells could migrate to the brain and lung and exert detrimental effects by secreting IL-17A^8,22. Moreover, considering our previous findings of significantly increased pulmonary γδ T cells in the early stages of sepsis, we hypothesized a significant expansion of small intestinal γδ T cells and their migration to the lung following sepsis. To investigate this, the small intestine of Kaede-tg mice underwent localized irradiation to track the migration of small intestinal lymphocytes (Supplementary Fig. 2a). The fluorescence results confirmed the effectiveness and high efficiency of ultraviolet light irradiation-induced fluorescence conversion (Supplementary Fig. 2b). We ruled out the impact of ultraviolet light irradiation on the expression of Kaede Red⁺ fluorescence in organs other than the small intestine (mesenteric lymph nodes (mLNs), blood, lung, and spleen) (Supplementary Fig. 2c). A notable surge in IL-17A and IL-17A-secreting γδ T cells in the lung was observed 24 h post-CLP, while S-KT treatment effectively suppressed these changes induced by sepsis (Fig. 3n–p). This further corroborated our previous findings in the wild-type mouse sepsis model. Importantly, we detected small intestinal-derived γδ T cells (Kaede red⁺ CD3⁺ γδ TCR⁺) in the lung 24 h post-CLP, with these cells comprising the major population among the IL-17A-secreting γδ T cells in the lung (Fig. 3n, q). However, we could not rule out the possibility that some small intestinal γδ T cells migrated to the lung before differentiating to γδ T17 cells.

Furthermore, by in-depth analysis of single-cell sequencing data from the small intestine at 24 h post-CLP, we identified predominant small intestinal γδ T cell populations originating from both the T cell and ILC groups (Supplementary Fig. 3a). Reclustered Trdc⁺ T cells exhibited high expression of CD8a and granzyme B (GZMB) (Supplementary Fig. 3b), suggesting potential effector and cytotoxic functions. Conversely, Trdc⁺ reclustered innate lymphoid-like cells (population 3), which expressed high levels of marker genes for γ δ T cells (CD3 and Trdc) and moderate levels of the ILC marker gene (Id2), displayed high expression of IL-7R but low expression of CD8a (Supplementary Fig. 3c, d). Subsequent merging and clustering of ILCs and T cells revealed that γδ T cells primarily segregate into two subtypes: γδ T C1 (CD44^high Ly6C^– IL-7R^high CD8^low γδ T cells) and γδ T C2 (CD44^low Ly6C^– IL-7R^low CD8^high γδ T cells) (Fig. 4a, b). Of note, the expression level of IL-17A is markedly higher in γδ T C1 relative to that in γδ T C2 (Fig. 4b). The ratio of γδ T C2 subset increased, while the ratio of γδ T C1 subset decreased in septic mice (Fig. 4c). Furthermore, these two subsets of γδ T cells exhibited distinct differentiation trajectories (Fig. 4d, e). In support of our single-cell sequencing data, a significant decrease in IL-7R^high CD8^low γδ T cell percentage and a significant increase in IL-7R^low CD8^high γδ T cell percentage were observed in the small intestine 24 h post-CLP (Fig. 4f). IL-7R^low CD8^high γδ T cells were the major cell population secreting IL-17A in the small intestine (Fig. 4g). In contrast, IL-7R^high CD8^low γδ T cell percentage significantly increased in the lung, while IL-7R^low CD8^high γδ T cell percentage significantly decreased (Fig. 4f), and IL-7R^high CD8^low γδ T cells constitute the dominant cell population secreting IL-17A in the lung (Fig. 4g). We verified our conjecture using Kaede-tg mice, and found that almost all small intestine-derived γδT17 cells in the lung are Ly6C^– IL-7R^high CD8^low memory γδ T cell subtypes (Fig. 4h). Similarly, we also could not excluded the possibility that some small intestinal Ly6C^– IL-7R^high CD8^low memory γδ T cells migrated to the lung before differentiating to Ly6C^– IL-7R^high CD8^low memory γδ T17 cells.

Fig. 4: Small intestinal IL-7R^high CD8^low γδ T cells are the dominant source of IL-17A in the mice lung after sepsis.

a T lymphocytes and ILCs in the small intestine were analyzed from sham-operated and septic mice and projected with UMAP plots with colors and the number in parenthesis indicating the identified cell cluster. b Violin plots demonstrate expression of the genes that identify each γδ T cell cluster. c The proportion of each subset of cells in the small intestine. d Diagram of pseudotime trajectory differentiation stages of γδ T cells. e Pseudotime trajectory analysis differentiation plot based on cell types. f Representative FCM plots and percentages of IL-7R and CD8 expression in γδ T cells in the small intestine and lung (n = 8 for sham, 10 for CLP). g Representative FCM plots and percentages of IL-7R and CD8 expression in γδ T17 (IL-17A⁺ γδ T) cells (n = 8 for sham, 10 for CLP). h FCM gate strategy of γδ T17 subgroup and analysis. Graph shows the percentages of IL-7R^low CD8^high γδ T17 cell and IL-7R^high CD8^low γδ T17 cell subsets in pulmonary Kaede red⁺ γδ T17 cells in septic mice (n = 7). i, j Representative H&E sections of lung and histological injury scores (n = 8). 20×, Scale bar = 100 μm. 40×, Scale bar = 50 μm. k The lung W/D weight ratio (n = 10), l pulmonary capillary leakage determined with Evans blue dye (n = 8), m total protein in the BALF (n = 8), n the protein expression of inflammatory cytokines IL-17A, IL-6 and TNF-α (n = 8), and o Kaplan–Meier analysis of survival (n = 22 for Il-7r^flox/flox, n = 21 for Trdc^creERT2 Il-7r^flox/flox) in septic Il-7r^flox/flox mice and Trdc^creERT2 Il-7r^flox/flox mice. Data are shown as mean ± SD. ^*P < 0.05, ^**P < 0.01, ^***P < 0.001, ^****P < 0.0001. P values were calculated using two-sided Student’s unpaired t-test (f–h, j–n), and the Log-rank (Mantel–Cox) test (o). Source data and exact P values are provided as a Source Data file.

Considering small intestinal Ly6C^– IL-7R^high CD8^low memory γδ T17 cells are the dominant source of IL-17A in the lung of septic mice 24 h post-CLP, we, therefore, aimed to assess whether the absence of IL-7R^high γδ T cells in the lung could alleviate the immune damage in septic mice, we crossed Trdc^CreERT2 mice to Il-7r^flox/flox mice to obtain Trdc^CreERT2 Il-7r^flox/flox mice in which Il-7r expression was predominantly abrogated in γδ T cells. Our data clearly demonstrate that Trdc^CreERT2 Il-7r^flox/flox mice exhibited substantial improvement in lung injury compared to control mice 24 h post-CLP (Fig. 4i–m). Additionally, there was a notable suppression of inflammatory responses in the lung of Trdc^CreERT2 Il-7r^flox/flox mice, as evidenced by a reduction in the levels of IL-17A, IL-6, and TNF-α in the bronchoalveolar lavage fluid (BALF) (Fig. 4n). In line with these findings, Trdc^CreERT2 Il-7r^flox/flox mice exhibited significantly enhanced 7-day survival rates (Fig. 4o). This work elucidates a hitherto unrecognized migration of a distinct small intestinal γδ T cell subset to the lung, associated with septic ALI.

Role of pulmonary Wnt signaling activation in intestinal γδ T cell migration

Subsequently, we investigated the factors that drive the migration of intestinal γδ T cells to the lung. Chemokines play a crucial role in γδ T cell migration²². We observed a significant increase in the chemokine CCL1 in the lung homogenates of mice 24 h post-CLP, with a homogeneous distribution within the group (Fig. 5a). We further confirmed the upregulation of CCL1 in the lung 24 h post-CLP, which was significantly reduced by S-KT treatment (Fig. 5b). Furthermore, we explored the cellular source of the increased CCL1 in the lung. In vitro stimulation of the three cell types with the highest proportions in the lung: A549 cell line (lung epithelial cells), RAW264.7 cell line (monocytes/macrophages), and MH-S (murine alveolar macrophages (AMs)). A549 and RAW264.7 cells failed to induce significant secretion of CCL1, except for MH-S cells stimulated with lipopolysaccharide (LPS), which showed a significant increase in CCL1 protein expression (Supplementary Fig. 4a–c). Furthermore, we isolated alveolar macrophages from BALF obtained from both septic mice and human patients (Supplementary Fig. 4d, e). Our analysis revealed a significant upregulation in the expression of CCL1 under septic conditions (Supplementary Fig. 4f, g). The result indicates that AMs may be the major cellular source of CCL1 during the early stages of sepsis, in line with previous reports that AMs are the major source of lung chemokines and produce inflammatory mediators^23,24.

Fig. 5: Pulmonary Wnt signaling activates CCL1 expression to attract γδ T cells after sepsis.

a Heatmap showing the chemokine expression signature in sham-operated and CLP mice (n = 4). b The protein levels of CCL1 in the murine lung after combined prophylactic and therapeutic administration of S-KT or 0.9% saline (n = 6). c Transwell assay was used to assess the ability of CCL1 to chemotaxis sorted-γδ T cells from spleen mononuclear cells (n = 6). d CCR8 expression by intestinal γδ T cells in sham-operated and CLP mice was measured by FCM analysis (n = 6 for sham, 7 for CLP). Sorted-γδ T cells from spleen mononuclear cells in sham-operated or CLP mice were co-cultured with AMs. Expression of e IL-17A and f IFN-γ by γδ T cells was measured by FCM analysis (n = 3). g Heatmap of the top 29 up-regulated DEGs in the lung between the sham-operated mice and CLP mice (n = 3). h The protein levels of the components of the Wnt signaling pathway (Wnt5a, FZD5, β-catenin, and lymphoid enhancer-binding factor 1 (Lef1)) in murine lungs after combined prophylactic and therapeutic administration of S-KT or 0.9% saline (n = 6). i Average plots and heat maps showing the signals at the TSS in the ChIP-seq data sets. j Pie charts showing the distribution of all Lef1 peak locations. k Distal intergenic region-specific binding peaks between Lef1 and CCL1 gene. Data are shown as mean ± SD. ^*P < 0.05, ^**P < 0.01, ^***P < 0.001, ^****P < 0.0001. ns not significant. P values were calculated using one-way ANOVA followed by Tukey’s multiple-comparison test (b, h), and two-sided Student’s unpaired t-test (c–f). Source data and exact P values are provided as a Source Data file.

We further investigated whether CCL1 could exert chemotactic effects on γδ T cell migration. Using magnetic bead sorting, we obtained highly purified γδ T cells (Supplementary Fig. 4h) and conducted Transwell migration assays, which revealed a clear chemotactic effect of CCL1 on γδ T cells (Fig. 5c). We found that CCL1 retained a strong chemotactic effect on γδ T cells within the mixed immune cell population of the splenic single-cell suspension (Supplementary Fig. 4i). Moreover, lung homogenates obtained from septic mice demonstrated enhanced chemotactic potential toward γδ T cells from splenic single-cell suspensions when compared to lung homogenates from sham-operated mice (Supplementary Fig. 4j). To further elucidate the role of pulmonary CCL1 in this process, we selectively reduced the levels of CCL1 in the lungs of septic mice by intratracheal administration of shRNA-CCL1 adenoviruses (Supplementary Fig. 4k). This intervention significantly attenuated the chemotactic capacity of septic lung homogenates towards γδ T cells (Supplementary Fig. 4j).

CCL1 plays a pivotal role in modulating tissue homeostasis by orchestrating the recruitment of immune cells to sites of inflammation, primarily through its interaction with the CCR8 receptor²⁵. To elucidate the involvement of this signaling axis in our study, we investigated the expression of CCR8 on small intestinal γδ T cells. Remarkably, we observed a significant upregulation of CCR8 expression on small intestinal γδ T cells at 24 h post-CLP (Fig. 5d). In light of the elevated expression of other chemokines such as CCL2, CCL7, CCL12 (ligands for CCR2), and CXCL13 (ligand for CXCR5) in septic lung homogenates (Fig. 5a), we also assessed the expression of CCR2 and CXCR5 on small intestinal γδ T cells. Our analysis revealed that only a small fraction of small intestinal γδ T cells expressed CCR2 and CXCR5, with no significant difference observed between septic and sham-operated groups (Supplementary Fig. 4l, m).

Given that the CCL1/CCR8 axis possesses direct chemotactic migration ability to γδ T cells, and the elevated CCL1 in the lung stems mainly from AMs, we next considered whether activated AMs could potentially direct the differentiation of γδ T cells. AMs from both sham-operated and septic mice were isolated and co-cultured with γδ T cells. Surprisingly, we discovered that the AMs from septic mice promoted the differentiation of γδ T cells into IL-17A-secreting γδ T cells, rather than IFN-γ-secreting γδ T cells (Fig. 5e, f). The result indicates that AMs might not only promote the migration of small intestinal γδ T cells or γδ T17 cells but also direct the differentiation of γδ T cells to γδ T17 cells.

Wnt signaling activation up-regulates CCL1 expression in the lung

To gain further insights into the underlying cause of the upregulation of pulmonary CCL1 following sepsis, we conducted an analysis of the proteomics data from lung tissues obtained 24 h post-CLP. Our investigation unveiled a notable increase in the protein level of FriZZled5 (FZD5) in septic lungs (Fig. 5g). FZD5, a critical receptor protein in the Wnt signaling pathway, acts in concert with Wnt family member 5 A (Wnt5a) as a signaling molecule in the immune response to infection²⁶. We observed a significant elevation in FZD5 level in the lung 24 h post-CLP, concomitant with increased expression of the upstream ligand protein Wnt5a, downstream protein β-catenin, and transcription factor Lef1 (Fig. 5h). Notably, a similar trend was also observed in AMs from both septic and non-septic patients (Supplementary Fig. 4n). Intriguingly, treatment with S-KT exhibited significant inhibition of Wnt signaling activation in the lungs of mice following sepsis (Fig. 5h).

To corroborate the interaction between S-KT and components of the Wnt signaling pathway and discern the main drug-target binding modes, we employed molecular docking to predict compound-target interactions between S-KT and Wnt5a, and FZD5. The calculated binding energy of S-KT to FZD5 was −5.9 kcal/mol. Notably, Tyr507 formed a hydrogen bond with the ketone group of S-KT, suggesting its involvement in the interaction’s electron transfer (Supplementary Fig. 5a). Additionally, acidic amino acids Asp384 and Glu471 formed salt bridges with the nitrogen ion of the substrate, contributing to the stabilization of the protein’s secondary structure. Tyr391 and Tyr468 participated in pi–pi stacking interactions with the phenyl ring of S-KT, further stabilizing the substrate in the binding pocket. The surrounding amino acids in the FZD5 binding pocket exhibited rigidity, implying a relatively rough binding pocket with a presumed narrower substrate range (Supplementary Fig. 5b). Similarly, the binding energy of S-KT to Wnt5a was calculated as −4.6 kcal/mol. Arg200 and Arg204 formed hydrogen bonds with the ketone group of S-KT, while Arg204 also engaged in pi–pi stacking with the benzene ring (Supplementary Fig. 5c). Additionally, Met128 and Tyr139 formed hydrogen bonds with the nitrogen ion, facilitating electron transfer in the reaction. The presence of basic and hydrophobic amino acids surrounding the Wnt5a binding pocket likely contributed to stabilizing the enzyme’s conformation and creating a suitable microenvironment for catalytic activity (Supplementary Fig. 5d). We postulate that the activation of the Wnt/β-catenin signaling pathway during the early stages of sepsis may play a pivotal role in the immunopathological processes underlying ALI.

Studies have shown that the TCF/Lef transcription factors in the canonical Wnt pathway directly regulate the transcription of various target genes, including several chemokines involved in immune cell recruitment to sites of infection²⁷. Therefore, we speculated that the activation of the Wnt/β-catenin pathway in the lung during the early stages of sepsis might transcriptionally regulate CCL1, thus promoting the migration of small intestinal γδ T cells to the lung. Considering the significant increase in CCL1 secretion by LPS-stimulated MH-S cells and the marked elevation of the Wnt signaling transcription factor Lef1 in the lung 24 h post-CLP, we performed chromatin immunoprecipitation followed by sequencing (ChIP-seq) experiments targeting Lef1 in MH-S cells to investigate the potential specific binding between Lef1 and CCL1. The ChIP-seq results exhibited high signal intensity and a high signal-to-noise ratio in the peak regions of MH-S cells (Fig. 5i). The distribution of peaks across different genomic feature types is shown in Fig. 5j. Our sequencing results revealed specific binding of Lef1 to the distal region of the CCL1 gene (Fig. 5k), suggesting that Lef1 may play a transcriptional regulatory role in CCL1 in AMs.

We further investigated the regulatory relationship between the activated Wnt signaling and CCL1 after sepsis. iCRT14, an inhibitor of β-catenin in the Wnt signaling, was administrated to inhibit the early activation of the Wnt signaling in the lung during sepsis. The inhibitory effects of iCRT14 on key proteins of the Wnt signaling were validated by western blotting (Fig. 6a). Importantly, a significant decrease in CCL1 protein levels in the lung of septic mice was observed 24 h post-CLP following iCRT14 treatment (Fig. 6b). Moreover, we conducted in vitro experiments to further validate the regulatory relationship between the Wnt signaling and CCL1. Pre-treatment of MH-S cells with iCRT14 significantly reduced the elevated expression of CCL1 induced by LPS stimulation (Fig. 6c, d). Our results demonstrate that the activation of the Wnt signaling after sepsis is involved in the positive regulation of CCL1. To further clarify the involvement of pulmonary CCL1 in the migration of intestinal γδ T cells, we selectively lowered CCL1 levels in the lungs of septic mice through intratracheal administration of shRNA-CCL1 adenoviruses and intranasal administration of anti-CCL1 antibodies. Our findings demonstrate that blockade of pulmonary CCL1 markedly diminished the migration of small intestinal γδ T17 cells to the lung (Fig. 6e, f).

Fig. 6: Inhibiting pulmonary Wnt signaling activation mitigates septic ALI by downregulating CCL1.

a The protein expression of β-catenin, Lef1, TCF1/7 and b CCL1 in the lung (n = 3 for Lef1, 6 for TCF1/7 and CCL1). c A diagram of MH-S cell line treated with iCRT14 or vehicle followed by stimulation with LPS or PBS for 24 h in vitro to analyze the protein level of CCL1. d The protein level of CCL1 in MH-S cells (n = 7). e Gating strategy and frequency for Kaede red⁺ γδ T17 cells in the lung of Kaede-tg septic mice treated with CCL1- shRNA (Adv-shRNA (CCL1)) or adenovirus-negative control (Adv-NC) (n = 6). f Gating strategy and frequency of Kaede red⁺ γδ T17 cells in the lung of Kaede-tg septic mice treated with anti-CCL1 mAb or control IgG (n = 6). g Gating strategy and quantification of γδ T cells in the small intestine (n = 8). h Representative FCM plots and absolute number of Kaede red⁺ γδ T17 cells in the lung of Kaede-tg septic mice treated with vehicle or iCRT14 (n = 6). i, j Representative H&E sections of lung and histological injury scores (n = 7). 20×, Scale bar = 100 μm. 40×, Scale bar = 50 μm. k The levels of IL-17A (n = 10 for sham + vehicle, 7 for CLP + vehicle, 6 for CLP + iCRT14), IL-6 (n = 10 for sham + vehicle, 8 for CLP + vehicle and CLP + iCRT14) and total protein (n = 6) in the BALF. l The protein level of IL-17A and IL-6 in the lung (n = 6). Data are shown as mean ± SD. ^*P < 0.05, ^**P < 0.01, ^***P < 0.001, ^****P < 0.0001. ns not significant. P values were calculated using one-way ANOVA followed by Tukey’s multiple-comparison test (a, b, d, k, l), and two-sided Student’s unpaired t-test (e–h, j). Source data and exact P values are provided as a Source Data file.

We next investigated the impact of iCRT14 on the migration of intestinal γδ T cells. iCRT14 did not affect the cell count of small intestinal γδ T cells but significantly suppressed the migration of IL-17A-producing γδ T cells from the intestine to the lung (Fig. 6g, h). Correspondingly, iCRT14 alleviated lung damage in septic mice (Fig. 6i, j), and also reduced the expression levels of IL-17A and IL-6 in the BALF of septic mice, as well as the overall protein extravasation (Fig. 6k). Moreover, iCRT14 inhibited the elevated protein levels of IL-6 and IL-17A in the lung 24 h post-CLP (Fig. 6l). This result suggests that the activation of Wnt signaling could mediate the upregulation of pulmonary CCL1 expression, thereby promoting the migration of small intestinal γδ T cells to the lung.

We next assessed the alterations of three key immune cell types in the early septic lung, including AMs, interstitial macrophages (IMs), and neutrophils (Supplementary Fig. 6). Our results revealed a significant increase in both AMs and IMs within the lung 24 h post-CLP (Supplementary Fig. 7a–c), accompanied by a marked elevation in M1 polarization and a notable reduction in M2 polarization for both macrophage types (Supplementary Fig. 7d–f). Notably, the influence of S-KT treatment on AMs was particularly pronounced, as it significantly attenuated the observed increase in AMs proportions and M1 polarization following sepsis (Supplementary Fig. 7a, c, d, f). It’s worth highlighting that neutrophil proportions within the lung substantially increased 24 h post-CLP, which was significantly curtailed after S-KT treatment (Supplementary Fig. 7g, h). The neutrophils are capable of exacerbating inflammatory responses by generating neutrophil extracellular traps (NETs). We unveiled a substantial increase in the NETs formation within the lung of septic mice 24 h post-CLP, which was significantly alleviated by S-KT treatment (Supplementary Fig. 7i, k).

The role of IL-17A in septic ALI

Previous animal experiments have demonstrated significant benefits from IL-17 blockade²⁸, however, there also exists paradoxical conclusions²⁹. This might be because systemic IL-17 depletion also blocks IL-17 signaling pathways in tissues and organs where the inflammatory response is not particularly strong, thereby inhibiting the body’s normal bacterial clearance processes. For intense and detrimental inflammatory host responses triggered by pathogenic infections, localized neutralization of the IL-17 axis might be more suitable. Therefore, we further investigated the roles of IL-17A on ALI by intratracheally administering adenoviruses for IL-17A to result in a specific intervention in pulmonary IL-17A expression (Supplementary Fig. 8a, b), without notable influence on the IL-17A levels in other vital organs (Supplementary Fig. 8c). Adenovirus-mediated overexpression of IL-17A in septic mice led to a significant increase in the proportion of AMs (Fig. 7a), a marked reduction in IMs proportion (Fig. 7b), a significant increase in neutrophil proportion (Fig. 7c), and increased M1 polarization of AMs and IMs (Fig. 6d, e) in the lung. Consistently, IL-17A overexpression also resulted in an increase in neutrophil proportion and M1 polarization of AMs in the lung of normal mice (Supplementary Fig. 9a–d), along with an increase in NETs formation (Supplementary Fig. 9e). It is noteworthy that IL-17A overexpression in septic mice exacerbated ALI (Fig. 7f). On the contrary, adenovirus-mediated downregulation of IL-17A in septic mice led to a significant decrease in the proportions of AMs and neutrophils in the lung (Fig. 7g, h), along with a remarkable improvement in lung injury (Fig. 7i) and decreased NETs formation (Fig. 7j). Simultaneously, the levels of inflammatory cytokines IL-6 and TNF-α in the lung were increased by IL-17A overexpression, but were decreased by IL-17A downregulation in septic mice (Fig. 7k). Importantly, we observed that the downregulation of IL-17A in septic mice significantly improved the 7-day survival rate, even though IL-17A overexpression failed to significantly lower the survival rate of septic mice (Fig. 7l). These data suggest that IL-17A could exacerbate post-septic ALI by the increased recruitment of neutrophils and formation of NETs, and increased M1 polarization of AMs and IMs.

Fig. 7: The impact of IL-17A on septic ALI.

a Representative FCM plots and frequency of AMs, b IMs, and c neutrophils in the lung of septic mice treated with IL-17A-expressing recombinant adenovirus (Adv-IL-17A) or control Adv (Adv-Ctrl) (n = 6 for CLP + Adv-Ctrl, 7 for CLP + Adv-IL-17A). d, e Representative FCM plots and percentage of CD86 and CD206 markers in AMs and IMs (n = 6). f Representative H&E sections of lung and histological injury scores in septic mice treated with Adv-Ctrl or Adv-IL-17A (n = 7). 20×, Scale bar = 100 μm. 40×, Scale bar = 50 μm. g, h Representative FCM plots and frequency of AMs and neutrophils in the lung of septic mice treated with adenovirus expressing IL-17A- shRNA (Adv-shRNA (IL-17A)) or adenovirus-negative control (Adv-NC) (n = 6). i Representative H&E sections of lung and histological injury scores. 20×, Scale bar = 100 μm. 40X, Scale bar = 50 μm. j Representative images and semi-quantitative analysis of NETs formation in the lung detected by immunofluorescence staining with Cit-H3 and MPO (n = 6). 20X, Scale bar = 50 μm. 40×, Scale bar = 25 μm. k The protein level of IL-6 and TNF-α in the BALF in septic mice treated with Adv-IL-17A (or Adv-Ctrl) or Adv-shRNA (IL-17A) (or Adv-NC) (n = 8 for IL-6 in CLP + Adv-Ctrl and CLP + Adv-NC, TNF-α in CLP + Adv-IL-17A and CLP + Adv-Ctrl, 7 for IL-6 in CLP + Adv-IL-17A and CLP + Adv-shRNA (IL-17A), TNF-α in CLP + Adv-Ctrl and CLP + Adv-shRNA (IL-17A)). l Kaplan–Meier analysis of survival in septic mice treated with Adv-IL-17A (or Adv-Ctrl) or Adv-shRNA (IL-17A) (or Adv-NC) (n = 20). Data are shown as mean ± SD. ^*P < 0.05, ^**P < 0.01, ^***P < 0.001, ^****P < 0.0001. P values were calculated using two-sided Student’s unpaired t-test (a–k), and the Log-rank (Mantel–Cox) test (l). Source data and exact P values are provided as a Source Data file.

Discussion

γδ T cells encompass various cellular subtypes³⁰. In this study, the pulmonary cells secreting IL-17A primarily stem from the Ly6C⁻ IL-7R^high CD8^low memory γδ T17 cell subset (CD44^high Ly6C⁻ IL-7R^high CD8^low) within the small intestine after sepsis. Ly6C⁻ IL-7R^high CD8^low memory γδ T17 were the major subtype of γδ T17 cells that migrated from the small intestine to the lung and aggravated septic ALI. Previous research has documented the promotion of allergic lung inflammation by pulmonary memory γδ T cells through the secretion of IL-17A³¹. Furthermore, within peripheral lymph nodes and mesenteric lymph nodes (mLNs), only CD44⁺ Ly6C^– γδ T cells could exhibit IL-17A secretion, but not CD44^low Ly6C^+/– γδ T cells and CD44⁺ Ly6C⁺ γδ T cells³⁰, in line with our data. IL-7 is shown to be requisite for intrinsic TCRγ gene rearrangement in fetal and adult thymus and for the survival of adult thymocytes. However, in peripheral γδ T cells, IL-7 is not obligatory for their survival or proliferation, although it can extend the lifespan of mature γδ T cells³². Il-7^−/− and Il-7rα^−/− mice both lack mature γδ T cells³². Through pseudo-temporal analysis from single-cell sequencing of the small intestine after sepsis, we inferred that the Ly6C^– IL-7R^high CD8^low γδ T cell subtype is likely indicative of mature γδ T cells, while Ly6C^– IL-7R^low CD8^high γδ T cells represent immature γδ T cells. Our results suggest that post-sepsis, mature Ly6C⁻ IL-7R^high CD8^low memory γδ T17 cells from the small intestine migrated to the lung and promoted ALI through the secretion of IL-17A, and that targeting of pulmonary IL-17A could serve as promising adjunct therapy against this challenging disease.

One characteristic hallmark of sepsis-induced lung injury is the infiltration of white blood cells, a phenomenon primarily guided by chemokines³³. Early response cytokines, adhesion molecules, and chemokines orchestrate the concerted recruitment of inflammatory cells into the lung parenchyma. Our investigation identifies CCL1 as a significant player in the pathological progression of acute lung injury in sepsis. Accumulating evidence supports alveolar macrophages (AMs) as pivotal sources of lung chemokines, concurrently producing inflammatory mediators such as IL-8 and GRO-related peptides^23,24. In line with these findings, our study underscores the importance of AMs as key contributors to CCL1 production during the initial phases of sepsis. Notably, levels of various other chemokines, including CCL2, CCL7, CCL12 (ligands for CCR2), and CXCL13 (ligand for CXCR5), also exhibit elevation in septic lung homogenates. Previous evidence indicates that γδ T17 cells in mucocutaneous tissues, constitutively express chemokine receptors CCR6 and CCR2. While CCR6 facilitates the recruitment of resting γδ T17 cells to the dermis, CCR2 drives their rapid influx to inflamed tissues during conditions such as autoimmunity, cancer, and infection³⁴. Thus, we investigated the expression of CCR2 and CXCR5 on small intestinal γδ T cells following sepsis. Our analysis revealed that only a minor fraction of small intestinal γδ T cells expressed CCR2 and CXCR5, with no significant disparity observed between septic and sham-operated groups. Additionally, prior studies have implicated CCL2 and CCL5, produced by cardiac tissue, in mediating the migration of myocardial γδ T cells during myopathic conditions³⁵, suggesting variations in the chemokine axes involved in driving γδ T cell migration across different disease models. However, limited research has focused on delineating the chemokines responsible for γδ T cell migration in sepsis-induced lung injury. Our findings provide novel insights by confirming the potent chemotactic effect of pulmonary CCL1 on small intestinal γδ T cells, suggesting that CCL1, predominantly produced by AMs and positively regulated by Wnt signaling pathway during the early stages of sepsis, may play a crucial role in orchestrating the migration of small intestinal γδ T17 cells to the lung parenchyma.

As one of the enantiomers of ketamine, S-KT exhibits similar mechanisms of action and pharmacological properties to ketamine. Clinical evidence suggests that intraoperative administration of ketamine can alleviate postoperative inflammatory responses following major surgeries, including cardiac and abdominal procedures^36,37. Ketamine appears to exert its anti-inflammatory activity under conditions of immune activation, acting as an anti-inflammatory agent rather than an immune suppressant³⁸. Studies indicate a close association between the antidepressant action of S-KT and its anti-inflammatory properties¹¹, however, the specific mechanisms under the anti-inflammatory effect of S-KT are not clear. In this study, we affirm the positive role of S-KT in sepsis-induced lung injury and shed light on its mechanism of action. Our findings indicate that S-KT exerts its effects by downregulating the activation of the pulmonary Wnt/β-catenin signaling, inhibiting the migration of small intestinal γδ T17 cells into the lung. This process helps to maintain the balance of the pulmonary local immune microenvironment during the early stages of sepsis and alleviates sepsis-induced lung injury. To further validate our findings, we employed molecular docking to assess the predicted compound-target interactions between S-KT and key components of the Wnt signaling pathway, including Wnt5a and FZD5. Our results suggest that S-KT may exert an inhibitory effect on the activated Wnt signaling pathway post-sepsis by specifically binding to Wnt5a and FZD5. However, further investigations are warranted to elucidate the precise mechanisms underlying these interactions.

Methods

Animals

Male wild-type (WT) C57BL/6 J mice, Trdc^CreERT2 mice, and Il-7r^flox/flox mice (C57BL/6 background) weighing between 23.0–25.0 g and aged 8–10 weeks, were obtained from Vital River Laboratory Animal Technology Co Ltd., Beijing, China. Trdc^CreERT2 mice were mated with Il-7r^flox/flox mice to generate Trdc^CreERT2 Il-7r^flox/flox mice. Kaede-transgenic (Kaede-Tg) mice (B6.Cg-Gt(ROSA)26Sor<tm1.1(CAG-kikGR) Kgwa>) were kindly provided by M. Tomura (Kyoto University). All mice were housed and maintained in specific pathogen-free conditions with a 12:12 light/dark schedule, controlled ambient temperature, and humidity, and were given free access to food and water. Ethical approval for all experiments was obtained from the committee of experimental animals of Tongji Medical College (Permission number: 3250) and followed the guidelines set by the National Institutes of Health Guide for the Care and Use of Laboratory Animals. The study was conducted in accordance with the ARRIVE guidelines (Animals in Research: Reporting In Vivo Experiments).

Animal model

Mice were subjected to xylazine-ketamine anesthesia and placed in a supine position. CLP was performed through a midline laparotomy. The abdominal area was shaved and disinfected. A 1-cm midline incision was made to expose the cecum, which was then ligated using a 4-0 silk suture, 1 cm proximal to the distal cecal extremity. Subsequently, the cecum was punctured using a 20-gauge needle, allowing for a small amount of cecal content to be extruded through both holes. After this procedure, the ligated cecum was returned to the peritoneal cavity, and the surgical wound was closed in layers. S-KT (15 mg kg^–1, Hengrui) or the 0.9% saline (10 ml kg^–1) was injected intravenously into mice in three different protocols as follows: (1) single-dose administration 24 h prior to CLP; (2) single-dose administration immediately after CLP; and (3) administration twice 24 h before and immediately after CLP. Inhibitor iCRT14 (Cat# S8704, Selleck) is administered at a concentration of 40 mg kg^–1. The compound was administered by i.p. injection three times a week for 3 weeks before the CLP procedure. The body weight, rectal temperature, and survival rate of mice in each group were observed and recorded for 21 days post-CLP surgery. The severity of sepsis was assessed using the Murine Sepsis Score (MSS) system³⁹.

To overexpress IL-17A, or knockdown IL-17A and CCL1 in the lung, Mice were subjected to intratracheal injection of adenoviruses 4 days prior to sham operation or CLP. The adenoviruses include: 1 × 10⁹ PFU IL-17A-expressing recombinant adenovirus (Adv-IL-17A, Cat# H27226, Obio), adenovirus expressing IL-17A-short hairpin RNA (shRNA) (Adv-shRNA (IL-17A), Cat# Y22590, Obio), and adenovirus expressing CCL1-shRNA (Adv-shRNA (CCL1), Cat# Y25327, Obio). Control mice received an equivalent amount of control adenoviruses. Detailed reagent information is provided in the Supplementary Table 1. Additionally, in another experimental group, CCL1 neutralizing antibody (75 μg, Cat# MAB845, R&D Systems) in sterile phosphate-buffered saline (PBS) was administered via intranasal administration daily for three days. Mice received three total doses of anti-CCL1 monoclonal antibodies, with two doses administered prior to the CLP challenge and one dose administered on the day immediately following the CLP challenge. Control mice received equivalent doses of normal hamster serum IgG.

Cell culture and treatment

A549 (human respiratory epithelial cell line, Cat# CL-0016), MH-S (murine alveolar macrophage cell line, Cat# CL-0597), and RAW264.7 (murine monocyte/macrophage cell line, Cat# CL-0190) were purchased from Punosai Life Science and Technology Co., Ltd. A549 cells were cultured in RPMI-1640 with 10% fetal bovine serum (FBS) at 37 °C containing 5% CO₂. A549 cells treated with 50 or 100 μg ml^–1 LPS for 24 h were regarded as the septic ALI model at the cellular level. RAW264.7 cells and MH-S cells were cultured in Dulbecco’s Modified Eagle Medium (DMEM) supplemented with 10% FBS and 1% penicillin–streptomycin, followed by incubation at 37 °C in 5% CO₂/95% air humidified atmosphere. For stimulation, Murine RAW264.7 cells were treated with 500 or 1000 ng ml^–1 LPS. MH-S cells were treated with 1000 ng ml^–1 LPS (Cat# L2880, Sigma–Aldrich). After 24 h of incubation in cultures, cells were harvested and processed for further measurements. To inhibit the activity of the Wnt pathway, MH-S cells were treated with iCRT14 (25 μM) for 4 h before LPS stimulation.

Isolation of primary mouse alveolar macrophages (mAMs)

mAMs were isolated from murine bronchoalveolar lavage fluid (BALF). Briefly, mice were anesthetized prior to lavage, and the trachea was carefully dissected. The lungs were then lavaged three times with 1 ml of ice-cold PBS, and the retained BALF was centrifuged at 420 × g and 4 °C for 6 min. The resulting pellet was collected, resuspended in PBS, and subsequently subjected to fluorescence-activated cell sorting (FACS, BD FACSAriaTM II cell sorter (BD biosciences)) to enrich CD45⁺ F4/80⁺ CD11c⁺ mAMs. Flow-sorted mAMs were divided for cell culture and protein extraction purposes.

Patient population and isolation of human alveolar macrophages (hAMs)

12 adult patients admitted to the Intensive Care Units (ICU) of Wuhan Union Hospital, including 6 septic patients (3 males and 3 females) who were diagnosed with sepsis according to the Third International Consensus Definitions for Sepsis and Septic Shock (Sepsis−3) within the first 24 h of admittance into the ICU and 6 non-septic patients (5 postoperative patients and 1 cardiovascular patients; 2 males and 4 females)¹, were included in our study. This project was reviewed and proved by the Ethics Committee of the Huazhong University of Science and Technology (2023-0143-01). hAMs were isolated from BALF from these septic and non-septic patients following investigative bronchoscopy with bronchoalveolar lavage for diagnostic reasons, after informed consent. Patients were sedated and administered 1% lidocaine for topical anesthesia. A fiberoptic bronchoscope was inserted into the proximal airways, and three 20- to 30-ml aliquots of sterile 0.9% normal saline were instilled into the right middle lobe medial or lateral segments and/or the lingula. BALF was collected via suction into sterile specimen containers, subsequently filtered through 40-µm cell strainers, and centrifuged at 420 × g for 10 min at 4 °C to pellet the cells. The retained pellet was washed twice with PBS and then stained with the appropriate antibodies: fixable viability dye (Cat# 564406, BD Biosciences), CD45 (0.5 ng μl^–1, Cat# 304027, Biolegend), CD169 (0.5 ng μl^–1, Cat# 307609, Biolegend), HLA-DR (0.5 ng μl^–1, Cat# 361611, Biolegend), followed by FACS to enrich CD45⁺ CD11b⁺ HLA-DR⁺ CD169⁺ hAMs.

Western blotting

Total proteins were extracted from tissues and cells. Protein concentrations were determined using the BCA method. Subsequently, Equal amount of protein was separated by SDS-PAGE and transferred onto PVDF membranes (Millipore). After blocking with 5% skimmed milk for 2 h at room temperature, the membranes were incubated overnight at 4 °C with primary antibodies against IL-17A (1:1000, Cat# sc-374218, Santa Cruz), IL-6 (1:1000, Cat# ab229381, Abcam), IL-1β (1:1000, Cat# A1112, ABclonal), Mucin-2 (MUC-2, 1:1000, Cat# 27675-1-AP, Proteintech), ZO-1 (1:1000, Cat# 21773-1-AP, Proteintech), Occludin (1:1500, Cat# 27260-1-AP, Proteintech), Wnt5a (1:1000, Cat# 619919, Zen Bio), FZD5 (1:1000, Cat# 222996, Zen Bio), β-catenin (1:2000, Cat# 51067-2-AP, Proteintech), Lef1 (1:500, Cat# 14972-1-AP, Proteintech), TCF1/7 (1:1000, Cat# A20835, ABclonal), anti-human CCL1 (1:1000, Cat# A10670, ABclonal), anti-mouse CCL1 (1:1000, Cat# MAB845, R&D Systems), GAPDH (1:2000, Cat# 60004-1-Ig, Proteintech) and β-actin (1:2000, Cat# 66009-1-Ig, Proteintech). The membranes were then incubated with HRP-conjugated anti-mouse IgG antibody (Cat# ANT019, AntGene). Protein bands were visualized using ECL Western Blotting Detection Reagents (Beyotime) and captured with a UVP gel documentation system (UVP, LLC, Phoenix). The band intensity was quantified using ImageJ software (National Institutes of Health, Bethesda). Detailed reagent information is provided in the Supplementary Table 1.

Hematoxylin and eosin (H&E) staining

The lung and small intestine tissues were harvested and promptly fixed in 4% paraformaldehyde (pH 7.4). Subsequently, the tissues were embedded in paraffin for sectioning at a thickness of 5 μm. The paraffin-embedded lung sections were then deparaffinized and rehydrated using xylene and alcohol, respectively. For histological visualization, the sections were stained with H&E to respectively label the nucleus and cytoplasm. Finally, the specimens were observed and images were captured under a light microscope (Olympus, Japan). For each section, five visual fields were randomly selected. The pathological damage of lung tissue and small intestine was then scored based on the observed features^40,41.

Lung wet/dry (W/D) weight ratio

Lung edema was assessed by calculating the W/D weight ratio of the lung tissues. The right lungs were excised and weighed immediately to determine the wet weight. Subsequently, the lung samples were placed in an incubator at 60 °C for 48 h to obtain the dry weight. The lung W/D ratio was then calculated as the ratio of the wet weight to the dry weight.

Bacterial counting

Mice were humanely euthanized under xylazine-ketamine anesthesia. Following thoracotomy, lungs were aseptically collected, weighed, and homogenized in 1 ml of sterile PBS. To determine the colony-forming units (CFU), 10 μl of lung homogenates were serially diluted (1:10) in sterile PBS, and 20 μl of the diluted tissue homogenates were plated onto LB broth agar plates. The plates were then incubated at 37 °C for 24 h, and the number of bacterial colonies was counted to quantify the bacterial load. Bacterial loads were expressed as CFU per ml.

Evans blue dye permeability assay

After 21 h following sepsis induction or sham operation in mice, 2% Evans Blue dye (Sigma–Aldrich, 4 ml kg^–1) was intravenously injected via the mouse tail vein. 3 h post-dye injection, the mice were transcardially perfused with PBS to ensure complete removal of Evans Blue dye from the pulmonary circulation. The entire lung tissue was then excised from the mice and homogenized in 1 ml of N,N-dimethylformamide (Sigma–Aldrich), incubated overnight at 55 °C in a water bath, and centrifuged at 12,000 × g for 20 min. The OD of the supernatants was measured at 620 nm by spectrophotometry. The assessment of vascular permeability was based on the amount of Evans Blue dye extravasation, calculated as follows: Evans Blue extravasation (μg g^–1) = (Evans Blue concentration (μg ml^–1) × 1 ml)/tissue wet weight (g).

Intestinal permeability

For the assessment of intestinal permeability, the relatively impermeant macromolecule 4-kDa fluorescein isothiocyanate (FITC)-dextran (FD4; Sigma–Aldrich) was employed. Mice were subjected to a 4-h fasting period, both for food and water, prior to and following an oral gavage with a 50 mg ml^–1 FITC-dextran solution at a dose of 0.5 mg g^–1 body weight. After 3 h, blood samples were collected via the retro-orbital route, and the plasma fluorescence intensity was quantified using fluorescence plates with an excitation wavelength of 485 nm and an emission wavelength of 535 nm. The fluorescence measurements were compared to a standard curve constructed using FD4 in normal serum.

Kaede photoconversion

For in vivo tracing of intestinal immune cells, photoconversion was performed. Immediately preceding CLP or sham surgery, the small intestine of Kaede-transgenic mice were exposed to a 405-nm laser for a duration of 10 min, while ensuring the surrounding tissue was shielded from light using aluminum foil.

Flow cytometry analysis

The spleen and mesenteric lymph nodes were carefully homogenized using the plunger of a 1 ml syringe, and cells were rinsed from the strainer with 5 ml of PBS. Cells from the spleen, mesenteric lymph nodes, and blood were then subjected to red blood cell lysis. The small intestine was meticulously cleaned, opened lengthwise, and washed with PBS before being cut into 1.0–1.5 cm segments. These segments were then incubated twice in 20 ml of HBSS with 10 mM HEPES, 8% FBS, 4 mM EDTA, and 0.5 mM dithiothreitol at 37 °C with constant agitation for 15 min. After thoroughly removing any residual EDTA, the remaining pieces of the small intestine and lung were finely chopped and digested in RPMI-1640 medium containing fetal bovine serum, collagenase IV (0.1 mg/mL), and DNase I (0.1 mg/mL) at 37 °C for 40 min. The digested tissues were converted into a single-cell suspension through vigorous shaking and filtration through a 40-μm cell strainer following the removal of red blood cells. 1 × 10⁶ isolated cells were counted and resuspended in 100 µl 1% BSA in PBS blocked with 5 ng μl^–1 of anti-CD16/CD32 antibody (Cat# 553141, BD Biosciences) for 5 min at 4 °C, and then stained with the appropriate antibodies. The following antibodies were used for extracellular staining: fixable viability dye (Cat# 564406, BD Biosciences), CD45 (0.5 ng μl^–1, Cat# 553079, BD Biosciences), CD3 (0.5 ng μl^–1, Cat# 552774, BD Biosciences), TCR γδ (2 ng μl^–1, Cat# 118115, Biolegend), CD44 (4 ng μl^–1, Cat# 563736, BD Biosciences), Ly6C (0.6 ng μl^–1, Cat# 561237, BD Biosciences), IL-7R (2 ng μl^–1, Cat# 416-1271-82, Thermo Fisher Scientific), CD8 (0.6 ng μl^–1, Cat# 100766, Biolegend), CD11c (2 ng μl^–1, Cat# 564079, BD Biosciences), F4/80 (2 ng μl^–1, Cat# 565411, BD Biosciences), CD206 (5 ng μl^–1, Cat# 141721, Biolegend), CD86 (5 ng μl^–1, Cat# 558703, BD Biosciences), CD11b (2 ng μl^–1, Cat# 561688, BD Biosciences), Ly6G (2 ng μl^–1, Cat# 560601, BD Biosciences), CCR8 (3 ng μl^–1, Cat# 150312, Biolegend), CCR2 (5 ng μl^–1, Cat# 150605, Biolegend) and CXCR5 (5 ng μl^–1, Cat# 145504, Biolegend). For intracellular staining, cells were first stained for surface markers as detailed above, fixed and permeabilized, then the following antibodies were used: IL-17A (4 ng μl^–1, Cat# 506925, Biolegend), IFN-γ (4 ng μl^–1, Cat# 554411, BD Biosciences), Ki67 (2 ng μl^–1, Cat# 652411, Biolegend) and Siglec-F (0.5 ng μl^–1, Cat# 552125, BD Biosciences). Samples were analyzed on a BD LSRFortessa X-20 cytometer (BD Biosciences, San Jose, CA) or Beckman CytoFLEX cytometer (Beckman Coulter, Brea, CA). Data analysis was conducted using FlowJo 10.0 software (FlowJo software, Oregon, USA). Detailed reagent information is provided in the Supplementary Table 1.

MPO activity determination

MPO activity was assessed using an MPO kit (Cat# A044-1-1, Nanjing Jiancheng, China), following the manufacturer’s recommended procedures. In brief, brain and lung tissues were excised and accurately weighed, and subsequently homogenized in MPO assay buffer (1:19, wt vol^–1). The resultant homogenates were introduced into the reaction system. MPO activity within the samples was quantified by measuring absorbance at 460 nm and expressed as units per gram of total protein (U g^–1).

Immunofluorescence staining

The mice were anesthetized and subjected to transcardial perfusion using PBS, followed by 4% paraformaldehyde in PBS. To minimize nonspecific binding, normal goat serum was utilized for blocking. Immunoassays were conducted overnight at 4 °C with the following primary antibodies: Occludin (1:100, Cat# 27260-1-AP, Proteintech), ZO-1 (1:100, Cat# 21773-1-AP, Proteintech), MUC-2 (1:100, Cat# 27675-1-AP, Proteintech), and IL-17A (1:50, Cat# sc-374218, Santa Cruz). Primary antibody binding was visualized using Dylight 549-conjugated goat anti-rabbit (1:200, Cat# ab150116, Abcam) or Dylight 488-conjugated goat anti-mouse IgG secondary antibodies (1:200, Cat# ab150077, Abcam). Nuclei were counterstained with DAPI, and the slides were examined under a fluorescence microscope (BX51; Olympus Corporation, Tokyo, Japan). For the immunofluorescence detection of Nets, the Tyramide Signal Amplification method was employed, which enabled double staining using antibodies from the same species. In brief, tissue sections were initially incubated with a primary antibody against Cit-H3 (1:400, Cat# ab5103, Abcam) overnight at 4 °C. Subsequently, they were treated with Goat Anti-Rabbit IgG H&L (HRP) secondary antibody (1:4000, Cat# ab205718, Abcam) for 30 min. Each tissue sample was then covered with a CY5 tyramide working solution and incubated in darkness at room temperature for 10 min. Afterward, the slides were rinsed and immersed in TBST. The tissue sections were subsequently placed in eluent at 42 °C for 20 min. To block nonspecific binding, tissue samples were covered with 10% goat serum and incubated for 30 min at 37 °C. Similarly, sections were then incubated with a primary antibody against MPO (1:5000, Cat# ab208670, Abcam), followed by the addition of CY3 tyramide stock solution to the sections. Each tissue section was counterstained with DAPI. The immune-stained slides were scanned using a whole-slide fluorescence scanner (Case Viewer, 3D Histech, Budapest, Hungary). Image analyses were realized with the Image J software (Image J 1.8.0, NIH, USA).

Determination of cytokine levels

Levels of various cytokines in mouse serum and BALF (clarified by centrifugation) were determined either using the BD mouse Th1/Th2/Th17 cytokine cytometric bead array kit (for IL-6, IL-10, TNF-α and IL-17, Cat# 560485, BD Biosciences) by FCM analysis, or using commercial ELISA kits according to the manufacturer’s instructions (for IL-6 (Cat# EMC004.96, Neobioscience), TNF-α (Cat# EMC102a.96, Neobioscience) and IL-17 (Cat# EMC008.96, Neobioscience)).

Luminex assay

Unperfused lung tissue was gathered and immersed in lysis buffer consisting of PBS supplemented with 1% protease inhibitor cocktail and 1% phosphatase inhibitor cocktail. These tissues were subsequently subjected to homogenization, sonication, and centrifugation at 10,000 × g for a duration of 15 min. Following centrifugation, lysate samples were standardized based on their total protein concentration, which was determined through a BCA assay. Bio-Plex Pro mouse chemokine panel 31-plex (Cat# 12009159, Bio-rad) was used for the Luminex assay.

Chemotaxis assay

Spleen γδ T cells (5 × 10⁴) were magnetically sorted (Cat# 130-092-125, Miltenyi) and placed in the upper chamber of Transwell tissue culture inserts (5.0-mm pore diameter, Corning) under sterile conditions. These inserts were positioned over individual wells of a 96-well plate containing either apyrogenic BSA (2 mg ml^–1, Cat# 23209, Thermo Scientific) or recombinant mouse (rm) CCL1 (10 ng ml^–1, Cat# 845-TC, R&D Systems). Incubation occurred for 3 h at 37 °C with 5% CO₂. Transmigrated cells were collected from the lower chambers and subjected to FCM analysis. Spleen mononuclear cells were isolated, washed, and resuspended in PBS to generate a single-cell suspension, followed by cell counting using a hemocytometer. Apyrogenic BSA, rmCCL1 (10 ng ml^–1), and total protein extracts from the lungs of sham-operated or CLP mice (2 mg ml^–1) were plated in the lower compartment of 5-μm-pore Transwell chambers. Spleen mononuclear cells were seeded in the upper compartment of Transwell chambers in PBS containing 1% FBS. After a 6-h incubation at 37 °C, migrated cells in the lower compartment were collected for FCM analysis.

Tandem mass tag labeling (TMT) quantitative proteomics analysis

Sample preparation

Mice from the CLP group and sham-operated group (n = 3 each) were chosen for proteomic analysis. The mice were anesthetized and the lung tissues were immediately removed, then snap-frozen in liquid nitrogen. The protein sample to be tested was weighed and added to the lysis buffer containing SDS L3 and EDTA. The mixture was then ground (25,000 × g, 4 °C, 5 min). Next, the mixture was supplemented with 10 mM DTT (final concentration 10 mM) and IAM (final concentration 45 mM) and incubated away from light for 45 min. The proteins were precipitated with cold acetone, followed by centrifugation (25,000 × g, 4 °C, 15 min). The pellet was air-dried, and an appropriate amount of SDS L3-free lysate was added to the proteins, ensuring thorough solubilization. After another centrifugation step (25,000 × g, 4 °C, 15 min), the extracted proteins were quantified using Bradford’s method. Subsequently, the proteins were enzymatically cleaved using Trypsin, followed by a desalting procedure. The resulting peptides were freeze-dried, and the peptide solution was subjected to separation using a Shimadzu LC-20AB liquid phase system.

High-performance liquid chromatography and mass spectrometry

The acquired peptides were separated by an Easy nLC 1200 system, employing a self-assembled C18 column for effective gradient separation. The separated peptides were then analyzed by tandem mass spectrometry in Data Dependent Acquisition mode. The TMT-proteomic analysis was conducted by BGI-Shenzhen, Shenzhen, China.

Bioinformatics analysis

The protein sequences of proteins exhibiting differential expression were retrieved in batches for the purpose of Gene Ontology (GO) mapping and annotation from the UniProtKB database (Release 2016_10) in FASTA format. The FASTA sequences of proteins that displayed significant changes in expression were subjected to a sequence alignment with the online Kyoto Encyclopedia of Genes and Genomes (KEGG) database (http://geneontology.org/) to identify their respective KEGG Orthology (KO) identifiers and were subsequently associated with specific pathways within KEGG12. The relevant KEGG pathways were then extracted and the acquired data on protein expression levels was visually represented as a hierarchical clustering heatmap.

Single-cell RNA sequencing (scRNA-seq)

Single-cell suspension preparation

Small intestines were harvested from euthanized mice. The intestines were carefully cleaned and opened longitudinally, washed with PBS, cut into 1.0–1.5 cm pieces, and incubated in 20 ml of HBSS/10 mM HEPES, 8% FBS, 4 mM EDTA, and 0.5 mM dithiothreitol at 37 °C with constant agitation for 15 min twice. After washing with Ca²⁺/Mg²⁺-PBS to remove EDTA, the remaining pieces of small intestines were minced and digested in RPMI-1640 medium containing FBS, collagenase IV (BioFroxx, 0.1 mg ml^–1), and DNase I (BioFroxx, 0.1 mg ml^–1) at 37 °C for 40 min. The tissue digestion was filtered through a 40 μm cell strainer, following by centrifugation at 400 × g for 5 min at 4 °C. Cells were collected and resuspended in PBS. CD45⁺ cells (as detected by an anti-CD45 antibody) were flow-sorted to isolate small intestinal immune cells. CD45⁺ cells were then stained using 0.4% trypan blue to evaluate cell counts and viability. The cell concentration was adjusted to 1000 cells per μl for library preparation.

Single-cell RNA sequencing

Prepared single-cell suspensions were loaded on a chip of the Single-Cell 3′ Chip Kit (10X Genomics) and subjected to a Chromium Single-Cell Instrument (10X Genomics) to generate single-cell gel beads in emulsions (GEMs). cDNA amplification and library construction underwent using a Chromium Single-Cell 3′ Reagent Kits v2 (10X Genomics). Libraries were sequenced on a BGISEQ-500 sequencer with the following strategy: 10 bp of barcodes, 28 bp of read-1, 100 bp of read-2, 2 lanes. The scRNA-seq was conducted by BGI-Shenzhen, Shenzhen, China.

scRNA-seq data processing

Sequencing data were aligned to the Macaca fascicularis 5.0 genome by STAR (v2.7.4a). Counts of aligned UMI were used to quantify downstream gene expression levels. For scRNA-seq data, cells with less than 200 genes and the top 10% of cells with the highest number of genes were removed. Additionally, cells with more than 15% mitochondrion reads were filtered out. DoubletFinder was used to remove potential doublets.

Unsupervised cell clustering

Seurat package (v3.2) in the R environment (v4.0.3) was used to conduct variable features screening, unsupervised clustering, and marker gene identification. First in each sample, genes detected in less than 5 cells were filtered out from downstream analysis, then NormalizeData and FindVariableFeatures (selection. method = “vst”, nfeatures = 2000) functions were performed. A set of 2000 most variable genes were selected for principal component analysis (PCA). FindIntegrationAnchors and IntegrateData functions were used to emerge data from two groups (CLP and Sham) as an integrated data assay. Cell clusters were identified via FindNeighbors and FindClusters functions with the top 20 principal components (resolution = 0.5). The uniform manifold approximation and projection (UMAP) was used to present the distance between cells in a two-dimensional panel. Finally, the FindAllMarkers function (only.pos = T, min.pct = 0.25, logfc.threshold = 0.25) was adopted to find all differentially expressed genes (DEGs).

Cell type annotation

Clusters were annotated according to the CellMarker database and marker genes of specific cell types reported in research. B cells (Cd19 and Mzb1), T cells (Cd3g and Cd5), dendritic cells (Cst3, Gzma and Ccr7), macrophages (C1qb, Cx3cr1, and Csf1r), neutrophil (S100a8 and S100a9), ILC1 (Il21r and Ccl4), ILC2 (Gata3), ILC3 (Il22), epithelium (Krt8 and Epcam), and fibroblast (Col1a1 and Adamdec1). γδT cells were identified according to the high expression of Trdc and Rorc.

Pseudotime trajectory analysis

Monocle2 R package was used to clarify the cell lineage trajectory of γδ T cells. CellDataSet (CDs) with UMI counts, phenoData, and featureData were constructed for downstream analysis. DifferentialGeneTest function was performed to detect DEGs between γδ T cell subclusters. Significant genes with a q value < 0.1 were used to order γδ T cells in pseudotime trajectory analysis. After the trajectory was constructed, the discriminative dimensionality reduction with trees (DDRTree) was used to visualize data in the two-dimensional panel.

Chromatin immunoprecipitation assay with sequencing (ChIP-seq) analysis

MH-S cells were cultured in 15-cm Petri dishes and subjected to the respective treatments upon reaching 60% confluence. After 24 h, the cells were harvested and DNA purification was carried out following the manufacturer’s instructions. High-throughput DNA sequencing libraries were prepared using the ThruPLEX DNA-seq kit. A range of 200–500 BPS libraries was enriched, quantified, and sequenced using a HiSeq 2500 sequencer (Illumina). The antibody was used for ChIP analysis from Cell Signaling Technology (anti-Lef1, Cat# 76010). The ChiP-seq analysis was conducted by Gene Read, Wuhan, China.

Co-culture of alveolar macrophages (AMs) and γδ T cells

γδ T cells were purified from splenic T cells via magnetic bead cell sorting (Milteny Biotec). mAMs isolated from both sham and CLP mice were co-cultured with purified γδ T cells from naive WT mice for 24 h (2.5 × 10⁵ mAMs and 2.5 × 10⁵ γδ T cells per well in 24-well plates pre-coated with 10 μg ml^–1 of TCR-γδ antibody (GL4, Cat# 553180, BD Biosciences) in RPMI-1640 supplemented with 10% FBS, 2 mM L-glutamine, 30 μM β-mercaptoethanol, non-essential amino acids (NEAAs), 1 mM sodium pyruvate, and penicillin/streptomycin (Punosai Life Sciences Ltd)). Subsequently, cells were washed with fresh medium and resuspended in RPMI-1640 containing 10% FBS, 100 ng ml^–1 PMA, 1 μg ml^–1 ionomycin, and 3 μg ml^–1 brefeldin (Sigma–Aldrich) and incubated for 4 h at 37 °C. Cells were then washed and stained with IL-17A (4 ng μl^–1, Cat# 506925, Biolegend) and IFN-γ (4 ng μl^–1, Cat# 554411, BD Biosciences) antibodies, and the percentage of IFN-γ⁺ and IL-17A⁺ cells in the γδ T cell population was determined.

Protein modeling and molecular docking

The structural formula of S-KT was obtained from the PubChem database. The crystal structure of the FZD5 protein was retrieved from the Protein Data Bank (PDB ID: 6WW2). The protein structure of Wnt5a was predicted using AlphaFold2⁴². Molecular docking analyses were conducted utilizing Discovery Studio software (version 19.1.0). The resulting models of FZD5 and Wnt5a were visualized, and figures were generated using the PyMOL Molecular Graphics System (version 4.6).

Statistical analysis

Data were presented as mean ± standard deviation (Mean ± SD) and analyzed using Prism 9 software (GraphPad). Heat maps in Fig. 5 were constructed and visualized using ChiPlot (https://www.chiplot.online/, accessed on 23 January 2023). Normality distribution of the data was determined by using D’Agostino & Pearson test. One-way analysis of variance (ANOVA) followed by post-hoc Tukey’s multiple-comparison test or Kruskal–Wallis with Dunn’s multiple-comparison tests were used for multiple-group comparisons. Statistical significance for the murine body weight changes was analyzed with two-way ANOVA. Unpaired Student’s t-test or Mann–Whitney test was employed for comparisons between the two groups. Survival curves were generated using the Kaplan–Meier method, and differences among groups were compared using the Log-rank (Mantel–Cox) test to assess the significance of the survival differences. A p-value of less than 0.05 was considered statistically significant, and ^*P < 0.05, ^**P < 0.01, ^***P < 0.001, ^****P < 0.0001.

Reporting summary

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