A pro-inflammatory neutrophil subpopulation drives intestinal ischemia–reperfusion injury via the ATF4-mediated endoplasmic reticulum stress pathway

Abstract

This study investigated the role of neutrophils in intestinal ischemia–reperfusion injury (IRI) in mice. We combined single-cell RNA sequencing (scRNA-seq) with in vivo and in vitro functional assays to characterize cellular dynamics. Single-cell RNA sequencing of IRI model tissue revealed a significant increase in neutrophils and inflammatory monocytes, alongside a decrease in T cells, B cells, and NK cells. In vivo neutrophil depletion markedly alleviated intestinal damage, as indicated by reduced serum diamine oxidase (DAO) and IL-6 levels, improved histopathological scores, and preserved Occludin protein integrity. Mechanistically, scRNA-seq identified a pro-inflammatory neutrophil subcluster (C5) characterized by enrichment of endoplasmic reticulum stress (ERS) markers, particularly the transcription factor ATF4. In vitro and in vivo studies confirmed that neutrophils exacerbate IRI severity by inducing ERS via the ATF4 pathway. Pharmacological inhibition of ERS or genetic ablation of ATF4 significantly attenuated neutrophil-driven inflammation and mucosal injury. These findings demonstrate that a specific neutrophil subpopulation aggravates intestinal IRI through the intrinsic ERS/ATF4 pathway, providing a novel perspective on IRI pathophysiology and highlighting a potential therapeutic target for mitigating intestinal damage.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-36938-9.

Introduction

The intestinal mucosal barrier is a critical defense system that prevents the translocation of harmful microorganisms and toxins into the body^1–4. It comprises physical, chemical, and immunological components, including mucus, antimicrobial peptides, and the mucosal immune system, which collectively maintain gut homeostasis^5–7. The interplay between this barrier, the microbiota, and immune cells is essential for gastrointestinal integrity^8–10.

Intestinal ischemia–reperfusion injury (IRI) represents a severe clinical challenge, often arising from shock, severe burns, or trauma^11–13. During IRI, the restoration of blood flow paradoxically exacerbates tissue damage, leading to a cascade of pro-inflammatory signaling, reactive oxygen species production, and intestinal epithelial cell death^14–16. This compromises barrier function, potentially resulting in systemic inflammation, sepsis, and multi-organ failure^17–19. Despite its clinical significance, effective therapeutic strategies for IRI remain limited.

Single-cell RNA sequencing (scRNA-seq) has emerged as a transformative technology for elucidating disease mechanisms^20–22. Unlike bulk sequencing, scRNA-seq enables the dissection of cellular heterogeneity by profiling gene expression in individual cells, providing unprecedented insights into cell types and states during disease progression^23–27.

Neutrophils are recognized as pivotal mediators in the early phase of IRI, where they exacerbate inflammation, increase intestinal permeability, and promote epithelial cell death. Neutrophils are rapidly recruited to the injury site and can be broadly categorized into pro-inflammatory (N1-like) and pro-resolving (N2-like) subtypes, with distinct kinetics and functions. In IRI, the predominant N1-like subset releases reactive oxygen species, proteases, and neutrophil extracellular traps, directly damaging the intestinal epithelium and amplifying the inflammatory cascade^28–30. However, neutrophils are now understood to be a heterogeneous population with context-dependent functions, where specific subsets may contribute to tissue damage while others facilitate repair. The precise mechanisms and diversity of neutrophil responses in intestinal IRI remain incompletely defined.

Additionally, endoplasmic reticulum (ER) stress has been implicated in the pathogenesis of IRI across various organs. The unfolded protein response (UPR), initiated by sensors like PERK, IRE1α, and ATF6, aims to restore ER homeostasis but can trigger apoptosis and inflammation under sustained stress. In intestinal IRI, ER stress contributes to epithelial cell death, barrier dysfunction, and immune cell activation^31,32. The interplay between neutrophil-driven inflammation and ER stress signaling in intestinal IRI, however, is poorly understood.

In this study, we employed scRNA-seq to profile the cellular landscape of the mouse small intestine following IRI. We focused on characterizing neutrophil heterogeneity and dynamics to identify specific subpopulations and mechanisms that drive mucosal damage. We anticipate that a deeper understanding of these processes will reveal novel therapeutic targets for mitigating intestinal IRI.

Methods

The Methods section has been reorganized into clearly defined subsections corresponding to major experimental approaches. Each subsection states the objective, experimental groups, and detailed protocols to enhance clarity and reproducibility.

Experimental design and animal models

Objective

To establish a murine model of intestinal ischemia–reperfusion injury (IRI) and to delineate the roles of neutrophils and endoplasmic reticulum (ER) stress through pharmacological modulation and genetic knockout approaches.

Experimental mice and groups

All animal procedures were approved by the Institutional Animal Care and Use Committee (IACUC) of Jiangsu Huitong Test & Evaluation Technology Group Co., Ltd (Approval Reference Number: HT-2025-LWFB-0088). Male C57BL/6 mice (6–8 weeks old, 20–22 g) were used, including wild-type and ATF4 knockout (C57BL/6JCya-ATF4em1/Cya) strains. Mice were purchased from Jiangsu Huitong Test & Evaluation Technology Group Co., Ltd. Mice were housed under specific pathogen-free conditions with a 12-h light/dark cycle and provided food and water ad libitum. After a one-week acclimatization period, they were randomly assigned to experimental groups (n = 10 per group).

The study included the following experimental groups for in vivo investigations

Sham Group: Underwent surgical procedure without superior mesenteric artery (SMA) occlusion.

IR Group: Underwent 45 min SMA occlusion followed by 240 min reperfusion.

IR + αLy6G Group: IR mice pre-treated with anti-Ly6G antibody for neutrophil depletion.

IR + Tunicamycin Group: IR mice pre-treated with the ER stress activator tunicamycin.

IR + 4-PBA Group: IR mice pre-treated with the ER stress inhibitor 4-PBA.

ATF4 KO + IR Group: ATF4 knockout mice subjected to IR.

ATF4 KO + IR + Tunicamycin Group: ATF4 knockout mice pre-treated with tunicamycin and subjected to IR.

To modulate endoplasmic reticulum stress, mice received intraperitoneal injections of tunicamycin (25 µg/mouse in 150 mM glucose) or 4-PBA (3 mg/mouse) 24 h prior to ischemic reperfusion modeling. All methods were performed in accordance with the relevant guidelines and regulations, including the ARRIVE guidelines for reporting experiments on live vertebrates.

Intestinal ischemia–reperfusion model

Mice were fasted for 12 h prior to the procedure. Under anesthesia, a midline laparotomy was performed. The superior mesenteric artery (SMA) was identified, carefully separated from surrounding tissue and the mesenteric vein, and occluded with a non-traumatic microvascular clip for 45 min to induce ischemia. Ischemia was confirmed by visual pallor of the small intestine. Reperfusion was initiated by clip removal and confirmed by the return of bowel color and pulsatile flow, lasting for 240 min. The durations of ischemia (45 min) and reperfusion (240 min) were selected based on established protocols that reliably induce significant yet sub-lethal intestinal injury, allowing for the study of inflammatory and cellular responses³³. Sham-operated control animals underwent the same surgical procedure without SMA occlusion. At the end of the reperfusion period, mice were euthanized for collection of serum and small intestinal tissues.

Single-cell RNA sequencing analysis

Objective: To characterize the transcriptional profiles and cellular heterogeneity within the small intestine in response to IRI using single-cell RNA sequencing.

Single-cell suspension preparation

Small intestines were harvested from C57BL/6 mice following IR modeling or sham operation. Tissues were washed in ice-cold RPMI-1640 and digested enzymatically with a cocktail of collagenase II, collagenase IV, and DNase I (all from Absin) [Reference for digestion protocol: Bomidi C, et al. Proc Natl Acad Sci U S A. 2021;118(9):e2025416118]. After lysing erythrocytes (Biosharp, BL503A), cell count and viability were assessed using a fluorescence cell analyzer (Countstar® Rigel S2) with AO/PI staining. Dead cells were removed (Miltenyi Biotec, 130–090-101), and viable cells were washed and resuspended in PBS containing 0.04% BSA at a density of 1 × 10^6 cells/mL.

Sequencing data processing and analysis

Libraries were sequenced on an Illumina NovaSeq 6000 platform. The unique molecular identifier (UMI) count matrix from 72,512 initial cells was processed using the Seurat package (v4.3.0) in R³⁴. Cells were filtered based on the following criteria: < 1,000 UMIs, < 500 detected genes, or > 25% mitochondrial UMIs. Genes detected in fewer than 10 cells were also removed. This quality control process resulted in a final dataset of 26,250 high-quality cells from two intestinal injury and two control samples.

Following quality control, the UMI count matrix was log-normalized, and the top 2,000 variable features were used for dataset integration. Integration was performed using the Seurat (v4.3.0) pipeline, which identifies integration anchors and generates an integrated data matrix. Principal component analysis (PCA) was conducted for dimensionality reduction, and the top 50 principal components were selected for downstream analysis. Cell clustering was performed using the FindClusters function at a resolution of 1.6, yielding 48 initial clusters. These clusters were visualized using UMAP. Cell types were annotated by first identifying cluster-specific marker genes with the FindAllMarkers function (logfc.threshold > 0.5, min.pct > 0.25, adjusted p-value < 0.05). These markers were then cross-referenced with an expanded set of published markers for mouse intestinal injury using the ScType automated annotation tool³⁵.

Differential expression and advanced analyses

Differentially expressed genes (DEGs) were identified using the FindAllMarkers function in Seurat, applying a Wilcoxon rank-sum test with Bonferroni correction. DEGs were defined by the following thresholds: |avg_log2FC|> 0.5, min.pct > 0.2, and adjusted p-value < 0.05.

We performed integrated analysis on the pre-processed Seurat object. Pseudotemporal ordering and branch trajectory analysis were conducted using Monocle3³⁶, with results visualized via UMAP. Genes differentially regulated along the trajectories were identified using the built-in Monocle3 toolkit. Subsequently, cell–cell communication analysis was performed using CellChat³⁷.

Neutrophil depletion and isolation

Objective: To assess the functional contribution of neutrophils to intestinal IRI pathogenesis through depletion and isolation studies.

Neutrophil depletion model

C57BL/6 mice intended for neutrophil depletion were intraperitoneally injected with 500 µg/20 g of LY6G-1A8 monoclonal antibody (eBioscience, San Diego, CA, USA) 12 h before establishing the model. Control mice for depletion experiments received an equivalent volume of isotype control antibody (Rat IgG2a, κ, eBioscience). Following injection, the mice were housed under standard conditions in the animal facility and allowed to rest until the start of the experiment.

Neutrophil isolation

Neutrophils were isolated from the peripheral blood of C57BL/6 mice using a Neutrophil Isolation Kit (SABIO-ROCHE, Beijing, China) according to the manufacturer’s protocol. Neutrophils were cultured in RPMI 1640 medium in 6-well plates.

Cell culture and co-culture experiments

Objective: To examine the direct cytopathic effects and molecular crosstalk between neutrophils and intestinal epithelial cells in a co-culture system.

Cell culture

MODE-K mouse intestinal epithelial cells (Fuheng Biotechnology) were maintained in RPMI 1640 medium supplemented with 10% FBS at 37 °C under 5% CO2. To establish intestinal monolayer barriers, cells were seeded at a density of 1 × 10^5 cells per well on 3-μm polyester membrane Transwell inserts (Corning, 723,021) in 6-well plates. The apical and basolateral chambers contained 1 mL and 1.5 mL of medium, respectively. The medium was replaced daily until confluent monolayers were formed (typically 5–7 days).

Co-culture and treatments

For co-culture experiments, MODE-K cells were co-cultured with isolated neutrophils (1:5 epithelial cell to neutrophil ratio) from control or IR mice in the presence or absence of pharmacological modulators. The ER stress activator tunicamycin (GLPBIO) was used at 1 µg/mL, and the inhibitor 4-PBA (Bioss) was used at 3 mM, based on preliminary dose–response experiments and prior literature³⁸.

Biochemical and molecular assays

Objective: To quantify injury biomarkers, inflammatory responses, and activation of relevant molecular pathways in vivo and in vitro.

ELISA for serum markers

Serum IL-6 and diamine oxidase (DAO) levels were measured using ELISA kits (Shanghai Jianglai Industrial Co., Ltd.) following the manufacturer’s instructions. IL-6 was selected as a representative early and central pro-inflammatory cytokine in IRI, reflecting neutrophil-mediated inflammation. While other cytokines (e.g., TNF-α, IL-1β) are involved, IL-6 serves as a reliable and sensitive indicator of the inflammatory cascade in our model³⁹. Optical density (OD) was measured at 450 nm using a microplate reader (BioTek Synergy H1).

Histological analysis and chiu’s scoring

Small intestinal tissues from C57BL/6 mice were fixed in 4% paraformaldehyde, embedded in paraffin, and sectioned at 4 μm. Sections were stained with hematoxylin and eosin (H&E) following deparaffinization and rehydration. Histopathological damage was assessed in a blinded manner using Chiu’s scoring system⁴⁰. Briefly, the scoring system (0–5) evaluates: 0 = normal mucosa; 1 = development of subepithelial space; 2 = extension of space with moderate epithelial lifting; 3 = massive epithelial lifting with denuded villi tips; 4 = denuded villi with exposed lamina propria; 5 = disintegration of lamina propria, hemorrhage, and ulceration. Scores from at least 10 randomly selected fields per sample were averaged.

Immunofluorescence staining

Paraffin-embedded small intestine sections were deparaffinized, rehydrated, and subjected to antigen retrieval (if applicable). After blocking with 1% BSA and 0.1% Triton X-100 for 10 min, sections were incubated overnight at 4 °C with the following primary antibodies: anti-Ly6G (Thermo Fisher, PA5-84,280; 1:1000), anti-ATF4 (Proteintech, 10,835–1-AP; 1:500), anti-CD45 (Affinity, AF3904; 1:500), anti-CD3 (Abcam, ab186735; 1:500), and anti-CD11b (Abcam, ab133504; 1:500). Following PBS washes, sections were incubated with fluorophore-conjugated secondary antibodies and counterstained with DAPI. Images were acquired using a fluorescence microscope (Olympus).

Western blot

Small intestinal tissues or cell pellets were homogenized in RIPA lysis buffer containing protease inhibitors on ice. Lysates were centrifuged at 12,000 × g for 10 min at 4 °C, and the supernatant protein concentration was determined by BCA assay (Beyotime). Equal amounts of protein were separated by SDS-PAGE and transferred to nitrocellulose membranes. In accordance with journal policy regarding digital image integrity, all Western blot membranes were processed intact without being cut prior to antibody hybridization. The original, uncropped images of all full-length blots used in this study, showing the membrane edges and molecular weight markers, are provided in the Supplementary Information file (Supplementary File). After blocking (Servicebio), membranes were incubated with primary antibodies against ATF4 (Proteintech, 10,835–1-AP; 1:1000), GRP78 (CST, 3177 T; 1:1000), Occludin (CST, 91131 T; 1:1000), and β-actin (CST, 4970 T; 1:3000). HRP-conjugated secondary antibodies were used for detection with a chemiluminescent substrate (Odyssey CLX). Band intensities were quantified using ImageJ software and normalized to β-actin.

Cell viability and apoptosis assays

Objective: To evaluate neutrophil-induced cytotoxicity, specifically focusing on epithelial cell viability and apoptosis.

Cell viability (CCK-8) assay

MODE-K cell viability was determined using the Cell Counting Kit-8 (CCK-8; Beyotime) according to the manufacturer’s instructions. Briefly, 5 × 10^3 MODE-K cells per well were seeded in 96-well plates. After treatments (co-culture, tunicamycin 1 µg/mL, or 4-PBA 3 mM), 10 μL of CCK-8 solution was added to each well. After a 2-h incubation, the absorbance at 450 nm was measured using a microplate reader (BioTek, Synergy H1). Viability was expressed as a percentage relative to control cells.

TUNEL apoptosis assay

Apoptosis in MODE-K cell paraffin sections was detected using a TUNEL assay kit (Meliunbio, MA0223) according to the manufacturer’s protocol. Following deparaffinization and rehydration, sections were digested with proteinase K (20 μg/mL) for 30 min at 37 °C. After PBS washes, sections were incubated with the TUNEL reaction mixture, followed by counterstaining with DAPI. Fluorescent images were acquired using an Olympus microscope. Apoptosis was quantified by determining the percentage of TUNEL-positive cells relative to the total number of DAPI-positive nuclei.

Transmission electron microscopy (TEM)

MODE-K cells were fixed in 2.5% glutaraldehyde and post-fixed in 2% osmium tetroxide. The samples were then dehydrated through a graded ethanol series, embedded in epoxy resin, and sectioned into 70-nm ultrathin slices. Sections were stained with uranyl acetate and lead citrate and imaged using an FEI Teneo VS transmission electron microscope.

Anesthesia and euthanasia

Mice were anesthetized using isoflurane (2%-3% concentration) via inhalation prior to surgical procedures. At the end of the study, euthanasia was performed via CO ~ 2 ~ inhalation followed by cervical dislocation to ensure humane endpoints in accordance with the AVMA Guidelines for the Euthanasia of Animals (2020).

Statistical analysis

All statistical analyses were performed using GraphPad Prism 9.3.1. Data are presented as mean ± SD or median with interquartile range, as appropriate. Normality and homogeneity of variance were assessed using the Shapiro–Wilk and Levene’s tests, respectively. For comparisons involving one, two, or three factors, one-, two-, or three-way ANOVA was applied, followed by Tukey’s post-hoc test. A p-value of less than 0.05 was considered statistically significant (*p < 0.05, **p < 0.01, ***p < 0.001).

Results

Overview of intestinal cell landscapes

Single-cell RNA sequencing of 72,512 cells from the small intestines of control and ischemia–reperfusion (IR) mice revealed distinct shifts in cellular composition following injury.Clustering identified 48 transcriptionally distinct subpopulations, which were annotated into 19 major cell types (Fig. 1A, E).

Fig. 1.

Single-cell RNA sequencing reveals a comprehensive atlas of intestinal cells. (A) UMAP visualization of cells between two groups and across different cell clusters. The x- and y-axes denote the first and second principal components of the gene-expression matrix across all cells, respectively. Each dot represents a single cell; colors indicate different samples (left panel) or distinct cell clusters (right panel). (B) Changes in cell cluster proportions between different groups. The x-axis shows the cell clusters; the y-axis shows the log2-transformed ratio of each cluster’s relative abundance between the two experimental groups. Bars in the upper-right quadrant indicate increased relative proportion, while those in the lower-left quadrant indicate decreased proportion. Greater fold-changes yield taller bars with darker colors. (C-E) UMAP of clustering trajectories for two groups obtained from pseudotime analysis. The cell-annotation results are displayed: each dot represents a single cell, and different colors denote distinct cell types. (F) Marker genes for each cell type. The x-axis displays canonical marker genes, and the y-axis shows cell clusters. The color intensity of each dot indicates the average expression level, while the dot size represents the proportion of cells within that cluster expressing the gene. (G) GO pathway analysis of top 3 marker genes in each cell cluster. For each cluster, the three most enriched GO biological-process terms among its marker genes were selected. The corresponding enrichment q-values for these terms were computed per cluster, column-wise z-scored, and visualized as a heatmap; within-column color intensity thus reflects the significance of the enriched pathways for that cluster’s markers. (H) KEGG pathway analysis of top 3 marker genes in each cell cluster. For each cell cluster, the three most enriched KEGG pathways among its marker genes were first selected. The enrichment q-values of these pathways were then calculated for every cluster, column-wise z-scored, and displayed as a heatmap. Within-column color intensity allows direct comparison of how significantly the marker genes of that cluster are enriched in each pathway. (I) Cell communication analysis. The network diagram illustrates the number/strength of intercellular communications, where nodes represent cell populations and edge width or color intensity corresponds to the quantity or intensity of signaling interactions between them. Data are presented as mean ± SD. Statistical significance is indicated as *P < 0.05, **P < 0.01, ***P < 0.001, **P < 0.0001. NC, normal control; IR, ischemia–reperfusion; Tun, tunicamycin; KO, knockout; neu, neutrophils.

IR induced a significant reduction in the median number of expressed genes per cell (1,368) compared to controls (1,716). Proportional analysis demonstrated a marked increase in neutrophils, inflammatory monocytes, stem cells, and enteroendocrine cells, alongside a decrease in T cells, B cells, and NK cells in the IR group (Fig. 1B).

Pseudotime analysis delineated divergent differentiation trajectories between groups (Fig. 1C), supported by cluster-specific marker gene expression (Fig. 1D, F) and pathway enrichment of these markers (Figs. 1G, H). Cell–cell communication analysis revealed enhanced interaction networks involving neutrophils, inflammatory monocytes, endocrine, and endothelial cells in IR mice, while signaling from tuft and other intestinal populations was reduced (Fig. 1I). This implicates neutrophils as potential key mediators in the intestinal IR response.

Neutrophil-mediated enhancement of intestinal I/R injury

To functionally validate the role of neutrophils inferred from scRNA-seq, we employed a neutrophil depletion model in mice subjected to intestinal I/R. Depletion of neutrophils significantly attenuated I/R-induced injury, as evidenced by reduced serum levels of the intestinal damage marker DAO and the inflammatory cytokine IL-6 compared to the non-depleted IR group (Figs. 2A, 2B).

Fig. 2.

Neutrophils contribute to ischemia–reperfusion injury of the small intestinal mucosa in mice. (A) The expression level of serum diamine oxidase (DAO) in mice was determined using an ELISA kit. (B) Serum Interleukin-6 (IL-6) expression in mice was assessed by ELISA. (C, D) Quantification of small intestinal length in mice. (E, F) Hematoxylin and eosin (H&E) staining of mouse small intestine and histological evaluation using Chiu’s scoring system. (G-I) Immunofluorescence staining of mouse small intestine for Ly6G and ATF4 expression. (J, K) Expression level of intestinal mucosal barrier protein Occludin in mouse small intestine. Data are presented as mean ± SD. Statistical significance is indicated as *P < 0.05, **P < 0.01, ***P < 0.001, **P < 0.0001. NC, normal control; IR, ischemia–reperfusion; Tun, tunicamycin; KO, knockout; neu, neutrophils.

Morphological and histological assessments further confirmed this protective effect. Neutrophil depletion mitigated the I/R-induced shortening of the small intestine (Figs. 2C, 2D) and ameliorated histopathological damage, resulting in significantly lower Chiu’s scores (Figs. 2E, 2F).

Immunofluorescence analysis revealed a substantial infiltration of Ly6G + neutrophils and other immune cells (CD45 + , CD3 + , CD11b +) in the intestinal tissue following I/R, which was markedly reduced upon neutrophil depletion (Figs. 2G-I, S1).

Consistent with the preserved tissue integrity, the expression of the tight junction protein occludin was significantly higher in the depletion group than in the IR group (Figs. 2J, 2K).

Cellular mechanisms of neutrophil-mediated mucosal injury

To delineate the direct cytotoxic effects of neutrophils, we co-cultured MODE-K intestinal epithelial cells with neutrophils isolated from control (NC neu) or I/R (IR neu) mice.

Co-culture with IR neu neutrophils significantly amplified the inflammatory response, elevating IL-6 secretion (Fig. 3A) and reducing MODE-K cell viability (Fig. 3B) compared to NC neu co-culture. This cytotoxicity was further evidenced by a marked increase in TUNEL-positive apoptotic cells (Figs. 3D, 3E) and a substantial loss of the tight junction protein occludin (Figs. 3F, G) in the IR neu group.

Fig. 3.

Neutrophils induce damage to MODE-K cells. (A) IL-6 expression in co-culture supernatant was quantified using an ELISA kit. (B) Cell viability of MODE-K cells was assessed using the CCK-8 assay. (C) Transmission electron microscopy (TEM) of MODE-K cells at magnifications of × 2.0 K, × 6.0 K, and × 12.0 K. (D, E) Apoptosis in MODE-K cells was detected by TUNEL staining. (F, G) Expression of the tight junction protein Occludin in MODE-K cells was detected by Western blot. Data are presented as mean ± SD. Statistical significance is indicated as *P < 0.05, **P < 0.01, ***P < 0.001, **P < 0.0001. NC, normal control; IR, ischemia–reperfusion; Tun, tunicamycin; KO, knockout; neu, neutrophils.

Transmission electron microscopy revealed severe ultrastructural damage in MODE-K cells co-cultured with IR neu neutrophils, characterized by mitochondrial outer membrane dissolution, cristae breakdown, and endoplasmic reticulum degradation (Fig. 3C). In contrast, cells co-cultured with NC neu neutrophils maintained normal organellar architecture.

Neutrophil heterogeneity and a pro-inflammatory ATF4 + subpopulation in I/R

Given the central role of neutrophils in I/R injury, we dissected their heterogeneity through subclustering, identifying six distinct subsets (C0-C5) (Fig. 4A). Differential expression analysis of these neutrophils revealed enrichment in inflammatory and chemotactic pathways (Fig. 4B).

Fig. 4.

Single-cell RNA sequencing enables in-depth profiling of neutrophil subpopulations. (A) Distribution proportions of different cell populations in Sham and IR groups. (B) GO analysis of differentially expressed genes in neutrophils. (C) Clustering analysis (UMAP) of neutrophil subpopulations. (D) Top 5 marker genes of neutrophil subpopulations. (E) Immune pathway analysis of neutrophil subpopulations. (F) Pseudotime analysis of neutrophils and distribution of neutrophil subpopulations along pseudotime. (G) Differentially expressed transcription factors in C5 subgroup. (H) GO analysis of differentially expressed genes in C5 subgroup cells. (I) ATF4-target gene co-expression network. Data are presented as mean ± SD. Statistical significance is indicated as *P < 0.05, **P < 0.01, ***P < 0.001, **P < 0.0001. NC, normal control; IR, ischemia–reperfusion; Tun, tunicamycin; KO, knockout; neu, neutrophils.

Pseudotime analysis reconstructed a continuous differentiation trajectory, positioning the C5 subpopulation at the start and the C1/C2 subpopulations at the endpoint (Figs. 4C, F). This trajectory was associated with dynamic shifts in the expression of key marker genes across subsets (Figs. 4D, E, S2). Functional annotation suggested distinct temporal roles: C1 in early immune response, C4 in mid-phase, and C3 in later stages, while C2 was associated with B-cell interaction.

Notably, the C5 subpopulation, characterized by high expression of inflammatory and immune genes, represented recently recruited neutrophils and was implicated in driving mucosal injury.

We therefore focused on C5 and constructed a transcription factor (TF) regulatory network. This analysis nominated ATF4 as a key regulator, whose target genes were significantly enriched in inflammatory and immune-related pathways (Figs. 4G-I). This positions the ATF4-driven C5 neutrophil subpopulation as a potential key mediator of intestinal I/R pathology.

ER stress mediates neutrophil-driven intestinal I/R injury

Building on the identification of ATF4 in neutrophil subclusters, we investigated the functional role of endoplasmic reticulum (ER) stress using pharmacological modulators.

Mice subjected to I/R were treated with the ER stress activator tunicamycin (IR + Tun) or inhibitor 4-PBA (IR + 4PBA). Augmenting ER stress in the IR + Tun group exacerbated intestinal injury, as shown by elevated serum DAO and IL-6 levels, greater shortening of the small intestine, and worsened histopathological scores compared to the IR group (Figs. 5A-F). Conversely, inhibiting ER stress with 4-PBA in the IR + 4PBA group significantly attenuated all these injury parameters.

Fig. 5.

Pharmacological modulation of ER stress influences I/R injury severity. (A) The expression level of serum DAO in mice was determined using an ELISA kit. (B) Serum IL-6 expression in mice was assessed by ELISA. (C, D) Quantification of small intestinal length in mice. (E, F) H&E staining of mouse small intestine and histological evaluation using Chiu’s scoring system. (G-I) Immunofluorescence staining of mouse small intestine for Ly6G and ATF4 expression. (J-M) Expression levels of the intestinal mucosal barrier protein Occludin and endoplasmic reticulum stress markers ATF4 and GRP78 in mouse intestinal tissue. Data are presented as mean ± SD. Statistical significance is indicated as *P < 0.05, **P < 0.01, ***P < 0.001, **P < 0.0001. NC, normal control; IR, ischemia–reperfusion; Tun, tunicamycin; KO, knockout; neu, neutrophils.

Immunofluorescence analysis demonstrated that this ER stress-dependent aggravation of injury was associated with enhanced infiltration of LyGG + neutrophils and other immune cells (CD45 + , CD3 + , CD11b +) in the IR + Tun group, which was suppressed by 4-PBA treatment (Figs. 5G-I, S3, S4 and S5).

At the molecular level, the expression of the tight junction protein occludin was inversely correlated with ER stress. It was most severely depleted in the IR + Tun group and significantly preserved in the IR + 4PBA group (Figs. 5J, K). Western blot analysis confirmed that the protein levels of the ER stress markers ATF4 and GRP78 were induced by I/R, further amplified by tunicamycin, and suppressed by 4-PBA (Figs. 5L, M).

Cellular mechanism of neutrophil-induced injury via ER stress

To establish a direct causal link, we modulated ER stress in a co-culture system of MODE-K cells with I/R-derived neutrophils (IR neu).

Enhancing ER stress with tunicamycin (IR neu + Tun) exacerbated the inflammatory response (elevated IL-6), reduced cell viability (CCK-8), and increased apoptosis (TUNEL) compared to the IR neu group. Conversely, inhibiting ER stress with 4-PBA (IR neu + 4PBA) significantly mitigated these cytotoxic effects (Figs. 6A, B, D, E).

Fig. 6.

Neutrophils induce MODE-K cell injury through endoplasmic reticulum stress. (A) IL-6 expression in co-culture supernatant was quantified using an ELISA kit. (B) Cell viability of MODE-K cells was assessed using the CCK-8 assay. (C) Transmission electron microscopy (TEM) of MODE-K cells. (D, E) Apoptosis in MODE-K cells was detected by TUNEL staining. (F-I) Expression levels of the tight junction protein Occludin and endoplasmic reticulum stress markers ATF4 and GRP78 in MODE-K cells. Data are presented as mean ± SD. Statistical significance is indicated as *P < 0.05, **P < 0.01, ***P < 0.001, **P < 0.0001. NC, normal control; IR, ischemia–reperfusion; Tun, tunicamycin; KO, knockout; neu, neutrophils.

Transmission electron microscopy revealed that ER stress potentiation severely disrupted mitochondrial integrity, inducing outer membrane dissolution and cristae breakdown. These ultrastructural damages were markedly rescued by ER stress inhibition (Fig. 6C).

At the molecular level, the loss of the tight junction protein occludin in the IR neu group was dramatically worsened by tunicamycin and prevented by 4-PBA (Figs. 6F, G). Western blot analysis confirmed that this phenotypic cascade was driven by the activation of the ER stress pathway, as shown by the corresponding increases and decreases in ATF4 and GRP78 protein levels across the treatment groups (Figs. 6H, I).

Genetic deletion of ATF4 abrogates neutrophil-mediated injury

To definitively establish the centrality of ATF4 in the observed ER stress pathway, we subjected ATF4 knockout (KO) mice and wild-type (NC) controls to I/R with or without tunicamycin (Tun) pre-treatment.

Augmenting ER stress with Tun in wild-type mice (NC + IR + Tun) severely exacerbated intestinal injury, as shown by elevated serum DAO and IL-6, greater intestinal shortening, and worsened histopathological scores compared to the NC + IR group (Figs. 7A-F). This was accompanied by enhanced infiltration of Ly6G + neutrophils and other immune cells, and a profound loss of the barrier protein occludin (Figs. 7G-K).

Fig. 7.

Genetic deletion of ATF4 protects against I/R injury exacerbated by ER stress. (A) The expression level of serum DAO in mice was determined using an ELISA kit. (B) Serum IL-6 expression in mice was assessed by ELISA. (C, D) Quantification of small intestinal length in mice. (E, F) H&E staining of mouse small intestine and histological evaluation using Chiu’s scoring system. (G-I) Immunofluorescence staining of mouse small intestine for Ly6G and ATF4 expression. (J, K) Expression level of intestinal mucosal barrier protein Occludin in mouse small intestine. Data are presented as mean ± SD. Statistical significance is indicated as *P < 0.05, **P < 0.01, ***P < 0.001, **P < 0.0001. NC, normal control; IR, ischemia–reperfusion; Tun, tunicamycin; KO, knockout; neu, neutrophils.

Critically, the genetic ablation of ATF4 completely abrogated the injurious effects of Tun. The KO + IR and KO + IR + Tun groups exhibited equally mild levels of injury across all parameters measured, which were significantly lower than those in the corresponding wild-type groups. The absence of significant differences between KO + IR and KO + IR + Tun demonstrates that ATF4 is an essential downstream mediator through which ER stress-activating signals drive neutrophil-mediated intestinal I/R injury.

• Neutrophil-Intrinsic ATF4 Drives Epithelial Injury via ER Stress

Neutrophil-intrinsic ATF4 drives epithelial injury via ER stress

To establish that ER stress acts through ATF4 within neutrophils to mediate their cytotoxicity, we co-cultured MODE-K cells with neutrophils isolated from either wild-type or ATF4 knockout (KO) mice, under I/R-mimicking conditions with or without tunicamycin (Tun).

Neutrophils from wild-type mice (IR neu) induced inflammatory cytokine release (IL-6), reduced MODE-K cell viability, and caused severe mitochondrial and endoplasmic reticulum damage. These cytotoxic effects were significantly exacerbated by Tun pre-treatment (IR neu + Tun) (Figs. 8A-C).

Fig. 8.

Neutrophil-intrinsic ATF4 is required for ER stress-mediated epithelial injury. (A) IL-6 expression in co-culture supernatant was quantified using an ELISA kit. (B) Cell viability of MODE-K cells was assessed using the CCK-8 assay. (C) Transmission electron microscopy (TEM) of MODE-K cells. (D, E) Apoptosis in MODE-K cells was detected by TUNEL staining. (F, G) Expression of the tight junction protein Occludin in MODE-K cells was detected by Western blotData are presented as mean ± SD. Statistical significance is indicated as *P < 0.05, **P < 0.01, ***P < 0.001, **P < 0.0001. NC, normal control; IR, ischemia–reperfusion; Tun, tunicamycin; KO, knockout; neu, neutrophils.

In stark contrast, neutrophils derived from ATF4 KO mice [(KO + IR) neu] exhibited a markedly blunted capacity to injure MODE-K cells. Critically, pre-treating these ATF4-deficient neutrophils with Tun [(KO + IR) neu + Tun] failed to exacerbate the injury. The two KO groups showed similarly low levels of IL-6, high cell viability, intact cellular ultrastructure, reduced apoptosis, and preserved occludin expression, with no significant differences between them (Figs. 8A-G).

These results conclusively demonstrate that the ER stress-ATF4 pathway intrinsic to neutrophils is essential for their full cytotoxic potential in driving intestinal epithelial injury during I/R.

Discussion

This study delineates a high-resolution landscape of neutrophil heterogeneity in intestinal ischemia–reperfusion (IR) injury, identifying a pathogenic subpopulation driven by the endoplasmic reticulum stress (ERS) sensor ATF4. While neutrophils are established early responders that exacerbate tissue damage through the release of cytotoxic mediators^41–43, our scRNA-seq analysis transcends this homogeneous view by revealing six distinct subclusters.

A key finding is the characterization of the C5 neutrophil subcluster, which exhibits a pronounced pro-inflammatory (N1-like) signature marked by effector cytokines (e.g., I11b, Tnf, CxcI2). This aligns with previous studies describing pro-inflammatory neutrophil subsets in other inflammatory contexts, such as atherosclerosis and acute lung injury, which are major drivers of tissue damage^44–46. This suggests a model where coordinated subpopulations mediate the injury response, with the C5 cluster acting as a primary amplifier of inflammation. Consequently, selectively targeting this defined subset, rather than global neutrophil depletion, may present a superior therapeutic strategy to curb collateral damage.

Mechanistically, we established a functional link between neutrophil activation and ERS. A prominent ERS signature, characterized by upregulation of the key transcription factor ATF4, was specifically enriched in the C5 subcluster. Our findings extend the known role of ER stress in IRI from parenchymal cells to immune cells. Prior works^47–50 have shown that ER stress in hepatocytes or cardiomyocytes contributes to IRI-induced apoptosis. Here, we demonstrate that ER stress within neutrophils themselves amplifies their cytotoxic function, creating a vicious cycle of inflammation and epithelial damage. Functional validation confirmed that both pharmacologic inhibition of ERS with 4-PBA and genetic ablation of ATF4 significantly attenuated neutrophil-driven inflammation and mucosal injury^51–53. These findings strongly implicate the ERS-ATF4 axis as a central pathway in this pathology, although contributions from other ERS branches like IRE1α-XBP1 warrant future investigation^54–56.

However, our study has several limitations. First, the translational relevance of targeting a specific neutrophil subpopulation in human IRI requires further validation, including the identification of conserved surface markers for the C5 subset [58,59]. Second, while we focused on the ATF4 branch, the potential crosstalk between different UPR pathways (PERK, IRE1α, ATF6 [60,61,62]) in neutrophils during IRI remains unexplored. Third, the long-term consequences of modulating the neutrophil ERS-ATF4 pathway on tissue repair and host defense against secondary infections were not assessed in this acute model [63]. Future studies should address these points to facilitate clinical translation.

In summary, our work identifies a pro-inflammatory, ERS-ATF4-driven neutrophil subcluster as a critical mediator of intestinal IR injury. Targeting this specific cellular mechanism offers a promising avenue for therapy. Future efforts should focus on identifying surface markers for the C5 subset and evaluating the long-term consequences of modulating this pathway on tissue repair and host defense to facilitate clinical translation.

Author contributions

Y.Y. and Q.Z. contributed equally to this work. Y.Y., Q.Z., J.W., and Y.X. conceived and designed the research. Y.Y., Q.Z., and J.W. performed the majority of the experiments and data acquisition. S.L. and B.W. provided assistance with statistical and computational analyses. Y.Y. and Q.Z. drafted the initial manuscript. Y.X. and J.W. supervised the project and critically revised the manuscript. All authors reviewed, edited, and approved the final version of the manuscript.

Funding

This research was funded by the Jiangsu Provincial Key Medical Discipline Cultivation Unit, grant number JSDW202250. The funding body had no involvement in the design of the study, data collection and analysis, interpretation of results, or preparation of the manuscript.

Data availability

The single-cell RNA sequencing datasets generated and analyzed in the current study are publicly accessible in the NCBI Gene Expression Omnibus (GEO) repository . The mouse intestinal epithelial cell line MODE-K was procured from Fuheng Biotechnology (Shanghai, China). All original, uncropped Western blot images are included in the Supplementary file. All other data supporting the findings of this study are available from the corresponding authors upon reasonable request.

Declarations

Competing interests

The authors declare no competing interests.

Ethics approval

The animal study protocol received ethical approval from the Institutional Animal Care and Use Committee (IACUC) of Jiangsu Huitong Test & Evaluation Technology Group Co., Ltd (Approval Reference Number: HT-2025-LWFB-0088). All experiments were performed in strict adherence to institutional and national guidelines for the care and use of laboratory animals.