OLR1 as a Key Inflammatory Regulator Associated with Poor Response to Neoadjuvant Immunochemotherapy in Resectable Lung Cancer Authors: Xinxin Chen, Siang Zhang, Xuefeng Ai, Yongfei Fan, Wenjie Chen Published: 2026 ABSTRACT BackgroundPoor pathological response to neoadjuvant immunotherapy combined with chemotherapy (nICT) limits long-term survival in resectable lung cancer. How to improve response rates and the regulatory mechanism of tumor inflammatory signaling in nICT efficacy remain unclear.MethodsIn this study, patients were stratified into molecular subgroups using unsupervised consensus clustering based on inflammatory response genes (IRGs) expression profiles. Single-cell RNA sequencing (scRNA-seq) and spatial transcriptomics were applied to characterize tumor inflammatory states, immune microenvironment heterogeneity, and spatial organization between major pathological response (MPR) and non-MPR (NMPR) patients. Differentially expressed gene (DEG) analysis was performed to identify potential therapeutic targets, followed by preliminary functional validation to explore their biological roles.ResultsConsensus clustering identified four molecular subgroups with distinct survival outcomes and clinical characteristics. scRNA-seq profiling uncovered that intratumoral inflammatory activation was markedly elevated in NMPR patients relative to MPR patients, with neutrophil subsets exhibiting the most robust inflammatory response among all immune cell populations. Further subset analysis identified two pro-inflammatory neutrophil populations (N3_CXCL8 and N4_VCAN), which were enriched in hypoxic tumor regions and contributed to immune evasion and angiogenesis. OLR1 was identified as a key regulatory hub associated with inflammatory activation, immunosuppression, and pro-angiogenic signaling. In addition, OLR1+ monocytes and macrophages exhibited consistent pro-tumor inflammatory phenotypes.ConclusionThis study identifies an OLR1-associated inflammatory regulatory axis that contributes to impaired response to nICT in lung cancer. Targeting OLR1 may represent a promising strategy to enhance immunotherapy efficacy and improve clinical outcomes in patients with resectable lung cancer. However, the study is limited by a relatively small validation cohort and a strong reliance on bioinformatics analyses, underscoring the need for further translational research and clinical validation to facilitate potential therapeutic application. Introduction Latest epidemiological statistics issued by the International Agency for Research on Cancer (IARC) highlight that lung cancer remains the most prevalent malignant tumor worldwide, topping the list of both cancer incidence and mortality across all neoplastic diseases.1 In line with the histological classification standards formulated by the World Health Organization (WHO), lung cancer is broadly divided into two major pathological categories: non-small cell lung cancer (NSCLC), which represents roughly 85% of all newly diagnosed lung cancer cases, and small cell lung cancer (SCLC), accounting for the remaining 15% of clinical presentations.2 In recent years, remarkable progress in contemporary medical science has substantially optimized long-term survival outcomes for lung cancer patients, supported by multiple technological and clinical breakthroughs. These advances include early tumor screening via low-dose chest computed tomography (CT), precise preoperative lesion staging via positron emission tomography-CT (PET-CT), upgraded surgical intervention techniques, as well as a host of adjuvant therapeutic modalities encompassing chemotherapy, radiotherapy, molecular targeted therapy and immunotherapy.3–6 Despite these cumulative clinical achievements, roughly 1.8 million people worldwide still die of lung cancer every year,1 underscoring the persistent urgency to explore more effective therapeutic strategies to further improve patient prognosis and clinical outcomes. Surgical resection stands as the optimal therapeutic choice for patients with early and intermediate-stage lung cancer, including those with stage I, stage II and part of stage III disease, and currently remains the only curative treatment modality for this malignancy.7 Nevertheless, patient prognosis exhibits remarkable heterogeneity closely correlated with tumor staging. Existing clinical studies have validated that the 5-year survival rate of stage I NSCLC patients reaches approximately 80%, whereas this rate drops to a range of 13% to 60% among patients with stage II or stage III NSCLC.3,8 Surgical treatment alone fails to achieve ideal survival benefits for patients with advanced local disease, which has prompted the integration of surgical intervention with multidisciplinary adjuvant therapy to become the mainstream clinical treatment protocol for lung cancer.9,10 The rise of neoadjuvant therapy has fundamentally reshaped the conventional surgical treatment paradigm for lung cancer, bringing a new breakthrough to perioperative clinical management.9,11 The combination of neoadjuvant therapy prior to surgical resection can effectively reduce tumor volume, improve the probability of radical complete resection, and enhance both the success rate and safety of surgical operation. Meanwhile, preoperative systemic therapy is capable of clearing microscopic metastatic lesions, thereby lowering the risks of postoperative local tumor recurrence and distant metastasis.9,11,12 Neoadjuvant immunotherapy combined with chemotherapy (nICT) takes full advantage of the direct cytotoxicity of chemotherapeutic drugs and the immune activation effect of immunotherapeutic agents, exerting synergistic therapeutic effects to achieve higher tumor regression rate and better pathological response, thus effectively improving the prognostic status of lung cancer patients.10,12 The Phase III CheckMate 816 clinical trial was conducted to compare the therapeutic efficacy of three cycles of neoadjuvant nivolumab combined with platinum-doublet chemotherapy versus chemotherapy alone in resectable stage IB to IIIA NSCLC patients. Results revealed that patients treated with nivolumab plus chemotherapy obtained significantly higher pathological complete response (pCR; 24% vs 2%; odds ratio (OR) = 14) and major pathological response (MPR; 37% vs 9%; OR = 5.7) rates than those treated with chemotherapy alone.13 Correspondingly, the Phase II NEOSTAR clinical trial showed that among resectable stage I–IIIA NSCLC patients treated with single-agent nivolumab, the MPR rate was merely 24% and the pCR rate was only 9%.14 These clinical studies collectively confirm that nICT exerts favorable therapeutic effects in improving patient prognosis. However, even with the combined application of immunotherapy and platinum-doublet chemotherapy, the pathological response rate of patients remains relatively low, which greatly limits the survival benefits for a large number of lung cancer patients.12,15 Accordingly, screening and identifying effective molecular targets to elevate neoadjuvant therapeutic response has become vital to expand survival benefits for more patients. Inflammatory response, which can be categorized into acute and chronic inflammation, exerts a bidirectional regulatory role in tumor development within the tumor microenvironment (TME), possessing both tumor-promoting and tumor-suppressing biological functions.16,17 Acute inflammation serves as a rapid defensive response to tissue injury or pathogenic infection, which can eliminate abnormal tumor cells and exert anti-tumor biological effects by activating immune cells such as macrophages, neutrophils and natural killer cells.17,18 Tumor-associated inflammation, a core hallmark of malignant tumors, mainly presents as chronic inflammation. Chronic inflammation in the TME is typically featured by sustained low-intensity immune activation and continuous cytokine secretion.19 Such microenvironmental state further induces the aggregation of a variety of immunosuppressive cells, including tumor-associated macrophages (TAMs), myeloid-derived suppressor cells (MDSCs) and regulatory T cells (Tregs), thereby constructing an immunosuppressive TME.17,18,20 In addition, tumor-associated chronic inflammation promotes tumor progression primarily by accelerating tumor angiogenesis. Inflammatory cell infiltration, especially TAM accumulation, plays a vital role in tumor angiogenesis by secreting vascular endothelial growth factor (VEGF); this pro-angiogenic effect can be further enhanced by inflammatory cytokines such as IL-1 and IL-6, while chemokines including CXCL8 and CCL5 also promote angiogenesis through direct or indirect pathways.17,19,21 Moreover, tumor-associated chronic inflammation can induce gene mutations, disturb the expression and transformation of oncogenes and tumor suppressor genes, and inhibit tumor cell apoptosis, ultimately accelerating the malignant progression of tumors.18 Therefore, targeting tumor-associated inflammatory response has become a novel and promising direction for tumor clinical treatment.17,21,22 Importantly, accumulating evidence suggests that persistent chronic inflammation also contributes to resistance to cancer immunotherapy by establishing an immunosuppressive and pro-angiogenic TME. However, the mechanisms by which chronic inflammatory states influence the pathological response to nICT in lung cancer remain poorly understood. We therefore hypothesized that distinct inflammatory states within the TME contribute to the heterogeneous responses to nICT by orchestrating immune evasion and angiogenic remodeling. In this study, a total of 1,842 lung cancer patients were enrolled and divided into four subgroups with distinct TME characteristics via consensus clustering analysis based on the expression levels of inflammatory response genes (IRGs) obtained from bulk RNA sequencing (RNA-seq) data. Single-cell RNA sequencing (scRNA-seq) data derived from post-nICT patient tissues were used to assess inflammatory response levels across different pathological response groups and various cell subpopulations. Combined with spatial transcriptomics analysis, we further identified a specific neutrophil subset with pro-angiogenic and immunosuppressive biological characteristics. Differential gene expression analysis demonstrated that OLR1 was significantly upregulated in immunosuppressive and pro-angiogenic subsets of neutrophils, macrophages, and monocytes. Recent evidence has highlighted that tumor-associated inflammation is tightly linked to systemic metabolic and nutritional dysregulation, which further contributes to immune remodeling and tumor progression, suggesting an important inflammation–metabolism–immune axis in cancer biology.23 Importantly, OLR1 was prioritized among candidate inflammation-related genes due to its central role in integrating lipid uptake, metabolic reprogramming, and inflammatory signaling, which distinguishes it from other differentially expressed candidates identified in the analysis pipeline.24,25 Subsequent functional analyses further confirmed that OLR1 promotes tumor-associated inflammation, immune escape, and angiogenesis. These findings provide novel mechanistic insights and potential therapeutic targets for improving pathological response to nICT in lung cancer. Material and Methods Patient Tissue Samples and Clinical Data Collection A total of 26 lung cancer patients who received nICT followed by radical surgical resection were enrolled at Shanghai Pulmonary Hospital, with the clinical time range spanning from July 1, 2023, to March 31, 2024. Detailed clinicopathological characteristics of all enrolled subjects were summarized in Supplementary Table 1. The inclusion criteria were set as follows: 1) Patients diagnosed with stage IIIA or IIIB lung cancer in line with the 9th edition of TNM staging system formulated by the International Association for the Study of Lung Cancer (IASLC) prior to surgical intervention;26 2) Patients assessed as unsuitable for direct surgical resection, thus requiring preoperative neoadjuvant treatment; 3) Patients receiving platinum-based combination immunotherapy as the neoadjuvant therapeutic regimen. Meanwhile, patients meeting the following criteria were excluded: 1) Individuals with poor surgical tolerance or those from whom valid clinical tissue specimens could not be collected; 2) Patients who had received other antitumor treatments including conventional chemotherapy, radiotherapy or molecular targeted therapy prior to neoadjuvant treatment. Immunohistochemical Staining Immunohistochemical (IHC) staining targeting OLR1 was carried out on collected clinical paraffin-embedded tissue specimens. Briefly, 4 µm thick paraffin sections were subjected to conventional deparaffinization and gradient rehydration. Antigen retrieval was performed via high-pressure heating in sodium citrate buffer (pH = 6.0) for 20 minutes, followed by natural cooling to room temperature. Subsequently, sections were incubated with 3% hydrogen peroxide solution for 10 minutes to block endogenous peroxidase activity, and rinsed three times with phosphate-buffered saline (PBS). After blocking with 10% normal goat serum at room temperature for 30 minutes, sections were incubated with OLR1 primary antibody (1:200, Proteintech, China) at 4°C overnight. On the next day, sections were washed with PBS three times, followed by incubation with horseradish peroxidase-conjugated secondary antibody (1:500, Proteintech, China) at room temperature for 1 hour. After additional PBS rinsing, color development was conducted using 3,3’-diaminobenzidine (DAB) chromogenic reagent for 8 minutes. Sections were then counterstained with hematoxylin, dehydrated via gradient ethanol, sealed with neutral balsam, and photographed under a Nikon optical microscope (Nikon, Japan). For semi-quantitative analysis, ImageJ software (version 1.8.0) was used. For each sample, three random fields of view (randomly selected non-overlapping regions) were analyzed. DAB-positive staining was quantified using a standardized threshold-based image analysis approach in ImageJ. The positive staining area was calculated as the proportion of DAB-positive area relative to total tissue area within each field. The final IHC score for each sample was defined as the average value across three randomly selected fields. Bulk RNA Sequencing Dataset Acquisition and Preprocessing Public bulk RNA-seq datasets were retrieved from The Cancer Genome Atlas (TCGA, https://portal.gdc.cancer.gov/) and Gene Expression Omnibus (GEO, https://www.ncbi.nlm.nih.gov/geo/) databases. Four eligible GEO datasets (GSE37745, GSE41271, GSE42127, GSE50081) with consistent sequencing depth and complete clinicopathological annotations were included in this study. Patients with overall survival time less than 30 days were excluded to eliminate interference from non-tumor related death. The baseline clinicopathological information of lung cancer patients with complete follow-up data was presented in Supplementary Table 2. The “limma” R package was applied to convert TCGA RNA-seq data from FPKM (Fragments Per Kilobase of transcript per Million) format to TPM (transcripts per million) format for standardization. GEO datasets were directly adopted using the provided log2-normalized expression matrix. The Combat function embedded in the “sva” R package was utilized to eliminate batch effects derived from different datasets, ensuring the consistency of subsequent integrated analysis. Consensus Clustering Analysis Based on the expression levels of IRGs in the integrated TCGA and GEO lung cancer cohort, consensus clustering analysis was conducted using the “ConsensusClusterPlus” R package. TPM-normalized data were employed for TCGA samples, while log2-normalized microarray data were used for GEO samples. The optimal clustering number was determined by combining the cumulative distribution function (CDF) curves of consensus scores and the distribution characteristics of consensus matrix heatmap. Pathway Enrichment Analysis Hallmark gene sets and IRGs were obtained from the Molecular Signatures Database (MSigDB, https://www.gsea-msigdb.org/gsea/msigdb/). Gene Set Variation Analysis (GSVA) was performed using the “GSVA” R package to estimate sample-level enrichment scores of Hallmark pathways across individual tumor samples. GSVA is a non-parametric, unsupervised method that calculates pathway activity based on the expression profile of each sample independently, without requiring any predefined reference group such as adjacent normal lung tissue. Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis was carried out using the “clusterProfiler” R package, with Homo sapiens selected as the research species based on the “org.Hs.eg.db” R package. Enrichment results were filtered with a threshold of adjusted p-value < 0.05. Differential expression results were adjusted for multiple testing using the Benjamini–Hochberg false discovery rate (FDR) correction method, and adjusted p-values were used to determine statistical significance where applicable. Metabolic pathway analysis was conducted using the “scMetabolism” R package with the parameter set as “metabolism type = KEGG”, to identify differential metabolic activity among subgroups. Functional enrichment analysis of IRGs was performed using the Metascape platform (https://metascape.org). The IRG list was analyzed using over-representation analysis (ORA), with the background set defined as all annotated human protein-coding genes. This analysis represents a gene set–based functional annotation and does not involve comparisons between tumor and adjacent normal tissues or between clinical subgroups. Immune Infiltration and Functional Evaluation Single-sample gene set enrichment analysis (ssGSEA) was performed to quantify the infiltration levels of 29 immune cell subsets and related immune functional scores in different subgroups.27 To facilitate cross-sample and cross-gene set comparison, ssGSEA scores were normalized. The normalization process scaled the original enrichment scores to a range of 0 to 1, where 0 represented the minimum enrichment score and 1 represented the maximum enrichment score for each gene set. Single-Cell RNA Sequencing Data Processing Public scRNA-seq data of lung cancer samples post nICT treatment were downloaded from the GEO database (GSE207422).28 Patients were divided into MPR group (≤10% residual viable tumor cells) and NMPR group (>10% residual viable tumor cells) according to postoperative pathological response. Each group contained 2 lung adenocarcinoma samples and 2 lung squamous carcinoma samples (Supplementary Table 3), and pCR samples were incorporated into the MPR group as they conformed to the MPR definition. Data integration and batch effect correction were conducted using the “Harmony” R package, followed by quality control, data normalization, dimensionality reduction, visualization and cell clustering via the “Seurat” R package. Low-quality cells were filtered out based on the following criteria: 1) total gene count per cell ≤ 1000; 2) detected gene number per cell ≤ 500; 3) mitochondrial gene expression ratio > 20%. The FindNeighbors function was applied with parameters set as reduction = “harmony” and dims = 1:15 to construct cell neighborhood graph and calculate cell similarity. The optimal clustering resolution was determined as 0.5 via clustree analysis. Cell Subpopulation Annotation The FindAllMarkers function was first used to screen specific marker genes for each cell cluster. Then, automatic cell annotation was performed using the “SingleR” R package with the HumanPrimaryCellAtlasData as the reference database. Finally, cell annotation results were further verified and optimized combined with classical cell subset marker genes published in previous literature. Cellular Communication Analysis Cell-cell communication analysis was conducted using the “CellChat” R package based on the gene expression matrix of each cell subset. The CellChatDB.human dataset was adopted as the reference library to analyze the ligand-receptor interaction and secreted signaling protein communication among different cell subsets. The netVisual_diffInteraction function was utilized to visualize the quantity and intensity of intercellular communication. Pseudotime Trajectory Analysis Pseudotime developmental trajectory analysis of cell subsets was performed using the “Monocle2” R package. Dimensionality reduction was conducted via the “DDRTree” function, and cell trajectory distribution was visualized using the plot_cell_trajectory function. The differentialGeneTest function was employed to conduct likelihood ratio test and screen genes with significant expression changes along the cell trajectory. The plot_genes_branched_heatmap function was applied to plot the heatmap of branching gene expression along the pseudotime axis. Differential Gene Expression Analysis The FindMarkers function embedded in the “Seurat” R package was used to identify marker genes of each cell subset. Differentially expressed genes (DEGs) between NMPR and MPR groups were screened with thresholds of p-value < 0.05 and |log2 fold change| > 1.5. Differential expression results were adjusted for multiple testing using the Benjamini–Hochberg FDR correction method, and adjusted p-values were used to determine statistical significance where applicable. The “VennDiagram” R package was used to intersect differential genes with inflammatory response and angiogenesis-related genes, and the intersection results were visualized via Venn diagram. Molecular Docking Simulation Target protein structures were retrieved from the Protein Data Bank (PDB, https://www.rcsb.org/) and processed using PyMOL software (version 2.5.0) to remove redundant water molecules and unrelated receptor residues. The molecular structure of corresponding small molecules was constructed using Chemdraw software, and the structure was optimized via the “Calculation-MM2 minimize energy” function. Molecular docking between target proteins and small molecule ligands was performed using PyRx software (version 0.9.1, https://pyrx.sourceforge.io/) with the built-in AutoDock Vina tool. The binding affinity (kcal/mol) was used to evaluate the binding stability between ligand and receptor, with lower affinity values representing stronger and more stable ligand-receptor binding. Spatial Transcriptome Analysis Spatial transcriptome data were obtained from the Comprehensive Repository of Spatial Transcriptomics (CROST, https://ngdc.cncb.ac.cn/crost/home). The human lung cancer tissue sample (ID: VISDS000889) in the Spatial Visualization module was selected as the research object. Spatial transcriptome deconvolution algorithm was applied to resolve mixed cell types, assign specific cell subsets to independent spatial spots, and obtain gene expression profiles of each spatial spot. To explore the spatial co-localization of key marker genes, FeaturePlot images derived from spatial transcriptome data were converted into 8-bit grayscale TIFF images with black background, and imported into ImageJ software for co-localization analysis using the Coloc 2 plugin. Manders’ overlap coefficients (M1 and M2) were calculated to quantify the spatial overlapping degree of different gene channels, and Pearson correlation coefficients were computed to evaluate the expression intensity correlation. Furthermore, distance-dependent correlation analysis was performed using GraphPad Prism software. The correlation between key gene expression intensity and spatial distance from the region of interest (ROI) was analyzed, and the distance attenuation curve of gene expression was plotted to quantify the spatial expression gradient. Immunofluorescence Staining Formalin-fixed paraffin-embedded (FFPE) tissue sections were collected from 3 NSCLC patients receiving nICT, including 1 pCR case, 1 MPR case and 1 NMPR case. Immunofluorescence (IF) staining was conducted to analyze the spatial distribution relationship between CD8⁺ T cells and OLR1-positive cells in the TME. Briefly, 4 μm thick tissue sections were deparaffinized with xylene, rehydrated via gradient ethanol, and subjected to heat-induced antigen retrieval in citrate buffer (pH = 6.0). Endogenous autofluorescence was quenched using TrueBlack reagent (Biotium, USA). After blocking with 5% bovine serum albumin (BSA) at room temperature for 1 hour, sections were incubated with CD8 antibody (1:200; Proteintech, Wuhan, China) and OLR1 antibody (1:200; Proteintech, Wuhan, China) at 4°C overnight. After rinsing, sections were incubated with Alexa Fluor-conjugated species-specific secondary antibodies in the dark at room temperature for 1 hour (Proteintech, Wuhan, China). Nuclei were counterstained with DAPI, and sections were sealed with antifade mounting medium. IF images were captured using a Zeiss LSM 880 laser scanning confocal microscope (Carl Zeiss Microscopy GmbH, Germany). For analysis, representative fields of view were randomly selected from each specimen to assess the spatial relationship between OLR1 expression and CD8⁺ T-cell infiltration. Cell Culture Human monocytic cell line THP-1 and human NSCLC cell line NCI-H2122 were purchased from Zhongqiao Xinzhuo Biotechnology Co., Ltd. (Shanghai, China). Cells were cultured in RPMI 1640 medium (Procell, Wuhan, China) supplemented with 10% fetal bovine serum (FBS; Procell, Wuhan, China), and incubated in a 37°C, 5% CO2 humidified incubator (Thermo Fisher Scientific, Waltham, MA, USA). All cell lines were authenticated via short tandem repeat (STR) profiling, and confirmed to be free of mycoplasma contamination. Cell Transfection The OLR1 overexpression plasmid (OV-OLR1) and corresponding empty vector negative control plasmid (OV-NC) were purchased from Jiman Biotechnology Co., Ltd. (Shanghai, China), and cell transfection was conducted using FanRi-Trans CE transfection reagent (FanRi® company, Xiamen, China). To induce macrophage differentiation, THP-1 cells were treated with 100 ng/mL phorbol 12‑myristate 13‑acetate (PMA; MedChemExpress, USA) for 24 h. THP-1 cells in logarithmic growth phase were seeded into 6-well plates at a density of 2.5 × 104 cells per well with 2 mL complete medium. After cell adherence, the medium was replaced with fresh medium, leaving 1.2 mL per well. According to the manufacturer’s instructions, 8 μL OLR1 plasmid and 8 μL transfection reagent were gently mixed and incubated at room temperature for 20 minutes, then added into each well dropwise. The culture medium was replaced with fresh complete medium after 24 hours of transfection. Western Blot To induce macrophage differentiation, THP-1 cells were treated with 100 ng/mL PMA for 24 h. THP-1 cells were divided into negative control group and OLR1 overexpression group, with three biological replicates in each group. Cell transfection was performed in 6-well plates following the aforementioned protocol. After 48 hours of transfection, total protein was extracted using RIPA lysis buffer supplemented with protease and phosphatase inhibitors (Beyotime, Shanghai, China). Protein concentration was quantified using a BCA protein assay kit (Beyotime, Shanghai, China). Protein samples were separated via 10% SDS-PAGE gel electrophoresis (Epizyme, Shanghai, China), and transferred onto PVDF membranes (Millipore, Billerica, MA, USA). Membranes were blocked with rapid blocking buffer (Epizyme, Shanghai, China) on a shaking table at room temperature for 15 minutes. Subsequently, membranes were incubated with OLR1 primary antibody (1:1000, Proteintech, Wuhan, China) at 4°C overnight. After washing three times with TBST buffer (5 minutes each time), membranes were incubated with horseradish peroxidase-conjugated secondary antibody (1:2000, Proteintech, Wuhan, China) at room temperature for 1.5 hours, followed by TBST rinsing. Protein bands were visualized using ECL chemiluminescent reagent (Epizyme, Shanghai, China) and a gel imaging system. Band gray values were quantified using ImageJ software (version 1.8.0). Enzyme-Linked Immunosorbent Assay To induce macrophage differentiation, THP-1 cells were treated with 100 ng/mL PMA for 24 h. THP-1 cells were grouped and transfected as described above, with three biological replicates per group. After 48 hours of transfection, the culture medium was replaced with fresh medium, and cells were cultured for another 24 hours. The cell culture supernatant was collected after 72 hours of total transfection for subsequent detection. Enzyme-linked immunosorbent assay (ELISA) was performed using TGF-β and VEGF ELISA kits (Boster Biological Technology, Wuhan, China) following the manufacturer’s protocols. Capture antibody was diluted and added into 96-well ELISA plates (Nunc, Thermo Fisher, USA) at 100 µL per well, followed by overnight incubation at 4°C. After washing with PBS-T buffer (PBS containing 0.05% Tween-20), plates were blocked with 200 µL 5% BSA blocking buffer at room temperature for 1–2 hours. After washing, 100 µL cell supernatant was added into each well and incubated at room temperature for 1–2 hours. Biotinylated detection antibody was added and incubated for another 1 hour, followed by incubation with streptavidin-HRP conjugate for 1 hour. Plates were thoroughly washed, incubated with TMB chromogenic substrate, and the reaction was terminated with stop solution. The absorbance value at 450 nm was detected using a TECAN microplate reader (TECAN, Männedorf, Switzerland). All ELISA measurements were performed based on three independent biological replicates per experimental group. Cell Co-Culture An indirect transwell co-culture system was established to evaluate the effect of OLR1-overexpressing macrophages on NCI-H2122 cell proliferation. To induce macrophage differentiation, THP-1 cells were treated with 100 ng/mL PMA for 24 h. THP-1 cells transfected with OLR1 plasmid or negative control plasmid were seeded into the upper chamber of 24-well transwell plates (0.4 μm pore size; Corning, USA). Meanwhile, NCI-H2122 cells were seeded into the lower chamber at a density of 1×105 cells per well. Cells were co-cultured for 48 hours to allow the exchange of soluble cytokines without direct cell contact. CCK-8 Proliferation Assay After co-culture, NCI-H2122 cells in the lower chamber were collected and re-seeded into 96-well plates at a density of 5 × 103 cells per well, with three replicate wells per group. Cell proliferation was detected at 0, 1, 2 and 3 days after seeding using Cell Counting Kit-8 (CCK-8; Beyotime, Shanghai, China) following the manufacturer’s instructions. The absorbance value at 450 nm was measured using an Infinite 200 Pro microplate reader (Tecan, Männedorf, Switzerland), and the relative proliferation rate was calculated based on OD values. All CCK-8 measurements were performed based on three independent biological replicates per experimental group. Clone Formation Assay After 48 hours of co-culture, NCI-H2122 cells were collected and seeded into 6-well plates at a density of 1000 cells per well. Cells were cultured in complete medium for 6 days, with medium renewal every 3 days. At the end of culture, cells were washed twice with PBS, fixed with 4% paraformaldehyde (Beyotime, Shanghai, China) for 20 minutes, and stained with 0.1% crystal violet (Beyotime, Shanghai, China) at room temperature for 15 minutes. Stained cell clones were photographed after washing and air drying. The number of cell clones was quantified using ImageJ software (version 10.8.1) with the Analyze Particles function after background subtraction and threshold adjustment. Three independent replicate wells were set for each group, and the final result was expressed as the mean clone number per well. Wound Healing Assay NCI-H2122 cells were collected after co-culture and seeded into 6-well plates at a density of 5×105 cells per well. When cell confluence reached approximately 90%, a sterile 200-μL pipette tip was used to make a straight scratch in the cell monolayer. Cells were gently washed twice with PBS to remove floating cells, and cultured with serum-free medium thereafter. Scratch images were captured at 0 and 24 hours using a Nikon phase-contrast microscope (NIB-610; Nikon, Tokyo, Japan). The scratch area was measured using ImageJ software (version 10.8.1), and the wound healing rate was calculated using the following formula: \documentclass[12pt]{minimal} \usepackage{wasysym} \usepackage[substack]{amsmath} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage[mathscr]{eucal} \usepackage{mathrsfs} \DeclareFontFamily{T1}{linotext}{} \DeclareFontShape{T1}{linotext}{m}{n} {linotext }{} \DeclareSymbolFont{linotext}{T1}{linotext}{m}{n} \DeclareSymbolFontAlphabet{\mathLINOTEXT}{linotext} \begin{document}$${\rm {Wound\ healing\ rate (\%) = }}{{{{\rm {A}}_0}{\rm { - }}{{\rm {A}}_\rm t}} \over {{\rm { }}{{\rm {A}}_0}}}{\rm { \times 100}}$$\end{document} where A0 represents the initial scratch area at 0 h, and At represents the residual scratch area at 24 h. Three independent visual fields were analyzed per well, with three replicate wells per group. Transwell Migration and Invasion Assays Transwell chambers with 8.0-μm pore size polycarbonate membranes (Corning, USA) were used to detect the migration and invasion ability of NCI-H2122 cells. For migration assay, 1×105 NCI-H2122 cells suspended in 200 μL serum-free medium were seeded into the upper chamber, and 600 μL complete medium supplemented with 20% FBS was added into the lower chamber as a chemoattractant. For invasion assay, the upper chamber was pre-coated with 60 μL Matrigel (1:8; Corning, USA) and incubated at 37°C for 1 hour before cell seeding, with other steps consistent with the migration assay. After 24 hours of incubation, non-migrated or non-invaded cells on the upper membrane surface were wiped off with a cotton swab. Cells on the lower membrane surface were fixed with 4% paraformaldehyde for 20 minutes and stained with 0.1% crystal violet for 15 minutes. Three randomly selected visual fields were photographed under an optical microscope, and the number of migrated or invaded cells was quantified using ImageJ software (version 10.8.1). Statistical Analysis All statistical analyses in this study were performed using R software (version 4.2.2) and GraphPad Prism software (version 10.1.2). The comparison between two groups was conducted using Student’s t-test or Wilcoxon rank-sum test. The correlation between variables was analyzed using Pearson or Spearman correlation coefficient. A p-value < 0.05 was considered statistically significant. Results Identification of Clustering Subgroups Based on the Expression of IRGs in Lung Cancer First, functional enrichment and clustering analyses were conducted via Metascape online platform to annotate the biological functions of IRGs retrieved from the MSigDB database (Supplementary Table 4). Enrichment outcomes verified that IRGs were predominantly enriched in inflammatory response-related biological processes, which constituted the core functional characteristic of these gene sets (Figure 1A and Supplementary Figure 1A). Accumulating evidence has validated that inflammatory response serves a dual regulatory role in biological processes, exerting both protective and detrimental effects on the organism.16,17,20 Tumor-related inflammation typically presents as chronic inflammatory activation, and functions as a critical regulator in remodeling the TME and mediating tumor progression.19 Further enrichment analysis revealed that IRGs were mainly enriched in IL-10, Toll-like receptor, IL-1, and JAK-STAT signaling pathways, all of which have been reported to exert bidirectional regulatory effects in modulating immune homeostasis and tumor malignant progression.29–32 Collectively, these findings elucidated the paradoxical regulatory properties of inflammatory response, and further underscored the intricate regulatory mechanism of IRGs in TME immune modulation and tumor development. Figure 1Consensus clustering and subgroup characterization based on IRGs expression in lung cancer. (A) Functional enrichment and clustering analysis of IRGs using Metascape. (B and C) Consensus clustering (B) subgroup classification and (C) CDF curves related to IRGs (n = 4). (D) Enrichment analysis of hallmark gene sets for four subgroups related to IRGs. (E) Immune cell infiltration and functional analysis of four subgroups related to IRGs. (F) Differential survival analyses of four subgroups related to IRGs. ***: p-value<0.001.This infographic analyzes IRG-based lung cancer subtypes. Panel A depicts biological processes like inflammatory response. Panel B shows a consensus matrix for k=4, indicating clustering. Panel C′s delta area plot for CDF curves shows clustering stability. Panel D′s heatmap illustrates TumorPurity, ESTIMATEScore, ImmuneScore and StromalScore across subtypes A, B, C and D, with pathways such as INTERFERONALPHARESPONSE. Subtype A has higher immune scores. Panel E′s box plot highlights immune infiltration differences, noting MDSCs and CD8 T cells, with Subtype C showing lower infiltration. Panel F′s survival plot reveals significant differences (p<0.05), with Subtype A having better survival rates. The layout logically progresses from clustering to subtype characterization, immune infiltration and survival analysis.Infographic on IRG-based lung cancer subtypes showing differences in immune environment and survival. Consensus clustering and subgroup characterization based on IRGs expression in lung cancer. (A) Functional enrichment and clustering analysis of IRGs using Metascape. (B and C) Consensus clustering (B) subgroup classification and (C) CDF curves related to IRGs (n = 4). (D) Enrichment analysis of hallmark gene sets for four subgroups related to IRGs. (E) Immune cell infiltration and functional analysis of four subgroups related to IRGs. (F) Differential survival analyses of four subgroups related to IRGs. ***: p-value<0.001. Subsequently, consensus clustering analysis was performed on a cohort of 1842 lung cancer patients, with patient grouping conducted based on the expression profile of IRGs. Combined with the CDF curve and consensus matrix validation, the enrolled patients were stratified into four distinct molecular subgroups: cluster A (n = 453), cluster B (n = 487), cluster C (n = 332), and cluster D (n = 570) (Figure 1B, C, Supplementary Figure 1B and Supplementary Table 5). GSVA enrichment analysis was subsequently performed to explore pathway activation discrepancies among subgroups. Results manifested that cluster A and cluster B exhibited prominent upregulation of both anti-tumor immune signals (including INF-α and INF-β pathways) and pro-tumor immune evasion signals, such as IL-6/JAK/STAT3, IL-2/STAT5, transforming growth factor-β (TGF-β), epithelial-mesenchymal transition (EMT) and angiogenesis-related pathways. Notably, the activation amplitude of both immune-killing and immunosuppressive pathways was significantly higher in cluster B than that in cluster A. By contrast, the aforementioned signaling pathways were notably downregulated in cluster C and cluster D. Additionally, glycolysis and hypoxia signaling pathways, which are closely associated with tumor hyperproliferation and immune suppression, were markedly upregulated in cluster B and cluster D (Figure 1D and Supplementary Figure 1C). Immune infiltration and functional assessment further illustrated that, relative to cluster C and cluster D, cluster A and cluster B presented higher infiltration levels of B cells, CD8⁺ T cells, macrophages and NK cells, accompanied by enhanced activation of antigen-presenting cell (APC) co-stimulation, T cell co-stimulation and interferon response-related immune functions (Figure 1E). Prognostic survival analysis revealed significant survival discrepancies among the four subgroups, with cluster A exhibiting the most favorable clinical prognosis (Figure 1F; p < 0.05). Notably, despite the elevated immune infiltration and pathway activation observed in cluster B, this subgroup failed to acquire corresponding survival benefits. Further comparative analysis demonstrated that the infiltration levels of CD8⁺ T cells and NK cells, the core effector cells mediating tumor killing, were higher in cluster A than those in cluster B, which may be attributed to the excessive activation of immune evasion-related signaling in cluster B (Figure 1D and E). Taken together, IRG-based consensus clustering successfully identified four molecular subgroups of lung cancer with distinct TME immune characteristics, and the classification was closely correlated with patient clinical prognosis. ScRNA-Seq Reveals Variability in Inflammatory Response States Between Different Pathologic Responses After quality control and data filtering, transcriptomics data from 53,072 cells were included in this study. Dimensionality reduction and clustering of these cells resulted in 25 clusters (Supplementary Figure 2), which were annotated based on the expression of typical marker genes, identifying them as T cells, macrophages, epithelial cells, monocytes, neutrophils, B cells, fibroblasts, mast cells, and endothelial cells (Figure 2A and Supplementary Table 6). The heatmap presents the top 5 marker genes for each cell subpopulation (Figure 2B). Figure 2C shows the distribution of cell subpopulations across samples, reflecting TME heterogeneity. Similarly, there is significant heterogeneity in the proportions of cell subpopulations between the MPR and NMPR groups. The MPR group had a notably higher proportion of T cells, B cells, and neutrophils, while the NMPR group had a significantly higher proportion of epithelial cells, fibroblasts, and endothelial cells (Figure 2D). We further explored TME heterogeneity between the NMPR and MPR groups using GSVA enrichment analysis. The results indicated that immune evasion signaling were highly enriched in the NMPR group, including IL-6/JAK/STAT3, IL-2/STAT5, EMT, and angiogenesis. Additionally, glycolysis and hypoxia signaling were upregulated in NMPR, while the MPR group was primarily characterized by INF-mediated immune killing. Interestingly, inflammatory response signaling was upregulated in the NMPR group, with significantly higher levels compared to the MPR group (Figure 2E–G; p value < 0.05). Subsequently, our analysis of cell subgroups indicated that the inflammatory response signaling was highest in neutrophils, followed by monocytes and macrophages (Figure 2H and I). In summary, scRNA-seq analysis revealed significant variability in inflammatory response signaling between the NMPR and MPR groups, which may affect patient’s responses to immunotherapy and prognosis. Figure 2scRNA-seq analysis of TME and assessment of inflammatory response in lung cancer patients receiving nICT. (A) UMAP plot visualising the TME of lung cancer patients, consisting of 9 cell subpopulations, each represented by a different color. (B) Heatmap displays the top 5 marker genes for each cell type. (C) Bar graph showing the proportions of the composition of cell types in different samples. Each colour represents a cell type. (D) Bar graph showing the proportion of cell type composition in the NMPR and MPR groups. Each colour represents a cell type. (E) GSVA enrichment analysis comparing the degree of enrichment of hallmark gene sets between the NMPR and MPR groups. The x-axis represents the t-values derived from the comparison of GSVA scores between NMPR and MPR. Color intensity indicates the magnitude of the t-value and does not directly reflect statistical significance. (F) UMAP plot exhibiting the distribution of inflammatory response scores in the NMPR and MPR groups. (G) Violin plot comparing the variability of inflammatory response scores in the NMPR and MPR groups. (H) UMAP plot showing the distribution of inflammatory response scores in cell types. (I) Violin plot displaying the inflammatory response scores in cell types. Each colour represents a cell type. ***: p-value < 0.001; Red box: highlights signals of an inflammatory response.The images analyze cell types and gene expressions. Image A: UMAP scatter plot with clusters of various cells. Image B: Heatmap showing gene expression levels (0-6) across cell types. Image C: Stacked bar chart of cell type ratios in samples. Image D: Stacked bar chart comparing cell type compositions between MPR and NMPR groups. Image E: Horizontal bar chart of t values for GSVA enrichment analysis comparing the degree of enrichment of hallmark gene sets between the NMPR and MPR groups. Image F: UMAP plots for MPR and NMPR, highlighting inflammation response levels. Image G: Violin plot of inflammatory response scores between MPR and NMPR, showing significant differences. Image H: UMAP plot with labeled cell types and inflammation response scale. Image I: Violin plot of inflammation response across cell types, with Neutrophils showing highest values.Multi-panel scRNA-seq infographic comparing TME cell types and inflammatory response in MPR vs NMPR. scRNA-seq analysis of TME and assessment of inflammatory response in lung cancer patients receiving nICT. (A) UMAP plot visualising the TME of lung cancer patients, consisting of 9 cell subpopulations, each represented by a different color. (B) Heatmap displays the top 5 marker genes for each cell type. (C) Bar graph showing the proportions of the composition of cell types in different samples. Each colour represents a cell type. (D) Bar graph showing the proportion of cell type composition in the NMPR and MPR groups. Each colour represents a cell type. (E) GSVA enrichment analysis comparing the degree of enrichment of hallmark gene sets between the NMPR and MPR groups. The x-axis represents the t-values derived from the comparison of GSVA scores between NMPR and MPR. Color intensity indicates the magnitude of the t-value and does not directly reflect statistical significance. (F) UMAP plot exhibiting the distribution of inflammatory response scores in the NMPR and MPR groups. (G) Violin plot comparing the variability of inflammatory response scores in the NMPR and MPR groups. (H) UMAP plot showing the distribution of inflammatory response scores in cell types. (I) Violin plot displaying the inflammatory response scores in cell types. Each colour represents a cell type. ***: p-value < 0.001; Red box: highlights signals of an inflammatory response. Inflammatory Responses Between Neutrophils and Endothelial Cells Promote Tumor Angiogenesis and Immune Evasion To further dissect the underlying mechanism accounting for elevated inflammatory signaling activation in the NMPR group, cell-cell communication analysis was performed to explore intercellular crosstalk within the TME. Based on the calculated communication weight, the results uncovered intensive and complex interaction networks among diverse cell subpopulations (Figure 3A). Given the markedly heightened inflammatory response activity within neutrophils, this cell subset was selected as the core focus for subsequent in-depth analysis. Figure 3ScRNA reveals neutrophil heterogeneity and intercellular communication networks associated with differential inflammatory responses between NMPR and MPR groups. (A) Cell-cell communication analysis among 9 cell subgroups in scRNA-seq data. The thickness of the lines represents the strength of cell-cell communication. (B) Visualization of cell-cell communication between neutrophils and other cell subgroups. (C) Visualization of cell-cell communication between endothelial cells and other cell subgroups. (D) UMAP visualization of neutrophil clustering. Each color represents a neutrophil subtype. (E) UMAP plot demonstrating normalized expression of markers for neutrophil subgroups. (F) Bar plot showing the percentage distribution of neutrophil subgroups in the NMPR and MPR groups. (G) The ridge plot illustrates the inflammatory response scores across neutrophil subgroups. Each color represents a neutrophil subtype. (H) Violin plot comparing differences of inflammatory response scores in the NMPR and MPR groups. (I) GSVA enrichment analysis depicting the enrichment degree of hallmark gene sets in neutrophil subgroups. (J–K) Violin plot comparing the difference in inflammatory response scores of (J) N3_CXCL8 and (J) N4_VCAN subgroups in NMPR and MPR groups. ***: p-value < 0.001.The infographic presents scRNA-seq data analysis of neutrophil subtypes and their communication networks, focusing on differences between MPR and NMPR groups. Top row: A shows a network diagram of cell interactions, B highlights neutrophil connections and C focuses on endothelial cells. Middle row: D is a UMAP plot of neutrophil clustering with subtypes N1 to N5, E shows marker expression for these subtypes and F is a bar chart indicating subtype distribution, with N1 and N3 enriched in NMPR. Bottom row: G is a ridge plot of inflammatory response scores, H and J are violin plots showing higher scores in NMPR and K compares N3CXCL8 and N4VCAN subgroups, with NMPR showing elevated responses. I is a heatmap of pathway enrichment, highlighting increased inflammatory signaling in NMPR. The layout follows a top-to-bottom reading order, summarizing key differences in inflammatory responses and subtype distributions.Infographic on scRNA-seq neutrophil subtypes, cell communication and inflammatory response in MPR vs NMPR. ScRNA reveals neutrophil heterogeneity and intercellular communication networks associated with differential inflammatory responses between NMPR and MPR groups. (A) Cell-cell communication analysis among 9 cell subgroups in scRNA-seq data. The thickness of the lines represents the strength of cell-cell communication. (B) Visualization of cell-cell communication between neutrophils and other cell subgroups. (C) Visualization of cell-cell communication between endothelial cells and other cell subgroups. (D) UMAP visualization of neutrophil clustering. Each color represents a neutrophil subtype. (E) UMAP plot demonstrating normalized expression of markers for neutrophil subgroups. (F) Bar plot showing the percentage distribution of neutrophil subgroups in the NMPR and MPR groups. (G) The ridge plot illustrates the inflammatory response scores across neutrophil subgroups. Each color represents a neutrophil subtype. (H) Violin plot comparing differences of inflammatory response scores in the NMPR and MPR groups. (I) GSVA enrichment analysis depicting the enrichment degree of hallmark gene sets in neutrophil subgroups. (J–K) Violin plot comparing the difference in inflammatory response scores of (J) N3_CXCL8 and (J) N4_VCAN subgroups in NMPR and MPR groups. ***: p-value < 0.001. Among all intercellular interactions, the crosstalk between neutrophils and endothelial cells was identified as the most prominent signaling axis (Figure 3B and C). Further ligand-receptor pair analysis demonstrated that the strongest interaction signal from neutrophils to endothelial cells was mediated by NAMPT-(ITGA5+ITGB1) axis, while the ANXA1-FPR1 axis dominated the reverse signaling transduction from endothelial cells to neutrophils (Supplementary Figure 3A). Accumulating previous studies have validated that these ligand-receptor pairs exert dual regulatory effects in modulating inflammatory responses, exerting both pro-inflammatory and anti-inflammatory functions to further affect tumor malignant progression, immune evasion and pathological angiogenesis.33–35 Notably, the interaction intensity between neutrophils and endothelial cells in the NMPR group was significantly stronger than that in the MPR group (Supplementary Figure 3B). Among the enriched inflammatory signaling pathways, IL-1, oncostatin M (OSM) and chemokine ligand (CCL) signaling pathways exerted dual regulatory roles in tumor biological behavior and immune response, participating in the bidirectional modulation of inflammatory activation and further mediating the formation of immunosuppressive and pro-angiogenic TME31,36,37 (Supplementary Figure 3C). To further elucidate the regulatory mechanism underlying neutrophil-endothelial cell crosstalk, dimensionality reduction and subgroup clustering were conducted on the total neutrophil population, and five distinct neutrophil subsets were identified: N1_PGLYRP1, N2_APOC1, N3_CXCL8, N4_VCAN, and N5_CCL5 (Figure 3D). The top 10 characteristic marker genes of each neutrophil subset are presented in Supplementary Table 7 and visualized in Supplementary Figure 3D, with the distribution of these subset-specific markers illustrated via UMAP plot (Figure 3E). Cell proportion analysis demonstrated that the percentages of N3_CXCL8 and N4_VCAN subsets were markedly elevated in the NMPR group relative to the MPR group (Figure 3F). IRGs enrichment scoring revealed that the N3_CXCL8 subset possessed the highest inflammatory activation score, followed by the N4_VCAN subset (Figure 3G and H), and GSVA enrichment analysis yielded consistent results (Figure 3I). Moreover, the inflammatory response scores of N3_CXCL8 and N4_VCAN subsets were both significantly higher in the NMPR group than those in the MPR group (Figure 3J and K; p < 0.05). Further pathway enrichment analysis indicated that the N3_CXCL8 subset exhibited prominent enrichment of multiple signaling pathways, including anti-tumor immune-killing pathways (INF-α and INF-β), immune evasion-related pathways (IL-6/JAK/STAT3, IL-2/STAT5, TGF-β, EMT and angiogenesis), as well as hypoxia and glycolysis metabolic pathways (Figure 3I). The pathway activation profile of this subset was highly consistent with that of cluster B identified via consensus clustering (Figure 1D). In contrast, the N4_VCAN subset presented notably downregulated immune-killing signaling, alongside lower enrichment of immune evasion pathways compared with the N3_CXCL8 subset; however, this subset was highly enriched in hypoxia and glycolysis signaling, which was consistent with the pathway characteristics of cluster D (Figure 1D, I and Supplementary Figure 3E and F). GSVA enrichment analysis targeting immune cell infiltration and immune functional genes further validated the above findings. The N3_CXCL8 subset showed significant upregulation and enrichment of immune cell-related markers and immune effector functions, whereas the N4_VCAN subset exhibited the opposite expression trend. These results were highly consistent with the TME characteristics of cluster B and cluster D, further confirming the pivotal regulatory role of inflammatory response in tumor immune microenvironment remodeling (Figure 1D, 3I and Supplementary Figure 3G). A recent report in Science described how tumor-infiltrating neutrophils aggregate and develop into a long-lived T3 cell subgroup. This subgroup is located in a unique hypoxia-glycolysis niche within the tumor and promotes angiogenesis.38 This is consistent with our findings of the biological characterization of the N3_CXCL8 and N4_VCAN subgroups. Subsequently, we selected CD11b and CD16 as marker genes for neutrophil localization, CXCL8 and IGKC as marker genes for the N3_CXCL8 subgroup, VCAN and THBS1 as marker genes for the N4_VCAN subgroup, and HIF1A as the gene for hypoxia localization. Spatial transcriptomics results indicated that the N3_CXCL8 and N4_VCAN subgroups were colocalized with hypoxic regions (Figure 4A). To further validate this observation, we extracted grayscale FeaturePlot images of representative markers and performed fluorescence-like co-localization analysis using ImageJ (Coloc 2 plugin). This analysis revealed strong spatial overlap between HIF1A and markers associated with the N3_CXCL8 and N4_VCAN subgroups (Supplementary Figure 3H and I). In addition, we conducted distance-based expression correlation analysis using GraphPad Prism, confirming the spatial consistency of N3_CXCL8 and N4_VCAN subgroups with proximity to hypoxic cores, thereby reinforcing their localization within hypoxic niches (Supplementary Figure 3J and K). Next, we further explored the pro-angiogenic functions of the N3_CXCL8 and N4_VCAN subgroups. Intercellular communication analysis revealed that the N3_CXCL8 and N4_VCAN subgroups had the most obvious interactions with endothelial cells (Supplementary Figure 4A and B). In terms of communication signals, we found that the hallmark VEGF signal for angiogenesis was significantly more active in the NMPR group compared to the MPR group. Additionally, the NMPR group exhibited more active immune evasion signals, including ANGPTL, OSM, IL-16, IL-1, CCL, SPP1, and CXCL signals (Figure 4B). Receptor-ligand analysis revealed that endothelial cells interact with neutrophil subgroups via ANGPTL2 - (ITGA5+ITGB1) (Figure 4C and Supplementary Figure 4C). The N4_VCAN subgroup interact with endothelial cells via VEGFA - VEGFR1 (Figure 4C and Supplementary Figure 4D). The N3_CXCL8 subgroup interact with endothelial cells via OSM - (LIFR+IL6ST) (Figure 4C and Supplementary Figure 4E). Studies have demonstrated that these interactions are involved in tumor-associated inflammation, immune evasion and angiogenesis, thereby promoting tumor progression.35,36,39–41 Metabolic analysis revealed that cells in the N4_VCAN subgroup had vigorous metabolic activity, whereas the N3_CXCL8 subgroup had a low level of metabolic activity. Notably, nitrogen metabolism was significantly up-regulated in both the N3_CXCL8 and N4_VCAN subgroups (Figure 4D). In the TME, active ammonia metabolism not only inhibits the proliferation and function of T cells and NK cells but also enhances the activity of Tregs and MDSCs, promoting immunosuppression. Moreover, ammonia metabolism alters the metabolic state of immune cells, affecting their energy supply and function, ultimately enabling tumors to suppress immune cells through metabolic competition and promote tumor growth and progression.42,43 In addition, the N3_CXCL8 and N4_VCAN subgroup marker gene signature scores indicated that the survival of the high-scoring group was significantly worse than the low-scoring group (Figure 4E and F). To infer neutrophil lineage dynamics, we constructed a pseudotime trajectory using the Monocle2 algorithm. Although Monocle2 supports automatic root state selection, the auto-inferred root in our dataset was located within the N3_CXCL8 cluster, which is biologically implausible given its activated phenotype and enrichment for pro-inflammatory pathways. Therefore, the root state was manually set to the N1_PGLYRP1 subpopulation, which exhibited the highest expression of PGLYRP1, a marker of immature or resting neutrophils, and minimal activation of inflammatory or angiogenic programs. Based on this root assignment, pseudotime analysis revealed a developmental path originating from N1_PGLYRP1 and diverging toward two terminal endpoints—N3_CXCL8 and N4_VCAN (Figure 5A–C). Branch-point analysis further demonstrated that the developmental endpoints of N3_CXCL8 and N4_VCAN were characterized by the presence of VEGF signaling, which promotes angiogenesis (Figure 5D and E). Based on the above analysis, we found that the high levels of inflammatory response in the N3_CXCL8 and N4_VCAN subgroups promote tumor angiogenesis and mediate immune evasion in lung cancer. Figure 4Spatial and prognostic characteristics analysis of CXCL8⁺ and VCAN⁺ neutrophil subgroups. (A) Spatial transcriptomics revealed the co-localization of the N3_CXCL8 and N4_VCAN subgroups with hypoxia in the TME. (B and C) Cell-cell communication signals between neutrophil subgroups and endothelial cells. (D) Bubble plot demonstrating metabolic pathway analysis of neutrophil subgroups. (E and F) Prognostic analysis comparing groups with high and low signature gene scores for the (E) N3_CXCL8 and (F) N4_VCAN subgroups in the TCGA dataset.This infographic analyzes neutrophil subgroups, focusing on spatial markers, signaling pathways, metabolic activity and survival outcomes. The layout includes multiple panels labeled A to F. Panel A shows spatial maps with markers CD11b, CD16, HIF1A, CXCL8, IGKC, VCAN and THBS, highlighting their co-localization. Panel B presents a bar chart comparing NMPR and MPR groups, with NMPR showing higher relative information flow in most pathways. Panel C illustrates pathway networks for ANGPTL, VEGF and OSM, connecting neutrophil subgroups and endothelial cells. Panel D features a bubble plot of metabolic pathways, indicating vigorous activity in sulfur metabolism and glycolysis. Panels E and F display Kaplan-Meier survival curves, showing significantly better survival for low signature score groups. Overall, the infographic highlights the role of neutrophil subgroups in tumor progression and immune evasion.Neutrophil subgroups: markers, pathways, metabolism, survival analysis. Figure 5Pseudotime trajectory and functional analysis of neutrophil subgroups. (A–C) Pseudotime trajectory analysis of neutrophil subgroups. (D–E) KEGG enrichment analysis of branch point genes along the pseudotime trajectory in the (D) N3_CXCL8 and (E) N4_VCAN subgroups.A) A scatter plot displays neutrophil subgroups on a two-component embedding with a branching trajectory. The X-axis is labeled Component 1 and the Y-axis is labeled Component 2. Points are grouped by cell type: N1PGLYRP1, N2APOC1, N3CXCL8, N4VCAN, N5CCL5. Three trajectory arms converge near the origin, with nodes labeled 1 to 4. N1PGLYRP1 forms a vertical arm, N4VCAN clusters at high Component 1, N2APOC1 follows the lower right arm, N3CXCL8 clusters at the vertical arm′s top and N5CCL5 forms a smaller cluster in the lower right. B) The same plot is colored by pseudotime, with lower values near the root and higher values at branch ends. C) Five scatter plots show individual cell types with the same trajectory and nodes. D) A heatmap titled ′KEGG enrichment analysis′ shows pathway names and a matrix of colored cells, divided into blocks along the trajectory. E) A second heatmap with a similar structure shows branch-specific gene sets. Units for components and pseudotime are not shown.Multi-panel scatter and heatmap plots of neutrophil pseudotime branching and KEGG enrichment patterns. Spatial and prognostic characteristics analysis of CXCL8⁺ and VCAN⁺ neutrophil subgroups. (A) Spatial transcriptomics revealed the co-localization of the N3_CXCL8 and N4_VCAN subgroups with hypoxia in the TME. (B and C) Cell-cell communication signals between neutrophil subgroups and endothelial cells. (D) Bubble plot demonstrating metabolic pathway analysis of neutrophil subgroups. (E and F) Prognostic analysis comparing groups with high and low signature gene scores for the (E) N3_CXCL8 and (F) N4_VCAN subgroups in the TCGA dataset. Pseudotime trajectory and functional analysis of neutrophil subgroups. (A–C) Pseudotime trajectory analysis of neutrophil subgroups. (D–E) KEGG enrichment analysis of branch point genes along the pseudotime trajectory in the (D) N3_CXCL8 and (E) N4_VCAN subgroups. Taken together, multiple lines of evidence support that inflammatory neutrophil–endothelial cell interactions contribute to both tumor angiogenesis and immune evasion in the NMPR group. These include enriched neutrophil–endothelial signaling via OSM–LIFR/IL6ST and VEGFA–VEGFR1 pairs, immunosuppressive and pro-angiogenic features of neutrophil subsets N3_CXCL8 and N4_VCAN, spatial colocalization with hypoxic niches, and pseudotime trajectories indicating their differentiation toward immunoregulatory endpoints. This integrative analysis highlights the pathogenic role of neutrophil–stromal crosstalk in poor pathological response. Overexpression of OLR1 in Macrophages and Monocytes Promotes Tumor Angiogenesis and Immune Evasion Based on these findings, we investigated potential targets to enhance the efficacy of combined immunotherapy and chemotherapy. Importantly, N3_CXCL8 and N4_VCAN neutrophil subpopulations, which were identified in our previous analyses, exhibited strong enrichment of inflammatory response and angiogenesis-related biological programs. Therefore, they were selected as the key cell populations for downstream candidate gene prioritization. We then performed differential gene expression analysis between the NMPR and MPR groups within the N3_CXCL8 and N4_VCAN subsets, using all DEGs. To prioritize biologically relevant candidates associated with tumor progression and immune regulation, these DEGs were intersected with predefined angiogenesis-related and inflammatory response-related gene sets from the MSigDB Hallmark collection. Importantly, this intersection strategy was designed as a candidate gene prioritization approach rather than a definitive inference of key regulatory drivers, aiming to narrow down genes simultaneously associated with pathological response, inflammation, and angiogenesis. This analysis identified OLR1 as a candidate gene for further validation (Figure 6A). We observed that OLR1⁺ neutrophils were predominantly enriched in the N3_CXCL8 and N4_VCAN subclusters (Figure 6B). To further assess OLR1 expression across treatment response groups, we projected the Harmony-integrated dataset (same as Figure 2A) onto separate UMAPs for MPR, NMPR, and pCR samples, including all annotated cell types. The shape variations among the UMAPs reflect underlying differences in cellular composition across the three groups. Consistently, the single-cell dot plot and split UMAP visualization showed a higher density and broader distribution of OLR1-positive cells in the NMPR group than in the MPR and pCR groups, suggesting an enrichment of OLR1-expressing myeloid populations in patients with poor pathological response (Figure 6C–E). To further delineate the individual-level distribution of OLR1 expression, we performed a sample-level analysis treating each patient as an independent statistical unit. The single-cell dot plot revealed substantial inter-patient heterogeneity in OLR1 expression, with NMPR patients exhibiting higher OLR1 levels and a greater proportion of OLR1-positive cells compared with MPR patients (Supplementary Figure 5A and Supplementary Table 3). Consistently, the bubble plot summarized the average OLR1 expression and the percentage of expressing cells per sample, further confirming elevated OLR1 expression in patients with poor pathological response at the individual level (Supplementary Figure 5B). IHC staining further demonstrated the differential expression of OLR1 in lung cancer samples with different pathological responses, and this differential expression was significant (Figure 6F–J and Supplementary Table 1; p value < 0.05). IF staining revealed that OLR1 expression was markedly higher in NMPR patient samples compared to MPR and pCR groups, whereas CD8⁺ T cell infiltration was significantly lower in NMPR samples than in the MPR and pCR groups (Figure 6K). In addition, searching the literature revealed that OLR1 can be used as a marker for polymorphonuclear myeloid-derived suppressor cells (PMN-MDSCs),44 which supports our above findings. Interestingly, macrophages showed the highest OLR1 expression among all cell subpopulations, followed by monocytes (Figure 6E). Figure 6Expression and functional analysis of OLR1 in nICT for TME in lung cancer. (A) Differential gene analysis of the N3_CXCL8 and N4_VCAN subgroups between the NMPR and MPR groups, and intersection with inflammatory response and angiogenesis gene sets. (B) UMAP plot presenting OLR1 expression levels in neutrophil subgroups. (C and D) (C) Scatter plot and (D) UMAP visualization of OLR1 expression across the pCR, MPR, and NMPR groups in neoadjuvant-treated patients. (E) Scatter plots demonstrating OLR1 expression levels in cell subgroups. (F–I) IHC staining comparing the differential expression levels of OLR1 in NMPR, MPR, pCR, and adjacent normal tissues. (J) Quantification of OLR1-positive staining area (%) in NMPR, MPR, pCR, and adjacent normal tissues using ImageJ analysis. (K) IF staining of CD8+ T cells (green) and OLR1 (red) in tumor samples from NMPR, MPR, and pCR patients. Scale bar: 50 μm. ****: p-value < 0.0001; ns: not significant.A research figure on OLR1 expression in lung cancer and treatment-response groups is presented. Panel A features a Venn diagram showing OLR1′s overlap in angiogenesis and inflammatory response gene sets. Panel B includes a UMAP plot of OLR1 expression in neutrophil subgroups. Panel C shows a scatter plot with higher OLR1 expression in NMPR compared to MPR and pCR. Panel D provides UMAP visualizations of OLR1 in MPR, NMPR and pCR groups. Panel E displays a scatter plot of OLR1 across cell types, highlighting macrophages and monocytes. Panels F to I compare IHC staining of OLR1 in NMPR, MPR, pCR and normal tissues at different magnifications. Panel J quantifies OLR1 expression, showing NMPR as highest. Panel K shows IF staining of CD8 T cells and OLR1 in tumor samples, indicating lower CD8 infiltration in NMPR. The layout emphasizes higher OLR1 in NMPR.Multi-panel figure on OLR1 expression in lung cancer, highlighting higher levels in NMPR vs MPR/pCR/normal groups. Expression and functional analysis of OLR1 in nICT for TME in lung cancer. (A) Differential gene analysis of the N3_CXCL8 and N4_VCAN subgroups between the NMPR and MPR groups, and intersection with inflammatory response and angiogenesis gene sets. (B) UMAP plot presenting OLR1 expression levels in neutrophil subgroups. (C and D) (C) Scatter plot and (D) UMAP visualization of OLR1 expression across the pCR, MPR, and NMPR groups in neoadjuvant-treated patients. (E) Scatter plots demonstrating OLR1 expression levels in cell subgroups. (F–I) IHC staining comparing the differential expression levels of OLR1 in NMPR, MPR, pCR, and adjacent normal tissues. (J) Quantification of OLR1-positive staining area (%) in NMPR, MPR, pCR, and adjacent normal tissues using ImageJ analysis. (K) IF staining of CD8+ T cells (green) and OLR1 (red) in tumor samples from NMPR, MPR, and pCR patients. Scale bar: 50 μm. ****: p-value < 0.0001; ns: not significant. To further investigate whether OLR1 has a similar function in macrophages and monocytes as in neutrophil subgroups, we performed dimensionality reduction clustering of macrophages and monocytes again. For macrophages, we obtained 6 subgroups (Supplementary Figure 6A, B and Supplementary Table 8), in which OLR1 was mainly expressed in macrophage_1, macrophage_2, macrophage_4 and macrophage_5 (Supplementary Figure 6C and D). Using the same GSVA enrichment analysis, we found that the macrophage_4 subgroup, similar to the N3_CXCL8 subgroup, exhibited the highest inflammatory response signaling and up-regulated tumor immune-killing, immune-evasion, hypoxia, glycolysis, and angiogenesis signals. The macrophage_1, macrophage_2 and macrophage_5 subgroups were all exhibited downregulation of tumor immune killing signals. Macrophage_1 and macrophage_2 subgroups were similar to the N4_VCAN subgroup, showing low enrichment of inflammatory response and immune-evasion signaling, but high enrichment of hypoxia, glycolysis, and angiogenesis pathways. However, the macrophage_5 subgroup only showed low enrichment in angiogenesis (Supplementary Figure 6E). GSVA enrichment analysis further revealed that immune cells and functional genes were enriched and downregulated in the macrophage_1 and macrophage_2 subgroups, enriched and upregulated in the macrophage_4 subgroup, while enrichment in the macrophage_5 subgroup was intermediate between these groups (Supplementary Figure 6F). For monocytes, dimensionality reduction clustering identified 7 cell subgroups (Supplementary Figure 7A, B and Supplementary Table 9), with OLR1 predominantly expressed in the monocyte_1 and monocyte_2 subgroups (Supplementary Figure 7C and D). Similarly, pathway enrichment analysis revealed that the monocyte_1 subgroup exhibited a consistent active state with the N3_CXCL8 subgroup, and the monocyte_2 subgroup exhibited a consistent active state with the N4_VCAN subgroup (Supplementary Figure 7E). Consistent phenomena were also observed in the immune cell and functional enrichment of monocyte subpopulations (Supplementary Figure 7F). Finally, using bulk RNA-seq data, we divided the lung cancer population into a high OLR1 expression group and a low OLR1 expression group, and compared the differences between these groups using a consistent GSVA enrichment approach. The results demonstrated that tumor immune killing, immune evasion and angiogenesis signals were gradually upregulated with increasing OLR1 expression levels in the population, suggesting that high OLR1 expression may promote tumor immune evasion and angiogenesis (Supplementary Figure 8). THP‑1 is a human monocytic cell line, which can differentiate into macrophages upon PMA induction. In vitro experiments demonstrated that after OLR1 transfection of THP-1 cells, WB showed a significant elevation of OLR1 protein expression (Figure 7A and B). ELISA experiments showed that THP-1 cells overexpressing the OLR1 gene had a significantly elevated ability to secrete TGF-β and VEGF (Figure 7C; p value < 0.05). To investigate the regulatory effect of OLR1⁺macrophages on tumor cell behavior, an indirect transwell co-culture system was established, in which OLR1-overexpressing THP-1 macrophages were seeded in the upper chamber and NCI-H2122 cells were seeded in the lower chamber (Figure 7D). As shown by the CCK-8 assay, NCI-H2122 cells co-cultured with OLR1-overexpressing macrophages exhibited significantly increased proliferation over time compared with the control group (Figure 7E; p < 0.05). This pro-proliferative effect was further confirmed by the colony formation assay, where a markedly higher number of crystal violet–stained colonies was observed in the OLR1 overexpression group (Figure 7F and G; p value < 0.05). In the wound healing assay, NCI-H2122 cells exposed to OLR1-overexpressing macrophages demonstrated accelerated scratch closure at 24 hours, indicating enhanced migratory capacity (Figure 7H and I; p value < 0.05). Transwell assays also revealed significantly increased numbers of migrated and invaded NCI-H2122 cells in the OLR1 overexpression group compared with controls (Figure 7J–L; p value < 0.05). Collectively, these findings suggest that macrophage-derived OLR1 may promote the proliferation, migration, and invasion of NCI-H2122 cells, potentially through the secretion of soluble factors within the TME. Figure 7Functional analysis of OLR1⁺ macrophages and their effects on the proliferation, migration, and invasion of NCI-H2122. (A and B) (A) WB assay for relative expression and (B) quantification of OLR1 protein in OV-OLR1 and OV-NC groups in THP-1 cells. (C) ELISA for TGF-β and VEGF secretion levels and quantification in OV-OLR1 and OV-NC groups in THP-1 cells. (D) Schematic diagram of the co-culture system: THP-1 macrophages transfected with OV-NC or OV-OLR1 were seeded in the upper chambers of a transwell insert, while NCI-H2122 cells were seeded in the lower chambers to allow indirect interaction via soluble factors. (E) CCK-8 assay showing the relative proliferation of NCI-H2122 cells co-cultured with OV-NC or OV-OLR1 THP-1 cells over 3 days (n = 3). (F and G) Colony formation assay of NCI-H2122 cells after co-culture. (F) Representative crystal violet–stained images and (G) quantification of colony number (n = 3). (H and I) Wound healing assay showing enhanced migration of NCI-H2122 cells following co-culture with OLR1-overexpressing macrophages. (H) Representative images at 0 h and 24 h; (I) quantification of wound closure percentage (n = 3). Scale bar: 200 μm. (J–L) Transwell invasion and migration assays. (J) Representative images showing invaded and migrated NCI-H2122 cells stained with crystal violet. (K and L) Quantification of invaded and migrated cell numbers (n = 3). Scale bar: 100 μm.The multi-panel experimental figure is labeled A through L and arranged in rows, read left to right and top to bottom. Group labels include OV-NC and OV-OLR1 and cell labels include THP-1 and NCI-H2122. A shows WB assay bands labeled OLR1 and GAPDH, with molecular weights 55 kDa and 36 kDa, comparing OV-OLR1 versus OV-NC. B is a bar chart titled THP-1 with y-axis label Relative protein expression and values from 0.00 to 0.05. Two bars labeled OV-NC and OV-OLR1 are shown with P equals 0.0338 and OV-OLR1 is higher. C contains two bar charts titled THP-1. Left chart y-axis label TGF-beta (pg/mL) with values 0 to 200, P equals 0.0010, OV-OLR1 higher than OV-NC. Right chart y-axis label VEGF (pg/mL) with values 0 to 250, P equals 0.0007, OV-OLR1 higher than OV-NC. D is a schematic diagram labeled THP-1, PMA, OLR1 plasmid, OLR1 overexpression, NCI-H2122 and Cell co-culture, showing THP-1 treated with PMA, transfected with OLR1 plasmid, then co-cultured with NCI-H2122 in a two-chamber insert setup. E is a line graph titled NCI-H2122 with x-axis label Time (day) and ticks 0, 1, 2, 3, 4 and y-axis label Relative Proliferation with ticks 0, 1, 2, 3. Two series labeled OV-NC and OV-OLR1 are shown with P equals 0.0010 and OV-OLR1 is higher across days 1 to 3. F shows two colony formation assay images labeled OV-NC and OV-OLR1. G is a bar chart titled NCI-H2122 with y-axis label Count and values 0 to 600, P equals 0.0145, with OV-OLR1 higher than OV-NC. H shows wound healing assay images labeled OV-NC and OV-OLR1 at 0 h and 24 h. I is a bar chart titled NCI-H2122 with y-axis label Healing rate (percent) and values 0 to 60, P equals 0.0009, with OV-OLR1 higher than OV-NC. J shows transwell assay images with row labels Invasion and Migration and column labels OV-NC and OV-OLR1. K is a bar chart titled NCI-H2122 with y-axis label Invasion (count) and values 0 to 800, P equals 0.0085, with OV-OLR1 higher than OV-NC. L is a bar chart titled NCI-H2122 with y-axis label Migration (count) and values 0 to 800, P equals 0.0435, with OV-OLR1 higher than OV-NC.A multi-panel experimental figure showing OV-OLR1 increases THP-1 secretion and NCI-H2122 behaviors. Functional analysis of OLR1⁺ macrophages and their effects on the proliferation, migration, and invasion of NCI-H2122. (A and B) (A) WB assay for relative expression and (B) quantification of OLR1 protein in OV-OLR1 and OV-NC groups in THP-1 cells. (C) ELISA for TGF-β and VEGF secretion levels and quantification in OV-OLR1 and OV-NC groups in THP-1 cells. (D) Schematic diagram of the co-culture system: THP-1 macrophages transfected with OV-NC or OV-OLR1 were seeded in the upper chambers of a transwell insert, while NCI-H2122 cells were seeded in the lower chambers to allow indirect interaction via soluble factors. (E) CCK-8 assay showing the relative proliferation of NCI-H2122 cells co-cultured with OV-NC or OV-OLR1 THP-1 cells over 3 days (n = 3). (F and G) Colony formation assay of NCI-H2122 cells after co-culture. (F) Representative crystal violet–stained images and (G) quantification of colony number (n = 3). (H and I) Wound healing assay showing enhanced migration of NCI-H2122 cells following co-culture with OLR1-overexpressing macrophages. (H) Representative images at 0 h and 24 h; (I) quantification of wound closure percentage (n = 3). Scale bar: 200 μm. (J–L) Transwell invasion and migration assays. (J) Representative images showing invaded and migrated NCI-H2122 cells stained with crystal violet. (K and L) Quantification of invaded and migrated cell numbers (n = 3). Scale bar: 100 μm. Overall, we found that OLR1 expression was predominantly distributed in a subset of myeloid cells and macrophages with immunosuppressive and pro-angiogenic effects, promising a potential target for improving the efficacy of neoadjuvant immunotherapy in combination with chemotherapy. Anti-Angiogenic Drug May Potentially Interact with OLR1 Classical VEGF signaling is a central regulator of tumor-associated inflammatory activation, immune suppression, and pathological angiogenesis, and VEGF-targeted anti-angiogenic agents have been widely applied in clinical cancer therapy.41,45,46 Given that OLR1 was identified in our study as a hub gene associated with angiogenesis-, inflammation-, and immunosuppression-related biological processes in nICT-treated lung cancer, we performed molecular docking as an exploratory in silico approach to evaluate the potential structural compatibility between OLR1 and clinically approved anti-angiogenic agents (bevacizumab and anlotinib). Molecular docking verified that bevacizumab binds well to the OLR1 receptor pocket, with a binding energy of –7.3 kcal/mol, indicating a stable interaction (Supplementary Figure 9A and Supplementary Table 10). Two stable hydrogen bonds were formed with key residues SER‑207 and PRO‑214 (Supplementary Figure 9B). The two‑dimensional interaction map revealed that van der Waals forces, involving residues Arg208, Ser207 and Leu216, are the main non‑covalent forces maintaining the binding conformation (Supplementary Figure 9C). In addition, extensive non‑bonding contacts, dominated by electrostatic potential and van der Waals forces, further stabilized the interaction (Supplementary Figure 9D). Anlotinib showed an even lower binding energy (–8.0 kcal/mol) with OLR1, suggesting a more stable ligand–receptor conformation (Supplementary Figure 9E and Supplementary Table 11). It formed three hydrogen bonds with TRP‑217, LEU‑216 and SER‑207 (Supplementary Figure 9F). The two‑dimensional interaction map illustrated these hydrogen bonds as well as non‑covalent interactions with surrounding residues including Arg209, Glu218 and Ser240 (Supplementary Figure 9G). Similarly, extensive non‑bonding contacts dominated by electrostatic potential and van der Waals forces were observed (Supplementary Figure 9H). However, it is important to emphasize that these computational findings reflect only theoretical molecular interactions and do not constitute evidence of direct pharmacological targeting or functional modulation. Rather, these in silico observations should be interpreted as preliminary and hypothesis-generating, providing supportive computational context for the OLR1-associated biological programs that emerged from our multi-omic profiling. Given the ongoing clinical interest in combining anti-angiogenic agents with nICT to overcome resistance, these computational results may offer a theoretical rationale for future mechanistic investigations. Therefore, our results may provide preliminary, hypothesis-generating support for further investigation of OLR1-associated pathways in the context of nICT combined with anti-angiogenic treatment strategies (Figure 8). Figure 8Schematic diagram of nICT with anti-OLR1 treatment. (A) DGE analysis identified OLR1 as a key target. OLR1⁺ neutrophils, monocytes, and macrophages were involved in tumor-associated inflammation, immune evasion, and angiogenesis. (B) Targeted inhibition of OLR1⁺ neutrophils, monocytes, and macrophages within the TME may enhance the efficacy of nICT in lung cancer patients by suppressing tumor angiogenesis, immune evasion, and tumor-associated inflammatory responses.The diagram illustrates the nICT with anti-OLR1 treatment. Image A shows tumor-associated inflammation, immune evasion and angiogenesis involving OLR1 superscript + neutrophils, monocytes and macrophages. Image B depicts targeted inhibition of these cells using an OLR1 inhibitor, leading to anti-angiogenesis, anti-immune evasion and antitumor inflammation. The process is linked to neoadjuvant therapy, combining immunotherapy and chemotherapy with anti-OLR1 therapy, targeting lung cancer treatment. Arrows indicate the flow from cellular interactions to therapeutic applications.Schematic diagram of nICT with anti‑OLR1 treatment in lung cancer tumor microenvironment. Schematic diagram of nICT with anti-OLR1 treatment. (A) DGE analysis identified OLR1 as a key target. OLR1⁺ neutrophils, monocytes, and macrophages were involved in tumor-associated inflammation, immune evasion, and angiogenesis. (B) Targeted inhibition of OLR1⁺ neutrophils, monocytes, and macrophages within the TME may enhance the efficacy of nICT in lung cancer patients by suppressing tumor angiogenesis, immune evasion, and tumor-associated inflammatory responses. Discussion Based on cellular origin and phenotypic characteristics, MDSCs can be further categorized into two major subsets: PMN-MDSCs differentiated from granulocytic myeloid cells and monocytic MDSCs (M-MDSCs) derived from monocytic myeloid lineages.47 Among them, PMN-MDSCs are widely acknowledged as the primary population representing immunosuppressive neutrophils in the tumor context.18 Increasing studies have highlighted that pathologically activated neutrophils act as the core effector cells driving tumor-associated inflammatory activation.48–50 Preclinical studies have revealed that antibody-mediated neutrophil depletion can effectively alleviate TME inflammatory infiltration, thereby suppressing tumor occurrence and progression in orthotopic tumor mouse models.51–53 Mechanistically, activated neutrophils exert immunosuppressive functions by secreting a variety of pro-inflammatory cytokines, growth factors and chemokines, inducing the high expression of immunosuppressive mediators such as ARG1, reactive oxygen species (ROS) and nitric oxide (NO), which further inhibit CD8⁺ T cell activation and proliferation, and ultimately promote the formation of immunosuppressive TME.49,53,54 Meanwhile, activated neutrophils can secrete a series of pro-angiogenic factors including BV8, MMP9 and VEGFA, mediating the angiogenic switch during early tumorigenesis and promoting pathological angiogenesis to support tumor growth and metastasis.54–57 In addition, circulating monocytes or M-MDSCs can rapidly infiltrate into tumor tissues and differentiate into TAMs.58 Classically, TAMs are functionally divided into M1-type and M2-type subtypes: M1 macrophages induced by IFN-γ, LPS or GM-CSF exert pro-inflammatory and anti-tumor immune effects; in contrast, M2-type TAMs induced by cytokines such as IL-4 and IL-13 predominantly secrete anti-inflammatory mediators (IL-10, TGF-β, etc)., which aggravate TME immunosuppression and accelerate tumor angiogenesis.59,60 Notably, recent studies have confirmed that TAMs differentiated from M-MDSCs retain stronger immunosuppressive activity than those derived from conventional circulating monocytes, highlighting the pivotal role of MDSC-derived TAMs in tumor immune escape.61 Therefore, targeted intervention against these immunosuppressive myeloid cell subsets is considered a promising strategy to improve the response rate of tumor immunotherapy.62,63 In this study, differential gene expression analysis between NMPR and MPR groups was performed to screen key regulatory targets, and OLR1 was ultimately identified as the hub gene. Further validation demonstrated that OLR1 was significantly highly expressed in neutrophils, macrophages and monocyte subsets with pro-inflammatory, immunosuppressive and pro-angiogenic functions. A previous landmark study by Condamine et al has validated that OLR1 can be utilized as a specific surface marker for human PMN-MDSCs, which provides a theoretical basis for our phenotypic identification.44 Moreover, existing studies have confirmed that OLR1 overexpression in tumor tissues can promote tumor inflammatory activation and TME immunosuppression, which is highly consistent with the functional characteristics of OLR1 revealed in this study.44,64 Notably, different from previous studies, the present research further uncovered the novel biological function of OLR1 in mediating tumor pathological angiogenesis, expanding the regulatory network of OLR1 in tumor progression. Critically, our multi-omic profiling reveals that NMPR patients are characterized by a coordinated OLR1-centered inflammatory–immune–angiogenic axis, which is largely orthogonal to the mechanisms targeted by existing nICT components. This observation identifies a specific unmet clinical niche: patients with pre-existing or early-onset OLR1-high inflammatory signatures may derive limited benefit from current regimens and could potentially be rescued by adjunctive OLR1-directed intervention. Accordingly, targeted inhibition of OLR1 is expected to exert synergistic anti-tumor effects by simultaneously reversing chronic tumor-associated inflammation, alleviating TME immunosuppression and blocking pathological angiogenesis, providing a novel target for optimizing tumor treatment regimens. Importantly, OLR1 may also serve as a potential biomarker for stratifying patients with distinct inflammatory and immune suppression states, and for predicting pathological response to nICT in lung cancer. Its expression could be further developed into clinically applicable assays, such as immunohistochemistry-based scoring or transcriptomic signature panels, to facilitate patient selection and therapeutic decision-making. Future prospective studies are required to validate its predictive performance in independent clinical cohorts. Further functional analysis revealed that OLR1 exhibits highly consistent biological functions with classical VEGF signaling, a core target of clinical anti-angiogenic therapy. VEGF signaling drives tumor malignant progression by promoting pathological angiogenesis, inhibiting T cell immune activation and inducing the infiltration of immunosuppressive cells such as Tregs and MDSCs.41,45 Currently, VEGF-targeted anti-angiogenic drugs have been widely applied in clinical practice and achieved remarkable efficacy in inhibiting advanced tumor progression.46 Particularly, the landmark Phase III IMpower150 clinical trial confirmed that the triple combination regimen of immunotherapy, chemotherapy and anti-angiogenic therapy can significantly improve the long-term prognosis of patients with advanced lung cancer.65 Molecular docking simulation in this study verified that clinical first-line VEGF inhibitors exhibited strong binding affinity with OLR1 protein, suggesting that OLR1 and VEGF may share a similar regulatory mechanism in promoting tumor progression, and preliminarily validating the feasibility of OLR1 as a novel therapeutic target. In addition, clinical practice has confirmed that long-term use of VEGF inhibitors is often accompanied by serious adverse reactions including hypertension, proteinuria, bleeding and thromboembolism, which severely affect the quality of life of patients.41 Critically, however, the therapeutic novelty of OLR1 targeting does not reside in merely recapitulating VEGF inhibition. Rather, its mechanistic distinction lies in the concurrent modulation of three pathologically intertwined processes—neutrophil-driven inflammation, MDSC or TAM-mediated immunosuppression, and pathological angiogenesis—converging upon a single molecular node. Unlike anti-VEGF agents, which primarily interrupt angiogenic signaling with only secondary effects on myeloid infiltration, or immune checkpoint inhibitors, which relieve T-cell exhaustion without counteracting upstream myeloid inflammatory networks, OLR1 blockade operates at the myeloid–vascular interface to dismantle a self-reinforcing vicious cycle. This multipronged regulatory profile distinguishes OLR1 from single-pathway inhibitors and establishes it as an upstream integrative hub, whose inhibition may yield synergistic effects unattainable by simply co-administering anti-VEGF and anti-PD-1 or PD-L1 agents. Moreover, unlike broad-spectrum myeloid depletion strategies (eg, anti-Gr-1 or anti-Ly6G) that entail substantial toxicity risks, OLR1-directed intervention affords greater selectivity, attributable to its preferential expression on pathologically activated pro-tumoral myeloid subsets rather than on steady-state circulating granulocytes. Collectively, these considerations suggest that OLR1-targeted therapy may serve as a mechanistically distinctive and potentially safer alternative, achieving broader anti-tumor efficacy via simultaneous abrogation of inflammatory, immunosuppressive and angiogenic pathways. Patients achieving pathological response after nICT can benefit from preoperative systemic immune activation, which induces the formation of immune memory, reduces the risk of postoperative tumor recurrence and improves long-term survival.12,66 Meanwhile, nICT can induce tumor regression and downstaging, expand surgical indications and improve the rate of radical resection.12,67 However, for patients who fail to achieve effective pathological response, long-term neoadjuvant treatment may lead to tumor distant metastasis in some cases, resulting in the loss of radical surgical opportunities.12 Notably, anti-angiogenic therapy has not been included in the standard nICT regimen for lung cancer. Based on the findings of this study, we innovatively propose a novel neoadjuvant treatment strategy combining immunotherapy, chemotherapy and OLR1-targeted therapy. Tumor-associated inflammation and pathological angiogenesis run through the entire process of tumor development, and targeted intervention on these two links can effectively remodel the TME and inhibit tumor progression.17,41 The addition of OLR1-targeted anti-angiogenic therapy can effectively reduce the formation of tumor neovascularization, cut off the nutrient and oxygen supply of tumor tissues, inhibit tumor proliferation during neoadjuvant treatment, and reduce the risk of disease progression.45 Simultaneously, blocking OLR1 signaling can reverse TME immunosuppression, activate systemic anti-tumor immune response, synergize with immunochemotherapy to improve pathological response rate, and reduce the risk of distant metastasis during neoadjuvant treatment.67 Based on the findings of this study, we innovatively propose that adding OLR1-targeted therapy to the existing nICT backbone may represent a rational strategy to simultaneously counteract inflammation, immunosuppression and angiogenesis. Beyond this conceptual framework, we further propose a pragmatic translational roadmap to bridge our correlative findings with future clinical implementation. First, OLR1 expression levels, assessed by immunohistochemistry on pretreatment biopsy specimens or through a customized transcriptomic signature, could be incorporated into prospective nICT trials to stratify patients at high risk of non-response, enabling enrichment for adjunctive OLR1-targeted regimens. Second, dynamic monitoring of circulating OLR1⁺ myeloid cell frequencies during the first cycle of nICT may serve as an early on-treatment pharmacodynamic biomarker to distinguish impending MPR from NMPR, facilitating adaptive treatment modifications before radiological progression becomes apparent. Third, given the availability of OLR1-targeting small molecules and monoclonal antibodies in early-phase development for other disease contexts, expedited evaluation in humanized lung cancer models or patient-derived organoid cultures that recapitulate the nICT setting is warranted, particularly in combination with PD-1 inhibitors and platinum-based doublets, to verify whether OLR1 blockade can indeed reverse TME immunosuppression, activate systemic anti-tumor immune response, and synergize with immunochemotherapy to improve pathological response rates while reducing the risk of distant metastasis during neoadjuvant treatment. Collectively, these steps provide a concrete framework for moving beyond correlation toward clinical utility, while honestly acknowledging that the present work remains hypothesis-generating and requires dedicated interventional validation. Several limitations should be acknowledged in this study. First, the core conclusions of this study are primarily based on high-throughput sequencing data analysis and bioinformatic inference. Although multi-omics analyses integrating bulk RNA-seq, scRNA-seq, and spatial transcriptomics were performed, the study still relies substantially on computational prediction, and lacks sufficient in-depth in vitro and in vivo functional validation, which warrants further experimental confirmation. This limitation is partly attributed to the inherent biological characteristics of primary myeloid cells, including the short survival time of primary neutrophils in vitro and the extremely low transfection efficiency of primary macrophages and monocytes, which together hinder the establishment of stable gain- or loss-of-function experimental models. In addition, there are significant species-specific differences between human and murine myeloid cells in transcriptional profiles, immune regulatory pathways, and responses to external stimuli, which further limit the direct translational applicability of preliminary in vitro and potential in vivo experimental findings. Second, the clinical validation cohort was relatively small, which may limit the statistical robustness and generalizability of the findings. This limitation is further attributed to the inherent difficulty in obtaining post-neoadjuvant immunochemotherapy lung cancer specimens, as such clinical samples are limited in availability due to strict surgical indications, treatment timing, and patient eligibility criteria in real-world clinical practice. Third, while comprehensive statistical analyses including differential expression and pathway enrichment were performed, residual confounding factors inherent to retrospective datasets cannot be completely excluded. Finally, independent large-scale and prospective clinical cohorts are required to further validate the prognostic and therapeutic relevance of OLR1 in lung cancer patients receiving nICT. In future studies, we aim to optimize gene delivery strategies in human primary monocytes and macrophages, such as lentiviral transduction or CRISPR activation-mediated gene editing, to construct stable functional models for mechanistic investigation. In parallel, CD14⁺ primary myeloid cells isolated from clinical peripheral blood will be used to further validate the regulatory role and molecular mechanisms of OLR1 in human-derived immune cells, thereby providing more robust evidence to support its clinical translation. Nevertheless, several potential challenges should be considered when translating OLR1-targeted therapy into clinical application. As OLR1 is involved in lipid uptake and metabolic regulation in normal myeloid and endothelial cells, systemic inhibition may potentially affect physiological lipid metabolism and immune homeostasis, raising concerns regarding on-target toxicity. In addition, although OLR1 is enriched in immunosuppressive myeloid populations within the TME, its expression in certain normal immune cell subsets may limit absolute tumor specificity, which warrants careful evaluation of therapeutic selectivity. Furthermore, the feasibility of clinical translation remains to be established, as effective targeting strategies such as antibody-based blockade, small-molecule inhibitors, or RNA-based approaches require further optimization to ensure efficient delivery and tumor-specific targeting in vivo. Conclusions In conclusion, this study systematically elucidates the critical regulatory role of tumor-associated inflammatory activation in mediating the efficacy of nICT for resectable lung cancer, and identifies OLR1 as a novel actionable target for optimizing treatment outcomes. Our integrated transcriptomic and scRNA-seq analyses demonstrate that aberrantly activated neutrophils, macrophages and monocyte subsets dominated by MDSCs drive robust inflammatory response, immune evasion and pathological angiogenesis in the TME, which is closely correlated with insufficient pathological response and poor clinical prognosis in lung cancer patients.