Exploring the mechanism of prognostic genes associated with neuroendocrine differentiation and immunotherapy resistance in non-small cell lung cancer based on the bulk transcriptome and single cell RNA sequencing

Exploring the mechanism of prognostic genes associated with neuroendocrine differentiation and immunotherapy resistance in non-small cell lung cancer based on the bulk transcriptome and single cell RNA sequencing | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Exploring the mechanism of prognostic genes associated with neuroendocrine differentiation and immunotherapy resistance in non-small cell lung cancer based on the bulk transcriptome and single cell RNA sequencing Tong Li, Meidan Luo, Sibo Peng, Fan Ren, Jingliang Gan, Huangsheng Xie, and 4 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8816614/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 8 You are reading this latest preprint version Abstract Background: The challenge of immunotherapy resistance has become a current research hotspot. Previous studies showed that neuroendocrine differentiation (NED) might contribute significantly to the initiation and development of non-small cell lung cancer (NSCLC). However, the interplay between immunotherapy resistance and NED of NSCLC remains unclear. This article mainly explored the mechanism of NED-related genes (NEDRGs) and immunotherapy resistance related genes in NSCLC. Methods: The NSCLC and control samples, and the NSCLC samples with PD-1 blockade were selected from the public databases to obtain differentially expressed genes (DEGs). Then, univariate Cox regression analysis and the proportional hazards (PH) assumption test were adopted to obtain prognostic genes based on the DEGs and NEDRGs, and a risk model was built and validated. Then, the nomogram was established. Subsequently, gene set enrichment analysis (GSEA), immune analysis and drug sensitivity were adopted. The key cells were ascertained in a single cell RNA sequencing (scRNA-seq) dataset. And we collected specimens of six patients from Tianjin Medical University General Hospital for validation using Quantitative Real-Time Polymerase Chain Reaction (qRT-PCR) and Immunohistochemistry (IHC). Results: A total of 3 prognostic genes including RRM2, WDR76 and PLEKHH2 were obtained and the risk model could predict the survival outcomes of NSCLC patients. The nomogram had good predictive ability of survival rate for NSCLC. the GSEA results revealed that notable pathways in the 2 risk groups included cell cycle, and prognostic genes were all enriched in this pathway. Furthermore, activated memory CD4 T cells might contribute significantly while 109 drugs such as AZD6738 might be the candidate drugs of NSCLC. Epithelials were identified as the key population, wherein the expression of prognostic genes altered significantly across different epithelials developmental stages. PCR and IHC validation confirmed that RRM2 and WDR76 expression increased in cancer versus normal tissues and with advancing stage, while PLEKHH2 showed opposite trends. Conclusion: This study identified 3 prognostic genes (RRM2, WDR76 and PLEKHH2) to build risk model, and the risk model exhibited superior predictive accuracy of NSCLC. These results might provide new ideas for investigating novel therapeutic targets in NSCLC. Non-small cell lung cancer Neuroendocrine resistance Immunotherapy Prognostic genes Risk model Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Figure 10 Figure 11 1. Introduction Lung cancer, which originates from the bronchial mucosa or pulmonary glands, is a highly prevalent and lethal malignancy globally ( 1 ). Non-small cell lung cancer (NSCLC) accounts for approximately 85% of all cases, with its primary subtypes being lung adenocarcinoma (LUAD) and lung squamous cell carcinoma (LUSC) ( 2 , 3 ). Known risk factors for NSCLC include smoking, family history, radiation exposure, and chronic lung diseases ( 4 , 5 ). Although advancements in treatment modalities such as surgery, radiotherapy, chemotherapy, and targeted therapy have improved clinical outcomes, the early asymptomatic nature of NSCLC often leads to late-stage diagnoses. Together with tumor heterogeneity and therapy resistance, these factors contribute to the persistently low overall survival rates ( 6 , 7 ). Thus, identifying novel prognostic biomarkers and elucidating the molecular mechanisms of NSCLC are urgently needed. In recent years, immunotherapy has emerged as a promising approach for cancer treatment. Immune checkpoint inhibitors (ICIs), particularly those targeting programmed death-1 (PD-1) and its ligand (PD-L1), have become a standard treatment for NSCLC, notably prolonging survival in advanced-stage patients either as monotherapy or in combination with chemotherapy ( 8 , 9 ). However, the majority of NSCLC patients exhibit limited response to ICIs, and primary or acquired resistance remains a major clinical challenge ( 10 ). Neuroendocrine differentiation (NED), characterized by the presence of tumor cells expressing neuroendocrine markers such as CD56, synaptophysin, and chromogranin A, has been implicated in therapy resistance and is an independent poor prognostic factor in NSCLC( 11 ). Although the role of NED is well-established in prostate cancer, its interplay with immunotherapy resistance in NSCLC is not fully understood ( 12 , 13 ). Therefore, investigating the mechanisms underlying NED and immunotherapy resistance is critical for developing personalized treatment strategies. Single-cell RNA sequencing (scRNA-seq) represents a transformative technology that enables the dissection of transcriptional heterogeneity at the cellular level. Unlike bulk RNA sequencing, scRNA-seq reveals cell-type-specific gene expression patterns, cellular subpopulations, and dynamic state transitions, providing unprecedented insights into tumor microenvironments and cell-cell communication networks ( 14 – 16 ). This approach has proven invaluable for identifying rare malignant cells and understanding cancer biology. The application of scRNA-seq technology has advanced our understanding of tumor heterogeneity, immune microenvironment, treatment response and drug resistance mechanisms in NSCLC, providing a more solid research foundation for the diagnosis and treatment of NSCLC. In this study, we integrated bulk transcriptomic data from public NSCLC datasets and scRNA-seq data to identify prognostic genes associated with NED and immunotherapy resistance. Using machine learning algorithms, we constructed a risk model and performed comprehensive analyses of biological functions, immune infiltration, and drug sensitivity. Collectively, our results offer novel insights into NSCLC pathogenesis and highlight potential targets for augmenting the efficacy of immunotherapeutic interventions. 2. Methods 2.1 Sample collection and sequencing We retrieved the TCGA-NSCLC dataset from The Cancer Genome Atlas ( https://xena.ucsc.edu/ ), which consisted of 989 tumor samples with survival data and 109 control samples. In addition, the clinical data (age, gender, stage, and TNM stage) and somatic mutation data of 989 NSCLC samples were downloaded, and the access time was March 14, 2025. A total of 13 NSCLC tissue samples before PD-1 blockade therapy and 29 NSCLC tissue samples after PD-1 blockade therapy from the GSE248249 dataset (GPL23126), a total of 196 tissue samples of NSCLC with survival information in the GSE37745 dataset (GPL570), and a total of 15 NSCLC samples from the GSE207422 dataset (GPL24676) were included from the Gene Expression Omnibus (GEO) database ( https://www.ncbi.nlm.nih.gov/geo/ ), comprising 3 NSCLC tissue samples PD-1 blockade therapy and 12 NSCLC tissue samples after PD-1 blockade therapy. The 1482 NEDRGs were obtained from the literature ( 17 ) ( Additional file 1 ). To validate the bioinformatic findings, we collected paired tumor and normal tissues from six NSCLC patients (three stage I patients and three stage III patients) undergoing surgery at Tianjin Medical University General Hospital for PCR and IHC validation. 2.2 Identification of differentially expressed genes (DEGs) and candidate genes First, we identified DEGs between NSCLC and control samples in the TCGA-NSCLC dataset using "DESeq2" (v 1.38.0( 18 )) (P 0.5), naming this gene set DEGs1. Then, we visualized DEGs1 in a volcano plot via the "ggplot2" package (v 3.4.1( 19 )), labeling the top 10 most significantly up- and down-regulated genes. Then, the top 10 up or down DEGs1 between the NSCLC and control samples were also depicted via the “ComplexHeatmap” package (v 2.14.0) ( 20 ). Using the "limma" package (v 3.54.1) ( 21 ), we identified DEGs between pre- and post-PD-1 blockade therapy samples in GSE248249 (P 0.5) and defined this gene set as DEGs2. The results were depicted by the same methods mentioned earlier. Candidate genes were defined as the common set resulting from the intersection of DEGs1, DEGs2, and NEDRGs, identified using the "ggvenn" package (v 0.1.9) ( 22 ). 2.3 Enrichment pathways and protein-protein interaction (PPI) network The Kyoto Encyclopedia of Genes and Genomes enrichment (KEGG) and Gene Ontology (GO) analyses were applied via the “clusterProfiler” package (v 4.2.2) ( 23 ) to explore the functions of candidate genes (P < 0.05). The top 5 notably enriched terms or all terms in GO and all enriched terms in KEGG were depicted in this study. To investigate protein-level interactions, a PPI network for the candidate genes was constructed using the Search Tool for the Retrieval of Interacting Genes (STRING) database ( https://string-db.org/ ) with a confidence score threshold > 0.4, followed by visualization in Cytoscape (v 3.7.1) ( 24 ). 2.4 Development and verification of the risk model Prognostic genes were screened by univariate Cox regression ("survival" v 3.7-0( 25 )) and visualized ("forestplot" v 3.1.1( 26 )); after PH assumption testing (P > 0.05), an RSF model ("randomForestSRC" v 3.2.2( 27 ); ntree = 5, mtry = 6) was built to calculate risk scores and stratify patients into High risk group (HG)/Low risk group (LG), with results displayed in risk and survival status plots. Survival differences were assessed by comparing K-M curves ("survminer" v 0.4.9) with the Log-rank test (P 0.6 threshold. The above analyses were all applied in the 989 NSCLC samples in the TCGA-NSCLC dataset. Additionally, a risk model was established in the GSE37745 cohort (n = 196) by constructing a risk score for each NSCLC sample. The predictive efficiency of the model in GSE37745 dataset was validated utilizing identical methodology. 2.5 The spread of risk scores in different clinical groups In the TCGA-NSCLC dataset, the distribution of risk scores across clinical subgroups was assessed using non-parametric tests: the Wilcoxon test for binary variables (age, gender, M stage) and the Kruskal-Wallis test for multi-category variables (N stage, T stage, overall stage), with a uniform significance threshold of P < 0.05. 2.6 Construction and analysis of nomogram To assess the independent prognostic value of the risk score and clinical features in early-stage NSCLC, we conducted univariate Cox regression (P 0.05), and multivariate Cox regression (P < 0.05, HR ≠ 1) using the "survival" package (v 3.7-0). Subsequently, a nomogram was developed with the "rms" package (v 6.5-0) ( 29 ) to predict 1-, 3-, and 5-year survival probabilities. In the nomogram, Each factor corresponded to each points, and the total points was obtained by adding the points of each factor. The survival rates of NSCLC were inferred by the total points. The nomogram's predictive performance was validated by employing a calibration curve ("rms" package, v 6.5-0) and a receiver operating characteristic (ROC) curve ("survivalROC" package, v 1.0.3.1( 30 )) to assess its accuracy and discriminative power (AUC > 0.7), respectively. The above analyses were all applied in the 989 NSCLC samples in the TCGA-NSCLC dataset. 2.7 gene set enrichment analysis (GSEA) analysis This study utilized the "c2.cp.kegg.v7.5.1.symbols.gmt" gene set from MSigDB ( https://www.gsea-msigdb.org/ ) as the reference. Differential expression analysis between risk groups ("DESeq2" v 1.38.0) provided log2FC values for standard GSEA. To specifically investigate the prognostic genes, their functions were elucidated by performing GSEA on gene lists ranked by Pearson correlation with each prognostic gene ("psych" package v 2.2.9( 31 )). Subsequently, GSEA was executed across all TCGA-NSCLC samples via the "clusterProfiler" package (v 4.2.2) under thresholds of adj.P 1, and False Discovery Rate (FDR) < 0.25, and the top 5 pathways (P < 0.05) were visualized with the "enrichplot" package (v 1.18.0) ( 32 ). 2.8 Immune analysis We characterized the immune landscape by first quantifying the infiltration of 22 immune cell types in the TCGA-NSCLC cohort using CIBERSORT ( 33 ) ( samples with P > 0.05 were excluded) and visualizing the results with "ggplot2" (v 3.4.1). Differential immune cells (DICs) between HG and LG groups were identified via the Wilcoxon test (P 0.3 and P < 0.05. In parallel, we compared the expression of 41 immune checkpoints from literature ( 34 ) between the risk groups (Wilcoxon test, P < 0.05) and retrieved TIDE-related scores (TIDE, exclusion, dysfunction) and MSI from the TIDE database ( https://tide.dfci.harvard.edu/ ). To evaluate disparities in immunotherapy response potential between the risk groups, we performed two comparative analyses using the Wilcoxon test (P < 0.05). First, we compared the TIDE-related scores between HG and LG. Second, we contrasted the immunophenotype score (IPS), which was acquired from the TCIA database ( https://tcia.at/home ), across the same groups. 2.9 The analysis of somatic mutation Using the "maftools" package (v 2.18.0( 34 )), we investigated somatic mutations in the TCGA-NSCLC dataset by profiling the top 20 mutated genes and comparing tumor mutational burden (TMB) between the HG and LG groups with the Wilcoxon test (P < 0.05). 2.10 Drug sensitivity analysis Drug data from the Genomics of Drug Sensitivity in Cancer (GDSC) database ( https://www.cancerrxgene.org/ ) were analyzed by predicting IC 50 values ("pRRophetic" v 0.5( 35 )); differences between HG and LG groups in TCGA-NSCLC were tested (Wilcoxon test, P < 0.05), and the top 20 most significant drugs were displayed. 2.11 Single cell RNA sequencing (scRNA-seq) analysis The scRNA-seq data were processed utilizing the “Seurat” package (v 5.1.0) ( 36 ). Cells were filtered based on three quality metrics: nFeature_RNA (200–3500), representing the number of detected genes; nCount_RNA (200–6000), indicating the total transcript count; and percent.mt (< 20%), reflecting the mitochondrial gene proportion. In addition, the genes covering at least 3 cells were retained. Single-cell data preprocessing in Seurat (v 5.1.0) included normalization of counts per cell (scale factor 10,000, log-transformed), identification and labeling of the top 2,000 (top 10 labeled) highly variable genes, and principal component analysis (PCA)-based assessment of batch effects. After that, normalization analysis of all samples was applied utilizing the Scale Data function of the “Seurat” package (v 5.1.0). The PCA analysis of hypervariable genes was applied via the “ExPosition” package (v 2.8.23) ( 37 ). The significance of principal components was further evaluated using the JackStrawPlot function. Then, the number of principal components (PCs) corresponding to the point where the curve starts to level off was selected for cell clustering (P < 0.05, dims = 30). Cell clustering was performed using the "Seurat" package (v 5.1.0) via the FindNeighbors and FindClusters functions (resolution = 0.25), followed by dimensionality reduction with the RunUMAP function for visualization. The FindAllMarkers function of the “Seurat” package (v 5.1.0) was utilized to obtain the marker genes of cells (parameter settings min.pct = 0.1, logfc.threshold = 0.25, only.pos = TRUE). Finally, the cell annotation was executed according to the marker genes acquired in the literature ( 9 ) and CellMarker database ( http://xteam.xbio.top/CellMarker/ ). To characterize the functional profiles of different cell types, we first visualized marker gene expression. We then performed single-cell functional enrichment using the analyze_sc_clusters function ("ReactomeGSA" v 1.12.0( 38 )). Subsequently, the pathways function was used to calculate the enrichment score range for each pathway across cell types. Finally, the top 10 most variable pathways were displayed using the plot_gsva_heatmap function. Proportions of cells per sample were compared using the Wilcoxon test (P < 0.05); meanwhile, prognostic gene expression in all cells was compared between before PD-1 blockade therapy and after PD-1 blockade therapy groups using the Wilcoxon test (P < 0.05). Then, the cells with different expression of genes and different proportion between 2 groups were selected as key cells. The interactions and communication between cells were explored via the “CellChat” package (v 1.6.1) ( 39 ) (P < 0.05). To reclassify key cells into different clusters, the same methods mentioned earlier were utilized and the “DDRTree” package (v 0.1.5) ( 40 ) was used to visualize the results of clusters. Afterwards, the clusterCells function of the “Seurat” package (v 5.1.0) was utilized to divide cluster into different subgroups. Pseudotemporal analysis of key cells was conducted with the "Monocle" package (v 2.30.1) ( 41 ), followed by visualization of prognostic gene expression along the pseudotime continuum. 2.12 Gene expression analysis Differential expression of prognostic genes was evaluated in the TCGA-NSCLC cohort by Wilcoxon test (P < 0.05). 2.13 Quantitative Real-Time Polymerase Chain Reaction (qRT-PCR) Total RNA was extracted from tissues using TRIzol (ThermoFisher, USA). Isolated RNA was reverse-transcribed into cDNA using a First Strand cDNA Synthesis Kit (Thermo). SYBR green-based qRT-PCR was performed on the 7900HT Fast Real-Time PCR System (Applied Biosystems/Life Technologies, Waltham, USA). Relative mRNA expression levels were calculated using the 2^−ΔΔCt method. Primer sequences were as follows: ① PLEKHH2-F: CCCGCTGCTGAGATAGACAG ② PLEKHH2-R: AGCTGTTGCATCTTCTCTGCT ③RRM2-F: TTTAAAGGCTGCTGGAGTGAGG ④RRM2-R: GCTGCTTTAGTTTTCGGCTCC ⑤WDR76-F: TCTTGCCCCCTTTTCACTCA ⑥WDR76-R: CATCTGCCGTTCTCTTGGGT 2.14 Immunohistochemistry (IHC) Using the kit (E-IR-R217, Elabscience, China), antigen retrieval was performed on tissue sections according to the instructions. The sections were incubated overnight with RRM2(Proteintech, 11661-1-AP), WDR76(Proteintech, 25528-1-AP) or PLEKHH2(Proteintech, 14204-1-AP) Polyclonal Antibody. The following day, the sections underwent haematoxylin staining and dehydration before being mounted. The nuclear H-score was calculated as the sum of the percentages of nuclei exhibiting low, medium, and high intensity staining. 2.15 Statistical analysis Statistical analyses were carried out in R (v 4.2.2), employing the Wilcoxon test for group comparisons with a significance threshold of P < 0.05. Statistical differences between two conditions were calculated by unpaired parametric t-test. 3. Results 3.1 Identification and exploration of candidate genes in NSCLC Differential expression analysis yielded 19,106 DEGs1 (13,197 up-regulated in NSCLC; Fig. 1 a, b) from the TCGA cohort and 201 DEGs2 (99 up-regulated pre-treatment; Fig. 1 c, d) from the PD-1 blockade therapy dataset. Ultimately, 14 candidate genes were obtained in this study (Fig. 1 e). The candidate genes were significantly enriched (P < 0.05) across 80 biological processes (BPs), 4 cellular components (CCs), and 15 molecular functions (MFs) in GO analysis, including DNA replication, intercellular bridge, and clathrin heavy chain binding, respectively (Fig. 1 f, Additional file 2 ). Our analysis identified significant enrichment of the candidate genes in 7 KEGG pathways, such as DNA replication and pyruvate metabolism (Fig. 1 g). Furthermore, in the PPI network, only 4 candidate genes had interactions with each other (Fig. 1 h). Comprehensively, RRM2 had an interactive relationship with WDR76, CDC7 and RNASEH2A. 3.2 Prediction accuracy of risk models A total of 3 prognostic genes including RRM2, WDR76 and PLEKHH2 were obtained (Fig. 2 a-d). Based on an established risk model, TCGA-NSCLC patients were classified into high-risk (HG, n = 555) and low-risk (LG, n = 434) groups using the optimal risk score cutoff of 49.27128. The results of risk graph indicated that the number of death cases increased with the increase of risk score, and the number of death for patients in LG was notably lower than that of patients in HG in TCGA-NSCLC dataset (Fig. 2 e and f ). Furthermore, the HG group was associated with significantly poorer survival (P 0.70, Fig. 2 h). Meanwhile, NSCLC patients in GSE37745 dataset were also classified into HG (126 patients) and LG (70 patients) with the best truncation value of risk score (46.78981). The risk stratification pattern observed in the GSE37745 dataset was consistent with that in the TCGA-NSCLC dataset (Fig. 2 i and j ). The NSCLC patients in HG had lower survival rates (Fig. 2 k) (P < 0.0027).The model achieved AUCs of 0.60 (1-year), 0.61 (3-year), and 0.61 (5-year) in time-dependent ROC analysis (Fig. 2 l). All these results suggested that the risk model exhibited substantial predictive accuracy for the survival rate of NSCLC. 3.3 Construction of nomogram with independent prognostic factors The risk score distribution differed significantly across T stages (T1-T4), N stages (N0-N2), and overall stages (I-IV) (P < 0.0001; Fig. 3 a). Consistent with these associations, multivariate Cox regression confirmed that both the risk score and the N stage served as independent prognostic factors (Fig. 3 b-f). In the nomogram, the total points were inversely correlated with NSCLC survival probability. The risk score held superior predictive weight over the N stage (Fig. 3 g). The favorable slope of the calibration curves (close to 1) confirmed the model's reliable predictive accuracy (Fig. 3 h). The all AUC were exceeding 0.7, indicating the nomogram had well predictive efficiency for NSCLC (Fig. 3 i). In summary, the constructed nomogram was validated as an accurate predictor of survival probability in NSCLC patients. 3.4 Enrichment pathways linked to prognostic genes and differences enrichment pathways in the HG and LG In this study, 14 prominent pathways such as cell cycle and DNA replication were found in the HG and LG (Fig. 4 a, Additional file 3 ). In addition, RRM2, WDR76 and PLEKHH2 were all enriched in the cell cycle, and WDR76 and PLEKHH2 were all enriched in the notably pathways such as DNA Replication (Fig. 4 b-d, Additional file 4). These results indicated that these notably enriched pathways might play an important role in NSCLC progression. 3.5 The DICs linked to prognostic genes and differences in immune therapy responsein the HG and LG Comparative analysis identified a markedly higher infiltration of resting CD4 + memory T cells in the LG than in the HG (Fig. 5 a). In addition, a total of 11 DICs such as activated memory CD4 T cells were obtained (P 0.30, P < 0.05) were notable positive linked to activated CD4 memory T cells (Fig. 5 c-e). These results indicated these DICs, especially activated memory CD4 T cells might have a notable impact with biomarkers on NSCLC. In addition, the expressions of the immune checkpoints such as CD276 and NRP1 in the HG were higher than that in the LG (Fig. 5 f). The TIDE score in the HG was higher than that in the LG (P < 0.05), indicating a lower success rate of immunotherapy in the HG (Fig. 5 g). A significantly higher exclusion prediction score was observed in the HG versus the LG (P < 0.05), implying a heightened state of T cell rejection in the high-risk group (Fig. 5 h). The LG exhibited a significantly lower T cell dysfunction score than the HG (P < 0.05, Fig. 5 i), indicative of better-preserved T cell function; however, the IPS was comparable between the groups (Fig. 5 j). In addition, the results showed that the score of IPS-CTLA4 (-)/PD-1(-) and IPS-CTLA4(+)/PD-1(-) had notable differences between the 2 groups, indicating that the expression of CTLA-4 might be a key factor driving differences in immune status in the tumor microenvironment lacking activation of the PD-1 pathway (Fig. 5 k-n). 3.6 The differences of somatic mutation The waterfall plot demonstrated that the mutation type in the 2 groups was missense mutation and multiple mutations (Fig. 6 a and b ).Mutation profiling identified TP53 as the most frequently mutated gene, with rates of 72.00% in the HG and 61.00% in the LG, highlighting the importance of gene mutations in NSCLC. Moreover, elevated TMB scores in the HG compared to the LG indicated potentially enhanced immunogenicity in high-risk patients (Fig. 6 c). 3.7 The differences of IC 50 with drugs in the HG and LG A total of 109 drugs with notable differences in IC 50 values between HG and LG (P < 0.05) ( Additional file 6) . For example, the IC 50 of drugs such as AZD6738, AZD7762 in the HG was notably lower than that in the HG (Fig. 7 ). In conclusion, the above drugs with different sensitivities in the HG and LG, indicating that patients in different groups need to choose different drugs. 3.8 Identification of epithelials After undergoing quality control, 64,245 cells and 24,292 genes were obtained ( Additional file 7a and b ). Then, the top 2,000 hypervariable genes such as IGKC were selected for PCA (Fig. 8 a). In the absence of detectable batch effects, cells were clustered into 20 distinct populations using the top 30 principal components (Fig. 8 b-e). Then, a total of 8 cells types (B cells, T cells, epithials, stromal cells, mast cells, myeloids, neutrophils and plasma cells) were obtained (Fig. 8 f). Cell type-specific marker genes, including MS4A1 for B cells, were predominantly expressed in their respective cell populations (Fig. 8 g). Enrichment analysis identified significant pathways, including COX reactions, in the eight cell clusters (Fig. 8 h), while cellular composition analysis showed that proportion of all cells differed between the two groups of samples (Fig. 8 i). Among the eight cell clusters, significant differences in PLEKHH2, RRM2, and WDR76 expression between Samples before and after treatment were exclusively observed in epithelials (Fig. 8 j). So epithelials were selected as key cell in this study. 3.9 The ommunication networks among cells and pseudo temporal analysis of epithelials Among the communication networks, the number of communication interactions between myeloids, epithelials and neutrophil cells were large relatively, and the communication strength between myeloids and neutrophils were relatively high in the 2 groups In addition, in the treated samples, the number of communication between endothelials and myeloid cells increased (Fig. 9 a-d). In addition, the communication strength of ANXA1-FPR1 between epithelials and neutrophil cells in the 2 groups was the highest (Fig. 9 e and f ). Then, epithelials were reclustered into 10 cell clusters (Fig. 9 g). The differentiation of epithelials was classified into 3 stages, and 5 clusters, 8 clusters, 7 clusters, and 9 clusters were the main clusters in the early stage (Fig. 9 h-k). RRM2 and WDR76 were mainly expressed in the later stages of differentiation, while PLEKHH2 were mainly expressed in the middle stage of the temporal sequence (Fig. 9 l).The results indicated that the development of key cells may also be associated with the expression of prognostic genes. 3.10 The expression of 3 prognostic genes The expression of 3 prognostic genes all had differences between NSCLC and control groups (P < 0.05) ( Fig. 10 ) . The expressions of RRM2 and WDR76 were all higher in the NSCLC group while the expression of PLEKHH2 were opposite. To further verify, normal lung tissues and corresponding lung cancer tissues were collected from six clinical patients, including three stage I patients and three stage III patients. The expression levels of RRM2, WDR76, and PLEKHH2 were validated using PCR(Fig. 11 a) and IHC(Fig. 11 b). Consistent with previous analyses, the results showed that RRM2 and WDR76 expression was significantly higher in cancer tissues compared to normal tissues, and higher in stage III patients than in stage I patients. In contrast, PLEKHH2 expression was significantly lower in cancer tissues than in normal tissues and lower in stage III patients than in stage I patients. These experimental findings further suggest that RRM2 and WDR76 may promote tumor malignant progression, whereas PLEKHH2 likely plays an inhibitory role. 4. Discussion Immunotherapy resistance in NSCLC is a major challenge currently faced in clinical practice ( 41 ). In recent years, NED has been recognized as playing a critical role in tumor progression and therapy resistance ( 41 ); however, its specific mechanism in mediating immune resistance in NSCLC remains unclear. In this study, we identified three prognostic genes, namely RRM2, WDR76, and PLEKHH2, which are closely associated with neuroendocrine differentiation and immunotherapy resistance in NSCLC. The risk model, constructed from these genes, was validated in both the TCGA and independent GEO cohorts, demonstrating consistent and robust predictive accuracy for patient survival. Furthermore, we elucidated the biological pathways, immune microenvironment characteristics, and potential therapeutic compounds associated with high- and low-risk groups. RRM2 has been implicated in oxidative stress response and chemoresistance. Under iron-rich conditions, RRM2 is prone to depolymerization and degradation, making it a potential therapeutic target ( 41 ). Previous studies have reported RRM2 overexpression in chemotherapy-resistant cancers, and its inhibition may enhance chemosensitivity ( 42 ). In NSCLC, elevated RRM2 expression correlates with distant metastasis and poorer survival, underscoring its value as both a diagnostic and prognostic biomarker ( 43 ). This study further reveals that RRM2 is significantly upregulated in high-risk NSCLC populations and is closely associated with cell cycle pathways. Given its central role in DNA synthesis, our findings, showing elevated RRM2 in high-risk patients and its enrichment in cell cycle pathways, lead us to hypothesize that RRM2-driven tumor proliferation may contribute to an immunosuppressive microenvironment, potentially through mechanisms such as T cell exhaustion or upregulation of immune checkpoints, which in turn could underlie immunotherapy resistance (e.g., G1/S transition) ( 44 ). Highly proliferative tumors often exhibit reduced T-cell infiltration or functional exhaustion and may upregulate immune checkpoint molecules, thereby weakening the response to immunotherapy ( 45 ). Additionally, RRM2-mediated genomic instability may increase tumor mutational burden (TMB) but could also promote immune editing and heterogeneous clonal evolution, ultimately leading to immune escape ( 46 ). Therefore, RRM2 is not only a biomarker for poor prognosis in NSCLC but also a potential therapeutic target. Inhibitors targeting RRM2 (such as 3-AP, COH29, etc.) have shown synergistic effects with chemotherapy and immunotherapy in preclinical studies. Based on the findings of this study, future research could explore combination therapy strategies of RRM2 inhibitors with PD-1/PD-L1 inhibitors in high-risk NSCLC populations, with the aim of reversing immune resistance and improving patient outcomes. WDR76, a WD40 repeat-containing protein, participates in DNA damage repair, ubiquitin-mediated degradation, and epigenetic regulation. Although its role in NSCLC is less studied, WDR76 has been shown to destabilize RAS proteins in liver cancer and sensitize colorectal cancer cells to 5-fluorouracil ( 47 , 48 ). Our study is the first to systematically elucidate the clinical significance and immunoregulatory function of WDR76 in NSCLC. As an adaptor of the CRL4 (Cullin4-RING E3 ubiquitin ligase) complex, WDR76 mediates the ubiquitination and degradation of various tumor suppressors (such as p53, p27, etc.), thereby promoting tumor cell proliferation and survival ( 47 , 49 ). Additionally, WDR76 is involved in maintaining genomic stability by regulating key proteins in the DNA damage response, influencing tumor cell sensitivity to chemotherapy and radiotherapy ( 50 ). Notably, our bioinformatic analysis identified a significant positive association between WDR76 expression and the infiltration level of activated memory CD4 + T cells, suggesting its potential role in influencing NSCLC progression via modulation of the immune microenvironment. In summary, our data suggest that WDR76, functioning as a molecular scaffold, may contribute to NSCLC progression by regulating both intrinsic tumor cell proliferation and the extrinsic immune microenvironment, as evidenced by its correlation with activated memory CD4 + T cells. Targeting WDR76 and its related pathways may provide novel strategies for enhancing the efficacy of immunotherapy in NSCLC. PLEKHH2 contains multiple functional domains, including PH, MyTH, and FERM domains, and is involved in cytoskeletal organization and signal transduction. Although initially studied in psychiatric disorders, recent evidence indicates that PLEKHH2 promotes malignant phenotypes in NSCLC by activating the FAK/PI3K/AKT pathway through interaction with β-arrestin1 ( 51 ). Our results suggest that PLEKHH2 may influence both neuroendocrine differentiation and T-cell-mediated immunity, thus establishing it as a key molecular node linking intrinsic tumor characteristics with immune microenvironment remodeling. Therefore, PLEKHH2 may act as a key molecular node that links the intrinsic characteristics of tumor cells with the remodeling of the immune microenvironment by regulating the signal transduction and cell adhesion of tumor cells. On the one hand, it induces neuroendocrine phenotypic transformation (associated with treatment resistance), and on the other hand, it affects the recruitment and activation of T cells in the tumor microenvironment. Pathway enrichment analysis revealed that cell cycle and DNA replication pathways are significantly activated in high-risk NSCLC patients. The three key prognostic genes, RRM2, WDR76, and PLEKHH2, are all closely associated with these pathways, suggesting their potential roles as core molecules synergistically driving dysregulated cell cycle and genomic instability in NSCLC. As a critical subunit of ribonucleotide reductase, RRM2 provides sufficient dNTP precursors for DNA synthesis, which is essential for proper S-phase progression ( 52 ). WDR76, functioning as an adaptor for the E3 ubiquitin ligase complex, may influence G1/S phase transition by regulating the stability of cyclins or cyclin-dependent kinase inhibitors such as p21 and p27 ( 47 ). Meanwhile, PLEKHH2 may indirectly affect cell cycle progression by integrating extracellular signals with intracellular cytoskeletal reorganization. The aberrant overexpression of these three genes collectively forms a molecular network supporting rapid tumor cell proliferation. Furthermore, aberrant activation of the cell cycle not only directly promotes tumor proliferation but also shapes the immunosuppressive microenvironment through multiple mechanisms. Rapidly proliferating tumor cells may upregulate immune checkpoint molecules like PD-L1 ( 53 ), induce T-cell exhaustion, and secrete immunosuppressive cytokines such as TGF-β and IL-10, which suppress effector immune cell function and recruit regulatory T cells (Tregs) ( 54 ). Increased genomic instability may enhance tumor antigen heterogeneity, enabling some tumor cell clones to evade immune recognition. These mechanisms collectively contribute to immunotherapy resistance. Therefore, our study not only elucidates the roles of RRM2, WDR76, and PLEKHH2 as key nodes in the cell cycle regulatory network during NSCLC progression but, more importantly, provides precise therapeutic directions for high-risk NSCLC patients. The combination of inhibitors targeting these pathways with immune checkpoint inhibitors may yield synergistic antitumor effects by simultaneously targeting intrinsic tumor proliferation mechanisms and extrinsic immune microenvironment regulation. This molecular stratification-based combination therapy strategy holds promise for overcoming immunotherapy resistance in NSCLC and improving patient outcomes. Our immune infiltration analysis revealed an intriguing phenomenon: activated memory CD4 + T cells showed a positive correlation with the expression of high-risk genes RRM2 and WDR76. Given the high heterogeneity of CD4 + T cell populations (including immunosuppressive regulatory T cells [Tregs] and functionally exhausted helper T cells) ( 55 ), this finding may suggest that these infiltrating CD4 + T cells are not exerting anti-tumor effects but rather constitute part of the immunosuppressive microenvironment. scRNA-seq analysis further established the central role of T cells in immunotherapy response. We observed that patients with a favorable treatment response (MPR) exhibited significantly weakened communication between T cells and stromal cells. This may result from effective immunotherapy disrupting the tumor-stromal barrier, enabling T cells to more efficiently target and kill tumor cells instead of being suppressed within the stromal network. Such alterations in cellular interaction patterns could potentially serve as a novel predictive biomarker for treatment efficacy. From a therapeutic perspective, drugs such as AZD6738 (ATR inhibitor) and AZD7762 (CHK1 inhibitor) showed higher sensitivity in the high-risk group. These compounds target DNA damage response pathways and may be particularly effective in tumors with cell cycle dysregulation ( 55 , 56 ). Our drug sensitivity analysis provides a rationale for personalized treatment strategies based on risk stratification. In conclusion, we successfully constructed and validated a risk model comprising three novel prognostic genes (RRM2, WDR76, and PLEKHH2) through integrated multi-omics data. The established model demonstrated good predictive capability for patient outcomes. Multivariate Cox regression analysis confirmed that both the risk score and N stage were independent prognostic factors. Our analysis identified significant enrichment of the prognostic genes in pathways governing cell cycle progression and DNA replication. Immune infiltration analysis identified a significant correlation between activated memory CD4 + T cells and the prognostic genes. Drug sensitivity analysis indicated differential responses to medications such as AZD6738 and AZD7762 between high- and low-risk groups. In conclusion, utilizing a series of bioinformatic approaches, we identified prognostic genes associated with NSCLC outcomes and further investigated their mechanistic roles in the pathological processes of NSCLC. Despite these insights, our study has several limitations. The expression and functional roles of RRM2, WDR76, and PLEKHH2 at the protein level were not validated in clinical samples. Moreover, the mechanisms by which these genes modulate immune cell function and therapy resistance require further experimental investigation. Future studies using in vitro and in vivo models are needed to confirm these findings and explore translational applications. Abbreviations Non-small cell lung cancer NSCLC lung adenocarcinoma LUAD lung squamous cell carcinoma LUSC Immune checkpoint inhibitors ICIs programmed death-1 PD-1 Neuroendocrine differentiation NED Single-cell RNA sequencing scRNA-seq Gene Expression Omnibus GEO major pathologic response MPR non-MPR NMPR differentially expressed genes DEGs Kyoto Encyclopedia of Genes and Genomes enrichment KEGG Gene Ontology GO protein-protein interaction PPI Search Tool for the Retrieval of Interacting Genes STRING proportional hazards PH receiver operating characteristic ROC area under the curve AUC gene set enrichment analysis GSEA fold change FC normalized enrichment score NES False Discovery Rate FDR Differential immune cells DICs High risk group HG Low risk group LG immunophenotype score IPS tumor mutational burden TMB Genomics of Drug Sensitivity in Cancer GDSC Single cell RNA sequencing scRNA-seq principal component analysis PCA principal components PCs Declarations Ethics approval and consent to participate All the samples were collected with informed consent from the patients and ethics committee of the Tianjin Medical University General Hospita（IRB2024-YX-045-01）. Consent for publication Not adapted. Availability of data and materials The dataset(TCGA-NSCLC) supporting the conclusions of this article is available in the [TCGA] repository, [https://xena.ucsc.edu/]. The datasets (GSE248249 and GSE37745) supporting the conclusions of this article is available in the [GEO] repository, [https://www.ncbi.nlm.nih.gov/geo/]. Competing Interest The authors declare that they have no competing interests. Funding This work was supported by Tianjin Municipal Health Commission Scientific Research Project on Integrative Traditional Chinese and Western Medicine(2023150)， National Natural Science Foundation of China(82303793) and Tianjin Science and Technology Plan Project (24ZXGZSY00010). Author Contributions LT, LMD and PSB contributed equally to this work. They were responsible for conceived the study, designed and performed the majority of the experiments, conducted data analysis. LT, LMD, PSB, RF, GJL, XHS, QJY, LCC contributed to data acquisition, data analysis and assays performance. LJand XS revised the manuscript and finally approved the version to be published. All authors read and approved the final manuscript. Acknowledgments The authors sincerely thank the technical support of the Department of Lung Cancer Surgery, Tianjin Medical University General Hospital and Nankai University. We are also grateful to Dr. Xu and Dr. Li. We also extend our appreciation to all participants who volunteered for this study. Contributor Information Song Xu, E-mail: [email protected] Jia Li, E-mail: [email protected] Tong Li, E-mail: [email protected] References Wéber A, Morgan E, Vignat J, Laversanne M, Pizzato M, Rumgay H, et al. Lung cancer mortality in the wake of the changing smoking epidemic: a descriptive study of the global burden in 2020 and 2040. BMJ Open. 2023;13(5):e065303. Zhou X, You L, Xin Z, Su H, Zhou J, Ma Y. Leveraging circulating microbiome signatures to predict tumor immune microenvironment and prognosis of patients with non-small cell lung cancer. J Transl Med. 2023;21(1):800. Zhang R, Zhu G, Li Z, Meng Z, Huang H, Ding C, et al. ITGAL expression in non-small-cell lung cancer tissue and its association with immune infiltrates. Front Immunol. 2024;15:1382231. Xiang Y, Liu X, Wang Y, Zheng D, Meng Q, Jiang L, et al. Mechanisms of resistance to targeted therapy and immunotherapy in non-small cell lung cancer: promising strategies to overcoming challenges. Front Immunol. 2024;15:1366260. Chen P, Liu Y, Wen Y, Zhou C. Non-small cell lung cancer in China. Cancer Commun (Lond). 2022;42(10):937–70. Wang Y, Wang Y, Wang Y, Zhang Y. Identification of prognostic signature of non-small cell lung cancer based on TCGA methylation data. Sci Rep. 2020;10(1):8575. Li Z, Zhou B, Zhu X, Yang F, Jin K, Dai J, et al. Differentiation-related genes in tumor-associated macrophages as potential prognostic biomarkers in non-small cell lung cancer. Front Immunol. 2023;14:1123840. Kang J, Zhang C, Zhong WZ. Neoadjuvant immunotherapy for non-small cell lung cancer: State of the art. Cancer Commun (Lond). 2021;41(4):287–302. Hu J, Zhang L, Xia H, Yan Y, Zhu X, Sun F, et al. Tumor microenvironment remodeling after neoadjuvant immunotherapy in non-small cell lung cancer revealed by single-cell RNA sequencing. Genome Med. 2023;15(1):14. Shi L, Lu J, Zhong D, Song M, Liu J, You W, et al. Clinicopathological and predictive value of MAIT cells in non-small cell lung cancer for immunotherapy. J Immunother Cancer. 2023;11(1). Ozaki Y, Miura S, Oki R, Morikawa T, Uchino K. Neuroendocrine Neoplasms of the Breast: The Latest WHO Classification and Review of the Literature. Cancers (Basel). 2021;14(1). Liu H, Han Y, Liu Z, Gao L, Yi T, Yu Y, et al. Depiction of neuroendocrine features associated with immunotherapy response using a novel one-class predictor in lung adenocarcinoma. Discov Oncol. 2023;14(1):71. Lin S, Dai Y, Han C, Han T, Zhao L, Wu R, et al. Single-cell transcriptomics reveal distinct immune-infiltrating phenotypes and macrophage-tumor interaction axes among different lineages of pituitary neuroendocrine tumors. Genome Med. 2024;16(1):60. Wang Z, Ding H, Zou Q. Identifying cell types to interpret scRNA-seq data: how, why and more possibilities. Brief Funct Genomics. 2020;19(4):286–91. Cheng C, Chen W, Jin H, Chen X. A Review of Single-Cell RNA-Seq Annotation, Integration, and Cell-Cell Communication. Cells. 2023;12(15). Li C, Song W, Zhang J, Luo Y. Single-cell transcriptomics reveals heterogeneity in esophageal squamous epithelial cells and constructs models for predicting patient prognosis and immunotherapy. Front Immunol. 2023;14:1322147. Zhang T, Zhao F, Lin Y, Liu M, Zhou H, Cui F, et al. Integrated analysis of single-cell and bulk transcriptomics develops a robust neuroendocrine cell-intrinsic signature to predict prostate cancer progression. Theranostics. 2024;14(3):1065–80. Love MI, Huber W, Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 2014;15(12):550. Gustavsson EK, Zhang D, Reynolds RH, Garcia-Ruiz S, Ryten M. ggtranscript: an R package for the visualization and interpretation of transcript isoforms using ggplot2. Bioinformatics. 2022;38(15):3844–6. Gu Z, Eils R, Schlesner M. Complex heatmaps reveal patterns and correlations in multidimensional genomic data. Bioinformatics. 2016;32(18):2847–9. Scalori A, Apale P, Panizzuti F, Mascoli N, Pioltelli P, Pozzi M, et al. Depression during interferon therapy for chronic viral hepatitis: early identification of patients at risk by means of a computerized test. Eur J Gastroenterol Hepatol. 2000;12(5):505–9. Mao W, Ding J, Li Y, Huang R, Wang B. Inhibition of cell survival and invasion by Tanshinone IIA via FTH1: A key therapeutic target and biomarker in head and neck squamous cell carcinoma. Exp Ther Med. 2022;24(2):521. Yu G, Wang LG, Han Y, He QY. clusterProfiler: an R package for comparing biological themes among gene clusters. Omics. 2012;16(5):284–7. Shannon P, Markiel A, Ozier O, Baliga NS, Wang JT, Ramage D, et al. Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome Res. 2003;13(11):2498–504. Shen L, Mo J, Yang C, Jiang Y, Ke L, Hou D, et al. SurvivalPath:A R package for conducting personalized survival path mapping based on time-series survival data. PLoS Comput Biol. 2023;19(1):e1010830. Jiang H, Li Y, Wang T. Comparison of Billroth I, Billroth II, and Roux-en-Y reconstructions following distal gastrectomy: A systematic review and network meta-analysis. Cir Esp (Engl Ed). 2021;99(6):412–20. Zhao P, Zhen H, Zhao H, Huang Y, Cao B. Identification of hub genes and potential molecular mechanisms related to radiotherapy sensitivity in rectal cancer based on multiple datasets. J Transl Med. 2023;21(1):176. Heagerty PJ, Lumley T, Pepe MS. Time-dependent ROC curves for censored survival data and a diagnostic marker. Biometrics. 2000;56(2):337–44. Li M, Wei X, Zhang SS, Li S, Chen SH, Shi SJ, et al. Recognition of refractory Mycoplasma pneumoniae pneumonia among Myocoplasma pneumoniae pneumonia in hospitalized children: development and validation of a predictive nomogram model. BMC Pulm Med. 2023;23(1):383. Zheng Y, Wen Y, Cao H, Gu Y, Yan L, Wang Y, et al. Global Characterization of Immune Infiltration in Clear Cell Renal Cell Carcinoma. Onco Targets Ther. 2021;14:2085–100. Robles-Jimenez LE, Aranda-Aguirre E, Castelan-Ortega OA, Shettino-Bermudez BS, Ortiz-Salinas R, Miranda M, et al. Worldwide Traceability of Antibiotic Residues from Livestock in Wastewater and Soil: A Systematic Review. Animals (Basel). 2021;12(1). Wang L, Wang D, Yang L, Zeng X, Zhang Q, Liu G, et al. Cuproptosis related genes associated with Jab1 shapes tumor microenvironment and pharmacological profile in nasopharyngeal carcinoma. Front Immunol. 2022;13:989286. Chen B, Khodadoust MS, Liu CL, Newman AM, Alizadeh AA. Profiling Tumor Infiltrating Immune Cells with CIBERSORT. Methods Mol Biol. 2018;1711:243–59. Jiang F, Lu DF, Zhan Z, Yuan GQ, Liu GJ, Gu JY, et al. SARS-CoV-2 Pattern Provides a New Scoring System and Predicts the Prognosis and Immune Therapeutic Response in Glioma. Cells. 2022;11(24). Wang W, Lu Z, Wang M, Liu Z, Wu B, Yang C, et al. The cuproptosis-related signature associated with the tumor environment and prognosis of patients with glioma. Front Immunol. 2022;13:998236. Hao Y, Hao S, Andersen-Nissen E, Mauck WM, 3rd, Zheng S, Butler A, et al. Integrated analysis of multimodal single-cell data. Cell. 2021;184(13):3573–87.e29. Zhang L, Wang X, Wang W, Ning E, Chen L, Li Z, et al. Metabolomic analysis reveals the changing trend and differential markers of volatile and nonvolatile components of Artemisiae argyi with different aging years. Phytochem Anal. 2024;35(6):1286–93. Grentner A, Ragueneau E, Gong C, Prinz A, Gansberger S, Oyarzun I, et al. ReactomeGSA: new features to simplify public data reuse. Bioinformatics. 2024;40(6). Jin S, Plikus MV, Nie Q. CellChat for systematic analysis of cell-cell communication from single-cell transcriptomics. Nat Protoc. 2025;20(1):180–219. Ren C, He D, Wang Q. Identification and Experimental Validation of Tumor Antigens and Hypoxia Subtypes of Osteosarcoma for Potential mRNA Vaccine Development. Curr Med Chem. 2025. Cao J, Spielmann M, Qiu X, Huang X, Ibrahim DM, Hill AJ, et al. The single-cell transcriptional landscape of mammalian organogenesis. Nature. 2019;566(7745):496–502. Zhan Y, Jiang L, Jin X, Ying S, Wu Z, Wang L, et al. Inhibiting RRM2 to enhance the anticancer activity of chemotherapy. Biomed Pharmacother. 2021;133:110996. Zhou D, Zhai X, Zhang R. Ribonucleotide reductase regulatory subunit M2 (RRM2) as a potential sero-diagnostic biomarker in non-small cell lung cancer. PLoS One. 2023;18(9):e0291461. Li J, Pang J, Liu Y, Zhang J, Zhang C, Shen G, et al. Suppression of RRM2 inhibits cell proliferation, causes cell cycle arrest and promotes the apoptosis of human neuroblastoma cells and in human neuroblastoma RRM2 is suppressed following chemotherapy. Oncol Rep. 2018;40(1):355–60. Huang L, Zhang C, Jiang A, Lin A, Zhu L, Mou W, et al. T-cell Senescence in the Tumor Microenvironment. Cancer Immunol Res. 2025;13(5):618–32. Wang Y, Chen R, Zhang J, Zeng P. A comprehensive analysis of ribonucleotide reductase subunit M2 for carcinogenesis in pan-cancer. PLoS One. 2024;19(4):e0299949. Jeong WJ, Park JC, Kim WS, Ro EJ, Jeon SH, Lee SK, et al. WDR76 is a RAS binding protein that functions as a tumor suppressor via RAS degradation. Nat Commun. 2019;10(1):295. Hu Y, Tan X, Zhang L, Zhu X, Wang X. WDR76 regulates 5-fluorouracil sensitivity in colon cancer via HRAS. Discov Oncol. 2023;14(1):45. Huang D, Li Q, Sun X, Sun X, Tang Y, Qu Y, et al. CRL4(DCAF8) dependent opposing stability control over the chromatin remodeler LSH orchestrates epigenetic dynamics in ferroptosis. Cell Death Differ. 2021;28(5):1593–609. Gilmore JM, Sardiu ME, Groppe BD, Thornton JL, Liu X, Dayebgadoh G, et al. WDR76 Co-Localizes with Heterochromatin Related Proteins and Rapidly Responds to DNA Damage. PLoS One. 2016;11(6):e0155492. Wang R, Wang S, Li Z, Luo Y, Zhao Y, Han Q, et al. PLEKHH2 binds β-arrestin1 through its FERM domain, activates FAK/PI3K/AKT phosphorylation, and promotes the malignant phenotype of non-small cell lung cancer. Cell Death Dis. 2022;13(10):858. Taricani L, Shanahan F, Malinao MC, Beaumont M, Parry D. A functional approach reveals a genetic and physical interaction between ribonucleotide reductase and CHK1 in mammalian cells. PLoS One. 2014;9(11):e111714. Yu J, Ling S, Hong J, Zhang L, Zhou W, Yin L, et al. TP53/mTORC1-mediated bidirectional regulation of PD-L1 modulates immune evasion in hepatocellular carcinoma. J Immunother Cancer. 2023;11(11). Das L, Levine AD. TGF-beta inhibits IL-2 production and promotes cell cycle arrest in TCR-activated effector/memory T cells in the presence of sustained TCR signal transduction. J Immunol. 2008;180(3):1490–8. Hu Y, Zhao Q, Qin Y, Mei S, Wang B, Zhou H, et al. CARD11 signaling regulates CD8(+) T cell tumoricidal function. Nat Immunol. 2025;26(7):1113–26. Chen M, Zhang S, Wang F, He J, Jiang W, Zhang L. DLGAP5 promotes lung adenocarcinoma growth via upregulating PLK1 and serves as a therapeutic target. J Transl Med. 2024;22(1):209. Additional Declarations No competing interests reported. Supplementary Files Additionalfile1.xlsx Additional file 1 NEDRGs information from the literature Additionalfile2.xlsx Additional file 2 Results of GO enrichment analysis Additionalfile3.xlsx Additional file 3 Results of KEGG enrichment analysis Additionalfile4.xlsx Additional file 4 Results of GSEA enrichment analysis of RRM2, WDR76 and PLEKHH2 Additionalfile5.xlsx Additional file 5 Infiltration levels of immune cells in DICs between HG group and LG group Additionalfile6.xlsx Additional file 6 Different drug IC50 values in HG group and LG group Additionalfile7.jpg Additional file 7 Single - cell data pre - processing. Before (a) and after (b) single - cell data quality control. Cite Share Download PDF Status: Under Review Version 1 posted Editorial decision: Revision requested 02 Mar, 2026 Reviewers agreed at journal 14 Feb, 2026 Reviews received at journal 12 Feb, 2026 Reviewers agreed at journal 10 Feb, 2026 Reviewers invited by journal 10 Feb, 2026 Editor assigned by journal 09 Feb, 2026 Submission checks completed at journal 09 Feb, 2026 First submitted to journal 07 Feb, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8816614","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":591333867,"identity":"fa18ccb5-9d30-49ae-8376-ce59ff85e11d","order_by":0,"name":"Tong Li","email":"","orcid":"","institution":"Department of Lung Cancer Surgery, Tianjin Medical University General Hospital","correspondingAuthor":false,"prefix":"","firstName":"Tong","middleName":"","lastName":"Li","suffix":""},{"id":591333869,"identity":"926f03f6-bf3a-473f-b9f0-993c5bf8b649","order_by":1,"name":"Meidan Luo","email":"","orcid":"","institution":"Department of Lung Cancer Surgery, Tianjin Medical University General Hospital","correspondingAuthor":false,"prefix":"","firstName":"Meidan","middleName":"","lastName":"Luo","suffix":""},{"id":591333873,"identity":"fe89d4af-1e61-4a1b-aa37-47dd5dc20a2b","order_by":2,"name":"Sibo Peng","email":"","orcid":"","institution":"Department of Lung Cancer Surgery, Tianjin Medical University General Hospital","correspondingAuthor":false,"prefix":"","firstName":"Sibo","middleName":"","lastName":"Peng","suffix":""},{"id":591333876,"identity":"ddadf1ad-bc79-410e-a6cd-2cdb7c00f6ef","order_by":3,"name":"Fan Ren","email":"","orcid":"","institution":"Department of Lung Cancer Surgery, Tianjin Medical University General Hospital","correspondingAuthor":false,"prefix":"","firstName":"Fan","middleName":"","lastName":"Ren","suffix":""},{"id":591333878,"identity":"07ab3f13-8410-418e-82c6-7705760cd820","order_by":4,"name":"Jingliang Gan","email":"","orcid":"","institution":"Department of Lung Cancer Surgery, Tianjin Medical University General Hospital","correspondingAuthor":false,"prefix":"","firstName":"Jingliang","middleName":"","lastName":"Gan","suffix":""},{"id":591333879,"identity":"069ba100-a27d-4234-9ebb-9f84b235f240","order_by":5,"name":"Huangsheng Xie","email":"","orcid":"","institution":"Department of Lung Cancer Surgery, Tianjin Medical University General Hospital","correspondingAuthor":false,"prefix":"","firstName":"Huangsheng","middleName":"","lastName":"Xie","suffix":""},{"id":591333880,"identity":"df9378b7-e131-4550-9898-f67dbe9084cc","order_by":6,"name":"Jiayue Qiao","email":"","orcid":"","institution":"Department of Lung Cancer Surgery, Tianjin Medical University General Hospital","correspondingAuthor":false,"prefix":"","firstName":"Jiayue","middleName":"","lastName":"Qiao","suffix":""},{"id":591333881,"identity":"222f3b3f-1183-4850-92ea-6efc052cd7d4","order_by":7,"name":"Chicheng Liu","email":"","orcid":"","institution":"Department of Lung Cancer Surgery, Tianjin Medical University General Hospital","correspondingAuthor":false,"prefix":"","firstName":"Chicheng","middleName":"","lastName":"Liu","suffix":""},{"id":591333883,"identity":"9a876c98-3778-4e28-8761-1c769439f11d","order_by":8,"name":"Jia Li","email":"","orcid":"","institution":"Department of Lung Cancer Surgery, Tianjin Medical University General Hospital","correspondingAuthor":false,"prefix":"","firstName":"Jia","middleName":"","lastName":"Li","suffix":""},{"id":591333886,"identity":"8a3ef7fd-916f-4cd8-83c6-a009d971b253","order_by":9,"name":"Song Xu","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA6ElEQVRIiWNgGAWjYPACCTl+9h4wi4ePSC0WxpI9ZxgYDgC1sBGppSJxw40csBYGgloMbqQ/fFzwSyJx5sy3Bx9/zLGTYWNgfvjoBh4tkjNyjI1n9kkY90vnJRsc3JYMdBibsXEOHi38Ejls0rw9ErIzZ+eYSRzcxgzUwsMmjU8Lm0T6M5AWxg03z4C01BPWwi+RYCbN80NCccMNHpCWw4S1SPa8MTbmbZAABnKOscHZbcd52JgJ+MXgODDEeP7UAaPyjOGDym3V9vzszQ8f49MCBoxtyDxmQsrB4A9RqkbBKBgFo2CkAgCLG0Qkj0e3hAAAAABJRU5ErkJggg==","orcid":"","institution":"Department of Lung Cancer Surgery, Tianjin Medical University General Hospital","correspondingAuthor":true,"prefix":"","firstName":"Song","middleName":"","lastName":"Xu","suffix":""}],"badges":[],"createdAt":"2026-02-07 15:38:58","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8816614/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8816614/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":102791235,"identity":"97b86026-b462-42a9-ad3d-6592b7824081","added_by":"auto","created_at":"2026-02-16 17:18:04","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":3587495,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eScreening and enrichment analysis of differentially expressed genes. \u003c/strong\u003e(a) Volcano plot of DEGs for differential gene screening in patient and normal control samples. Yellow represents up - regulation, and blue represents down - regulation. (b) Heatmap of DEGs for differential gene screening in patient and normal control samples. Red indicates higher expression levels, and blue indicates lower expression levels. (c) Volcano plot of DEGs for differential gene screening in samples before and after immunotherapy. Yellow represents up - regulation, and blue represents down - regulation. (d) Heatmap of DEGs for differential gene screening in samples before and after immunotherapy. Red indicates higher expression levels, and blue indicates lower expression levels. (e) Venn diagram of candidate gene screening. (f) GO enrichment analysis of candidate genes. (g) KEGG pathway analysis of candidate genes. (h) PPI network of candidate genes.\u003c/p\u003e","description":"","filename":"Figure1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8816614/v1/cd150c6871077d0f14fba496.jpg"},{"id":102963106,"identity":"5e77adf9-be4a-4a51-90b1-aae7d76a329a","added_by":"auto","created_at":"2026-02-19 04:13:32","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":2769230,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eMachine learning for screening prognostic genes. \u003c/strong\u003e(a) Forest plot of univariate Cox regression analysis. Prognosis-related genes PLEKHH2 (b), RRM2 (c) and WDR76 (d) obtained from univariate Cox regression analysis, and their PH test plots. Distribution map of patient survival status (e) and risk curve (f) calculated in the training set. Orange dots represent high-risk samples, and green dots represent low-risk samples. (g) Kaplan-Meier survival curve plot of patients calculated by constructing in the training set. (h) ROC curve to evaluate the effectiveness of the model constructed in the training set. Distribution map of patient survival status (e) and risk curve (f) calculated in the validation set. Orange dots represent high-risk samples, and green dots represent low-risk samples. (g) Kaplan-Meier survival curve plot of patients calculated by constructing in the validation set. (h) ROC curve to evaluate the effectiveness of the model in the validation set.\u003c/p\u003e","description":"","filename":"Figure2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8816614/v1/85862cc85d37bb328d6bd7e0.jpg"},{"id":102791237,"identity":"32f9c2fb-af83-46ce-9704-0863df01dbd9","added_by":"auto","created_at":"2026-02-16 17:18:05","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":2668535,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eCorrelation analysis of clinical features and construction of Nomogram plot. \u003c/strong\u003e(a) Difference plot of risk scores among different clinical feature groups. (b) Forest plot of univariate Cox analysis of risk scores and different clinical features. Univariate significant factors m_stage (c), n_stage (d), and riskScore (e) for PH test. (f) Multivariate Cox analysis of factors satisfying the PH test. (g) Construction of the Nomogram plot. Each predictor corresponds to a horizontal line, and different values correspond to different scores. (h) Calibration curve of the Nomogram plot. (i) ROC curve of the Nomogram plot.\u003c/p\u003e","description":"","filename":"Figure3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8816614/v1/785165c1fc831dae210dfac4.jpg"},{"id":102962909,"identity":"10cef99d-fbb2-45e2-947e-9d34811e91a8","added_by":"auto","created_at":"2026-02-19 04:12:05","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":6442372,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eGSEA enrichment analysis. \u003c/strong\u003e(a) GSEA enrichment analysis of high- and low-risk groups. GSEA enrichment analysis of prognostic genes PLEKHH2 (b), RRM2 (c) and WDR76 (d). The figure is divided into three parts. The top part is the enrichment score line chart. Each line represents a pathway, and the peak of each line is the enrichment score of that pathway. The genes before the peak are the core genes under the gene set of that pathway. If the peak is in the upper left corner, it means that the core genes are mainly upregulated genes based on the differential analysis of high- and low-risk groups. If the peak is in the lower right corner, it means that the core genes are mainly downregulated genes based on the differential analysis of high- and low-risk groups. The second part marks the genes located in the gene set with lines. The third part is the rank value distribution of all genes.\u003c/p\u003e","description":"","filename":"Figure4.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8816614/v1/2299a56fff838844551ae8ba.jpg"},{"id":102791252,"identity":"22285315-c628-4153-8ead-27d58736839e","added_by":"auto","created_at":"2026-02-16 17:18:06","extension":"jpg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":9014486,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eImmune infiltration and immune checkpoint analysis. \u003c/strong\u003e(a) Heatmap of immune infiltration abundance in high- and low-risk groups. Different colors represent different immune cell types. (b) Box plots of differences in immune cell infiltration between high- and low-risk groups. ns represents insignificant, *p\u0026lt;0.05, **p\u0026lt;0.01, ***p\u0026lt;0.001, ****p\u0026lt;0.0001. Bubble plots of the correlation between prognostic genes PLEKHH2 (c), RRM2 (d) and WDR76 (e) and differentially infiltrated immune cells. (f) Box plots of the differential expression of immune checkpoint genes in high- and low-risk groups. ns represents insignificant, *p\u0026lt;0.05, **p\u0026lt;0.01, ***p\u0026lt;0.001, ****p\u0026lt;0.0001. Box plots of differences in TIDE (g), Exclusion (h), Dysfunction (i), and MSI (j) between high- and low-risk groups. (k-n) Difference plots of IPS scores in high- and low-risk groups. ns represents insignificant, *p\u0026lt;0.05, **p\u0026lt;0.01.\u003c/p\u003e","description":"","filename":"Figure5.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8816614/v1/1ac0ac410cb2c061bfbb2e4d.jpg"},{"id":102962463,"identity":"f8a1ce67-5870-4d91-9dce-54c25989b6e9","added_by":"auto","created_at":"2026-02-19 04:08:59","extension":"jpg","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":7699282,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSomatic mutation analysis. \u003c/strong\u003eWaterfall plot of TMB analysis in the high-risk group (a) and low-risk group (b). (c) Violin plot of TMB differences between the high- and low-risk groups in the training set, ***p\u0026lt;0.001.\u003c/p\u003e","description":"","filename":"Figure6.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8816614/v1/07fc77f3d25a4c3f77fe25cf.jpg"},{"id":102962984,"identity":"209920bd-3105-47c0-a242-9cfba1c7c410","added_by":"auto","created_at":"2026-02-19 04:12:35","extension":"jpg","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":3189573,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eDrug sensitivity analysis. \u003c/strong\u003eBox plot of the difference in IC50 of anti-tumor drugs between the high- and low-risk groups in the training set, *p\u0026lt;0.05, **p\u0026lt;0.01, ***p\u0026lt;0.001, ****p\u0026lt;0.0001.\u003c/p\u003e","description":"","filename":"Figure7.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8816614/v1/e4de7b3d96b444ef2e3863ca.jpg"},{"id":102791244,"identity":"f75cf419-1724-423b-a59c-d9bc38166cc6","added_by":"auto","created_at":"2026-02-16 17:18:05","extension":"jpg","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":3210153,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSingle - cell analysis. \u003c/strong\u003e(a) Screening of highly variable genes, and highly variable genes are shown in red. (b) PCA sample cell distribution map. (c) PCA scree plot. (d) PCA dimensionality reduction map. (e) UMAP algorithm dimensionality reduction and clustering visualization of cell cluster classification results map. (f) UMAP dimensionality reduction map after annotating cell clusters as different types of cells according to marker genes. (g) Expression bubble plot of marker genes in different cell clusters. (h) Functional enrichment analysis of cell clusters, with colors ranging from blue (low enrichment) to red (high enrichment) representing different degrees of pathway enrichment. (i) Statistical chart of differences in cell content in different groups. (j) Expression difference map of prognostic genes between the NSCLC group and the control group in different types of cells, where ns represents insignificant, *p \u0026lt; 0.05, **p \u0026lt; 0.01, ***p \u0026lt; 0.001.\u003c/p\u003e","description":"","filename":"Figure8.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8816614/v1/2c1b3652597eb43bf9294383.jpg"},{"id":102791250,"identity":"be1f6d99-efc4-49ca-8110-b727ff5ff301","added_by":"auto","created_at":"2026-02-16 17:18:05","extension":"jpg","order_by":9,"title":"Figure 9","display":"","copyAsset":false,"role":"figure","size":5532294,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eCell communication analysis and pseudotemporal analysis of key cell clusters. \u003c/strong\u003eNumber of cell communication networks between different cell types in before PD-1 blockade therapy (a) and after PD-1 blockade therapy (b). The size of the circles of various colors on the periphery represents the number of cells. The larger the circle, the more cells there are. The cells that send out arrows express ligands, and the cells pointed by the arrows express receptors. The more ligand-receptor pairs there are, the thicker the line. Intensity of cell communication networks between different cell types in before PD-1 blockade therapy (c) and after PD-1 blockade therapy (d). Receptor-ligand communication maps between different cell types in before PD-1 blockade therapy (e) and after PD-1 blockade therapy (f). (g) UMAP clustering map of key cells. (h) Pseudotemporal analysis at different differentiation times. (i) Pseudotemporal analysis of the differentiation states of cells at each stage of key cells. (j) Pseudotemporal analysis of cell differentiation in different samples. (k) Pseudotemporal analysis of cell differentiation in different subclusters. (l) Heatmap of the expression of prognostic genes in different developmental states of cells. The color change in the figure reflects the relative intensity of gene expression. The red area indicates a relatively high expression level of the gene at this stage, while the blue area indicates a relatively low expression level of the gene at this stage.\u003c/p\u003e","description":"","filename":"Figure9.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8816614/v1/4c65438695dd82aa37c8e7a3.jpg"},{"id":102791248,"identity":"0dc1fa13-8efd-428d-885c-f0da48a5f330","added_by":"auto","created_at":"2026-02-16 17:18:05","extension":"jpg","order_by":10,"title":"Figure 10","display":"","copyAsset":false,"role":"figure","size":3309765,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eBox plot of prognostic gene expression level analysis, ****p\u0026lt;0.0001\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"Figure10.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8816614/v1/aa8c2aeffc4f0d47af64e2b6.jpg"},{"id":102791251,"identity":"2b33e8d4-6f10-4bf8-aa3f-eb5169f3367a","added_by":"auto","created_at":"2026-02-16 17:18:05","extension":"jpg","order_by":11,"title":"Figure 11","display":"","copyAsset":false,"role":"figure","size":14999030,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003ePCR and IHC of patients’ tissue. \u003c/strong\u003e(a) qRT-PCR validation of RRM2, WDR76 and PLEKHH2 expression levels in normal and tumor tissue; (b) IHC validation of RRM2, WDR76 and PLEKHH2 expression levels in normal and tumor tissue. *P\u0026lt;0.05；**P\u0026lt;0.01；***P\u0026lt;0.001；****P\u0026lt;0.001.\u003c/p\u003e","description":"","filename":"Figure11.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8816614/v1/a60684340712e8fc3485a3e9.jpg"},{"id":103508775,"identity":"461ce828-bc8e-42ce-9146-237428f2978b","added_by":"auto","created_at":"2026-02-26 13:54:07","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":63963307,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8816614/v1/81a8f2d3-4ca3-4c68-9705-0b07f134cc92.pdf"},{"id":102963003,"identity":"5b107518-4d89-409a-aa2d-a2d51eff7b6f","added_by":"auto","created_at":"2026-02-19 04:12:44","extension":"xlsx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":28213,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eAdditional file 1 NEDRGs information from the literature\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"Additionalfile1.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8816614/v1/abbe558e0fffffdc761f1f44.xlsx"},{"id":102962492,"identity":"0d522740-a66e-4cb4-a9e2-ff38b63fa859","added_by":"auto","created_at":"2026-02-19 04:09:23","extension":"xlsx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":19645,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eAdditional file 2 Results of GO enrichment analysis\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"Additionalfile2.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8816614/v1/5862dc90b244b178a360e49d.xlsx"},{"id":102962920,"identity":"fabdde26-155f-41c2-8781-154d0059b190","added_by":"auto","created_at":"2026-02-19 04:12:10","extension":"xlsx","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":13809,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eAdditional file 3 Results of KEGG enrichment analysis\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"Additionalfile3.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8816614/v1/c7494f1b208e6e84cee39765.xlsx"},{"id":102791242,"identity":"60eabcde-cd01-4104-b1c8-5d942f10bdc3","added_by":"auto","created_at":"2026-02-16 17:18:05","extension":"xlsx","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":83227,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eAdditional file 4 Results of GSEA enrichment analysis of RRM2, WDR76 and PLEKHH2\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"Additionalfile4.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8816614/v1/bc2192c757353d9fef585dad.xlsx"},{"id":103503909,"identity":"cc1f8226-6257-4f7b-8314-3b8f0b71e633","added_by":"auto","created_at":"2026-02-26 13:04:37","extension":"xlsx","order_by":5,"title":"","display":"","copyAsset":false,"role":"supplement","size":11441,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eAdditional file 5 Infiltration levels of immune cells in DICs between HG group and LG group\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"Additionalfile5.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8816614/v1/ac43b5ade383b6bcc29765c1.xlsx"},{"id":102962743,"identity":"65a9789a-b2dd-4370-a70b-11c55dc89c49","added_by":"auto","created_at":"2026-02-19 04:10:52","extension":"xlsx","order_by":6,"title":"","display":"","copyAsset":false,"role":"supplement","size":25821,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eAdditional file 6 Different drug IC50 values in HG group and LG group\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"Additionalfile6.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-8816614/v1/2f7b056f9d0b8873a22100cd.xlsx"},{"id":102791246,"identity":"b40d7b00-2544-44dc-ad08-3a1bcf59ee4f","added_by":"auto","created_at":"2026-02-16 17:18:05","extension":"jpg","order_by":7,"title":"","display":"","copyAsset":false,"role":"supplement","size":1847637,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eAdditional file 7 Single - cell data pre - processing. Before (a) and after (b) single - cell data quality control.\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"Additionalfile7.jpg","url":"https://assets-eu.researchsquare.com/files/rs-8816614/v1/fec85e225cbbc695367d7010.jpg"}],"financialInterests":"No competing interests reported.","formattedTitle":"Exploring the mechanism of prognostic genes associated with neuroendocrine differentiation and immunotherapy resistance in non-small cell lung cancer based on the bulk transcriptome and single cell RNA sequencing","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eLung cancer, which originates from the bronchial mucosa or pulmonary glands, is a highly prevalent and lethal malignancy globally (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e). Non-small cell lung cancer (NSCLC) accounts for approximately 85% of all cases, with its primary subtypes being lung adenocarcinoma (LUAD) and lung squamous cell carcinoma (LUSC) (\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e, \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e). Known risk factors for NSCLC include smoking, family history, radiation exposure, and chronic lung diseases (\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e). Although advancements in treatment modalities such as surgery, radiotherapy, chemotherapy, and targeted therapy have improved clinical outcomes, the early asymptomatic nature of NSCLC often leads to late-stage diagnoses. Together with tumor heterogeneity and therapy resistance, these factors contribute to the persistently low overall survival rates (\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e, \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e). Thus, identifying novel prognostic biomarkers and elucidating the molecular mechanisms of NSCLC are urgently needed.\u003c/p\u003e \u003cp\u003eIn recent years, immunotherapy has emerged as a promising approach for cancer treatment. Immune checkpoint inhibitors (ICIs), particularly those targeting programmed death-1 (PD-1) and its ligand (PD-L1), have become a standard treatment for NSCLC, notably prolonging survival in advanced-stage patients either as monotherapy or in combination with chemotherapy (\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e, \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e). However, the majority of NSCLC patients exhibit limited response to ICIs, and primary or acquired resistance remains a major clinical challenge (\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e). Neuroendocrine differentiation (NED), characterized by the presence of tumor cells expressing neuroendocrine markers such as CD56, synaptophysin, and chromogranin A, has been implicated in therapy resistance and is an independent poor prognostic factor in NSCLC(\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e). Although the role of NED is well-established in prostate cancer, its interplay with immunotherapy resistance in NSCLC is not fully understood (\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e, \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e). Therefore, investigating the mechanisms underlying NED and immunotherapy resistance is critical for developing personalized treatment strategies.\u003c/p\u003e \u003cp\u003eSingle-cell RNA sequencing (scRNA-seq) represents a transformative technology that enables the dissection of transcriptional heterogeneity at the cellular level. Unlike bulk RNA sequencing, scRNA-seq reveals cell-type-specific gene expression patterns, cellular subpopulations, and dynamic state transitions, providing unprecedented insights into tumor microenvironments and cell-cell communication networks (\u003cspan additionalcitationids=\"CR15\" citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e). This approach has proven invaluable for identifying rare malignant cells and understanding cancer biology. The application of scRNA-seq technology has advanced our understanding of tumor heterogeneity, immune microenvironment, treatment response and drug resistance mechanisms in NSCLC, providing a more solid research foundation for the diagnosis and treatment of NSCLC.\u003c/p\u003e \u003cp\u003eIn this study, we integrated bulk transcriptomic data from public NSCLC datasets and scRNA-seq data to identify prognostic genes associated with NED and immunotherapy resistance. Using machine learning algorithms, we constructed a risk model and performed comprehensive analyses of biological functions, immune infiltration, and drug sensitivity. Collectively, our results offer novel insights into NSCLC pathogenesis and highlight potential targets for augmenting the efficacy of immunotherapeutic interventions.\u003c/p\u003e"},{"header":"2. Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003e2.1 Sample collection and sequencing\u003c/h2\u003e \u003cp\u003eWe retrieved the TCGA-NSCLC dataset from The Cancer Genome Atlas (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://xena.ucsc.edu/\u003c/span\u003e\u003cspan address=\"https://xena.ucsc.edu/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e), which consisted of 989 tumor samples with survival data and 109 control samples. In addition, the clinical data (age, gender, stage, and TNM stage) and somatic mutation data of 989 NSCLC samples were downloaded, and the access time was March 14, 2025. A total of 13 NSCLC tissue samples before PD-1 blockade therapy and 29 NSCLC tissue samples after PD-1 blockade therapy from the GSE248249 dataset (GPL23126), a total of 196 tissue samples of NSCLC with survival information in the GSE37745 dataset (GPL570), and a total of 15 NSCLC samples from the GSE207422 dataset (GPL24676) were included from the Gene Expression Omnibus (GEO) database (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.ncbi.nlm.nih.gov/geo/\u003c/span\u003e\u003cspan address=\"https://www.ncbi.nlm.nih.gov/geo/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e), comprising 3 NSCLC tissue samples PD-1 blockade therapy and 12 NSCLC tissue samples after PD-1 blockade therapy. The 1482 NEDRGs were obtained from the literature (\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e) (\u003cb\u003eAdditional file 1\u003c/b\u003e). To validate the bioinformatic findings, we collected paired tumor and normal tissues from six NSCLC patients (three stage I patients and three stage III patients) undergoing surgery at Tianjin Medical University General Hospital for PCR and IHC validation.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003e2.2 Identification of differentially expressed genes (DEGs) and candidate genes\u003c/h2\u003e \u003cp\u003eFirst, we identified DEGs between NSCLC and control samples in the TCGA-NSCLC dataset using \"DESeq2\" (v 1.38.0(\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e)) (P\u0026thinsp;\u0026lt;\u0026thinsp;0.05, |log2 fold change (FC)| \u0026gt; 0.5), naming this gene set DEGs1. Then, we visualized DEGs1 in a volcano plot via the \"ggplot2\" package (v 3.4.1(\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e)), labeling the top 10 most significantly up- and down-regulated genes. Then, the top 10 up or down DEGs1 between the NSCLC and control samples were also depicted via the \u0026ldquo;ComplexHeatmap\u0026rdquo; package (v 2.14.0) (\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eUsing the \"limma\" package (v 3.54.1) (\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e), we identified DEGs between pre- and post-PD-1 blockade therapy samples in GSE248249 (P\u0026thinsp;\u0026lt;\u0026thinsp;0.05, |log2FC| \u0026gt; 0.5) and defined this gene set as DEGs2. The results were depicted by the same methods mentioned earlier. Candidate genes were defined as the common set resulting from the intersection of DEGs1, DEGs2, and NEDRGs, identified using the \"ggvenn\" package (v 0.1.9) (\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003e2.3 Enrichment pathways and protein-protein interaction (PPI) network\u003c/h2\u003e \u003cp\u003eThe Kyoto Encyclopedia of Genes and Genomes enrichment (KEGG) and Gene Ontology (GO) analyses were applied via the \u0026ldquo;clusterProfiler\u0026rdquo; package (v 4.2.2) (\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e) to explore the functions of candidate genes (P\u0026thinsp;\u0026lt;\u0026thinsp;0.05). The top 5 notably enriched terms or all terms in GO and all enriched terms in KEGG were depicted in this study. To investigate protein-level interactions, a PPI network for the candidate genes was constructed using the Search Tool for the Retrieval of Interacting Genes (STRING) database (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://string-db.org/\u003c/span\u003e\u003cspan address=\"https://string-db.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) with a confidence score threshold\u0026thinsp;\u0026gt;\u0026thinsp;0.4, followed by visualization in Cytoscape (v 3.7.1) (\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003e2.4 Development and verification of the risk model\u003c/h2\u003e \u003cp\u003ePrognostic genes were screened by univariate Cox regression (\"survival\" v 3.7-0(\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e)) and visualized (\"forestplot\" v 3.1.1(\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e)); after PH assumption testing (P\u0026thinsp;\u0026gt;\u0026thinsp;0.05), an RSF model (\"randomForestSRC\" v 3.2.2(\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e); ntree\u0026thinsp;=\u0026thinsp;5, mtry\u0026thinsp;=\u0026thinsp;6) was built to calculate risk scores and stratify patients into High risk group (HG)/Low risk group (LG), with results displayed in risk and survival status plots. Survival differences were assessed by comparing K-M curves (\"survminer\" v 0.4.9) with the Log-rank test (P\u0026thinsp;\u0026lt;\u0026thinsp;0.05); meanwhile, predictive accuracy for 1-, 3-, and 5-year OS was evaluated via time-dependent ROC analysis (\"timeROC\" v 0.4(\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e)) with an area under the curve (AUC)\u0026thinsp;\u0026gt;\u0026thinsp;0.6 threshold. The above analyses were all applied in the 989 NSCLC samples in the TCGA-NSCLC dataset. Additionally, a risk model was established in the GSE37745 cohort (n\u0026thinsp;=\u0026thinsp;196) by constructing a risk score for each NSCLC sample. The predictive efficiency of the model in GSE37745 dataset was validated utilizing identical methodology.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003e2.5 The spread of risk scores in different clinical groups\u003c/h2\u003e \u003cp\u003eIn the TCGA-NSCLC dataset, the distribution of risk scores across clinical subgroups was assessed using non-parametric tests: the Wilcoxon test for binary variables (age, gender, M stage) and the Kruskal-Wallis test for multi-category variables (N stage, T stage, overall stage), with a uniform significance threshold of P\u0026thinsp;\u0026lt;\u0026thinsp;0.05.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003e2.6 Construction and analysis of nomogram\u003c/h2\u003e \u003cp\u003eTo assess the independent prognostic value of the risk score and clinical features in early-stage NSCLC, we conducted univariate Cox regression (P\u0026thinsp;\u0026lt;\u0026thinsp;0.05, HR\u0026thinsp;\u0026ne;\u0026thinsp;1), proportional hazards (PH) assumption testing (P\u0026thinsp;\u0026gt;\u0026thinsp;0.05), and multivariate Cox regression (P\u0026thinsp;\u0026lt;\u0026thinsp;0.05, HR\u0026thinsp;\u0026ne;\u0026thinsp;1) using the \"survival\" package (v 3.7-0). Subsequently, a nomogram was developed with the \"rms\" package (v 6.5-0) (\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e) to predict 1-, 3-, and 5-year survival probabilities. In the nomogram, Each factor corresponded to each points, and the total points was obtained by adding the points of each factor. The survival rates of NSCLC were inferred by the total points. The nomogram's predictive performance was validated by employing a calibration curve (\"rms\" package, v 6.5-0) and a receiver operating characteristic (ROC) curve (\"survivalROC\" package, v 1.0.3.1(\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e)) to assess its accuracy and discriminative power (AUC\u0026thinsp;\u0026gt;\u0026thinsp;0.7), respectively. The above analyses were all applied in the 989 NSCLC samples in the TCGA-NSCLC dataset.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003e2.7 gene set enrichment analysis (GSEA) analysis\u003c/h2\u003e \u003cp\u003eThis study utilized the \"c2.cp.kegg.v7.5.1.symbols.gmt\" gene set from MSigDB (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.gsea-msigdb.org/\u003c/span\u003e\u003cspan address=\"https://www.gsea-msigdb.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) as the reference. Differential expression analysis between risk groups (\"DESeq2\" v 1.38.0) provided log2FC values for standard GSEA. To specifically investigate the prognostic genes, their functions were elucidated by performing GSEA on gene lists ranked by Pearson correlation with each prognostic gene (\"psych\" package v 2.2.9(\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e)). Subsequently, GSEA was executed across all TCGA-NSCLC samples via the \"clusterProfiler\" package (v 4.2.2) under thresholds of adj.P\u0026thinsp;\u0026lt;\u0026thinsp;0.05, |normalized enrichment score (NES)| \u0026gt; 1, and False Discovery Rate (FDR)\u0026thinsp;\u0026lt;\u0026thinsp;0.25, and the top 5 pathways (P\u0026thinsp;\u0026lt;\u0026thinsp;0.05) were visualized with the \"enrichplot\" package (v 1.18.0) (\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec10\" class=\"Section2\"\u003e \u003ch2\u003e2.8 Immune analysis\u003c/h2\u003e \u003cp\u003eWe characterized the immune landscape by first quantifying the infiltration of 22 immune cell types in the TCGA-NSCLC cohort using CIBERSORT (\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e) ( samples with P\u0026thinsp;\u0026gt;\u0026thinsp;0.05 were excluded) and visualizing the results with \"ggplot2\" (v 3.4.1). Differential immune cells (DICs) between HG and LG groups were identified via the Wilcoxon test (P\u0026thinsp;\u0026lt;\u0026thinsp;0.05). Subsequently, the correlations between DICs and prognostic genes were assessed by Spearman analysis (\"psych\" package v 2.2.9(\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e)) using thresholds of |cor| \u0026gt; 0.3 and P\u0026thinsp;\u0026lt;\u0026thinsp;0.05. In parallel, we compared the expression of 41 immune checkpoints from literature (\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e) between the risk groups (Wilcoxon test, P\u0026thinsp;\u0026lt;\u0026thinsp;0.05) and retrieved TIDE-related scores (TIDE, exclusion, dysfunction) and MSI from the TIDE database (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://tide.dfci.harvard.edu/\u003c/span\u003e\u003cspan address=\"https://tide.dfci.harvard.edu/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). To evaluate disparities in immunotherapy response potential between the risk groups, we performed two comparative analyses using the Wilcoxon test (P\u0026thinsp;\u0026lt;\u0026thinsp;0.05). First, we compared the TIDE-related scores between HG and LG. Second, we contrasted the immunophenotype score (IPS), which was acquired from the TCIA database (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://tcia.at/home\u003c/span\u003e\u003cspan address=\"https://tcia.at/home\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e), across the same groups.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003e2.9 The analysis of somatic mutation\u003c/h2\u003e \u003cp\u003eUsing the \"maftools\" package (v 2.18.0(\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e)), we investigated somatic mutations in the TCGA-NSCLC dataset by profiling the top 20 mutated genes and comparing tumor mutational burden (TMB) between the HG and LG groups with the Wilcoxon test (P\u0026thinsp;\u0026lt;\u0026thinsp;0.05).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003e2.10 Drug sensitivity analysis\u003c/h2\u003e \u003cp\u003eDrug data from the Genomics of Drug Sensitivity in Cancer (GDSC) database (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.cancerrxgene.org/\u003c/span\u003e\u003cspan address=\"https://www.cancerrxgene.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) were analyzed by predicting IC\u003csub\u003e50\u003c/sub\u003e values (\"pRRophetic\" v 0.5(\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e)); differences between HG and LG groups in TCGA-NSCLC were tested (Wilcoxon test, P\u0026thinsp;\u0026lt;\u0026thinsp;0.05), and the top 20 most significant drugs were displayed.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003e2.11 Single cell RNA sequencing (scRNA-seq) analysis\u003c/h2\u003e \u003cp\u003eThe scRNA-seq data were processed utilizing the \u0026ldquo;Seurat\u0026rdquo; package (v 5.1.0) (\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e). Cells were filtered based on three quality metrics: nFeature_RNA (200\u0026ndash;3500), representing the number of detected genes; nCount_RNA (200\u0026ndash;6000), indicating the total transcript count; and percent.mt (\u0026lt;\u0026thinsp;20%), reflecting the mitochondrial gene proportion. In addition, the genes covering at least 3 cells were retained. Single-cell data preprocessing in Seurat (v 5.1.0) included normalization of counts per cell (scale factor 10,000, log-transformed), identification and labeling of the top 2,000 (top 10 labeled) highly variable genes, and principal component analysis (PCA)-based assessment of batch effects. After that, normalization analysis of all samples was applied utilizing the Scale Data function of the \u0026ldquo;Seurat\u0026rdquo; package (v 5.1.0). The PCA analysis of hypervariable genes was applied via the \u0026ldquo;ExPosition\u0026rdquo; package (v 2.8.23) (\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e). The significance of principal components was further evaluated using the JackStrawPlot function. Then, the number of principal components (PCs) corresponding to the point where the curve starts to level off was selected for cell clustering (P\u0026thinsp;\u0026lt;\u0026thinsp;0.05, dims\u0026thinsp;=\u0026thinsp;30). Cell clustering was performed using the \"Seurat\" package (v 5.1.0) via the FindNeighbors and FindClusters functions (resolution\u0026thinsp;=\u0026thinsp;0.25), followed by dimensionality reduction with the RunUMAP function for visualization. The FindAllMarkers function of the \u0026ldquo;Seurat\u0026rdquo; package (v 5.1.0) was utilized to obtain the marker genes of cells (parameter settings min.pct\u0026thinsp;=\u0026thinsp;0.1, logfc.threshold\u0026thinsp;=\u0026thinsp;0.25, only.pos\u0026thinsp;=\u0026thinsp;TRUE). Finally, the cell annotation was executed according to the marker genes acquired in the literature (\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e) and CellMarker database (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://xteam.xbio.top/CellMarker/\u003c/span\u003e\u003cspan address=\"http://xteam.xbio.top/CellMarker/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). To characterize the functional profiles of different cell types, we first visualized marker gene expression. We then performed single-cell functional enrichment using the analyze_sc_clusters function (\"ReactomeGSA\" v 1.12.0(\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e)). Subsequently, the pathways function was used to calculate the enrichment score range for each pathway across cell types. Finally, the top 10 most variable pathways were displayed using the plot_gsva_heatmap function.\u003c/p\u003e \u003cp\u003eProportions of cells per sample were compared using the Wilcoxon test (P\u0026thinsp;\u0026lt;\u0026thinsp;0.05); meanwhile, prognostic gene expression in all cells was compared between before PD-1 blockade therapy and after PD-1 blockade therapy groups using the Wilcoxon test (P\u0026thinsp;\u0026lt;\u0026thinsp;0.05). Then, the cells with different expression of genes and different proportion between 2 groups were selected as key cells. The interactions and communication between cells were explored via the \u0026ldquo;CellChat\u0026rdquo; package (v 1.6.1) (\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e) (P\u0026thinsp;\u0026lt;\u0026thinsp;0.05).\u003c/p\u003e \u003cp\u003eTo reclassify key cells into different clusters, the same methods mentioned earlier were utilized and the \u0026ldquo;DDRTree\u0026rdquo; package (v 0.1.5) (\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e) was used to visualize the results of clusters. Afterwards, the clusterCells function of the \u0026ldquo;Seurat\u0026rdquo; package (v 5.1.0) was utilized to divide cluster into different subgroups. Pseudotemporal analysis of key cells was conducted with the \"Monocle\" package (v 2.30.1) (\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e), followed by visualization of prognostic gene expression along the pseudotime continuum.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003e2.12 Gene expression analysis\u003c/h2\u003e \u003cp\u003eDifferential expression of prognostic genes was evaluated in the TCGA-NSCLC cohort by Wilcoxon test (P\u0026thinsp;\u0026lt;\u0026thinsp;0.05).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003e2.13 Quantitative Real-Time Polymerase Chain Reaction (qRT-PCR)\u003c/h2\u003e \u003cp\u003eTotal RNA was extracted from tissues using TRIzol (ThermoFisher, USA). Isolated RNA was reverse-transcribed into cDNA using a First Strand cDNA Synthesis Kit (Thermo). SYBR green-based qRT-PCR was performed on the 7900HT Fast Real-Time PCR System (Applied Biosystems/Life Technologies, Waltham, USA). Relative mRNA expression levels were calculated using the 2^\u0026minus;ΔΔCt method. Primer sequences were as follows:\u003c/p\u003e \u003cp\u003e① PLEKHH2-F: CCCGCTGCTGAGATAGACAG\u003c/p\u003e \u003cp\u003e② PLEKHH2-R: AGCTGTTGCATCTTCTCTGCT\u003c/p\u003e \u003cp\u003e③RRM2-F: TTTAAAGGCTGCTGGAGTGAGG\u003c/p\u003e \u003cp\u003e④RRM2-R: GCTGCTTTAGTTTTCGGCTCC\u003c/p\u003e \u003cp\u003e⑤WDR76-F: TCTTGCCCCCTTTTCACTCA\u003c/p\u003e \u003cp\u003e⑥WDR76-R: CATCTGCCGTTCTCTTGGGT\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003e2.14 Immunohistochemistry (IHC)\u003c/h2\u003e \u003cp\u003eUsing the kit (E-IR-R217, Elabscience, China), antigen retrieval was performed on tissue sections according to the instructions. The sections were incubated overnight with RRM2(Proteintech, 11661-1-AP), WDR76(Proteintech, 25528-1-AP) or PLEKHH2(Proteintech, 14204-1-AP) Polyclonal Antibody. The following day, the sections underwent haematoxylin staining and dehydration before being mounted. The nuclear H-score was calculated as the sum of the percentages of nuclei exhibiting low, medium, and high intensity staining.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec17\" class=\"Section2\"\u003e \u003ch2\u003e\u003cb\u003e2.15 Statistical analysis\u003c/b\u003e\u003c/h2\u003e \u003cp\u003eStatistical analyses were carried out in R (v 4.2.2), employing the Wilcoxon test for group comparisons with a significance threshold of P\u0026thinsp;\u0026lt;\u0026thinsp;0.05. Statistical differences between two conditions were calculated by unpaired parametric t-test.\u003c/p\u003e \u003c/div\u003e"},{"header":"3. Results","content":"\u003cdiv id=\"Sec19\" class=\"Section2\"\u003e \u003ch2\u003e3.1 Identification and exploration of candidate genes in NSCLC\u003c/h2\u003e \u003cp\u003eDifferential expression analysis yielded 19,106 DEGs1 (13,197 up-regulated in NSCLC; Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ea, b) from the TCGA cohort and 201 DEGs2 (99 up-regulated pre-treatment; Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ec, d) from the PD-1 blockade therapy dataset. Ultimately, 14 candidate genes were obtained in this study (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ee).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThe candidate genes were significantly enriched (P\u0026thinsp;\u0026lt;\u0026thinsp;0.05) across 80 biological processes (BPs), 4 cellular components (CCs), and 15 molecular functions (MFs) in GO analysis, including DNA replication, intercellular bridge, and clathrin heavy chain binding, respectively (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ef, \u003cb\u003eAdditional file 2\u003c/b\u003e). Our analysis identified significant enrichment of the candidate genes in 7 KEGG pathways, such as DNA replication and pyruvate metabolism (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eg). Furthermore, in the PPI network, only 4 candidate genes had interactions with each other (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eh). Comprehensively, RRM2 had an interactive relationship with WDR76, CDC7 and RNASEH2A.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec20\" class=\"Section2\"\u003e \u003ch2\u003e3.2 Prediction accuracy of risk models\u003c/h2\u003e \u003cp\u003eA total of 3 prognostic genes including RRM2, WDR76 and PLEKHH2 were obtained (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ea-d). Based on an established risk model, TCGA-NSCLC patients were classified into high-risk (HG, n\u0026thinsp;=\u0026thinsp;555) and low-risk (LG, n\u0026thinsp;=\u0026thinsp;434) groups using the optimal risk score cutoff of 49.27128. The results of risk graph indicated that the number of death cases increased with the increase of risk score, and the number of death for patients in LG was notably lower than that of patients in HG in TCGA-NSCLC dataset (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ee \u003cb\u003eand f\u003c/b\u003e). Furthermore, the HG group was associated with significantly poorer survival (P\u0026thinsp;\u0026lt;\u0026thinsp;0.0001, Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eg), corroborated by the model's strong predictive power across 1-, 3-, and 5-year timepoints (AUC: 0.78, 0.77, 0.79; all \u0026gt;\u0026thinsp;0.70, Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eh). Meanwhile, NSCLC patients in GSE37745 dataset were also classified into HG (126 patients) and LG (70 patients) with the best truncation value of risk score (46.78981). The risk stratification pattern observed in the GSE37745 dataset was consistent with that in the TCGA-NSCLC dataset (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ei \u003cb\u003eand j\u003c/b\u003e). The NSCLC patients in HG had lower survival rates (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ek) (P\u0026thinsp;\u0026lt;\u0026thinsp;0.0027).The model achieved AUCs of 0.60 (1-year), 0.61 (3-year), and 0.61 (5-year) in time-dependent ROC analysis (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003el). All these results suggested that the risk model exhibited substantial predictive accuracy for the survival rate of NSCLC.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec21\" class=\"Section2\"\u003e \u003ch2\u003e3.3 Construction of nomogram with independent prognostic factors\u003c/h2\u003e \u003cp\u003eThe risk score distribution differed significantly across T stages (T1-T4), N stages (N0-N2), and overall stages (I-IV) (P\u0026thinsp;\u0026lt;\u0026thinsp;0.0001; Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ea). Consistent with these associations, multivariate Cox regression confirmed that both the risk score and the N stage served as independent prognostic factors (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eb-f). In the nomogram, the total points were inversely correlated with NSCLC survival probability. The risk score held superior predictive weight over the N stage (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eg). The favorable slope of the calibration curves (close to 1) confirmed the model's reliable predictive accuracy (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eh). The all AUC were exceeding 0.7, indicating the nomogram had well predictive efficiency for NSCLC (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ei). In summary, the constructed nomogram was validated as an accurate predictor of survival probability in NSCLC patients.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec22\" class=\"Section2\"\u003e \u003ch2\u003e3.4 Enrichment pathways linked to prognostic genes and differences enrichment pathways in the HG and LG\u003c/h2\u003e \u003cp\u003eIn this study, 14 prominent pathways such as cell cycle and DNA replication were found in the HG and LG (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ea, \u003cb\u003eAdditional file 3\u003c/b\u003e). In addition, RRM2, WDR76 and PLEKHH2 were all enriched in the cell cycle, and WDR76 and PLEKHH2 were all enriched in the notably pathways such as DNA Replication (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eb-d, \u003cb\u003eAdditional file 4).\u003c/b\u003e These results indicated that these notably enriched pathways might play an important role in NSCLC progression.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cb\u003e3.5 The DICs linked to prognostic genes and differences in immune therapy responsein the HG and LG\u003c/b\u003e \u003c/p\u003e \u003cp\u003eComparative analysis identified a markedly higher infiltration of resting CD4\u0026thinsp;+\u0026thinsp;memory T cells in the LG than in the HG (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ea). In addition, a total of 11 DICs such as activated memory CD4 T cells were obtained (P\u0026thinsp;\u0026lt;\u0026thinsp;0.05) (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eb, \u003cb\u003eAdditional file 5\u003c/b\u003e). Moreover, RRM2 and WDR76 (cor\u0026thinsp;\u0026gt;\u0026thinsp;0.30, P\u0026thinsp;\u0026lt;\u0026thinsp;0.05) were notable positive linked to activated CD4 memory T cells (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ec-e). These results indicated these DICs, especially activated memory CD4 T cells might have a notable impact with biomarkers on NSCLC. In addition, the expressions of the immune checkpoints such as CD276 and NRP1 in the HG were higher than that in the LG (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ef). The TIDE score in the HG was higher than that in the LG (P\u0026thinsp;\u0026lt;\u0026thinsp;0.05), indicating a lower success rate of immunotherapy in the HG (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eg). A significantly higher exclusion prediction score was observed in the HG versus the LG (P\u0026thinsp;\u0026lt;\u0026thinsp;0.05), implying a heightened state of T cell rejection in the high-risk group (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eh). The LG exhibited a significantly lower T cell dysfunction score than the HG (P\u0026thinsp;\u0026lt;\u0026thinsp;0.05, Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ei), indicative of better-preserved T cell function; however, the IPS was comparable between the groups (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ej). In addition, the results showed that the score of IPS-CTLA4 (-)/PD-1(-) and IPS-CTLA4(+)/PD-1(-) had notable differences between the 2 groups, indicating that the expression of CTLA-4 might be a key factor driving differences in immune status in the tumor microenvironment lacking activation of the PD-1 pathway (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ek-n).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec23\" class=\"Section2\"\u003e \u003ch2\u003e3.6 The differences of somatic mutation\u003c/h2\u003e \u003cp\u003eThe waterfall plot demonstrated that the mutation type in the 2 groups was missense mutation and multiple mutations (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003ea \u003cb\u003eand b\u003c/b\u003e).Mutation profiling identified TP53 as the most frequently mutated gene, with rates of 72.00% in the HG and 61.00% in the LG, highlighting the importance of gene mutations in NSCLC. Moreover, elevated TMB scores in the HG compared to the LG indicated potentially enhanced immunogenicity in high-risk patients (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003ec).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec24\" class=\"Section2\"\u003e \u003ch2\u003e3.7 The differences of IC\u003csub\u003e50\u003c/sub\u003e with drugs in the HG and LG\u003c/h2\u003e \u003cp\u003eA total of 109 drugs with notable differences in IC\u003csub\u003e50\u003c/sub\u003e values between HG and LG (P\u0026thinsp;\u0026lt;\u0026thinsp;0.05) (\u003cb\u003eAdditional file 6)\u003c/b\u003e. For example, the IC\u003csub\u003e50\u003c/sub\u003e of drugs such as AZD6738, AZD7762 in the HG was notably lower than that in the HG (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003e). In conclusion, the above drugs with different sensitivities in the HG and LG, indicating that patients in different groups need to choose different drugs.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec25\" class=\"Section2\"\u003e \u003ch2\u003e3.8 Identification of epithelials\u003c/h2\u003e \u003cp\u003eAfter undergoing quality control, 64,245 cells and 24,292 genes were obtained (\u003cb\u003eAdditional file 7a and b\u003c/b\u003e). Then, the top 2,000 hypervariable genes such as IGKC\u003c/p\u003e \u003cp\u003ewere selected for PCA (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003ea). In the absence of detectable batch effects, cells were clustered into 20 distinct populations using the top 30 principal components (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eb-e). Then, a total of 8 cells types (B cells, T cells, epithials, stromal cells, mast cells, myeloids, neutrophils and plasma cells) were obtained (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003ef). Cell type-specific marker genes, including MS4A1 for B cells, were predominantly expressed in their respective cell populations (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eg). Enrichment analysis identified significant pathways, including COX reactions, in the eight cell clusters (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eh), while cellular composition analysis showed that proportion of all cells differed between the two groups of samples (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003ei). Among the eight cell clusters, significant differences in PLEKHH2, RRM2, and WDR76 expression between Samples before and after treatment were exclusively observed in epithelials (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003ej). So epithelials were selected as key cell in this study.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec26\" class=\"Section2\"\u003e \u003ch2\u003e3.9 The ommunication networks among cells and pseudo temporal analysis of epithelials\u003c/h2\u003e \u003cp\u003eAmong the communication networks, the number of communication interactions between myeloids, epithelials and neutrophil cells were large relatively, and the communication strength between myeloids and neutrophils were relatively high in the 2 groups In addition, in the treated samples, the number of communication between endothelials and myeloid cells increased (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003ea-d). In addition, the communication strength of ANXA1-FPR1 between epithelials and neutrophil cells in the 2 groups was the highest (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003ee \u003cb\u003eand f\u003c/b\u003e). Then, epithelials were reclustered into 10 cell clusters (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003eg). The differentiation of epithelials was classified into 3 stages, and 5 clusters, 8 clusters, 7 clusters, and 9 clusters were the main clusters in the early stage (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003eh-k). RRM2 and WDR76 were mainly expressed in the later stages of differentiation, while PLEKHH2 were mainly expressed in the middle stage of the temporal sequence (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003el).The results indicated that the development of key cells may also be associated with the expression of prognostic genes.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec27\" class=\"Section2\"\u003e \u003ch2\u003e3.10 The expression of 3 prognostic genes\u003c/h2\u003e \u003cp\u003eThe expression of 3 prognostic genes all had differences between NSCLC and control groups (P\u0026thinsp;\u0026lt;\u0026thinsp;0.05)\u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e10\u003c/span\u003e\u003cb\u003e)\u003c/b\u003e. The expressions of RRM2 and WDR76 were all higher in the NSCLC group while the expression of PLEKHH2 were opposite. To further verify, normal lung tissues and corresponding lung cancer tissues were collected from six clinical patients, including three stage I patients and three stage III patients. The expression levels of RRM2, WDR76, and PLEKHH2 were validated using PCR(Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e11\u003c/span\u003ea) and IHC(Fig.\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e11\u003c/span\u003eb). Consistent with previous analyses, the results showed that RRM2 and WDR76 expression was significantly higher in cancer tissues compared to normal tissues, and higher in stage III patients than in stage I patients. In contrast, PLEKHH2 expression was significantly lower in cancer tissues than in normal tissues and lower in stage III patients than in stage I patients. These experimental findings further suggest that RRM2 and WDR76 may promote tumor malignant progression, whereas PLEKHH2 likely plays an inhibitory role.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"4. Discussion","content":"\u003cp\u003eImmunotherapy resistance in NSCLC is a major challenge currently faced in clinical practice (\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e). In recent years, NED has been recognized as playing a critical role in tumor progression and therapy resistance (\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e); however, its specific mechanism in mediating immune resistance in NSCLC remains unclear.\u003c/p\u003e \u003cp\u003eIn this study, we identified three prognostic genes, namely RRM2, WDR76, and PLEKHH2, which are closely associated with neuroendocrine differentiation and immunotherapy resistance in NSCLC. The risk model, constructed from these genes, was validated in both the TCGA and independent GEO cohorts, demonstrating consistent and robust predictive accuracy for patient survival. Furthermore, we elucidated the biological pathways, immune microenvironment characteristics, and potential therapeutic compounds associated with high- and low-risk groups.\u003c/p\u003e \u003cp\u003eRRM2 has been implicated in oxidative stress response and chemoresistance. Under iron-rich conditions, RRM2 is prone to depolymerization and degradation, making it a potential therapeutic target (\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e). Previous studies have reported RRM2 overexpression in chemotherapy-resistant cancers, and its inhibition may enhance chemosensitivity (\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e). In NSCLC, elevated RRM2 expression correlates with distant metastasis and poorer survival, underscoring its value as both a diagnostic and prognostic biomarker (\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e). This study further reveals that RRM2 is significantly upregulated in high-risk NSCLC populations and is closely associated with cell cycle pathways. Given its central role in DNA synthesis, our findings, showing elevated RRM2 in high-risk patients and its enrichment in cell cycle pathways, lead us to hypothesize that RRM2-driven tumor proliferation may contribute to an immunosuppressive microenvironment, potentially through mechanisms such as T cell exhaustion or upregulation of immune checkpoints, which in turn could underlie immunotherapy resistance (e.g., G1/S transition) (\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e). Highly proliferative tumors often exhibit reduced T-cell infiltration or functional exhaustion and may upregulate immune checkpoint molecules, thereby weakening the response to immunotherapy (\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e). Additionally, RRM2-mediated genomic instability may increase tumor mutational burden (TMB) but could also promote immune editing and heterogeneous clonal evolution, ultimately leading to immune escape (\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e). Therefore, RRM2 is not only a biomarker for poor prognosis in NSCLC but also a potential therapeutic target. Inhibitors targeting RRM2 (such as 3-AP, COH29, etc.) have shown synergistic effects with chemotherapy and immunotherapy in preclinical studies. Based on the findings of this study, future research could explore combination therapy strategies of RRM2 inhibitors with PD-1/PD-L1 inhibitors in high-risk NSCLC populations, with the aim of reversing immune resistance and improving patient outcomes.\u003c/p\u003e \u003cp\u003eWDR76, a WD40 repeat-containing protein, participates in DNA damage repair, ubiquitin-mediated degradation, and epigenetic regulation. Although its role in NSCLC is less studied, WDR76 has been shown to destabilize RAS proteins in liver cancer and sensitize colorectal cancer cells to 5-fluorouracil (\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e, \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e). Our study is the first to systematically elucidate the clinical significance and immunoregulatory function of WDR76 in NSCLC. As an adaptor of the CRL4 (Cullin4-RING E3 ubiquitin ligase) complex, WDR76 mediates the ubiquitination and degradation of various tumor suppressors (such as p53, p27, etc.), thereby promoting tumor cell proliferation and survival (\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e, \u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e). Additionally, WDR76 is involved in maintaining genomic stability by regulating key proteins in the DNA damage response, influencing tumor cell sensitivity to chemotherapy and radiotherapy (\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e). Notably, our bioinformatic analysis identified a significant positive association between WDR76 expression and the infiltration level of activated memory CD4\u003csup\u003e+\u003c/sup\u003e T cells, suggesting its potential role in influencing NSCLC progression via modulation of the immune microenvironment. In summary, our data suggest that WDR76, functioning as a molecular scaffold, may contribute to NSCLC progression by regulating both intrinsic tumor cell proliferation and the extrinsic immune microenvironment, as evidenced by its correlation with activated memory CD4\u0026thinsp;+\u0026thinsp;T cells. Targeting WDR76 and its related pathways may provide novel strategies for enhancing the efficacy of immunotherapy in NSCLC.\u003c/p\u003e \u003cp\u003ePLEKHH2 contains multiple functional domains, including PH, MyTH, and FERM domains, and is involved in cytoskeletal organization and signal transduction. Although initially studied in psychiatric disorders, recent evidence indicates that PLEKHH2 promotes malignant phenotypes in NSCLC by activating the FAK/PI3K/AKT pathway through interaction with β-arrestin1 (\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e). Our results suggest that PLEKHH2 may influence both neuroendocrine differentiation and T-cell-mediated immunity, thus establishing it as a key molecular node linking intrinsic tumor characteristics with immune microenvironment remodeling. Therefore, PLEKHH2 may act as a key molecular node that links the intrinsic characteristics of tumor cells with the remodeling of the immune microenvironment by regulating the signal transduction and cell adhesion of tumor cells. On the one hand, it induces neuroendocrine phenotypic transformation (associated with treatment resistance), and on the other hand, it affects the recruitment and activation of T cells in the tumor microenvironment.\u003c/p\u003e \u003cp\u003ePathway enrichment analysis revealed that cell cycle and DNA replication pathways are significantly activated in high-risk NSCLC patients. The three key prognostic genes, RRM2, WDR76, and PLEKHH2, are all closely associated with these pathways, suggesting their potential roles as core molecules synergistically driving dysregulated cell cycle and genomic instability in NSCLC. As a critical subunit of ribonucleotide reductase, RRM2 provides sufficient dNTP precursors for DNA synthesis, which is essential for proper S-phase progression (\u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e52\u003c/span\u003e). WDR76, functioning as an adaptor for the E3 ubiquitin ligase complex, may influence G1/S phase transition by regulating the stability of cyclins or cyclin-dependent kinase inhibitors such as p21 and p27 (\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e). Meanwhile, PLEKHH2 may indirectly affect cell cycle progression by integrating extracellular signals with intracellular cytoskeletal reorganization. The aberrant overexpression of these three genes collectively forms a molecular network supporting rapid tumor cell proliferation. Furthermore, aberrant activation of the cell cycle not only directly promotes tumor proliferation but also shapes the immunosuppressive microenvironment through multiple mechanisms. Rapidly proliferating tumor cells may upregulate immune checkpoint molecules like PD-L1 (\u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e53\u003c/span\u003e), induce T-cell exhaustion, and secrete immunosuppressive cytokines such as TGF-β and IL-10, which suppress effector immune cell function and recruit regulatory T cells (Tregs) (\u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e54\u003c/span\u003e). Increased genomic instability may enhance tumor antigen heterogeneity, enabling some tumor cell clones to evade immune recognition. These mechanisms collectively contribute to immunotherapy resistance. Therefore, our study not only elucidates the roles of RRM2, WDR76, and PLEKHH2 as key nodes in the cell cycle regulatory network during NSCLC progression but, more importantly, provides precise therapeutic directions for high-risk NSCLC patients. The combination of inhibitors targeting these pathways with immune checkpoint inhibitors may yield synergistic antitumor effects by simultaneously targeting intrinsic tumor proliferation mechanisms and extrinsic immune microenvironment regulation. This molecular stratification-based combination therapy strategy holds promise for overcoming immunotherapy resistance in NSCLC and improving patient outcomes.\u003c/p\u003e \u003cp\u003eOur immune infiltration analysis revealed an intriguing phenomenon: activated memory CD4\u0026thinsp;+\u0026thinsp;T cells showed a positive correlation with the expression of high-risk genes RRM2 and WDR76. Given the high heterogeneity of CD4\u0026thinsp;+\u0026thinsp;T cell populations (including immunosuppressive regulatory T cells [Tregs] and functionally exhausted helper T cells) (\u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e55\u003c/span\u003e), this finding may suggest that these infiltrating CD4\u0026thinsp;+\u0026thinsp;T cells are not exerting anti-tumor effects but rather constitute part of the immunosuppressive microenvironment. scRNA-seq analysis further established the central role of T cells in immunotherapy response. We observed that patients with a favorable treatment response (MPR) exhibited significantly weakened communication between T cells and stromal cells. This may result from effective immunotherapy disrupting the tumor-stromal barrier, enabling T cells to more efficiently target and kill tumor cells instead of being suppressed within the stromal network. Such alterations in cellular interaction patterns could potentially serve as a novel predictive biomarker for treatment efficacy.\u003c/p\u003e \u003cp\u003eFrom a therapeutic perspective, drugs such as AZD6738 (ATR inhibitor) and AZD7762 (CHK1 inhibitor) showed higher sensitivity in the high-risk group. These compounds target DNA damage response pathways and may be particularly effective in tumors with cell cycle dysregulation (\u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e55\u003c/span\u003e, \u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e56\u003c/span\u003e). Our drug sensitivity analysis provides a rationale for personalized treatment strategies based on risk stratification.\u003c/p\u003e \u003cp\u003eIn conclusion, we successfully constructed and validated a risk model comprising three novel prognostic genes (RRM2, WDR76, and PLEKHH2) through integrated multi-omics data. The established model demonstrated good predictive capability for patient outcomes. Multivariate Cox regression analysis confirmed that both the risk score and N stage were independent prognostic factors. Our analysis identified significant enrichment of the prognostic genes in pathways governing cell cycle progression and DNA replication. Immune infiltration analysis identified a significant correlation between activated memory CD4\u0026thinsp;+\u0026thinsp;T cells and the prognostic genes. Drug sensitivity analysis indicated differential responses to medications such as AZD6738 and AZD7762 between high- and low-risk groups. In conclusion, utilizing a series of bioinformatic approaches, we identified prognostic genes associated with NSCLC outcomes and further investigated their mechanistic roles in the pathological processes of NSCLC. Despite these insights, our study has several limitations. The expression and functional roles of RRM2, WDR76, and PLEKHH2 at the protein level were not validated in clinical samples. Moreover, the mechanisms by which these genes modulate immune cell function and therapy resistance require further experimental investigation. Future studies using in vitro and in vivo models are needed to confirm these findings and explore translational applications.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\" width=\"210\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eNon-small cell lung cancer\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eNSCLC\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003elung adenocarcinoma\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eLUAD\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003elung squamous cell carcinoma\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eLUSC\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eImmune checkpoint inhibitors\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eICIs\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eprogrammed death-1\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ePD-1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eNeuroendocrine differentiation\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eNED\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eSingle-cell RNA sequencing\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003escRNA-seq\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eGene Expression Omnibus\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eGEO\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003emajor pathologic response\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eMPR\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003enon-MPR\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eNMPR\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003edifferentially expressed genes\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eDEGs\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eKyoto Encyclopedia of Genes and Genomes enrichment\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eKEGG\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eGene Ontology\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eGO\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eprotein-protein interaction\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ePPI\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eSearch Tool for the Retrieval of Interacting Genes\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eSTRING\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eproportional hazards\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ePH\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003ereceiver operating characteristic\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eROC\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003earea under the curve\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eAUC\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003egene set enrichment analysis\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eGSEA\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003efold change\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eFC\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003enormalized enrichment score\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eNES\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eFalse Discovery Rate\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eFDR\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eDifferential immune cells\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eDICs\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eHigh risk group\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eHG\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eLow risk group\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eLG\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eimmunophenotype score\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eIPS\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003etumor mutational burden\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eTMB\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eGenomics of Drug Sensitivity in Cancer\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eGDSC\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eSingle cell RNA sequencing\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003escRNA-seq\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eprincipal component analysis\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ePCA\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eprincipal components\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ePCs\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll the samples were collected with informed consent from the patients and ethics committee of the Tianjin Medical University General Hospita（IRB2024-YX-045-01）.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot adapted.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and materials\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe dataset(TCGA-NSCLC) supporting the conclusions of this article is available in the [TCGA] repository, [https://xena.ucsc.edu/]. The datasets (GSE248249 and GSE37745) supporting the conclusions of this article is available in the [GEO] repository, [https://www.ncbi.nlm.nih.gov/geo/].\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting Interest\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that they have no competing interests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis work was supported by Tianjin Municipal Health Commission Scientific Research Project on Integrative Traditional Chinese and Western Medicine(2023150)，\u003c/p\u003e\n\u003cp\u003eNational Natural Science Foundation of China(82303793) \u0026nbsp;and Tianjin Science and Technology Plan Project (24ZXGZSY00010).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor Contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eLT, LMD and PSB\u0026nbsp;contributed equally to this work. They were responsible for conceived the study, designed and performed the majority of the experiments, conducted data analysis.\u0026nbsp;LT, LMD, PSB, RF, GJL, XHS, QJY, LCC\u0026nbsp;contributed to data acquisition, data analysis and assays performance.\u0026nbsp;LJand XS\u0026nbsp;revised the manuscript and finally approved the version to be published. All authors read and approved the final manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgments\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors sincerely thank the technical support of the Department of Lung Cancer Surgery, Tianjin Medical University General Hospital and Nankai University. We are also grateful to Dr. Xu and Dr. Li. We also extend our appreciation to all participants who volunteered for this study.\u003c/p\u003e\n\u003ch2\u003e\u003cstrong\u003eContributor Information\u003c/strong\u003e\u003c/h2\u003e\n\u003cp\u003eSong Xu, E-mail: [email protected]\u003c/p\u003e\n\u003cp\u003eJia Li, E-mail: [email protected]\u003c/p\u003e\n\u003cp\u003eTong Li, E-mail: [email protected]\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eW\u0026eacute;ber A, Morgan E, Vignat J, Laversanne M, Pizzato M, Rumgay H, et al. Lung cancer mortality in the wake of the changing smoking epidemic: a descriptive study of the global burden in 2020 and 2040. BMJ Open. 2023;13(5):e065303.\u003c/li\u003e\n\u003cli\u003eZhou X, You L, Xin Z, Su H, Zhou J, Ma Y. Leveraging circulating microbiome signatures to predict tumor immune microenvironment and prognosis of patients with non-small cell lung cancer. J Transl Med. 2023;21(1):800.\u003c/li\u003e\n\u003cli\u003eZhang R, Zhu G, Li Z, Meng Z, Huang H, Ding C, et al. ITGAL expression in non-small-cell lung cancer tissue and its association with immune infiltrates. Front Immunol. 2024;15:1382231.\u003c/li\u003e\n\u003cli\u003eXiang Y, Liu X, Wang Y, Zheng D, Meng Q, Jiang L, et al. Mechanisms of resistance to targeted therapy and immunotherapy in non-small cell lung cancer: promising strategies to overcoming challenges. Front Immunol. 2024;15:1366260.\u003c/li\u003e\n\u003cli\u003eChen P, Liu Y, Wen Y, Zhou C. Non-small cell lung cancer in China. Cancer Commun (Lond). 2022;42(10):937\u0026ndash;70.\u003c/li\u003e\n\u003cli\u003eWang Y, Wang Y, Wang Y, Zhang Y. Identification of prognostic signature of non-small cell lung cancer based on TCGA methylation data. Sci Rep. 2020;10(1):8575.\u003c/li\u003e\n\u003cli\u003eLi Z, Zhou B, Zhu X, Yang F, Jin K, Dai J, et al. Differentiation-related genes in tumor-associated macrophages as potential prognostic biomarkers in non-small cell lung cancer. Front Immunol. 2023;14:1123840.\u003c/li\u003e\n\u003cli\u003eKang J, Zhang C, Zhong WZ. Neoadjuvant immunotherapy for non-small cell lung cancer: State of the art. Cancer Commun (Lond). 2021;41(4):287\u0026ndash;302.\u003c/li\u003e\n\u003cli\u003eHu J, Zhang L, Xia H, Yan Y, Zhu X, Sun F, et al. Tumor microenvironment remodeling after neoadjuvant immunotherapy in non-small cell lung cancer revealed by single-cell RNA sequencing. Genome Med. 2023;15(1):14.\u003c/li\u003e\n\u003cli\u003eShi L, Lu J, Zhong D, Song M, Liu J, You W, et al. Clinicopathological and predictive value of MAIT cells in non-small cell lung cancer for immunotherapy. J Immunother Cancer. 2023;11(1).\u003c/li\u003e\n\u003cli\u003eOzaki Y, Miura S, Oki R, Morikawa T, Uchino K. Neuroendocrine Neoplasms of the Breast: The Latest WHO Classification and Review of the Literature. Cancers (Basel). 2021;14(1).\u003c/li\u003e\n\u003cli\u003eLiu H, Han Y, Liu Z, Gao L, Yi T, Yu Y, et al. Depiction of neuroendocrine features associated with immunotherapy response using a novel one-class predictor in lung adenocarcinoma. Discov Oncol. 2023;14(1):71.\u003c/li\u003e\n\u003cli\u003eLin S, Dai Y, Han C, Han T, Zhao L, Wu R, et al. Single-cell transcriptomics reveal distinct immune-infiltrating phenotypes and macrophage-tumor interaction axes among different lineages of pituitary neuroendocrine tumors. Genome Med. 2024;16(1):60.\u003c/li\u003e\n\u003cli\u003eWang Z, Ding H, Zou Q. Identifying cell types to interpret scRNA-seq data: how, why and more possibilities. Brief Funct Genomics. 2020;19(4):286\u0026ndash;91.\u003c/li\u003e\n\u003cli\u003eCheng C, Chen W, Jin H, Chen X. A Review of Single-Cell RNA-Seq Annotation, Integration, and Cell-Cell Communication. Cells. 2023;12(15).\u003c/li\u003e\n\u003cli\u003eLi C, Song W, Zhang J, Luo Y. Single-cell transcriptomics reveals heterogeneity in esophageal squamous epithelial cells and constructs models for predicting patient prognosis and immunotherapy. Front Immunol. 2023;14:1322147.\u003c/li\u003e\n\u003cli\u003eZhang T, Zhao F, Lin Y, Liu M, Zhou H, Cui F, et al. Integrated analysis of single-cell and bulk transcriptomics develops a robust neuroendocrine cell-intrinsic signature to predict prostate cancer progression. Theranostics. 2024;14(3):1065\u0026ndash;80.\u003c/li\u003e\n\u003cli\u003eLove MI, Huber W, Anders S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 2014;15(12):550.\u003c/li\u003e\n\u003cli\u003eGustavsson EK, Zhang D, Reynolds RH, Garcia-Ruiz S, Ryten M. ggtranscript: an R package for the visualization and interpretation of transcript isoforms using ggplot2. Bioinformatics. 2022;38(15):3844\u0026ndash;6.\u003c/li\u003e\n\u003cli\u003eGu Z, Eils R, Schlesner M. Complex heatmaps reveal patterns and correlations in multidimensional genomic data. Bioinformatics. 2016;32(18):2847\u0026ndash;9.\u003c/li\u003e\n\u003cli\u003eScalori A, Apale P, Panizzuti F, Mascoli N, Pioltelli P, Pozzi M, et al. Depression during interferon therapy for chronic viral hepatitis: early identification of patients at risk by means of a computerized test. Eur J Gastroenterol Hepatol. 2000;12(5):505\u0026ndash;9.\u003c/li\u003e\n\u003cli\u003eMao W, Ding J, Li Y, Huang R, Wang B. Inhibition of cell survival and invasion by Tanshinone IIA via FTH1: A key therapeutic target and biomarker in head and neck squamous cell carcinoma. Exp Ther Med. 2022;24(2):521.\u003c/li\u003e\n\u003cli\u003eYu G, Wang LG, Han Y, He QY. clusterProfiler: an R package for comparing biological themes among gene clusters. Omics. 2012;16(5):284\u0026ndash;7.\u003c/li\u003e\n\u003cli\u003eShannon P, Markiel A, Ozier O, Baliga NS, Wang JT, Ramage D, et al. Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome Res. 2003;13(11):2498\u0026ndash;504.\u003c/li\u003e\n\u003cli\u003eShen L, Mo J, Yang C, Jiang Y, Ke L, Hou D, et al. SurvivalPath:A R package for conducting personalized survival path mapping based on time-series survival data. PLoS Comput Biol. 2023;19(1):e1010830.\u003c/li\u003e\n\u003cli\u003eJiang H, Li Y, Wang T. Comparison of Billroth I, Billroth II, and Roux-en-Y reconstructions following distal gastrectomy: A systematic review and network meta-analysis. Cir Esp (Engl Ed). 2021;99(6):412\u0026ndash;20.\u003c/li\u003e\n\u003cli\u003eZhao P, Zhen H, Zhao H, Huang Y, Cao B. Identification of hub genes and potential molecular mechanisms related to radiotherapy sensitivity in rectal cancer based on multiple datasets. J Transl Med. 2023;21(1):176.\u003c/li\u003e\n\u003cli\u003eHeagerty PJ, Lumley T, Pepe MS. Time-dependent ROC curves for censored survival data and a diagnostic marker. Biometrics. 2000;56(2):337\u0026ndash;44.\u003c/li\u003e\n\u003cli\u003eLi M, Wei X, Zhang SS, Li S, Chen SH, Shi SJ, et al. Recognition of refractory Mycoplasma pneumoniae pneumonia among Myocoplasma pneumoniae pneumonia in hospitalized children: development and validation of a predictive nomogram model. BMC Pulm Med. 2023;23(1):383.\u003c/li\u003e\n\u003cli\u003eZheng Y, Wen Y, Cao H, Gu Y, Yan L, Wang Y, et al. Global Characterization of Immune Infiltration in Clear Cell Renal Cell Carcinoma. Onco Targets Ther. 2021;14:2085\u0026ndash;100.\u003c/li\u003e\n\u003cli\u003eRobles-Jimenez LE, Aranda-Aguirre E, Castelan-Ortega OA, Shettino-Bermudez BS, Ortiz-Salinas R, Miranda M, et al. Worldwide Traceability of Antibiotic Residues from Livestock in Wastewater and Soil: A Systematic Review. Animals (Basel). 2021;12(1).\u003c/li\u003e\n\u003cli\u003eWang L, Wang D, Yang L, Zeng X, Zhang Q, Liu G, et al. Cuproptosis related genes associated with Jab1 shapes tumor microenvironment and pharmacological profile in nasopharyngeal carcinoma. Front Immunol. 2022;13:989286.\u003c/li\u003e\n\u003cli\u003eChen B, Khodadoust MS, Liu CL, Newman AM, Alizadeh AA. Profiling Tumor Infiltrating Immune Cells with CIBERSORT. Methods Mol Biol. 2018;1711:243\u0026ndash;59.\u003c/li\u003e\n\u003cli\u003eJiang F, Lu DF, Zhan Z, Yuan GQ, Liu GJ, Gu JY, et al. SARS-CoV-2 Pattern Provides a New Scoring System and Predicts the Prognosis and Immune Therapeutic Response in Glioma. Cells. 2022;11(24).\u003c/li\u003e\n\u003cli\u003eWang W, Lu Z, Wang M, Liu Z, Wu B, Yang C, et al. The cuproptosis-related signature associated with the tumor environment and prognosis of patients with glioma. Front Immunol. 2022;13:998236.\u003c/li\u003e\n\u003cli\u003eHao Y, Hao S, Andersen-Nissen E, Mauck WM, 3rd, Zheng S, Butler A, et al. Integrated analysis of multimodal single-cell data. Cell. 2021;184(13):3573\u0026ndash;87.e29.\u003c/li\u003e\n\u003cli\u003eZhang L, Wang X, Wang W, Ning E, Chen L, Li Z, et al. Metabolomic analysis reveals the changing trend and differential markers of volatile and nonvolatile components of Artemisiae argyi with different aging years. Phytochem Anal. 2024;35(6):1286\u0026ndash;93.\u003c/li\u003e\n\u003cli\u003eGrentner A, Ragueneau E, Gong C, Prinz A, Gansberger S, Oyarzun I, et al. ReactomeGSA: new features to simplify public data reuse. Bioinformatics. 2024;40(6).\u003c/li\u003e\n\u003cli\u003eJin S, Plikus MV, Nie Q. CellChat for systematic analysis of cell-cell communication from single-cell transcriptomics. Nat Protoc. 2025;20(1):180\u0026ndash;219.\u003c/li\u003e\n\u003cli\u003eRen C, He D, Wang Q. Identification and Experimental Validation of Tumor Antigens and Hypoxia Subtypes of Osteosarcoma for Potential mRNA Vaccine Development. Curr Med Chem. 2025.\u003c/li\u003e\n\u003cli\u003eCao J, Spielmann M, Qiu X, Huang X, Ibrahim DM, Hill AJ, et al. The single-cell transcriptional landscape of mammalian organogenesis. Nature. 2019;566(7745):496\u0026ndash;502.\u003c/li\u003e\n\u003cli\u003eZhan Y, Jiang L, Jin X, Ying S, Wu Z, Wang L, et al. Inhibiting RRM2 to enhance the anticancer activity of chemotherapy. Biomed Pharmacother. 2021;133:110996.\u003c/li\u003e\n\u003cli\u003eZhou D, Zhai X, Zhang R. Ribonucleotide reductase regulatory subunit M2 (RRM2) as a potential sero-diagnostic biomarker in non-small cell lung cancer. PLoS One. 2023;18(9):e0291461.\u003c/li\u003e\n\u003cli\u003eLi J, Pang J, Liu Y, Zhang J, Zhang C, Shen G, et al. Suppression of RRM2 inhibits cell proliferation, causes cell cycle arrest and promotes the apoptosis of human neuroblastoma cells and in human neuroblastoma RRM2 is suppressed following chemotherapy. Oncol Rep. 2018;40(1):355\u0026ndash;60.\u003c/li\u003e\n\u003cli\u003eHuang L, Zhang C, Jiang A, Lin A, Zhu L, Mou W, et al. T-cell Senescence in the Tumor Microenvironment. Cancer Immunol Res. 2025;13(5):618\u0026ndash;32.\u003c/li\u003e\n\u003cli\u003eWang Y, Chen R, Zhang J, Zeng P. A comprehensive analysis of ribonucleotide reductase subunit M2 for carcinogenesis in pan-cancer. PLoS One. 2024;19(4):e0299949.\u003c/li\u003e\n\u003cli\u003eJeong WJ, Park JC, Kim WS, Ro EJ, Jeon SH, Lee SK, et al. WDR76 is a RAS binding protein that functions as a tumor suppressor via RAS degradation. Nat Commun. 2019;10(1):295.\u003c/li\u003e\n\u003cli\u003eHu Y, Tan X, Zhang L, Zhu X, Wang X. WDR76 regulates 5-fluorouracil sensitivity in colon cancer via HRAS. Discov Oncol. 2023;14(1):45.\u003c/li\u003e\n\u003cli\u003eHuang D, Li Q, Sun X, Sun X, Tang Y, Qu Y, et al. CRL4(DCAF8) dependent opposing stability control over the chromatin remodeler LSH orchestrates epigenetic dynamics in ferroptosis. Cell Death Differ. 2021;28(5):1593\u0026ndash;609.\u003c/li\u003e\n\u003cli\u003eGilmore JM, Sardiu ME, Groppe BD, Thornton JL, Liu X, Dayebgadoh G, et al. WDR76 Co-Localizes with Heterochromatin Related Proteins and Rapidly Responds to DNA Damage. PLoS One. 2016;11(6):e0155492.\u003c/li\u003e\n\u003cli\u003eWang R, Wang S, Li Z, Luo Y, Zhao Y, Han Q, et al. PLEKHH2 binds \u0026beta;-arrestin1 through its FERM domain, activates FAK/PI3K/AKT phosphorylation, and promotes the malignant phenotype of non-small cell lung cancer. Cell Death Dis. 2022;13(10):858.\u003c/li\u003e\n\u003cli\u003eTaricani L, Shanahan F, Malinao MC, Beaumont M, Parry D. A functional approach reveals a genetic and physical interaction between ribonucleotide reductase and CHK1 in mammalian cells. PLoS One. 2014;9(11):e111714.\u003c/li\u003e\n\u003cli\u003eYu J, Ling S, Hong J, Zhang L, Zhou W, Yin L, et al. TP53/mTORC1-mediated bidirectional regulation of PD-L1 modulates immune evasion in hepatocellular carcinoma. J Immunother Cancer. 2023;11(11).\u003c/li\u003e\n\u003cli\u003eDas L, Levine AD. TGF-beta inhibits IL-2 production and promotes cell cycle arrest in TCR-activated effector/memory T cells in the presence of sustained TCR signal transduction. J Immunol. 2008;180(3):1490\u0026ndash;8.\u003c/li\u003e\n\u003cli\u003eHu Y, Zhao Q, Qin Y, Mei S, Wang B, Zhou H, et al. CARD11 signaling regulates CD8(+) T cell tumoricidal function. Nat Immunol. 2025;26(7):1113\u0026ndash;26.\u003c/li\u003e\n\u003cli\u003eChen M, Zhang S, Wang F, He J, Jiang W, Zhang L. DLGAP5 promotes lung adenocarcinoma growth via upregulating PLK1 and serves as a therapeutic target. J Transl Med. 2024;22(1):209.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"biological-procedures-online","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"bpro","sideBox":"Learn more about [Biological Procedures Online](http://biologicalproceduresonline.biomedcentral.com/)","snPcode":"12575","submissionUrl":"https://submission.nature.com/new-submission/12575/3","title":"Biological Procedures Online","twitterHandle":"@MedicalEvidence","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Non-small cell lung cancer, Neuroendocrine resistance, Immunotherapy, Prognostic genes, Risk model","lastPublishedDoi":"10.21203/rs.3.rs-8816614/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8816614/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eBackground:\u003c/strong\u003e The challenge of immunotherapy resistance has become a current research hotspot. Previous studies showed that neuroendocrine differentiation (NED) might contribute significantly to the initiation and development of non-small cell lung cancer (NSCLC). However, the interplay between immunotherapy resistance and NED of NSCLC remains unclear. This article mainly explored the mechanism of NED-related genes (NEDRGs) and immunotherapy resistance related genes in NSCLC.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods:\u003c/strong\u003e The NSCLC and control samples, and the NSCLC samples with PD-1 blockade were selected from the public databases to obtain differentially expressed genes (DEGs). Then, univariate Cox regression analysis and the proportional hazards (PH) assumption test were adopted to obtain prognostic genes based on the DEGs and NEDRGs, and a risk model was built and validated. Then, the nomogram was established. Subsequently, gene set enrichment analysis (GSEA), immune analysis and drug sensitivity were adopted. The key cells were ascertained in a single cell RNA sequencing (scRNA-seq) dataset. And we collected specimens of six patients from Tianjin Medical University General Hospital for validation using Quantitative Real-Time Polymerase Chain Reaction (qRT-PCR) and Immunohistochemistry (IHC).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults:\u003c/strong\u003e A total of 3 prognostic genes including RRM2, WDR76 and PLEKHH2 were obtained and the risk model could predict the survival outcomes of NSCLC patients. The nomogram had good predictive ability of survival rate for NSCLC. the GSEA results revealed that notable pathways in the 2 risk groups included cell cycle, and prognostic genes were all enriched in this pathway. Furthermore, activated memory CD4 T cells might contribute significantly while 109 drugs such as AZD6738 might be the candidate drugs of NSCLC. Epithelials were identified as the key population, wherein the expression of prognostic genes altered significantly across different epithelials developmental stages. PCR and IHC validation confirmed that RRM2 and WDR76 expression increased in cancer versus normal tissues and with advancing stage, while PLEKHH2 showed opposite trends.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusion:\u003c/strong\u003e This study identified 3 prognostic genes (RRM2, WDR76 and PLEKHH2) to build risk model, and the risk model exhibited superior predictive accuracy of NSCLC. These results might provide new ideas for investigating novel therapeutic targets in NSCLC.\u003c/p\u003e","manuscriptTitle":"Exploring the mechanism of prognostic genes associated with neuroendocrine differentiation and immunotherapy resistance in non-small cell lung cancer based on the bulk transcriptome and single cell RNA sequencing","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-02-16 17:17:59","doi":"10.21203/rs.3.rs-8816614/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2026-03-02T22:07:19+00:00","index":"","fulltext":""},{"type":"reviewerAgreed","content":"199581330487698684077610927749630432859","date":"2026-02-14T11:38:44+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-02-12T12:16:16+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"214252670435972779229785955582107951605","date":"2026-02-10T23:48:30+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-02-10T22:30:20+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-02-09T12:29:21+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-02-09T12:27:50+00:00","index":"","fulltext":""},{"type":"submitted","content":"Biological Procedures Online","date":"2026-02-07T15:34:02+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"biological-procedures-online","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"bpro","sideBox":"Learn more about [Biological Procedures Online](http://biologicalproceduresonline.biomedcentral.com/)","snPcode":"12575","submissionUrl":"https://submission.nature.com/new-submission/12575/3","title":"Biological Procedures Online","twitterHandle":"@MedicalEvidence","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"0e69758b-704d-4be3-adbc-ef9dcd76a381","owner":[],"postedDate":"February 16th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-05-04T20:38:13+00:00","versionOfRecord":[],"versionCreatedAt":"2026-02-16 17:17:59","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8816614","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8816614","identity":"rs-8816614","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}