Immune Profiling and Drug Sensitivity in LUAD: Insights from m5C-Related Ferroptosis Gene Analysis

Immune Profiling and Drug Sensitivity in LUAD: Insights from m5C-Related Ferroptosis Gene Analysis | 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 Immune Profiling and Drug Sensitivity in LUAD: Insights from m5C-Related Ferroptosis Gene Analysis Qi Wang, Heng Zou, Danping Liu, Jinlong Zhong This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8096492/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 10 You are reading this latest preprint version Abstract Background Recent research highlights the pivotal role of 5-methylcytosine (m5C) modification and ferroptosis in the progression of various cancers. However, the prognostic value of m5C-related ferroptosis genes in lung adenocarcinoma (LUAD) remains unclear. This study aims to establish a prognostic framework centered on m5C-related ferroptosis genes to improve the accuracy of prognosis prediction in LUAD patients, thereby optimizing targeted therapeutic strategies. Methods The mRNA expression profiles, along with clinicopathological information of LUAD patients were obtained from The Cancer Genome Atlas (TCGA). Differential expression and weighted gene co-expression network analysis (WGCNA) identified prognosis-related modules. Ferroptosis-related genes and m5C regulators were integrated to identify m5C-associated ferroptosis genes. Machine learning and Cox regression were applied to construct a prognostic model. Finally, functional enrichment, immune infiltration, and drug sensitivity analyses were performed. Results Two key gene modules significantly correlated with LUAD prognosis were identified, yielding 29 m5C-related ferroptosis genes. Ten hub genes were selected via machine learning, and the prognostic risk model achieved strong predictive performance. Immune infiltration profiling revealed close associations between the model and tumor microenvironment features. Among the hub genes, SLC2A1, RRM2, and KIF20A were markedly overexpressed in tumor tissues and associated with poor survival. Notably, SLC2A1 expression correlated with enhanced immunotherapy response. Drug sensitivity analysis and molecular docking identified nine small-molecule compounds with favorable binding affinities to SLC2A1. Conclusions This study delineates the critical prognostic significance of m5C-related ferroptosis genes in LUAD and establishes a clinically relevant prognostic model. The identification of candidate SLC2A1 inhibitors offers promising avenues for targeted therapy and personalized treatment strategies in LUAD. Lung adenocarcinoma m5C ferroptosis tumor immunity Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 1 Background Cancer remains a leading global public health challenge and a primary cause of mortality before the age of 70 across most nations. Lung cancer is the most prevalent malignancy at diagnosis, accounting for 11.6% of all cancer cases, and also ranks first in cancer-related mortality at 18.4%[ 1 ]. It is categorized into two major types: small-cell lung cancer (SCLC) and non-small-cell lung cancer (NSCLC); NSCLC further encompasses three subtypes, including lung adenocarcinoma (LUAD), lung squamous cell carcinoma, and large cell carcinoma[ 2 ]. LUAD is the most common subtype of lung cancer, representing approximately 40% of NSCLC cases[ 3 ]. Early-stage LUAD exhibits mild symptoms and lacks robust diagnostic modalities, leading to most patients being diagnosed at an advanced stage with poor prognosis. Despite advances in multimodal therapies (including chemotherapy, radiotherapy, and molecular targeted therapy) for lung cancer, the 5-year survival rate remains only 15%[ 4 , 5 ]. Thus, investigating the pathogenesis of LUAD remains a scientific priority, and identifying key genes that regulate LUAD progression, alongside the development and validation of biomarkers, which is critical for addressing this disease. RNA modifications have become pivotal post-transcriptional regulators governing gene expression programs, with in excess of 170 distinct types identified to date[ 6 ]. These primarily include N6-methyladenosine (m6A), 5-methylcytosine (m5C), and N1-methyladenosine (m1A)[ 7 ]. m5C is a widespread mRNA modification that acts on the untranslated regions of mRNA transcripts, exerting diverse regulatory roles, including facilitating mRNA export, ribosome assembly, and translation[ 8 ]. Importantly, m5C has been linked to multiple cancers, such as gastric cancer[ 9 ], colorectal cancer[ 10 ], bladder cancer[ 11 ], and lung cancer[ 12 , 13 ]. Mechanistically, YBX1 recognizes m5C modifications in CHD3 mRNA and sustains mRNA stability by recruiting the PABPC1 protein, enabling tumor cells to resist platinum-induced apoptotic stress[ 14 ]. In prostate cancer cells, CDK13 interacts with the RNA methyltransferase NSUN5 to catalyze m5C modification of ACC1 mRNA, thereby enhancing its stability and nuclear export[ 15 ]. In lung cancer, hypermethylation of RNA m5C and the NSUN2 gene are significantly associated with intrinsic resistance to EGFR tyrosine kinase inhibitors[ 9 ]. Additionally, NSUN4-mediated m5C modification of circERI3 promotes lung cancer progression by altering mitochondrial energy metabolism[ 16 ]. Ferroptosis is a type of non-apoptotic programmed cell death linked to oxidative damage[ 17 ], characterized by iron-dependent accumulation of lipid peroxides and subsequent plasma membrane impairment[ 3 ]. Genes associated with ferroptosis represent promising therapeutic targets in anticancer drug research and cancer treatment[ 18 ]. Notably, ferroptosis plays a critical role in tumor cell elimination and growth suppression, and has been implicated in cell death processes driving tumors such as breast cancer, NSCLC, and other malignancies[ 19 ]. In recent years, emerging evidence has revealed potential associations between m5C molecules and ferroptosis-related genes during tumor progression, highlighting an increasingly apparent intimate link between ferroptosis and m5C modification. For instance, m5C modification of GCLC reduces lipid peroxidation and confers a resistant phenotype against doxorubicin-induced ferroptosis[ 20 ]. NSUN2 promotes m5C modification of SLC7A11 mRNA, leading to enhanced SLC7A11 mRNA stability and thereby endowing endometrial cancer cells with ferroptosis resistance[ 21 ]. Additionally, m5C modification of MALAT1 facilitates sorafenib resistance in hepatocellular carcinoma by promoting ferroptosis via the ELAVL1/SLC7A11 axis[ 22 ]. An increasing body of research has revealed interactions between ferroptosis and tumor immunity[ 23 ]. Ferroptosis can synergistically enhance the efficacy of immunotherapies, providing a theoretical basis for designing combination therapies to combat cancer[ 24 ]. Ferroptosis inducers not only inhibit tumor cells but also alter the function and activity of immune cells within the tumor microenvironment (TME)[ 25 ]. Furthermore, m5C modification is also capable of regulating the infiltration level of immune cells in the TME, thereby modulating the efficacy of immunotherapies[ 26 , 27 ]. However, there remains a paucity of reports on the crosstalk among ferroptosis, m5C modification, and anti-tumor immunity. Therefore, in-depth investigation into the interactions between ferroptosis, m5C modification, and anti-tumor immunity would facilitate the development of therapeutic strategies targeting LUAD. In the present study, we first performed comprehensive transcriptomic analyses using datasets from TCGA to identify gene modules significantly associated with the prognosis of LUAD. Subsequently, leveraging integrated machine learning algorithm, we constructed a prognostic signature by combining differentially expressed genes (DEGs) related to m5C regulators and ferroptosis, and further investigated their potential role in diagnosis. Furthermore, we explored the relationship between the risk score and key signaling pathways, and evaluated their impact on immune cell infiltration within LUAD. Focusing on the SLC2A1, which is a key m5C/ferroptosis-related gene with significant prognostic implications, and found that SLC2A1 expression was associated with the response of LUAD patients to immunotherapy. Finally, we also identified potential SLC2A1 inhibitors through pharmacogenomic screening. This study aims to fill the existing knowledge gap by synergistically analyzing the mechanisms underlying m5C modification, ferroptosis, and immune infiltration. By constructing a visualized bioinformatics model, the study seeks to uncover potential prognostic biomarkers and novel therapeutic avenues for LUAD, thereby providing a more comprehensive and multi-dimensional perspective for further exploration. 2. Materials and methods 2.1 Data collection Transcriptome profiles and corresponding clinical data of LUAD patients were retrieved from The Cancer Genome Atlas (TCGA) database ( https://portal.gdc.cancer.gov/ ), including 515 tumor samples and 59 matched adjacent normal tissues. After quality control measures, 2 cases with incomplete clinical records were excluded, resulting in 513 eligible LUAD samples for subsequent analyses. To enhance the representativeness of normal tissues, mRNA expression data of 288 healthy lung tissue samples were obtained from the Genotype-Tissue Expression (GTEx) project via the UCSC Xena browser ( https://xenabrowser.net/datapages/ ). For validating the independent dataset, four independent microarray datasets (GSE27262, GSE43458, GSE116959, and GSE229705) were curated from the Gene Expression Omnibus (GEO) database ( http://www.ncbi.nlm.nih.gov/geo/ ). DEGs between TCGA-LUAD tumor and normal tissues were identified using the R package DESeq2, with significant differential expression defined as an absolute log₂ fold change (|log₂FC|) ≥ 1 and an adjusted p-value ≤ 0.05. Parallel DEGs analysis in GEO datasets was performed using the GEO2R web tool, with the same statistical thresholds (|log₂FC| ≥ 1, adjusted p ≤ 0.05) applied to ensure consistency across platforms. 2.2 WGCNA construction and module identification A scale-free co-expression network was constructed from TCGA-LUAD expression profiles integrated with clinical outcomes using the Weighted Gene Co-expression Network Analysis (WGCNA) R package (version 1.72-1). The soft-thresholding parameter (β) was determined as the minimum value enabling the scale-free topology fitting index (R²) to reach ≥ 0.9, which ensures network connectivity follows a biologically meaningful power-law distribution. Module detection was conducted with a minimum module size threshold of 30 genes set to enhance functional interpretability. The association between modules and traits was evaluated by calculating the Pearson correlation coefficient between module eigengenes and clinical parameters (overall survival and status). Modules with the strongest clinical relevance were identified based on statistical significance and association strength. Additionally, Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis was performed on the selected DEGs using the DAVID bioinformatics resource ( https://david.ncifcrf.gov/ ). 2.3 Calculation of signature from machine learning The integrated LUAD RNA-seq dataset was randomly split into a training set and a validation set at a 7:3 ratio. To identify genes most closely associated with prognosis, eight distinct machine learning algorithms, including Random Forest (RF), Support Vector Machine (SVM), Gradient Boosting Machine (GBM), Neural Network (NNET), Least Absolute Shrinkage and Selection Operator (LASSO) regression, K-Nearest Neighbors (KNN), Decision Tree (DT), and Generalized Linear Model (GLM), were applied to the training set. Each algorithm was evaluated via 10-fold cross-validation (CV), with model performance determined by calculating the Area Under the Curve (AUC). All models were tested for overall survival (OS) across 513 LUAD samples. Specifically, GBM, NNET, DT, GLM, and KNN analyses were implemented using the R package "caret"; LASSO regression via "glmnet"; RF analysis via "randomForest"; SVM via "kernlab"; and model interpretation via "DALEX". 2.4 Definition of the m5c-related ferroptosis genes prognostic model To further validate the association between survival time and m5C-related ferroptosis genes, univariate and multivariate Cox regression analyses were performed using the R packages survival (version 2.44.1.1). First, univariate Cox regression was conducted to determine the correlation between the expression levels of m5C-related ferroptosis genes and OS, with a significance threshold set at P < 0.05. The m5C-related ferroptosis genes meeting this criterion were then subjected to multivariate Cox regression analysis, combined with clinical variables including age, gender, pathologic stage, and smoking status. This analysis aimed to evaluate their independent prognostic contribution and hazard ratio (HR), with 95% confidence intervals (CIs) calculated for key m5C-related ferroptosis genes. LUAD samples were stratified into high-risk and low-risk groups based on risk scores derived from the regression results. Time-dependent receiver operating characteristic (ROC) curves were constructed, and the AUC was calculated to assess the performance of the prognostic model. Kaplan-Meier survival analysis was performed to plot the survival time. To facilitate the clinical application of prognostic model, a comprehensive nomogram was developed using the R package "rms", integrating molecular risk scores and established clinical variables (age, gender, pathologic stage, and smoking status). A corresponding calibration plot was also constructed to compare the consistency between predicted and actual survival probabilities. 2.5 Enrichment analysis and Immunological correlation evaluation To identify the enriched pathways correlated with Gene Ontology-Biological Process (GO-BP) and KEGG pathways in high-risk and low-risk groups, Gene Set Variation Analysis (GSVA) was employed. This part of the analysis was mainly implemented using the clusterProfiler R package and the SangerBox online platform ( http://sangerbox.com/index.html ). The proportions of tumor-infiltrating immune cell types within the mixed cell populations in tumor tissues were estimated using three distinct algorithms, namely the ESTIMATE algorithm, single-sample Gene Set Enrichment Analysis (ssGSEA) algorithm, and CIBERSORT deconvolution algorithm, respectively. Additionally, the Tumor Immune Estimation Resource (TIMER) ( http://timer.comp-genomics.org/ ) was utilized to analyze the correlation between gene expression and the infiltration of several immune cell types in LUAD. 2.6 Drug sensitivity analysis and molecular docking The GDSC2 database ( https://www.cancerrxgene.org/ ) was used to detect the association between SLC2A1 and half-maximal inhibitory concentration (IC50) of drugs. In this study, differences in IC50 values for each of the 198 drugs were included, which these drugs were categorized based on gene expression levels, and the analysis was performed using the R package oncoPredict. The online tool "ROC Plotter" ( https://rocplot.com/ ) was utilized to analyze the relationship between SLC2A1 expression and immune checkpoint inhibitors. The online platform "CB-DOCK2" ( https://cadd.labshare.cn/cb-dock2/php/index.php ) was employed to conduct molecular docking and visualization between SLC2A1 and small-molecule drugs. The chemical structures of small-molecule drugs were obtained from PubChem ( https://pubchem.ncbi.nlm.nih.gov/#query ). Additionally, the protein structure of SLC2A1 was predicted using the AlphaFold3 online database ( https://alphafoldserver.com/ ). 2.7 Statistical analysis Data processing, statistical analysis, and graphing were all performed using R software (Version 4.2.1), GraphPad Prism (Version 9.0), and the SangerBox platform. Student’s t-test was used for statistical significance analysis for intergroup comparisons. The Wilcoxon rank-sum test was applied to evaluate the correlation between continuous variables. The Spearman correlation coefficient was used to assess the ordinal correlation between variables. Significance was defined as p < 0.05. (*P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001). 3 Results 3.1 Identification of prognostic-related modules using WGCNA in LUAD RNA sequencing data comprising 59 adjacent normal lung tissues and 523 LUAD tissues from TCGA were retrieved for comprehensive transcriptomic analysis. With the screening criteria of |log₂FC| >1 and adj-p < 0.05, a total of 5573 DEGs were identified, including 3374 upregulated genes and 2199 downregulated genes (Fig. 1 A). Next, WGCNA was performed on all DEGs. 513 LUAD samples with complete clinical annotations were retained after excluding cases with incomplete records. Based on the scale-free network topology with R² = 0.95, the Pearson correlation matrix of genes was converted into a weighted adjacency matrix via power transformation with β = 3 (Fig. 1 B-E). Using the dynamic tree cutting algorithm, all selected genes were clustered based on different metrics of the topological overlap matrix (TOM), and the tree could be divided into 29 modules marked with different colors (Fig. 1 F). According to the similarity between modules shown in the clustering tree, similar modules were merged with a merging threshold of 0.6, and 26 modules were retained and marked with different colors (merge dynamic) (Fig. 1 G). Subsequently, gene co-expression was summarized using eigengenes, and the correlation between each eigengene and clinical features was calculated, including age, sex, smoking status, pathologic stage, primary tumor stage (stage_T), distant metastasis stage (stage_M), regional lymph node stage (stage_N), survival status (status), and survival time (Fig. 1 H). The module-trait relationship plot showed that the co-expressed "purple module" (containing 123 genes) was significantly positively correlated with pathologic stage (R = 0.14, p = 0.002), stage_T (R = 0.18, p = 3e − 05), and status (R = 0.13, p = 0.004), while negatively correlated with survival time (R = -0.09, p = 0.04) (Fig. 1 H). The co-expressed "turquoise module" (containing 1289 genes) was significantly positively correlated with pathologic stage (R = 0.17, p = 1e − 04), stage_T (R = 0.12, p = 0.008), and status (R = 0.16, p = 2e − 04) (Fig. 1 H). Thus, these two modules may be key modules for determining the prognosis and metastasis of LUAD, and the genes within them were extracted for further analysis (Fig. 1 H). KEGG pathway enrichment analyses indicated these genes were significant enrichment in the cell cycle, DNA replication, biosynthesis of antibiotics, fanconi anemia pathway and p53 signaling pathway (Fig. 1 I). 3.2 Exploration of m5C-related ferroptosis gene in prognosis via machine learning in LUAD To identify m5C-related ferroptosis gene in prognosis in LUAD, we conducted bioinformatics data mining. Based on the intersections between 1412 DEGs related to prognosis derived from WGCNA and ferroptosis-related genes yielded in the FerrDb database, 45 ferroptosis-related and prognosis-related DEGs were identified (Fig. 2 A). Meanwhile, we selected 14 m5C regulators, which are writer: NOP2, NSUN2, NSUN3, NSUN4, NSUN5, NSUN6, DNMT1, TRDMT1, DNMT3A, and DNMT3B; reader: ALYREF; eraser: TET2 and YBX1 (Fig. 2 A). Pearson correlation analysis was performed between 14 m5C regulators and the 45 ferroptosis-related and prognosis-related DEGs identified. Under the criteria of |Pearson coefficient| >0.4 and P adj ≤ 0.01, 29 m5C-related ferroptosis genes in prognosis in LUAD were obtained (Fig. 2 B). Subsequently, eight machine learning algorithms were employed to further narrow 29 m5C-related ferroptosis genes. These algorithms included RF, SVM, GBM, NNET, LASSO regression, KNN, DT, and GLM (Fig. 2 C-D). Among the eight evaluated machine learning models, including the RF, SVM, and LASSO exhibited more excellent performance (Fig. 2 E). Thus, we focused on these three machine learning methods for in-depth investigation. In the RF algorithm, we ranked the important of the 29 m5C-related ferroptosis genes, select the top twenty genes as the significant genes, and listed the top 10 m5C-related ferroptosis genes (Fig. 2 F-G). Using the LASSO algorithm, we identified 12 m5C-related ferroptosis genes (Fig. 2 H-I). Additionally, under the SVM algorithm, we identified a total of 28 m5C-related ferroptosis genes (Fig. 2 J-K). 3.3 Development of prognostic model using the m5C-related ferroptosis genes Based on the aforementioned three machine learning algorithms, the overlapping m5C-related ferroptosis genes (SLC2A1, RRM2, ZFP69B, KIF20A, EZH2, STMN1, PARP1, GPX2, SLC16A1, and SLC7A11) were selected as robust prognostic factors (Fig. 3 A). Subsequently, univariate and multivariate Cox regression analyses were performed on 10 key m5C-related ferroptosis genes. The results indicated that SLC2A1, RRM2, KIF20A, GPX2, and SLC7A11 were statistically significant (Fig. 3 B-C). Based on the constructed prognostic risk score model, 513 TCGA-LUAD samples were divided into high-risk and low-risk groups according to the median risk score (Fig. 3 D). Time-dependent ROC curves showed that the risk score model had stable performance, with areas under the curve (AUC) of 0.731, 0.749, and 0.704 for 1-year, 3-year, and 5-year survival rates, respectively (Fig. 3 E). Kaplan-Meier survival curves for OS revealed that the predicted survival time of the high-risk group was significantly shorter than that of the low-risk group (Fig. 3 F). Furthermore, to forecast OS in LUAD patients, a nomogram was developed, incorporating age, gender, pathologic stage, and the expression levels of m5C-related ferroptosis genes (Fig. 3 G). Calibration curves were plotted for 1-year, 3-year, and 5-year survival rates, which confirmed that this nomogram could predict the OS of LUAD patients (Fig. 3 H). To evaluate the distribution of biological processes between the high-risk and low-risk groups, we performed GSVA enrichment analysis. For gene sets defined by GO-BP, the high-risk group was mainly associated with pathways such as NEGATIVE_T_CELL_SELECTION, MAST_CELL_ACTIVATION, and IMMUNOLOGICAL_MEMORY_PROCESS (Fig. 4 A). In contrast, the low-risk group was mainly associated with pathways including DNA_DEPENDENT_DNA_REPLICATION, SISTER_CHROMATID_SEGREGATION, and DNA_REPLICATION (Fig. 4 B). For gene sets defined by KEGG, the high-risk group was primarily correlated with pathways such as ALLOGRAFT_REJECTION, AUTOIMMUNE_THYROID_DISEASE, and INTESTINAL_IMMUNE_NETWORK_FOR_IGA_PRODUCTION (Fig. 4 C). Meanwhile, the low-risk group was primarily correlated with pathways including CELL_CYCLE, DNA_REPLICATION, and OOCYTE_MEIOSIS (Fig. 4 D). Subsequently, we applied several immune infiltration algorithms to estimate the infiltration levels of different types of immune cells in the TME. We used the ESTIMATE algorithm to calculate the distribution of stromal, immune, and estimate scores between patients in the high-risk and low-risk groups. Compared with the high-risk group, the low-risk group had higher immune, stromal, and ESTIMATE scores (Fig. 4 E). The ssGSEA algorithm showed that the high-risk group had a higher level of Th2 cells, while the low-risk group had higher levels of B cells, CD8 T cells, Cytotoxic cells, DC, Eosinophils, iDC, Macrophages, Mast cells, NK cells, T cells, Tcm, Tem, TFH, Th1 cells, and Th17 cells (Fig. 4 F). The CIBERSORT algorithm revealed that the high-risk group had higher levels of resting NK cells, M0 Macrophages, and M1 Macrophages, whereas the low-risk group had higher levels of memory B cells, resting CD4 memory T cells, Monocytes, resting Dendritic cells, and resting Mast cells. These results collectively indicate that the prognostic model is closely associated with immune cell infiltration in the TME (Fig. 4 G). 3.4 Expression levels of m5C-related ferroptosis genes in LUAD Then, we analyzed the strength of correlations between key m5C-related ferroptosis genes and various m5C regulators. The results showed that GPX2 exhibited a correlation solely with the DNMT1 (R = − 0.243, p < 0.001) (Fig. 5 A), while SLC7A11 only correlated with the NOP2 (R = 0.185, p < 0.001) (Fig. 5 B), both exhibiting relatively weak correlations. In contrast, SLC2A1 displayed strong correlations with three m5C regulators: NSUN2 (R = 0.383, p < 0.001), ALYREF (R = 0.509, p < 0.001), and YBX1 (R = 0.470, p < 0.001) (Fig. 5 C). RRM2 correlated with six m5C regulators, namely NOP2 (R = 0.180, p < 0.001), NSUN2 (R = 0.447, p < 0.001), DNMT1 (R = 0.452, p < 0.001), DNMT3B (R = 0.513, p < 0.001), ALYREF (R = 0.688, p < 0.001), and YBX1 (R = 0.517, p < 0.001) (Fig. 5 D). Notably, KIF20A displayed the strongest correlations with m5C regulators, including NOP2 (R = 0.162, p < 0.001), NSUN2 (R = 0.474, p < 0.001), DNMT1 (R = 0.523, p < 0.001), DNMT3A (R = 0.404, p < 0.001), DNMT3B (R = 0.551, p < 0.001), ALYREF (R = 0.655, p < 0.001), and YBX1 (R = 0.496, p < 0.001) (Fig. 5 E). Therefore, we selected these three m5C-related ferroptosis genes (SLC2A1, RRM2, and KIF20A) for further analysis. Subsequently, we analyzed the expression levels of SLC2A1, RRM2, and KIF20A in a larger cohort of LUAD samples from the XENA database ( https://xenabrowser.net/datapages/ ), which included 347 normal tissue samples and 515 tumor tissue samples. The results showed that SLC2A1, RRM2, and KIF20A exhibited significantly higher expression levels in tumor tissues (Fig. 6 A-C). Meanwhile, we incorporated the GEO database for validation, the results demonstrated that SLC2A1, RRM2, and KIF20A consistently maintained high expression levels in tumor tissues across the GSE27262, GSE43458, GSE116959, and GSE229705 datasets (Fig. 6 D-G). Additionally, we showed that high expression of SLC2A1, RRM2, and KIF20A was associated with poor prognosis in LUAD patients (Fig. 6 H-J). We also utilized the Human Protein Atlas ( https://www.proteinatlas.org/ ) to perform immunohistochemical (IHC) staining in patients diagnosed with LUAD. Our analysis revealed that the expression levels of SLC2A1, RRM2, and KIF20A in LUAD tissues were statistically significantly higher compared to normal tissues (Fig. 6 K-M). These results indicated that the expression of SLC2A1, RRM2, and KIF20A is closely associated with the progression of LUAD. 3.5 Identification of potential SLC2A1 inhibitors through pharmacogenetic screening Furthermore, the TIMER database demonstrated that the expression of three m5C-related ferroptosis genes (SLC2A1, RRM2, and KIF20A) was significantly correlated with the infiltration of several immune cell types in LUAD, including B cells, macrophages, neutrophils, and CD4 + T cells. Specifically, the results revealed that KIF20A expression was positively correlated with neutrophils while negatively correlated with B cells and macrophages (Fig. 7 A). RRM2 expression showed a positive correlation with neutrophils and a negative correlation with B cells and CD4 + T cells (Fig. 7 B-C). For SLC2A1, its expression was positively correlated with neutrophils and negatively correlated with B cells (Fig. 7 D). Meanwhile, we analyzed the correlation between the expression of SLC2A1, RRM2, KIF20A and immunotherapy response using ROC Plotter ( https://rocplot.com/ ). Intriguingly, only high expression of SLC2A1 was significantly associated with a higher response rate to anti-PD-1 and anti-PD-L1 therapies (Fig. 7 E-F), and it was also correlated with the overall therapeutic efficacy of immune checkpoint inhibitors (Fig. 7 G). In contrast, RRM2 and KIF20A exhibited no correlation with the therapeutic effect of immune checkpoint inhibitors. To identify potential therapeutic drugs targeting SLC2A1 for LUAD, we screened the Genomics of Drug Sensitivity in Cancer (GDSC) database to select agents that exhibit enhanced efficacy under conditions of high SLC2A1 expression. Drug sensitivity analysis performed using the oncoPredict package identified nine small-molecule drugs, namely BI.2536, Dabrafenib, IWP.2, Leflunomide, LJI308, MN.64, RO.3306, Sinularin, and WEHI.539, with altered potency (Fig. 8 A-I). Namely, LUAD patients with high SLC2A1 expression showed increased sensitivity to these drugs. Furthermore, we predicted the protein structure of SLC2A1 using the AlphaFold3 online platform ( https://golgi.sandbox.google.com/ ) (Fig. 8 J). Subsequently, we employed the CB-DOCK2 online tool to predict the potential binding sites and binding regions of SLC2A1 when interacting with the nine small-molecule compounds (Fig. 8 K-S). The results of this structural prediction analysis indicated that SLC2A1 and the nine small-molecule compounds exhibit favorable binding efficiency. These results indicate that these drugs have the potential to be repurposed as anti-cancer agents targeting SLC2A1 to inhibit tumor progression in LUAD patients. 4 Discussion and Conclusion LUAD remains a leading cause of cancer-related mortality worldwide, with marked heterogeneity in its clinical manifestations, this limits the efficacy of standardized therapies. Although RNA m5C modification, has been demonstrated to be associated with the development and drug resistance of LUAD, and ferroptosis is essential for tumor suppression, their interaction in shaping LUAD prognosis remains unelucidated. Herein, we address this gap using WGCNA and multimodal machine learning to identify m5C-related ferroptosis genes as prognostic biomarkers and therapeutic targets. Concurrently, we constructed a prognostic risk model to screen for genes closely associated with LUAD progression, and validated this prognostic model. Furthermore, we developed a novel survival probability prediction model based on the clinicopathological features of LUAD to support clinicians in designing personalized treatment regimens. Finally, through our multi-level analyses, we identified the key prognostic gene SLC2A1, which links m5C modification, ferroptosis, and immune infiltration to LUAD progression. Among these prognosis-related genes, SLC2A1, RRM2, and KIF20A stand out, given their strong correlations with m5C regulators and ubiquitous high expression in LUAD. Existing studies demonstrate that ALYREF mediates m5C methylation of KIF20A mRNA, stabilizing KIF20A expression to confer ferroptosis resistance in cervical cancer cells[ 28 ], and this is a finding fully consistent with our observation that KIF20A exhibits the strongest correlation with ALYREF (R = 0.655, p < 0.001). Additionally, KIF20A knockdown enhances the synergistic antiproliferative effects of gemcitabine and the ferroptosis inducer on LUAD cells[ 29 ]. Inhibition of KIF20A boosts hepatocellular carcinoma immunotherapy efficacy by promoting c-Myc ubiquitination[ 30 ], while comprehensive bioinformatics analyses have validated KIF20A as a prognostic factor and therapeutic target for LUAD[ 31 ]. Collectively, these data establish KIF20A as a robust candidate for drug development, given its tight associations with ferroptosis, m5C modification, and immunotherapy. For RRM2, no studies linking it to m5C modification have been reported to date. However, as a key ferroptosis regulator, RRM2 is closely associated with the progression of hepatocellular carcinoma[ 32 ], lung cancer[ 33 ], ovarian cancer[ 34 ], and sepsis [ 35 ]. It also correlates strongly with immune infiltration levels across pan-cancer contexts: RRM2 knockdown enhances the antitumor efficiency of PD-1 blockade in renal cell carcinoma[ 36 ], and RRM2 has been characterized as a critical ferroptosis regulator in LUAD, promoting tumor immune infiltration via ferroptosis inhibition[ 37 ]. These results strongly corroborate our analysis, supporting that RRM2 may modulate LUAD progression through mechanisms involving ferroptosis, m5C modification, and immune infiltration. SLC2A1, also known as glucose transporter type 1 (GLUT1), is a facilitative glucose transporter responsible for the continuous uptake and transport of glucose. Studies have demonstrated that SLC2A1 is overexpressed in multiple cancers, including breast cancer, lung cancer, hepatocellular carcinoma, colorectal cancer, and gastric cancer[ 38 ]. SLC2A1 is deeply involved in regulating TME remodeling across different tumors[ 39 , 40 ]; it also serves as a classic ferroptosis-related gene and a diagnostic biomarker for immune infiltration in colorectal cancer[ 41 ]. Additionally, evidence has linked SLC2A1 to the ferroptosis mechanism mediated by m6A modification[ 3 ]. In the present study, we observed that SLC2A1 expression in LUAD was positively correlated with neutrophils and negatively correlated with B cells. Meanwhile, high SLC2A1 expression was significantly associated with a higher response rate to anti-PD-1 and anti-PD-L1 therapies—findings that support its potential as a predictive biomarker for immune checkpoint inhibitors (ICIs). Via pharmacogenomic screening using the GDSC database, we further identified enhanced efficacy of nine small-molecule drugs (e.g., BI.2536, dabrafenib) in LUAD with high SLC2A1 expression, and structural prediction additionally confirmed favorable binding between SLC2A1 and these small-molecule compounds, supporting SLC2A1 as a target for drug repurposing. Collectively, our study further validates SLC2A1 as a suitable molecular target that connects m5C modification, ferroptosis, and immunotherapy. Based on genes mediating the m5C-ferroptosis crosstalk, the present study established a prognostic signature for LUAD, which exhibits high accuracy in OS prediction and TME stratification. Notably, SLC2A1 was identified as a dual-functional biomarker with both prognostic value and predictive utility for ICIs responses, as well as a therapeutic target. Additionally, repurposable small-molecule drugs targeting SLC2A1 were identified. These findings deepen the understanding of m5C-ferroptosis interactions in LUAD and pave the way for precision oncology strategies. Nevertheless, this study has several limitations that remain to be addressed. First, given the reliance on retrospective data, prospective cohort studies are required to validate the utility of the identified biomarker. Second, mechanistic studies are needed to elucidate how m5C regulates the expression or function of SLC2A1, RRM2, and KIF20A, which would extend the current understanding of m5C-mediated post-transcriptional regulation in tumorigenesis. Third, in vitro and in vivo experiments are necessary to validate the efficacy of SLC2A1-targeting drugs and their synergistic effects with ICIs, as preclinical validation is critical for translating molecular findings to clinical applications. Finally, intratumoral heterogeneity may impact the performance of the biomarker, requires further investigation. Declarations Acknowledgements We want to express our gratitude to the Shanghai Public Health Clinical Center Study group and thank the research teams involved in maintaining and sharing this valuable resource. All lab members are acknowledged for stimulating discussions. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Funding This research received no funding. Data Availability Statement The data used in this study is publicly available and can be accessed through the links provided. Author Contributions Jinlong Zhong and Danping Liu designed the study and approved the final version of the submitted article. Qi Wang and Heng Zou developed methodology, performed the experiments and researched the data. Qi Wang analyzed and interpreted the data. Qi Wang, Heng Zou and Danping Liu wrote and edited the manuscript. All authors read and approved the contents of the manuscript and its publication. Ethics approval Not applicable. Consent to participate Not applicable. Consent to publish Not applicable. Clinical Trial Number Not applicable. References Bray F, Ferlay J, Soerjomataram I et al. Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin 2018; 68: 394-424. Thai AA, Solomon BJ, Sequist LV et al. Lung cancer. Lancet 2021; 398: 535-554. Li D, Chen T, Li QG. Identification of a m(6)A-related ferroptosis signature as a potential predictive biomarker for lung adenocarcinoma. BMC Pulm Med 2023; 23: 128. Li Y, Yan B, He S. Advances and challenges in the treatment of lung cancer. Biomed Pharmacother 2023; 169: 115891. Grant C, Hagopian G, Nagasaka M. Neoadjuvant therapy in non-small cell lung cancer. Crit Rev Oncol Hematol 2023; 190: 104080. Li R, Chen H, Li C et al. The prognostic value and immune landscaps of m6A/m5C-related lncRNAs signature in the low grade glioma. BMC Bioinformatics 2023; 24: 274. Ma C, Gu Z, Yang Y. Development of m6A/m5C/m1A regulated lncRNA signature for prognostic prediction, personalized immune intervention and drug selection in LUAD. J Cell Mol Med 2024; 28: e18282. Haruehanroengra P, Zheng YY, Zhou Y et al. RNA modifications and cancer. RNA Biol 2020; 17: 1560-1575. Wang Y, Wei J, Feng L et al. Aberrant m5C hypermethylation mediates intrinsic resistance to gefitinib through NSUN2/YBX1/QSOX1 axis in EGFR-mutant non-small-cell lung cancer. Mol Cancer 2023; 22: 81. Li H, Yu K, Hu H et al. METTL17 coordinates ferroptosis and tumorigenesis by regulating mitochondrial translation in colorectal cancer. Redox Biol 2024; 71: 103087. Yu M, Ni M, Xu F et al. NSUN6-mediated 5-methylcytosine modification of NDRG1 mRNA promotes radioresistance in cervical cancer. Mol Cancer 2024; 23: 139. Yang Y, Fan H, Liu H et al. NOP2 facilitates EZH2-mediated epithelial-mesenchymal transition by enhancing EZH2 mRNA stability via m5C methylation in lung cancer progression. Cell Death Dis 2024; 15: 506. Cai D, Chen X, Xu H et al. m5C-modified circRREB1 promotes lung cancer progression by inducing mitophagy. J Exp Clin Cancer Res 2025; 44: 203. Meng H, Miao H, Zhang Y et al. YBX1 promotes homologous recombination and resistance to platinum-induced stress in ovarian cancer by recognizing m5C modification. Cancer Lett 2024; 597: 217064. Zhang Y, Chen XN, Zhang H et al. CDK13 promotes lipid deposition and prostate cancer progression by stimulating NSUN5-mediated m5C modification of ACC1 mRNA. Cell Death Differ 2023; 30: 2462-2476. Wu J, Zhao Q, Chen S et al. NSUN4-mediated m5C modification of circERI3 promotes lung cancer development by altering mitochondrial energy metabolism. Cancer Lett 2024; 605: 217266. Chen X, Comish PB, Tang D, Kang R. Characteristics and Biomarkers of Ferroptosis. Front Cell Dev Biol 2021; 9: 637162. Ji X, Qian J, Rahman SMJ et al. xCT (SLC7A11)-mediated metabolic reprogramming promotes non-small cell lung cancer progression. Oncogene 2018; 37: 5007-5019. Li M, Shi H, Dong J et al. Mechanisms of Ferroptosis-Related Genes in Gallbladder Cancer Based on Bioinformatics Analysis. Mol Biotechnol 2025; 67: 1814-1825. Niu K, Chen Z, Li M et al. NSUN2 lactylation drives cancer cell resistance to ferroptosis through enhancing GCLC-dependent glutathione synthesis. Redox Biol 2025; 79: 103479. Chen SJ, Zhang J, Zhou T et al. Epigenetically upregulated NSUN2 confers ferroptosis resistance in endometrial cancer via m(5)C modification of SLC7A11 mRNA. Redox Biol 2024; 69: 102975. Shi CJ, Pang FX, Lei YH et al. 5-methylcytosine methylation of MALAT1 promotes resistance to sorafenib in hepatocellular carcinoma through ELAVL1/SLC7A11-mediated ferroptosis. Drug Resist Updat 2025; 78: 101181. Yang Y, Hao L, Guiyang L, Haozhe P. Multifaceted bioinformatic analysis of m6A-related ferroptosis and its link with gene signatures and tumour-infiltrating immune cells in gliomas. J Cell Mol Med 2024; 28: e70060. Liao P, Wang W, Wang W et al. CD8(+) T cells and fatty acids orchestrate tumor ferroptosis and immunity via ACSL4. Cancer Cell 2022; 40: 365-378 e366. Xu H, Ye D, Ren M et al. Ferroptosis in the tumor microenvironment: perspectives for immunotherapy. Trends Mol Med 2021; 27: 856-867. Li Y, Jin H, Li Q et al. The role of RNA methylation in tumor immunity and its potential in immunotherapy. Mol Cancer 2024; 23: 130. Chen Z, Li Q, Lin Y et al. m5C regulator-mediated methylation modification phenotypes characterized by distinct tumor microenvironment immune heterogenicity in colorectal cancer. Sci Rep 2023; 13: 11950. Gao Y, Zou T. Hypoxia Triggers ALYREF-Mediated m(5)C Methylation of KIF20A to Activate KIF20A/BUB1 for Generating Ferroptosis Resistance in Cervical Cancer Cells. Kaohsiung J Med Sci 2025; e70093. He H, Liang L, Huang J et al. KIF20A is associated with clinical prognosis and synergistic effect of gemcitabine combined with ferroptosis inducer in lung adenocarcinoma. Front Pharmacol 2022; 13: 1007429. Chen S, Zhao L, Liu J et al. Inhibition of KIF20A enhances the immunotherapeutic effect of hepatocellular carcinoma by enhancing c-Myc ubiquitination. Cancer Lett 2024; 598: 217105. Sun D, Zhang H, Zhang C, Wang L. An evaluation of KIF20A as a prognostic factor and therapeutic target for lung adenocarcinoma using integrated bioinformatics analysis. Front Bioeng Biotechnol 2022; 10: 993820. Yang Y, Lin J, Guo S et al. RRM2 protects against ferroptosis and is a tumor biomarker for liver cancer. Cancer Cell Int 2020; 20: 587. Tang B, Xu W, Wang Y et al. Identification of critical ferroptosis regulators in lung adenocarcinoma that RRM2 facilitates tumor immune infiltration by inhibiting ferroptotic death. Clin Immunol 2021; 232: 108872. Feng L, Zou X, Tang L et al. CircASH1L inhibits ferroptosis and enhances cisplatin resistance by sponging miR-515-5p to regulate cell cycle-related CDCA7/RRM2 in ovarian cancer cells. Front Pharmacol 2025; 16: 1563869. He S, He Y, Deng L et al. Identification of RRM2 as a key ferroptosis-related gene in sepsis. Inflamm Res 2024; 73: 459-473. Xiong W, Zhang B, Yu H et al. RRM2 Regulates Sensitivity to Sunitinib and PD-1 Blockade in Renal Cancer by Stabilizing ANXA1 and Activating the AKT Pathway. Adv Sci (Weinh) 2021; 8: e2100881. Ma C, Luo H, Cao J et al. Independent prognostic implications of RRM2 in lung adenocarcinoma. J Cancer 2020; 11: 7009-7022. Xi X, Gong X, Liu Y et al. SLC2A1 and MPST as diagnostic and prognostic biomarkers of potential endometrial cancer. Front Immunol 2025; 16: 1575916. Shi Y, Kotchetkov IS, Dobrin A et al. GLUT1 overexpression enhances CAR T cell metabolic fitness and anti-tumor efficacy. Mol Ther 2024; 32: 2393-2405. Yang S, Qian L, Li Z et al. Integrated Multi-Omics Landscape of Liver Metastases. Gastroenterology 2023; 164: 407-423 e417. Liu XS, Yang JW, Zeng J et al. SLC2A1 is a Diagnostic Biomarker Involved in Immune Infiltration of Colorectal Cancer and Associated With m6A Modification and ceRNA. Front Cell Dev Biol 2022; 10: 853596. Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Review Version 1 posted Editorial decision: Revision requested 05 Jan, 2026 Reviews received at journal 04 Jan, 2026 Reviewers agreed at journal 26 Dec, 2025 Reviews received at journal 25 Dec, 2025 Reviewers agreed at journal 23 Dec, 2025 Reviewers invited by journal 04 Dec, 2025 Editor invited by journal 01 Dec, 2025 Editor assigned by journal 13 Nov, 2025 Submission checks completed at journal 13 Nov, 2025 First submitted to journal 12 Nov, 2025 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-8096492","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":555513575,"identity":"eff548a3-a7a4-4951-8320-3c8afc763639","order_by":0,"name":"Qi Wang","email":"","orcid":"","institution":"Jinshan Hospital of Fudan University","correspondingAuthor":false,"prefix":"","firstName":"Qi","middleName":"","lastName":"Wang","suffix":""},{"id":555513576,"identity":"0750da6f-f812-48d9-baa0-062d46b637fa","order_by":1,"name":"Heng Zou","email":"","orcid":"","institution":"School of Life Sciences, Shanghai University","correspondingAuthor":false,"prefix":"","firstName":"Heng","middleName":"","lastName":"Zou","suffix":""},{"id":555513577,"identity":"76fb7f9e-f7de-4cb3-aeb0-f4514a6a0f6c","order_by":2,"name":"Danping Liu","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA/ElEQVRIiWNgGAWjYDCCA2CSGYgZGxgYCiQY+JmZDz4goAWkFKbFQIJBsp0t2YBILSAAVGtwnsdMAJ8OvuPNzx983GMtby7d3Pbwi4FF4ubDDGYMDDU20bi0SJ45Ztg441m64c45B9uNZQwkErcdZkh7wHAsLbcBhxaDGzmMzTwHDjNuuJHYJi0B0XLcgLHhMEEt9nAtm5sZ2ySI0ZII0iL5AahlAzMzG14tIL/MnHEgPXnDnYNt0sBANp5xmI3ZIAGPX4Ah9uDDhwPWthtutz+T/FFRJ9vff/7jgw81Nji1IIAEMHZ4YJwEgsqhWhh/EKVyFIyCUTAKRhoAACQ+YkmICo/QAAAAAElFTkSuQmCC","orcid":"","institution":"Shanghai Public Health Clinical Center, Fudan University","correspondingAuthor":true,"prefix":"","firstName":"Danping","middleName":"","lastName":"Liu","suffix":""},{"id":555513579,"identity":"dc711c8f-dd87-4fa2-be12-e758c18aaae3","order_by":3,"name":"Jinlong Zhong","email":"","orcid":"","institution":"Jinshan Hospital of Fudan University","correspondingAuthor":false,"prefix":"","firstName":"Jinlong","middleName":"","lastName":"Zhong","suffix":""}],"badges":[],"createdAt":"2025-11-12 12:53:24","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8096492/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8096492/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":97697291,"identity":"9c5a0c27-b166-4161-b790-88e41a85d1de","added_by":"auto","created_at":"2025-12-08 11:42:43","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":109611,"visible":true,"origin":"","legend":"","description":"","filename":"RevisedManuscript.docx","url":"https://assets-eu.researchsquare.com/files/rs-8096492/v1/fd935d7bdb5c28554600dd03.docx"},{"id":97697275,"identity":"4cd7a200-6481-4ac1-9bbf-5d8f7608ef7a","added_by":"auto","created_at":"2025-12-08 11:42:43","extension":"tif","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1702904,"visible":true,"origin":"","legend":"","description":"","filename":"Fig1.tif","url":"https://assets-eu.researchsquare.com/files/rs-8096492/v1/212dc52e3f3d67cdd9a5aa3b.tif"},{"id":97697279,"identity":"78dd3edb-7da0-432e-b883-ea95b5b5da9c","added_by":"auto","created_at":"2025-12-08 11:42:43","extension":"tif","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1815328,"visible":true,"origin":"","legend":"","description":"","filename":"Fig2.tif","url":"https://assets-eu.researchsquare.com/files/rs-8096492/v1/15dd0ccce297d4c6858633d0.tif"},{"id":97697309,"identity":"3a9b56b8-2ee8-4197-adb4-8db31e8cbdaf","added_by":"auto","created_at":"2025-12-08 11:42:46","extension":"tif","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":2060494,"visible":true,"origin":"","legend":"","description":"","filename":"fig3.tif","url":"https://assets-eu.researchsquare.com/files/rs-8096492/v1/1726872d98b8cfb2b1ab1f34.tif"},{"id":97894305,"identity":"6a63b5fa-25c6-4b03-986e-bde451cae110","added_by":"auto","created_at":"2025-12-10 15:32:16","extension":"tif","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":2196126,"visible":true,"origin":"","legend":"","description":"","filename":"fig4.tif","url":"https://assets-eu.researchsquare.com/files/rs-8096492/v1/71a55eb11e7d4b4ca2c96e44.tif"},{"id":97697293,"identity":"ea156285-4a75-44ca-a260-59c0c4bb4b74","added_by":"auto","created_at":"2025-12-08 11:42:43","extension":"tif","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":2774914,"visible":true,"origin":"","legend":"","description":"","filename":"fig5.tif","url":"https://assets-eu.researchsquare.com/files/rs-8096492/v1/cead3fa5e6446ab3eead92a2.tif"},{"id":97894448,"identity":"1a9af2e4-a94a-45ab-86d5-2664d2be6d42","added_by":"auto","created_at":"2025-12-10 15:32:31","extension":"tif","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":4388580,"visible":true,"origin":"","legend":"","description":"","filename":"fig6.tif","url":"https://assets-eu.researchsquare.com/files/rs-8096492/v1/84ffc9bf62f5ab23e73934d8.tif"},{"id":97697281,"identity":"33959ae9-c453-4628-b4c2-c6ea16f3f38c","added_by":"auto","created_at":"2025-12-08 11:42:43","extension":"tif","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":2104666,"visible":true,"origin":"","legend":"","description":"","filename":"fig7.tif","url":"https://assets-eu.researchsquare.com/files/rs-8096492/v1/0c22d196a86588f27a15aa5f.tif"},{"id":97697295,"identity":"f50576cb-ca5a-4539-bbea-f9596fa74469","added_by":"auto","created_at":"2025-12-08 11:42:43","extension":"tif","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":3053126,"visible":true,"origin":"","legend":"","description":"","filename":"fig8.tif","url":"https://assets-eu.researchsquare.com/files/rs-8096492/v1/87454a8f7ed6282abd607326.tif"},{"id":97697284,"identity":"d0882d72-f419-4c22-85d0-35a0983c9b74","added_by":"auto","created_at":"2025-12-08 11:42:43","extension":"json","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":6625,"visible":true,"origin":"","legend":"","description":"","filename":"44f625345f1b4db98d1a9d9c52c208f8.json","url":"https://assets-eu.researchsquare.com/files/rs-8096492/v1/28b60f9d26a57cf19fc061e4.json"},{"id":97697280,"identity":"dac2f73e-05e7-4831-96c1-90b84cd7b7c2","added_by":"auto","created_at":"2025-12-08 11:42:43","extension":"xml","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":102068,"visible":true,"origin":"","legend":"","description":"","filename":"44f625345f1b4db98d1a9d9c52c208f81enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-8096492/v1/f4878e07c6353858af04d8a8.xml"},{"id":97893233,"identity":"6242ab3a-c658-446d-89ad-c483bd0a8897","added_by":"auto","created_at":"2025-12-10 15:29:03","extension":"tif","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1702904,"visible":true,"origin":"","legend":"","description":"","filename":"Fig1.tif","url":"https://assets-eu.researchsquare.com/files/rs-8096492/v1/9afa487aab67abb1d2fa3bcb.tif"},{"id":97892737,"identity":"21bd0fe2-40b6-4236-b6dc-f7257a7e9208","added_by":"auto","created_at":"2025-12-10 15:20:05","extension":"tif","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1815328,"visible":true,"origin":"","legend":"","description":"","filename":"Fig2.tif","url":"https://assets-eu.researchsquare.com/files/rs-8096492/v1/3cb95c1cd4c3da0a3a5b1607.tif"},{"id":97893217,"identity":"5313fe1d-ce78-45de-bc57-42445c6b596f","added_by":"auto","created_at":"2025-12-10 15:28:44","extension":"tif","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":2060494,"visible":true,"origin":"","legend":"","description":"","filename":"fig3.tif","url":"https://assets-eu.researchsquare.com/files/rs-8096492/v1/e9dabb6daac1d66c93e1eec9.tif"},{"id":97697301,"identity":"d64e47bd-3b02-46b0-8b92-f3eb4458bf69","added_by":"auto","created_at":"2025-12-08 11:42:43","extension":"tif","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":2196126,"visible":true,"origin":"","legend":"","description":"","filename":"fig4.tif","url":"https://assets-eu.researchsquare.com/files/rs-8096492/v1/a945b693f689106ffda89bed.tif"},{"id":97892936,"identity":"ef0d7bc9-962f-431b-928d-1b1dc812c4e9","added_by":"auto","created_at":"2025-12-10 15:24:20","extension":"tif","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":2774914,"visible":true,"origin":"","legend":"","description":"","filename":"fig5.tif","url":"https://assets-eu.researchsquare.com/files/rs-8096492/v1/4ffa74521b76d52cb6f9f403.tif"},{"id":97697304,"identity":"f203be7a-ddd1-4adf-80ea-ae9d41f1a72b","added_by":"auto","created_at":"2025-12-08 11:42:43","extension":"tif","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":4388580,"visible":true,"origin":"","legend":"","description":"","filename":"fig6.tif","url":"https://assets-eu.researchsquare.com/files/rs-8096492/v1/478c7d6b0636a15e8bdaf4a0.tif"},{"id":97697286,"identity":"014dfe0d-7994-45c6-8117-04629db01a01","added_by":"auto","created_at":"2025-12-08 11:42:43","extension":"tif","order_by":17,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":2104666,"visible":true,"origin":"","legend":"","description":"","filename":"fig7.tif","url":"https://assets-eu.researchsquare.com/files/rs-8096492/v1/2441c76d99eff8d50aa4786a.tif"},{"id":97697297,"identity":"ea0027da-82a1-48fc-b054-0e398712fbf3","added_by":"auto","created_at":"2025-12-08 11:42:43","extension":"tif","order_by":18,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":3053126,"visible":true,"origin":"","legend":"","description":"","filename":"fig8.tif","url":"https://assets-eu.researchsquare.com/files/rs-8096492/v1/a4fbea9170520e56efa8c972.tif"},{"id":97697288,"identity":"fc013bca-0bf3-4055-ada7-cdf3cd02139c","added_by":"auto","created_at":"2025-12-08 11:42:43","extension":"png","order_by":19,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":169958,"visible":true,"origin":"","legend":"","description":"","filename":"OnlineFig1.png","url":"https://assets-eu.researchsquare.com/files/rs-8096492/v1/1d4796d4e22baafe820ea351.png"},{"id":97893185,"identity":"b9c421c4-96e5-4181-8736-142ca11cd9bb","added_by":"auto","created_at":"2025-12-10 15:28:27","extension":"png","order_by":20,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":160169,"visible":true,"origin":"","legend":"","description":"","filename":"OnlineFig2.png","url":"https://assets-eu.researchsquare.com/files/rs-8096492/v1/7f70b09ad77f9e68554d912e.png"},{"id":97893174,"identity":"0a762c6f-eb31-4d5b-96b0-d46d2e5a0682","added_by":"auto","created_at":"2025-12-10 15:28:18","extension":"png","order_by":21,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":183078,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefig3.png","url":"https://assets-eu.researchsquare.com/files/rs-8096492/v1/65a7714fdba3a2d90d4acdb7.png"},{"id":97697283,"identity":"e826858e-9e4c-45c2-ab11-46ac06ab1e9c","added_by":"auto","created_at":"2025-12-08 11:42:43","extension":"png","order_by":22,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":216523,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefig4.png","url":"https://assets-eu.researchsquare.com/files/rs-8096492/v1/2eaf5543e6ee2fbc2f09d637.png"},{"id":97697308,"identity":"5367423e-e309-46e2-9727-992b5a91de00","added_by":"auto","created_at":"2025-12-08 11:42:45","extension":"png","order_by":23,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":480948,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefig5.png","url":"https://assets-eu.researchsquare.com/files/rs-8096492/v1/1be10ccc78dd872ca787425e.png"},{"id":97893205,"identity":"d5729a1a-3c1f-4574-b8f3-7987726e2e3e","added_by":"auto","created_at":"2025-12-10 15:28:37","extension":"png","order_by":24,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":562858,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefig6.png","url":"https://assets-eu.researchsquare.com/files/rs-8096492/v1/76744842ae1484858029259d.png"},{"id":97697292,"identity":"4492c883-f010-4359-8252-f276fde465e2","added_by":"auto","created_at":"2025-12-08 11:42:43","extension":"png","order_by":25,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":192962,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefig7.png","url":"https://assets-eu.researchsquare.com/files/rs-8096492/v1/414bb2434778dd7b7b15c56a.png"},{"id":97893443,"identity":"c9f228e2-3560-40e3-9d64-8eee9819ca0c","added_by":"auto","created_at":"2025-12-10 15:30:27","extension":"png","order_by":26,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":385936,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefig8.png","url":"https://assets-eu.researchsquare.com/files/rs-8096492/v1/ba89335459703c6f683168f7.png"},{"id":97893159,"identity":"a606d7eb-6329-4df3-915e-7286a1f33139","added_by":"auto","created_at":"2025-12-10 15:28:03","extension":"xml","order_by":27,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":102900,"visible":true,"origin":"","legend":"","description":"","filename":"44f625345f1b4db98d1a9d9c52c208f81structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-8096492/v1/dd366eee44b3d9efc7c19552.xml"},{"id":97892743,"identity":"39726044-b5cf-49ed-90ce-926aea337f79","added_by":"auto","created_at":"2025-12-10 15:20:10","extension":"html","order_by":28,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":111004,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-8096492/v1/bea6ac267c31d4037c0df4f6.html"},{"id":97697273,"identity":"dfd41642-cbc0-4068-beaf-52d8d79ec7f2","added_by":"auto","created_at":"2025-12-08 11:42:42","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":1625823,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eIdentification of prognostic-related modules using WGCNA in LUAD. (A) \u003c/strong\u003eVolcano plots for DEGs in LUAD. \u003cstrong\u003e(B-E)\u003c/strong\u003e This section describes related parameters of WGCNA analysis.\u003cstrong\u003e (F) \u003c/strong\u003eNetwork heatmap of the TOM from WGCNA. \u003cstrong\u003e(G) \u003c/strong\u003eGene dendrograms with varying similarity based on topological overlap, along with the specified module colors. \u003cstrong\u003e(H) \u003c/strong\u003eWGCNA identified co-expression gene modules associated with prognosis. \u003cstrong\u003e(I) \u003c/strong\u003eKEGG enrichment analyses for prognostic related DEGs.\u003c/p\u003e","description":"","filename":"Fig1.png","url":"https://assets-eu.researchsquare.com/files/rs-8096492/v1/0c1f1210b4e6e7444d7c6422.png"},{"id":97697274,"identity":"bd785d3f-f4ab-48f7-b531-03e1791696ad","added_by":"auto","created_at":"2025-12-08 11:42:42","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":1655208,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eExploration of ferroptosis‑and m5C‑related prognostic genes via machine learning in LUAD. (A) \u003c/strong\u003eVenn diagrams showing numbers of DEGs between prognostic related DEGs and ferroptosis-related genes and genes belonging to m5C regulators. \u003cstrong\u003e(B)\u003c/strong\u003e Heatmap plots of the correlations of the 14 m5C molecules with the 29 prognostic m5C‑related ferroptosis genes.\u003cstrong\u003e (C) \u003c/strong\u003eReverse cumulative distribution of the residual in the 8 ML methods. \u003cstrong\u003e(D) \u003c/strong\u003eBoxplots displaying the residuals of the sample in the 8 ML methods, with the red dot indicating the root mean square of the residuals. \u003cstrong\u003e(E) \u003c/strong\u003eROC evaluation of the performance of the 8 ML methods. \u003cstrong\u003e(F-G) \u003c/strong\u003eThe correlation between the total number of trees in the RF and the error rate is assessed, with ranking based on the relative importance of the genes.\u003cstrong\u003e (H-I) \u003c/strong\u003eConstruction of the LASSO regression model, cross-validation for parameter optimization in this model. Each curve represents a single gene. \u003cstrong\u003e(J-K) \u003c/strong\u003eFeature gene selection method based on SVM-RFE.\u003c/p\u003e","description":"","filename":"Fig2.png","url":"https://assets-eu.researchsquare.com/files/rs-8096492/v1/44be95a93991ddc49387f647.png"},{"id":97894360,"identity":"dc23a706-adc9-4eb5-92ab-ca28814ae473","added_by":"auto","created_at":"2025-12-10 15:32:23","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":1823824,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eDevelopment of prognostic model using the m5C-related ferroptosis genes. (A) \u003c/strong\u003eVenn diagram showing the m5C-related ferroptosis genes that are common to RF, SVM-RFE, and LASSO. \u003cstrong\u003e(B)\u003c/strong\u003e Forest map of univariate cox regression analysis showed HR of the ten selected m5C-related ferroptosis genes.\u003cstrong\u003e (C) \u003c/strong\u003eForest map of multivariate cox regression analyses to assess the prognostic value of m5C-related ferroptosis genes in relation to other clinical features. \u003cstrong\u003e(D) \u003c/strong\u003eThe scatter plot illustrates the distribution of risk scores and survival status. \u003cstrong\u003e(E) \u003c/strong\u003eThe ROC curves over time predict the probability of patient mortality at 1, 3 and 5 years. \u003cstrong\u003e(F) \u003c/strong\u003eThe Kaplan-Meier survival curve is based on patient risk scores and survival status.\u003cstrong\u003e (G) \u003c/strong\u003eConstruct a nomogram that integrates risk scores with clinical information. \u003cstrong\u003e(H) \u003c/strong\u003eThe calibration curves for 1-year, 3-year, and 5-year survival were used to evaluate the predictive robustness of the model.\u003c/p\u003e","description":"","filename":"fig3.png","url":"https://assets-eu.researchsquare.com/files/rs-8096492/v1/2a0ff39d4a6d07733efcc33c.png"},{"id":97697276,"identity":"1b7a6bb5-e45f-4d22-a56e-a2b49f6768d8","added_by":"auto","created_at":"2025-12-08 11:42:43","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":2180411,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eDevelopment of prognostic model using the m5C-related ferroptosis genes. \u003c/strong\u003eGSVA analysis showed relationship between GO-BP pathways (A-B), KEGG pathways (C-D) and the prognostic model.\u003cstrong\u003e \u003c/strong\u003eComparisons of infiltration levels of immune cells between high-risk and low-risk groups with the ESTIMATE (E), ssGSEA (F), and CIBERSORT (G) algorithm.\u003c/p\u003e","description":"","filename":"fig4.png","url":"https://assets-eu.researchsquare.com/files/rs-8096492/v1/4ca2a0864b0c7e2ce05e3214.png"},{"id":97894050,"identity":"e693e6bb-08a2-4d0c-87ed-32db43b314f7","added_by":"auto","created_at":"2025-12-10 15:31:52","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":4817554,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eCorrelations between key m5C-related ferroptosis genes and m5C regulators in the TCGA-LUAD cohort. \u003c/strong\u003e(A) the correlation between GPX2 and DNMT1 expression level. (B) the correlation between SLC7A11 and NOP2 expression level. (C) the correlation between SLC2A1 and NSUN2, ALYREF, YBX1 expression level. (D) the correlation between RRM2 and NOP2, NSUN2, DNMT1, DNMT3B, ALYREF, YBX1 expression level. (E) the correlation between KIF20A and NOP2, NSUN2, DNMT1, DNMT3A, DNMT3B, ALYREF, YBX1 expression level.\u003c/p\u003e","description":"","filename":"fig5.png","url":"https://assets-eu.researchsquare.com/files/rs-8096492/v1/9b9ab20e2339a15a3145e1b5.png"},{"id":97893982,"identity":"50e5b295-e7a4-4411-a44c-ec0850df1dea","added_by":"auto","created_at":"2025-12-10 15:31:45","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":7076671,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eExpression levels of key m5C-related ferroptosis genes in the LUAD.\u003c/strong\u003e The mRNA expression level of KIF20A\u003cstrong\u003e \u003c/strong\u003e(A), RRM2\u003cstrong\u003e \u003c/strong\u003e(B) and SLC2A1\u003cstrong\u003e \u003c/strong\u003e(C) in the TCGA-LUAD cohort. The KIF20A, RRM2 and SLC2A1 mRNA expression level in the GSE27262 (D), GSE43458 (E), GSE116959 (F), and GSE229705 (G) datasets. Kaplan−Meier curves for the OS of KIF20A (H), RRM2 (I) and SLC2A1 (J). The representative IHC staining of KIF20A (K), RRM2 (L) and SLC2A1 (M) on patients diagnosed with LUAD.\u003c/p\u003e","description":"","filename":"fig6.png","url":"https://assets-eu.researchsquare.com/files/rs-8096492/v1/175936f0e91fd8c132ba66fe.png"},{"id":97697307,"identity":"575f8f75-9067-4457-8de1-425fd46609c3","added_by":"auto","created_at":"2025-12-08 11:42:44","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":2215776,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eCorrelation analysis of key m5C-related ferroptosis genes expression and tumor-infiltrating immune cells.\u003c/strong\u003e (A) Correlation analysis of KIF20A expression and B Cell, macrophage and neutrophil. (B-C) Correlation analysis of RRM2 expression and B Cell, CD4+ T Cell and Neutrophil. (D) Correlation analysis of SLC2A1 expression and B Cell and Neutrophil. (E-G) The correlation between SLC2A1 expression and immunotherapy drug sensitivity.\u003c/p\u003e","description":"","filename":"fig7.png","url":"https://assets-eu.researchsquare.com/files/rs-8096492/v1/786a2ba964e593e537c85d84.png"},{"id":97697306,"identity":"84a31a8e-230d-442b-88c9-b3b66c28dfb9","added_by":"auto","created_at":"2025-12-08 11:42:43","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":4064569,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eIdentification of potential SLC2A1 inhibitors through pharmacogenetic screening.\u003c/strong\u003e (A-I) Depict the sensitivity for nine drugs, namely BI.2536, Dabrafenib, IWP.2, Leflunomide, LJI308, MN.64, RO.3306, Sinularin, and WEHI.539, that showed altered potency. (J) The protein structure of SLC2A1 using the AlphaFold3 online platform. (K-S) The potential binding sites and binding regions of SLC2A1 when interacting with the nine small-molecule compounds.\u003c/p\u003e","description":"","filename":"fig8.png","url":"https://assets-eu.researchsquare.com/files/rs-8096492/v1/16017d9ca094fc0c61e2ed50.png"},{"id":97902480,"identity":"9f0de68c-4402-4bef-adea-75926508e020","added_by":"auto","created_at":"2025-12-10 15:52:28","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":26742004,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8096492/v1/1cb3fc84-b45f-47d6-a5bb-80f04c420b6c.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Immune Profiling and Drug Sensitivity in LUAD: Insights from m5C-Related Ferroptosis Gene Analysis","fulltext":[{"header":"1 Background","content":"\u003cp\u003eCancer remains a leading global public health challenge and a primary cause of mortality before the age of 70 across most nations. Lung cancer is the most prevalent malignancy at diagnosis, accounting for 11.6% of all cancer cases, and also ranks first in cancer-related mortality at 18.4%[\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]. It is categorized into two major types: small-cell lung cancer (SCLC) and non-small-cell lung cancer (NSCLC); NSCLC further encompasses three subtypes, including lung adenocarcinoma (LUAD), lung squamous cell carcinoma, and large cell carcinoma[\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. LUAD is the most common subtype of lung cancer, representing approximately 40% of NSCLC cases[\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. Early-stage LUAD exhibits mild symptoms and lacks robust diagnostic modalities, leading to most patients being diagnosed at an advanced stage with poor prognosis. Despite advances in multimodal therapies (including chemotherapy, radiotherapy, and molecular targeted therapy) for lung cancer, the 5-year survival rate remains only 15%[\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]. Thus, investigating the pathogenesis of LUAD remains a scientific priority, and identifying key genes that regulate LUAD progression, alongside the development and validation of biomarkers, which is critical for addressing this disease.\u003c/p\u003e\u003cp\u003eRNA modifications have become pivotal post-transcriptional regulators governing gene expression programs, with in excess of 170 distinct types identified to date[\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]. These primarily include N6-methyladenosine (m6A), 5-methylcytosine (m5C), and N1-methyladenosine (m1A)[\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e]. m5C is a widespread mRNA modification that acts on the untranslated regions of mRNA transcripts, exerting diverse regulatory roles, including facilitating mRNA export, ribosome assembly, and translation[\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]. Importantly, m5C has been linked to multiple cancers, such as gastric cancer[\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e], colorectal cancer[\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e], bladder cancer[\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e], and lung cancer[\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e, \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]. Mechanistically, YBX1 recognizes m5C modifications in CHD3 mRNA and sustains mRNA stability by recruiting the PABPC1 protein, enabling tumor cells to resist platinum-induced apoptotic stress[\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]. In prostate cancer cells, CDK13 interacts with the RNA methyltransferase NSUN5 to catalyze m5C modification of ACC1 mRNA, thereby enhancing its stability and nuclear export[\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e]. In lung cancer, hypermethylation of RNA m5C and the NSUN2 gene are significantly associated with intrinsic resistance to EGFR tyrosine kinase inhibitors[\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]. Additionally, NSUN4-mediated m5C modification of circERI3 promotes lung cancer progression by altering mitochondrial energy metabolism[\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eFerroptosis is a type of non-apoptotic programmed cell death linked to oxidative damage[\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e], characterized by iron-dependent accumulation of lipid peroxides and subsequent plasma membrane impairment[\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. Genes associated with ferroptosis represent promising therapeutic targets in anticancer drug research and cancer treatment[\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]. Notably, ferroptosis plays a critical role in tumor cell elimination and growth suppression, and has been implicated in cell death processes driving tumors such as breast cancer, NSCLC, and other malignancies[\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e]. In recent years, emerging evidence has revealed potential associations between m5C molecules and ferroptosis-related genes during tumor progression, highlighting an increasingly apparent intimate link between ferroptosis and m5C modification. For instance, m5C modification of GCLC reduces lipid peroxidation and confers a resistant phenotype against doxorubicin-induced ferroptosis[\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]. NSUN2 promotes m5C modification of SLC7A11 mRNA, leading to enhanced SLC7A11 mRNA stability and thereby endowing endometrial cancer cells with ferroptosis resistance[\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]. Additionally, m5C modification of MALAT1 facilitates sorafenib resistance in hepatocellular carcinoma by promoting ferroptosis via the ELAVL1/SLC7A11 axis[\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eAn increasing body of research has revealed interactions between ferroptosis and tumor immunity[\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]. Ferroptosis can synergistically enhance the efficacy of immunotherapies, providing a theoretical basis for designing combination therapies to combat cancer[\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e]. Ferroptosis inducers not only inhibit tumor cells but also alter the function and activity of immune cells within the tumor microenvironment (TME)[\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e]. Furthermore, m5C modification is also capable of regulating the infiltration level of immune cells in the TME, thereby modulating the efficacy of immunotherapies[\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e]. However, there remains a paucity of reports on the crosstalk among ferroptosis, m5C modification, and anti-tumor immunity. Therefore, in-depth investigation into the interactions between ferroptosis, m5C modification, and anti-tumor immunity would facilitate the development of therapeutic strategies targeting LUAD.\u003c/p\u003e\u003cp\u003eIn the present study, we first performed comprehensive transcriptomic analyses using datasets from TCGA to identify gene modules significantly associated with the prognosis of LUAD. Subsequently, leveraging integrated machine learning algorithm, we constructed a prognostic signature by combining differentially expressed genes (DEGs) related to m5C regulators and ferroptosis, and further investigated their potential role in diagnosis. Furthermore, we explored the relationship between the risk score and key signaling pathways, and evaluated their impact on immune cell infiltration within LUAD. Focusing on the SLC2A1, which is a key m5C/ferroptosis-related gene with significant prognostic implications, and found that SLC2A1 expression was associated with the response of LUAD patients to immunotherapy. Finally, we also identified potential SLC2A1 inhibitors through pharmacogenomic screening. This study aims to fill the existing knowledge gap by synergistically analyzing the mechanisms underlying m5C modification, ferroptosis, and immune infiltration. By constructing a visualized bioinformatics model, the study seeks to uncover potential prognostic biomarkers and novel therapeutic avenues for LUAD, thereby providing a more comprehensive and multi-dimensional perspective for further exploration.\u003c/p\u003e"},{"header":"2. Materials and methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\u003ch2\u003e2.1 Data collection\u003c/h2\u003e\u003cp\u003eTranscriptome profiles and corresponding clinical data of LUAD patients were retrieved from The Cancer Genome Atlas (TCGA) database (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://portal.gdc.cancer.gov/\u003c/span\u003e\u003cspan address=\"https://portal.gdc.cancer.gov/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e), including 515 tumor samples and 59 matched adjacent normal tissues. After quality control measures, 2 cases with incomplete clinical records were excluded, resulting in 513 eligible LUAD samples for subsequent analyses. To enhance the representativeness of normal tissues, mRNA expression data of 288 healthy lung tissue samples were obtained from the Genotype-Tissue Expression (GTEx) project via the UCSC Xena browser (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://xenabrowser.net/datapages/\u003c/span\u003e\u003cspan address=\"https://xenabrowser.net/datapages/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). For validating the independent dataset, four independent microarray datasets (GSE27262, GSE43458, GSE116959, and GSE229705) were curated from the Gene Expression Omnibus (GEO) database (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://www.ncbi.nlm.nih.gov/geo/\u003c/span\u003e\u003cspan address=\"http://www.ncbi.nlm.nih.gov/geo/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eDEGs between TCGA-LUAD tumor and normal tissues were identified using the R package DESeq2, with significant differential expression defined as an absolute log₂ fold change (|log₂FC|)\u0026thinsp;\u0026ge;\u0026thinsp;1 and an adjusted p-value\u0026thinsp;\u0026le;\u0026thinsp;0.05. Parallel DEGs analysis in GEO datasets was performed using the GEO2R web tool, with the same statistical thresholds (|log₂FC| \u0026ge; 1, adjusted p\u0026thinsp;\u0026le;\u0026thinsp;0.05) applied to ensure consistency across platforms.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec4\" class=\"Section2\"\u003e\u003ch2\u003e2.2 WGCNA construction and module identification\u003c/h2\u003e\u003cp\u003eA scale-free co-expression network was constructed from TCGA-LUAD expression profiles integrated with clinical outcomes using the Weighted Gene Co-expression Network Analysis (WGCNA) R package (version 1.72-1). The soft-thresholding parameter (β) was determined as the minimum value enabling the scale-free topology fitting index (R\u0026sup2;) to reach\u0026thinsp;\u0026ge;\u0026thinsp;0.9, which ensures network connectivity follows a biologically meaningful power-law distribution. Module detection was conducted with a minimum module size threshold of 30 genes set to enhance functional interpretability. The association between modules and traits was evaluated by calculating the Pearson correlation coefficient between module eigengenes and clinical parameters (overall survival and status). Modules with the strongest clinical relevance were identified based on statistical significance and association strength. Additionally, Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis was performed on the selected DEGs using the DAVID bioinformatics resource (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://david.ncifcrf.gov/\u003c/span\u003e\u003cspan address=\"https://david.ncifcrf.gov/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e).\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec5\" class=\"Section2\"\u003e\u003ch2\u003e2.3 Calculation of signature from machine learning\u003c/h2\u003e\u003cp\u003eThe integrated LUAD RNA-seq dataset was randomly split into a training set and a validation set at a 7:3 ratio. To identify genes most closely associated with prognosis, eight distinct machine learning algorithms, including Random Forest (RF), Support Vector Machine (SVM), Gradient Boosting Machine (GBM), Neural Network (NNET), Least Absolute Shrinkage and Selection Operator (LASSO) regression, K-Nearest Neighbors (KNN), Decision Tree (DT), and Generalized Linear Model (GLM), were applied to the training set. Each algorithm was evaluated via 10-fold cross-validation (CV), with model performance determined by calculating the Area Under the Curve (AUC). All models were tested for overall survival (OS) across 513 LUAD samples. Specifically, GBM, NNET, DT, GLM, and KNN analyses were implemented using the R package \"caret\"; LASSO regression via \"glmnet\"; RF analysis via \"randomForest\"; SVM via \"kernlab\"; and model interpretation via \"DALEX\".\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec6\" class=\"Section2\"\u003e\u003ch2\u003e2.4 Definition of the m5c-related ferroptosis genes prognostic model\u003c/h2\u003e\u003cp\u003eTo further validate the association between survival time and m5C-related ferroptosis genes, univariate and multivariate Cox regression analyses were performed using the R packages survival (version 2.44.1.1). First, univariate Cox regression was conducted to determine the correlation between the expression levels of m5C-related ferroptosis genes and OS, with a significance threshold set at P\u0026thinsp;\u0026lt;\u0026thinsp;0.05. The m5C-related ferroptosis genes meeting this criterion were then subjected to multivariate Cox regression analysis, combined with clinical variables including age, gender, pathologic stage, and smoking status. This analysis aimed to evaluate their independent prognostic contribution and hazard ratio (HR), with 95% confidence intervals (CIs) calculated for key m5C-related ferroptosis genes. LUAD samples were stratified into high-risk and low-risk groups based on risk scores derived from the regression results. Time-dependent receiver operating characteristic (ROC) curves were constructed, and the AUC was calculated to assess the performance of the prognostic model. Kaplan-Meier survival analysis was performed to plot the survival time. To facilitate the clinical application of prognostic model, a comprehensive nomogram was developed using the R package \"rms\", integrating molecular risk scores and established clinical variables (age, gender, pathologic stage, and smoking status). A corresponding calibration plot was also constructed to compare the consistency between predicted and actual survival probabilities.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec7\" class=\"Section2\"\u003e\u003ch2\u003e2.5 Enrichment analysis and Immunological correlation evaluation\u003c/h2\u003e\u003cp\u003eTo identify the enriched pathways correlated with Gene Ontology-Biological Process (GO-BP) and KEGG pathways in high-risk and low-risk groups, Gene Set Variation Analysis (GSVA) was employed. This part of the analysis was mainly implemented using the clusterProfiler R package and the SangerBox online platform (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://sangerbox.com/index.html\u003c/span\u003e\u003cspan address=\"http://sangerbox.com/index.html\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). The proportions of tumor-infiltrating immune cell types within the mixed cell populations in tumor tissues were estimated using three distinct algorithms, namely the ESTIMATE algorithm, single-sample Gene Set Enrichment Analysis (ssGSEA) algorithm, and CIBERSORT deconvolution algorithm, respectively. Additionally, the Tumor Immune Estimation Resource (TIMER) (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://timer.comp-genomics.org/\u003c/span\u003e\u003cspan address=\"http://timer.comp-genomics.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) was utilized to analyze the correlation between gene expression and the infiltration of several immune cell types in LUAD.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e\u003ch2\u003e2.6 Drug sensitivity analysis and molecular docking\u003c/h2\u003e\u003cp\u003eThe GDSC2 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) was used to detect the association between SLC2A1 and half-maximal inhibitory concentration (IC50) of drugs. In this study, differences in IC50 values for each of the 198 drugs were included, which these drugs were categorized based on gene expression levels, and the analysis was performed using the R package oncoPredict. The online tool \"ROC Plotter\" (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://rocplot.com/\u003c/span\u003e\u003cspan address=\"https://rocplot.com/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) was utilized to analyze the relationship between SLC2A1 expression and immune checkpoint inhibitors. The online platform \"CB-DOCK2\" (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://cadd.labshare.cn/cb-dock2/php/index.php\u003c/span\u003e\u003cspan address=\"https://cadd.labshare.cn/cb-dock2/php/index.php\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) was employed to conduct molecular docking and visualization between SLC2A1 and small-molecule drugs. The chemical structures of small-molecule drugs were obtained from PubChem (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://pubchem.ncbi.nlm.nih.gov/#query\u003c/span\u003e\u003cspan address=\"https://pubchem.ncbi.nlm.nih.gov/#query\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). Additionally, the protein structure of SLC2A1 was predicted using the AlphaFold3 online database (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://alphafoldserver.com/\u003c/span\u003e\u003cspan address=\"https://alphafoldserver.com/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e).\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec9\" class=\"Section2\"\u003e\u003ch2\u003e2.7 Statistical analysis\u003c/h2\u003e\u003cp\u003eData processing, statistical analysis, and graphing were all performed using R software (Version 4.2.1), GraphPad Prism (Version 9.0), and the SangerBox platform. Student\u0026rsquo;s t-test was used for statistical significance analysis for intergroup comparisons. The Wilcoxon rank-sum test was applied to evaluate the correlation between continuous variables. The Spearman correlation coefficient was used to assess the ordinal correlation between variables. Significance was defined as p\u0026thinsp;\u0026lt;\u0026thinsp;0.05. (*P\u0026thinsp;\u0026lt;\u0026thinsp;0.05; **P\u0026thinsp;\u0026lt;\u0026thinsp;0.01; ***P\u0026thinsp;\u0026lt;\u0026thinsp;0.001; ****P\u0026thinsp;\u0026lt;\u0026thinsp;0.0001).\u003c/p\u003e\u003c/div\u003e"},{"header":"3 Results","content":"\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e\u003ch2\u003e3.1 Identification of prognostic-related modules using WGCNA in LUAD\u003c/h2\u003e\u003cp\u003eRNA sequencing data comprising 59 adjacent normal lung tissues and 523 LUAD tissues from TCGA were retrieved for comprehensive transcriptomic analysis. With the screening criteria of |log₂FC| \u0026gt;1 and adj-p\u0026thinsp;\u0026lt;\u0026thinsp;0.05, a total of 5573 DEGs were identified, including 3374 upregulated genes and 2199 downregulated genes (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA). Next, WGCNA was performed on all DEGs. 513 LUAD samples with complete clinical annotations were retained after excluding cases with incomplete records. Based on the scale-free network topology with R\u0026sup2; = 0.95, the Pearson correlation matrix of genes was converted into a weighted adjacency matrix via power transformation with β\u0026thinsp;=\u0026thinsp;3 (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eB-E). Using the dynamic tree cutting algorithm, all selected genes were clustered based on different metrics of the topological overlap matrix (TOM), and the tree could be divided into 29 modules marked with different colors (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eF).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eAccording to the similarity between modules shown in the clustering tree, similar modules were merged with a merging threshold of 0.6, and 26 modules were retained and marked with different colors (merge dynamic) (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eG). Subsequently, gene co-expression was summarized using eigengenes, and the correlation between each eigengene and clinical features was calculated, including age, sex, smoking status, pathologic stage, primary tumor stage (stage_T), distant metastasis stage (stage_M), regional lymph node stage (stage_N), survival status (status), and survival time (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eH). The module-trait relationship plot showed that the co-expressed \"purple module\" (containing 123 genes) was significantly positively correlated with pathologic stage (R\u0026thinsp;=\u0026thinsp;0.14, p\u0026thinsp;=\u0026thinsp;0.002), stage_T (R\u0026thinsp;=\u0026thinsp;0.18, p\u0026thinsp;=\u0026thinsp;3e\u0026thinsp;\u0026minus;\u0026thinsp;05), and status (R\u0026thinsp;=\u0026thinsp;0.13, p\u0026thinsp;=\u0026thinsp;0.004), while negatively correlated with survival time (R = -0.09, p\u0026thinsp;=\u0026thinsp;0.04) (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eH). The co-expressed \"turquoise module\" (containing 1289 genes) was significantly positively correlated with pathologic stage (R\u0026thinsp;=\u0026thinsp;0.17, p\u0026thinsp;=\u0026thinsp;1e\u0026thinsp;\u0026minus;\u0026thinsp;04), stage_T (R\u0026thinsp;=\u0026thinsp;0.12, p\u0026thinsp;=\u0026thinsp;0.008), and status (R\u0026thinsp;=\u0026thinsp;0.16, p\u0026thinsp;=\u0026thinsp;2e\u0026thinsp;\u0026minus;\u0026thinsp;04) (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eH). Thus, these two modules may be key modules for determining the prognosis and metastasis of LUAD, and the genes within them were extracted for further analysis (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eH). KEGG pathway enrichment analyses indicated these genes were significant enrichment in the cell cycle, DNA replication, biosynthesis of antibiotics, fanconi anemia pathway and p53 signaling pathway (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eI).\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec12\" class=\"Section2\"\u003e\u003ch2\u003e3.2 Exploration of m5C-related ferroptosis gene in prognosis via machine learning in LUAD\u003c/h2\u003e\u003cp\u003eTo identify m5C-related ferroptosis gene in prognosis in LUAD, we conducted bioinformatics data mining. Based on the intersections between 1412 DEGs related to prognosis derived from WGCNA and ferroptosis-related genes yielded in the FerrDb database, 45 ferroptosis-related and prognosis-related DEGs were identified (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA). Meanwhile, we selected 14 m5C regulators, which are writer: NOP2, NSUN2, NSUN3, NSUN4, NSUN5, NSUN6, DNMT1, TRDMT1, DNMT3A, and DNMT3B; reader: ALYREF; eraser: TET2 and YBX1 (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA). Pearson correlation analysis was performed between 14 m5C regulators and the 45 ferroptosis-related and prognosis-related DEGs identified. Under the criteria of |Pearson coefficient| \u0026gt;0.4 and P adj\u0026thinsp;\u0026le;\u0026thinsp;0.01, 29 m5C-related ferroptosis genes in prognosis in LUAD were obtained (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eB).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eSubsequently, eight machine learning algorithms were employed to further narrow 29 m5C-related ferroptosis genes. These algorithms included RF, SVM, GBM, NNET, LASSO regression, KNN, DT, and GLM (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eC-D). Among the eight evaluated machine learning models, including the RF, SVM, and LASSO exhibited more excellent performance (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eE). Thus, we focused on these three machine learning methods for in-depth investigation. In the RF algorithm, we ranked the important of the 29 m5C-related ferroptosis genes, select the top twenty genes as the significant genes, and listed the top 10 m5C-related ferroptosis genes (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eF-G). Using the LASSO algorithm, we identified 12 m5C-related ferroptosis genes (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eH-I). Additionally, under the SVM algorithm, we identified a total of 28 m5C-related ferroptosis genes (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eJ-K).\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec13\" class=\"Section2\"\u003e\u003ch2\u003e3.3 Development of prognostic model using the m5C-related ferroptosis genes\u003c/h2\u003e\u003cp\u003eBased on the aforementioned three machine learning algorithms, the overlapping m5C-related ferroptosis genes (SLC2A1, RRM2, ZFP69B, KIF20A, EZH2, STMN1, PARP1, GPX2, SLC16A1, and SLC7A11) were selected as robust prognostic factors (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eA). Subsequently, univariate and multivariate Cox regression analyses were performed on 10 key m5C-related ferroptosis genes. The results indicated that SLC2A1, RRM2, KIF20A, GPX2, and SLC7A11 were statistically significant (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eB-C). Based on the constructed prognostic risk score model, 513 TCGA-LUAD samples were divided into high-risk and low-risk groups according to the median risk score (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eD). Time-dependent ROC curves showed that the risk score model had stable performance, with areas under the curve (AUC) of 0.731, 0.749, and 0.704 for 1-year, 3-year, and 5-year survival rates, respectively (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eE). Kaplan-Meier survival curves for OS revealed that the predicted survival time of the high-risk group was significantly shorter than that of the low-risk group (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eF). Furthermore, to forecast OS in LUAD patients, a nomogram was developed, incorporating age, gender, pathologic stage, and the expression levels of m5C-related ferroptosis genes (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eG). Calibration curves were plotted for 1-year, 3-year, and 5-year survival rates, which confirmed that this nomogram could predict the OS of LUAD patients (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eH).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eTo evaluate the distribution of biological processes between the high-risk and low-risk groups, we performed GSVA enrichment analysis. For gene sets defined by GO-BP, the high-risk group was mainly associated with pathways such as NEGATIVE_T_CELL_SELECTION, MAST_CELL_ACTIVATION, and IMMUNOLOGICAL_MEMORY_PROCESS (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA). In contrast, the low-risk group was mainly associated with pathways including DNA_DEPENDENT_DNA_REPLICATION, SISTER_CHROMATID_SEGREGATION, and DNA_REPLICATION (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eB). For gene sets defined by KEGG, the high-risk group was primarily correlated with pathways such as ALLOGRAFT_REJECTION, AUTOIMMUNE_THYROID_DISEASE, and INTESTINAL_IMMUNE_NETWORK_FOR_IGA_PRODUCTION (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eC). Meanwhile, the low-risk group was primarily correlated with pathways including CELL_CYCLE, DNA_REPLICATION, and OOCYTE_MEIOSIS (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eD). Subsequently, we applied several immune infiltration algorithms to estimate the infiltration levels of different types of immune cells in the TME. We used the ESTIMATE algorithm to calculate the distribution of stromal, immune, and estimate scores between patients in the high-risk and low-risk groups. Compared with the high-risk group, the low-risk group had higher immune, stromal, and ESTIMATE scores (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eE). The ssGSEA algorithm showed that the high-risk group had a higher level of Th2 cells, while the low-risk group had higher levels of B cells, CD8 T cells, Cytotoxic cells, DC, Eosinophils, iDC, Macrophages, Mast cells, NK cells, T cells, Tcm, Tem, TFH, Th1 cells, and Th17 cells (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eF). The CIBERSORT algorithm revealed that the high-risk group had higher levels of resting NK cells, M0 Macrophages, and M1 Macrophages, whereas the low-risk group had higher levels of memory B cells, resting CD4 memory T cells, Monocytes, resting Dendritic cells, and resting Mast cells. These results collectively indicate that the prognostic model is closely associated with immune cell infiltration in the TME (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eG).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec14\" class=\"Section2\"\u003e\u003ch2\u003e3.4 Expression levels of m5C-related ferroptosis genes in LUAD\u003c/h2\u003e\u003cp\u003eThen, we analyzed the strength of correlations between key m5C-related ferroptosis genes and various m5C regulators. The results showed that GPX2 exhibited a correlation solely with the DNMT1 (R\u0026thinsp;=\u0026thinsp;\u0026minus;\u0026thinsp;0.243, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001) (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA), while SLC7A11 only correlated with the NOP2 (R\u0026thinsp;=\u0026thinsp;0.185, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001) (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eB), both exhibiting relatively weak correlations. In contrast, SLC2A1 displayed strong correlations with three m5C regulators: NSUN2 (R\u0026thinsp;=\u0026thinsp;0.383, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), ALYREF (R\u0026thinsp;=\u0026thinsp;0.509, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), and YBX1 (R\u0026thinsp;=\u0026thinsp;0.470, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001) (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eC). RRM2 correlated with six m5C regulators, namely NOP2 (R\u0026thinsp;=\u0026thinsp;0.180, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), NSUN2 (R\u0026thinsp;=\u0026thinsp;0.447, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), DNMT1 (R\u0026thinsp;=\u0026thinsp;0.452, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), DNMT3B (R\u0026thinsp;=\u0026thinsp;0.513, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), ALYREF (R\u0026thinsp;=\u0026thinsp;0.688, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), and YBX1 (R\u0026thinsp;=\u0026thinsp;0.517, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001) (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eD). Notably, KIF20A displayed the strongest correlations with m5C regulators, including NOP2 (R\u0026thinsp;=\u0026thinsp;0.162, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), NSUN2 (R\u0026thinsp;=\u0026thinsp;0.474, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), DNMT1 (R\u0026thinsp;=\u0026thinsp;0.523, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), DNMT3A (R\u0026thinsp;=\u0026thinsp;0.404, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), DNMT3B (R\u0026thinsp;=\u0026thinsp;0.551, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), ALYREF (R\u0026thinsp;=\u0026thinsp;0.655, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), and YBX1 (R\u0026thinsp;=\u0026thinsp;0.496, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001) (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eE). Therefore, we selected these three m5C-related ferroptosis genes (SLC2A1, RRM2, and KIF20A) for further analysis.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eSubsequently, we analyzed the expression levels of SLC2A1, RRM2, and KIF20A in a larger cohort of LUAD samples from the XENA database (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://xenabrowser.net/datapages/\u003c/span\u003e\u003cspan address=\"https://xenabrowser.net/datapages/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e), which included 347 normal tissue samples and 515 tumor tissue samples. The results showed that SLC2A1, RRM2, and KIF20A exhibited significantly higher expression levels in tumor tissues (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eA-C). Meanwhile, we incorporated the GEO database for validation, the results demonstrated that SLC2A1, RRM2, and KIF20A consistently maintained high expression levels in tumor tissues across the GSE27262, GSE43458, GSE116959, and GSE229705 datasets (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eD-G). Additionally, we showed that high expression of SLC2A1, RRM2, and KIF20A was associated with poor prognosis in LUAD patients (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eH-J). We also utilized the Human Protein Atlas (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.proteinatlas.org/\u003c/span\u003e\u003cspan address=\"https://www.proteinatlas.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) to perform immunohistochemical (IHC) staining in patients diagnosed with LUAD. Our analysis revealed that the expression levels of SLC2A1, RRM2, and KIF20A in LUAD tissues were statistically significantly higher compared to normal tissues (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eK-M). These results indicated that the expression of SLC2A1, RRM2, and KIF20A is closely associated with the progression of LUAD.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec15\" class=\"Section2\"\u003e\u003ch2\u003e3.5 Identification of potential SLC2A1 inhibitors through pharmacogenetic screening\u003c/h2\u003e\u003cp\u003eFurthermore, the TIMER database demonstrated that the expression of three m5C-related ferroptosis genes (SLC2A1, RRM2, and KIF20A) was significantly correlated with the infiltration of several immune cell types in LUAD, including B cells, macrophages, neutrophils, and CD4\u0026thinsp;+\u0026thinsp;T cells. Specifically, the results revealed that KIF20A expression was positively correlated with neutrophils while negatively correlated with B cells and macrophages (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eA). RRM2 expression showed a positive correlation with neutrophils and a negative correlation with B cells and CD4\u0026thinsp;+\u0026thinsp;T cells (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eB-C). For SLC2A1, its expression was positively correlated with neutrophils and negatively correlated with B cells (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eD). Meanwhile, we analyzed the correlation between the expression of SLC2A1, RRM2, KIF20A and immunotherapy response using ROC Plotter (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://rocplot.com/\u003c/span\u003e\u003cspan address=\"https://rocplot.com/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). Intriguingly, only high expression of SLC2A1 was significantly associated with a higher response rate to anti-PD-1 and anti-PD-L1 therapies (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eE-F), and it was also correlated with the overall therapeutic efficacy of immune checkpoint inhibitors (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eG). In contrast, RRM2 and KIF20A exhibited no correlation with the therapeutic effect of immune checkpoint inhibitors.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eTo identify potential therapeutic drugs targeting SLC2A1 for LUAD, we screened the Genomics of Drug Sensitivity in Cancer (GDSC) database to select agents that exhibit enhanced efficacy under conditions of high SLC2A1 expression. Drug sensitivity analysis performed using the oncoPredict package identified nine small-molecule drugs, namely BI.2536, Dabrafenib, IWP.2, Leflunomide, LJI308, MN.64, RO.3306, Sinularin, and WEHI.539, with altered potency (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eA-I). Namely, LUAD patients with high SLC2A1 expression showed increased sensitivity to these drugs. Furthermore, we predicted the protein structure of SLC2A1 using the AlphaFold3 online platform (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://golgi.sandbox.google.com/\u003c/span\u003e\u003cspan address=\"https://golgi.sandbox.google.com/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eJ). Subsequently, we employed the CB-DOCK2 online tool to predict the potential binding sites and binding regions of SLC2A1 when interacting with the nine small-molecule compounds (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eK-S). The results of this structural prediction analysis indicated that SLC2A1 and the nine small-molecule compounds exhibit favorable binding efficiency. These results indicate that these drugs have the potential to be repurposed as anti-cancer agents targeting SLC2A1 to inhibit tumor progression in LUAD patients.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e"},{"header":"4 Discussion and Conclusion","content":"\u003cp\u003eLUAD remains a leading cause of cancer-related mortality worldwide, with marked heterogeneity in its clinical manifestations, this limits the efficacy of standardized therapies. Although RNA m5C modification, has been demonstrated to be associated with the development and drug resistance of LUAD, and ferroptosis is essential for tumor suppression, their interaction in shaping LUAD prognosis remains unelucidated. Herein, we address this gap using WGCNA and multimodal machine learning to identify m5C-related ferroptosis genes as prognostic biomarkers and therapeutic targets. Concurrently, we constructed a prognostic risk model to screen for genes closely associated with LUAD progression, and validated this prognostic model. Furthermore, we developed a novel survival probability prediction model based on the clinicopathological features of LUAD to support clinicians in designing personalized treatment regimens. Finally, through our multi-level analyses, we identified the key prognostic gene SLC2A1, which links m5C modification, ferroptosis, and immune infiltration to LUAD progression.\u003c/p\u003e\u003cp\u003eAmong these prognosis-related genes, SLC2A1, RRM2, and KIF20A stand out, given their strong correlations with m5C regulators and ubiquitous high expression in LUAD. Existing studies demonstrate that ALYREF mediates m5C methylation of KIF20A mRNA, stabilizing KIF20A expression to confer ferroptosis resistance in cervical cancer cells[\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e], and this is a finding fully consistent with our observation that KIF20A exhibits the strongest correlation with ALYREF (R\u0026thinsp;=\u0026thinsp;0.655, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001). Additionally, KIF20A knockdown enhances the synergistic antiproliferative effects of gemcitabine and the ferroptosis inducer on LUAD cells[\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e]. Inhibition of KIF20A boosts hepatocellular carcinoma immunotherapy efficacy by promoting c-Myc ubiquitination[\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e], while comprehensive bioinformatics analyses have validated KIF20A as a prognostic factor and therapeutic target for LUAD[\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e]. Collectively, these data establish KIF20A as a robust candidate for drug development, given its tight associations with ferroptosis, m5C modification, and immunotherapy. For RRM2, no studies linking it to m5C modification have been reported to date. However, as a key ferroptosis regulator, RRM2 is closely associated with the progression of hepatocellular carcinoma[\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e], lung cancer[\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e], ovarian cancer[\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e], and sepsis [\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e]. It also correlates strongly with immune infiltration levels across pan-cancer contexts: RRM2 knockdown enhances the antitumor efficiency of PD-1 blockade in renal cell carcinoma[\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e], and RRM2 has been characterized as a critical ferroptosis regulator in LUAD, promoting tumor immune infiltration via ferroptosis inhibition[\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e]. These results strongly corroborate our analysis, supporting that RRM2 may modulate LUAD progression through mechanisms involving ferroptosis, m5C modification, and immune infiltration.\u003c/p\u003e\u003cp\u003eSLC2A1, also known as glucose transporter type 1 (GLUT1), is a facilitative glucose transporter responsible for the continuous uptake and transport of glucose. Studies have demonstrated that SLC2A1 is overexpressed in multiple cancers, including breast cancer, lung cancer, hepatocellular carcinoma, colorectal cancer, and gastric cancer[\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e]. SLC2A1 is deeply involved in regulating TME remodeling across different tumors[\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e, \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e]; it also serves as a classic ferroptosis-related gene and a diagnostic biomarker for immune infiltration in colorectal cancer[\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e]. Additionally, evidence has linked SLC2A1 to the ferroptosis mechanism mediated by m6A modification[\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. In the present study, we observed that SLC2A1 expression in LUAD was positively correlated with neutrophils and negatively correlated with B cells. Meanwhile, high SLC2A1 expression was significantly associated with a higher response rate to anti-PD-1 and anti-PD-L1 therapies\u0026mdash;findings that support its potential as a predictive biomarker for immune checkpoint inhibitors (ICIs). Via pharmacogenomic screening using the GDSC database, we further identified enhanced efficacy of nine small-molecule drugs (e.g., BI.2536, dabrafenib) in LUAD with high SLC2A1 expression, and structural prediction additionally confirmed favorable binding between SLC2A1 and these small-molecule compounds, supporting SLC2A1 as a target for drug repurposing. Collectively, our study further validates SLC2A1 as a suitable molecular target that connects m5C modification, ferroptosis, and immunotherapy.\u003c/p\u003e\u003cp\u003eBased on genes mediating the m5C-ferroptosis crosstalk, the present study established a prognostic signature for LUAD, which exhibits high accuracy in OS prediction and TME stratification. Notably, SLC2A1 was identified as a dual-functional biomarker with both prognostic value and predictive utility for ICIs responses, as well as a therapeutic target. Additionally, repurposable small-molecule drugs targeting SLC2A1 were identified. These findings deepen the understanding of m5C-ferroptosis interactions in LUAD and pave the way for precision oncology strategies. Nevertheless, this study has several limitations that remain to be addressed. First, given the reliance on retrospective data, prospective cohort studies are required to validate the utility of the identified biomarker. Second, mechanistic studies are needed to elucidate how m5C regulates the expression or function of SLC2A1, RRM2, and KIF20A, which would extend the current understanding of m5C-mediated post-transcriptional regulation in tumorigenesis. Third, in vitro and in vivo experiments are necessary to validate the efficacy of SLC2A1-targeting drugs and their synergistic effects with ICIs, as preclinical validation is critical for translating molecular findings to clinical applications. Finally, intratumoral heterogeneity may impact the performance of the biomarker, requires further investigation.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe want to express our gratitude to the Shanghai Public Health Clinical Center Study group and thank the research teams involved in maintaining and sharing this valuable resource. All lab members are acknowledged for stimulating discussions.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDeclaration of competing interest\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis research received no funding.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData Availability Statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe data used in this study is publicly available and can be accessed through the links provided.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor Contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eJinlong Zhong and Danping Liu designed the study and approved the final version of the submitted article. Qi Wang and Heng Zou developed methodology, performed the experiments and researched the data. Qi Wang analyzed and interpreted the data. Qi Wang, Heng Zou and Danping Liu wrote and edited the manuscript. All authors read and approved the contents of the manuscript and its publication.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics approval\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent to participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent to publish\u003cbr\u003e\u0026nbsp;\u003c/strong\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eClinical Trial Number\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eBray F, Ferlay J, Soerjomataram I et al. Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin 2018; 68: 394-424.\u003c/li\u003e\n\u003cli\u003eThai AA, Solomon BJ, Sequist LV et al. Lung cancer. Lancet 2021; 398: 535-554.\u003c/li\u003e\n\u003cli\u003eLi D, Chen T, Li QG. Identification of a m(6)A-related ferroptosis signature as a potential predictive biomarker for lung adenocarcinoma. BMC Pulm Med 2023; 23: 128.\u003c/li\u003e\n\u003cli\u003eLi Y, Yan B, He S. Advances and challenges in the treatment of lung cancer. Biomed Pharmacother 2023; 169: 115891.\u003c/li\u003e\n\u003cli\u003eGrant C, Hagopian G, Nagasaka M. Neoadjuvant therapy in non-small cell lung cancer. Crit Rev Oncol Hematol 2023; 190: 104080.\u003c/li\u003e\n\u003cli\u003eLi R, Chen H, Li C et al. The prognostic value and immune landscaps of m6A/m5C-related lncRNAs signature in the low grade glioma. BMC Bioinformatics 2023; 24: 274.\u003c/li\u003e\n\u003cli\u003eMa C, Gu Z, Yang Y. Development of m6A/m5C/m1A regulated lncRNA signature for prognostic prediction, personalized immune intervention and drug selection in LUAD. J Cell Mol Med 2024; 28: e18282.\u003c/li\u003e\n\u003cli\u003eHaruehanroengra P, Zheng YY, Zhou Y et al. RNA modifications and cancer. RNA Biol 2020; 17: 1560-1575.\u003c/li\u003e\n\u003cli\u003eWang Y, Wei J, Feng L et al. Aberrant m5C hypermethylation mediates intrinsic resistance to gefitinib through NSUN2/YBX1/QSOX1 axis in EGFR-mutant non-small-cell lung cancer. Mol Cancer 2023; 22: 81.\u003c/li\u003e\n\u003cli\u003eLi H, Yu K, Hu H et al. METTL17 coordinates ferroptosis and tumorigenesis by regulating mitochondrial translation in colorectal cancer. Redox Biol 2024; 71: 103087.\u003c/li\u003e\n\u003cli\u003eYu M, Ni M, Xu F et al. NSUN6-mediated 5-methylcytosine modification of NDRG1 mRNA promotes radioresistance in cervical cancer. Mol Cancer 2024; 23: 139.\u003c/li\u003e\n\u003cli\u003eYang Y, Fan H, Liu H et al. NOP2 facilitates EZH2-mediated epithelial-mesenchymal transition by enhancing EZH2 mRNA stability via m5C methylation in lung cancer progression. Cell Death Dis 2024; 15: 506.\u003c/li\u003e\n\u003cli\u003eCai D, Chen X, Xu H et al. m5C-modified circRREB1 promotes lung cancer progression by inducing mitophagy. J Exp Clin Cancer Res 2025; 44: 203.\u003c/li\u003e\n\u003cli\u003eMeng H, Miao H, Zhang Y et al. YBX1 promotes homologous recombination and resistance to platinum-induced stress in ovarian cancer by recognizing m5C modification. Cancer Lett 2024; 597: 217064.\u003c/li\u003e\n\u003cli\u003eZhang Y, Chen XN, Zhang H et al. CDK13 promotes lipid deposition and prostate cancer progression by stimulating NSUN5-mediated m5C modification of ACC1 mRNA. Cell Death Differ 2023; 30: 2462-2476.\u003c/li\u003e\n\u003cli\u003eWu J, Zhao Q, Chen S et al. NSUN4-mediated m5C modification of circERI3 promotes lung cancer development by altering mitochondrial energy metabolism. Cancer Lett 2024; 605: 217266.\u003c/li\u003e\n\u003cli\u003eChen X, Comish PB, Tang D, Kang R. Characteristics and Biomarkers of Ferroptosis. Front Cell Dev Biol 2021; 9: 637162.\u003c/li\u003e\n\u003cli\u003eJi X, Qian J, Rahman SMJ et al. xCT (SLC7A11)-mediated metabolic reprogramming promotes non-small cell lung cancer progression. Oncogene 2018; 37: 5007-5019.\u003c/li\u003e\n\u003cli\u003eLi M, Shi H, Dong J et al. Mechanisms of Ferroptosis-Related Genes in Gallbladder Cancer Based on Bioinformatics Analysis. Mol Biotechnol 2025; 67: 1814-1825.\u003c/li\u003e\n\u003cli\u003eNiu K, Chen Z, Li M et al. NSUN2 lactylation drives cancer cell resistance to ferroptosis through enhancing GCLC-dependent glutathione synthesis. Redox Biol 2025; 79: 103479.\u003c/li\u003e\n\u003cli\u003eChen SJ, Zhang J, Zhou T et al. Epigenetically upregulated NSUN2 confers ferroptosis resistance in endometrial cancer via m(5)C modification of SLC7A11 mRNA. Redox Biol 2024; 69: 102975.\u003c/li\u003e\n\u003cli\u003eShi CJ, Pang FX, Lei YH et al. 5-methylcytosine methylation of MALAT1 promotes resistance to sorafenib in hepatocellular carcinoma through ELAVL1/SLC7A11-mediated ferroptosis. Drug Resist Updat 2025; 78: 101181.\u003c/li\u003e\n\u003cli\u003eYang Y, Hao L, Guiyang L, Haozhe P. Multifaceted bioinformatic analysis of m6A-related ferroptosis and its link with gene signatures and tumour-infiltrating immune cells in gliomas. J Cell Mol Med 2024; 28: e70060.\u003c/li\u003e\n\u003cli\u003eLiao P, Wang W, Wang W et al. CD8(+) T cells and fatty acids orchestrate tumor ferroptosis and immunity via ACSL4. Cancer Cell 2022; 40: 365-378 e366.\u003c/li\u003e\n\u003cli\u003eXu H, Ye D, Ren M et al. Ferroptosis in the tumor microenvironment: perspectives for immunotherapy. Trends Mol Med 2021; 27: 856-867.\u003c/li\u003e\n\u003cli\u003eLi Y, Jin H, Li Q et al. The role of RNA methylation in tumor immunity and its potential in immunotherapy. Mol Cancer 2024; 23: 130.\u003c/li\u003e\n\u003cli\u003eChen Z, Li Q, Lin Y et al. m5C regulator-mediated methylation modification phenotypes characterized by distinct tumor microenvironment immune heterogenicity in colorectal cancer. Sci Rep 2023; 13: 11950.\u003c/li\u003e\n\u003cli\u003eGao Y, Zou T. Hypoxia Triggers ALYREF-Mediated m(5)C Methylation of KIF20A to Activate KIF20A/BUB1 for Generating Ferroptosis Resistance in Cervical Cancer Cells. Kaohsiung J Med Sci 2025; e70093.\u003c/li\u003e\n\u003cli\u003eHe H, Liang L, Huang J et al. KIF20A is associated with clinical prognosis and synergistic effect of gemcitabine combined with ferroptosis inducer in lung adenocarcinoma. Front Pharmacol 2022; 13: 1007429.\u003c/li\u003e\n\u003cli\u003eChen S, Zhao L, Liu J et al. Inhibition of KIF20A enhances the immunotherapeutic effect of hepatocellular carcinoma by enhancing c-Myc ubiquitination. Cancer Lett 2024; 598: 217105.\u003c/li\u003e\n\u003cli\u003eSun D, Zhang H, Zhang C, Wang L. An evaluation of KIF20A as a prognostic factor and therapeutic target for lung adenocarcinoma using integrated bioinformatics analysis. Front Bioeng Biotechnol 2022; 10: 993820.\u003c/li\u003e\n\u003cli\u003eYang Y, Lin J, Guo S et al. RRM2 protects against ferroptosis and is a tumor biomarker for liver cancer. Cancer Cell Int 2020; 20: 587.\u003c/li\u003e\n\u003cli\u003eTang B, Xu W, Wang Y et al. Identification of critical ferroptosis regulators in lung adenocarcinoma that RRM2 facilitates tumor immune infiltration by inhibiting ferroptotic death. Clin Immunol 2021; 232: 108872.\u003c/li\u003e\n\u003cli\u003eFeng L, Zou X, Tang L et al. CircASH1L inhibits ferroptosis and enhances cisplatin resistance by sponging miR-515-5p to regulate cell cycle-related CDCA7/RRM2 in ovarian cancer cells. Front Pharmacol 2025; 16: 1563869.\u003c/li\u003e\n\u003cli\u003eHe S, He Y, Deng L et al. Identification of RRM2 as a key ferroptosis-related gene in sepsis. Inflamm Res 2024; 73: 459-473.\u003c/li\u003e\n\u003cli\u003eXiong W, Zhang B, Yu H et al. RRM2 Regulates Sensitivity to Sunitinib and PD-1 Blockade in Renal Cancer by Stabilizing ANXA1 and Activating the AKT Pathway. Adv Sci (Weinh) 2021; 8: e2100881.\u003c/li\u003e\n\u003cli\u003eMa C, Luo H, Cao J et al. Independent prognostic implications of RRM2 in lung adenocarcinoma. J Cancer 2020; 11: 7009-7022.\u003c/li\u003e\n\u003cli\u003eXi X, Gong X, Liu Y et al. SLC2A1 and MPST as diagnostic and prognostic biomarkers of potential endometrial cancer. Front Immunol 2025; 16: 1575916.\u003c/li\u003e\n\u003cli\u003eShi Y, Kotchetkov IS, Dobrin A et al. GLUT1 overexpression enhances CAR T cell metabolic fitness and anti-tumor efficacy. Mol Ther 2024; 32: 2393-2405.\u003c/li\u003e\n\u003cli\u003eYang S, Qian L, Li Z et al. Integrated Multi-Omics Landscape of Liver Metastases. Gastroenterology 2023; 164: 407-423 e417.\u003c/li\u003e\n\u003cli\u003eLiu XS, Yang JW, Zeng J et al. SLC2A1 is a Diagnostic Biomarker Involved in Immune Infiltration of Colorectal Cancer and Associated With m6A Modification and ceRNA. Front Cell Dev Biol 2022; 10: 853596.\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":"discover-oncology","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"dion","sideBox":"Learn more about [Discover Oncology](https://www.springer.com/12672)","snPcode":"","submissionUrl":"","title":"Discover Oncology","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Discover Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Lung adenocarcinoma, m5C, ferroptosis, tumor immunity","lastPublishedDoi":"10.21203/rs.3.rs-8096492/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8096492/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e\u003cp\u003eRecent research highlights the pivotal role of 5-methylcytosine (m5C) modification and ferroptosis in the progression of various cancers. However, the prognostic value of m5C-related ferroptosis genes in lung adenocarcinoma (LUAD) remains unclear. This study aims to establish a prognostic framework centered on m5C-related ferroptosis genes to improve the accuracy of prognosis prediction in LUAD patients, thereby optimizing targeted therapeutic strategies.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e\u003cp\u003eThe mRNA expression profiles, along with clinicopathological information of LUAD patients were obtained from The Cancer Genome Atlas (TCGA). Differential expression and weighted gene co-expression network analysis (WGCNA) identified prognosis-related modules. Ferroptosis-related genes and m5C regulators were integrated to identify m5C-associated ferroptosis genes. Machine learning and Cox regression were applied to construct a prognostic model. Finally, functional enrichment, immune infiltration, and drug sensitivity analyses were performed.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e\u003cp\u003eTwo key gene modules significantly correlated with LUAD prognosis were identified, yielding 29 m5C-related ferroptosis genes. Ten hub genes were selected via machine learning, and the prognostic risk model achieved strong predictive performance. Immune infiltration profiling revealed close associations between the model and tumor microenvironment features. Among the hub genes, SLC2A1, RRM2, and KIF20A were markedly overexpressed in tumor tissues and associated with poor survival. Notably, SLC2A1 expression correlated with enhanced immunotherapy response. Drug sensitivity analysis and molecular docking identified nine small-molecule compounds with favorable binding affinities to SLC2A1.\u003c/p\u003e\u003ch2\u003eConclusions\u003c/h2\u003e\u003cp\u003eThis study delineates the critical prognostic significance of m5C-related ferroptosis genes in LUAD and establishes a clinically relevant prognostic model. The identification of candidate SLC2A1 inhibitors offers promising avenues for targeted therapy and personalized treatment strategies in LUAD.\u003c/p\u003e","manuscriptTitle":"Immune Profiling and Drug Sensitivity in LUAD: Insights from m5C-Related Ferroptosis Gene Analysis","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-12-08 11:42:37","doi":"10.21203/rs.3.rs-8096492/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2026-01-06T04:56:04+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-01-04T06:06:17+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"94264777282441702768389424149556311156","date":"2025-12-26T11:30:12+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-12-25T16:11:34+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"216001808651637417694622850719804450285","date":"2025-12-23T23:23:39+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-12-04T14:41:55+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2025-12-01T08:03:14+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-11-14T03:01:54+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-11-14T03:00:33+00:00","index":"","fulltext":""},{"type":"submitted","content":"Discover Oncology","date":"2025-11-12T12:45:39+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"discover-oncology","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"dion","sideBox":"Learn more about [Discover Oncology](https://www.springer.com/12672)","snPcode":"","submissionUrl":"","title":"Discover Oncology","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Discover Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"f16f09e5-e1f1-4d01-a2cb-e3bb1e36dbca","owner":[],"postedDate":"December 8th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-03-31T08:42:08+00:00","versionOfRecord":[],"versionCreatedAt":"2025-12-08 11:42:37","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8096492","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8096492","identity":"rs-8096492","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}