Targeting Macrophage-Associated Core Genes for Prognostic Prediction and Therapeutic Insights in Bladder Cancer

Targeting Macrophage-Associated Core Genes for Prognostic Prediction and Therapeutic Insights in Bladder Cancer | 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 Targeting Macrophage-Associated Core Genes for Prognostic Prediction and Therapeutic Insights in Bladder Cancer HaoLin Liu, Yuanqi Chu, Jian Hou, Yumin Wang, Junxiong Li, Jingbo Qin, and 5 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6452577/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Background: Bladder cancer (Bca) is a highly malignant tumor characterized by a high recurrence and metastasis rate. Macrophages crucially affect tumor progression and immunotherapy response, while researches have not well explored their precise mechanisms of action against Bca. The study aims at investigating the function of macrophage-related genes (MRGs) within the Bca immune microenvironment and exploring their potential value in prognosis prediction and therapeutic decision-making. Method: This study integrated Bca transcriptomic data from the TCGA and GEO databases along with single-cell RNA sequencing (scRNA-seq) data to systematically identify key MRGs. Differential expression analysis, weighted gene co-expression network analysis (WGCNA), and single-cell sequencing analysis served for screening for core MRGs. The results from LASSO Cox regression analysis were used for constructing a survival risk prediction model, together with the evaluation of the model’s predictive accuracy. Besides, core MRGs were subjected to immune cell infiltration and drug sensitivity analyses for the elucidation of their roles in immune regulation and therapeutic response. Furthermore, key genes in the prognostic model were validated using PCR, Western blot, and immunohistochemistry. Result: This study identified 11 core genes significantly associated with macrophages and developed a risk prediction model based on ANXA1, ST3GAL5, and VIM. The model demonstrated high predictive accuracy across all samples (AUC = 0.682). Immune analysis revealed that high-risk patients exhibited a distinctly immunosuppressive tumor microenvironment (TME), characterized by increased infiltration of M2 macrophages and neutrophils, along with a significant reduction in effector immune cells of CD8⁺ T cells and NK cells. Additionally, high-risk patients displayed greater sensitivity to targeted therapies (e.g., EGFR and HER2 inhibitors) but reduced sensitivity to conventional chemotherapy. According to in vitro and in vivo experiments, ST3GAL5 overexpression significantly promoted Bca cell proliferation and tumor growth, underscoring its potential role in tumor progression. Conclusion: This study highlights the crucial impact of MRGs on the TME of Bca and constructs a risk prediction model that effectively predicts patient survival outcomes, providing a theoretical foundation and practical guidance for personalized treatment strategies. Future researches are suggested to more deeply elucidate the functional mechanisms of these core genes as well as explore their potential as therapeutic targets for Bca. Bladder cancer Macrophages Survival risk model Immune microenvironment Drug sensitivity analysis Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 1. Introduction Bladder cancer (Bca) is a representative and lethal malignancy of the urinary system worldwide, with over 170,000 related deaths annually. It ranks 4th among all cancers for male and is one of the most frequently diagnosed malignancies for female [ 1 , 2 ]. Based on pathological characteristics, bladder cancer (BC) falls into non-muscle-invasive bladder cancer (NMIBC) and muscle-invasive bladder cancer (MIBC). Among them, MIBC is more aggressive and exhibits a worse prognosis[ 3 ]. Despite the widespread application of surgical resection, chemotherapy, and immunotherapy, MIBC patients still exhibit a low 5-year survival rate[ 4 , 5 ]. Hence, efforts shall be made to confirm key molecular mechanisms participating in Bca development and progression to provide novel diagnostic and therapeutic strategies for patients. Recent research on the tumor microenvironment (TME) has highlighted its central role in tumor initiation, progression, and treatment response[ 6 ]. The TME encompasses tumor cells, stromal cells, immune cells, blood vessels, and the extracellular matrix (ECM). The interactions among these components dictate tumor growth dynamics and influence treatment sensitivity[ 7 , 8 ]. Among the various immune cell types, macrophages exhibit strong heterogeneity and functional diversity, thus being widely concerned. Studies have shown that tumor-associated macrophages (TAMs) fall into pro-inflammatory M1 macrophages and anti-inflammatory M2 macrophages. The latter typically promote tumor growth and metastasis by facilitating immunosuppression and angiogenesis[ 9 ]. However, the specific distribution patterns of macrophage subpopulations within different TMEs and their underlying molecular mechanisms remain unclear. Moreover, whether macrophage-related genes (MRGs) can serve as biomarkers for risk stratification and precision therapy in Bca patients requires further investigation. The discovery of single-cell RNA sequencing (scRNA-seq) has assisted in deciphering tumor heterogeneity and immune microenvironment complexity[ 10 ]. This technology enables high-throughput transcriptomic analysis of different cell populations within tumor tissues at single-cell resolution, thereby revealing cellular heterogeneity and subpopulation-specific gene expression patterns. This study gives a systematic examination of the heterogeneity of macrophage subpopulations and the functional roles of key MRGs in Bca by integrating bulk transcriptomic data from the TCGA-BLCA cohort and scRNA-seq data from the GEO database. Through scRNA-seq analysis, the study delineates the spatial distribution of macrophages within the TME and reveals their subpopulation-specific characteristics. Differential expression analysis (DEA) together with weighted gene co-expression network analysis (WGCNA) served for identifying prognostically relevant MRGs, followed by KEGG pathway enrichment analysis, which highlighted their potential involvement in key signaling pathways (Wnt, TGF-β, and PI3K/Akt). These MRGs were utilized for the construction of a survival risk prediction model, with the model’s predictive performance being evaluated for patient risk stratification and clinical prognosis. To further validate the clinical applicability of this model, immune infiltration analysis and drug sensitivity analysis were conducted for the assessment of how MRGs affected immunotherapy and chemotherapy responses. Finally, by integrating single-cell transcriptomic data, the study uncovers the cell type-specific expression patterns of ANXA1, ST3GAL5, and VIM, assisting in understanding their potential immunoregulatory roles in Bca progression from new perspectives. 2. Materials and Methods 2.1 Data Sources and Processing Figure 1 illustrates the analytical workflow. The TCGA-BLCA RNA-seq dataset, comprising 408 bladder tumor tissues and 19 adjacent non-tumor tissues, was obtained from the GDC portal of the TCGA database ( http://cancergenome.nih.gov ) using the "TCGAbiolinks" R package. The raw data were converted into Transcripts Per Million (TPM) for normalization. To ensure data balance and eliminate potential biases introduced by random matching, a chi-square test compared the two groups in terms of the clinical characteristics. Additionally, distinct cellular subpopulations within the Bca TME were deeply analyzed by virtue of scRNA-seq data from the GEO database (GSE145281) [ 11 ]. Data preprocessing relied on the Seurat R package, including quality control (QC) steps to remove low-expression cells ( 10%), and high gene count cells (> 5000 genes) to ensure data reliability. Batch effect correction was carried out using Harmony, followed by principal component analysis (PCA) for reducing the dimensionality. Subsequently, t-distributed stochastic neighbor embedding (t-SNE) together with uniform manifold approximation and projection (UMAP) served for the cell clustering analysis, allowing for the identification and spatial characterization of macrophage subpopulations within the Bca TME. This study employed the online TISCH database ( http://tisch.comp-genomics.org/home/ ) for the examination of the relationships between macrophage-associated key molecules and varying cell types within the Bca TME at the single-cell level. Subsequently, the macrophage genes extracted from the single-cell analysis, the differentially expressed genes (DEGs) identified from the TCGA-BLCA dataset, and the genes from the WGCNA modules were intersected. Thus, we obtained TCGA-BLCA macrophage-related DEGs for further analyses. 2.2 DEG Extraction and Functional Enrichment Analysis Bulk RNA-seq data were subjected to differential gene expression analysis under the assistance of the "limma" R package [ 12 ]. Genes with |logFC| > 1 and p < 0.05 were identified as DEGs. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses explored the potential roles of core genes in critical signaling pathways of Wnt, TGF-β, and PI3K/Akt by virtue of the "clusterProfiler" R package. 2.3 Proportion of Immune Cell Infiltration (ICI) in the TME CIBERSORT, as a powerful bioinformatics tool, works on predicting the proportions of 22 infiltrating immune cell types in the TME [ 13 ]. In this study, we applied CIBERSORT to the TCGA-BLCA dataset to analyze immune infiltration, allowing us to observe the ICI in Bca tissues. p < 0.05 indicated statistical significance. Next, the proportions of TAMs in above two groups of tissues were displayed by using the ggplot2 R package. Additionally, quanTIseq is another algorithm quantifying tumor immunity based on human RNA-seq data [ 14 ]. We applied the Immunedeconv R package with the quanTIseq algorithm for calculating TAM proportions in each sample. 2.4 WGCNA Analysis We employed the "WGCNA" R package for constructing a gene co-expression network [ 15 ], including genes with a median absolute deviation (MAD) higher than the top 25%, and excluding genes lacking significant expression. We built the gene matrix by computing the Pearson’s correlation coefficients, and built a highly connected, scale-free network by selecting β = 5 as an optimal soft threshold. In addition, the generated topological overlap matrix (TOM) served for estimating network connectivity. We then created a hierarchical clustering dendrogram pertaining to the TOM matrix, setting the minimum module size to 30. Highly similar modules were merged by using a merge cut height of < 0.25. Next, the quanTIseq analysis results were integrated with the module eigengenes (MEs). Gene significance (GS) was computed to reflect correlations between genes and immune cell subpopulations, whereas module membership (MM) indicated relationships between individual genes and their respective modules. Finally, genes closely related to macrophage subpopulations and patient prognosis were identified based on their GS and MM values. 2.5 Construction and Validation of a Survival Risk Prediction Model We employed LASSO Cox regression analysis for identifying macrophage-related core genes that could significantly affect Bca patients’ prognosis, followed by constructing a survival risk prediction model. Using these genes, a risk score model was built accordingly, with relevant calculating formula as follows: Risk Score=∑(βi​×expression level of genei​) Where βi is the Cox regression coefficient of gene iii. The median risk score in the training cohort was taken into account to classify patients in the validation cohort into high-risk or low-risk groups. Subsequently, we performed Kaplan–Meier (K-M) survival analysis to compare the overall survival (OS) between these two groups. To evaluate the model’s predictive accuracy at 1, 3, and 5 years, we employed receiver operating characteristic (ROC) curves changing temporally. Furthermore, the model’s independent prognostic value was confirmed via Cox proportional hazards regression, thereby supporting its use in risk stratification and personalized treatment decisions for Bca patients. 2.6 Construction of a Nomogram Model for Clinical Application Genes underwent univariate and multivariate Cox regression analyses in terms of the clinicopathological features and risk scores. The multivariate Cox model selected variables with p < 0.05p < 0.05p < 0.05 to build a nomogram to predict Bca prognosis utilizing the “rms” R package. We then generated a calibration curve for assessing the nomogram’s predictive power, with its clinical utility and reliability being evaluated via the decision curve analysis (DCA). 2.7 Correlation of Signatures, Genotypes, and the TME Related to DEGs We utilized the CIBERSORT algorithm for estimating the infiltration proportions of the 23 immune cell subpopulations in glioma, thereby evaluating the TME at the immune cell level. To further gauge inflammation within the TME, we applied a single-sample gene set enrichment analysis (ssGSEA). Besides, we worked on assessing the immune and tumor purity scores by using the “ESTIMATE” package. 2.8 Single Nucleotide Polymorphism (SNP) and Copy Number Variation (CNV) Analyses We analyzed glioma SNP data (Map Toolkit) to generate a waterfall plot for the top 20 mutated genes derived from the 10 key genes detailed in Section 2.12, thereby comparing the two groups in terms of the SNP expression. For CNV analysis, the relevant files were labeled and imported into the GenePattern software; the resulting data were subsequently visualized with the Map Toolkit. 2.9 Drug Sensitivity Analysis To assess drug sensitivity, we employed the pRRophetic algorithm, which applies a ridge regression model for the prediction of the half-maximal inhibitory concentration (IC50) based on gene expression data. The TCGA dataset served as the evaluation set, while cell line data from the Genomics of Drug Sensitivity in Cancer (GDSC; https://www.cancerrxgene.org/ ) served as the training set. By examining the correlation between mRNA expression and the IC50 values for cisplatin and other common drugs in the TCGA dataset, we adopted Spearman’s correlation for the prediction of IC50 values. 2.10 Cell culture 5637 and RT4 cells (Procell Life Science & Technology, Wuhan, China) were cultured at 37°C in DMEM (HyClone, Logan, USA) that contained 10% FBS (Invitrogen, Carlsbad, CA, USA) with 5% CO 2 . 5637 and RT4 cells were induced by concentration multiplication. The initial concentration of temozolomide was 5 µM, the final concentration was 640 µM, and each concentration was maintained for one month. Cells at the logarithmic growth stage were selected for the experimental study. 2.11 Cell transfection Lentiviral ST3GAL5-shRNA came from Genechem (Shanghai, China). shRNA sequences were: si-ST3GAL5-1: GCACTGTTTAACCTTATCA, si-ST3GAL5-2: ATAGCAGCCATGCATTGAC oe-ST3GAL5. ATGGGCCGCGGCCCCCCGCGGCGGGAGCAGGCCGCCG.5637 and RT4 cells were inoculated into 6-well plates (3-5x104 cells/ml) to undergo 16–24 h of culture at 37°C until cell fusion reached 30–50%, followed by the addition of mentivirus and Infection Enhancement Solution as per the instructions and the change of the solution after 16 hours to continue the culture. 2.12 qRT‒PCR analysis TRIzol (Invitrogen) was employed for isolating total RNA from 5637 and RT4 cells, followed by the reverse transcription of cDNA by an mRNA Reverse Transcription Kit (Roche) as per the producer’s protocol. A SYBR Green RNA Kit (Applied Biosystems, Foster City, USA) served for the quantitative real-time PCR (qRT‒PCR). The PCR cycle conditions included: 95°C for 30 s and 45 cycles at 95°C for 10 s and 60°C for 30 s. Calculation of the relative mRNA expression level relied on the 2 −ΔΔCq method. GAPDH served as a reference. 2.13 Western blot analysis We collected treated cells from each group. Two times of PBS wash were followed by the addition of phosphatase inhibitor-containing RIPA lysis buffer. The lysate underwent half an hour of lysis in an ice water bath, together with the lysis of the supernatant under an ultrasonic cell fragmentation apparatus. The supernatant underwent centrifugation before the concentration detection by the BCA method. The denatured proteins underwent SDS‒PAGE separation (loading amount of 60 µg/well), and were then moved to PVDF membranes. The collected PVDF membranes received 2 h of immersion in 5% skim milk powder, followed by one night of incubation in primary antibodies at 4°C. Then, after three times of 1×PBST wash, the membranes received 2 h of incubation in secondary antibody (Beyotime) at room temperature. An enhanced chemiluminescence kit (Beyotime, Shanghai, China) was then employed for visualizing the membranes before another three times of wash in 1× PBST. Protein band analysis relied on ImageJ software. GAPDH served as a reference. 2.14 Immunohistochemistry (IHC) Different grades of Bca specimens and normal brain tissue underwent fixation in formalin, followed by paraffin-embedding. After slicing, the 4µm sections received pretreatment (dewaxing and rehydration) for retrieving antigens in citrate buffer and underwent endogenous peroxidase quenching using 3% hydrogen peroxide (H2O2). After the blockage of nonspecific antigenic sites with 10% normal goat serum, sections underwent one night of incubation using ANXA1 antibody (1:100, Abcam),ST3GAL5 antibody (1:100, Abcam) and Vimentin antibody (1:500, CST) at 4℃, followed by certain period of incubation with secondary antibody (goat anti-rabbit IgG, 1:5000, Proteintech) and DAB and hematoxylin staining. We acquired the IHC images, and employed Image J for the calculation of the score of protein expression. 2.15 Cell viability analysis Cells plated in 96-well plates (2x10 3 cells/well) in DMEM underwent 24 and 48 h of culture after different treatments. Each well added with 10 microliters of CCK8 reagent (Glpbio, California, USA) received 1 h of incubation at 37°C. A microplate reader served for the absorbance measurement of each well at 450 nm. 2.16 Colony formation assay Suspended 5637 and RT4 cells were plated at a density of 1000 cells/6-well plate and cultured for 14 days with different treatments. After PBS wash and methanol fixation, the cells received 0.1% crystal violet staining. We counted colonies consisting of ≥ 50 cells under a microscope. 2.17 Flow Cytometry Detection of Bca Cell Cycle Cell cycle analysis relied on PI staining. After treatment, the collected cells underwent two times of PBS wash, followed by treatment with **0.25% trypsin** to generate a single-cell suspension, one night of fixation in 70% ethanol, PBS wash and half an hour of incubation in PBS solution that contained RNase (10 µg/mL) at 37°C in succession. Next, cells were subjected to half an hour of incubation in PI staining (50 µg/mL) at 4°C in the dark. Cell cycle distribution was measured via a flow cytometer (BD FACSCanto II), together with the determination of the proportions of cells in G0/G1, S, and G2/M phases. After analyzing flow cytometry data with FlowJo software, we plotted for the apoptosis rates and cell cycle distributions. Experiments were repeated ≥ 3 times, with results following the mean ± standard deviation (SD) format. T-test assisted in the between-group comparisons. 2.18 Invasion assay 40 µl of BD Matrigel (Corning, USA) were coated on chamber inserts for 1 h of solidification at 37°C. About 5×10 4 cells were resuspended in 500 µl FBS-free DMEM and plated into the top chamber of the insert, followed by being placed into a 24-well plate that contained 750 µl FBS-containing DMEM. 24 h of the invasion assay later, the migrated cells underwent fixation treatment in 4% paraformaldehyde, and 0.05% crystal violet staining, followed by being counted under microscopy. 2.19 In Vivo Tumor Progression Assay Using ST3GAL5-Deficient Cell Lines Two Bca cell lines with lower ST3GAL5 expression were subjected to in vivo tumor progression assay, which were harvested as well expanded to be further analyzed. Tumor cells were harvested by trypsin digestion, neutralized, and counted. PBS was used to adjust the cell concentration to the desired density. Next, the modified cells were subcutaneously implanted into 4-week-old BALB/c nude mice. Specifically, experimenters subcutaneously injected 120 µL of the prepared tumor cell suspension that contained 2 × 10^7 cells achieving exponential growth into the mouse’s right axillary region with a sterile syringe and needle. Tumor growth, overall health, and signs of distress were monitored regularly. We used calipers to measure the tumor dimensions. Equation below interprets the calculation of tumor volumes. V = 1/2×L×W 2 V: tumor volume (mm³); L: long axis of the tumor (mm); W: short axis of the tumor (mm).One month after implantation was followed by the euthanization of the mice, and the excision, weight, and photographing of the tumors in succession. All animal experiments obeyed the protocols of the Institutional Animal Care and Use Committee of the First Affiliated Hospital of Kunming Medical University. 2.20 Statistical Analysis All computational and statistical analyses relied on R software (version 4.1.2) and GraphPad Prism 9. The Wilcoxon rank-sum test served for two-group comparison, and the Kruskal–Wallis test served for the comparison among ≥ 3 groups. The K-M method assisted in the OS analysis, with log-rank tests evaluating statistical significance. Results of LASSO and univariate/multivariate Cox regression analyses were taken into account for constructing a prognostic model for Bca patients. Correlation assessment among key genes depended on the Pearson’s correlation. P < 0.05 denoted statistical significance. 3. Result 3.1 Identification of Macrophage-Associated Core Genes and Key Signaling Pathways in Bca The study retrieved transcriptomic data and clinical information from 408 Bca patients in the TCGA-BLCA cohort. Among the 4,271 identified DEGs, 1,614 presented upregulation and 2,657 presented downregulation. Their expression patterns were visualized via heatmap and volcano plot (Fig. 2 . A-B; Supplementary Table 1). With the objective of examining the roles of MRGs in Bca progression, we compared tumor and normal tissues in terms of their ICI utilizing the CIBERSORT algorithm, focusing on macrophage infiltration. We then conducted WGCNA, selecting an optimal soft threshold (Fig. 2 . C) to generate a scale-free network. Using dynamic tree cutting, we combined highly similar gene modules (Fig. 2 . D) and identified two modules (brown and turquoise) that were strongly correlated with macrophage features (Fig. 2 . E). From these modules, 2,474 MRGs were extracted (Supplementary Table 2). Next, we analyzed the GSE145281 single-cell RNA-sequencing dataset, identifying nine distinct cell types following log normalization and dimensionality reduction (Fig. 2 . F), yielding 761 macrophage-related key genes (Supplementary Table 3). Finally, by intersecting three gene sets—DEGs (DIFF), WGCNA module genes, and single-cell analysis genes—we pinpointed 11 overlapping genes, which may serve as core functional markers of macrophage subpopulations (Fig. 2 . G). GO enrichment analysis ascertained the participation of the DEGs in muscle contraction, ECM remodeling, collagen formation, and regulation at the biological process (BP) level. Hence, macrophages could crucially affect the tissue remodeling, cell migration, and intercellular communication. At the cellular component (CC) level, genes were highly enriched in the muscle contractile complex, ECM, and collagen triple helix, indicating their importance in structural maintenance. At the molecular function (MF) level, the DEGs participated in ligand binding, ion channel activity, and receptor interactions (Fig. 3 . A–D). In parallel, in KEGG pathway analysis, these genes clustered in calcium signaling pathways, cell adhesion molecules, neurotransmitter ligand–receptor interactions, and cardiomyopathy-related pathways, implying pivotal functions in signal transduction, cellular interactions, and tissue-specific processes. Notably, the highest enrichment scores (in muscle cell skeleton, calcium signaling pathway, and cell adhesion molecules) highlight the fundamental involvement of macrophages in controlling migration, signal coordination, and tissue remodeling within the TME (Fig. 3 . E-F). 3.2 Construction and Validation of an MRG-Based Prognostic Risk Model Using LASSO Cox regression, we identified differentially expressed MRGs with strong predictive capacity for Bca prognosis and built a risk scoring model that we comprehensively validated in the entire cohort, as well as in separate training and testing subsets. During model construction, key feature genes were selected via LASSO to ensure both simplicity and high efficiency (Fig. 4 . A-B). According to K-M survival analysis, the high-risk group presented remarkably worse OS versus the low-risk group, regardless of whether we considered all patients (ALL; Fig. 4 . C), the training set ( Fig. 4 . H), or the testing set (Fig. 4 . M). On these accounts, the risk scoring model, constructed from differentially expressed MRGs, excels in well distinguishing the two risk groups. The risk distribution plots further underscored this point: the high-risk group consistently presented higher mortality rates in the entire cohort (Fig. 4 . D), the training subset (Fig. 4 . I), and the testing subset (Fig. 4 . N). Survival status scatter plots demonstrated a clear positive correlation between increased risk scores and mortality, with more frequent deaths clustered among high-risk individuals. Additionally, heatmaps depicting the expression of core genes showed a pronounced difference in gene expression patterns between two risk groups, reflecting distinct tumor biology and hinting at potentially valuable biomarkers for predicting outcomes and guiding treatment (Fig. 4 . G,L,Q). Finally, the model achieved area under the curve (AUC) values between 0.60 and 0.70 at 1, 3, and 5 years in all three cohorts in the ROC curve analysis (ALL, Train, and Test) (Fig. 4 . F,K,P), demonstrating its favorable classification performance, robust generalizability, and importance for the short- and long-term prognostic assessments. Univariate and multivariate Cox regression analyses worked on more deeply evaluating the MCG-based risk score in terms of the independent prognostic performance and comprehensive accuracy. Risk score could be used for independently predicting OS, exhibiting the highest hazard ratio (HR = 3.045, 95% CI: 1.594–5.817) among all variables in the univariate analysis, including stage, T stage, and N stage. In multivariate analysis, with other clinical characteristics (age, sex, and tumor stage) being adjusted, the risk score still showed the best performance in independently predicting disease prognosis (HR = 2.529, 95% CI: 1.243–5.144, p = 0.01), underscoring its robustness (Fig. 5 . A-B). Comparisons via ROC curves likewise ascertained the better performance of the risk score (AUC = 0.682) versus conventional clinical features of age (AUC = 0.559), sex (AUC = 0.518), and tumor stage (AUC = 0.605) (Fig. 5 . C). In a nomogram that integrated risk score, tumor stage, age, sex, etc., the risk score contributed substantially to the total prognostic score (Fig. 5 . D). The 1-, 3-, and 5-year survival calibration curves exhibited outstanding concordance between predicted and observed outcomes (Fig. 5 . E). Furthermore, cumulative hazard plots showed a markedly higher risk accumulation over time among high-risk patients, validating the model’s discriminative power (Fig. 5 . F). Lastly, as evidenced by the C-index curves, used for dynamic comparison with traditional clinical characteristics, the risk score consistently outperformed age, sex, and tumor stage in predicting long-term survival (Fig. 5 . G). Taken together, these results establish the MCG-based risk model as a highly effective prognostic tool, offering valuable guidance for patient stratification and personalized treatment in Bca. 3.3 Association Between the MRG-Based Risk Score and ICI in the TME ICI comparison revealed significantly higher levels of immunosuppressive cells (e.g., neutrophils and M2 macrophages) in the high-risk group, and obviously lower level of effector immune cells (e.g., activated CD8^+ T cells and activated NK cells) in the low-risk group (Fig. 6 . A). Hence, high-risk patients’ immune microenvironment exhibits a more pronounced immunosuppressive profile, which could impair antitumor immune responses. In subsequent heatmap analysis, the risk score exhibited a positive relevance to the infiltration of immunosuppressive cell types (neutrophils, M2 macrophages, and resting dendritic cells), but a negative relevance to the infiltration of effector immune cells (activated CD8^+ T cells, activated NK cells, and CD4^+ memory T cells) (Fig. 6 . B). This distribution pattern highlights the immunosuppressive features of TME in the high-risk group and underscores the predictive utility of risk score for immune status. Further correlation analyses via scatter plots reinforced these observations, with infiltration by M2 macrophages (R = 0.39, p = 2.3e − 08) and neutrophils (R = 0.23, p = 0.007) showing significant positive correlations with the risk score. In contrast, activated CD8^+ T cells (R = − 0.25, p = 0.0004) and other effector cells demonstrated a negative relevance to the risk score (Fig. 6 . C). Therefore, high-risk patients not only exhibit elevated levels of immunosuppressive cells but also display reduced proportions of effector immune cells, further indicating a potent immunosuppressive microenvironment. Lastly, according to violin plots comparing the risk score to key TME parameters—StromalScore, ImmuneScore, and ESTIMATEScore, high-risk patients presented dramatically higher StromalScore and ImmuneScore versus their low-risk counterparts (p < 0.001), indicating an increase in both stromal components and immunosuppressive cell infiltration (Fig. 6 . D). This integrated set of results provides robust evidence of a significantly immunosuppressive TME in the high-risk cohort, indicative of the enhancement of immune evasion through the inhibition of effector immune cells’ function. 3.4 Association of Risk Score, Tumor Mutation Burden (TMB), and Patient Survival To investigate how the gene mutation profile and TMB differ between the two risk groups, as well as further examine their associations with patient prognosis, we generated mutation frequency (MF) plots for both groups (Supplementary Fig. 1. A-B). Genes exhibiting the largest MF in the high-risk group were TP53, TTN, MUC16, PIK3CA, and KMT2D, with TP53 mutations occurring in 45% of patients, notably higher than the 40% observed in the low-risk group. That is to say, TP53 may affect the tumor malignant progression in high-risk patients. Additionally, genes such as MUC16 and KMT2D presented an obviously higher MF in the high-risk group, indicating a possible link between these mutations and increased tumor proliferation, metastasis, and immune evasion. Collectively, the molecular features of high-risk tumors may confer greater aggressiveness and adaptability. We next explored how TMB levels varied between the two risk cohorts. Although the high-risk group exhibited a marginally higher TMB level, the difference showed only borderline significance (p = 0.068; Supplementary Fig. 1. C). This indicates that while high-risk patients may have a heavier mutational burden, other molecular factors might also influence tumor behavior. Notably, elevated TMB is frequently associated with genomic instability, which can drive TME variation and bolster the tumor’s capacity for immune evasion. According to K-M curves that more deeply illustrate the TMB-OS relationship, low-TMB patients presented a remarkably worse OS versus low-TMB patients (p < 0.001; Supplementary Fig. 1. D), suggesting that TMB can be used to independently predict the disease prognosis. The poorer outcomes in high-TMB patients could reflect a heightened complexity and adaptability of tumor cells. Combining TMB and the MCG-based risk score, we subdivided patients into four groups: High TMB + High Risk, High TMB + Low Risk, Low TMB + High Risk, and Low TMB + Low Risk. As shown in Supplementary Fig. 1. E, the High TMB + High Risk group and the Low TMB + Low Risk group had the worst and the best OS, respectively (p < 0.001). These findings not only underscore the complementary prognostic value of TMB and the MCG-based risk score but also demonstrate that integrating both indicators significantly enhances the accuracy of patient stratification for survival outcomes. 3.5 Correlation between the MRG-Based Risk Score and Drug Sensitivity The study focused on evaluating the correlation between core genes in the prognostic risk model and drug sensitivity data from the GDSC, and the drug response difference between two risk groups, aiming at more deeply elucidating the potential utility of the model in guiding precision oncology. According to the correlation analyses, genes such as ANXA1 , ST3GAL5 , and VIM may play pivotal roles in modulating therapeutic responses (Supplementary Fig. 2). Notably, ANXA1 overexpression was associated with increased sensitivity to several agents, including PI3K inhibitors and DNA damage repair inhibitors, whereas elevated VIM levels correlated with resistance to certain antimetabolites. These observations suggest that these core genes may influence drug response by regulating specific signaling pathways, thereby providing a molecular basis for more effective anticancer therapies. Further comparisons between the two risk cohorts in the drug sensitivity indicated significant distinctions in their responses to multiple anticancer agents, underscoring the practical relevance of the MCG-based risk score for clinical decision-making (Supplementary Fig. 2–3). For instance, high-risk patients displayed greater sensitivity to targeted therapies such as EGFR inhibitors (e.g., gefitinib) and HER2 inhibitors (e.g., lapatinib), possibly reflecting increased pathway activity that renders them more responsive to these drugs. Conversely, high-risk patients exhibited reduced sensitivity to conventional chemotherapeutic agents (e.g., temozolomide and gemcitabine), suggesting a higher likelihood of chemoresistance in this subgroup. These disparities may be linked to the distinct molecular and microenvironmental characteristics of high-risk tumors, thereby informing potential optimization strategies for chemotherapy. Taken together, these findings hold promising clinical implications. First, the expression of core genes (e.g., ANXA1 and VIM ) may serve as predictive biomarkers for tumor drug sensitivity, informing individualized treatment decisions. Second, the robust differences in the drug sensitivity underscore the potential of the risk score as an adjunct tool for therapeutic selection. For example, high-risk patients may derive greater benefit from targeted therapies (e.g., EGFR and HER2 inhibitors), whereas low-risk patients might respond more favorably to traditional chemotherapy. Moreover, prioritizing targeted therapies could help overcome resistance in high-risk tumors, ultimately enhancing treatment outcomes. 3.6 Single-Cell Analysis Reveals Expression Patterns of Macrophage-Associated Hub Genes in Bca To more deeply reveal the roles of the model’s core genes ( ANXA1 , ST3GAL5 , and VIM ) within the Bca TEM, we carried out an integrated single-cell RNA-sequencing analysis using the GSE149652 dataset. This approach enabled us to delineate the cellular heterogeneity within tumor tissues and examine the spatial distribution and density of diverse immune cell populations (CD4^+ T cells, CD8^+ T cells, exhausted T cells (CD8Tex), NK cells, proliferative T cells (Tprolif), and regulatory T cells (Treg cell)) (Fig. 7 . A). All these highlight the functional specialization of each cell subset in tumor progression and immune modulation. We next investigated the expression profiles of ANXA1 , ST3GAL5 , and VIM across different immune cell populations (Fig. 7 . B–D). ANXA1 was highly expressed in CD8^+ T cells and CD8Tex, implying a potential role in immunosuppression and T-cell exhaustion regulation. ST3GAL5 had relatively uniform expression but was slightly enriched in CD8^+ T cells, suggesting a broad involvement in immune regulation. Conversely, VIM was highly expressed in both Treg and NK cells, pointing to its contribution to immune evasion and NK cell activity modulation. Quantitative comparisons (Fig. 7 . E) reinforced these observations, underscoring the possibility that ANXA1 overexpression in CD8Tex could be a pivotal driver of tumor immune escape, while VIM upregulation in Treg cells may potentiate the immunosuppressive milieu. Finally, cell–cell communication analyses underscored the intricate signaling networks between CD8^+ T cells and other immune cell types. Notably, communications between CD8^+ T cells, Treg, and Tprolif cells were significantly enhanced, indicating extensive signal exchange in the TME. These interactions may orchestrate tumor immune evasion and inflammation; e.g., Treg cells could suppress the antitumor functions of CD8^+ T cells, whereas T-prolif cells may bolster tumor proliferation and accelerate disease progression. 3.7 Differential Expression of ANXA1, ST3GAL5, and Waveform Proteins in Bca Cell Lines Western blotting and IHC analyses revealed obviously different expressions of ANXA1, ST3GAL5, and waveform proteins in various Bca cell lines (SVHUC1, 5637, RT4, and TCCSUP). ANXA1 and ST3GAL5 were significantly elevated in SVHUC1 cells compared to other cell lines, and surprisingly, mRNA expression data further confirmed these findings, suggesting a possible association between these proteins and Bca progression. On these accounts, ANXA1, ST3GAL5 and Vimentin may be potential biomarkers for Bca (Fig. 8 ). 3.8 The Role of ST3GAL5 Overexpression in Bca Cell Proliferation and Tumor Growth As illustrated in Fig. 9 , experimental data support the notion that ST3GAL5 overexpression markedly enhances Bca cell proliferation and tumor growth. In both 5637 and RT4 cell lines, ST3GAL5 overexpression (oe-ST3GAL5) led to significantly increased cell viability and colony formation compared to the control (NC) and siRNA- mediated knockdown (si-ST3GAL5) groups. Cell viability assays showed that the oe-ST3GAL5 group exhibited notably higher cell proliferation at 24 h and 36 h (p < 0.01). Moreover, flow cytometry revealed that ST3GAL5 overexpression induced a greater proportion of cells in the S and G2 phases, hinting accelerated cell-cycle progression. In vivo, mice injected with 5637 cells overexpressing ST3GAL5 developed significantly larger and heavier tumors than mice in the NC group (p < 0.01). Collectively, all these highlight the key function of ST3GAL5 in bolstering Bca cell proliferation and tumor growth, confirming ST3GAL5 as a promising therapeutic target specific to Bca. 4. Discussion The present study integrated bulk transcriptomic data and scRNA-seq data to investigate the macrophage heterogeneity in the Bca TME and to characterize the functions of related genes. By identifying macrophage-related core genes and developing a survival risk prediction model, we not only demonstrated the crucial impact of macrophages on tumor progression and immunoregulation but also emphasized the potential clinical utility of these core genes in immunotherapy and precision medicine. Macrophages are key players in the TME, orchestrating immune responses and fostering tumor development. Our findings confirmed the heterogeneity of macrophages in Bca tissues, showing that different subpopulations (e.g., M1- and M2-like macrophages) exhibit distinct spatial distributions and functional properties. Notably, high-risk patients presented obviously abundant M2-like macrophages, in line with their known roles in promoting immunosuppression and tumor progression [ 16 , 17 ]. DEA further revealed that core genes such as ANXA1 , ST3GAL5 , and VIM are differentially expressed across macrophage subsets, possibly facilitating immune escape and enhancing tumor aggressiveness via pathways like Wnt and TGF-β[ 18 , 19 ]. The enrichment of M2-like macrophages alongside neutrophils in high-risk patients supports a highly immunosuppressive microenvironment, whereas a marked decrease in CD8^+ T cells and NK cells could weaken antitumor immunity [ 20 ]. Future studies should elucidate the mechanistic interplay between M2-like macrophages and other immune cells, and define how these core genes modulate the TME. Through the identification of macrophage-related core genes and subsequent model construction, we developed a robust prognostic tool for Bca. The risk prediction model achieved an AUC of 0.682, indicating satisfactory accuracy in prognostic stratification. Consistently, high-risk patients exhibited an obviously worse OS versus low-risk patients. One advantage of our model is that it integrates MRG expression with clinical features to offer both refined risk stratification and actionable information for individualized treatment decisions. Additionally, each of the core genes in this model appears to have clinical relevance. For example, ANXA1 was highly expressed in CD8^+ T cells and CD8Tex, hinting its immunosuppressive action [ 21 ], whereas VIM exhibited elevated expression in Treg cells, potentially contributing to immune evasion [ 22 ]. Targeting these molecules could open new therapeutic avenues that may improve outcomes for high-risk patients. We also performed drug sensitivity analyses to explore how these core genes influence treatment responses and to assess their utility in precision oncology. Notably, high-risk patients showed greater sensitivity to targeted therapies (e.g., the EGFR inhibitor gefitinib and the HER2 inhibitor lapatinib) while demonstrating reduced sensitivity to chemotherapeutic agents such as gemcitabine (consistent with enhanced chemoresistance). These differences might reflect the higher expression of core genes in high-risk patients and their regulatory effects on key signaling pathways, a notion supported by previous reports indicating that ANXA1 and VIM modulate cell-cycle and immune signaling pathways[ 23 – 25 ]. Meanwhile, ST3GAL5 could affect metabolic pathways and thereby alter chemotherapy efficacy, aligning with observations in other malignancies[ 26 – 28 ]. Besides, the higher TMB observed in high-risk individuals points to the potential value of immunotherapeutic strategies (e.g., PD-1/PD-L1 inhibitors) for this subgroup. Integrating the risk score with TMB further improved patient stratification and offers a theoretical foundation for the design of more targeted immunotherapies. In summary, our study illuminates the functional heterogeneity of macrophages in Bca, identifies a robust macrophage-related risk model for patient stratification, and underscores the critical contributions of core genes to immunoregulation and drug responsiveness. All these make us more deeply comprehend Bca tumor biology and contribute to the formulation of innovative and more effective treatment strategies aimed at harnessing or modifying the TME. Future work, including mechanistic studies and prospective clinical trials, is warranted to validate these findings and to translate our risk model and proposed targets into routine clinical practice. Furthermore, functional assays in the present study confirmed that ST3GAL5 overexpression can substantially enhance Bca cell proliferation and tumor growth, as evidenced by increased cell viability, colony formation, and tumor volume in vivo. Given that ST3GAL5 encodes a key enzyme essential for ganglioside biosynthesis, it may influence multiple signaling pathways related to cell proliferation, migration, and interactions with the TME. Elevated ST3GAL5 expression could thus promote a more aggressive tumor phenotype through enhanced glycan-mediated cell signaling and potential modulation of ICI. All these underscore the oncogenic role of ST3GAL5 in Bca progression and indicate that targeting ST3GAL5 or its downstream pathways can be a valuable treatment option—particularly for patients whose tumors exhibit high ST3GAL5 expression. Future investigations are suggested to pay attention to delineating the exact molecular mechanisms for ST3GAL5 to drive tumor growth and determining whether inhibiting its activity could improve responses to conventional therapies or synergize with emerging immunotherapeutic and targeted approaches. Despite shedding light on the critical roles of MRGs in Bca, this study has certain limitations. First, the construction and validation of our model predominantly relied on publicly available datasets; hence, further validation in larger clinical cohorts is necessary to confirm its generalizability. Second, the exact functional mechanisms of these core genes operating within the tumor immune microenvironment remain to be elucidated through experimental research. Moreover, although our model exhibited a reasonably favorable predictive performance (AUC = 0.682), there is still room for improvement. Future studies could integrate proteomics, metabolomics, and other multi-omics data to further refine and optimize the model. Overall, this study offers new insights into stratified management and personalized treatment for Bca patients, while providing a solid theoretical basis and practical references for future research on tumor immunology. 5. Conclusion This study integrated transcriptomic and scRNA-seq data for the systematic elucidation of the functional characteristics of MRGs within the Bca TME, and the constructed survival risk model demonstrated promising predictive performance to support personalized treatment strategies. Future researches are suggested to pay more attention to exploring the molecular regulatory mechanisms of MRGs against tumor immunity and validate the model in larger clinical cohorts, thereby advancing precision medicine in Bca and other tumor immunology studies. Declarations Data Availability Statement Bladder cancer (Bca) transcriptional datasets were obtained from the TCGA and GEO databases (accessible at https://portal.gdc.cancer.gov ). Conflict of Interest The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Author Contributions Haolin Liu, Jian Hou, Yumin Wang and Yuanqi Chu collaboratively drafted the manuscript. Junxiong Li, Jingbo Qin and Pinyao Liang did the cell experiments. Peng Gu and Xiaodong Liu performed the data visualization. Guoqiang Liao, Xiangyang Wen collected resources and finished the data analysis. Xiangyang Wen and Xiaodong Liu conceived and supervised this study. All authors read and approved the final version of the manuscript. Funding This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Clinical trial number not applicable. Ethics declaration not applicable. Consent to Publish declaration not applicable. Consent to Participate declaration not applicable. Acknowledgments We gratefully acknowledge the contributions from the TCGA and GEO project, which provided valuable data and resources for this research. References Patel VA-O, Oh WK and Galsky MD. Treatment of muscle-invasive and advanced bladder cancer in 2020. Lenis AT, Lec PM, Chamie K Fau - Mshs MD and Mshs MD. Bladder Cancer: A Review. Song X, Xin S, Zhang Y, Mao J, Duan C, Cui K, Chen L, Li F, Liu Z, Wang T, Liu J, Liu X and Song W. Identification and Quantification of Iron Metabolism Landscape on Therapy and Prognosis in Bladder Cancer. Siracusano SA-O, Rizzetto R and Porcaro AB. Bladder cancer genomics. Zhang ZL, Dong P Fau - Li Y-H, Li Yh Fau - Liu Z-W, Liu Zw Fau - Yao K, Yao K Fau - Han H, Han H Fau - Qin Z-K, Qin Zk Fau - Zhou F-J and Zhou FJ. Radical cystectomy for bladder cancer: oncologic outcome in 271 Chinese patients. Jiang Y, Wang C and Zhou S. Targeting tumor microenvironment in ovarian cancer: Premise and promise. Xiao Y and Yu D. Tumor microenvironment as a therapeutic target in cancer. Huang T, Song X, Xu D, Tiek D, Goenka A, Wu B, Sastry N, Hu B and Cheng SY. Stem cell programs in cancer initiation, progression, and therapy resistance. Pan Y, Yu Y, Wang X and Zhang T. Tumor-Associated Macrophages in Tumor Immunity. Lu J, Sheng Y, Qian W, Pan M, Zhao X and Ge QA-O. scRNA-seq data analysis method to improve analysis performance. Yuen KC, Liu LF, Gupta V, Madireddi S, Keerthivasan S, Li C, Rishipathak D, Williams P, Kadel EErA-O, Koeppen H, Chen YJ, Modrusan Z, Grogan JA-O, Banchereau RA-O, Leng N, Thastrom A, Shen X, Hashimoto K, Tayama D, van der Heijden MS, Rosenberg JE, McDermott DA-O, Powles T, Hegde PS, Huseni MA-O and Mariathasan SA-OX. High systemic and tumor-associated IL-8 correlates with reduced clinical benefit of PD-L1 blockade. Ritchie ME, Phipson B, Wu D, Hu Y, Law CW, Shi W and Smyth GK. limma powers differential expression analyses for RNA-sequencing and microarray studies. Newman AM, Liu CL, Green MA-O, Gentles AA-O, Feng W, Xu Y, Hoang CD, Diehn M and Alizadeh AA-O. Robust enumeration of cell subsets from tissue expression profiles. Finotello FA-O and Trajanoski ZA-O. Quantifying tumor-infiltrating immune cells from transcriptomics data. Langfelder P and Horvath S. WGCNA: an R package for weighted correlation network analysis. Yunna C, Mengru H, Lei W and Weidong C. Macrophage M1/M2 polarization. Li M, Yang Y, Xiong L, Jiang P, Wang J and Li C. Metabolism, metabolites, and macrophages in cancer. Zhou Y, Xu J, Luo H, Meng X, Chen M and Zhu D. Wnt signaling pathway in cancer immunotherapy. Peng D, Fu M, Wang M, Wei Y and Wei X. Targeting TGF-β signal transduction for fibrosis and cancer therapy. Hu J, Zhang L, Xia H, Yan Y, Zhu X, Sun F, Sun L, Li S, Li D, Wang J, Han Y, Zhang J, Bian D, Yu H, Chen Y, Fan P, Ma Q, Jiang G, Wang C and Zhang P. Tumor microenvironment remodeling after neoadjuvant immunotherapy in non-small cell lung cancer revealed by single-cell RNA sequencing. Li L, Wang B, Zhao S, Xiong Q and Cheng A. The role of ANXA1 in the tumor microenvironment. Ridge KA-O, Eriksson JE, Pekny MA-O and Goldman RA-O. Roles of vimentin in health and disease. Shi X, Wu Y, Tang L, Ni H and Xu Y. Downregulated annexin A1 expression correlates with poor prognosis, metastasis, and immunosuppressive microenvironment in Ewing's sarcoma. Chen J, Lu T, Chen C, Zheng W, Lu L and Li N. Elevation of ANXA1 associated with potential protective mechanism against ferroptosis and immune cell infiltration in age-related macular degeneration. Ghantasala S, Pai MGJ, Biswas D, Gahoi N, Mukherjee S, Kp M, Nissa MU, Srivastava A, Epari S, Shetty P, Moiyadi A and Srivastava S. Multiple Reaction Monitoring-Based Targeted Assays for the Validation of Protein Biomarkers in Brain Tumors. Manoochehri J, Dastgheib SA, Khamirani HA-O, Mollaie M, Sharifi Z, Zoghi S, Tabei SMB, Mohammadi S, Dehghanian F, Farbod Z and Dianatpour M. A novel frameshift pathogenic variant in ST3GAL5 causing salt and pepper developmental regression syndrome (SPDRS): A case report. Heide S, Jacquemont ML, Cheillan D, Renouil M, Tallot M, Schwartz CE, Miquel J, Bintner M, Rodriguez D, Darcel F, Buratti J, Haye D, Passemard S, Gras D, Perrin L, Capri Y, Gérard B, Piton A, Keren B, Thauvin-Robinet C, Duffourd Y, Faivre L, Poe C, Pervillé A, Héron D, Thévenon J, Arnaud L, LeGuern E, La Selva L, Vetro A, Guerrini R, Nava C and Mignot C. GM3 synthase deficiency in non-Amish patients. van der Haar Àvila IA-O, Zhang TA-O, Lorrain VA-O, de Bruin F, Spreij TA-O, Nakayama HA-O, Iwabuchi KA-O, García-Vallejo JA-O, Wuhrer MA-O, van Kooyk YA-O and van Vliet SA-O. Limited impact of cancer-derived gangliosides on anti-tumor immunity in colorectal cancer. LID - 10.1093/glycob/cwae036 [doi] LID - cwae036. Additional Declarations No competing interests reported. Supplementary Files SupplementaryMateriallegends.docx SupplementaryTable1.xls SupplementaryTable2.xls SupplementaryTable3.xls SupplementaryFigure1.jpeg SupplementaryFigure2.tif SupplementaryFigure3.jpeg Cite Share Download PDF Status: Posted Version 1 posted 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-6452577","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":454087296,"identity":"9738a12b-8de1-490f-b048-d913892cc636","order_by":0,"name":"HaoLin Liu","email":"","orcid":"","institution":"West China Hospital of Sichuan University","correspondingAuthor":false,"prefix":"","firstName":"HaoLin","middleName":"","lastName":"Liu","suffix":""},{"id":454087297,"identity":"af254922-4e91-4728-9d75-ec4189c01a6e","order_by":1,"name":"Yuanqi Chu","email":"","orcid":"","institution":"XD Group Hospital,Xi'an,Shanxi Province","correspondingAuthor":false,"prefix":"","firstName":"Yuanqi","middleName":"","lastName":"Chu","suffix":""},{"id":454087298,"identity":"6980c33f-2bcc-43a4-9dd2-3d6d6a676a42","order_by":2,"name":"Jian Hou","email":"","orcid":"","institution":"First Affiliated Hospital of Kunming Medical University","correspondingAuthor":false,"prefix":"","firstName":"Jian","middleName":"","lastName":"Hou","suffix":""},{"id":454087299,"identity":"9d60d3fc-d0e2-4279-bc15-abef5db43d49","order_by":3,"name":"Yumin Wang","email":"","orcid":"","institution":"First Affiliated Hospital of Kunming Medical University","correspondingAuthor":false,"prefix":"","firstName":"Yumin","middleName":"","lastName":"Wang","suffix":""},{"id":454087300,"identity":"368007cc-2b0b-432c-bd75-7ee1572bbc54","order_by":4,"name":"Junxiong Li","email":"","orcid":"","institution":"First Affiliated Hospital of Kunming Medical University","correspondingAuthor":false,"prefix":"","firstName":"Junxiong","middleName":"","lastName":"Li","suffix":""},{"id":454087301,"identity":"d33d80fd-a588-4b40-b6f8-fda87b51e505","order_by":5,"name":"Jingbo Qin","email":"","orcid":"","institution":"First Affiliated Hospital of Kunming Medical University","correspondingAuthor":false,"prefix":"","firstName":"Jingbo","middleName":"","lastName":"Qin","suffix":""},{"id":454087302,"identity":"604553d4-87cd-48bb-9122-94da6738f38e","order_by":6,"name":"Pinyao Liang","email":"","orcid":"","institution":"First Affiliated Hospital of Kunming Medical University","correspondingAuthor":false,"prefix":"","firstName":"Pinyao","middleName":"","lastName":"Liang","suffix":""},{"id":454087303,"identity":"9969fbc2-8d5d-4e62-bd2d-730fe8aceca6","order_by":7,"name":"Guoqiang Liao","email":"","orcid":"","institution":"Shenzhen Longgang Second People's Hospital","correspondingAuthor":false,"prefix":"","firstName":"Guoqiang","middleName":"","lastName":"Liao","suffix":""},{"id":454087304,"identity":"dcf584f7-901c-4ee2-9a81-fa270a261547","order_by":8,"name":"Peng Gu","email":"","orcid":"","institution":"First Affiliated Hospital of Kunming Medical University","correspondingAuthor":false,"prefix":"","firstName":"Peng","middleName":"","lastName":"Gu","suffix":""},{"id":454087305,"identity":"28ea2037-0ba6-449f-8645-4836085b27ab","order_by":9,"name":"XiaoDong Liu","email":"","orcid":"","institution":"First Affiliated Hospital of Kunming Medical University","correspondingAuthor":false,"prefix":"","firstName":"XiaoDong","middleName":"","lastName":"Liu","suffix":""},{"id":454087306,"identity":"e1b6718d-bd66-4ff9-b192-b3b99ec9902f","order_by":10,"name":"Xiangyang Wen","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAuklEQVRIiWNgGAWjYNCCChs5Nvb2A6RoOZNmzMdzJoEEHYxthxPnSTgYEKfa4HjvMWneNub0NgmGBIYfFduI0HLmXLIxzzm23DbpxgOMPWduE9ZidiPH8DFPGU9um8yBBGbGNmK03H9jcJiHTSKdTSLBgEgtN3iAtrQZJBCvxf5MjrHhnDMJhm3AQD5IlF8k28+YSbyp+C8v395+8MGPCiK0gAATD5RxgDj1QMD4g2ilo2AUjIJRMCIBAJJIOzkpYaO7AAAAAElFTkSuQmCC","orcid":"","institution":"Shenzhen Longgang Second People's Hospital","correspondingAuthor":true,"prefix":"","firstName":"Xiangyang","middleName":"","lastName":"Wen","suffix":""}],"badges":[],"createdAt":"2025-04-15 08:23:26","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6452577/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6452577/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":82608734,"identity":"66462ac1-730b-4292-a4c6-95ffcb0640f4","added_by":"auto","created_at":"2025-05-13 10:24:22","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":2458895,"visible":true,"origin":"","legend":"\u003cp\u003eWorkflow of the Study\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-6452577/v1/9ddabcf165f01b1bfdbfaf69.png"},{"id":82608723,"identity":"e7696d8d-d2db-4ceb-a5cb-4d330a8eca3b","added_by":"auto","created_at":"2025-05-13 10:24:21","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":1019971,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eIdentification of key genes through multi-omics analysis: (A) \u003c/strong\u003eHeatmap of differentially expressed genes (DEGs) showing expression patterns across samples. \u003cstrong\u003e(B) \u003c/strong\u003eVolcano plot of DEGs, with log2 fold change (x-axis) and -log10(P-value) (y-axis); yellow and blue points indicate upregulated and downregulated genes, respectively. (\u003cstrong\u003eC\u003c/strong\u003e) Soft-threshold selection in weighted gene co-expression network analysis (WGCNA), displaying scale independence (left) and mean connectivity (right). (\u003cstrong\u003eD\u003c/strong\u003e) WGCNA gene clustering dendrogram, with different colors representing distinct co-expression modules. (\u003cstrong\u003eE\u003c/strong\u003e) t-SNE plot of single-cell RNA sequencing (scRNA-seq) data, illustrating different cell populations in reduced-dimensional space. (\u003cstrong\u003eF\u003c/strong\u003e) Correlation heatmap of WGCNA gene modules and clinical traits, where red indicates a positive correlation and blue indicates a negative correlation. (\u003cstrong\u003eG\u003c/strong\u003e) Venn diagram showing the overlap of genes identified by DEGs analysis (DIFF), WGCNA, and scRNA-seq, highlighting 11 key genes.\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-6452577/v1/7df22c677b0195d78479e14a.png"},{"id":82608724,"identity":"aeabee41-b874-43e7-9d24-44635d4edfde","added_by":"auto","created_at":"2025-05-13 10:24:22","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":2411036,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eFunctional enrichment analysis of differentially expressed genes (DEGs): (A) \u003c/strong\u003eGene Ontology (GO) enrichment analysis results, showing the top enriched biological processes (BP), cellular components (CC), and molecular functions (MF). \u003cstrong\u003e(B-D)\u003c/strong\u003eCircular diagrams representing GO term relationships for BP (B), CC (C), and MF (E), respectively. \u003cstrong\u003e(E) \u003c/strong\u003eKEGG pathway enrichment analysis, displaying significantly enriched pathways with enrichment scores and p-values. \u003cstrong\u003e(F) \u003c/strong\u003eKEGG circular diagram illustrating the relationship between enriched pathways and associated genes.\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-6452577/v1/fe1fcff258f91e79ae00ba3c.png"},{"id":82608726,"identity":"9c1fce88-06d8-40c5-8301-5171ccfa3421","added_by":"auto","created_at":"2025-05-13 10:24:22","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":987829,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eConstruction and validation of the prognostic model: (A-B)\u003c/strong\u003e LASSO Cox regression analysis for feature selection, with (A) showing the cross-validation plot for selecting the optimal lambda and (B) displaying the coefficient profiles of selected genes. \u003cstrong\u003e(C-E) \u003c/strong\u003eKaplan-Meier survival curves comparing high- and low-risk groups in the entire cohort (C), training set (D), and test set (E), with p-values indicating statistical significance. (\u003cstrong\u003eF-H\u003c/strong\u003e) Risk score distribution plots for all patients (F), training set (G), and test set (H), where high-risk patients are shown in red and low-risk patients in blue. (\u003cstrong\u003eI-K\u003c/strong\u003e) Survival status scatter plots, illustrating the distribution of deceased (red dots) and surviving patients (blue dots) across the risk score spectrum for the entire cohort (I), training set (J), and test set (K). (\u003cstrong\u003eL-N\u003c/strong\u003e) Heatmaps of prognostic gene expression patterns in the entire cohort (L), training set (M), and test set (N), with red indicating high expression and blue indicating low expression. (\u003cstrong\u003eO-Q\u003c/strong\u003e) Time-dependent ROC curves for evaluating the predictive accuracy of the model at 1, 3, and 5 years in the entire cohort (O), training set (P), and test set (Q), with AUC values indicating the model's performance.\u003c/p\u003e","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-6452577/v1/c76f24e230c996326a9e7651.png"},{"id":82608727,"identity":"f8b78472-0a91-44c5-aaa8-f3a53fcd593b","added_by":"auto","created_at":"2025-05-13 10:24:22","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":648270,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eIndependent prognostic analysis and nomogram-based survival prediction: \u003c/strong\u003e(\u003cstrong\u003eA-B\u003c/strong\u003e) Univariate (\u003cstrong\u003eA\u003c/strong\u003e) and multivariate (\u003cstrong\u003eB\u003c/strong\u003e) Cox regression analyses evaluating the independent prognostic value of risk score and clinical factors (age, gender, stage, T, M, and N classifications). Hazard ratios (HR) with 95% confidence intervals are shown. (\u003cstrong\u003eC\u003c/strong\u003e) Receiver operating characteristic (ROC) curves comparing the predictive performance of risk score, age, gender, grade, and stage, with corresponding area under the curve (AUC) values. (\u003cstrong\u003eD\u003c/strong\u003e) Nomogram integrating risk score and clinical parameters to predict 1-year, 3-year, and 5-year overall survival (OS). (\u003cstrong\u003eE\u003c/strong\u003e) Calibration curves assessing the accuracy of the nomogram-predicted OS compared to actual survival outcomes. (\u003cstrong\u003eF\u003c/strong\u003e) Decision curve analysis (DCA) evaluating the clinical utility of the risk score-based model. (\u003cstrong\u003eG\u003c/strong\u003e) Concordance index (C-index) curves for different prognostic factors over time, showing the predictive consistency of risk score, age, gender, grade, and stage.\u003c/p\u003e","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-6452577/v1/7dde0705a6105af31db80bdf.png"},{"id":82609759,"identity":"4d74a93f-adf8-417d-bc1f-baf63b2a18ee","added_by":"auto","created_at":"2025-05-13 10:32:22","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":1071633,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eImmune infiltration analysis and its correlation with risk score: \u003c/strong\u003e(\u003cstrong\u003eA\u003c/strong\u003e) Boxplot showing the distribution of immune cell infiltration levels between high-risk (orange) and low-risk (blue) groups. (\u003cstrong\u003eB\u003c/strong\u003e) Heatmap illustrating the correlation between immune cell subtypes and risk scores, with deeper red indicating stronger correlations. (\u003cstrong\u003eC\u003c/strong\u003e) Violin plot comparing different immune microenvironment scores (stromal score, immune score, and ESTIMATE score) between high-risk and low-risk groups. (\u003cstrong\u003eD\u003c/strong\u003e) Scatter plots depicting the correlation between risk score and various immune cell populations, with fitted regression lines and correlation coefficients.\u003c/p\u003e","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-6452577/v1/69f205d0fd4c9d7939770845.png"},{"id":82609754,"identity":"4a4845e1-a162-4709-8d6b-bcf770395595","added_by":"auto","created_at":"2025-05-13 10:32:22","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":1038930,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSingle-cell RNA sequencing analysis of BLCA (GSE149652 dataset): \u003c/strong\u003e(\u003cstrong\u003eA\u003c/strong\u003e) t-SNE plot showing the distribution of major immune cell lineages in the BLCA dataset, with different colors representing distinct cell types, including CD8+ T cells, CD8+ exhausted T cells, NK cells, proliferative cells, and regulatory T cells (Tregs). (\u003cstrong\u003eB-D\u003c/strong\u003e) Feature plots displaying the expression patterns of \u003cstrong\u003eST3GAL5 (\u003c/strong\u003eB\u003cstrong\u003e), VIM (\u003c/strong\u003eC\u003cstrong\u003e), and ANXA1 (\u003c/strong\u003eD\u003cstrong\u003e)\u003c/strong\u003e in the single-cell dataset, with darker colors indicating higher expression levels. (\u003cstrong\u003eE\u003c/strong\u003e) Violin plots illustrating the expression distribution of \u003cstrong\u003eANXA1, ST3GAL5, and VIM\u003c/strong\u003e across different immune cell subtypes.\u003c/p\u003e","description":"","filename":"floatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-6452577/v1/096d799dfd0444cae6077089.png"},{"id":82608737,"identity":"3023d435-fe70-41e7-a43a-c89574f20e77","added_by":"auto","created_at":"2025-05-13 10:24:22","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":5146723,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eValidation of ANXA1, ST3GAL5, and Vimentin expression in bladder cancer: \u003c/strong\u003e(\u003cstrong\u003eLeft panel\u003c/strong\u003e) Western blot analysis of ANXA1, ST3GAL5, and Vimentin protein expression in bladder cancer cell lines (SW1116, 5637, RT4, and TCCSUP) with GAPDH as a loading control. (\u003cstrong\u003eMiddle panel\u003c/strong\u003e) Quantification of relative protein levels of ANXA1, ST3GAL5, and Vimentin, along with their corresponding mRNA expression levels across different cell lines. (\u003cstrong\u003eRight panel\u003c/strong\u003e) Immunohistochemical (IHC) staining of ANXA1, ST3GAL5, and Vimentin in bladder tumor (left column) and adjacent normal tissue (right column), showing differential protein expression between cancerous and normal tissues\u003c/p\u003e","description":"","filename":"floatimage8.png","url":"https://assets-eu.researchsquare.com/files/rs-6452577/v1/2f17a880ced7e4449b92f5a0.png"},{"id":82608758,"identity":"5adc09bf-3b9c-4f22-b9c6-a21bea279584","added_by":"auto","created_at":"2025-05-13 10:24:23","extension":"png","order_by":9,"title":"Figure 9","display":"","copyAsset":false,"role":"figure","size":23328850,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eFunctional validation of ST3GAL5 in bladder cancer cells and xenograft models: \u003c/strong\u003e(\u003cstrong\u003eTop left\u003c/strong\u003e) qRT-PCR analysis of ST3GAL5 mRNA expression in bladder cancer cell lines (5637 and RT4) after knockdown (si-ST3GAL5) or overexpression (oe-ST3GAL5), compared to the negative control (NC). (\u003cstrong\u003eTop middle\u003c/strong\u003e) Cell viability assay showing the proliferation rates of different groups over time. (\u003cstrong\u003eTop right\u003c/strong\u003e) Flow cytometry analysis of cell cycle distribution in 5637 and RT4 cells under different ST3GAL5 expression conditions. (\u003cstrong\u003eMiddle left\u003c/strong\u003e) Transwell migration assay images and quantification, showing cell migration ability in different groups. (\u003cstrong\u003eMiddle right\u003c/strong\u003e) Colony formation assay results in 5637 and RT4 cells after ST3GAL5 overexpression. (\u003cstrong\u003eBottom\u003c/strong\u003e) In vivo tumorigenicity assay in a mouse xenograft model, with representative images of tumors (left), tumor volume growth curves (top right), and final tumor weight comparisons (bottom right), showing a significant difference between the NC and ST3GAL5-OE groups.\u003c/p\u003e","description":"","filename":"floatimage9.png","url":"https://assets-eu.researchsquare.com/files/rs-6452577/v1/071e095c6c9210bf9d9995fa.png"},{"id":96605301,"identity":"25b1fc09-5d6c-495f-8c30-1b30a8bd10f5","added_by":"auto","created_at":"2025-11-24 09:22:10","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":56752198,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6452577/v1/dfc6eb76-7a2e-4b1b-923f-d9f98c4eece5.pdf"},{"id":82608719,"identity":"1b4a9527-51e2-4ad2-b38d-33f846fae38f","added_by":"auto","created_at":"2025-05-13 10:24:21","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":14848,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryMateriallegends.docx","url":"https://assets-eu.researchsquare.com/files/rs-6452577/v1/fab7d2b91e57be05ac69af8a.docx"},{"id":82609753,"identity":"ed814585-c940-446b-921d-f6bcd101fdf5","added_by":"auto","created_at":"2025-05-13 10:32:22","extension":"xls","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":942080,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryTable1.xls","url":"https://assets-eu.researchsquare.com/files/rs-6452577/v1/6070326d056e8dbbbcd62f8c.xls"},{"id":82608722,"identity":"886f7937-4944-44ea-b5ad-4c34b9bc8d74","added_by":"auto","created_at":"2025-05-13 10:24:21","extension":"xls","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":60928,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryTable2.xls","url":"https://assets-eu.researchsquare.com/files/rs-6452577/v1/bcc661ed1feb62ea2f744daa.xls"},{"id":82608743,"identity":"2a13aaae-a119-4f32-9a7a-ba780bab84c1","added_by":"auto","created_at":"2025-05-13 10:24:22","extension":"xls","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":144384,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryTable3.xls","url":"https://assets-eu.researchsquare.com/files/rs-6452577/v1/0ad52f0cc665ff23ab9200b9.xls"},{"id":82608729,"identity":"26b73916-48ee-466d-9206-a9cf501ddbb5","added_by":"auto","created_at":"2025-05-13 10:24:22","extension":"jpeg","order_by":5,"title":"","display":"","copyAsset":false,"role":"supplement","size":1389852,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryFigure1.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-6452577/v1/7968a389a7dc6889f83176c6.jpeg"},{"id":82608756,"identity":"203635f3-87b6-41ef-a4f2-5bc68fadac16","added_by":"auto","created_at":"2025-05-13 10:24:23","extension":"tif","order_by":6,"title":"","display":"","copyAsset":false,"role":"supplement","size":17656276,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryFigure2.tif","url":"https://assets-eu.researchsquare.com/files/rs-6452577/v1/2b5ee48eff557bd65ee3e15e.tif"},{"id":82608736,"identity":"3688be84-1e4a-4d32-ba08-7fba768bbd2b","added_by":"auto","created_at":"2025-05-13 10:24:22","extension":"jpeg","order_by":7,"title":"","display":"","copyAsset":false,"role":"supplement","size":603742,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryFigure3.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-6452577/v1/497fac466180063f9d846296.jpeg"}],"financialInterests":"No competing interests reported.","formattedTitle":"Targeting Macrophage-Associated Core Genes for Prognostic Prediction and Therapeutic Insights in Bladder Cancer","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eBladder cancer (Bca) is a representative and lethal malignancy of the urinary system worldwide, with over 170,000 related deaths annually. It ranks 4th among all cancers for male and is one of the most frequently diagnosed malignancies for female [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. Based on pathological characteristics, bladder cancer (BC) falls into non-muscle-invasive bladder cancer (NMIBC) and muscle-invasive bladder cancer (MIBC). Among them, MIBC is more aggressive and exhibits a worse prognosis[\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. Despite the widespread application of surgical resection, chemotherapy, and immunotherapy, MIBC patients still exhibit a low 5-year survival rate[\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]. Hence, efforts shall be made to confirm key molecular mechanisms participating in Bca development and progression to provide novel diagnostic and therapeutic strategies for patients.\u003c/p\u003e \u003cp\u003eRecent research on the tumor microenvironment (TME) has highlighted its central role in tumor initiation, progression, and treatment response[\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]. The TME encompasses tumor cells, stromal cells, immune cells, blood vessels, and the extracellular matrix (ECM). The interactions among these components dictate tumor growth dynamics and influence treatment sensitivity[\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e, \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]. Among the various immune cell types, macrophages exhibit strong heterogeneity and functional diversity, thus being widely concerned. Studies have shown that tumor-associated macrophages (TAMs) fall into pro-inflammatory M1 macrophages and anti-inflammatory M2 macrophages. The latter typically promote tumor growth and metastasis by facilitating immunosuppression and angiogenesis[\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]. However, the specific distribution patterns of macrophage subpopulations within different TMEs and their underlying molecular mechanisms remain unclear. Moreover, whether macrophage-related genes (MRGs) can serve as biomarkers for risk stratification and precision therapy in Bca patients requires further investigation. The discovery of single-cell RNA sequencing (scRNA-seq) has assisted in deciphering tumor heterogeneity and immune microenvironment complexity[\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. This technology enables high-throughput transcriptomic analysis of different cell populations within tumor tissues at single-cell resolution, thereby revealing cellular heterogeneity and subpopulation-specific gene expression patterns.\u003c/p\u003e \u003cp\u003eThis study gives a systematic examination of the heterogeneity of macrophage subpopulations and the functional roles of key MRGs in Bca by integrating bulk transcriptomic data from the TCGA-BLCA cohort and scRNA-seq data from the GEO database. Through scRNA-seq analysis, the study delineates the spatial distribution of macrophages within the TME and reveals their subpopulation-specific characteristics. Differential expression analysis (DEA) together with weighted gene co-expression network analysis (WGCNA) served for identifying prognostically relevant MRGs, followed by KEGG pathway enrichment analysis, which highlighted their potential involvement in key signaling pathways (Wnt, TGF-β, and PI3K/Akt). These MRGs were utilized for the construction of a survival risk prediction model, with the model\u0026rsquo;s predictive performance being evaluated for patient risk stratification and clinical prognosis. To further validate the clinical applicability of this model, immune infiltration analysis and drug sensitivity analysis were conducted for the assessment of how MRGs affected immunotherapy and chemotherapy responses. Finally, by integrating single-cell transcriptomic data, the study uncovers the cell type-specific expression patterns of ANXA1, ST3GAL5, and VIM, assisting in understanding their potential immunoregulatory roles in Bca progression from new perspectives.\u003c/p\u003e"},{"header":"2. Materials and Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003e2.1 Data Sources and Processing\u003c/h2\u003e \u003cp\u003eFigure \u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e illustrates the analytical workflow. The TCGA-BLCA RNA-seq dataset, comprising 408 bladder tumor tissues and 19 adjacent non-tumor tissues, was obtained from the GDC portal of the TCGA database (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://cancergenome.nih.gov\u003c/span\u003e\u003cspan address=\"http://cancergenome.nih.gov\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) using the \"TCGAbiolinks\" R package. The raw data were converted into Transcripts Per Million (TPM) for normalization. To ensure data balance and eliminate potential biases introduced by random matching, a chi-square test compared the two groups in terms of the clinical characteristics. Additionally, distinct cellular subpopulations within the Bca TME were deeply analyzed by virtue of scRNA-seq data from the GEO database (GSE145281) [\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e]. Data preprocessing relied on the Seurat R package, including quality control (QC) steps to remove low-expression cells (\u0026lt;\u0026thinsp;200 genes), high mitochondrial gene expression cells (\u0026gt;\u0026thinsp;10%), and high gene count cells (\u0026gt;\u0026thinsp;5000 genes) to ensure data reliability. Batch effect correction was carried out using Harmony, followed by principal component analysis (PCA) for reducing the dimensionality. Subsequently, t-distributed stochastic neighbor embedding (t-SNE) together with uniform manifold approximation and projection (UMAP) served for the cell clustering analysis, allowing for the identification and spatial characterization of macrophage subpopulations within the Bca TME.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThis study employed the online TISCH database (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://tisch.comp-genomics.org/home/\u003c/span\u003e\u003cspan address=\"http://tisch.comp-genomics.org/home/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) for the examination of the relationships between macrophage-associated key molecules and varying cell types within the Bca TME at the single-cell level. Subsequently, the macrophage genes extracted from the single-cell analysis, the differentially expressed genes (DEGs) identified from the TCGA-BLCA dataset, and the genes from the WGCNA modules were intersected. Thus, we obtained TCGA-BLCA macrophage-related DEGs for further analyses.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003e2.2 DEG Extraction and Functional Enrichment Analysis\u003c/h2\u003e \u003cp\u003eBulk RNA-seq data were subjected to differential gene expression analysis under the assistance of the \"limma\" R package [\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]. Genes with |logFC| \u0026gt; 1 and p\u0026thinsp;\u0026lt;\u0026thinsp;0.05 were identified as DEGs. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses explored the potential roles of core genes in critical signaling pathways of Wnt, TGF-β, and PI3K/Akt by virtue of the \"clusterProfiler\" R package.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003e2.3 Proportion of Immune Cell Infiltration (ICI) in the TME\u003c/h2\u003e \u003cp\u003eCIBERSORT, as a powerful bioinformatics tool, works on predicting the proportions of 22 infiltrating immune cell types in the TME [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]. In this study, we applied CIBERSORT to the TCGA-BLCA dataset to analyze immune infiltration, allowing us to observe the ICI in Bca tissues. p\u0026thinsp;\u0026lt;\u0026thinsp;0.05 indicated statistical significance. Next, the proportions of TAMs in above two groups of tissues were displayed by using the ggplot2 R package. Additionally, quanTIseq is another algorithm quantifying tumor immunity based on human RNA-seq data [\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]. We applied the Immunedeconv R package with the quanTIseq algorithm for calculating TAM proportions in each sample.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003e2.4 WGCNA Analysis\u003c/h2\u003e \u003cp\u003eWe employed the \"WGCNA\" R package for constructing a gene co-expression network [\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e], including genes with a median absolute deviation (MAD) higher than the top 25%, and excluding genes lacking significant expression. We built the gene matrix by computing the Pearson\u0026rsquo;s correlation coefficients, and built a highly connected, scale-free network by selecting β\u0026thinsp;=\u0026thinsp;5 as an optimal soft threshold. In addition, the generated topological overlap matrix (TOM) served for estimating network connectivity. We then created a hierarchical clustering dendrogram pertaining to the TOM matrix, setting the minimum module size to 30. Highly similar modules were merged by using a merge cut height of \u0026lt;\u0026thinsp;0.25. Next, the quanTIseq analysis results were integrated with the module eigengenes (MEs). Gene significance (GS) was computed to reflect correlations between genes and immune cell subpopulations, whereas module membership (MM) indicated relationships between individual genes and their respective modules. Finally, genes closely related to macrophage subpopulations and patient prognosis were identified based on their GS and MM values.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003e2.5 Construction and Validation of a Survival Risk Prediction Model\u003c/h2\u003e \u003cp\u003eWe employed LASSO Cox regression analysis for identifying macrophage-related core genes that could significantly affect Bca patients\u0026rsquo; prognosis, followed by constructing a survival risk prediction model. Using these genes, a risk score model was built accordingly, with relevant calculating formula as follows:\u003c/p\u003e \u003cp\u003eRisk Score=\u0026sum;(βi​\u0026times;expression level of genei​)\u003c/p\u003e \u003cp\u003eWhere βi is the Cox regression coefficient of gene iii. The median risk score in the training cohort was taken into account to classify patients in the validation cohort into high-risk or low-risk groups. Subsequently, we performed Kaplan\u0026ndash;Meier (K-M) survival analysis to compare the overall survival (OS) between these two groups. To evaluate the model\u0026rsquo;s predictive accuracy at 1, 3, and 5 years, we employed receiver operating characteristic (ROC) curves changing temporally. Furthermore, the model\u0026rsquo;s independent prognostic value was confirmed via Cox proportional hazards regression, thereby supporting its use in risk stratification and personalized treatment decisions for Bca patients.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003e2.6 Construction of a Nomogram Model for Clinical Application\u003c/h2\u003e \u003cp\u003eGenes underwent univariate and multivariate Cox regression analyses in terms of the clinicopathological features and risk scores. The multivariate Cox model selected variables with p\u0026thinsp;\u0026lt;\u0026thinsp;0.05p\u0026thinsp;\u0026lt;\u0026thinsp;0.05p\u0026thinsp;\u0026lt;\u0026thinsp;0.05 to build a nomogram to predict Bca prognosis utilizing the \u0026ldquo;rms\u0026rdquo; R package. We then generated a calibration curve for assessing the nomogram\u0026rsquo;s predictive power, with its clinical utility and reliability being evaluated via the decision curve analysis (DCA).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003e2.7 Correlation of Signatures, Genotypes, and the TME Related to DEGs\u003c/h2\u003e \u003cp\u003eWe utilized the CIBERSORT algorithm for estimating the infiltration proportions of the 23 immune cell subpopulations in glioma, thereby evaluating the TME at the immune cell level. To further gauge inflammation within the TME, we applied a single-sample gene set enrichment analysis (ssGSEA). Besides, we worked on assessing the immune and tumor purity scores by using the \u0026ldquo;ESTIMATE\u0026rdquo; package.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec10\" class=\"Section2\"\u003e \u003ch2\u003e2.8 Single Nucleotide Polymorphism (SNP) and Copy Number Variation (CNV) Analyses\u003c/h2\u003e \u003cp\u003eWe analyzed glioma SNP data (Map Toolkit) to generate a waterfall plot for the top 20 mutated genes derived from the 10 key genes detailed in Section 2.12, thereby comparing the two groups in terms of the SNP expression. For CNV analysis, the relevant files were labeled and imported into the GenePattern software; the resulting data were subsequently visualized with the Map Toolkit.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003e2.9 Drug Sensitivity Analysis\u003c/h2\u003e \u003cp\u003eTo assess drug sensitivity, we employed the pRRophetic algorithm, which applies a ridge regression model for the prediction of the half-maximal inhibitory concentration (IC50) based on gene expression data. The TCGA dataset served as the evaluation set, while cell line data from the Genomics of Drug Sensitivity in Cancer (GDSC; \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) served as the training set. By examining the correlation between mRNA expression and the IC50 values for cisplatin and other common drugs in the TCGA dataset, we adopted Spearman\u0026rsquo;s correlation for the prediction of IC50 values.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003e2.10 Cell culture\u003c/h2\u003e \u003cp\u003e5637 and RT4 cells (Procell Life Science \u0026amp; Technology, Wuhan, China) were cultured at 37\u0026deg;C in DMEM (HyClone, Logan, USA) that contained 10% FBS (Invitrogen, Carlsbad, CA, USA) with 5% CO\u003csub\u003e2\u003c/sub\u003e. 5637 and RT4 cells were induced by concentration multiplication. The initial concentration of temozolomide was 5 \u0026micro;M, the final concentration was 640 \u0026micro;M, and each concentration was maintained for one month. Cells at the logarithmic growth stage were selected for the experimental study.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003e2.11 Cell transfection\u003c/h2\u003e \u003cp\u003eLentiviral ST3GAL5-shRNA came from Genechem (Shanghai, China). shRNA sequences were: si-ST3GAL5-1: GCACTGTTTAACCTTATCA, si-ST3GAL5-2: ATAGCAGCCATGCATTGAC oe-ST3GAL5. ATGGGCCGCGGCCCCCCGCGGCGGGAGCAGGCCGCCG.5637 and RT4 cells were inoculated into 6-well plates (3-5x104 cells/ml) to undergo 16\u0026ndash;24 h of culture at 37\u0026deg;C until cell fusion reached 30\u0026ndash;50%, followed by the addition of mentivirus and Infection Enhancement Solution as per the instructions and the change of the solution after 16 hours to continue the culture.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003e2.12 qRT‒PCR analysis\u003c/h2\u003e \u003cp\u003eTRIzol (Invitrogen) was employed for isolating total RNA from 5637 and RT4 cells, followed by the reverse transcription of cDNA by an mRNA Reverse Transcription Kit (Roche) as per the producer\u0026rsquo;s protocol. A SYBR Green RNA Kit (Applied Biosystems, Foster City, USA) served for the quantitative real-time PCR (qRT‒PCR). The PCR cycle conditions included: 95\u0026deg;C for 30 s and 45 cycles at 95\u0026deg;C for 10 s and 60\u0026deg;C for 30 s. Calculation of the relative mRNA expression level relied on the 2\u003csup\u003e\u0026minus;ΔΔCq\u003c/sup\u003e method. GAPDH served as a reference.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003e2.13 Western blot analysis\u003c/h2\u003e \u003cp\u003eWe collected treated cells from each group. Two times of PBS wash were followed by the addition of phosphatase inhibitor-containing RIPA lysis buffer. The lysate underwent half an hour of lysis in an ice water bath, together with the lysis of the supernatant under an ultrasonic cell fragmentation apparatus. The supernatant underwent centrifugation before the concentration detection by the BCA method. The denatured proteins underwent SDS‒PAGE separation (loading amount of 60 \u0026micro;g/well), and were then moved to PVDF membranes. The collected PVDF membranes received 2 h of immersion in 5% skim milk powder, followed by one night of incubation in primary antibodies at 4\u0026deg;C. Then, after three times of 1\u0026times;PBST wash, the membranes received 2 h of incubation in secondary antibody (Beyotime) at room temperature. An enhanced chemiluminescence kit (Beyotime, Shanghai, China) was then employed for visualizing the membranes before another three times of wash in 1\u0026times; PBST. Protein band analysis relied on ImageJ software. GAPDH served as a reference.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003e2.14 Immunohistochemistry (IHC)\u003c/h2\u003e \u003cp\u003eDifferent grades of Bca specimens and normal brain tissue underwent fixation in formalin, followed by paraffin-embedding. After slicing, the 4\u0026micro;m sections received pretreatment (dewaxing and rehydration) for retrieving antigens in citrate buffer and underwent endogenous peroxidase quenching using 3% hydrogen peroxide (H2O2). After the blockage of nonspecific antigenic sites with 10% normal goat serum, sections underwent one night of incubation using ANXA1 antibody (1:100, Abcam),ST3GAL5 antibody (1:100, Abcam) and Vimentin antibody (1:500, CST) at 4℃, followed by certain period of incubation with secondary antibody (goat anti-rabbit IgG, 1:5000, Proteintech) and DAB and hematoxylin staining. We acquired the IHC images, and employed Image J for the calculation of the score of protein expression.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec17\" class=\"Section2\"\u003e \u003ch2\u003e2.15 Cell viability analysis\u003c/h2\u003e \u003cp\u003eCells plated in 96-well plates (2x10\u003csup\u003e3\u003c/sup\u003e cells/well) in DMEM underwent 24 and 48 h of culture after different treatments. Each well added with 10 microliters of CCK8 reagent (Glpbio, California, USA) received 1 h of incubation at 37\u0026deg;C. A microplate reader served for the absorbance measurement of each well at 450 nm.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec18\" class=\"Section2\"\u003e \u003ch2\u003e2.16 Colony formation assay\u003c/h2\u003e \u003cp\u003eSuspended 5637 and RT4 cells were plated at a density of 1000 cells/6-well plate and cultured for 14 days with different treatments. After PBS wash and methanol fixation, the cells received 0.1% crystal violet staining. We counted colonies consisting of \u0026ge;\u0026thinsp;50 cells under a microscope.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec19\" class=\"Section2\"\u003e \u003ch2\u003e2.17 Flow Cytometry Detection of Bca Cell Cycle\u003c/h2\u003e \u003cp\u003eCell cycle analysis relied on PI staining. After treatment, the collected cells underwent two times of PBS wash, followed by treatment with **0.25% trypsin** to generate a single-cell suspension, one night of fixation in 70% ethanol, PBS wash and half an hour of incubation in PBS solution that contained RNase (10 \u0026micro;g/mL) at 37\u0026deg;C in succession. Next, cells were subjected to half an hour of incubation in PI staining (50 \u0026micro;g/mL) at 4\u0026deg;C in the dark. Cell cycle distribution was measured via a flow cytometer (BD FACSCanto II), together with the determination of the proportions of cells in G0/G1, S, and G2/M phases. After analyzing flow cytometry data with FlowJo software, we plotted for the apoptosis rates and cell cycle distributions. Experiments were repeated\u0026thinsp;\u0026ge;\u0026thinsp;3 times, with results following the mean\u0026thinsp;\u0026plusmn;\u0026thinsp;standard deviation (SD) format. T-test assisted in the between-group comparisons.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec20\" class=\"Section2\"\u003e \u003ch2\u003e2.18 Invasion assay\u003c/h2\u003e \u003cp\u003e40 \u0026micro;l of BD Matrigel (Corning, USA) were coated on chamber inserts for 1 h of solidification at 37\u0026deg;C. About 5\u0026times;10\u003csup\u003e4\u003c/sup\u003e cells were resuspended in 500 \u0026micro;l FBS-free DMEM and plated into the top chamber of the insert, followed by being placed into a 24-well plate that contained 750 \u0026micro;l FBS-containing DMEM. 24 h of the invasion assay later, the migrated cells underwent fixation treatment in 4% paraformaldehyde, and 0.05% crystal violet staining, followed by being counted under microscopy.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec21\" class=\"Section2\"\u003e \u003ch2\u003e2.19 In Vivo Tumor Progression Assay Using ST3GAL5-Deficient Cell Lines\u003c/h2\u003e \u003cp\u003eTwo Bca cell lines with lower ST3GAL5 expression were subjected to in vivo tumor progression assay, which were harvested as well expanded to be further analyzed. Tumor cells were harvested by trypsin digestion, neutralized, and counted. PBS was used to adjust the cell concentration to the desired density. Next, the modified cells were subcutaneously implanted into 4-week-old BALB/c nude mice. Specifically, experimenters subcutaneously injected 120 \u0026micro;L of the prepared tumor cell suspension that contained 2 \u0026times; 10^7 cells achieving exponential growth into the mouse\u0026rsquo;s right axillary region with a sterile syringe and needle. Tumor growth, overall health, and signs of distress were monitored regularly. We used calipers to measure the tumor dimensions. Equation below interprets the calculation of tumor volumes.\u003c/p\u003e \u003cp\u003eV\u0026thinsp;=\u0026thinsp;1/2\u0026times;L\u0026times;W\u003csup\u003e2\u003c/sup\u003e\u003c/p\u003e \u003cp\u003eV: tumor volume (mm\u0026sup3;); L: long axis of the tumor (mm); W: short axis of the tumor (mm).One month after implantation was followed by the euthanization of the mice, and the excision, weight, and photographing of the tumors in succession. All animal experiments obeyed the protocols of the Institutional Animal Care and Use Committee of the First Affiliated Hospital of Kunming Medical University.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec22\" class=\"Section2\"\u003e \u003ch2\u003e2.20 Statistical Analysis\u003c/h2\u003e \u003cp\u003eAll computational and statistical analyses relied on R software (version 4.1.2) and GraphPad Prism 9. The Wilcoxon rank-sum test served for two-group comparison, and the Kruskal\u0026ndash;Wallis test served for the comparison among \u0026ge;\u0026thinsp;3 groups. The K-M method assisted in the OS analysis, with log-rank tests evaluating statistical significance. Results of LASSO and univariate/multivariate Cox regression analyses were taken into account for constructing a prognostic model for Bca patients. Correlation assessment among key genes depended on the Pearson\u0026rsquo;s correlation. P\u0026thinsp;\u0026lt;\u0026thinsp;0.05 denoted statistical significance.\u003c/p\u003e \u003c/div\u003e"},{"header":"3. Result","content":"\u003cdiv id=\"Sec24\" class=\"Section2\"\u003e \u003ch2\u003e3.1 Identification of Macrophage-Associated Core Genes and Key Signaling Pathways in Bca\u003c/h2\u003e \u003cp\u003eThe study retrieved transcriptomic data and clinical information from 408 Bca patients in the TCGA-BLCA cohort. Among the 4,271 identified DEGs, 1,614 presented upregulation and 2,657 presented downregulation. Their expression patterns were visualized via heatmap and volcano plot (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e. A-B; Supplementary Table\u0026nbsp;1). With the objective of examining the roles of MRGs in Bca progression, we compared tumor and normal tissues in terms of their ICI utilizing the CIBERSORT algorithm, focusing on macrophage infiltration. We then conducted WGCNA, selecting an optimal soft threshold (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e. C) to generate a scale-free network. Using dynamic tree cutting, we combined highly similar gene modules (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e. D) and identified two modules (brown and turquoise) that were strongly correlated with macrophage features (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e. E). From these modules, 2,474 MRGs were extracted (Supplementary Table\u0026nbsp;2). Next, we analyzed the GSE145281 single-cell RNA-sequencing dataset, identifying nine distinct cell types following log normalization and dimensionality reduction (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e. F), yielding 761 macrophage-related key genes (Supplementary Table\u0026nbsp;3). Finally, by intersecting three gene sets\u0026mdash;DEGs (DIFF), WGCNA module genes, and single-cell analysis genes\u0026mdash;we pinpointed 11 overlapping genes, which may serve as core functional markers of macrophage subpopulations (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e. G).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eGO enrichment analysis ascertained the participation of the DEGs in muscle contraction, ECM remodeling, collagen formation, and regulation at the biological process (BP) level. Hence, macrophages could crucially affect the tissue remodeling, cell migration, and intercellular communication. At the cellular component (CC) level, genes were highly enriched in the muscle contractile complex, ECM, and collagen triple helix, indicating their importance in structural maintenance. At the molecular function (MF) level, the DEGs participated in ligand binding, ion channel activity, and receptor interactions (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e. A\u0026ndash;D). In parallel, in KEGG pathway analysis, these genes clustered in calcium signaling pathways, cell adhesion molecules, neurotransmitter ligand\u0026ndash;receptor interactions, and cardiomyopathy-related pathways, implying pivotal functions in signal transduction, cellular interactions, and tissue-specific processes. Notably, the highest enrichment scores (in muscle cell skeleton, calcium signaling pathway, and cell adhesion molecules) highlight the fundamental involvement of macrophages in controlling migration, signal coordination, and tissue remodeling within the TME (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e. E-F).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec25\" class=\"Section2\"\u003e \u003ch2\u003e3.2 Construction and Validation of an MRG-Based Prognostic Risk Model\u003c/h2\u003e \u003cp\u003eUsing LASSO Cox regression, we identified differentially expressed MRGs with strong predictive capacity for Bca prognosis and built a risk scoring model that we comprehensively validated in the entire cohort, as well as in separate training and testing subsets. During model construction, key feature genes were selected via LASSO to ensure both simplicity and high efficiency (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e. A-B). According to K-M survival analysis, the high-risk group presented remarkably worse OS versus the low-risk group, regardless of whether we considered all patients (ALL; Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e. C), the training set ( Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e. H), or the testing set (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e. M). On these accounts, the risk scoring model, constructed from differentially expressed MRGs, excels in well distinguishing the two risk groups. The risk distribution plots further underscored this point: the high-risk group consistently presented higher mortality rates in the entire cohort (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e. D), the training subset (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e. I), and the testing subset (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e. N). Survival status scatter plots demonstrated a clear positive correlation between increased risk scores and mortality, with more frequent deaths clustered among high-risk individuals. Additionally, heatmaps depicting the expression of core genes showed a pronounced difference in gene expression patterns between two risk groups, reflecting distinct tumor biology and hinting at potentially valuable biomarkers for predicting outcomes and guiding treatment (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e. G,L,Q). Finally, the model achieved area under the curve (AUC) values between 0.60 and 0.70 at 1, 3, and 5 years in all three cohorts in the ROC curve analysis (ALL, Train, and Test) (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e. F,K,P), demonstrating its favorable classification performance, robust generalizability, and importance for the short- and long-term prognostic assessments.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eUnivariate and multivariate Cox regression analyses worked on more deeply evaluating the MCG-based risk score in terms of the independent prognostic performance and comprehensive accuracy. Risk score could be used for independently predicting OS, exhibiting the highest hazard ratio (HR\u0026thinsp;=\u0026thinsp;3.045, 95% CI: 1.594\u0026ndash;5.817) among all variables in the univariate analysis, including stage, T stage, and N stage. In multivariate analysis, with other clinical characteristics (age, sex, and tumor stage) being adjusted, the risk score still showed the best performance in independently predicting disease prognosis (HR\u0026thinsp;=\u0026thinsp;2.529, 95% CI: 1.243\u0026ndash;5.144, p\u0026thinsp;=\u0026thinsp;0.01), underscoring its robustness (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e. A-B). Comparisons via ROC curves likewise ascertained the better performance of the risk score (AUC\u0026thinsp;=\u0026thinsp;0.682) versus conventional clinical features of age (AUC\u0026thinsp;=\u0026thinsp;0.559), sex (AUC\u0026thinsp;=\u0026thinsp;0.518), and tumor stage (AUC\u0026thinsp;=\u0026thinsp;0.605) (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e. C). In a nomogram that integrated risk score, tumor stage, age, sex, etc., the risk score contributed substantially to the total prognostic score (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e. D). The 1-, 3-, and 5-year survival calibration curves exhibited outstanding concordance between predicted and observed outcomes (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e. E). Furthermore, cumulative hazard plots showed a markedly higher risk accumulation over time among high-risk patients, validating the model\u0026rsquo;s discriminative power (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e. F). Lastly, as evidenced by the C-index curves, used for dynamic comparison with traditional clinical characteristics, the risk score consistently outperformed age, sex, and tumor stage in predicting long-term survival (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e. G). Taken together, these results establish the MCG-based risk model as a highly effective prognostic tool, offering valuable guidance for patient stratification and personalized treatment in Bca.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec26\" class=\"Section2\"\u003e \u003ch2\u003e3.3 Association Between the MRG-Based Risk Score and ICI in the TME\u003c/h2\u003e \u003cp\u003eICI comparison revealed significantly higher levels of immunosuppressive cells (e.g., neutrophils and M2 macrophages) in the high-risk group, and obviously lower level of effector immune cells (e.g., activated CD8^+ T cells and activated NK cells) in the low-risk group (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e. A). Hence, high-risk patients\u0026rsquo; immune microenvironment exhibits a more pronounced immunosuppressive profile, which could impair antitumor immune responses. In subsequent heatmap analysis, the risk score exhibited a positive relevance to the infiltration of immunosuppressive cell types (neutrophils, M2 macrophages, and resting dendritic cells), but a negative relevance to the infiltration of effector immune cells (activated CD8^+ T cells, activated NK cells, and CD4^+ memory T cells) (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e. B). This distribution pattern highlights the immunosuppressive features of TME in the high-risk group and underscores the predictive utility of risk score for immune status.\u003c/p\u003e \u003cp\u003eFurther correlation analyses via scatter plots reinforced these observations, with infiltration by M2 macrophages (R\u0026thinsp;=\u0026thinsp;0.39, p\u0026thinsp;=\u0026thinsp;2.3e\u0026thinsp;\u0026minus;\u0026thinsp;08) and neutrophils (R\u0026thinsp;=\u0026thinsp;0.23, p\u0026thinsp;=\u0026thinsp;0.007) showing significant positive correlations with the risk score. In contrast, activated CD8^+ T cells (R\u0026thinsp;=\u0026thinsp;\u0026minus;\u0026thinsp;0.25, p\u0026thinsp;=\u0026thinsp;0.0004) and other effector cells demonstrated a negative relevance to the risk score (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e. C). Therefore, high-risk patients not only exhibit elevated levels of immunosuppressive cells but also display reduced proportions of effector immune cells, further indicating a potent immunosuppressive microenvironment. Lastly, according to violin plots comparing the risk score to key TME parameters\u0026mdash;StromalScore, ImmuneScore, and ESTIMATEScore, high-risk patients presented dramatically higher StromalScore and ImmuneScore versus their low-risk counterparts (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), indicating an increase in both stromal components and immunosuppressive cell infiltration (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e. D). This integrated set of results provides robust evidence of a significantly immunosuppressive TME in the high-risk cohort, indicative of the enhancement of immune evasion through the inhibition of effector immune cells\u0026rsquo; function.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec27\" class=\"Section2\"\u003e \u003ch2\u003e3.4 Association of Risk Score, Tumor Mutation Burden (TMB), and Patient Survival\u003c/h2\u003e \u003cp\u003eTo investigate how the gene mutation profile and TMB differ between the two risk groups, as well as further examine their associations with patient prognosis, we generated mutation frequency (MF) plots for both groups (Supplementary Fig.\u0026nbsp;1. A-B). Genes exhibiting the largest MF in the high-risk group were TP53, TTN, MUC16, PIK3CA, and KMT2D, with TP53 mutations occurring in 45% of patients, notably higher than the 40% observed in the low-risk group. That is to say, TP53 may affect the tumor malignant progression in high-risk patients. Additionally, genes such as MUC16 and KMT2D presented an obviously higher MF in the high-risk group, indicating a possible link between these mutations and increased tumor proliferation, metastasis, and immune evasion. Collectively, the molecular features of high-risk tumors may confer greater aggressiveness and adaptability.\u003c/p\u003e \u003cp\u003eWe next explored how TMB levels varied between the two risk cohorts. Although the high-risk group exhibited a marginally higher TMB level, the difference showed only borderline significance (p\u0026thinsp;=\u0026thinsp;0.068; Supplementary Fig.\u0026nbsp;1. C). This indicates that while high-risk patients may have a heavier mutational burden, other molecular factors might also influence tumor behavior. Notably, elevated TMB is frequently associated with genomic instability, which can drive TME variation and bolster the tumor\u0026rsquo;s capacity for immune evasion.\u003c/p\u003e \u003cp\u003eAccording to K-M curves that more deeply illustrate the TMB-OS relationship, low-TMB patients presented a remarkably worse OS versus low-TMB patients (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001; Supplementary Fig.\u0026nbsp;1. D), suggesting that TMB can be used to independently predict the disease prognosis. The poorer outcomes in high-TMB patients could reflect a heightened complexity and adaptability of tumor cells. Combining TMB and the MCG-based risk score, we subdivided patients into four groups: High TMB\u0026thinsp;+\u0026thinsp;High Risk, High TMB\u0026thinsp;+\u0026thinsp;Low Risk, Low TMB\u0026thinsp;+\u0026thinsp;High Risk, and Low TMB\u0026thinsp;+\u0026thinsp;Low Risk. As shown in Supplementary Fig.\u0026nbsp;1. E, the High TMB\u0026thinsp;+\u0026thinsp;High Risk group and the Low TMB\u0026thinsp;+\u0026thinsp;Low Risk group had the worst and the best OS, respectively (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001). These findings not only underscore the complementary prognostic value of TMB and the MCG-based risk score but also demonstrate that integrating both indicators significantly enhances the accuracy of patient stratification for survival outcomes.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec28\" class=\"Section2\"\u003e \u003ch2\u003e3.5 Correlation between the MRG-Based Risk Score and Drug Sensitivity\u003c/h2\u003e \u003cp\u003eThe study focused on evaluating the correlation between core genes in the prognostic risk model and drug sensitivity data from the GDSC, and the drug response difference between two risk groups, aiming at more deeply elucidating the potential utility of the model in guiding precision oncology. According to the correlation analyses, genes such as \u003cb\u003eANXA1\u003c/b\u003e, \u003cb\u003eST3GAL5\u003c/b\u003e, and \u003cb\u003eVIM\u003c/b\u003e may play pivotal roles in modulating therapeutic responses (Supplementary Fig.\u0026nbsp;2). Notably, \u003cb\u003eANXA1\u003c/b\u003e overexpression was associated with increased sensitivity to several agents, including PI3K inhibitors and DNA damage repair inhibitors, whereas elevated \u003cb\u003eVIM\u003c/b\u003e levels correlated with resistance to certain antimetabolites. These observations suggest that these core genes may influence drug response by regulating specific signaling pathways, thereby providing a molecular basis for more effective anticancer therapies.\u003c/p\u003e \u003cp\u003eFurther comparisons between the two risk cohorts in the drug sensitivity indicated significant distinctions in their responses to multiple anticancer agents, underscoring the practical relevance of the MCG-based risk score for clinical decision-making (Supplementary Fig.\u0026nbsp;2\u0026ndash;3). For instance, high-risk patients displayed greater sensitivity to targeted therapies such as EGFR inhibitors (e.g., gefitinib) and HER2 inhibitors (e.g., lapatinib), possibly reflecting increased pathway activity that renders them more responsive to these drugs. Conversely, high-risk patients exhibited reduced sensitivity to conventional chemotherapeutic agents (e.g., temozolomide and gemcitabine), suggesting a higher likelihood of chemoresistance in this subgroup. These disparities may be linked to the distinct molecular and microenvironmental characteristics of high-risk tumors, thereby informing potential optimization strategies for chemotherapy.\u003c/p\u003e \u003cp\u003eTaken together, these findings hold promising clinical implications. First, the expression of core genes (e.g., \u003cb\u003eANXA1\u003c/b\u003e and \u003cb\u003eVIM\u003c/b\u003e) may serve as predictive biomarkers for tumor drug sensitivity, informing individualized treatment decisions. Second, the robust differences in the drug sensitivity underscore the potential of the risk score as an adjunct tool for therapeutic selection. For example, high-risk patients may derive greater benefit from targeted therapies (e.g., EGFR and HER2 inhibitors), whereas low-risk patients might respond more favorably to traditional chemotherapy. Moreover, prioritizing targeted therapies could help overcome resistance in high-risk tumors, ultimately enhancing treatment outcomes.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec29\" class=\"Section2\"\u003e \u003ch2\u003e3.6 Single-Cell Analysis Reveals Expression Patterns of Macrophage-Associated Hub Genes in Bca\u003c/h2\u003e \u003cp\u003eTo more deeply reveal the roles of the model\u0026rsquo;s core genes (\u003cb\u003eANXA1\u003c/b\u003e, \u003cb\u003eST3GAL5\u003c/b\u003e, and \u003cb\u003eVIM\u003c/b\u003e) within the Bca TEM, we carried out an integrated single-cell RNA-sequencing analysis using the GSE149652 dataset. This approach enabled us to delineate the cellular heterogeneity within tumor tissues and examine the spatial distribution and density of diverse immune cell populations (CD4^+ T cells, CD8^+ T cells, exhausted T cells (CD8Tex), NK cells, proliferative T cells (Tprolif), and regulatory T cells (Treg cell)) (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003e. A). All these highlight the functional specialization of each cell subset in tumor progression and immune modulation. We next investigated the expression profiles of \u003cb\u003eANXA1\u003c/b\u003e, \u003cb\u003eST3GAL5\u003c/b\u003e, and \u003cb\u003eVIM\u003c/b\u003e across different immune cell populations (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003e. B\u0026ndash;D). \u003cb\u003eANXA1\u003c/b\u003e was highly expressed in CD8^+ T cells and CD8Tex, implying a potential role in immunosuppression and T-cell exhaustion regulation. \u003cb\u003eST3GAL5\u003c/b\u003e had relatively uniform expression but was slightly enriched in CD8^+ T cells, suggesting a broad involvement in immune regulation. Conversely, \u003cb\u003eVIM\u003c/b\u003e was highly expressed in both Treg and NK cells, pointing to its contribution to immune evasion and NK cell activity modulation. Quantitative comparisons (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003e. E) reinforced these observations, underscoring the possibility that \u003cb\u003eANXA1\u003c/b\u003e overexpression in CD8Tex could be a pivotal driver of tumor immune escape, while \u003cb\u003eVIM\u003c/b\u003e upregulation in Treg cells may potentiate the immunosuppressive milieu. Finally, cell\u0026ndash;cell communication analyses underscored the intricate signaling networks between CD8^+ T cells and other immune cell types. Notably, communications between CD8^+ T cells, Treg, and Tprolif cells were significantly enhanced, indicating extensive signal exchange in the TME. These interactions may orchestrate tumor immune evasion and inflammation; e.g., Treg cells could suppress the antitumor functions of CD8^+ T cells, whereas T-prolif cells may bolster tumor proliferation and accelerate disease progression.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec30\" class=\"Section2\"\u003e \u003ch2\u003e3.7 Differential Expression of ANXA1, ST3GAL5, and Waveform Proteins in Bca Cell Lines\u003c/h2\u003e \u003cp\u003eWestern blotting and IHC analyses revealed obviously different expressions of ANXA1, ST3GAL5, and waveform proteins in various Bca cell lines (SVHUC1, 5637, RT4, and TCCSUP). ANXA1 and ST3GAL5 were significantly elevated in SVHUC1 cells compared to other cell lines, and surprisingly, mRNA expression data further confirmed these findings, suggesting a possible association between these proteins and Bca progression. On these accounts, ANXA1, ST3GAL5 and Vimentin may be potential biomarkers for Bca (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec31\" class=\"Section2\"\u003e \u003ch2\u003e3.8 The Role of ST3GAL5 Overexpression in Bca Cell Proliferation and Tumor Growth\u003c/h2\u003e \u003cp\u003eAs illustrated in Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003e, experimental data support the notion that ST3GAL5 overexpression markedly enhances Bca cell proliferation and tumor growth. In both 5637 and RT4 cell lines, ST3GAL5 overexpression (oe-ST3GAL5) led to significantly increased cell viability and colony formation compared to the control (NC) and siRNA- mediated knockdown (si-ST3GAL5) groups. Cell viability assays showed that the oe-ST3GAL5 group exhibited notably higher cell proliferation at 24 h and 36 h (p\u0026thinsp;\u0026lt;\u0026thinsp;0.01). Moreover, flow cytometry revealed that ST3GAL5 overexpression induced a greater proportion of cells in the S and G2 phases, hinting accelerated cell-cycle progression. In vivo, mice injected with 5637 cells overexpressing ST3GAL5 developed significantly larger and heavier tumors than mice in the NC group (p\u0026thinsp;\u0026lt;\u0026thinsp;0.01). Collectively, all these highlight the key function of ST3GAL5 in bolstering Bca cell proliferation and tumor growth, confirming ST3GAL5 as a promising therapeutic target specific to Bca.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"4. Discussion","content":"\u003cp\u003eThe present study integrated bulk transcriptomic data and scRNA-seq data to investigate the macrophage heterogeneity in the Bca TME and to characterize the functions of related genes. By identifying macrophage-related core genes and developing a survival risk prediction model, we not only demonstrated the crucial impact of macrophages on tumor progression and immunoregulation but also emphasized the potential clinical utility of these core genes in immunotherapy and precision medicine.\u003c/p\u003e \u003cp\u003eMacrophages are key players in the TME, orchestrating immune responses and fostering tumor development. Our findings confirmed the heterogeneity of macrophages in Bca tissues, showing that different subpopulations (e.g., M1- and M2-like macrophages) exhibit distinct spatial distributions and functional properties. Notably, high-risk patients presented obviously abundant M2-like macrophages, in line with their known roles in promoting immunosuppression and tumor progression [\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e]. DEA further revealed that core genes such as \u003cb\u003eANXA1\u003c/b\u003e, \u003cb\u003eST3GAL5\u003c/b\u003e, and \u003cb\u003eVIM\u003c/b\u003e are differentially expressed across macrophage subsets, possibly facilitating immune escape and enhancing tumor aggressiveness via pathways like Wnt and TGF-β[\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e, \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e]. The enrichment of M2-like macrophages alongside neutrophils in high-risk patients supports a highly immunosuppressive microenvironment, whereas a marked decrease in CD8^+ T cells and NK cells could weaken antitumor immunity [\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]. Future studies should elucidate the mechanistic interplay between M2-like macrophages and other immune cells, and define how these core genes modulate the TME.\u003c/p\u003e \u003cp\u003eThrough the identification of macrophage-related core genes and subsequent model construction, we developed a robust prognostic tool for Bca. The risk prediction model achieved an AUC of 0.682, indicating satisfactory accuracy in prognostic stratification. Consistently, high-risk patients exhibited an obviously worse OS versus low-risk patients. One advantage of our model is that it integrates MRG expression with clinical features to offer both refined risk stratification and actionable information for individualized treatment decisions. Additionally, each of the core genes in this model appears to have clinical relevance. For example, \u003cb\u003eANXA1\u003c/b\u003e was highly expressed in CD8^+ T cells and CD8Tex, hinting its immunosuppressive action [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e], whereas \u003cb\u003eVIM\u003c/b\u003e exhibited elevated expression in Treg cells, potentially contributing to immune evasion [\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e]. Targeting these molecules could open new therapeutic avenues that may improve outcomes for high-risk patients.\u003c/p\u003e \u003cp\u003eWe also performed drug sensitivity analyses to explore how these core genes influence treatment responses and to assess their utility in precision oncology. Notably, high-risk patients showed greater sensitivity to targeted therapies (e.g., the EGFR inhibitor gefitinib and the HER2 inhibitor lapatinib) while demonstrating reduced sensitivity to chemotherapeutic agents such as gemcitabine (consistent with enhanced chemoresistance). These differences might reflect the higher expression of core genes in high-risk patients and their regulatory effects on key signaling pathways, a notion supported by previous reports indicating that \u003cb\u003eANXA1\u003c/b\u003e and \u003cb\u003eVIM\u003c/b\u003e modulate cell-cycle and immune signaling pathways[\u003cspan additionalcitationids=\"CR24\" citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e]. Meanwhile, \u003cb\u003eST3GAL5\u003c/b\u003e could affect metabolic pathways and thereby alter chemotherapy efficacy, aligning with observations in other malignancies[\u003cspan additionalcitationids=\"CR27\" citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e]. Besides, the higher TMB observed in high-risk individuals points to the potential value of immunotherapeutic strategies (e.g., PD-1/PD-L1 inhibitors) for this subgroup. Integrating the risk score with TMB further improved patient stratification and offers a theoretical foundation for the design of more targeted immunotherapies.\u003c/p\u003e \u003cp\u003eIn summary, our study illuminates the functional heterogeneity of macrophages in Bca, identifies a robust macrophage-related risk model for patient stratification, and underscores the critical contributions of core genes to immunoregulation and drug responsiveness. All these make us more deeply comprehend Bca tumor biology and contribute to the formulation of innovative and more effective treatment strategies aimed at harnessing or modifying the TME. Future work, including mechanistic studies and prospective clinical trials, is warranted to validate these findings and to translate our risk model and proposed targets into routine clinical practice.\u003c/p\u003e \u003cp\u003eFurthermore, functional assays in the present study confirmed that \u003cb\u003eST3GAL5\u003c/b\u003e overexpression can substantially enhance Bca cell proliferation and tumor growth, as evidenced by increased cell viability, colony formation, and tumor volume in vivo. Given that ST3GAL5 encodes a key enzyme essential for ganglioside biosynthesis, it may influence multiple signaling pathways related to cell proliferation, migration, and interactions with the TME. Elevated ST3GAL5 expression could thus promote a more aggressive tumor phenotype through enhanced glycan-mediated cell signaling and potential modulation of ICI. All these underscore the oncogenic role of ST3GAL5 in Bca progression and indicate that targeting ST3GAL5 or its downstream pathways can be a valuable treatment option\u0026mdash;particularly for patients whose tumors exhibit high ST3GAL5 expression. Future investigations are suggested to pay attention to delineating the exact molecular mechanisms for ST3GAL5 to drive tumor growth and determining whether inhibiting its activity could improve responses to conventional therapies or synergize with emerging immunotherapeutic and targeted approaches.\u003c/p\u003e \u003cp\u003eDespite shedding light on the critical roles of MRGs in Bca, this study has certain limitations. First, the construction and validation of our model predominantly relied on publicly available datasets; hence, further validation in larger clinical cohorts is necessary to confirm its generalizability. Second, the exact functional mechanisms of these core genes operating within the tumor immune microenvironment remain to be elucidated through experimental research. Moreover, although our model exhibited a reasonably favorable predictive performance (AUC\u0026thinsp;=\u0026thinsp;0.682), there is still room for improvement. Future studies could integrate proteomics, metabolomics, and other multi-omics data to further refine and optimize the model. Overall, this study offers new insights into stratified management and personalized treatment for Bca patients, while providing a solid theoretical basis and practical references for future research on tumor immunology.\u003c/p\u003e"},{"header":"5. Conclusion","content":"\u003cp\u003eThis study integrated transcriptomic and scRNA-seq data for the systematic elucidation of the functional characteristics of MRGs within the Bca TME, and the constructed survival risk model demonstrated promising predictive performance to support personalized treatment strategies. Future researches are suggested to pay more attention to exploring the molecular regulatory mechanisms of MRGs against tumor immunity and validate the model in larger clinical cohorts, thereby advancing precision medicine in Bca and other tumor immunology studies.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eData Availability Statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eBladder cancer (Bca) transcriptional datasets were obtained from the TCGA and GEO databases (accessible at https://portal.gdc.cancer.gov ).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;\u003cstrong\u003eConflict of Interest\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;\u003cstrong\u003eAuthor Contributions\u003c/strong\u003e\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eHaolin Liu, Jian Hou, Yumin Wang and Yuanqi Chu collaboratively drafted the manuscript. Junxiong Li, Jingbo Qin and Pinyao Liang did the cell experiments. Peng Gu and Xiaodong Liu performed the data visualization. Guoqiang Liao, Xiangyang Wen collected resources and finished the data analysis. Xiangyang Wen and Xiaodong Liu conceived and supervised this study. All authors read and approved the final version of the manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003eFunding\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;\u003cstrong\u003eClinical trial number\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003enot applicable.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003eEthics declaration\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003enot applicable.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003eConsent to Publish declaration\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003enot applicable.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003eConsent to Participate declaration\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003enot applicable.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003eAcknowledgments\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe gratefully acknowledge the contributions from the TCGA and GEO project, which provided valuable data and resources for this research.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003ePatel VA-O, Oh WK and Galsky MD. Treatment of muscle-invasive and advanced bladder cancer in 2020. \u003c/li\u003e\n\u003cli\u003eLenis AT, Lec PM, Chamie K Fau - Mshs MD and Mshs MD. Bladder Cancer: A Review. \u003c/li\u003e\n\u003cli\u003eSong X, Xin S, Zhang Y, Mao J, Duan C, Cui K, Chen L, Li F, Liu Z, Wang T, Liu J, Liu X and Song W. Identification and Quantification of Iron Metabolism Landscape on Therapy and Prognosis in Bladder Cancer. \u003c/li\u003e\n\u003cli\u003eSiracusano SA-O, Rizzetto R and Porcaro AB. Bladder cancer genomics. \u003c/li\u003e\n\u003cli\u003eZhang ZL, Dong P Fau - Li Y-H, Li Yh Fau - Liu Z-W, Liu Zw Fau - Yao K, Yao K Fau - Han H, Han H Fau - Qin Z-K, Qin Zk Fau - Zhou F-J and Zhou FJ. Radical cystectomy for bladder cancer: oncologic outcome in 271 Chinese patients. \u003c/li\u003e\n\u003cli\u003eJiang Y, Wang C and Zhou S. Targeting tumor microenvironment in ovarian cancer: Premise and promise. \u003c/li\u003e\n\u003cli\u003eXiao Y and Yu D. Tumor microenvironment as a therapeutic target in cancer. \u003c/li\u003e\n\u003cli\u003eHuang T, Song X, Xu D, Tiek D, Goenka A, Wu B, Sastry N, Hu B and Cheng SY. Stem cell programs in cancer initiation, progression, and therapy resistance. \u003c/li\u003e\n\u003cli\u003ePan Y, Yu Y, Wang X and Zhang T. Tumor-Associated Macrophages in Tumor Immunity. \u003c/li\u003e\n\u003cli\u003eLu J, Sheng Y, Qian W, Pan M, Zhao X and Ge QA-O. scRNA-seq data analysis method to improve analysis performance. \u003c/li\u003e\n\u003cli\u003eYuen KC, Liu LF, Gupta V, Madireddi S, Keerthivasan S, Li C, Rishipathak D, Williams P, Kadel EErA-O, Koeppen H, Chen YJ, Modrusan Z, Grogan JA-O, Banchereau RA-O, Leng N, Thastrom A, Shen X, Hashimoto K, Tayama D, van der Heijden MS, Rosenberg JE, McDermott DA-O, Powles T, Hegde PS, Huseni MA-O and Mariathasan SA-OX. High systemic and tumor-associated IL-8 correlates with reduced clinical benefit of PD-L1 blockade. \u003c/li\u003e\n\u003cli\u003eRitchie ME, Phipson B, Wu D, Hu Y, Law CW, Shi W and Smyth GK. limma powers differential expression analyses for RNA-sequencing and microarray studies. \u003c/li\u003e\n\u003cli\u003eNewman AM, Liu CL, Green MA-O, Gentles AA-O, Feng W, Xu Y, Hoang CD, Diehn M and Alizadeh AA-O. Robust enumeration of cell subsets from tissue expression profiles. \u003c/li\u003e\n\u003cli\u003eFinotello FA-O and Trajanoski ZA-O. Quantifying tumor-infiltrating immune cells from transcriptomics data. \u003c/li\u003e\n\u003cli\u003eLangfelder P and Horvath S. WGCNA: an R package for weighted correlation network analysis. \u003c/li\u003e\n\u003cli\u003eYunna C, Mengru H, Lei W and Weidong C. Macrophage M1/M2 polarization. \u003c/li\u003e\n\u003cli\u003eLi M, Yang Y, Xiong L, Jiang P, Wang J and Li C. Metabolism, metabolites, and macrophages in cancer. \u003c/li\u003e\n\u003cli\u003eZhou Y, Xu J, Luo H, Meng X, Chen M and Zhu D. Wnt signaling pathway in cancer immunotherapy. \u003c/li\u003e\n\u003cli\u003ePeng D, Fu M, Wang M, Wei Y and Wei X. Targeting TGF-\u0026beta; signal transduction for fibrosis and cancer therapy. \u003c/li\u003e\n\u003cli\u003eHu J, Zhang L, Xia H, Yan Y, Zhu X, Sun F, Sun L, Li S, Li D, Wang J, Han Y, Zhang J, Bian D, Yu H, Chen Y, Fan P, Ma Q, Jiang G, Wang C and Zhang P. Tumor microenvironment remodeling after neoadjuvant immunotherapy in non-small cell lung cancer revealed by single-cell RNA sequencing. \u003c/li\u003e\n\u003cli\u003eLi L, Wang B, Zhao S, Xiong Q and Cheng A. The role of ANXA1 in the tumor microenvironment. \u003c/li\u003e\n\u003cli\u003eRidge KA-O, Eriksson JE, Pekny MA-O and Goldman RA-O. Roles of vimentin in health and disease. \u003c/li\u003e\n\u003cli\u003eShi X, Wu Y, Tang L, Ni H and Xu Y. Downregulated annexin A1 expression correlates with poor prognosis, metastasis, and immunosuppressive microenvironment in Ewing\u0026apos;s sarcoma. \u003c/li\u003e\n\u003cli\u003eChen J, Lu T, Chen C, Zheng W, Lu L and Li N. Elevation of ANXA1 associated with potential protective mechanism against ferroptosis and immune cell infiltration in age-related macular degeneration. \u003c/li\u003e\n\u003cli\u003eGhantasala S, Pai MGJ, Biswas D, Gahoi N, Mukherjee S, Kp M, Nissa MU, Srivastava A, Epari S, Shetty P, Moiyadi A and Srivastava S. Multiple Reaction Monitoring-Based Targeted Assays for the Validation of Protein Biomarkers in Brain Tumors. \u003c/li\u003e\n\u003cli\u003eManoochehri J, Dastgheib SA, Khamirani HA-O, Mollaie M, Sharifi Z, Zoghi S, Tabei SMB, Mohammadi S, Dehghanian F, Farbod Z and Dianatpour M. A novel frameshift pathogenic variant in ST3GAL5 causing salt and pepper developmental regression syndrome (SPDRS): A case report. \u003c/li\u003e\n\u003cli\u003eHeide S, Jacquemont ML, Cheillan D, Renouil M, Tallot M, Schwartz CE, Miquel J, Bintner M, Rodriguez D, Darcel F, Buratti J, Haye D, Passemard S, Gras D, Perrin L, Capri Y, G\u0026eacute;rard B, Piton A, Keren B, Thauvin-Robinet C, Duffourd Y, Faivre L, Poe C, Pervill\u0026eacute; A, H\u0026eacute;ron D, Th\u0026eacute;venon J, Arnaud L, LeGuern E, La Selva L, Vetro A, Guerrini R, Nava C and Mignot C. GM3 synthase deficiency in non-Amish patients. \u003c/li\u003e\n\u003cli\u003evan der Haar \u0026Agrave;vila IA-O, Zhang TA-O, Lorrain VA-O, de Bruin F, Spreij TA-O, Nakayama HA-O, Iwabuchi KA-O, Garc\u0026iacute;a-Vallejo JA-O, Wuhrer MA-O, van Kooyk YA-O and van Vliet SA-O. Limited impact of cancer-derived gangliosides on anti-tumor immunity in colorectal cancer. LID - 10.1093/glycob/cwae036 [doi] LID - cwae036. \u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Bladder cancer, Macrophages, Survival risk model, Immune microenvironment, Drug sensitivity analysis","lastPublishedDoi":"10.21203/rs.3.rs-6452577/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6452577/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eBackground: Bladder cancer (Bca) is a highly malignant tumor characterized by a high recurrence and metastasis rate. Macrophages crucially affect tumor progression and immunotherapy response, while researches have not well explored their precise mechanisms of action against Bca. The study aims at investigating the function of macrophage-related genes (MRGs) within the Bca immune microenvironment and exploring their potential value in prognosis prediction and therapeutic decision-making.\u003c/p\u003e\n\u003cp\u003eMethod: This study integrated Bca transcriptomic data from the TCGA and GEO databases along with single-cell RNA sequencing (scRNA-seq) data to systematically identify key MRGs. Differential expression analysis, weighted gene co-expression network analysis (WGCNA), and single-cell sequencing analysis served for screening for core MRGs. The results from LASSO Cox regression analysis were used for constructing a survival risk prediction model, together with the evaluation of the model’s predictive accuracy. Besides, core MRGs were subjected to immune cell infiltration and drug sensitivity analyses for the elucidation of their roles in immune regulation and therapeutic response. Furthermore, key genes in the prognostic model were validated using PCR, Western blot, and immunohistochemistry.\u003c/p\u003e\n\u003cp\u003eResult: This study identified 11 core genes significantly associated with macrophages and developed a risk prediction model based on ANXA1, ST3GAL5, and VIM. The model demonstrated high predictive accuracy across all samples (AUC = 0.682). Immune analysis revealed that high-risk patients exhibited a distinctly immunosuppressive tumor microenvironment (TME), characterized by increased infiltration of M2 macrophages and neutrophils, along with a significant reduction in effector immune cells of CD8⁺ T cells and NK cells. Additionally, high-risk patients displayed greater sensitivity to targeted therapies (e.g., EGFR and HER2 inhibitors) but reduced sensitivity to conventional chemotherapy. According to in vitro and in vivo experiments, ST3GAL5 overexpression significantly promoted Bca cell proliferation and tumor growth, underscoring its potential role in tumor progression.\u003c/p\u003e\n\u003cp\u003eConclusion: This study highlights the crucial impact of MRGs on the TME of Bca and constructs a risk prediction model that effectively predicts patient survival outcomes, providing a theoretical foundation and practical guidance for personalized treatment strategies. Future researches are suggested to more deeply elucidate the functional mechanisms of these core genes as well as explore their potential as therapeutic targets for Bca.\u003c/p\u003e","manuscriptTitle":"Targeting Macrophage-Associated Core Genes for Prognostic Prediction and Therapeutic Insights in Bladder Cancer","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-05-13 10:24:17","doi":"10.21203/rs.3.rs-6452577/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"902c1aaa-1136-4e1f-9013-e31d8e52ea71","owner":[],"postedDate":"May 13th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2025-11-23T01:23:14+00:00","versionOfRecord":[],"versionCreatedAt":"2025-05-13 10:24:17","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6452577","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6452577","identity":"rs-6452577","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}