FREM1 Serves as a Novel Therapeutic Target in Breast Cancer through Basement Membrane-Based Prognostic Modeling with Integrated Bioinformatics and Experimental Validation

FREM1 Serves as a Novel Therapeutic Target in Breast Cancer through Basement Membrane-Based Prognostic Modeling with Integrated Bioinformatics and Experimental Validation | 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 FREM1 Serves as a Novel Therapeutic Target in Breast Cancer through Basement Membrane-Based Prognostic Modeling with Integrated Bioinformatics and Experimental Validation Chao Li, Pingming Gong, Junfeng Hu, Chengyu Luo This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7206387/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 10 You are reading this latest preprint version Abstract Background​​ : Breast cancer remains the leading cause of cancer-related mortality in women worldwide, with metastatic disease posing significant therapeutic challenges. While immunotherapy has shown promise, tumor immune evasion limits its efficacy. The basement membrane (BM), a specialized extracellular matrix structure, plays a crucial yet understudied role in breast cancer progression and immune modulation. This study aims to investigate the prognostic value and therapeutic potential of BM-related genes in breast cancer. ​​ Methods​​ : We integrated transcriptomic data from TCGA and GEO databases to construct a BM-related gene signature. Unsupervised clustering stratified patients into molecular subtypes, while differential expression analysis identified key BM-associated genes. Functional enrichment analyses (GO, KEGG, GSEA) elucidated biological pathways, and immune microenvironment characterization was performed using ESTIMATE and CIBERSORT. Machine learning approaches pinpointed critical BM-related genes, which were subsequently validated through in vitro experiments. ​​ Results​​ : Breast cancer patients were classified into high- and low-BM groups, with the low-BM cohort exhibiting worse prognosis. Pathway analysis revealed significant enrichment in immune regulation, ECM remodeling, and cytokine signaling. FREM1 emerged as a top protective gene through machine learning. Experimental validation using low-FREM1-expressing breast cancer cell lines demonstrated that FREM1 overexpression (confirmed by qPCR and Western blot) significantly suppressed tumor cell proliferation, as evidenced by decreased Ki-67 expression and reduced EdU incorporation. ​​ Conclusion​​ : Our study establishes BM-related genes as novel prognostic biomarkers and therapeutic targets in breast cancer. FREM1 in particular functions as a tumor suppressor by inhibiting cancer cell proliferation, highlighting its potential for therapeutic exploitation. These findings provide critical insights into BM-mediated tumor progression and suggest new avenues for targeted breast cancer therapy. Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Introduction Breast cancer (BC) is the most prevalent malignancy among women worldwide, with an estimated 2.3 million new cases diagnosed each year (1), making it the leading cause of cancer incidence globally. Breast cancer is the leading cause of cancer-related deaths among women globally (2), accounting for the fifth highest mortality rate across all cancers (1). Breast cancer management currently involves a range of treatments, including surgery (3), chemotherapy (4), radiotherapy (5), endocrine therapy (6), and targeted therapies (7). Despite these options, their effectiveness is often limited, especially in cases of metastatic breast cancer, where long-term survival outcomes remain poor (8, 9). Recent advancements in immunotherapy have significantly impacted breast cancer treatment (10), offering particularly promising results for triple-negative breast cancer (TNBC) (11) and HER2-positive breast cancer (12). The emergence of immune checkpoint inhibitors has opened new avenues for strengthening localized anti-tumor immune responses (13). Nonetheless, while certain patients experience notable clinical benefits, the majority continue to face challenges of disease progression, largely attributed to the development of primary or acquired resistance (14). Enhancing treatment efficacy requires a deeper understanding of the interaction between breast cancer cells and the immune microenvironment, along with the mechanisms underlying immune evasion. Additionally, the discovery of reliable biomarkers can significantly advance the development of immunotherapy strategies. The extracellular matrix (ECM) is central to the tumor microenvironment, with the basement membrane (BM) acting as a barrier that prevents cancer cells from spreading (15). Abnormal regulation of the BM can, therefore, facilitate tumor invasion and metastasis (16). Numerous studies have demonstrated that basement membrane-related genes are linked to the prognosis of various cancers, including Lung adenocarcinoma (17) and hepatocellular carcinoma (18). Emerging studies indicate that the progression of breast cancer is driven not only by the tumor cells themselves but also by a significantly altered tumor microenvironment (19). The extracellular matrix (ECM), particularly the basement membrane (BM), is pivotal in regulating various aspects of tumor biology, including cell migration and invasion (19). Alterations in the ECM, like collagen IV breakdown, commonly induced by immune cells and cancer-associated fibroblasts, can create a pathway for tumor cells to invade through the basement membrane (20). These alterations point to a potential yet underexplored connection between the BM and tumor immune infiltration. In this study, we developed a BM-related gene signature to stratify breast cancer prognosis, rigorously validating its predictive performance using TCGA and GEO datasets. The model demonstrated independent prognostic value and revealed distinct tumor microenvironment profiles, including differential immune infiltration patterns that may influence immunotherapy responses. Through machine learning, we identified FREM1 as a key BM-associated molecular target, subsequently validating its tumor-suppressive role via functional assays. This integrated approach not only enhances prognostic precision but also uncovers novel therapeutic targets for breast cancer and its immune modulation. Materials and Methods Data Collection and Processing The breast cancer (BRCA) RNA-seq transcriptome data and clinical information were obtained from The Cancer Genome Atlas (TCGA, https://cancergenome.nih.gov/ ) as the primary dataset. To ensure the robustness of survival analysis, samples lacking survival data were excluded. The data were downloaded in fragments per kilobase million (FPKM) format for consistency in downstream analyses. For external validation, we utilized the GSE131769 dataset from the Gene Expression Omnibus (GEO, http://www.ncbi.nlm.nih.gov/geo ), which provides survival information and clinical outcomes. Additionally, two single-cell RNA-seq datasets, GSE148673 and GSE176078, were acquired for investigating the expression patterns of target genes within the breast cancer microenvironment. Based on a thorough review of the literature, we also compiled a list of 222 basement membrane-related genes for further analysis in this study (15). Molecular Subtyping via Consensus Clustering​ To delineate clinically relevant molecular subtypes based on basement membrane gene expression patterns, we implemented an unsupervised consensus clustering approach. The analysis was conducted using the "ConsensusClusterPlus" R package with the following parameters: 100 resampling iterations to ensure robust cluster stability, a pItem value of 0.8 to maintain sample assignment consistency, and Euclidean distance metric with hierarchical clustering. Systematic evaluation of the consensus matrix and cumulative distribution function curves revealed optimal separation into two distinct molecular subtypes. Subsequent survival analysis demonstrated significant differences in overall survival between these subtypes (log-rank p < 0.05), confirming the clinical relevance of the basement membrane-related gene signatures in breast cancer prognosis. Differential Expression and Functional Enrichment Analysis To characterize the molecular distinctions between the identified breast cancer subtypes, we performed differential gene expression analysis using the "limma" package in R, applying thresholds of |logFC| >1 and adjusted p-value < 0.05. Significantly differentially expressed genes (DEGs) were then functionally annotated using the "clusterProfiler" package. Gene Ontology (GO) analysis categorized DEGs into biological processes, cellular components, and molecular functions, while KEGG pathway analysis identified enriched signaling pathways. Additionally, Gene Set Enrichment Analysis (GSEA) was conducted using the "c5.go.symbols.gmt" and "c2.cp.kegg.symbols.gmt" databases from the MSigDB collection to further validate pathway-level associations. This integrated approach provided a systematic framework for interpreting the biological relevance of subtype-specific gene signatures. Analysis of Tumor Microenvironment Components To characterize the tumor microenvironment across molecular subtypes, we employed a comprehensive analytical approach. First, we calculated ESTIMATE (Est​​imation of ​​St​​romal and ​​Imm​​une cells in ​​M​​alignant ​​T​​umor tissues using ​​E​​xpression data) scores to quantify stromal and immune cell infiltration levels as well as tumor purity using standardized parameters. Next, we performed immune cell deconvolution using the CIBERSORTx (C​​ell ​​I​​dentity ​​B​​y ​​E​​stimated ​​R​​egression ​​S​​orting ​​O​​f ​​R​​NA ​​T​​ranscripts) algorithm (v1.06) with the LM22 signature matrix and 1000 permutations to estimate the relative proportions of 22 immune cell subtypes, applying a significance threshold of p < 0.05 for reliable detection. Additionally, we systematically evaluated the expression profiles of immunomodulatory markers, including HLA (​​H​​uman ​​L​​eukocyte ​​A​​ntigen) gene family members and key immune checkpoint molecules, using one-way ANOVA with Benjamini-Hochberg correction for multiple testing. This integrated analysis enabled quantitative assessment of the tumor-immune interface characteristics across different molecular subtypes. Development and Validation of a Basement Membrane-Related Prognostic Signature To establish a robust prognostic signature, we first performed univariate Cox regression analysis to identify basement membrane (BM)-related genes significantly associated with overall survival (p < 0.05). These candidate genes were subsequently subjected to Lasso regression with 10-fold cross-validation to prevent overfitting and generate a refined risk score model. Patients were dichotomized into high- and low-risk groups using the median risk score as the cutoff value. The model's predictive performance was systematically evaluated through Kaplan-Meier survival curves and risk score distribution visualization. External validation was conducted using an independent GEO cohort to confirm clinical applicability. To enhance biomarker discovery, we integrated machine learning approaches by combining Lasso regression with support vector machine (SVM) recursive feature elimination. This dual-selection strategy identified consensus prognostic genes exhibiting both statistical significance and biological relevance. The machine learning framework enabled detection of complex genomic patterns that conventional statistical methods might miss, thereby improving the model's predictive accuracy for personalized treatment stratification. Comprehensive Analysis of Key Genes in Breast Cancer Expression Prognosis and Tumor Microenvironment To refine the selection of clinically relevant genes, we systematically analyzed the machine learning-derived candidate genes using TCGA data. We initially conducted Kaplan-Meier survival analysis to investigate potential correlations between gene expression patterns and clinical outcomes in breast cancer patients. We subsequently examined differential gene expression between tumor and adjacent normal tissues to elucidate their possible involvement in breast cancer pathogenesis. This sequential analytical strategy - progressing from clinical relevance to mechanistic exploration - enabled systematic identification of the most promising candidate genes for further investigation. Furthermore, we performed single-cell RNA sequencing analysis to delineate the precise cellular localization and expression patterns of these genes across different cell populations in breast tumors. This comprehensive approach provided critical insights into the potential functional roles of these genes in tumor biology and disease progression. Cell Line Acquisition and Culture Conditions​ We utilized two breast cell lines for our experiments: the highly metastatic MDA-MB-231 cancer cells and non-tumorigenic MCF-10A normal epithelial cells, both obtained from Procell Life Science & Technology Co., Ltd (Wuhan, China). These cell lines were maintained under standard culture conditions (37°C, 5% CO 2 ) using distinct growth media formulations. The MDA-MB-231 line was propagated in high-glucose DMEM supplemented with 10% FBS and antibiotic solution. The MCF-10A normal epithelial cells required a specialized DMEM/F12-based medium containing 5% equine serum along with specific growth factors including EGF (20 ng/mL), hydrocortisone (0.5 µg/mL), insulin (10 µg/mL), and NEAA supplementation. Plasmid Construction and Transfection The FREM1 coding sequence was PCR-amplified from MDA-MB-231 cell cDNA using specific primers containing KpnI and XhoI restriction sites (Forward: 5'-GCCGGTACCGCCACCATGGTGACACAAGAATCCATGCTG-3'; Reverse: 5'-GGCCTCGAGTTACTTGTCATCGTCGTCCTTGTAATCGAGTTTTCTGGAACACAC-3'). The amplification was carried out for 30 cycles under standard conditions: 95°C for 15 seconds, 68°C for 15 seconds, and 72°C for 1 minute per cycle. Following amplification, the PCR product was digested with the corresponding restriction enzymes (Transgen, Beijing, China) at 37°C for 30 minutes to generate compatible ends for subsequent cloning. The resulting FREM1 fragment was then ligated into the expression vector and transformed into competent cells for plasmid propagation. For functional studies, the verified FREM1 overexpression plasmid was introduced into MDA-MB-231 cells using Lipo8000 transfection reagent (Beyotime, Shanghai, China) following the manufacturer's recommended protocol. Quantitative PCR Analysis for Cell Line Screening and Transduction Validation To identify breast cancer cell lines with low endogenous FREM1 expression and validate successful plasmid transfection, we performed quantitative real-time PCR (qPCR) analysis. Total RNA was extracted from cell lines using TRIzol reagent (Servicebio) followed by DNase I treatment to eliminate genomic DNA contamination. cDNA synthesis was carried out with 1 µg total RNA using HiScript II Reverse Transcriptase (Servicebio) and oligo (dT) primers. For qPCR amplification, reactions were prepared in triplicate with TB Green Premix Ex Taq II (Servicebio) and run on a LightCycler 480 system (Roche) under the following conditions: 95°C for 30 sec, followed by 40 cycles of 95°C for 5 sec and 60°C for 30 sec. FREM1-specific primers (Forward: 5'-GCCTGTGGTAACCAGGAACAA-3'; Reverse: 5'-CGCAGGTGTATCAGGGTCG-3') were designed to span exon-exon junctions, with GAPDH (Forward: 5'-GGAGCGAGATCCCTCCAAAAT-3'; Reverse: 5'-GGCTGTTGTCATACTTCTCATGG-3') serving as the endogenous control. Melting curve analysis confirmed primer specificity. Relative expression levels were calculated using the 2 −ΔΔCt method. For transduction validation, FREM1 overexpression was confirmed by ≥ 1.5-fold increase in mRNA levels compared to empty vector controls (p < 0.05). Western Blot Analysis for FREM1 Expression Screening and Overexpression Validation​​ ​​To complement our qPCR findings and further verify FREM1 protein expression levels, we performed comprehensive western blot analysis. Total proteins were extracted from breast cancer cell line, normal epithelial cell line, and transduced cells using ice-cold RIPA lysis buffer (Servicebio) supplemented with protease inhibitors. Following centrifugation at 15,000 × g for 30 min at 4°C, protein concentrations were determined using the BCA Protein Assay Kit (Servicebio). For FREM1 detection, 48 µg of total protein per sample was separated on 10% SDS-PAGE gels (Servicebio) and transferred to 0.45 µm PVDF membranes (Millipore Sigma). After blocking with 5% non-fat milk in TBST for 2 h at room temperature, membranes were incubated overnight at 4°C with primary antibodies: anti-FREM1 (1:1000, Proteintech, 13086-1-AP) and anti-GAPDH (1:5000, Proteintech, 60004-1-Ig) as loading control. Following three washes with TBST, membranes were incubated with HRP-conjugated secondary antibodies (1:5000, ZSGB-BIO, ZB-2301) for 1 h at room temperature. Protein signals were detected using the SuperPico ECL Chemiluminescent Substrate (Servicebio) and quantified using ImageJ software (NIH). Successful FREM1 overexpression was confirmed by ≥ 1.5-fold increase in protein expression compared to vector controls (p < 0.05), with all experiments performed in three independent biological replicates. EdU and Ki-67 Immunofluorescence Analysis of FREM1-Mediated Proliferation Inhibition in Breast Cancer Cells To assess the effect of FREM1 overexpression on breast cancer cell proliferation, we performed EdU incorporation and Ki-67 immunofluorescence assays. For EdU labeling, cells were seeded on coverslips in 48-well plates at a density of 5×10^4 cells/well and cultured for 8 hours. The cells were then incubated with 10 µM EdU (MCE) for 2 hours at 37°C, fixed with 4% paraformaldehyde for 30 minutes, and processed using the Click-iT EdU Alexa Fluor 594 Imaging Kit according to the manufacturer's protocol. For Ki-67 detection, cells grown on coverslips were fixed with 4% paraformaldehyde, permeabilized with 0.3% Triton X-100, and blocked with 5% Donkey Serum. The cells were then incubated overnight at 4°C with anti-Ki-67 antibody (Abclnal, A20018; 1:400 dilution), followed by incubation with Alexa Fluor 594-conjugated secondary antibody for 2 hours at room temperature. Nuclei were counterstained with DAPI (1 µg/mL) for 5 minutes. Fluorescence images were acquired using a confocal microscope (Leica TCS SP8), with at least five random fields captured per sample. The proliferation index was calculated as the percentage of EdU-positive or Ki-67-positive cells relative to the total number of DAPI-stained nuclei. Three independent biological replicates were performed for each experiment. Statistical Analysis All data analyses were conducted with R statistical software (v4.4.0). Continuous variables were analyzed using appropriate tests based on data distribution characteristics: parametric comparisons employed Student's t-test (two groups) or ANOVA (multiple groups), while nonparametric analyses utilized the Wilcoxon rank-sum test (two groups) or Kruskal-Wallis test (multiple groups). Bivariate associations were examined through correlation analysis, with Pearson's coefficient applied to normally distributed variables and Spearman's rank correlation used for non-normal distributions. Survival analysis was evaluated via Kaplan-Meier methodology, with between-group differences tested using the log-rank statistic. All tests were two-sided, and significance was set at p < 0.05 unless otherwise specified. Results ​​Identification and Characterization of Basement Membrane-Based Molecular Subtypes in Breast Cancer Figure 1 outlines our analytical workflow (Fig. 1 ). Consensus clustering of TCGA-BRCA samples using 222 basement membrane-related genes robustly segregated patients into two subtypes: a basement membrane-high group (C1) showing elevated expression of core components, and a basement membrane-low group (C2) with significantly reduced expression (Fig. 2 A). Kaplan-Meier analysis demonstrated significantly poorer overall survival in the C2 versus C1 group (p = 0.029, Fig. 2 B). The hierarchical clustering heatmap (Fig. 2 C) displays the expression profiles of the top 50 most significantly differentially expressed genes (|log2FC| >1, FDR < 0.05) between C1 and C2 subtypes, revealing distinct molecular signatures characteristic of each subgroup. Functional Annotation of BM-Associated DEGs Uncovers Their Involvement in Breast Cancer Microenvironment To investigate the biological significance of differentially expressed genes (DEGs) across basement membrane-associated subgroups, we performed integrated functional enrichment analyses including Gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG), and Gene Set Enrichment Analysis (GSEA). The GO analysis revealed significant enrichment in Biological Processes, with key pathways including epidermis and skin development, extracellular matrix and structure organization, chemokine-mediated signaling, antimicrobial humoral response, keratinization, and intermediate filament organization (Fig. 2 D). In the Cellular Component category, notable enrichments were observed in the cornified envelope, collagen-containing extracellular matrix, keratin and intermediate filaments, neuronal cell body, endoplasmic reticulum lumen, and neuron projection terminus (Fig. 2 D). Molecular Function analysis identified associations with cytokine and chemokine activity, structural constituents of skin epidermis, chemokine receptor binding, glycosaminoglycan binding, G protein-coupled receptor binding, and endopeptidase inhibitor activity (Fig. 2 D). KEGG pathway analysis revealed the top 10 significant pathways, including Cytokine-cytokine receptor interaction, Viral protein interaction with cytokine and cytokine receptor, Amoebiasis, IL-17 signaling pathway, Protein digestion and absorption, Rheumatoid arthritis, TNF signaling pathway, Chemokine signaling pathway, Neuroactive ligand-receptor interaction, and Hematopoietic cell lineage, as shown in Fig. 2 E. To identify pathway-level differences between BM expression subgroups, we conducted gene set enrichment analysis (GSEA) utilizing the Molecular Signatures Database (MSigDB) collections, including the GO gene sets (c5.go.symbols.gmt) and canonical KEGG pathways (c2.cp.kegg.symbols.gmt). In the C2 group, significant enrichment was observed in several immune-related KEGG pathways, including Cardiac Muscle Contraction, Drug Metabolism Other Enzymes, Glycosylphosphatidylinositol GPI Anchor Biosynthesis, Oxidative Phosphorylation, and Parkinson’s Disease. Additionally, the C2 group showed enrichment in GO terms such as Mitochondrial Respiratory Chain Complex Assembly, DNA Packaging Complex, Inner Mitochondrial Membrane Protein Complex, Respirasome, and Oxido-reduction Driven Active Transmembrane Transporter Activity (Fig. 2 F, G). In contrast, the C1 group was significantly enriched in KEGG pathways such as Cell Adhesion Molecules, Chemokine Signaling Pathway, Cytokine-Cytokine Receptor Interaction, Natural Killer Cell-Mediated Cytotoxicity, and Primary Immunodeficiency. The C1 group also showed enrichment in GO terms such as Granulocyte Chemotaxis, Leukocyte Chemotaxis, Keratin Filament, Chemokine Receptor Binding, and Cytokine Activity. These analyses suggest that basement membrane-related genes may play a role in modulating breast cancer microenvironment and metabolism. These analyses suggest that basement membrane-related genes may influence the breast cancer microenvironment and metabolism (Fig. 2 F, G). Comprehensive immune profiling revealed significant differences in the tumor microenvironment between BM-defined subgroups The C1 group demonstrated elevated ESTIMATE, immune, and stromal scores alongside reduced tumor purity compared to the C2 group (Fig. 3 A). Further analysis using CIBERSORT revealed substantial differences in the infiltration levels of various immune cells between the two basement membrane groups. These included naïve B cells, CD8 + T cells, memory resting CD4 + T cells, memory activated CD4 + T cells, follicular helper T cells, M1 macrophages, M2 macrophages, resting and activated dendritic cells, resting mast cells, and neutrophils (Fig. 3 B). These results highlight the crucial role of the basement membrane in influencing the breast cancer immune microenvironment. Furthermore, we observed significant differential expression of HLA genes and immune checkpoint molecules between the groups (Figs. 3 C, 3 D), highlighting the crucial role of basement membrane components in shaping the breast cancer immune landscape. These findings collectively demonstrate that BM-related gene expression patterns significantly influence both the cellular composition and immunoregulatory characteristics of the tumor microenvironment. Construction and Multi-Dimensional Validation of a Prognostic BMS Model in Breast Cancer​ Through univariate Cox regression analysis, we initially identified 18 basement membrane-related genes significantly associated with patient prognosis (Fig. 4 A). Subsequent LASSO regression analysis refined this set to 14 key prognostic genes, which were used to construct a predictive model named the Basement Membrane Signature (BMS) (Fig. 4 B). Each breast cancer sample was assigned a BMS risk score, with patients dichotomized into high- and low-risk groups based on the median score. Survival analysis confirmed significantly worse outcomes in the high-risk group, with robust reproducibility across both the TCGA (Fig. 4 C) training cohort (p < 0.001) and independent GEO (Fig. 4 D) validation dataset (p = 0.001), demonstrating the prognostic validity of this signature. Risk score distributions exhibited consistent patterns between TCGA (Fig. 4 E) and GEO (Fig. 4 F) cohorts, demonstrating the reproducibility of our risk stratification model. Comprehensive Cox regression analyses revealed distinct prognostic patterns across datasets. In the TCGA cohort, univariate and multivariate analyses consistently identified age, TNM stage, and BMS risk score as significant independent prognostic factors (all p < 0.001, Fig. 5 A). However, validation in the GEO dataset demonstrated that only BMS risk score maintained its prognostic significance (p = 0.001, Fig. 5 B), underscoring its robust predictive value across diverse patient populations. Discriminative ability was assessed through receiver operating characteristic (ROC) analysis. Comparative evaluation of age, TNM stage, and the BMS risk score yielded AUC values of 0.623, 0.651, and 0.705, respectively (Fig. 5 C). Time-dependent ROC analysis further demonstrated the stable predictive capacity of the BMS risk score across clinical endpoints, with AUCs of 0.659 (1-year), 0.707 (3-year), and 0.705 (5-year) survival (Fig. 5 D). Analysis of TCGA breast cancer samples demonstrated that patients in the high-risk group had significantly poorer outcomes in progression-free survival (PFS, p < 0.001) (Fig. 5 E), disease-free survival (DFS, p < 0.001) (Fig. 5 F), and disease-specific survival (DSS, p < 0.001) compared to the low-risk group (Fig. 5 G). To further validate these findings, we constructed prognostic nomograms and calibration curves for the predictive models incorporating age, TNM stage and the Basement Membrane Signature (BMS) risk score. These nomograms and calibration curves were developed to assess the predictive accuracy for overall survival (OS) (Supplementary Fig. 1A), progression free survival (PFS) (Supplementary Fig. 1B), disease free survival (DFS) (Supplementary Fig. 1C) and disease specific survival (DSS) (Supplementary Fig. 1D) outcomes. The close alignment between predicted and observed outcomes in both the nomograms and validation curves strongly supported the reliability of these prognostic models. Identification and Characterization of FREM1 as a Prognostic Biomarker in Cancer To refine our gene signature, we employed a machine learning approach using support vector machines (SVM), which identified a set of candidate genes (Fig. 5 H). Intersection of these SVM-derived candidates with our earlier LASSO-selected genes revealed six core prognostic genes: SEMA3B, ITGAX, FREM1, ADAM9, ADAMTS8, and UNC5A (Fig. 5 I). Subsequent validation in the TCGA cohort demonstrated that while all four genes (SEMA3B, ITGAX, FREM1, and ADAMTS8) showed significant associations with overall survival (OS) in Kaplan-Meier analysis (log-rank p < 0.05), FREM1 exhibited the most pronounced survival benefit (lowest p-value) (Supplementary Fig. 2A). Differential expression analysis between tumor and adjacent normal tissues revealed that FREM1 showed the most significant downregulation among all candidate genes (p < 0.05), with a consistent protective pattern (Supplementary Fig. 2B). Given its strongest prognostic performance, most remarkable tumor-specific expression pattern, and putative tumor-suppressive role, we prioritized FREM1 as our primary target for mechanistic investigation. Finally, using single-cell RNA sequencing data from two independent datasets-GSE148673 (Fig. 5 J) and GSE176078 (Fig. 5 K)-we performed a comprehensive analysis of FREM1 expression across diverse cell populations within the tumor microenvironment. The results demonstrated that FREM1 was predominantly expressed in cancer-associated fibroblasts (CAFs), with minimal expression in other cell types. This fibroblast-specific expression pattern suggests FREM1 may play a crucial role in tumor-stroma interactions and extracellular matrix remodeling. These findings highlight FREM1's unique position in shaping the tumor microenvironment and support its potential as both a stromal biomarker and therapeutic target in breast cancer progression. To functionally characterize FREM1, we performed in vitro studies using triple-negative MDA-MB-231 breast cancer cells compared to normal mammary epithelial cells (MCF-10A). Both qPCR (Fig. 6 A) and Western blot (Fig. 6 B, C) analyses confirmed significantly lower FREM1 expression in cancer cells, with > 1.44-fold reduction at mRNA level (p 1.77-fold decrease at protein level (p < 0.01) relative to normal controls. To investigate its functional role, we transfected MDA-MB-231 cells with a FREM1 overexpression plasmid, achieving 1.48-fold higher mRNA levels (p < 0.001) (Fig. 6 D) and 2.42-fold increased protein expression (Fig. 6 E, F) (p < 0.001) compared to vector controls. Functional assays demonstrated that FREM1 overexpression significantly inhibited cell proliferation, with Ki-67 immunofluorescence (Fig. 6 G, H) revealing 33.3% fewer proliferating cells (p < 0.001) and EdU incorporation assays (Fig. 6 I, J) showing 34.8% reduction in S-phase cells (p < 0.001). These results collectively demonstrate that FREM1 acts as a tumor suppressor in breast cancer by inhibiting cellular proliferation. Discussion Our study establishes a novel prognostic signature based on basement membrane (BM)-related genes, which effectively stratifies breast cancer patients into distinct risk groups with significant differences in clinical outcomes. This BM-related gene signature (BMS) not only demonstrates robust predictive accuracy for survival but also reveals the critical role of BM components in shaping tumor progression and the immune microenvironment. By integrating multi-omics data and functional validation, we highlight the potential of BMS as a valuable tool for prognosis assessment and personalized treatment strategies in breast cancer. Breast cancer represents a highly heterogeneous disease with diverse molecular subtypes and clinical outcomes, underscoring the need for precise prognostic biomarkers (21). Emerging evidence has highlighted the basement membrane (BM) as a dynamic structural and functional component that actively participates in tumor progression and metastasis (22, 23). Unlike traditional views of the BM as a passive barrier, recent studies demonstrate its crucial role in regulating cancer cell invasion, immune cell infiltration, and therapeutic resistance (19, 24). The BM's unique composition of collagen IV, laminins, and other glycoproteins forms a specialized extracellular matrix niche that influences tumor-stroma crosstalk and metastatic dissemination (25, 26). Our identification of BM-related gene signatures builds upon these findings by providing a systematic framework to quantify BM remodeling patterns and their clinical implications. This approach offers new opportunities to understand breast cancer heterogeneity beyond conventional molecular subtyping, potentially guiding more personalized treatment strategies. Our study systematically evaluated the prognostic value of basement membrane (BM)-related genes in breast cancer through a multi-stage analytical approach. Initially, consensus clustering analysis of BM-related gene expression patterns revealed two distinct molecular subtypes with significant clinical differences. The observed disparity in clinical outcomes between risk groups may be mechanistically linked to distinct immunological characteristics within the tumor microenvironment. These findings suggested that BM composition actively shapes both tumor behavior and immune responses (27, 28). To refine these observations into a clinically applicable tool, we employed machine learning techniques including LASSO regression and Cox proportional hazards modeling. This process identified a robust 14-gene signature that effectively stratified patients by risk across multiple independent cohorts. The signature showed significant prognostic discrimination, with high-risk patients exhibiting substantially worse clinical outcomes than low-risk patients. Importantly, the signature maintained its predictive power even after adjusting for conventional clinical parameters such as age and tumor stage in multivariate analyses. The clinical utility of this signature was further enhanced through the development of comprehensive nomograms that integrate BM-related gene expression with standard prognostic factors. These tools showed excellent predictive accuracy for both short-term and long-term outcomes, providing clinicians with a quantitative framework for risk assessment (29, 30). Mechanistically, the signature captures key biological processes in cancer progression, including extracellular matrix remodeling and Cytokine-cytokine receptor interaction. The strong association between signature scores and specific immune cell populations suggests potential utility in predicting response to immunotherapy. These findings significantly advance our understanding of BM biology in breast cancer, transforming it from a passive structural component to an active regulator of tumor progression and microenvironment modulation. The signature provides both prognostic value and biological insights, offering opportunities for more personalized treatment approaches. Future studies should explore its potential in guiding therapeutic decisions, particularly for aggressive subtypes where current prognostic tools remain limited. The integration of BM-related gene expression patterns into clinical decision-making represents a promising avenue for improving breast cancer management and outcomes. To identify clinically relevant targets from our basement membrane-related gene signature, we employed a machine learning pipeline combining LASSO regression and support vector machine (SVM) algorithms, which identified six high-confidence candidate genes. Among these, FREM1 emerged as the most consistently downregulated gene in breast tumor tissues compared to matched normal samples. To functionally validate these findings, we established FREM1-overexpressing MDA-MB-231 cells (MDA-FREM1) by transfecting with the FREM1 expression plasmid, with empty vector-transfected cells (MDA-EV) serving as controls. Successful overexpression was confirmed through qPCR analysis of mRNA levels and western blot detection of protein expression. Subsequent functional analyses revealed that FREM1 overexpression significantly suppressed tumor cell proliferation, as demonstrated by reduced Ki-67 expression and decreased EdU incorporation. These experimental findings are supported by accumulating clinical evidence establishing FREM1 as a tumor suppressor in breast cancer. Multiple independent cohorts have confirmed the significant downregulation of FREM1 in malignant versus normal breast epithelium. Importantly, reduced FREM1 expression correlates with aggressive clinicopathological features including advanced TNM stage and poorer survival outcomes, while also associating with an immunosuppressive tumor microenvironment marked by diminished cytotoxic lymphocyte infiltration and increased immunosuppressive cell populations (31, 32). Collectively, these data position FREM1 as a key regulator of breast cancer progression, with both prognostic implications and potential therapeutic relevance for overcoming tumor immunosuppression. Our study establishes the Basement Membrane Signature (BMS) as a novel prognostic tool in breast cancer, demonstrating its ability to stratify patients into distinct risk groups with significant differences in clinical outcomes. By integrating multi-omics data and machine learning approaches, we developed a robust 14-gene signature that effectively predicts survival and reflects key biological processes, including extracellular matrix remodeling and immune microenvironment modulation. However, several limitations should be acknowledged. First, our findings are based on retrospective analyses of publicly available datasets, which may be subject to selection biases and technical variability. Although we validated the BMS in both TCGA and GEO cohorts, further confirmation in prospective, multi-center studies is needed to ensure its clinical applicability. Second, while we identified FREM1 as a potential tumor suppressor through in vitro experiments, the functional roles of other signature genes in breast cancer progression remain to be fully elucidated. Additional mechanistic studies, including in vivo models and patient-derived xenografts, are required to validate their biological significance and therapeutic potential. Moreover, the molecular pathways linking basement membrane remodeling to tumor progression and immune evasion warrant further investigation. Future research should explore whether targeting these pathways could enhance treatment efficacy, particularly in aggressive subtypes. Despite these limitations, our work provides a foundation for incorporating basement membrane biology into breast cancer prognostication and therapy. The BMS model offers a clinically actionable tool that complements existing prognostic markers, while the identified genes present new opportunities for therapeutic development. Continued validation and functional studies will be essential to translate these findings into improved patient outcomes. Conclusion Our study establishes the Basement Membrane Signature (BMS) as a robust prognostic tool in breast cancer, effectively stratifying patients into distinct risk groups with significant survival differences. By integrating multi-omics data and machine learning, we developed a 14-gene signature that reflects key biological processes including extracellular matrix remodeling and immune microenvironment modulation. The BMS demonstrated consistent predictive accuracy across multiple cohorts, with high-risk patients showing significantly poorer prognosis compared to low-risk patients. Functional validation identified FREM1 as a potential tumor suppressor, providing mechanistic insights into basement membrane biology. While retrospective in nature, our findings highlight the clinical potential of BMS for risk assessment and treatment guidance. Future studies should focus on prospective validation, mechanistic exploration of basement membrane-immune interactions, and therapeutic targeting of identified pathways to improve patient outcomes. This work advances our understanding of breast cancer heterogeneity and provides a foundation for more personalized management strategies. Declarations Acknowledgments: Not applicable. Author contributions: C.L.and C.L. wrote the main manuscript text. C.L. and P.G. performed comprehensive data analysis. C.L. and J.H. created visualizations. All authors reviewed the manuscript. Funding: None. Conflicts of Interest: The authors have no conflicts of interest to declare. Clinical Trial Number: Not applicable. Consent to Publish Declaration: Not applicable. Consent to Participate Declaration: Not applicable. Ethics Declaration: Not applicable. Data Availability Statements： The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request. References Sung H, Ferlay J, Siegel RL, Laversanne M, Soerjomataram I, Jemal A, et al. Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. CA Cancer J Clin. 2021;71(3):209-49. Liu Y, Ouyang W, Huang H, Tan Y, Zhang Z, Yu Y, et al. Identification of a tumor immune-inflammation signature predicting prognosis and immune status in breast cancer. Front Oncol. 2022;12:960579. Pasta V, Monti M, Cialini M, Vergine M, Urciuoli P, Iacovelli A, et al. Primitive sarcoma of the breast: new insight on the proper surgical management. J Exp Clin Cancer Res. 2015;34(1):72. Liao Q, Deng D, Xie Q, Gong X, Meng X, Xia Y, et al. Clinical characteristics, pregnancy outcomes and ovarian function of pregnancy-associated breast cancer patients: a retrospective age-matched study. BMC Cancer. 2022;22(1):152. Dong J, Yang Y, Han D, Zhao Q, Liu C, Sun H, et al. Hypofractionated Simultaneous Integrated Boost Radiotherapy Versus Conventional Fractionation Radiotherapy of Early Breast Cancer After Breast-Conserving Surgery: Clinical Observation and Analysis. Technol Cancer Res Treat. 2021;20:15330338211064719. Toivonen KI, Williamson TM, Carlson LE, Walker LM, Campbell TS. Potentially Modifiable Factors Associated with Adherence to Adjuvant Endocrine Therapy among Breast Cancer Survivors: A Systematic Review. Cancers (Basel). 2020;13(1). Kinnel B, Singh SK, Oprea-Ilies G, Singh R. Targeted Therapy and Mechanisms of Drug Resistance in Breast Cancer. Cancers (Basel). 2023;15(4). Dai G, Yang Y, Liu S, Liu H. Hypoxic Breast Cancer Cell-Derived Exosomal SNHG1 Promotes Breast Cancer Growth and Angiogenesis via Regulating miR-216b-5p/JAK2 Axis. Cancer Manag Res. 2022;14:123-33. Li W, Liu J, Zhang B, Bie Q, Qian H, Xu W. Transcriptome Analysis Reveals Key Genes and Pathways Associated with Metastasis in Breast Cancer. Onco Targets Ther. 2020;13:323-35. Du X, Zhang J, Liu L, Xu B, Han H, Dai W, et al. A novel anticancer property of Lycium barbarum polysaccharide in triggering ferroptosis of breast cancer cells. J Zhejiang Univ Sci B. 2022;23(4):286-99. Li L, Wei D, Zhang J, Deng R, Tang J, Su D. miR-641 Inhibited Cell Proliferation and Induced Apoptosis by Targeting NUCKS1/PI3K/AKT Signaling Pathway in Breast Cancer. Comput Math Methods Med. 2022;2022:5203839. Agostinetto E, Montemurro F, Puglisi F, Criscitiello C, Bianchini G, Del Mastro L, et al. Immunotherapy for HER2-Positive Breast Cancer: Clinical Evidence and Future Perspectives. Cancers (Basel). 2022;14(9). Mills MN, Figura NB, Arrington JA, Yu H-HM, Etame AB, Vogelbaum MA, et al. Management of brain metastases in breast cancer: a review of current practices and emerging treatments. Breast Cancer Res Treat. 2020;180(2):279-300. Gómez-Aleza C, Nguyen B, Yoldi G, Ciscar M, Barranco A, Hernández-Jiménez E, et al. Inhibition of RANK signaling in breast cancer induces an anti-tumor immune response orchestrated by CD8+ T cells. Nat Commun. 2020;11(1):6335. Jayadev R, Morais MRPT, Ellingford JM, Srinivasan S, Naylor RW, Lawless C, et al. A basement membrane discovery pipeline uncovers network complexity, regulators, and human disease associations. Sci Adv. 2022;8(20):eabn2265. Jayadev R, Sherwood DR. Basement membranes. Curr Biol. 2017;27(6):R207-R11. Zhang Z, Zhu H, Wang X, Lin S, Ruan C, Wang Q. A novel basement membrane-related gene signature for prognosis of lung adenocarcinomas. Comput Biol Med. 2023;154:106597. Sun W, Wang J, Wang Z, Xu M, Lin Q, Sun P, et al. Combining WGCNA and machine learning to construct basement membrane-related gene index helps to predict the prognosis and tumor microenvironment of HCC patients and verifies the carcinogenesis of key gene CTSA. Front Immunol. 2023;14:1185916. Tian W, Luo Y, Tang Y, Kong Y, Wu L, Zheng S, et al. Novel Implication of the Basement Membrane for Breast Cancer Outcome and Immune Infiltration. Int J Biol Sci. 2023;19(5):1645-63. Lu P, Weaver VM, Werb Z. The extracellular matrix: a dynamic niche in cancer progression. J Cell Biol. 2012;196(4):395-406. Hossain FM, Danos DM, Fu Q, Wang X, Scribner RA, Chu ST, et al. Association of Obesity and Diabetes With the Incidence of Breast Cancer in Louisiana. Am J Prev Med. 2022;63(1 Suppl 1):S83-S92. Tang X, Liu Y, Zhao J, Fu C, Yang W. Subtyping of gastric cancer based on basement membrane genes that stratifies the prognosis, immune infiltration and therapeutic response. Discov Oncol. 2024;15(1):362. Chang J, Saraswathibhatla A, Song Z, Varma S, Sanchez C, Alyafei NHK, et al. Cell volume expansion and local contractility drive collective invasion of the basement membrane in breast cancer. Nat Mater. 2024;23(5):711-22. Cai J, Zhang X, Xie W, Li Z, Liu W, Liu A. Identification of a basement membrane-related gene signature for predicting prognosis and estimating the tumor immune microenvironment in breast cancer. Front Endocrinol (Lausanne). 2022;13:1065530. Stricker AM, Hutson MS, Page-McCaw A. Piezo-dependent surveillance of matrix stiffness generates transient cells that repair the basement membrane. Dev Cell. 2025. Wang J, Lin S, Wei Y, Ye Z. Recombinant human collagen XVII protects skin basement membrane integrity by inhibiting the MAPK and Wnt signaling pathways. Mol Med Rep. 2025;31(4). Walter C, Davis JT, Mathur J, Pathak A. Physical defects in basement membrane-mimicking collagen-IV matrices trigger cellular EMT and invasion. Integr Biol (Camb). 2018;10(6):342-55. Wu Y, Liu X, Zhu Y, Qiao Y, Gao Y, Chen J, et al. Type IV collagen α5 chain promotes luminal breast cancer progression through c-Myc-driven glycolysis. J Mol Cell Biol. 2023;14(10). Xie L, Zhang Y, Niu X, Jiang X, Kang Y, Diao X, et al. A nomogram for predicting cancer-specific survival in patients with locally advanced unresectable esophageal cancer: development and validation study. Front Immunol. 2025;16:1524439. Hu J, Shi Y, Jin L, Yi S, Chen J, Wan D, et al. Predicting survival rates: the power of prognostic nomograms in distal cholangiocarcinoma. Front Oncol. 2025;15:1478836. Xu X-Y, Guo W-J, Pan S-H, Zhang Y, Gao F-L, Wang J-T, et al. TILRR (FREM1 isoform 2) is a prognostic biomarker correlated with immune infiltration in breast cancer. Aging (Albany NY). 2020;12(19):19335-51. Li H-N, Li X-R, Lv Z-T, Cai M-M, Wang G, Yang Z-F. Elevated expression of FREM1 in breast cancer indicates favorable prognosis and high-level immune infiltration status. Cancer Med. 2020;9(24):9554-70. Additional Declarations No competing interests reported. Supplementary Files floatimage7.png Supplymentary Figure 1. Prognostic nomograms and calibration curves for survival outcomes in breast cancer. (A) Overall Survival (OS) Nomogram and Calibration Curve. (B) Progression-Free Survival (PFS) Nomogram and Calibration Curve​. (C) Disease-Free Survival (DFS) Nomogram and Calibration Curve. (D) Disease-Specific Survival (DSS) Nomogram and Calibration Curve. floatimage8.png Supplymentary Figure 2. Prognostic and differential expression analysis of six candidate genes ( SEMA3B, ITGAX, FREM1, ADAM9, ADAMTS8 , and UNC5A ) in breast cancer. (A) Kaplan-Meier survival curves for overall survival (OS) stratified by high (red) and low (blue) expression levels of each gene in the TCGA-BRCA cohort. (B) Comparative analysis of gene expression between primary breast tumor tissues (red) and adjacent normal tissues (blue) from TCGA-BRCA. Cite Share Download PDF Status: Under Review Version 1 posted Editorial decision: Revision requested 28 Aug, 2025 Reviews received at journal 25 Aug, 2025 Reviews received at journal 23 Aug, 2025 Reviewers agreed at journal 18 Aug, 2025 Reviewers agreed at journal 18 Aug, 2025 Reviewers invited by journal 13 Aug, 2025 Editor assigned by journal 08 Aug, 2025 Editor invited by journal 08 Aug, 2025 Submission checks completed at journal 02 Aug, 2025 First submitted to journal 02 Aug, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7206387","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":501938771,"identity":"7ebdecf0-a13f-48d6-94fc-b8f4420b7684","order_by":0,"name":"Chao Li","email":"","orcid":"","institution":"Beijing Anzhen Hospital","correspondingAuthor":false,"prefix":"","firstName":"Chao","middleName":"","lastName":"Li","suffix":""},{"id":501938772,"identity":"ac9b53a8-e62f-47d6-900e-e32ae5ceef2b","order_by":1,"name":"Pingming Gong","email":"","orcid":"","institution":"Beijing Luhe Hospital Affiliated to Capital Medical University","correspondingAuthor":false,"prefix":"","firstName":"Pingming","middleName":"","lastName":"Gong","suffix":""},{"id":501938773,"identity":"60829775-4c2c-4fdb-8672-2114479651cb","order_by":2,"name":"Junfeng Hu","email":"","orcid":"","institution":"Beijing Luhe Hospital Affiliated to Capital Medical University","correspondingAuthor":false,"prefix":"","firstName":"Junfeng","middleName":"","lastName":"Hu","suffix":""},{"id":501938774,"identity":"40194ec8-1b01-45da-9bd0-9865236e15d1","order_by":3,"name":"Chengyu Luo","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAxElEQVRIiWNgGAWjYBACPiBmhjCZDxz48IMILWwILWyJB2f2kKaFx/gwBxsxWiSSD34uqLhjt+FGzofDDDwM8vxiBwhpSUuWnnHmWfKGG7kbDhdYMBjOnJ1ASEuOGTNv2+FkA5CWGTwMCQa3idLyD6Ql58FhHjaitTQctgNqYSBSC8+zZGmeY4cTJM88MwAGsgRhv/CzA0OMp+awPd/x5McfPvywkeeXJqAFBhIXHADTEsQpBwF7+QbiFY+CUTAKRsEIAwBwlUTtJUuoowAAAABJRU5ErkJggg==","orcid":"","institution":"Beijing Anzhen Hospital","correspondingAuthor":true,"prefix":"","firstName":"Chengyu","middleName":"","lastName":"Luo","suffix":""}],"badges":[],"createdAt":"2025-07-24 14:08:12","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7206387/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7206387/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":89837833,"identity":"a6a4b68e-34c5-4cbc-9a57-e7e878137f29","added_by":"auto","created_at":"2025-08-25 15:00:32","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":300783,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eFlowchart of the study design.\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7206387/v1/42ab825387658b85b795ec6b.png"},{"id":89839120,"identity":"bc8f7a24-0565-467d-9e13-0c20c077bd43","added_by":"auto","created_at":"2025-08-25 15:08:33","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":624431,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eMolecular subtyping and functional characterization of basement membrane-related genes in breast cancer.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e(A) Unsupervised clustering of TCGA-BRCA samples based on 222 basement membrane-related genes identified two distinct molecular subtypes (C1 and C2), demonstrating significant transcriptional heterogeneity.\u003c/p\u003e\n\u003cp\u003e(B) Kaplan-Meier survival analysis revealed markedly different clinical outcomes between subtypes (log-rank p \u0026lt; 0.05), establishing the prognostic significance of basement membrane gene expression patterns.\u003c/p\u003e\n\u003cp\u003e(C) Hierarchical clustering heatmap visualized the differential expression profiles of top 50 basement membrane-related genes across subtypes. (D-G) Functional enrichment analyses (GO, KEGG, and GSEA) of differentially expressed genes revealed significant associations with extracellular matrix remodeling and immune regulation processes.\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7206387/v1/e8074389c1306c775d20db93.png"},{"id":89837835,"identity":"ffd3fb92-447a-48f8-a580-0f34921dd6d9","added_by":"auto","created_at":"2025-08-25 15:00:32","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":359643,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eTumor-immune microenvironment disparities between basement membrane-defined subgroups in breast cancer.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e(A) Comparative assessment of tumor microenvironment components using ESTIMATE algorithm-derived scores.\u003c/p\u003e\n\u003cp\u003e(B) Differential infiltration patterns of 22 immune cell subtypes identified by CIBERSORT analysis.\u003c/p\u003e\n\u003cp\u003e(C) Distinct expression profiles of HLA (Human Leukocyte Antigen) genes, indicating altered immunogenicity between clusters.\u003c/p\u003e\n\u003cp\u003e(D) Cluster-specific immune checkpoint expression patterns associated with potential evasion mechanisms.\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7206387/v1/240d2bf17b3be8e6de4bca3c.png"},{"id":89837839,"identity":"e9c96af3-a9b8-4f6c-a104-dd356ed80dc1","added_by":"auto","created_at":"2025-08-25 15:00:32","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":517764,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eConstruction and validation of the basement membrane-related prognostic signature in breast cancer​.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e(A) Univariate Cox regression of potential biomarkers.\u003c/p\u003e\n\u003cp\u003e(B) LASSO regression-derived 14-gene signature.\u003c/p\u003e\n\u003cp\u003e(C-D) Survival curves comparing high- and low-risk groups in TCGA (left) and GEO (right) datasets.\u003c/p\u003e\n\u003cp\u003e(E) Association between risk stratification and clinical outcomes in the TCGA cohort, along with the expression heatmap of the 14-gene signature across high- and low-risk groups.\u003c/p\u003e\n\u003cp\u003e(F) External independent validation of the prognostic model in the GEO cohort.​\u003c/p\u003e","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7206387/v1/b33e3c535ab72eb57d13a13b.png"},{"id":89837842,"identity":"af408e99-4a10-44aa-af89-0badd4c07655","added_by":"auto","created_at":"2025-08-25 15:00:33","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":348947,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eComprehensive analysis of prognostic factors and tumor microenvironment in breast cancer.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e(A) Univariate and multivariate Cox regression analyses of TCGA-derived clinical features (Age, Stage, and Riskscore) identifying independent prognostic factors in AML.\u003cbr\u003e\n(B) Validation of prognostic factors through univariate and multivariate Cox regression using GEO database.\u003cbr\u003e\n(C, D) ROC curve analyses​​ demonstrated the predictive performance of Age, Stage, and Riskscore, followed by time-dependent evaluation (1-/3-/5-year) using Riskscore alone. ​​\u003cbr\u003e\n(E-G) Survival outcomes​​ were compared between high- and low-risk groups, including progression-free survival (PFS), disease-free survival (DFS), and disease-specific survival (DSS).\u003cbr\u003e\n(H) Support vector machine (SVM) algorithm for feature selection.\u003cbr\u003e\n(I) Venn diagram showing overlapping genes identified by both LASSO regression and SVM.\u003cbr\u003e\n(J, K) Single-cell RNA-seq analysis of FREM1 expression distribution in breast cancer microenvironment from GSE148673 and GSE176078 datasets.\u003c/p\u003e","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-7206387/v1/558a19f33102f5e570148b05.png"},{"id":89837853,"identity":"7bf65921-dddf-4256-b8f4-ffdd5623083c","added_by":"auto","created_at":"2025-08-25 15:00:33","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":368576,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eExperimental validation of FREM1 overexpression and functional characterization in breast cancer cells.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e(A) qPCR analysis of FREM1 transcriptional levels in MCF-7 and MDA-MB-231 cell lines.\u003c/p\u003e\n\u003cp\u003e(B) Western blot validation of FREM1 protein expression.\u003c/p\u003e\n\u003cp\u003e(C) Quantitative analysis of Western blot results from panel B.\u003c/p\u003e\n\u003cp\u003e(D) qPCR confirmation of FREM1 overexpression following plasmid transfection in MDA-MB-231 cells.\u003c/p\u003e\n\u003cp\u003e(E) Western blot verification of FREM1 protein overexpression.\u003c/p\u003e\n\u003cp\u003e(F) Statistical analysis of Western blot data from panel E.\u003c/p\u003e\n\u003cp\u003e(G) Ki-67 staining for proliferation evaluation.\u003c/p\u003e\n\u003cp\u003e(H) Statistical analysis of the proportion of Ki-67 positive cells.\u003c/p\u003e\n\u003cp\u003e(I) Cell proliferation assessment by EdU assay.\u003c/p\u003e\n\u003cp\u003e(J) Quantitative analysis of the proportion of EdU-positive cells.\u003c/p\u003e","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-7206387/v1/c94c47eca4de528424da7ea1.png"},{"id":89841476,"identity":"df968422-311f-4c92-b3d0-5380160456f0","added_by":"auto","created_at":"2025-08-25 15:24:34","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":3762652,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7206387/v1/d2cd0ff3-b2df-4f49-9479-ceb581fa2b75.pdf"},{"id":89837834,"identity":"b494fca4-de47-4fe0-8c60-638885de0886","added_by":"auto","created_at":"2025-08-25 15:00:32","extension":"png","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":314874,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSupplymentary Figure 1. Prognostic nomograms and calibration curves for survival outcomes in breast cancer.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e(A) Overall Survival (OS) Nomogram and Calibration Curve.\u003c/p\u003e\n\u003cp\u003e(B) Progression-Free Survival (PFS) Nomogram and Calibration Curve​.\u003c/p\u003e\n\u003cp\u003e(C) Disease-Free Survival (DFS) Nomogram and Calibration Curve.\u003c/p\u003e\n\u003cp\u003e(D) Disease-Specific Survival (DSS) Nomogram and Calibration Curve.\u003c/p\u003e","description":"","filename":"floatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-7206387/v1/1ce1fb5d845f3eca9a59e74f.png"},{"id":89839121,"identity":"2a10604d-bec3-433c-81d8-7a97f11868c5","added_by":"auto","created_at":"2025-08-25 15:08:33","extension":"png","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":452798,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSupplymentary Figure 2. Prognostic and differential expression analysis of six candidate genes (\u003c/strong\u003e\u003cem\u003e\u003cstrong\u003eSEMA3B, ITGAX, FREM1, ADAM9, ADAMTS8\u003c/strong\u003e\u003c/em\u003e\u003cstrong\u003e, and \u003c/strong\u003e\u003cem\u003e\u003cstrong\u003eUNC5A\u003c/strong\u003e\u003c/em\u003e\u003cstrong\u003e) in breast cancer.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e(A) Kaplan-Meier survival curves for overall survival (OS) stratified by high (red) and low (blue) expression levels of each gene in the TCGA-BRCA cohort.\u003c/p\u003e\n\u003cp\u003e(B) Comparative analysis of gene expression between primary breast tumor tissues (red) and adjacent normal tissues (blue) from TCGA-BRCA.\u003c/p\u003e","description":"","filename":"floatimage8.png","url":"https://assets-eu.researchsquare.com/files/rs-7206387/v1/c4faacc15203b9cdba2361b5.png"}],"financialInterests":"No competing interests reported.","formattedTitle":"FREM1 Serves as a Novel Therapeutic Target in Breast Cancer through Basement Membrane-Based Prognostic Modeling with Integrated Bioinformatics and Experimental Validation","fulltext":[{"header":"Introduction","content":"\u003cp\u003eBreast cancer (BC) is the most prevalent malignancy among women worldwide, with an estimated 2.3\u0026nbsp;million new cases diagnosed each year (1), making it the leading cause of cancer incidence globally. Breast cancer is the leading cause of cancer-related deaths among women globally (2), accounting for the fifth highest mortality rate across all cancers (1). Breast cancer management currently involves a range of treatments, including surgery (3), chemotherapy (4), radiotherapy (5), endocrine therapy (6), and targeted therapies (7). Despite these options, their effectiveness is often limited, especially in cases of metastatic breast cancer, where long-term survival outcomes remain poor (8, 9). Recent advancements in immunotherapy have significantly impacted breast cancer treatment (10), offering particularly promising results for triple-negative breast cancer (TNBC) (11) and HER2-positive breast cancer (12). The emergence of immune checkpoint inhibitors has opened new avenues for strengthening localized anti-tumor immune responses (13). Nonetheless, while certain patients experience notable clinical benefits, the majority continue to face challenges of disease progression, largely attributed to the development of primary or acquired resistance (14). Enhancing treatment efficacy requires a deeper understanding of the interaction between breast cancer cells and the immune microenvironment, along with the mechanisms underlying immune evasion. Additionally, the discovery of reliable biomarkers can significantly advance the development of immunotherapy strategies.\u003c/p\u003e\u003cp\u003eThe extracellular matrix (ECM) is central to the tumor microenvironment, with the basement membrane (BM) acting as a barrier that prevents cancer cells from spreading (15). Abnormal regulation of the BM can, therefore, facilitate tumor invasion and metastasis (16). Numerous studies have demonstrated that basement membrane-related genes are linked to the prognosis of various cancers, including Lung adenocarcinoma (17) and hepatocellular carcinoma (18). Emerging studies indicate that the progression of breast cancer is driven not only by the tumor cells themselves but also by a significantly altered tumor microenvironment (19). The extracellular matrix (ECM), particularly the basement membrane (BM), is pivotal in regulating various aspects of tumor biology, including cell migration and invasion (19). Alterations in the ECM, like collagen IV breakdown, commonly induced by immune cells and cancer-associated fibroblasts, can create a pathway for tumor cells to invade through the basement membrane (20). These alterations point to a potential yet underexplored connection between the BM and tumor immune infiltration.\u003c/p\u003e\u003cp\u003eIn this study, we developed a BM-related gene signature to stratify breast cancer prognosis, rigorously validating its predictive performance using TCGA and GEO datasets. The model demonstrated independent prognostic value and revealed distinct tumor microenvironment profiles, including differential immune infiltration patterns that may influence immunotherapy responses. Through machine learning, we identified FREM1 as a key BM-associated molecular target, subsequently validating its tumor-suppressive role via functional assays. This integrated approach not only enhances prognostic precision but also uncovers novel therapeutic targets for breast cancer and its immune modulation.\u003c/p\u003e"},{"header":"Materials and Methods","content":"\u003cp\u003e\u003cb\u003eData Collection and Processing\u003c/b\u003e\u003c/p\u003e\u003cp\u003eThe breast cancer (BRCA) RNA-seq transcriptome data and clinical information were obtained from The Cancer Genome Atlas (TCGA, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://cancergenome.nih.gov/\u003c/span\u003e\u003cspan address=\"https://cancergenome.nih.gov/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) as the primary dataset. To ensure the robustness of survival analysis, samples lacking survival data were excluded. The data were downloaded in fragments per kilobase million (FPKM) format for consistency in downstream analyses. For external validation, we utilized the GSE131769 dataset from the Gene Expression Omnibus (GEO, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://www.ncbi.nlm.nih.gov/geo\u003c/span\u003e\u003cspan address=\"http://www.ncbi.nlm.nih.gov/geo\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e), which provides survival information and clinical outcomes. Additionally, two single-cell RNA-seq datasets, GSE148673 and GSE176078, were acquired for investigating the expression patterns of target genes within the breast cancer microenvironment. Based on a thorough review of the literature, we also compiled a list of 222 basement membrane-related genes for further analysis in this study (15).\u003c/p\u003e\u003cp\u003e\u003cb\u003eMolecular Subtyping via Consensus Clustering​\u003c/b\u003e\u003c/p\u003e\u003cp\u003eTo delineate clinically relevant molecular subtypes based on basement membrane gene expression patterns, we implemented an unsupervised consensus clustering approach. The analysis was conducted using the \"ConsensusClusterPlus\" R package with the following parameters: 100 resampling iterations to ensure robust cluster stability, a pItem value of 0.8 to maintain sample assignment consistency, and Euclidean distance metric with hierarchical clustering. Systematic evaluation of the consensus matrix and cumulative distribution function curves revealed optimal separation into two distinct molecular subtypes. Subsequent survival analysis demonstrated significant differences in overall survival between these subtypes (log-rank p\u0026thinsp;\u0026lt;\u0026thinsp;0.05), confirming the clinical relevance of the basement membrane-related gene signatures in breast cancer prognosis.\u003c/p\u003e\u003cp\u003e\u003cb\u003eDifferential Expression and Functional Enrichment Analysis\u003c/b\u003e\u003c/p\u003e\u003cp\u003eTo characterize the molecular distinctions between the identified breast cancer subtypes, we performed differential gene expression analysis using the \"limma\" package in R, applying thresholds of |logFC| \u0026gt;1 and adjusted p-value\u0026thinsp;\u0026lt;\u0026thinsp;0.05. Significantly differentially expressed genes (DEGs) were then functionally annotated using the \"clusterProfiler\" package. Gene Ontology (GO) analysis categorized DEGs into biological processes, cellular components, and molecular functions, while KEGG pathway analysis identified enriched signaling pathways. Additionally, Gene Set Enrichment Analysis (GSEA) was conducted using the \"c5.go.symbols.gmt\" and \"c2.cp.kegg.symbols.gmt\" databases from the MSigDB collection to further validate pathway-level associations. This integrated approach provided a systematic framework for interpreting the biological relevance of subtype-specific gene signatures.\u003c/p\u003e\u003cp\u003e\u003cb\u003eAnalysis of Tumor Microenvironment Components\u003c/b\u003e\u003c/p\u003e\u003cp\u003eTo characterize the tumor microenvironment across molecular subtypes, we employed a comprehensive analytical approach. First, we calculated ESTIMATE (Est​​imation of ​​St​​romal and ​​Imm​​une cells in ​​M​​alignant ​​T​​umor tissues using ​​E​​xpression data) scores to quantify stromal and immune cell infiltration levels as well as tumor purity using standardized parameters. Next, we performed immune cell deconvolution using the CIBERSORTx (C​​ell ​​I​​dentity ​​B​​y ​​E​​stimated ​​R​​egression ​​S​​orting ​​O​​f ​​R​​NA ​​T​​ranscripts) algorithm (v1.06) with the LM22 signature matrix and 1000 permutations to estimate the relative proportions of 22 immune cell subtypes, applying a significance threshold of p\u0026thinsp;\u0026lt;\u0026thinsp;0.05 for reliable detection. Additionally, we systematically evaluated the expression profiles of immunomodulatory markers, including HLA (​​H​​uman ​​L​​eukocyte ​​A​​ntigen) gene family members and key immune checkpoint molecules, using one-way ANOVA with Benjamini-Hochberg correction for multiple testing. This integrated analysis enabled quantitative assessment of the tumor-immune interface characteristics across different molecular subtypes.\u003c/p\u003e\u003cp\u003e\u003cb\u003eDevelopment and Validation of a Basement Membrane-Related Prognostic Signature\u003c/b\u003e\u003c/p\u003e\u003cp\u003eTo establish a robust prognostic signature, we first performed univariate Cox regression analysis to identify basement membrane (BM)-related genes significantly associated with overall survival (p\u0026thinsp;\u0026lt;\u0026thinsp;0.05). These candidate genes were subsequently subjected to Lasso regression with 10-fold cross-validation to prevent overfitting and generate a refined risk score model. Patients were dichotomized into high- and low-risk groups using the median risk score as the cutoff value. The model's predictive performance was systematically evaluated through Kaplan-Meier survival curves and risk score distribution visualization. External validation was conducted using an independent GEO cohort to confirm clinical applicability.\u003c/p\u003e\u003cp\u003eTo enhance biomarker discovery, we integrated machine learning approaches by combining Lasso regression with support vector machine (SVM) recursive feature elimination. This dual-selection strategy identified consensus prognostic genes exhibiting both statistical significance and biological relevance. The machine learning framework enabled detection of complex genomic patterns that conventional statistical methods might miss, thereby improving the model's predictive accuracy for personalized treatment stratification.\u003c/p\u003e\u003cp\u003e\u003cb\u003eComprehensive Analysis of Key Genes in Breast Cancer Expression Prognosis and Tumor Microenvironment\u003c/b\u003e\u003c/p\u003e\u003cp\u003eTo refine the selection of clinically relevant genes, we systematically analyzed the machine learning-derived candidate genes using TCGA data. We initially conducted Kaplan-Meier survival analysis to investigate potential correlations between gene expression patterns and clinical outcomes in breast cancer patients. We subsequently examined differential gene expression between tumor and adjacent normal tissues to elucidate their possible involvement in breast cancer pathogenesis. This sequential analytical strategy - progressing from clinical relevance to mechanistic exploration - enabled systematic identification of the most promising candidate genes for further investigation. Furthermore, we performed single-cell RNA sequencing analysis to delineate the precise cellular localization and expression patterns of these genes across different cell populations in breast tumors. This comprehensive approach provided critical insights into the potential functional roles of these genes in tumor biology and disease progression.\u003c/p\u003e\u003cp\u003e\u003cb\u003eCell Line Acquisition and Culture Conditions​\u003c/b\u003e\u003c/p\u003e\u003cp\u003eWe utilized two breast cell lines for our experiments: the highly metastatic MDA-MB-231 cancer cells and non-tumorigenic MCF-10A normal epithelial cells, both obtained from Procell Life Science \u0026amp; Technology Co., Ltd (Wuhan, China). These cell lines were maintained under standard culture conditions (37\u0026deg;C, 5% CO\u003csub\u003e2\u003c/sub\u003e) using distinct growth media formulations. The MDA-MB-231 line was propagated in high-glucose DMEM supplemented with 10% FBS and antibiotic solution. The MCF-10A normal epithelial cells required a specialized DMEM/F12-based medium containing 5% equine serum along with specific growth factors including EGF (20 ng/mL), hydrocortisone (0.5 \u0026micro;g/mL), insulin (10 \u0026micro;g/mL), and NEAA supplementation.\u003c/p\u003e\u003cp\u003e\u003cb\u003ePlasmid Construction and Transfection\u003c/b\u003e\u003c/p\u003e\u003cp\u003eThe FREM1 coding sequence was PCR-amplified from MDA-MB-231 cell cDNA using specific primers containing KpnI and XhoI restriction sites (Forward: 5'-GCCGGTACCGCCACCATGGTGACACAAGAATCCATGCTG-3'; Reverse: 5'-GGCCTCGAGTTACTTGTCATCGTCGTCCTTGTAATCGAGTTTTCTGGAACACAC-3'). The amplification was carried out for 30 cycles under standard conditions: 95\u0026deg;C for 15 seconds, 68\u0026deg;C for 15 seconds, and 72\u0026deg;C for 1 minute per cycle. Following amplification, the PCR product was digested with the corresponding restriction enzymes (Transgen, Beijing, China) at 37\u0026deg;C for 30 minutes to generate compatible ends for subsequent cloning. The resulting FREM1 fragment was then ligated into the expression vector and transformed into competent cells for plasmid propagation. For functional studies, the verified FREM1 overexpression plasmid was introduced into MDA-MB-231 cells using Lipo8000 transfection reagent (Beyotime, Shanghai, China) following the manufacturer's recommended protocol.\u003c/p\u003e\u003cp\u003e\u003cb\u003eQuantitative PCR Analysis for Cell Line Screening and Transduction Validation\u003c/b\u003e\u003c/p\u003e\u003cp\u003eTo identify breast cancer cell lines with low endogenous FREM1 expression and validate successful plasmid transfection, we performed quantitative real-time PCR (qPCR) analysis. Total RNA was extracted from cell lines using TRIzol reagent (Servicebio) followed by DNase I treatment to eliminate genomic DNA contamination. cDNA synthesis was carried out with 1 \u0026micro;g total RNA using HiScript II Reverse Transcriptase (Servicebio) and oligo (dT) primers.\u003c/p\u003e\u003cp\u003eFor qPCR amplification, reactions were prepared in triplicate with TB Green Premix Ex Taq II (Servicebio) and run on a LightCycler 480 system (Roche) under the following conditions: 95\u0026deg;C for 30 sec, followed by 40 cycles of 95\u0026deg;C for 5 sec and 60\u0026deg;C for 30 sec. FREM1-specific primers (Forward: 5'-GCCTGTGGTAACCAGGAACAA-3'; Reverse: 5'-CGCAGGTGTATCAGGGTCG-3') were designed to span exon-exon junctions, with GAPDH (Forward: 5'-GGAGCGAGATCCCTCCAAAAT-3'; Reverse: 5'-GGCTGTTGTCATACTTCTCATGG-3') serving as the endogenous control. Melting curve analysis confirmed primer specificity. Relative expression levels were calculated using the 2\u003csup\u003e\u0026minus;ΔΔCt\u003c/sup\u003e method. For transduction validation, FREM1 overexpression was confirmed by \u0026ge;\u0026thinsp;1.5-fold increase in mRNA levels compared to empty vector controls (p\u0026thinsp;\u0026lt;\u0026thinsp;0.05).\u003c/p\u003e\u003cp\u003e\u003cb\u003eWestern Blot Analysis for FREM1 Expression Screening and Overexpression Validation​​\u003c/b\u003e\u003c/p\u003e\u003cp\u003e​​To complement our qPCR findings and further verify FREM1 protein expression levels, we performed comprehensive western blot analysis. Total proteins were extracted from breast cancer cell line, normal epithelial cell line, and transduced cells using ice-cold RIPA lysis buffer (Servicebio) supplemented with protease inhibitors. Following centrifugation at 15,000 \u0026times; g for 30 min at 4\u0026deg;C, protein concentrations were determined using the BCA Protein Assay Kit (Servicebio).\u003c/p\u003e\u003cp\u003eFor FREM1 detection, 48 \u0026micro;g of total protein per sample was separated on 10% SDS-PAGE gels (Servicebio) and transferred to 0.45 \u0026micro;m PVDF membranes (Millipore Sigma). After blocking with 5% non-fat milk in TBST for 2 h at room temperature, membranes were incubated overnight at 4\u0026deg;C with primary antibodies: anti-FREM1 (1:1000, Proteintech, 13086-1-AP) and anti-GAPDH (1:5000, Proteintech, 60004-1-Ig) as loading control. Following three washes with TBST, membranes were incubated with HRP-conjugated secondary antibodies (1:5000, ZSGB-BIO, ZB-2301) for 1 h at room temperature.\u003c/p\u003e\u003cp\u003eProtein signals were detected using the SuperPico ECL Chemiluminescent Substrate (Servicebio) and quantified using ImageJ software (NIH). Successful FREM1 overexpression was confirmed by \u0026ge;\u0026thinsp;1.5-fold increase in protein expression compared to vector controls (p\u0026thinsp;\u0026lt;\u0026thinsp;0.05), with all experiments performed in three independent biological replicates.\u003c/p\u003e\u003cp\u003e\u003cb\u003eEdU and Ki-67 Immunofluorescence Analysis of FREM1-Mediated Proliferation Inhibition in Breast Cancer Cells\u003c/b\u003e\u003c/p\u003e\u003cp\u003eTo assess the effect of FREM1 overexpression on breast cancer cell proliferation, we performed EdU incorporation and Ki-67 immunofluorescence assays. For EdU labeling, cells were seeded on coverslips in 48-well plates at a density of 5\u0026times;10^4 cells/well and cultured for 8 hours. The cells were then incubated with 10 \u0026micro;M EdU (MCE) for 2 hours at 37\u0026deg;C, fixed with 4% paraformaldehyde for 30 minutes, and processed using the Click-iT EdU Alexa Fluor 594 Imaging Kit according to the manufacturer's protocol.\u003c/p\u003e\u003cp\u003eFor Ki-67 detection, cells grown on coverslips were fixed with 4% paraformaldehyde, permeabilized with 0.3% Triton X-100, and blocked with 5% Donkey Serum. The cells were then incubated overnight at 4\u0026deg;C with anti-Ki-67 antibody (Abclnal, A20018; 1:400 dilution), followed by incubation with Alexa Fluor 594-conjugated secondary antibody for 2 hours at room temperature. Nuclei were counterstained with DAPI (1 \u0026micro;g/mL) for 5 minutes.\u003c/p\u003e\u003cp\u003eFluorescence images were acquired using a confocal microscope (Leica TCS SP8), with at least five random fields captured per sample. The proliferation index was calculated as the percentage of EdU-positive or Ki-67-positive cells relative to the total number of DAPI-stained nuclei. Three independent biological replicates were performed for each experiment.\u003c/p\u003e\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\u003ch2\u003eStatistical Analysis\u003c/h2\u003e\u003cp\u003eAll data analyses were conducted with R statistical software (v4.4.0). Continuous variables were analyzed using appropriate tests based on data distribution characteristics: parametric comparisons employed Student's t-test (two groups) or ANOVA (multiple groups), while nonparametric analyses utilized the Wilcoxon rank-sum test (two groups) or Kruskal-Wallis test (multiple groups). Bivariate associations were examined through correlation analysis, with Pearson's coefficient applied to normally distributed variables and Spearman's rank correlation used for non-normal distributions. Survival analysis was evaluated via Kaplan-Meier methodology, with between-group differences tested using the log-rank statistic. All tests were two-sided, and significance was set at p\u0026thinsp;\u0026lt;\u0026thinsp;0.05 unless otherwise specified.\u003c/p\u003e\u003c/div\u003e"},{"header":"Results","content":"\u003cp\u003e\u003cb\u003e​​Identification and Characterization of Basement Membrane-Based Molecular Subtypes in Breast Cancer\u003c/b\u003e\u003c/p\u003e\u003cp\u003eFigure \u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e outlines our analytical workflow (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). Consensus clustering of TCGA-BRCA samples using 222 basement membrane-related genes robustly segregated patients into two subtypes: a basement membrane-high group (C1) showing elevated expression of core components, and a basement membrane-low group (C2) with significantly reduced expression (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA). Kaplan-Meier analysis demonstrated significantly poorer overall survival in the C2 versus C1 group (p\u0026thinsp;=\u0026thinsp;0.029, Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eB). The hierarchical clustering heatmap (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eC) displays the expression profiles of the top 50 most significantly differentially expressed genes (|log2FC| \u0026gt;1, FDR\u0026thinsp;\u0026lt;\u0026thinsp;0.05) between C1 and C2 subtypes, revealing distinct molecular signatures characteristic of each subgroup.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003cb\u003eFunctional Annotation of BM-Associated DEGs Uncovers Their Involvement in Breast Cancer Microenvironment\u003c/b\u003e\u003c/p\u003e\u003cp\u003eTo investigate the biological significance of differentially expressed genes (DEGs) across basement membrane-associated subgroups, we performed integrated functional enrichment analyses including Gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG), and Gene Set Enrichment Analysis (GSEA). The GO analysis revealed significant enrichment in Biological Processes, with key pathways including epidermis and skin development, extracellular matrix and structure organization, chemokine-mediated signaling, antimicrobial humoral response, keratinization, and intermediate filament organization (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eD). In the Cellular Component category, notable enrichments were observed in the cornified envelope, collagen-containing extracellular matrix, keratin and intermediate filaments, neuronal cell body, endoplasmic reticulum lumen, and neuron projection terminus (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eD). Molecular Function analysis identified associations with cytokine and chemokine activity, structural constituents of skin epidermis, chemokine receptor binding, glycosaminoglycan binding, G protein-coupled receptor binding, and endopeptidase inhibitor activity (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eD).\u003c/p\u003e\u003cp\u003eKEGG pathway analysis revealed the top 10 significant pathways, including Cytokine-cytokine receptor interaction, Viral protein interaction with cytokine and cytokine receptor, Amoebiasis, IL-17 signaling pathway, Protein digestion and absorption, Rheumatoid arthritis, TNF signaling pathway, Chemokine signaling pathway, Neuroactive ligand-receptor interaction, and Hematopoietic cell lineage, as shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eE.\u003c/p\u003e\u003cp\u003eTo identify pathway-level differences between BM expression subgroups, we conducted gene set enrichment analysis (GSEA) utilizing the Molecular Signatures Database (MSigDB) collections, including the GO gene sets (c5.go.symbols.gmt) and canonical KEGG pathways (c2.cp.kegg.symbols.gmt). In the C2 group, significant enrichment was observed in several immune-related KEGG pathways, including Cardiac Muscle Contraction, Drug Metabolism Other Enzymes, Glycosylphosphatidylinositol GPI Anchor Biosynthesis, Oxidative Phosphorylation, and Parkinson\u0026rsquo;s Disease. Additionally, the C2 group showed enrichment in GO terms such as Mitochondrial Respiratory Chain Complex Assembly, DNA Packaging Complex, Inner Mitochondrial Membrane Protein Complex, Respirasome, and Oxido-reduction Driven Active Transmembrane Transporter Activity (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eF, G). In contrast, the C1 group was significantly enriched in KEGG pathways such as Cell Adhesion Molecules, Chemokine Signaling Pathway, Cytokine-Cytokine Receptor Interaction, Natural Killer Cell-Mediated Cytotoxicity, and Primary Immunodeficiency. The C1 group also showed enrichment in GO terms such as Granulocyte Chemotaxis, Leukocyte Chemotaxis, Keratin Filament, Chemokine Receptor Binding, and Cytokine Activity. These analyses suggest that basement membrane-related genes may play a role in modulating breast cancer microenvironment and metabolism. These analyses suggest that basement membrane-related genes may influence the breast cancer microenvironment and metabolism (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eF, G).\u003c/p\u003e\u003cp\u003e\u003cb\u003eComprehensive immune profiling revealed significant differences in the tumor microenvironment between BM-defined subgroups\u003c/b\u003e\u003c/p\u003e\u003cp\u003eThe C1 group demonstrated elevated ESTIMATE, immune, and stromal scores alongside reduced tumor purity compared to the C2 group (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eA). Further analysis using CIBERSORT revealed substantial differences in the infiltration levels of various immune cells between the two basement membrane groups. These included na\u0026iuml;ve B cells, CD8\u0026thinsp;+\u0026thinsp;T cells, memory resting CD4\u0026thinsp;+\u0026thinsp;T cells, memory activated CD4\u0026thinsp;+\u0026thinsp;T cells, follicular helper T cells, M1 macrophages, M2 macrophages, resting and activated dendritic cells, resting mast cells, and neutrophils (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eB). These results highlight the crucial role of the basement membrane in influencing the breast cancer immune microenvironment. Furthermore, we observed significant differential expression of HLA genes and immune checkpoint molecules between the groups (Figs.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eC, \u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eD), highlighting the crucial role of basement membrane components in shaping the breast cancer immune landscape. These findings collectively demonstrate that BM-related gene expression patterns significantly influence both the cellular composition and immunoregulatory characteristics of the tumor microenvironment.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003cb\u003eConstruction and Multi-Dimensional Validation of a Prognostic BMS Model in Breast Cancer​\u003c/b\u003e\u003c/p\u003e\u003cp\u003eThrough univariate Cox regression analysis, we initially identified 18 basement membrane-related genes significantly associated with patient prognosis (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA). Subsequent LASSO regression analysis refined this set to 14 key prognostic genes, which were used to construct a predictive model named the Basement Membrane Signature (BMS) (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eB). Each breast cancer sample was assigned a BMS risk score, with patients dichotomized into high- and low-risk groups based on the median score. Survival analysis confirmed significantly worse outcomes in the high-risk group, with robust reproducibility across both the TCGA (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eC) training cohort (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001) and independent GEO (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eD) validation dataset (p\u0026thinsp;=\u0026thinsp;0.001), demonstrating the prognostic validity of this signature. Risk score distributions exhibited consistent patterns between TCGA (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eE) and GEO (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eF) cohorts, demonstrating the reproducibility of our risk stratification model.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eComprehensive Cox regression analyses revealed distinct prognostic patterns across datasets. In the TCGA cohort, univariate and multivariate analyses consistently identified age, TNM stage, and BMS risk score as significant independent prognostic factors (all p\u0026thinsp;\u0026lt;\u0026thinsp;0.001, Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA). However, validation in the GEO dataset demonstrated that only BMS risk score maintained its prognostic significance (p\u0026thinsp;=\u0026thinsp;0.001, Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eB), underscoring its robust predictive value across diverse patient populations.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eDiscriminative ability was assessed through receiver operating characteristic (ROC) analysis. Comparative evaluation of age, TNM stage, and the BMS risk score yielded AUC values of 0.623, 0.651, and 0.705, respectively (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eC). Time-dependent ROC analysis further demonstrated the stable predictive capacity of the BMS risk score across clinical endpoints, with AUCs of 0.659 (1-year), 0.707 (3-year), and 0.705 (5-year) survival (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eD).\u003c/p\u003e\u003cp\u003eAnalysis of TCGA breast cancer samples demonstrated that patients in the high-risk group had significantly poorer outcomes in progression-free survival (PFS, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001) (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eE), disease-free survival (DFS, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001) (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eF), and disease-specific survival (DSS, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001) compared to the low-risk group (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eG). To further validate these findings, we constructed prognostic nomograms and calibration curves for the predictive models incorporating age, TNM stage and the Basement Membrane Signature (BMS) risk score. These nomograms and calibration curves were developed to assess the predictive accuracy for overall survival (OS) (Supplementary Fig.\u0026nbsp;1A), progression free survival (PFS) (Supplementary Fig.\u0026nbsp;1B), disease free survival (DFS) (Supplementary Fig.\u0026nbsp;1C) and disease specific survival (DSS) (Supplementary Fig.\u0026nbsp;1D) outcomes. The close alignment between predicted and observed outcomes in both the nomograms and validation curves strongly supported the reliability of these prognostic models.\u003c/p\u003e\u003cp\u003e\u003cb\u003eIdentification and Characterization of FREM1 as a Prognostic Biomarker in Cancer\u003c/b\u003e\u003c/p\u003e\u003cp\u003eTo refine our gene signature, we employed a machine learning approach using support vector machines (SVM), which identified a set of candidate genes (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eH). Intersection of these SVM-derived candidates with our earlier LASSO-selected genes revealed six core prognostic genes: SEMA3B, ITGAX, FREM1, ADAM9, ADAMTS8, and UNC5A (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eI). Subsequent validation in the TCGA cohort demonstrated that while all four genes (SEMA3B, ITGAX, FREM1, and ADAMTS8) showed significant associations with overall survival (OS) in Kaplan-Meier analysis (log-rank p\u0026thinsp;\u0026lt;\u0026thinsp;0.05), FREM1 exhibited the most pronounced survival benefit (lowest p-value) (Supplementary Fig.\u0026nbsp;2A). Differential expression analysis between tumor and adjacent normal tissues revealed that FREM1 showed the most significant downregulation among all candidate genes (p\u0026thinsp;\u0026lt;\u0026thinsp;0.05), with a consistent protective pattern (Supplementary Fig.\u0026nbsp;2B). Given its strongest prognostic performance, most remarkable tumor-specific expression pattern, and putative tumor-suppressive role, we prioritized FREM1 as our primary target for mechanistic investigation.\u003c/p\u003e\u003cp\u003eFinally, using single-cell RNA sequencing data from two independent datasets-GSE148673 (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eJ) and GSE176078 (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eK)-we performed a comprehensive analysis of FREM1 expression across diverse cell populations within the tumor microenvironment. The results demonstrated that FREM1 was predominantly expressed in cancer-associated fibroblasts (CAFs), with minimal expression in other cell types. This fibroblast-specific expression pattern suggests FREM1 may play a crucial role in tumor-stroma interactions and extracellular matrix remodeling. These findings highlight FREM1's unique position in shaping the tumor microenvironment and support its potential as both a stromal biomarker and therapeutic target in breast cancer progression.\u003c/p\u003e\u003cp\u003eTo functionally characterize FREM1, we performed in vitro studies using triple-negative MDA-MB-231 breast cancer cells compared to normal mammary epithelial cells (MCF-10A). Both qPCR (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eA) and Western blot (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eB, C) analyses confirmed significantly lower FREM1 expression in cancer cells, with \u0026gt;\u0026thinsp;1.44-fold reduction at mRNA level (p\u0026thinsp;\u0026lt;\u0026thinsp;0.01) and \u0026gt;\u0026thinsp;1.77-fold decrease at protein level (p\u0026thinsp;\u0026lt;\u0026thinsp;0.01) relative to normal controls. To investigate its functional role, we transfected MDA-MB-231 cells with a FREM1 overexpression plasmid, achieving 1.48-fold higher mRNA levels (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001) (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eD) and 2.42-fold increased protein expression (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eE, F) (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001) compared to vector controls. Functional assays demonstrated that FREM1 overexpression significantly inhibited cell proliferation, with Ki-67 immunofluorescence (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eG, H) revealing 33.3% fewer proliferating cells (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001) and EdU incorporation assays (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eI, J) showing 34.8% reduction in S-phase cells (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001). These results collectively demonstrate that FREM1 acts as a tumor suppressor in breast cancer by inhibiting cellular proliferation.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eOur study establishes a novel prognostic signature based on basement membrane (BM)-related genes, which effectively stratifies breast cancer patients into distinct risk groups with significant differences in clinical outcomes. This BM-related gene signature (BMS) not only demonstrates robust predictive accuracy for survival but also reveals the critical role of BM components in shaping tumor progression and the immune microenvironment. By integrating multi-omics data and functional validation, we highlight the potential of BMS as a valuable tool for prognosis assessment and personalized treatment strategies in breast cancer.\u003c/p\u003e\u003cp\u003eBreast cancer represents a highly heterogeneous disease with diverse molecular subtypes and clinical outcomes, underscoring the need for precise prognostic biomarkers (21). Emerging evidence has highlighted the basement membrane (BM) as a dynamic structural and functional component that actively participates in tumor progression and metastasis (22, 23). Unlike traditional views of the BM as a passive barrier, recent studies demonstrate its crucial role in regulating cancer cell invasion, immune cell infiltration, and therapeutic resistance (19, 24). The BM's unique composition of collagen IV, laminins, and other glycoproteins forms a specialized extracellular matrix niche that influences tumor-stroma crosstalk and metastatic dissemination (25, 26). Our identification of BM-related gene signatures builds upon these findings by providing a systematic framework to quantify BM remodeling patterns and their clinical implications. This approach offers new opportunities to understand breast cancer heterogeneity beyond conventional molecular subtyping, potentially guiding more personalized treatment strategies.\u003c/p\u003e\u003cp\u003eOur study systematically evaluated the prognostic value of basement membrane (BM)-related genes in breast cancer through a multi-stage analytical approach. Initially, consensus clustering analysis of BM-related gene expression patterns revealed two distinct molecular subtypes with significant clinical differences. The observed disparity in clinical outcomes between risk groups may be mechanistically linked to distinct immunological characteristics within the tumor microenvironment. These findings suggested that BM composition actively shapes both tumor behavior and immune responses (27, 28).\u003c/p\u003e\u003cp\u003eTo refine these observations into a clinically applicable tool, we employed machine learning techniques including LASSO regression and Cox proportional hazards modeling. This process identified a robust 14-gene signature that effectively stratified patients by risk across multiple independent cohorts. The signature showed significant prognostic discrimination, with high-risk patients exhibiting substantially worse clinical outcomes than low-risk patients. Importantly, the signature maintained its predictive power even after adjusting for conventional clinical parameters such as age and tumor stage in multivariate analyses.\u003c/p\u003e\u003cp\u003eThe clinical utility of this signature was further enhanced through the development of comprehensive nomograms that integrate BM-related gene expression with standard prognostic factors. These tools showed excellent predictive accuracy for both short-term and long-term outcomes, providing clinicians with a quantitative framework for risk assessment (29, 30). Mechanistically, the signature captures key biological processes in cancer progression, including extracellular matrix remodeling and Cytokine-cytokine receptor interaction. The strong association between signature scores and specific immune cell populations suggests potential utility in predicting response to immunotherapy.\u003c/p\u003e\u003cp\u003eThese findings significantly advance our understanding of BM biology in breast cancer, transforming it from a passive structural component to an active regulator of tumor progression and microenvironment modulation. The signature provides both prognostic value and biological insights, offering opportunities for more personalized treatment approaches. Future studies should explore its potential in guiding therapeutic decisions, particularly for aggressive subtypes where current prognostic tools remain limited. The integration of BM-related gene expression patterns into clinical decision-making represents a promising avenue for improving breast cancer management and outcomes.\u003c/p\u003e\u003cp\u003eTo identify clinically relevant targets from our basement membrane-related gene signature, we employed a machine learning pipeline combining LASSO regression and support vector machine (SVM) algorithms, which identified six high-confidence candidate genes. Among these, FREM1 emerged as the most consistently downregulated gene in breast tumor tissues compared to matched normal samples. To functionally validate these findings, we established FREM1-overexpressing MDA-MB-231 cells (MDA-FREM1) by transfecting with the FREM1 expression plasmid, with empty vector-transfected cells (MDA-EV) serving as controls. Successful overexpression was confirmed through qPCR analysis of mRNA levels and western blot detection of protein expression. Subsequent functional analyses revealed that FREM1 overexpression significantly suppressed tumor cell proliferation, as demonstrated by reduced Ki-67 expression and decreased EdU incorporation.\u003c/p\u003e\u003cp\u003eThese experimental findings are supported by accumulating clinical evidence establishing FREM1 as a tumor suppressor in breast cancer. Multiple independent cohorts have confirmed the significant downregulation of FREM1 in malignant versus normal breast epithelium. Importantly, reduced FREM1 expression correlates with aggressive clinicopathological features including advanced TNM stage and poorer survival outcomes, while also associating with an immunosuppressive tumor microenvironment marked by diminished cytotoxic lymphocyte infiltration and increased immunosuppressive cell populations (31, 32). Collectively, these data position FREM1 as a key regulator of breast cancer progression, with both prognostic implications and potential therapeutic relevance for overcoming tumor immunosuppression.\u003c/p\u003e\u003cp\u003eOur study establishes the Basement Membrane Signature (BMS) as a novel prognostic tool in breast cancer, demonstrating its ability to stratify patients into distinct risk groups with significant differences in clinical outcomes. By integrating multi-omics data and machine learning approaches, we developed a robust 14-gene signature that effectively predicts survival and reflects key biological processes, including extracellular matrix remodeling and immune microenvironment modulation. However, several limitations should be acknowledged. First, our findings are based on retrospective analyses of publicly available datasets, which may be subject to selection biases and technical variability. Although we validated the BMS in both TCGA and GEO cohorts, further confirmation in prospective, multi-center studies is needed to ensure its clinical applicability. Second, while we identified FREM1 as a potential tumor suppressor through in vitro experiments, the functional roles of other signature genes in breast cancer progression remain to be fully elucidated. Additional mechanistic studies, including in vivo models and patient-derived xenografts, are required to validate their biological significance and therapeutic potential. Moreover, the molecular pathways linking basement membrane remodeling to tumor progression and immune evasion warrant further investigation. Future research should explore whether targeting these pathways could enhance treatment efficacy, particularly in aggressive subtypes.\u003c/p\u003e\u003cp\u003eDespite these limitations, our work provides a foundation for incorporating basement membrane biology into breast cancer prognostication and therapy. The BMS model offers a clinically actionable tool that complements existing prognostic markers, while the identified genes present new opportunities for therapeutic development. Continued validation and functional studies will be essential to translate these findings into improved patient outcomes.\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eOur study establishes the Basement Membrane Signature (BMS) as a robust prognostic tool in breast cancer, effectively stratifying patients into distinct risk groups with significant survival differences. By integrating multi-omics data and machine learning, we developed a 14-gene signature that reflects key biological processes including extracellular matrix remodeling and immune microenvironment modulation. The BMS demonstrated consistent predictive accuracy across multiple cohorts, with high-risk patients showing significantly poorer prognosis compared to low-risk patients. Functional validation identified FREM1 as a potential tumor suppressor, providing mechanistic insights into basement membrane biology. While retrospective in nature, our findings highlight the clinical potential of BMS for risk assessment and treatment guidance. Future studies should focus on prospective validation, mechanistic exploration of basement membrane-immune interactions, and therapeutic targeting of identified pathways to improve patient outcomes. This work advances our understanding of breast cancer heterogeneity and provides a foundation for more personalized management strategies.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgments:\u003c/strong\u003e Not applicable.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor contributions:\u003c/strong\u003e C.L.and C.L. wrote the main manuscript text. C.L. and P.G. \u0026nbsp;performed comprehensive data analysis. C.L. and J.H. created visualizations. All authors reviewed the manuscript.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding:\u003c/strong\u003e None.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConflicts of Interest:\u003c/strong\u003e The authors have no conflicts of interest to declare.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eClinical Trial Number:\u003c/strong\u003e Not applicable. \u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent to Publish Declaration:\u003c/strong\u003e Not applicable. \u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent to Participate Declaration:\u003c/strong\u003e Not applicable. \u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics Declaration:\u003c/strong\u003e Not applicable.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData Availability Statements：\u003c/strong\u003eThe datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eSung H, Ferlay J, Siegel RL, Laversanne M, Soerjomataram I, Jemal A, et al. Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. CA Cancer J Clin. 2021;71(3):209-49.\u003c/li\u003e\n\u003cli\u003eLiu Y, Ouyang W, Huang H, Tan Y, Zhang Z, Yu Y, et al. Identification of a tumor immune-inflammation signature predicting prognosis and immune status in breast cancer. Front Oncol. 2022;12:960579.\u003c/li\u003e\n\u003cli\u003ePasta V, Monti M, Cialini M, Vergine M, Urciuoli P, Iacovelli A, et al. Primitive sarcoma of the breast: new insight on the proper surgical management. J Exp Clin Cancer Res. 2015;34(1):72.\u003c/li\u003e\n\u003cli\u003eLiao Q, Deng D, Xie Q, Gong X, Meng X, Xia Y, et al. Clinical characteristics, pregnancy outcomes and ovarian function of pregnancy-associated breast cancer patients: a retrospective age-matched study. BMC Cancer. 2022;22(1):152.\u003c/li\u003e\n\u003cli\u003eDong J, Yang Y, Han D, Zhao Q, Liu C, Sun H, et al. Hypofractionated Simultaneous Integrated Boost Radiotherapy Versus Conventional Fractionation Radiotherapy of Early Breast Cancer After Breast-Conserving Surgery: Clinical Observation and Analysis. Technol Cancer Res Treat. 2021;20:15330338211064719.\u003c/li\u003e\n\u003cli\u003eToivonen KI, Williamson TM, Carlson LE, Walker LM, Campbell TS. Potentially Modifiable Factors Associated with Adherence to Adjuvant Endocrine Therapy among Breast Cancer Survivors: A Systematic Review. Cancers (Basel). 2020;13(1).\u003c/li\u003e\n\u003cli\u003eKinnel B, Singh SK, Oprea-Ilies G, Singh R. Targeted Therapy and Mechanisms of Drug Resistance in Breast Cancer. Cancers (Basel). 2023;15(4).\u003c/li\u003e\n\u003cli\u003eDai G, Yang Y, Liu S, Liu H. Hypoxic Breast Cancer Cell-Derived Exosomal SNHG1 Promotes Breast Cancer Growth and Angiogenesis via Regulating miR-216b-5p/JAK2 Axis. Cancer Manag Res. 2022;14:123-33.\u003c/li\u003e\n\u003cli\u003eLi W, Liu J, Zhang B, Bie Q, Qian H, Xu W. Transcriptome Analysis Reveals Key Genes and Pathways Associated with Metastasis in Breast Cancer. Onco Targets Ther. 2020;13:323-35.\u003c/li\u003e\n\u003cli\u003eDu X, Zhang J, Liu L, Xu B, Han H, Dai W, et al. A novel anticancer property of Lycium barbarum polysaccharide in triggering ferroptosis of breast cancer cells. J Zhejiang Univ Sci B. 2022;23(4):286-99.\u003c/li\u003e\n\u003cli\u003eLi L, Wei D, Zhang J, Deng R, Tang J, Su D. miR-641 Inhibited Cell Proliferation and Induced Apoptosis by Targeting NUCKS1/PI3K/AKT Signaling Pathway in Breast Cancer. Comput Math Methods Med. 2022;2022:5203839.\u003c/li\u003e\n\u003cli\u003eAgostinetto E, Montemurro F, Puglisi F, Criscitiello C, Bianchini G, Del Mastro L, et al. Immunotherapy for HER2-Positive Breast Cancer: Clinical Evidence and Future Perspectives. Cancers (Basel). 2022;14(9).\u003c/li\u003e\n\u003cli\u003eMills MN, Figura NB, Arrington JA, Yu H-HM, Etame AB, Vogelbaum MA, et al. Management of brain metastases in breast cancer: a review of current practices and emerging treatments. Breast Cancer Res Treat. 2020;180(2):279-300.\u003c/li\u003e\n\u003cli\u003eG\u0026oacute;mez-Aleza C, Nguyen B, Yoldi G, Ciscar M, Barranco A, Hern\u0026aacute;ndez-Jim\u0026eacute;nez E, et al. Inhibition of RANK signaling in breast cancer induces an anti-tumor immune response orchestrated by CD8+ T cells. Nat Commun. 2020;11(1):6335.\u003c/li\u003e\n\u003cli\u003eJayadev R, Morais MRPT, Ellingford JM, Srinivasan S, Naylor RW, Lawless C, et al. A basement membrane discovery pipeline uncovers network complexity, regulators, and human disease associations. Sci Adv. 2022;8(20):eabn2265.\u003c/li\u003e\n\u003cli\u003eJayadev R, Sherwood DR. Basement membranes. Curr Biol. 2017;27(6):R207-R11.\u003c/li\u003e\n\u003cli\u003eZhang Z, Zhu H, Wang X, Lin S, Ruan C, Wang Q. A novel basement membrane-related gene signature for prognosis of lung adenocarcinomas. Comput Biol Med. 2023;154:106597.\u003c/li\u003e\n\u003cli\u003eSun W, Wang J, Wang Z, Xu M, Lin Q, Sun P, et al. Combining WGCNA and machine learning to construct basement membrane-related gene index helps to predict the prognosis and tumor microenvironment of HCC patients and verifies the carcinogenesis of key gene CTSA. Front Immunol. 2023;14:1185916.\u003c/li\u003e\n\u003cli\u003eTian W, Luo Y, Tang Y, Kong Y, Wu L, Zheng S, et al. Novel Implication of the Basement Membrane for Breast Cancer Outcome and Immune Infiltration. Int J Biol Sci. 2023;19(5):1645-63.\u003c/li\u003e\n\u003cli\u003eLu P, Weaver VM, Werb Z. The extracellular matrix: a dynamic niche in cancer progression. J Cell Biol. 2012;196(4):395-406.\u003c/li\u003e\n\u003cli\u003eHossain FM, Danos DM, Fu Q, Wang X, Scribner RA, Chu ST, et al. Association of Obesity and Diabetes With the Incidence of Breast Cancer in Louisiana. Am J Prev Med. 2022;63(1 Suppl 1):S83-S92.\u003c/li\u003e\n\u003cli\u003eTang X, Liu Y, Zhao J, Fu C, Yang W. Subtyping of gastric cancer based on basement membrane genes that stratifies the prognosis, immune infiltration and therapeutic response. Discov Oncol. 2024;15(1):362.\u003c/li\u003e\n\u003cli\u003eChang J, Saraswathibhatla A, Song Z, Varma S, Sanchez C, Alyafei NHK, et al. Cell volume expansion and local contractility drive collective invasion of the basement membrane in breast cancer. Nat Mater. 2024;23(5):711-22.\u003c/li\u003e\n\u003cli\u003eCai J, Zhang X, Xie W, Li Z, Liu W, Liu A. Identification of a basement membrane-related gene signature for predicting prognosis and estimating the tumor immune microenvironment in breast cancer. Front Endocrinol (Lausanne). 2022;13:1065530.\u003c/li\u003e\n\u003cli\u003eStricker AM, Hutson MS, Page-McCaw A. Piezo-dependent surveillance of matrix stiffness generates transient cells that repair the basement membrane. Dev Cell. 2025.\u003c/li\u003e\n\u003cli\u003eWang J, Lin S, Wei Y, Ye Z. Recombinant human collagen XVII protects skin basement membrane integrity by inhibiting the MAPK and Wnt signaling pathways. Mol Med Rep. 2025;31(4).\u003c/li\u003e\n\u003cli\u003eWalter C, Davis JT, Mathur J, Pathak A. Physical defects in basement membrane-mimicking collagen-IV matrices trigger cellular EMT and invasion. Integr Biol (Camb). 2018;10(6):342-55.\u003c/li\u003e\n\u003cli\u003eWu Y, Liu X, Zhu Y, Qiao Y, Gao Y, Chen J, et al. Type IV collagen \u0026alpha;5 chain promotes luminal breast cancer progression through c-Myc-driven glycolysis. J Mol Cell Biol. 2023;14(10).\u003c/li\u003e\n\u003cli\u003eXie L, Zhang Y, Niu X, Jiang X, Kang Y, Diao X, et al. A nomogram for predicting cancer-specific survival in patients with locally advanced unresectable esophageal cancer: development and validation study. Front Immunol. 2025;16:1524439.\u003c/li\u003e\n\u003cli\u003eHu J, Shi Y, Jin L, Yi S, Chen J, Wan D, et al. Predicting survival rates: the power of prognostic nomograms in distal cholangiocarcinoma. Front Oncol. 2025;15:1478836.\u003c/li\u003e\n\u003cli\u003eXu X-Y, Guo W-J, Pan S-H, Zhang Y, Gao F-L, Wang J-T, et al. TILRR (FREM1 isoform 2) is a prognostic biomarker correlated with immune infiltration in breast cancer. Aging (Albany NY). 2020;12(19):19335-51.\u003c/li\u003e\n\u003cli\u003eLi H-N, Li X-R, Lv Z-T, Cai M-M, Wang G, Yang Z-F. Elevated expression of FREM1 in breast cancer indicates favorable prognosis and high-level immune infiltration status. Cancer Med. 2020;9(24):9554-70.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"discover-oncology","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"dion","sideBox":"Learn more about [Discover Oncology](https://www.springer.com/12672)","snPcode":"","submissionUrl":"","title":"Discover Oncology","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Discover Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-7206387/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7206387/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cb\u003eBackground​​\u003c/b\u003e: Breast cancer remains the leading cause of cancer-related mortality in women worldwide, with metastatic disease posing significant therapeutic challenges. While immunotherapy has shown promise, tumor immune evasion limits its efficacy. The basement membrane (BM), a specialized extracellular matrix structure, plays a crucial yet understudied role in breast cancer progression and immune modulation. This study aims to investigate the prognostic value and therapeutic potential of BM-related genes in breast cancer.\u003c/p\u003e\u003cp\u003e​​\u003cb\u003eMethods​​\u003c/b\u003e: We integrated transcriptomic data from TCGA and GEO databases to construct a BM-related gene signature. Unsupervised clustering stratified patients into molecular subtypes, while differential expression analysis identified key BM-associated genes. Functional enrichment analyses (GO, KEGG, GSEA) elucidated biological pathways, and immune microenvironment characterization was performed using ESTIMATE and CIBERSORT. Machine learning approaches pinpointed critical BM-related genes, which were subsequently validated through in vitro experiments.\u003c/p\u003e\u003cp\u003e​​\u003cb\u003eResults​​\u003c/b\u003e: Breast cancer patients were classified into high- and low-BM groups, with the low-BM cohort exhibiting worse prognosis. Pathway analysis revealed significant enrichment in immune regulation, ECM remodeling, and cytokine signaling. FREM1 emerged as a top protective gene through machine learning. Experimental validation using low-FREM1-expressing breast cancer cell lines demonstrated that FREM1 overexpression (confirmed by qPCR and Western blot) significantly suppressed tumor cell proliferation, as evidenced by decreased Ki-67 expression and reduced EdU incorporation.\u003c/p\u003e\u003cp\u003e​​\u003cb\u003eConclusion​​\u003c/b\u003e: Our study establishes BM-related genes as novel prognostic biomarkers and therapeutic targets in breast cancer. FREM1 in particular functions as a tumor suppressor by inhibiting cancer cell proliferation, highlighting its potential for therapeutic exploitation. These findings provide critical insights into BM-mediated tumor progression and suggest new avenues for targeted breast cancer therapy.\u003c/p\u003e","manuscriptTitle":"FREM1 Serves as a Novel Therapeutic Target in Breast Cancer through Basement Membrane-Based Prognostic Modeling with Integrated Bioinformatics and Experimental Validation","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-08-25 15:00:28","doi":"10.21203/rs.3.rs-7206387/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2025-08-28T09:03:06+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-08-26T00:21:12+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-08-23T14:50:33+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"258745452324862279222735833742898641550","date":"2025-08-18T11:53:45+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"193782355788276727163452440975624621513","date":"2025-08-18T10:58:04+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-08-13T10:24:58+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-08-08T17:15:02+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2025-08-08T16:42:56+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-08-02T15:32:26+00:00","index":"","fulltext":""},{"type":"submitted","content":"Discover Oncology","date":"2025-08-02T15:11:45+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"discover-oncology","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"dion","sideBox":"Learn more about [Discover Oncology](https://www.springer.com/12672)","snPcode":"","submissionUrl":"","title":"Discover Oncology","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Discover Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"ce269a72-ffa4-49be-a8a0-5edcd8744993","owner":[],"postedDate":"August 25th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2025-11-17T17:23:18+00:00","versionOfRecord":[],"versionCreatedAt":"2025-08-25 15:00:28","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7206387","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7206387","identity":"rs-7206387","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}