Propylparaben Promotes Estrogen Receptor-Positive Breast Cancer Risk through PIK3R1: A Comprehensive Integrative Study

Propylparaben Promotes Estrogen Receptor-Positive Breast Cancer Risk through PIK3R1: A Comprehensive Integrative Study | 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 Propylparaben Promotes Estrogen Receptor-Positive Breast Cancer Risk through PIK3R1: A Comprehensive Integrative Study Kaiwen Wang, Guang Yao, Xiaobin Zhang, Ruiting Ma, Xinyi Li, Weize Kong, and 6 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7400738/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 9 You are reading this latest preprint version Abstract Background Propylparaben (PP), a preservative in cosmetics, food, and pharmaceuticals, is a potential endocrine disruptor. However, the molecular mechanisms linking PP exposure to estrogen receptor-positive (ER⁺) breast cancer (BC) remain unclear. Methods This study integrated network toxicology with The Cancer Genome Atlas (TCGA) data to identify carcinogenic targets of PP. Protein-protein interaction (PPI) networks were constructed, and machine learning algorithms identified core genes. Diagnostic performance was evaluated by Receiver Operating Characteristic (ROC) analysis. Mendelian Randomization (MR) assessed causal links between core genes and BC risk. Molecular docking verified the binding affinity between PP and PIK3R1. Immune infiltration was analyzed using single-sample Gene Set Enrichment Analysis. Results We identified 50 candidate genes related to BC. PPI analysis revealed 20 key genes, and machine learning narrowed it to 6 core genes. ROC analysis showed excellent diagnostic performance (AUC > 0.90). MR analysis showed that decreased PIK3R1 expression significantly increased BC risk (OR = 0.869, p = 0.017). Molecular docking confirmed strong binding between PP and PIK3R1. Immune analysis suggested a correlation between PIK3R1 expression and immune cell abundance. Conclusion PP may promote ER⁺ BC progression by binding to and suppressing PIK3R1, suggesting a potential carcinogenic effect, warranting further investigation in cohort studies. Propylparaben Breast cancer PIK3R1 Machine learning MR Molecular docking Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 1. Introduction Breast cancer (BC) remains one of the most common and lethal malignancies among women worldwide, posing a substantial public health burden. According to the World Health Organization, approximately 2.3 million women worldwide were diagnosed with breast cancer in 2022, with an estimated 670,000 deaths attributed to the disease( 1 ). The incidence of breast cancer exhibits significant gender and age differences, occurring predominantly in women over the age of 50. Notably, about 70% of BC cases are classified as estrogen receptor-positive (ER + ) subtypes( 2 ). As one of the three most prevalent cancers globally, BC prevention and control have garnered increasing attention( 3 ). In addition to lifestyle-related risk factors such as alcohol consumption, obesity, and physical inactivity( 4 ), environmental exposures-particularly to endocrine-disrupting chemicals (EDCs) have emerged as critical external contributors to BC risk( 5 ). Previous studies have shown that exogenous EDCs, such as parabens and bisphenols, can disrupt estrogen signaling pathways, thereby promoting the initiation and progression of hormone receptor-related tumors( 6 , 7 ). Elucidating the toxicological mechanisms of EDCs with potential breast carcinogenic activity has become a key focus in environmental toxicology research. Propylparaben (PP) is one of the most commonly used paraben compounds, widely incorporated into everyday consumer products, particularly as an effective preservative in cosmetics( 8 ). Due to its potent antimicrobial properties, PP is also extensively used in food preservation( 9 ) and as a preservative in pharmaceutical injectable formulations( 10 ). In recent years, its applications have expanded into industrial manufacturing and environmental remediation( 11 , 12 ). PP can enter the human body through dermal absorption or dietary intake, where it undergoes hydrolysis by esterases into p-hydroxybenzoic acid and propanol. These metabolites are subsequently conjugated with glucuronic acid, sulfate, or glycine in hepatic microsomes to form water-soluble derivatives, which are then excreted via urine( 13 ). Despite its widespread usage, the metabolism and toxicological safety of PP remain controversial. Multiple studies have suggested that PP may be a potential carcinogen, showing associations with cancers such as lung and prostate cancer( 14 – 16 ). However, evidence linking PP to BC remains limited. As an emerging EDC, PP exhibits a molecular structure like that of estrogen( 17 ). According to a scientific opinion by the Scientific Committee on Consumer Safety of the European Union, both in vivo and in vitro studies have confirmed the estrogen-like activity of PP ( 18 ). In breast tissue, imbalances in estrogen and progesterone levels can lead to DNA damage accumulation, gene mutations, and the formation of malignant cells. Estrogen also promotes the proliferation of malignant cells and the formation of a tumor-supportive microenvironment, thereby facilitating cancer progression ( 19 ). Although previous studies have suggested that PP, as an environmental estrogenic compound, may be involved in the initiation and progression of BC( 20 ), its precise molecular mechanisms remain unclear due to limitations of conventional research methods( 21 ). As a novel environmental contaminant, the association between PP and breast cancer, as well as the underlying molecular pathways, warrants further investigation. Traditional toxicological approaches often emphasize the direct interactions between a compound and its targets, overlooking the complex biological network through which diseases are induced. In this study, we employed a network toxicology framework, integrating bioinformatics, machine learning algorithms, Mendelian randomization (MR), and molecular docking, to systematically elucidate the molecular mechanisms by which PP may contribute to BC development. The findings are expected to provide theoretical insights and scientific reference for future research on environmentally induced tumorigenesis. 2. Methods 2.1 Toxicity Analysis and Target Screening of Propylparaben(PP) In this study, the standardized molecular structure and corresponding SMILES (Simplified Molecular Input Line Entry System) notation of PP were obtained by querying the PubChem database ( https://pubchem.ncbi.nlm.nih.gov/)(22) using the keyword "propyl p-hydroxybenzoate." Subsequently, PP was subjected to toxicity prediction using the ProTox 3.0 platform ( https://tox.charite.de/)(23) . By integrating molecular structural features, toxicological target identification, and pharmacokinetic property evaluation, we identified potential toxicological pathways and target organs associated with PP. These results provide a theoretical basis for subsequent mechanistic investigations. In this study, we employed the SuperPred database to predict the potential targets of PP. SuperPred is based on machine learning models and can predict the ATC classification codes and potential targets of compounds by integrating known drug information( 24 ). To complement and refine the target data, we additionally queried the STITCH ( http://stitch.embl.de/)(25) and SwissTargetPrediction ( http://www.swisstargetprediction.ch/)(26) databases, and standardized the names of candidate targets to retrieve those not covered by SuperPred. Finally, we integrated the results to construct an initial gene library of PP targets, providing essential data support for the systematic investigation of PP’s potential carcinogenic mechanisms. 2.2 Identification of Differentially Expressed Genes in Breast Cancer To identify differentially expressed genes (DEGs), we extracted the gene expression matrix related to BC from The Cancer Genome Atlas (TCGA) database( 27 ). During data preprocessing, we excluded recurrent cases and retained only primary solid tumor samples (n = 1086) and their corresponding normal breast tissue samples (n = 99) to ensure the representativeness and accuracy of the analysis results. After identifying DEGs and the toxicological targets of PP, we conducted Venn diagram analysis to identify overlapping genes. These intersecting genes were regarded as potential key targets involved in BC development associated with PP exposure, providing a foundation for subsequent functional enrichment analysis and network construction. 2.3 Functional and Pathway Enrichment Analysis of Target Genes To further explore the biological functions of potential gene targets involved in BC related to PP exposure, Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses were conducted using R software (version 4.4.0) GO analysis was performed across three categories: Biological Process (BP), Cellular Component (CC), and Molecular Function (MF)-to identify biological processes potentially associated with PP. KEGG enrichment analysis was used to determine pathways related to BC development associated with PP exposure, thereby elucidating its potential mechanism of action. The potential target genes associated with PP were then input into the STRING database( 28 ), with the species set to "Homo sapiens" and a minimum interaction score of > 0.15 to ensure a comprehensive mapping of gene interactions. The resulting data were subsequently imported into Cytoscape software (version 3.10.3) for further analysis. Cytoscape is a widely used bioinformatics visualization tool capable of calculating topological attributes of network nodes and constructing protein-protein interaction (PPI) networks( 29 ).We ranked nodes based on their degree values and enhanced the visualization through color gradients. The top 20 hub genes, ranked by degree, were identified as key targets for further investigation due to their high connectivity and potential regulatory significance within the network. 2.4 Machine Learning-Based Selection of Core Genes To enhance the robustness and accuracy of target gene identification, we applied three machine learning algorithms-Random Forest, Least Absolute Shrinkage and Selection Operator (LASSO) regression, and Support Vector Machine-Recursive Feature Elimination (SVM-RFE)-based on the expression matrix of hub genes derived from the TCGA dataset. Feature importance was assessed using the Random Forest algorithm with ten-fold cross-validation( 30 ). This approach was employed to identify genes with significant contributions to the classification model, even in the presence of nonlinear relationships. LASSO regression with a binomial distribution was applied( 31 ), and the optimal regularization parameter (λ) was determined through ten-fold cross-validation. Genes with non-zero coefficients were retained to reduce overfitting and enhance interpretability. SVM-RFE with a radial basis function kernel was used to iteratively refine feature selection( 32 ). Cross-validation error guided the elimination process, and informative genes were identified based on their contribution to predictive accuracy. To identify core targets, we conducted intersection analysis of the gene sets identified by each algorithm. The shared genes, identified via a Venn diagram, were regarded as core candidate targets involved in BC related to PP exposure, serving as a foundation for subsequent causal inference and mechanistic studies. Finally, to evaluate the diagnostic performance of the identified core genes, Receiver Operating Characteristic (ROC) curve analysis was performed. The area under the curve (AUC) was calculated for each gene to assess its ability to discriminate between exposed and control samples. All selected genes exhibited high diagnostic accuracy, with AUC values exceeding 0.90, underscoring their potential as robust biomarkers for BC linked to PP. 2.5 Mendelian Randomization Analysis To further investigate the causal relationship between core target genes and BC, this study employed Mendelian Randomization (MR) to evaluate the potential causal effects of candidate gene expression levels on breast cancer development. Based on previously identified core genes, we extracted expression quantitative trait locis (eQTLs) significantly associated with gene expression from the IEU Open GWAS database ( https://gwas.mrcieu.ac.uk/ ) to serve as instrumental variables (IVs). To ensure the strength and independence of IVs, the following criteria were applied: ( 1 ) statistically significant association with gene expression (P < 1×10⁻⁵); ( 2 ) linkage disequilibrium threshold set at r² 10. Given that PP is widely recognized as an endocrine-disrupting chemical, primarily acting via ER, we selected ER + breast cancer as the outcome phenotype. The outcome data were obtained from the IEU Open GWAS database (dataset ieu-a-1132 , https://opengwas.io/datasets/ieu-a-1132 ), which is based on a study by Michailidou K et al. (2017, Nature , n = 83,691 European female participants, BCAC) ( 33 ) and comprises 38,197 breast cancer cases and 45,494 healthy controls. MR analysis was primarily conducted using the inverse-variance weighted (IVW) method, with additional validation using MR-Egger regression, weighted median, simple mode, and weighted mode approaches to enhance the robustness of causal inference. To further evaluate the reliability and robustness of the findings, sensitivity analyses were performed, including: ( 1 ) Cochran’s Q test to assess heterogeneity among IVs; ( 2 ) MR-Egger intercept test to detect potential horizontal pleiotropy; and ( 3 ) leave-one-out analysis to exclude the influence of any single SNP dominating the causal effect. 2.6 Differential Expression and Prognostic Analysis To evaluate the differential expression of the key target gene in BC tissues, we extracted the expression data of PIK3R1 from the transcriptomic dataset of breast cancer patients in TCGA database. Expression levels were compared between groups based on corresponding clinicopathological information. During data preprocessing, we selected primary BC samples with complete clinical annotation (Tumor = 1086) and normal breast tissue samples (Normal = 99). To further explore the potential prognostic value of this gene in BC, we conducted survival analysis using the KM Plotter online platform ( https://kmplot.com/analysis/ ), which integrates clinical and gene expression data from multiple public BC cohorts( 34 ). A total of breast cancer samples were included in the analysis (n = 4929), with low expression group (n = 2467) and high expression group (n = 2462). The association between gene expression and patient survival outcomes was quantified using odds ratios (ORs) and 95% confidence intervals (CIs). Finally, survival results were visualized to assess the potential of this gene as a prognostic biomarker in BC. 2.7 Molecular Docking Analysis Molecular docking predicts the binding mode and affinity between small molecules and target proteins by identifying the lowest-energy conformation. To explore the potential carcinogenic mechanism of PP, docking was performed with key BC target proteins. First, the 3D structure of PP was downloaded from the PubChem Compound Database ( https://pubchem.ncbi.nlm.nih.gov/ ), while the crystal structures of key target proteins were obtained from UniProt ( https://www.uniprot.org/)(35) and the Protein Data Bank (PDB, https://www.rcsb.org/)(36) . Protein structures were preprocessed using PyMOL software( 37 ) by removing water molecules and original ligands, and saved in PDB format. Next, ligand preparation was conducted using AutoDock Tools 1.5.7( 36 ), followed by saving the ligand in PDBQT format. The protein structures were modified by adding hydrogen atoms and defined as receptors, then also saved in PDBQT format. The processed ligand and receptor were imported into AutoDock, and a docking grid box covering the entire protein surface was constructed. Docking was performed using AutoDock Vina( 38 ) with the number of docking poses set to 50. The docking results were ranked based on binding energy, and the protein-ligand complex with the lowest energy was selected for further analysis. The docked complex was uploaded to the PLIP Web Tool ( https://plip-tool.biotec.tu-dresden.de/plip-web/plip/index)(38) to identify key interaction sites, such as hydrogen bonds and hydrophobic interactions and binding conformations were visualized via PyMOL and Discovery Studio for 3D and 2D interaction mapping, respectively. 2.8 Immune infiltration To evaluate the potential role of the key target in the tumor immune microenvironment, single-sample Gene Set Enrichment Analysis (ssGSEA) was performed to quantify the relative abundance of various immune cell populations( 39 ). Cell-type-specific gene sets were obtained from the CellMarker database and formatted into GMT files for analysis. ssGSEA was applied to the normalized expression matrix to generate enrichment scores for each sample from TCGA. To further investigate the relationship between the key target gene and immune infiltration, samples were divided into high and low expression groups based on the median expression level of the gene in tumor tissues. Subsequently, the Wilcoxon rank-sum test was applied to compare differences in immune cell infiltration between the two groups, identifying immune cell types with significant variation under different expression statuses. Additionally, Spearman correlation analysis was performed to assess the associations between the key gene expression and ssGSEA scores of various immune cells, thereby revealing its potential role in modulating the tumor immune microenvironment. 2.9 statistical analysis In this study, data analysis was conducted using R software (version 4.4.0). Differential expression analysis was performed using the "limma" package in R software (version 4.4.0), with the filtering criteria set to p-value < 0.05 and |log2FC| ≥ 0.585, in order to obtain a comprehensive set of DEGs. GO and KEGG enrichment analyses were conducted using the “clusterProfiler”, “org.Hs.eg.db”, “enrichplot”, and “ggplot2” packages, with a significance cutoff of p < 0.05 for enriched term selection. To identify key genes, multiple machine learning algorithms were applied for feature selection. The LASSO regression model was constructed using the “glmnet” package, the random forest model with the “randomForest” package, and the SVM-RFE model with the “e1071”, “kernlab”, and “caret” packages. The predictive performance of all models was assessed by plotting ROC curves and calculating the AUC using the “pROC” package. For MR analysis, we sequentially employed the “VariantAnnotation”, “gwasglue”, and “TwoSampleMR” packages. Results were reported as odds ratios (ORs) with 95% confidence intervals (CIs), and statistical significance was defined as p < 0.05. Immune infiltration analysis was performed using the "GSVA" package for ssGSEA scoring, with immune cell annotations converted to GMT format via the "GSEABase" package. Wilcoxon rank-sum test compared immune cell abundance between groups. Data preprocessing and correlation analysis were done using "dplyr" and Spearman correlation assessed the relationship between gene expression and immune cell scores. Additionally, visualization tools included "ggvenn" for Venn diagrams, "ggplot2" and "ggrepel" for volcano plots and correlation plots, "pheatmap" for heatmaps, "grid", "readr" and "forestploter" for forest plots, and "ggpubr" for boxplots. 3. Results 3.1 Identification of Toxic Target Genes Associated with PP-Induced Breast Cancer In this study, the standard two-dimensional (2D) and three-dimensional (3D) molecular structures of PP were first retrieved from the PubChem database (Figs. 1 A-B). Subsequently, ProTox 3.0 platform was used to predict the toxicological features of PP (Fig. 1 C), including acute toxicity, toxicity classification, and structural similarity. Additionally, potential toxicological signaling pathways associated with PP were explored (Supplementary Table 1). Radar plot visualization (Fig. 1 D) revealed that PP exhibits multiple biological activities, particularly demonstrating a high activity probability in the ER signaling pathway, suggesting that PP may exert estrogen-like effects and potentially promote BC development through ER pathway activation. To comprehensively identify the potential molecular targets of PP, we integrated predictions from three databases-SuperPred, STITCH, and Swiss Target Prediction, resulting in a total of 144 candidate target genes associated with PP (Fig. 1 E). Next, differential expression analysis was performed using transcriptomic data from BC patients in the TCGA database. A volcano plot was generated to illustrate the overall distribution of DEGs (Fig. 2 A), identifying a total of 4,788 significant DEGs. By intersecting the 144 predicted PP target genes with the DEGs, 50 overlapping genes were identified (Fig. 2 B). These genes were preliminarily considered potential key regulatory targets through which PP may influence BC progression. Furthermore, a heatmap was constructed to display the expression patterns of these 50 genes in BC tissues versus normal tissues (Fig. 2 C), revealing distinct differential expression characteristics. Additionally, a compound–target interaction network was constructed to map the interactions between PP and its potential target genes (Fig. 2 D), providing a visual representation of how PP may exert its carcinogenic effects by modulating multiple key targets. 3.2 GO Functional Annotation and KEGG Pathway Enrichment Analysis of Target Genes To further investigate the biological functions of the potential PP target genes in BC, GO functional annotation and KEGG pathway enrichment analyses were performed on the 50 intersecting genes. The analyses were conducted using the R, with the species restricted to Homo sapiens , and included both GO and KEGG enrichment assessments. GO analysis was performed across three dimensions: BP, CC, and MF. The top ten most significantly enriched terms from each category were visualized (Fig. 2 E). In the BP category, rhythmic process was significantly enriched. Disruption of circadian rhythms shapes the tumor microenvironment, promotes epithelial-mesenchymal transition (EMT), and thereby influences the progression and metastasis of BC( 40 ). Closely associated with normal mammary gland development, its dysregulation is considered a hallmark of early pathological changes in BC( 41 ). In the MF category, the regulation of nuclear steroid receptor activity pathway was significantly enriched, suggesting that PP may influence typical nuclear steroid receptors such as estrogen receptors (ERα/ERβ), thereby contributing to the development of ER + BC( 2 ). Moreover, the enrichment of estrogen response element (ERE) binding activity further indicates that these genes may participate in the regulation of ER binding to promoter regions of target genes, playing a critical role in the progression of BC( 42 ). Further KEGG pathway enrichment analysis (Fig. 2 F) revealed that these target genes are significant enriched in multiple cancer-related signaling pathways, with the proteoglycans in cancer being the most prominent. Proteoglycans are complexes composed of a core protein and glycosaminoglycans (GAGs), broadly distributed on the cell membrane and within the extracellular matrix (ECM). They play a key role in regulating cell adhesion, migration, and interactions between ECM components and cells. The genetic signatures of proteoglycans have emerged as important biomarkers for the diagnosis, prognosis, and targeted therapy of BC( 43 ). The estrogen signaling pathway is a central driving mechanism in ER⁺ BC( 44 ). These findings further support the hypothesis that PP, as an estrogen-like compound, may promote breast cancer development by influencing proteoglycan expression and modulating estrogen signaling. Additionally, these genes were also significantly enriched in the MAPK signaling pathway, lipid and atherosclerosis, and other key signaling cascades that are closely involved in regulating cell proliferation, apoptosis, and the tumor microenvironment( 45 ). 3.3 Construction of the PPI Protein–Protein Interaction Network To further explore the potential molecular mechanisms linking PP and BC, we constructed a PPI network based on the 50 intersecting genes identified earlier. The network was generated using the STRING database and visualized with Cytoscape software for topological analysis. In the PPI network, each node represents a candidate target gene, and edges indicate known or predicted protein–protein interactions. We performed topological analysis by ranking nodes based on their degree values (i.e., the number of direct interactions with other proteins). The resulting network exhibited a clear hub structure, with several genes displaying high connectivity, suggesting their potential roles in mediating PP-related signaling pathways in BC (Fig. 3 A). To identify biologically relevant hub regulators, we selected the top 20 genes ranked by degree as potential hub genes (Supplementary Table 2). These hub genes exhibited high connectivity in the PPI network and are likely to play pivotal roles in PP-induced breast cancer progression, providing a theoretical foundation for subsequent machine learning-based gene prioritization and survival analysis. 3.4 Identification of Core Genes via Machine Learning Algorithms To further identify key regulatory target genes potentially involved in PP-mediated BC development, we applied three commonly used machine learning algorithms-LASSO regression, Random Forest and SVM-RFE to the top 20 hub genes ranked by degree in the previously constructed PPI network. In the LASSO regression analysis, the model exhibited minimal fluctuation around λ = λ.min. Based on non-zero regression coefficients, 13 candidate genes were selected for further analysis (Fig. 3 B). A Random Forest classification model was developed, and model performance was evaluated by plotting the error convergence curve as the number of trees increased (Fig. 3 C). When the number of trees reached 500, the model achieved optimal convergence, balancing accuracy and computational efficiency. Genes with importance scores (Mean Decrease Gini) greater than 5 were retained, resulting in 8 candidate genes (Fig. 3 D). SVM-RFE analysis identified 8 optimal genes with the lowest root mean squared error, as shown in the feature selection curve. These top-ranked genes were retained based on feature importance (Fig. 3 E). To integrate the results from the three algorithms, an intersection analysis of their respective candidate gene sets was performed, visualized using a Venn diagram (Fig. 3 F). A total of 6 genes were identified as common to all three methods and were preliminarily considered as the most promising regulatory targets in PP-induced BC development. To further characterize these 6 genes, boxplots were generated to display their expression patterns among different groups (Fig. 3 G). To assess their diagnostic performance, ROC curves were constructed and calculated the AUC was calculated for each gene (Fig. 3 H). Furthermore, to evaluate the overall diagnostic performance and robustness of the combined model based on the 6 candidate genes, an integrated ROC curve was constructed. The model demonstrated excellent discriminative ability ( AUC = 0.996,95%CI:0.992–0.999 ) (Supplementary Fig. 1). All 6 genes demonstrated strong diagnostic power for distinguishing between the disease and control groups, with AUC values of 0.988 for CA4 , 0.975 for CDK1 , 0.946 for CDK5 , 0.948 for MADA , 0.965 for NR3C1 , and 0.904 for PIK3R1. These results suggest that the identified genes may serve as promising biomarkers for BC diagnosis related to PP exposure and warrant further functional and clinical validation. Subsequent studies will focus on these genes for mechanistic investigation and functional validation. 3.5 Mendelian Randomization Analysis To validate the causal relationships between PP and key BC target genes, we performed MR analysis. Expression quantitative trait loci (eQTLs) for the candidate genes were retrieved from the IEU Open GWAS database, and suitable instrumental variables (IVs) were selected for MR based on the following criteria: P-value < 1×10⁻⁵; Linkage disequilibrium r² 10. Using these standards, valid eQTLs were identified for two genes: PIK3R1 and CA4. The remaining four genes were excluded from MR analysis due to the lack of qualified SNPs (Supplementary Table 3). By integrating the selected eQTLs with the GWAS dataset for ER⁺ BC (ieu-a-1127), MR analysis was performed using the Inverse Variance Weighted (IVW) method as the primary approach. The results demonstrated a significant causal association between decreased PIK3R1 expression and increased BC risk ( OR IVW = 0.869, 95% CI: 0.774–0.975, P = 0.017 ) (Fig. 4 A). Collectively, these results suggest that PIK3R1 may exert a protective role in the development of PP-induced ER⁺ BC, and that reduced expression of PIK3R1 may significantly elevate the risk of disease onset. The Mendelian randomization analysis did not show a statistically significant causal association for CA4 (Supplementary Fig. 2). To assess the robustness of the MR analysis, sensitivity analyses were performed. Cochran’s Q test revealed no significant heterogeneity in the IVW model ( Q = 7.15, P = 0.13 ), and residual heterogeneity in the MR-Egger model was also non-significant ( Q = 3.35, P = 0.34 ), indicating good model stability. The funnel plot showed a relatively symmetrical distribution of SNP effects, indicating no substantial directional bias and suggesting an absence of publication or selection bias (Fig. 4 B). In addition, the intercept test from MR-Egger regression yielded a P-value of 0.16, suggesting no evidence of horizontal pleiotropy. In addition, we plotted a funnel plot. In the scatter plot (Fig. 4 C), the slopes of fitted lines from multiple MR methods were largely consistent, indicating that the estimated causal effect of PIK3R1 expression on ER⁺ BC risk was directionally robust across different analytical approaches. Sensitivity analyses were also performed, the forest plot (Fig. 4 D) displayed individual causal estimates for each SNP and their contribution to the overall MR result, allowing visual assessment of effect direction and confidence intervals. Leave-one-out sensitivity analysis (Fig. 4 E) revealed that exclusion of any single SNP did not substantially alter the overall MR estimate, supporting the robustness of the results despite minor heterogeneity potentially introduced by specific variants. These results support the validity of the instrumental variables and enhance confidence in the causal inference. 3.6 Differential Expression and Prognostic Analysis of PIK3R1 in Breast Cancer Based on RNA-seq data from the TCGA database, including 1086 BC tissues and 99 adjacent normal tissues, the expression difference of PIK3R1 was evaluated using the Mann-Whitney U test. The results showed that PIK3R1 expression was significantly lower in BC tissues compared to normal tissues ( P < 0.05 ), suggesting possible downregulation in tumors (Fig. 5 A). To further explore the clinical relevance of PIK3R1 expression in breast cancer prognosis, the KM Plotter online platform ( https://kmplot.com/ ) was used to analyze the association between PIK3R1 expression levels and OS. A total of 4929 BC cases were included, divided into a high-expression group (n = 2462) and a low-expression group (n = 2467) based on median expression. Kaplan–Meier survival analysis indicated that patients in the high-expression group had significantly better OS compared to the low-expression group ( OR = 0.70, 95% CI: 0.63–0.78 ) (Fig. 5 B). These findings suggest that PIK3R1 may function as a protective factor in BC, and its high expression is closely associated with improved survival outcomes. 3.7 Molecular Docking Analysis of PIK3R1 To further evaluate the potential interaction between PP and the key BC target protein PIK3R1, molecular docking analysis was performed. The docking results showed that the lowest binding free energy between PP and PIK3R1 was − 6.37 kcal/mol, indicating a strong binding affinity (Fig. 5 C). Subsequent interaction analysis using the PLIP Web Tool revealed that PP stably binds within the active pocket of PIK3R1 through multiple non-covalent interactions. Specifically, PP formed several stable hydrogen bonds with residues GLU650, GLN654, and ASN630, contributing to the complex’s stability. In addition, hydrophobic interactions were observed with SER628, SER629, and ARG649, and a π–π stacking interaction occurred between PP and SER651, further enhancing the binding strength. Auxiliary polar interactions were also noted with residues ARG631 and ASN632. Three-dimensional structural visualization showed that PP is stably embedded in the hydrophobic pocket of PIK3R1, with a clear binding conformation and a dense distribution of surrounding residues, providing a structural basis for potential functional modulation. Collectively, these findings support the hypothesis that PP may exert pathogenic effects in BC by directly binding to and regulating the activity of PIK3R1. 3.8 Association of PIK3R1 Expression with Immune Infiltration and Tumor Microenvironment Based on the single-sample Gene Set Enrichment Analysis (ssGSEA), comparison between groups revealed that multiple immune cell infiltrations were significantly increased in the PIK3R1 high-expression group, including activated B cells, effector memory CD8⁺ T cells, eosinophils, immature B cells, mast cells, natural killer cells, neutrophils, regulatory T cells, follicular helper T cells, Th1 cells, and Th2 cells. In contrast, activated CD8⁺ T cells, CD56 dim natural killer cells, and monocytes showed higher infiltration levels in the low-expression group (Fig. 6 A). Spearman correlation analysis further revealed significant associations between PIK3R1 expression and various immune cells (Fig. 6 B). Specifically, PIK3R1 expression was positively correlated with mast cells (ρ = 0.29), plasmacytoid dendritic cells (ρ = 0.42), memory B cells (ρ = 0.21), regulatory T cells (ρ = 0.17), follicular helper T cells (ρ = 0.24), Th1 cells (ρ = 0.22), and Th2 cells (ρ = 0.17), while showing negative correlations with CD56 dim natural killer cell (ρ = −0.20), activated CD8⁺ T cells (ρ = −0.16), and monocytes (ρ = −0.30) (all p < 0.001). Notably, regulatory T cells, Th1 cells, Th2 cells, and mast cells exhibited consistent trends in both the differential and correlation analyses, with higher infiltration levels observed in the PIK3R1 high-expression group. These findings suggest that PIK3R1 may play a protective role in maintaining immune microenvironment homeostasis by promoting the enrichment of antitumor immune-related cells. 4. Discussion As a potential emerging environmental contaminant, PP has not been systematically investigated for its association with BC, particularly ER⁺ subtypes. In this study, we employed a comprehensive approach integrating network toxicology, bioinformatics, machine learning, Mendelian randomization, and molecular docking to systematically elucidate the potential molecular mechanisms by which PP may contribute to breast cancer development. Fiftyputative PP target genes were identified through toxicity prediction and integration with TCGA datasets. Prioritized through GO/KEGG enrichment analysis and PPI network construction, six core genes were selected by LASSO regression, Random Forest, and SVM-RFE. ROC analysis showed excellent diagnostic performance of the core genes (AUC > 0.90). PIK3R1 was further confirmed as a causal risk gene for breast cancer using MR analysis. At the experimental level, differential expression, survival analysis, and molecular docking demonstrated that PP downregulates PIK3R1 expression and promotes BC cell proliferation. Immune infiltration analysis revealed PIK3R1’s close association with antitumor immunity. These findings consistently support PIK3R1 as a key molecular target mediating BC related to PP progression, uncovering a novel mechanism of environmental exposure in ER⁺ breast cancer. Parabens, including PP, are widely used as preservatives in daily consumer products. Humans are exposed to these compounds through various routes, such as the use of cosmetics, consumption of food and pharmaceuticals, and contact with industrial goods. Previous studies have suggested that parabens may possess carcinogenic potential( 46 ). An epidemiological study comparing normal and malignant breast tissues in patients with BC revealed a significant association between paraben accumulation and BC risk, particularly in ER + and PR + cases( 47 ). Among parabens, PP is the most used and has attracted extensive attention. PP has been reported to adversely affect multiple human physiological systems, posing broad health risks. For instance, PP may interfere with neurodevelopment in children, increasing the risk of attention deficit hyperactivity disorder, with notable sex-specific differences in effect( 48 ). Additionally, PP exposure has been associated with elevated levels of xanthine, hypoxanthine, uric acid, and acylcarnitines, indicating disruptions in purine metabolism and fatty acid β-oxidation, which may contribute to energy metabolic disorders and a heightened risk of diabetes( 49 ). PP has also been implicated in reproductive toxicity, with evidence suggesting reduced fertility following exposure( 50 ). As an EDC, PP structurally resembles estrogen and exhibits estrogenic activity, potentially playing a crucial role in the initiation and progression of BC( 51 ). Our findings support the notion that PP exposure increases breast cancer risk. To further explore the underlying mechanisms, GO and KEGG enrichment analyses were conducted on 50 potential BCtarget genes. These genes were significantly enriched in the estrogen signaling pathway. The estrogen signaling pathway is closely linked to breast carcinogenesis. Its oncogenic mechanisms include the interaction of 17β-estradiol with estrogen receptors, the genotoxic effects of estrogen metabolites, and epigenetic modifications. As a xenoestrogen, PP can activate this pathway, thereby trigger carcinogenic processes and disrupt normal breast tissue growth, proliferation, and differentiation, ultimately contributing to malignancy development( 52 ). Moreover, the enrichment analysis suggested that biological rhythms, and their disruption influences tumor angiogenesis and the immune microenvironment, thereby increasing the risk of BC( 53 ). PP may also influence cancer invasiveness via proteoglycan pathways. Core proteoglycans have been shown to inhibit tumor growth and invasion in inflammatory breast cancer by destabilizing E-cadherin and suppressing EGFR/ERK signaling( 54 ). Given these findings, it is necessary to adopt precautionary measures, such as reducing the use of PP in preservative systems and implementing more comprehensive and standardized assessments of its application. Furthermore, increasing public awareness of the potential health risks associated with preservative ingredients is essential to minimize adverse effects caused by exposure to PP on human health. To further identify the key pathways involved in PP-induced BC, we integrated multiple public databases with TCGA data to screen for potential molecular targets. By combining PPI analysis, machine learning algorithms, and MR, we progressively narrowed down key candidate genes. Multilevel validation, including prognostic analysis, and molecular docking, consistently highlighted PIK3R1 as the core gene mediating PP-induced breast carcinogenesis. The PIK3R1 gene, located on chromosome 5q13.1, encodes the regulatory subunit p85α of class I PI3K, a well-recognized tumor suppressor. Its mutation, amplification, or deletion has been widely reported as a molecular hallmark of metastatic BC( 55 ). Together with the p110 catalytic subunit encoded by PIK3CA, p85α forms the PI3K complex, which plays a central role in estrogen signaling( 56 ). While PIK3CA mutations are typically early events in breast tumorigenesis, PIK3R1 dysregulation is more often associated with tumor invasion( 57 ). Abnormalities in either gene can lead to sustained activation of the PI3K/Akt signaling pathway, thereby promoting cell proliferation, angiogenesis, and apoptosis resistance. The PI3K/Akt pathway has been implicated in various malignancies, including breast, lung, and cervical cancers( 45 , 58 ). In ER⁺ BC, estrogen or growth factors bind to tyrosine kinase receptors or G protein–coupled estrogen receptors, initiating a cascade that activates PI3K. Activated PI3K catalyzes the generation of PIP3, which recruits and activates Akt, triggering downstream phosphorylation events that regulate cell proliferation, survival, and motility( 59 ). Previous studies have shown that targeting PIK3R1 can inhibit PI3K/Akt pathway activation and reverse epithelial–mesenchymal transition, a process critical for cancer invasion, metastasis, and stemness acquisition( 60 , 61 ). Under physiological conditions, p85α binds tightly to p110 and suppresses its catalytic activity( 62 ). However, under PP exposure, downregulation of PIK3R1 may lead to reduced levels of free p85α, relieving inhibition on p110, and consequently enhancing PI3K/Akt signaling( 63 ). This aberrant activation may promote tumor cell growth, survival, and angiogenesis, ultimately accelerating breast cancer progression. Despite the comprehensive integration of multi-omics data and the application of advanced analytical methods to elucidate the molecular mechanisms underlying breast cancer related to PP exposure, this study has several limitations. First, although network toxicology analyses identified multiple candidate targets and signaling pathways associated with PP, the subsequent analyses primarily focused on PIK3R1, which may have led to the oversight of other potentially important regulatory genes or pathways. Second, although molecular docking and in vitro assays preliminarily validated the interaction between PP and PIK3R1, the lack of in vivo functional validation limits the depth and completeness of mechanistic insight. Future research should incorporate animal models, multidimensional functional assays, and large-scale, multi-ethnic, and multicenter cohort data to strengthen the causal evidence and enhance the generalizability and robustness of the conclusions. Nevertheless, this study presents several important strengths. First, it is methodologically innovative and systematically designed, integrating diverse approaches, including network toxicology, transcriptomic analysis, protein–protein interaction network construction, machine learning–based gene prioritization, Mendelian randomization, molecular docking, and cellular validation-into a coherent research framework. This comprehensive strategy greatly enhances the reliability and scientific rigor of the findings. Second, beyond computational identification of candidate genes, the study incorporates clinical prognostic analysis and experimental validation, ensuring the biological and translational relevance of the results. Notably, this is the first study to clearly identify PIK3R1 as a core mediator of PP-induced breast carcinogenesis, expanding the theoretical understanding of environmental carcinogenesis and offering a novel molecular target for future preventive and therapeutic strategies. In conclusion, this study is the first to systematically reveal that exposure to PP may promote the initiation and progression of BC by downregulating PIK3R1 expression and activating the PI3K/Akt signaling pathway. These findings not only deepen our understanding of the molecular carcinogenic mechanisms of environmental chemicals, but also highlight PIK3R1 as a key tumor suppressor involved in PP-induced breast carcinogenesis. This provides a theoretical foundation for future risk assessment and the development of targeted prevention and therapeutic strategies for environmentally induced cancers. 5. Conclusion This study systematically elucidates the potential molecular mechanism by which PP exposure promotes BC initiation and progression via downregulation of PIK3R1 and activation of the PI3K/Akt signaling pathway. By integrating multiple approaches—including network toxicology, transcriptomic differential expression analysis, PPI network construction, machine learning–based gene prioritization, Mendelian randomization, molecular docking, and cellular validation-PIK3R1 was confirmed as a key regulatory gene in breast carcinogenesis related to PP exposure. These findings not only expand our molecular understanding of the carcinogenic mechanisms of environmental chemicals but also highlight the potential health risks posed by PP as an EDC. Moreover, they underscore the biological significance of PIK3R1 as a putative tumor suppressor. Collectively, this study provides a theoretical basis for the development of targeted prevention and therapeutic strategies and offers scientific insight for future research on environmentally induced cancers and the risk assessment and control of environmental carcinogens. Declarations Ethics approval and consent to participate Not applicable. Consent for publication Not applicable. Competing interests The authors declare that they have no competing interests. Author contributors Kaiwen Wang : Conceptualization, Methodology, Writing-original draft, Writing-review & editing; Guang Yao : Formal Analysis, Resources, Writing-original draft; Xiaobin Zhang : Investigation, Software, Visualization, Writing-original draft; Ruiting Ma : Visualization, Writing-original draft; Xinyi Li : Methodology, Writing-review & editing; Weize Kong : Supervision, Validation; Xu Hui : Data Curation, Software; Qian Liu : Formal Analysis, Investigation; Yanan Wu : Software, Visualization; Yi Xiao : Methodology, Supervision; Jingru Yang : Funding Acquisition, Resources, Writing-review & editing; Yongbin Lu : Funding Acquisition, Resources, Writing-review & editing. Funding This work was supported by the Natural Science Foundation of Gansu Province (22JR5RA907); the Postdoctoral Fellowship Program of CPSF under Grant Number (No. GZC20240624); the Science and Technology Program of Gansu Province (No. 24JRRA508); the Fundamental Research Funds for The Central Universities (lzujbky-2024-pd01); the Wu Jieping Medical Foundation Clinical Research Special Fund (320.6750, 2022-21-61); and the United Research Foundation of Gansu Province (25JRRA1250). Availability of data and materials Data will be made available on request. Requests for specific analyses or data can be submitted by email to [email protected] . Acknowledgements Not applicable. References Bray F, Laversanne M, Sung H, Ferlay J, Siegel RL, Soerjomataram I, et al. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 2024;74(3):229-63. DeSantis CE, Ma J, Gaudet MM, Newman LA, Miller KD, Goding Sauer A, et al. Breast cancer statistics, 2019. CA Cancer J Clin. 2019;69(6):438-51. Harbeck N, Penault-Llorca F, Cortes J, Gnant M, Houssami N, Poortmans P, et al. Breast cancer. Nat Rev Dis Primers. 2019;5(1):66. Xu H, Xu B. Breast cancer: Epidemiology, risk factors and screening. Chin J Cancer Res. 2023;35(6):565-83. Katsura C, Ogunmwonyi I, Kankam HK, Saha S. Breast cancer: presentation, investigation and management. Br J Hosp Med (Lond). 2022;83(2):1-7. Prossnitz ER, Barton M. The G protein-coupled oestrogen receptor GPER in health and disease: an update. Nat Rev Endocrinol. 2023;19(7):407-24. Sang C, Song Y, Jin TW, Zhang S, Fu L, Zhao Y, et al. Bisphenol A induces ovarian cancer cell proliferation and metastasis through estrogen receptor-α pathways. Environ Sci Pollut Res Int. 2021;28(27):36060-8. Głaz P, Rosińska A, Woźniak S, Boguszewska-Czubara A, Biernasiuk A, Matosiuk D. Effect of Commonly Used Cosmetic Preservatives on Healthy Human Skin Cells. Cells. 2023;12(7). Vandenberg LN, Bugos J. Assessing the Public Health Implications of the Food Preservative Propylparaben: Has This Chemical Been Safely Used for Decades. Curr Environ Health Rep. 2021;8(1):54-70. Petric Z, Ružić J, Žuntar I. The controversies of parabens - an overview nowadays. Acta Pharm. 2021;71(1):17-32. Räsänen A, Albrecht E, Annala M, Aro L, Laine AM, Maanavilja L, et al. After-use of peat extraction sites - A systematic review of biodiversity, climate, hydrological and social impacts. Sci Total Environ. 2023;882:163583. Gonzales I, Sengupta R, Mani S, Correa-Soto CE, Bejgum B, Alvarez-Nunez F, et al. Understanding the sorption of paraben on plastics using molecular dynamics simulations. Int J Pharm. 2025;671:125256. Géniès C, Jacques-Jamin C, Duplan H, Rothe H, Ellison C, Cubberley R, et al. Comparison of the metabolism of 10 cosmetics-relevant chemicals in EpiSkin™ S9 subcellular fractions and in vitro human skin explants. J Appl Toxicol. 2020;40(2):313-26. Mao W, Qu J, Zhong S, Wu X, Mao K, Liao K, et al. Associations between urinary parabens and lung cancer. Environmental science and pollution research international. 2023;30(24):66186-94. Fernández-Martínez NF, Rodríguez-Barranco M, Zamora-Ros R, Guevara M, Colorado-Yohar SM, Jiménez-Zabala A, et al. Relationship between exposure to parabens and benzophenones and prostate cancer risk in the EPIC-Spain cohort. Environmental science and pollution research international. 2024;31(4):6186-99. Wróbel AM, Gregoraszczuk E. Action of methyl-, propyl- and butylparaben on GPR30 gene and protein expression, cAMP levels and activation of ERK1/2 and PI3K/Akt signaling pathways in MCF-7 breast cancer cells and MCF-10A non-transformed breast epithelial cells. Toxicology letters. 2015;238(2):110-6. Liang J, Liu QS, Ren Z, Min K, Yang X, Hao F, et al. Studying paraben-induced estrogen receptor- and steroid hormone-related endocrine disruption effects via multi-level approaches. Sci Total Environ. 2023;869:161793. Opinion on Parabens. 2010. Xiong X, Zheng LW, Ding Y, Chen YF, Cai YW, Wang LP, et al. Breast cancer: pathogenesis and treatments. Signal Transduct Target Ther. 2025;10(1):49. Okubo T, Yokoyama Y, Kano K, Kano I. ER-dependent estrogenic activity of parabens assessed by proliferation of human breast cancer MCF-7 cells and expression of ERalpha and PR. Food Chem Toxicol. 2001;39(12):1225-32. Byford JR, Shaw LE, Drew MG, Pope GS, Sauer MJ, Darbre PD. Oestrogenic activity of parabens in MCF7 human breast cancer cells. J Steroid Biochem Mol Biol. 2002;80(1):49-60. Kim S, Chen J, Cheng T, Gindulyte A, He J, He S, et al. PubChem 2025 update. Nucleic Acids Res. 2025;53(D1):D1516-d25. Banerjee P, Kemmler E, Dunkel M, Preissner R. ProTox 3.0: a webserver for the prediction of toxicity of chemicals. Nucleic Acids Res. 2024;52(W1):W513-w20. Gallo K, Goede A, Preissner R, Gohlke BO. SuperPred 3.0: drug classification and target prediction-a machine learning approach. Nucleic Acids Res. 2022;50(W1):W726-w31. Szklarczyk D, Santos A, von Mering C, Jensen LJ, Bork P, Kuhn M. STITCH 5: augmenting protein-chemical interaction networks with tissue and affinity data. Nucleic Acids Res. 2016;44(D1):D380-4. Daina A, Michielin O, Zoete V. SwissTargetPrediction: updated data and new features for efficient prediction of protein targets of small molecules. Nucleic Acids Res. 2019;47(W1):W357-w64. Tomczak K, Czerwińska P, Wiznerowicz M. The Cancer Genome Atlas (TCGA): an immeasurable source of knowledge. Contemp Oncol (Pozn). 2015;19(1a):A68-77. Szklarczyk D, Nastou K, Koutrouli M, Kirsch R, Mehryary F, Hachilif R, et al. The STRING database in 2025: protein networks with directionality of regulation. Nucleic Acids Res. 2025;53(D1):D730-d7. Majeed A, Mukhtar S. Protein-Protein Interaction Network Exploration Using Cytoscape. Methods Mol Biol. 2023;2690:419-27. Svetnik V, Liaw A, Tong C, Culberson JC, Sheridan RP, Feuston BP. Random forest: a classification and regression tool for compound classification and QSAR modeling. J Chem Inf Comput Sci. 2003;43(6):1947-58. Fontanarosa JB, Dai Y. Using LASSO regression to detect predictive aggregate effects in genetic studies. BMC Proc. 2011;5 Suppl 9(Suppl 9):S69. Duan KB, Rajapakse JC, Wang H, Azuaje F. Multiple SVM-RFE for gene selection in cancer classification with expression data. IEEE Trans Nanobioscience. 2005;4(3):228-34. Michailidou K, Lindström S, Dennis J, Beesley J, Hui S, Kar S, et al. Association analysis identifies 65 new breast cancer risk loci. Nature. 2017;551(7678):92-4. Jia B, Zhao X, Wang Y, Wang J, Wang Y, Yang Y. Prognostic roles of MAGE family members in breast cancer based on KM-Plotter Data. Oncol Lett. 2019;18(4):3501-16. UniProt: the Universal Protein Knowledgebase in 2025. Nucleic Acids Res. 2025;53(D1):D609-d17. Berman HM, Burley SK. Protein Data Bank (PDB): Fifty-three years young and having a transformative impact on science and society. Q Rev Biophys. 2025;58:e9. Seeliger D, de Groot BL. Ligand docking and binding site analysis with PyMOL and Autodock/Vina. J Comput Aided Mol Des. 2010;24(5):417-22. Eberhardt J, Santos-Martins D, Tillack AF, Forli S. AutoDock Vina 1.2.0: New Docking Methods, Expanded Force Field, and Python Bindings. J Chem Inf Model. 2021;61(8):3891-8. Charoentong P, Finotello F, Angelova M, Mayer C, Efremova M, Rieder D, et al. Pan-cancer Immunogenomic Analyses Reveal Genotype-Immunophenotype Relationships and Predictors of Response to Checkpoint Blockade. Cell Rep. 2017;18(1):248-62. Li DD, Zhou T, Gao J, Wu GL, Yang GR. Circadian rhythms and breast cancer: from molecular level to therapeutic advancements. J Cancer Res Clin Oncol. 2024;150(9):419. Ingthorsson S, Traustadottir GA, Gudjonsson T. Breast Morphogenesis: From Normal Development to Cancer. Adv Exp Med Biol. 2025;1464:29-44. Toska E. Epigenetic mechanisms of cancer progression and therapy resistance in estrogen-receptor (ER+) breast cancer. Biochim Biophys Acta Rev Cancer. 2024;1879(3):189097. Buraschi S, Pascal G, Liberatore F, Iozzo RV. Comprehensive investigation of proteoglycan gene expression in breast cancer: Discovery of a unique proteoglycan gene signature linked to the malignant phenotype. Proteoglycan Res. 2025;3(1). Treeck O, Haerteis S, Ortmann O. Non-Coding RNAs Modulating Estrogen Signaling and Response to Endocrine Therapy in Breast Cancer. Cancers (Basel). 2023;15(6). Wang D, Liu X, Hong W, Xiao T, Xu Y, Fang X, et al. Muscone abrogates breast cancer progression through tumor angiogenic suppression via VEGF/PI3K/Akt/MAPK signaling pathways. Cancer Cell Int. 2024;24(1):214. Wang L, Chen L, Schlenk D, Li F, Liu J. Parabens promotes invasive properties of multiple human cells: A potential cancer-associated adverse outcome pathway. Sci Total Environ. 2024;926:172015. Amin MM, Tabatabaeian M, Chavoshani A, Amjadi E, Hashemi M, Ebrahimpour K, et al. Paraben Content in Adjacent Normal-malignant Breast Tissues from Women with Breast Cancer. Biomed Environ Sci. 2019;32(12):893-904. Tung CJ, Chen MH, Lin CC, Chen PC. Association between parabens exposure and neurodevelopment in children. Environ Int. 2024;188:108671. Zhao H, Zheng Y, Zhu L, Xiang L, Zhou Y, Li J, et al. Paraben Exposure Related To Purine Metabolism and Other Pathways Revealed by Mass Spectrometry-Based Metabolomics. Environ Sci Technol. 2020;54(6):3447-54. Ao J, Qiu W, Huo X, Wang Y, Wang W, Zhang Q, et al. Paraben exposure and couple fecundity: a preconception cohort study. Hum Reprod. 2023;38(4):726-38. Khanna S, Dash PR, Darbre PD. Exposure to parabens at the concentration of maximal proliferative response increases migratory and invasive activity of human breast cancer cells in vitro. J Appl Toxicol. 2014;34(9):1051-9. Starek-Świechowicz B, Budziszewska B, Starek A. Endogenous estrogens-breast cancer and chemoprevention. Pharmacol Rep. 2021;73(6):1497-512. Yan Y, Su L, Huang S, He Q, Lu J, Luo H, et al. Circadian rhythms and breast cancer: unraveling the biological clock's role in tumor microenvironment and ageing. Front Immunol. 2024;15:1444426. Ye H, Shen X, Li Y, Zou W, Hassan SSU, Feng Y, et al. Proteomic and metabolomic characterization of bone, liver, and lung metastases in plasma of breast cancer patients. Proteomics Clin Appl. 2024;18(5):e2300136. Wheler JJ, Parker BA, Lee JJ, Atkins JT, Janku F, Tsimberidou AM, et al. Unique molecular signatures as a hallmark of patients with metastatic breast cancer: implications for current treatment paradigms. Oncotarget. 2014;5(9):2349-54. Vallejo-Díaz J, Chagoyen M, Olazabal-Morán M, González-García A, Carrera AC. The Opposing Roles of PIK3R1/p85α and PIK3R2/p85β in Cancer. Trends Cancer. 2019;5(4):233-44. Miron A, Varadi M, Carrasco D, Li H, Luongo L, Kim HJ, et al. PIK3CA mutations in in situ and invasive breast carcinomas. Cancer Res. 2010;70(14):5674-8. Moghbeli M. PI3K/AKT pathway as a pivotal regulator of epithelial-mesenchymal transition in lung tumor cells. Cancer Cell Int. 2024;24(1):165. Khatpe AS, Adebayo AK, Herodotou CA, Kumar B, Nakshatri H. Nexus between PI3K/AKT and Estrogen Receptor Signaling in Breast Cancer. Cancers. 2021;13(3). Ansieau S. EMT in breast cancer stem cell generation. Cancer Lett. 2013;338(1):63-8. Yan LX, Liu YH, Xiang JW, Wu QN, Xu LB, Luo XL, et al. PIK3R1 targeting by miR-21 suppresses tumor cell migration and invasion by reducing PI3K/AKT signaling and reversing EMT, and predicts clinical outcome of breast cancer. Int J Oncol. 2016;48(2):471-84. Thorpe LM, Spangle JM, Ohlson CE, Cheng H, Roberts TM, Cantley LC, et al. PI3K-p110α mediates the oncogenic activity induced by loss of the novel tumor suppressor PI3K-p85α. Proceedings of the National Academy of Sciences of the United States of America. 2017;114(27):7095-100. Zhang HP, Jiang RY, Zhu JY, Sun KN, Huang Y, Zhou HH, et al. PI3K/AKT/mTOR signaling pathway: an important driver and therapeutic target in triple-negative breast cancer. Breast Cancer. 2024;31(4):539-51. Additional Declarations No competing interests reported. Supplementary Files supplementarytable1.docx Supplementary Table 1. ProTox-3.0 - Prediction of TOXicity of PP supplementarytable2.docx Supplementary Table 2. String Node Degrees supplementarytable3.docx Supplementary Table 3. Expression Quantitative Trait Locis of CA4 and PIK3R1 SupplementaryFigure1.tif Supplementary Figure 1. The integrated ROC curve of the six-gene model SupplementaryFigure2.tif Supplementary Figure 2. The forest plot about the MR results of CA4 Cite Share Download PDF Status: Under Review Version 1 posted Editorial decision: Revision requested 09 Feb, 2026 Reviews received at journal 05 Feb, 2026 Reviews received at journal 04 Feb, 2026 Reviewers agreed at journal 29 Jan, 2026 Reviewers agreed at journal 23 Jan, 2026 Reviewers invited by journal 20 Oct, 2025 Editor assigned by journal 15 Oct, 2025 Submission checks completed at journal 07 Oct, 2025 First submitted to journal 07 Oct, 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-7400738","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":532584517,"identity":"33ac2d51-8586-4415-a478-5f860610fc36","order_by":0,"name":"Kaiwen Wang","email":"","orcid":"","institution":"Lanzhou University","correspondingAuthor":false,"prefix":"","firstName":"Kaiwen","middleName":"","lastName":"Wang","suffix":""},{"id":532584518,"identity":"53f55dac-41c4-481e-bcf4-88ded5a2b0b1","order_by":1,"name":"Guang Yao","email":"","orcid":"","institution":"Lanzhou University","correspondingAuthor":false,"prefix":"","firstName":"Guang","middleName":"","lastName":"Yao","suffix":""},{"id":532584519,"identity":"bace09ba-8c63-43b7-9707-071d2d006d55","order_by":2,"name":"Xiaobin Zhang","email":"","orcid":"","institution":"The First Hospital of Lanzhou University","correspondingAuthor":false,"prefix":"","firstName":"Xiaobin","middleName":"","lastName":"Zhang","suffix":""},{"id":532584520,"identity":"86af7042-395c-4133-a067-1b7f0da0da31","order_by":3,"name":"Ruiting Ma","email":"","orcid":"","institution":"Lanzhou University","correspondingAuthor":false,"prefix":"","firstName":"Ruiting","middleName":"","lastName":"Ma","suffix":""},{"id":532584521,"identity":"e8b4dfdc-e64b-41f4-96db-972ca46eca20","order_by":4,"name":"Xinyi Li","email":"","orcid":"","institution":"Lanzhou University","correspondingAuthor":false,"prefix":"","firstName":"Xinyi","middleName":"","lastName":"Li","suffix":""},{"id":532584525,"identity":"a9a7c4dc-d23e-4740-b438-7f27ef4c979c","order_by":5,"name":"Weize Kong","email":"","orcid":"","institution":"Lanzhou University","correspondingAuthor":false,"prefix":"","firstName":"Weize","middleName":"","lastName":"Kong","suffix":""},{"id":532584526,"identity":"8453ae96-1841-4ec5-b316-135be23aba38","order_by":6,"name":"Xu Hui","email":"","orcid":"","institution":"Lanzhou University","correspondingAuthor":false,"prefix":"","firstName":"Xu","middleName":"","lastName":"Hui","suffix":""},{"id":532584528,"identity":"2dde2e5a-edd0-4c46-aed4-bdb3d871283d","order_by":7,"name":"Qian Liu","email":"","orcid":"","institution":"Lanzhou University","correspondingAuthor":false,"prefix":"","firstName":"Qian","middleName":"","lastName":"Liu","suffix":""},{"id":532584530,"identity":"72c6239a-b030-4415-8fb0-553bfaf6b865","order_by":8,"name":"Yanan Wu","email":"","orcid":"","institution":"Lanzhou University","correspondingAuthor":false,"prefix":"","firstName":"Yanan","middleName":"","lastName":"Wu","suffix":""},{"id":532584531,"identity":"a5c12be9-387f-4380-93f3-58f8e0b2eb2d","order_by":9,"name":"Yi Xiao","email":"","orcid":"","institution":"The First Hospital of Lanzhou University","correspondingAuthor":false,"prefix":"","firstName":"Yi","middleName":"","lastName":"Xiao","suffix":""},{"id":532584532,"identity":"cdc03f42-d0e6-4452-9bb4-47b2a6d601cd","order_by":10,"name":"Yongbin Lu","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAxUlEQVRIiWNgGAWjYBACPgYGNoaEChsZMI+HGC1sIPThTBoPaVoYZ7YdJkWLRI7ZY9628zzyMxIYH7xtY5A3J0KLuTHPuds8jDMSmA3ntjEY7mwgwhZpnrLbPMwSCWzSvG0MCQYHiNLCdo6HTSKB/TfRWiRntB3g4QHawkycFp5n5QYfziTzSPA8bJacc07CcAMhLfzsydseJFTYycm3Jx/88KbMRp6gLQwMHAZQBmMDkJAgqB4I2B8Qo2oUjIJRMApGMgAAb0s072fXEIcAAAAASUVORK5CYII=","orcid":"","institution":"Lanzhou University","correspondingAuthor":true,"prefix":"","firstName":"Yongbin","middleName":"","lastName":"Lu","suffix":""},{"id":532584533,"identity":"4050ba57-7b8b-4704-93b8-2580480ba954","order_by":11,"name":"Jingru Yang","email":"","orcid":"","institution":"Lanzhou University","correspondingAuthor":false,"prefix":"","firstName":"Jingru","middleName":"","lastName":"Yang","suffix":""}],"badges":[],"createdAt":"2025-08-18 14:38:13","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7400738/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7400738/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":94985428,"identity":"a490c1c5-ceaa-409b-9a82-eae831afb51c","added_by":"auto","created_at":"2025-11-03 06:58:10","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1950858,"visible":true,"origin":"","legend":"","description":"","filename":"Manuscript.docx","url":"https://assets-eu.researchsquare.com/files/rs-7400738/v1/b9f576cde1d5a748a79caf94.docx"},{"id":94846590,"identity":"8acd36ad-a64e-411a-bbe6-8e309413ed9b","added_by":"auto","created_at":"2025-10-31 10:18:46","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":11683,"visible":true,"origin":"","legend":"","description":"","filename":"dbc4fd69c3e8477795cd0ab1ebf6ddff.json","url":"https://assets-eu.researchsquare.com/files/rs-7400738/v1/547777e54710e18e8eeaf0ba.json"},{"id":94846596,"identity":"3eaff43a-00e1-4ba2-a069-1a23476cca5c","added_by":"auto","created_at":"2025-10-31 10:18:46","extension":"tif","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":95732,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryFigure1.tif","url":"https://assets-eu.researchsquare.com/files/rs-7400738/v1/36d12207ce440261e10c458d.tif"},{"id":94846606,"identity":"c6115a5a-4a4c-4184-8159-0c86cbb923f2","added_by":"auto","created_at":"2025-10-31 10:18:46","extension":"tif","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":273106,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryFigure2.tif","url":"https://assets-eu.researchsquare.com/files/rs-7400738/v1/cc8fb37ee5fd5dc93ef80d95.tif"},{"id":94846599,"identity":"4e630b26-f9d5-499b-ae8e-0b57e2a20114","added_by":"auto","created_at":"2025-10-31 10:18:46","extension":"docx","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":14883,"visible":true,"origin":"","legend":"","description":"","filename":"supplementarytable1.docx","url":"https://assets-eu.researchsquare.com/files/rs-7400738/v1/35c446fc9a481219e5307913.docx"},{"id":94985284,"identity":"9829612e-685b-42f6-af20-bf0d4e2e3176","added_by":"auto","created_at":"2025-11-03 06:57:51","extension":"docx","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":15424,"visible":true,"origin":"","legend":"","description":"","filename":"supplementarytable2.docx","url":"https://assets-eu.researchsquare.com/files/rs-7400738/v1/cb0d3a8e5522cc2468520507.docx"},{"id":94846603,"identity":"b413f66b-5d2b-4589-8ff0-ea1f32f4b55d","added_by":"auto","created_at":"2025-10-31 10:18:46","extension":"docx","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":19747,"visible":true,"origin":"","legend":"","description":"","filename":"supplementarytable3.docx","url":"https://assets-eu.researchsquare.com/files/rs-7400738/v1/05f1d44f40e18b779d51e683.docx"},{"id":94985799,"identity":"ff5ce3e1-f327-41d6-a8b5-5255738efb9b","added_by":"auto","created_at":"2025-11-03 06:59:02","extension":"xml","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":157210,"visible":true,"origin":"","legend":"","description":"","filename":"dbc4fd69c3e8477795cd0ab1ebf6ddff1enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-7400738/v1/c1020b4576885ad7252f5360.xml"},{"id":94846618,"identity":"2aa67c5e-7dd3-4e2a-8993-1c1ac467a7eb","added_by":"auto","created_at":"2025-10-31 10:18:46","extension":"jpeg","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":317480,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7400738/v1/f7e13986ec542b83a8c2c3cf.jpeg"},{"id":94985761,"identity":"db84d388-cee6-4589-832c-2e6af607906e","added_by":"auto","created_at":"2025-11-03 06:58:51","extension":"jpeg","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":512172,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage2.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7400738/v1/5dc868047a1ab074eda7d0bf.jpeg"},{"id":94846616,"identity":"70caf7be-8752-414b-b924-299a7d8a560a","added_by":"auto","created_at":"2025-10-31 10:18:46","extension":"jpeg","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":192978,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage3.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7400738/v1/0d29b534449265e5e661ac20.jpeg"},{"id":94846615,"identity":"af57e77e-a1c1-4395-90da-ed64921b6a00","added_by":"auto","created_at":"2025-10-31 10:18:46","extension":"jpeg","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":236518,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage4.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7400738/v1/d41ada9d9442e5543e7a06f0.jpeg"},{"id":94846613,"identity":"434fc89a-194a-4032-b0b7-438d7b21b6bd","added_by":"auto","created_at":"2025-10-31 10:18:46","extension":"jpeg","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":342928,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage5.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7400738/v1/64ea7521744cef4a7e8d1927.jpeg"},{"id":94846611,"identity":"67987b99-189e-4359-a120-217d3320c504","added_by":"auto","created_at":"2025-10-31 10:18:46","extension":"jpeg","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":182172,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage6.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7400738/v1/31a4e00148a0091ee5b8f005.jpeg"},{"id":94846620,"identity":"d8f473bd-52e6-4357-aedd-12a127982b96","added_by":"auto","created_at":"2025-10-31 10:18:47","extension":"png","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":55875,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7400738/v1/f40eb259953cf0f0b8d37c86.png"},{"id":94985122,"identity":"c21b46d5-a932-4c85-a466-945795686d5c","added_by":"auto","created_at":"2025-11-03 06:57:31","extension":"png","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":60281,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7400738/v1/b37d51494096eec214d03570.png"},{"id":94985011,"identity":"6fd94e5c-0617-437a-bf88-2a7667a7aad4","added_by":"auto","created_at":"2025-11-03 06:57:13","extension":"png","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":39742,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7400738/v1/56fa67a3ddf8373bd5748d76.png"},{"id":94846607,"identity":"46794617-3835-4675-9357-c75e55667c30","added_by":"auto","created_at":"2025-10-31 10:18:46","extension":"png","order_by":17,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":42889,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7400738/v1/cd86ca424ea8cf9b9d48a771.png"},{"id":94846619,"identity":"89df7092-e3bb-4145-9cff-08d0410f4fbe","added_by":"auto","created_at":"2025-10-31 10:18:46","extension":"png","order_by":18,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":63494,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-7400738/v1/6c13419663a3588b6c93f001.png"},{"id":94846612,"identity":"ca2f0703-a105-4a4c-81a7-e4a0dab0fde1","added_by":"auto","created_at":"2025-10-31 10:18:46","extension":"png","order_by":19,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":30462,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-7400738/v1/cc391be29fccd9170cbcf60a.png"},{"id":94846621,"identity":"3bad99d8-51c9-4c36-b583-2a1c7e59381f","added_by":"auto","created_at":"2025-10-31 10:18:47","extension":"xml","order_by":20,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":156932,"visible":true,"origin":"","legend":"","description":"","filename":"dbc4fd69c3e8477795cd0ab1ebf6ddff1structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-7400738/v1/5f0a5a2762b5baab0247ee54.xml"},{"id":94985380,"identity":"94c034a2-8263-47e0-97ea-0f407af33f6e","added_by":"auto","created_at":"2025-11-03 06:58:05","extension":"html","order_by":21,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":168124,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-7400738/v1/bdaed558ceb457c9f9ed43c8.html"},{"id":94846589,"identity":"ee57dddf-939a-4372-895e-35649c2fa8c0","added_by":"auto","created_at":"2025-10-31 10:18:46","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":132669,"visible":true,"origin":"","legend":"\u003cp\u003eStructural features, toxicity prediction, and potential target analysis of propyl p-hydroxybenzoate (PP). (A) 2D chemical structure of PP; (B) 3D molecular conformation of PP; (C) Toxicity prediction results from the ProTox 3.0 platform; (D) Radar plot visualizing predicted toxicological pathways associated with PP, highlighting high activity probability in the estrogen receptor (ER) signaling pathway; (E) Venn diagram illustrating the integration of candidate target genes from multiple databases for constructing the initial target gene library.\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-7400738/v1/3f74e83666fb907f445ac1cd.png"},{"id":94985604,"identity":"3929ceef-6014-4de9-8ff1-41970a319013","added_by":"auto","created_at":"2025-11-03 06:58:29","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":185666,"visible":true,"origin":"","legend":"\u003cp\u003eIntegrated analysis and functional annotation of PP target genes and breast cancer DEGs.(A) Volcano plot of differentially expressed genes (DEGs) based on TCGA breast cancer data; (B) Venn diagram showing the overlap between PP-predicted targets and DEGs; (C) Heatmap of the 50 intersecting genes based on TCGA data, comparing expression in tumor versus normal samples; (D) Compound–target–disease interaction network illustrating the relationships between PP, candidate targets, and breast cancer; (E) GO enrichment results for intersecting genes, showing the top 10 significant terms in Biological Process (BP), Cellular Component (CC), and Molecular Function (MF); (F) KEGG pathway enrichment results highlighting key cancer-related signaling pathways associated with the intersecting genes.\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-7400738/v1/90c59e9e554aae549d0bc1e7.png"},{"id":94846592,"identity":"88b465d6-f3d9-4778-ae2d-c47ef9c304eb","added_by":"auto","created_at":"2025-10-31 10:18:46","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":98399,"visible":true,"origin":"","legend":"\u003cp\u003ePPI network construction and machine learning–based screening of potential PP target genes. (A) PPI network based on the 50 intersecting genes between PP targets and breast cancer DEGs; (B) Cross-validation path plot of the LASSO regression model for λ parameter selection, with λ.min as the optimal value; (C) Relationship between the number of trees and model error in the Random Forest analysis; (D) Importance ranking of candidate genes in the Random Forest model, evaluated by the Mean Decrease Gini index; (E) SVM-RFE curve showing the relationship between the number of features and cross-validation error (RMSE), with the optimal feature set determined at the lowest error point; (F) Venn diagram of candidate genes identified by the three machine learning algorithms, highlighting shared core genes; (G) Expression levels of six core genes in control and treatment groups, with all genes showing significant differential expression (***p \u0026lt; 0.001). (H) ROC curves of the six core genes, all showing high diagnostic performance with AUC values above 0.90.\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-7400738/v1/76e9e5ebd124ff3426948034.png"},{"id":94846595,"identity":"3a1808f7-d2e8-4d84-81e8-5330cfc57f53","added_by":"auto","created_at":"2025-10-31 10:18:46","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":104682,"visible":true,"origin":"","legend":"\u003cp\u003eMendelian randomization analysis of the causal relationship between PIK3R1 expression and breast cancer risk. (A) Forest plot showing combined effect estimates from multiple MR methods, including IVW, MR-Egger, weighted median, weighted mode, and simple mode; (B) Scatter plot comparing effect estimates (β) and standard errors across MR methods, with slope consistency supporting robustness; (C) Regression plot showing each SNP's effect on the exposure (PIK3R1 expression) and the outcome (ER⁺ breast cancer), illustrating causal direction; (D) Forest plot of individual SNP estimates and their 95% confidence intervals; (E) Leave-one-out sensitivity analysis indicating the stability of MR estimates after sequential exclusion of individual SNPs\u003cstrong\u003e.\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-7400738/v1/09ccbdefcae141aa247201ed.png"},{"id":94846602,"identity":"5b38c7e7-ac0b-448c-8c38-3871f786a06c","added_by":"auto","created_at":"2025-10-31 10:18:46","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":144938,"visible":true,"origin":"","legend":"\u003cp\u003eDifferential expression, survival analysis, and molecular docking of PIK3R1 in breast cancer. (A) Box plot of PIK3R1 expression in breast tumor versus normal tissues based on TCGA data; (B) Kaplan–Meier survival curve from KM Plotter showing significantly improved overall survival in the high PIK3R1 expression group; (C) Molecular docking structure illustrating the 3D binding conformation and interaction sites between PP and PIK3R1 protein.\u003c/p\u003e","description":"","filename":"5.png","url":"https://assets-eu.researchsquare.com/files/rs-7400738/v1/7f65340ef7e4115e2797648d.png"},{"id":94846594,"identity":"baec234b-905b-4e58-9893-edf5433e978c","added_by":"auto","created_at":"2025-10-31 10:18:46","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":80313,"visible":true,"origin":"","legend":"\u003cp\u003eCorrelation of PIK3R1 expression with immune cell infiltration and comparison of immune cell abundance between high and low expression groups. (A) Differences in immune cell infiltration between PIK3R1 high- and low-expression groups based on ssGSEA. (B) Spearman correlation analysis showing significant associations between PIK3R1 expression and various immune cell types.\u003c/p\u003e","description":"","filename":"6.png","url":"https://assets-eu.researchsquare.com/files/rs-7400738/v1/c31cee906849e6067ef10cda.png"},{"id":95000948,"identity":"38928b32-960d-4eca-b00e-7058e70e4c44","added_by":"auto","created_at":"2025-11-03 09:00:57","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1437832,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7400738/v1/ff7058e7-9bcd-4632-ac04-8c107ab8445e.pdf"},{"id":94985628,"identity":"6ecc398a-2960-471b-9cd2-a339c763ab3a","added_by":"auto","created_at":"2025-11-03 06:58:31","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":14883,"visible":true,"origin":"","legend":"\u003cp\u003eSupplementary Table 1. ProTox-3.0 - Prediction of TOXicity of PP\u003c/p\u003e","description":"","filename":"supplementarytable1.docx","url":"https://assets-eu.researchsquare.com/files/rs-7400738/v1/e792343486ccd02226458059.docx"},{"id":94985433,"identity":"fbbf2ccb-62b1-4d98-badc-24d4fa2fc5ce","added_by":"auto","created_at":"2025-11-03 06:58:10","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":15424,"visible":true,"origin":"","legend":"\u003cp\u003eSupplementary Table 2.\u003cstrong\u003e \u003c/strong\u003eString Node Degrees\u003c/p\u003e","description":"","filename":"supplementarytable2.docx","url":"https://assets-eu.researchsquare.com/files/rs-7400738/v1/8096e41b3279568a524321ca.docx"},{"id":94985755,"identity":"3d9ac3b1-3f0d-44b6-bdc4-7768e04472b2","added_by":"auto","created_at":"2025-11-03 06:58:51","extension":"docx","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":19747,"visible":true,"origin":"","legend":"\u003cp\u003eSupplementary Table 3. Expression Quantitative Trait Locis of CA4 and PIK3R1\u003c/p\u003e","description":"","filename":"supplementarytable3.docx","url":"https://assets-eu.researchsquare.com/files/rs-7400738/v1/65b6bd855b4258a90de52be7.docx"},{"id":94984907,"identity":"6ae6b645-d53c-4553-a650-de0783fd4ed7","added_by":"auto","created_at":"2025-11-03 06:56:53","extension":"tif","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":95732,"visible":true,"origin":"","legend":"\u003cp\u003eSupplementary Figure 1. The integrated ROC curve of the six-gene model\u003c/p\u003e","description":"","filename":"SupplementaryFigure1.tif","url":"https://assets-eu.researchsquare.com/files/rs-7400738/v1/3c910c23135387b02181449e.tif"},{"id":94846614,"identity":"d6e45afb-292c-4e0a-9e46-e8b650d7de3e","added_by":"auto","created_at":"2025-10-31 10:18:46","extension":"tif","order_by":5,"title":"","display":"","copyAsset":false,"role":"supplement","size":273106,"visible":true,"origin":"","legend":"\u003cp\u003eSupplementary Figure 2. The forest plot about the MR results of CA4\u003c/p\u003e","description":"","filename":"SupplementaryFigure2.tif","url":"https://assets-eu.researchsquare.com/files/rs-7400738/v1/14fa5c23ecd78981b66ba219.tif"}],"financialInterests":"No competing interests reported.","formattedTitle":"Propylparaben Promotes Estrogen Receptor-Positive Breast Cancer Risk through PIK3R1: A Comprehensive Integrative Study","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eBreast cancer (BC) remains one of the most common and lethal malignancies among women worldwide, posing a substantial public health burden. According to the World Health Organization, approximately 2.3\u0026nbsp;million women worldwide were diagnosed with breast cancer in 2022, with an estimated 670,000 deaths attributed to the disease(\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e). The incidence of breast cancer exhibits significant gender and age differences, occurring predominantly in women over the age of 50. Notably, about 70% of BC cases are classified as estrogen receptor-positive (ER\u003csup\u003e+\u003c/sup\u003e) subtypes(\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e). As one of the three most prevalent cancers globally, BC prevention and control have garnered increasing attention(\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e). In addition to lifestyle-related risk factors such as alcohol consumption, obesity, and physical inactivity(\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e), environmental exposures-particularly to endocrine-disrupting chemicals (EDCs) have emerged as critical external contributors to BC risk(\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e). Previous studies have shown that exogenous EDCs, such as parabens and bisphenols, can disrupt estrogen signaling pathways, thereby promoting the initiation and progression of hormone receptor-related tumors(\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e, \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e). Elucidating the toxicological mechanisms of EDCs with potential breast carcinogenic activity has become a key focus in environmental toxicology research.\u003c/p\u003e\u003cp\u003ePropylparaben (PP) is one of the most commonly used paraben compounds, widely incorporated into everyday consumer products, particularly as an effective preservative in cosmetics(\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e). Due to its potent antimicrobial properties, PP is also extensively used in food preservation(\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e) and as a preservative in pharmaceutical injectable formulations(\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e). In recent years, its applications have expanded into industrial manufacturing and environmental remediation(\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e). PP can enter the human body through dermal absorption or dietary intake, where it undergoes hydrolysis by esterases into p-hydroxybenzoic acid and propanol. These metabolites are subsequently conjugated with glucuronic acid, sulfate, or glycine in hepatic microsomes to form water-soluble derivatives, which are then excreted via urine(\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e). Despite its widespread usage, the metabolism and toxicological safety of PP remain controversial. Multiple studies have suggested that PP may be a potential carcinogen, showing associations with cancers such as lung and prostate cancer(\u003cspan additionalcitationids=\"CR15\" citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e). However, evidence linking PP to BC remains limited. As an emerging EDC, PP exhibits a molecular structure like that of estrogen(\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e). According to a scientific opinion by the Scientific Committee on Consumer Safety of the European Union, both in vivo and in vitro studies have confirmed the estrogen-like activity of PP (\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e). In breast tissue, imbalances in estrogen and progesterone levels can lead to DNA damage accumulation, gene mutations, and the formation of malignant cells. Estrogen also promotes the proliferation of malignant cells and the formation of a tumor-supportive microenvironment, thereby facilitating cancer progression (\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eAlthough previous studies have suggested that PP, as an environmental estrogenic compound, may be involved in the initiation and progression of BC(\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e), its precise molecular mechanisms remain unclear due to limitations of conventional research methods(\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e). As a novel environmental contaminant, the association between PP and breast cancer, as well as the underlying molecular pathways, warrants further investigation. Traditional toxicological approaches often emphasize the direct interactions between a compound and its targets, overlooking the complex biological network through which diseases are induced. In this study, we employed a network toxicology framework, integrating bioinformatics, machine learning algorithms, Mendelian randomization (MR), and molecular docking, to systematically elucidate the molecular mechanisms by which PP may contribute to BC development. The findings are expected to provide theoretical insights and scientific reference for future research on environmentally induced tumorigenesis.\u003c/p\u003e"},{"header":"2. Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\u003ch2\u003e2.1 Toxicity Analysis and Target Screening of Propylparaben(PP)\u003c/h2\u003e\u003cp\u003eIn this study, the standardized molecular structure and corresponding SMILES (Simplified Molecular Input Line Entry System) notation of PP were obtained by querying the PubChem database (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://pubchem.ncbi.nlm.nih.gov/)(22)\u003c/span\u003e\u003cspan address=\"https://pubchem.ncbi.nlm.nih.gov/)(22)\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e using the keyword \"propyl p-hydroxybenzoate.\" Subsequently, PP was subjected to toxicity prediction using the ProTox 3.0 platform (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://tox.charite.de/)(23)\u003c/span\u003e\u003cspan address=\"https://tox.charite.de/)(23)\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. By integrating molecular structural features, toxicological target identification, and pharmacokinetic property evaluation, we identified potential toxicological pathways and target organs associated with PP. These results provide a theoretical basis for subsequent mechanistic investigations.\u003c/p\u003e\u003cp\u003eIn this study, we employed the SuperPred database to predict the potential targets of PP. SuperPred is based on machine learning models and can predict the ATC classification codes and potential targets of compounds by integrating known drug information(\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e). To complement and refine the target data, we additionally queried the STITCH (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://stitch.embl.de/)(25)\u003c/span\u003e\u003cspan address=\"http://stitch.embl.de/)(25)\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e and SwissTargetPrediction (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://www.swisstargetprediction.ch/)(26)\u003c/span\u003e\u003cspan address=\"http://www.swisstargetprediction.ch/)(26)\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e databases, and standardized the names of candidate targets to retrieve those not covered by SuperPred. Finally, we integrated the results to construct an initial gene library of PP targets, providing essential data support for the systematic investigation of PP\u0026rsquo;s potential carcinogenic mechanisms.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec4\" class=\"Section2\"\u003e\u003ch2\u003e2.2 Identification of Differentially Expressed Genes in Breast Cancer\u003c/h2\u003e\u003cp\u003eTo identify differentially expressed genes (DEGs), we extracted the gene expression matrix related to BC from The Cancer Genome Atlas (TCGA) database(\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e). During data preprocessing, we excluded recurrent cases and retained only primary solid tumor samples (n\u0026thinsp;=\u0026thinsp;1086) and their corresponding normal breast tissue samples (n\u0026thinsp;=\u0026thinsp;99) to ensure the representativeness and accuracy of the analysis results. After identifying DEGs and the toxicological targets of PP, we conducted Venn diagram analysis to identify overlapping genes. These intersecting genes were regarded as potential key targets involved in BC development associated with PP exposure, providing a foundation for subsequent functional enrichment analysis and network construction.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec5\" class=\"Section2\"\u003e\u003ch2\u003e2.3 Functional and Pathway Enrichment Analysis of Target Genes\u003c/h2\u003e\u003cp\u003eTo further explore the biological functions of potential gene targets involved in BC related to PP exposure, Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses were conducted using R software (version 4.4.0) GO analysis was performed across three categories: Biological Process (BP), Cellular Component (CC), and Molecular Function (MF)-to identify biological processes potentially associated with PP. KEGG enrichment analysis was used to determine pathways related to BC development associated with PP exposure, thereby elucidating its potential mechanism of action.\u003c/p\u003e\u003cp\u003eThe potential target genes associated with PP were then input into the STRING database(\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e), with the species set to \"Homo sapiens\" and a minimum interaction score of \u0026gt;\u0026thinsp;0.15 to ensure a comprehensive mapping of gene interactions. The resulting data were subsequently imported into Cytoscape software (version 3.10.3) for further analysis. Cytoscape is a widely used bioinformatics visualization tool capable of calculating topological attributes of network nodes and constructing protein-protein interaction (PPI) networks(\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e).We ranked nodes based on their degree values and enhanced the visualization through color gradients. The top 20 hub genes, ranked by degree, were identified as key targets for further investigation due to their high connectivity and potential regulatory significance within the network.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec6\" class=\"Section2\"\u003e\u003ch2\u003e2.4 Machine Learning-Based Selection of Core Genes\u003c/h2\u003e\u003cp\u003eTo enhance the robustness and accuracy of target gene identification, we applied three machine learning algorithms-Random Forest, Least Absolute Shrinkage and Selection Operator (LASSO) regression, and Support Vector Machine-Recursive Feature Elimination (SVM-RFE)-based on the expression matrix of hub genes derived from the TCGA dataset.\u003c/p\u003e\u003cp\u003eFeature importance was assessed using the Random Forest algorithm with ten-fold cross-validation(\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e). This approach was employed to identify genes with significant contributions to the classification model, even in the presence of nonlinear relationships. LASSO regression with a binomial distribution was applied(\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e), and the optimal regularization parameter (λ) was determined through ten-fold cross-validation. Genes with non-zero coefficients were retained to reduce overfitting and enhance interpretability. SVM-RFE with a radial basis function kernel was used to iteratively refine feature selection(\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e). Cross-validation error guided the elimination process, and informative genes were identified based on their contribution to predictive accuracy.\u003c/p\u003e\u003cp\u003eTo identify core targets, we conducted intersection analysis of the gene sets identified by each algorithm. The shared genes, identified via a Venn diagram, were regarded as core candidate targets involved in BC related to PP exposure, serving as a foundation for subsequent causal inference and mechanistic studies.\u003c/p\u003e\u003cp\u003eFinally, to evaluate the diagnostic performance of the identified core genes, Receiver Operating Characteristic (ROC) curve analysis was performed. The area under the curve (AUC) was calculated for each gene to assess its ability to discriminate between exposed and control samples. All selected genes exhibited high diagnostic accuracy, with AUC values exceeding 0.90, underscoring their potential as robust biomarkers for BC linked to PP.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec7\" class=\"Section2\"\u003e\u003ch2\u003e2.5 Mendelian Randomization Analysis\u003c/h2\u003e\u003cp\u003eTo further investigate the causal relationship between core target genes and BC, this study employed Mendelian Randomization (MR) to evaluate the potential causal effects of candidate gene expression levels on breast cancer development. Based on previously identified core genes, we extracted expression quantitative trait locis (eQTLs) significantly associated with gene expression from the IEU Open GWAS database (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://gwas.mrcieu.ac.uk/\u003c/span\u003e\u003cspan address=\"https://gwas.mrcieu.ac.uk/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) to serve as instrumental variables (IVs). To ensure the strength and independence of IVs, the following criteria were applied: (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e) statistically significant association with gene expression (P\u0026thinsp;\u0026lt;\u0026thinsp;1\u0026times;10⁻⁵); (\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e) linkage disequilibrium threshold set at r\u0026sup2; \u0026lt; 0.01 within a 10,000 kb window; and (\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e) exclusion of weak instruments, requiring an F-statistic\u0026thinsp;\u0026gt;\u0026thinsp;10.\u003c/p\u003e\u003cp\u003eGiven that PP is widely recognized as an endocrine-disrupting chemical, primarily acting via ER, we selected ER\u003csup\u003e+\u003c/sup\u003e breast cancer as the outcome phenotype. The outcome data were obtained from the IEU Open GWAS database (dataset \u003cb\u003eieu-a-1132\u003c/b\u003e, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://opengwas.io/datasets/ieu-a-1132\u003c/span\u003e\u003cspan address=\"https://opengwas.io/datasets/ieu-a-1132\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e), which is based on a study by Michailidou K et al. (2017, \u003cem\u003eNature\u003c/em\u003e, n\u0026thinsp;=\u0026thinsp;83,691 European female participants, BCAC) (\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e) and comprises 38,197 breast cancer cases and 45,494 healthy controls.\u003c/p\u003e\u003cp\u003eMR analysis was primarily conducted using the inverse-variance weighted (IVW) method, with additional validation using MR-Egger regression, weighted median, simple mode, and weighted mode approaches to enhance the robustness of causal inference. To further evaluate the reliability and robustness of the findings, sensitivity analyses were performed, including: (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e) Cochran\u0026rsquo;s Q test to assess heterogeneity among IVs; (\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e) MR-Egger intercept test to detect potential horizontal pleiotropy; and (\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e) leave-one-out analysis to exclude the influence of any single SNP dominating the causal effect.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e\u003ch2\u003e2.6 Differential Expression and Prognostic Analysis\u003c/h2\u003e\u003cp\u003eTo evaluate the differential expression of the key target gene in BC tissues, we extracted the expression data of PIK3R1 from the transcriptomic dataset of breast cancer patients in TCGA database. Expression levels were compared between groups based on corresponding clinicopathological information. During data preprocessing, we selected primary BC samples with complete clinical annotation (Tumor\u0026thinsp;=\u0026thinsp;1086) and normal breast tissue samples (Normal\u0026thinsp;=\u0026thinsp;99).\u003c/p\u003e\u003cp\u003eTo further explore the potential prognostic value of this gene in BC, we conducted survival analysis using the KM Plotter online platform (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://kmplot.com/analysis/\u003c/span\u003e\u003cspan address=\"https://kmplot.com/analysis/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e), which integrates clinical and gene expression data from multiple public BC cohorts(\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e). A total of breast cancer samples were included in the analysis (n\u0026thinsp;=\u0026thinsp;4929), with low expression group (n\u0026thinsp;=\u0026thinsp;2467) and high expression group (n\u0026thinsp;=\u0026thinsp;2462). The association between gene expression and patient survival outcomes was quantified using odds ratios (ORs) and 95% confidence intervals (CIs). Finally, survival results were visualized to assess the potential of this gene as a prognostic biomarker in BC.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec9\" class=\"Section2\"\u003e\u003ch2\u003e2.7 Molecular Docking Analysis\u003c/h2\u003e\u003cp\u003eMolecular docking predicts the binding mode and affinity between small molecules and target proteins by identifying the lowest-energy conformation. To explore the potential carcinogenic mechanism of PP, docking was performed with key BC target proteins.\u003c/p\u003e\u003cp\u003eFirst, the 3D structure of PP was downloaded from the PubChem Compound Database (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://pubchem.ncbi.nlm.nih.gov/\u003c/span\u003e\u003cspan address=\"https://pubchem.ncbi.nlm.nih.gov/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e), while the crystal structures of key target proteins were obtained from UniProt (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.uniprot.org/)(35)\u003c/span\u003e\u003cspan address=\"https://www.uniprot.org/)(35)\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e and the Protein Data Bank (PDB, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.rcsb.org/)(36)\u003c/span\u003e\u003cspan address=\"https://www.rcsb.org/)(36)\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Protein structures were preprocessed using PyMOL software(\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e) by removing water molecules and original ligands, and saved in PDB format. Next, ligand preparation was conducted using AutoDock Tools 1.5.7(\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e), followed by saving the ligand in PDBQT format. The protein structures were modified by adding hydrogen atoms and defined as receptors, then also saved in PDBQT format. The processed ligand and receptor were imported into AutoDock, and a docking grid box covering the entire protein surface was constructed. Docking was performed using AutoDock Vina(\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e) with the number of docking poses set to 50. The docking results were ranked based on binding energy, and the protein-ligand complex with the lowest energy was selected for further analysis. The docked complex was uploaded to the PLIP Web Tool (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://plip-tool.biotec.tu-dresden.de/plip-web/plip/index)(38)\u003c/span\u003e\u003cspan address=\"https://plip-tool.biotec.tu-dresden.de/plip-web/plip/index)(38)\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e to identify key interaction sites, such as hydrogen bonds and hydrophobic interactions and binding conformations were visualized via PyMOL and Discovery Studio for 3D and 2D interaction mapping, respectively.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec10\" class=\"Section2\"\u003e\u003ch2\u003e\u003cb\u003e2.8 Immune infiltration\u003c/b\u003e\u003c/h2\u003e\u003cp\u003eTo evaluate the potential role of the key target in the tumor immune microenvironment, single-sample Gene Set Enrichment Analysis (ssGSEA) was performed to quantify the relative abundance of various immune cell populations(\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e). Cell-type-specific gene sets were obtained from the CellMarker database and formatted into GMT files for analysis. ssGSEA was applied to the normalized expression matrix to generate enrichment scores for each sample from TCGA. To further investigate the relationship between the key target gene and immune infiltration, samples were divided into high and low expression groups based on the median expression level of the gene in tumor tissues. Subsequently, the Wilcoxon rank-sum test was applied to compare differences in immune cell infiltration between the two groups, identifying immune cell types with significant variation under different expression statuses. Additionally, Spearman correlation analysis was performed to assess the associations between the key gene expression and ssGSEA scores of various immune cells, thereby revealing its potential role in modulating the tumor immune microenvironment.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e\u003ch2\u003e2.9 statistical analysis\u003c/h2\u003e\u003cp\u003eIn this study, data analysis was conducted using R software (version 4.4.0). Differential expression analysis was performed using the \"limma\" package in R software (version 4.4.0), with the filtering criteria set to p-value\u0026thinsp;\u0026lt;\u0026thinsp;0.05 and |log2FC| \u0026ge; 0.585, in order to obtain a comprehensive set of DEGs. GO and KEGG enrichment analyses were conducted using the \u0026ldquo;clusterProfiler\u0026rdquo;, \u0026ldquo;org.Hs.eg.db\u0026rdquo;, \u0026ldquo;enrichplot\u0026rdquo;, and \u0026ldquo;ggplot2\u0026rdquo; packages, with a significance cutoff of p\u0026thinsp;\u0026lt;\u0026thinsp;0.05 for enriched term selection. To identify key genes, multiple machine learning algorithms were applied for feature selection. The LASSO regression model was constructed using the \u0026ldquo;glmnet\u0026rdquo; package, the random forest model with the \u0026ldquo;randomForest\u0026rdquo; package, and the SVM-RFE model with the \u0026ldquo;e1071\u0026rdquo;, \u0026ldquo;kernlab\u0026rdquo;, and \u0026ldquo;caret\u0026rdquo; packages. The predictive performance of all models was assessed by plotting ROC curves and calculating the AUC using the \u0026ldquo;pROC\u0026rdquo; package. For MR analysis, we sequentially employed the \u0026ldquo;VariantAnnotation\u0026rdquo;, \u0026ldquo;gwasglue\u0026rdquo;, and \u0026ldquo;TwoSampleMR\u0026rdquo; packages. Results were reported as odds ratios (ORs) with 95% confidence intervals (CIs), and statistical significance was defined as p\u0026thinsp;\u0026lt;\u0026thinsp;0.05. Immune infiltration analysis was performed using the \"GSVA\" package for ssGSEA scoring, with immune cell annotations converted to GMT format via the \"GSEABase\" package. Wilcoxon rank-sum test compared immune cell abundance between groups. Data preprocessing and correlation analysis were done using \"dplyr\" and Spearman correlation assessed the relationship between gene expression and immune cell scores. Additionally, visualization tools included \"ggvenn\" for Venn diagrams, \"ggplot2\" and \"ggrepel\" for volcano plots and correlation plots, \"pheatmap\" for heatmaps, \"grid\", \"readr\" and \"forestploter\" for forest plots, and \"ggpubr\" for boxplots.\u003c/p\u003e\u003c/div\u003e"},{"header":"3. Results","content":"\u003cdiv id=\"Sec13\" class=\"Section2\"\u003e\u003ch2\u003e3.1 Identification of Toxic Target Genes Associated with PP-Induced Breast Cancer\u003c/h2\u003e\u003cp\u003eIn this study, the standard two-dimensional (2D) and three-dimensional (3D) molecular structures of PP were first retrieved from the PubChem database (Figs.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA-B). Subsequently, ProTox 3.0 platform was used to predict the toxicological features of PP (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eC), including acute toxicity, toxicity classification, and structural similarity. Additionally, potential toxicological signaling pathways associated with PP were explored (Supplementary Table\u0026nbsp;1). Radar plot visualization (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eD) revealed that PP exhibits multiple biological activities, particularly demonstrating a high activity probability in the ER signaling pathway, suggesting that PP may exert estrogen-like effects and potentially promote BC development through ER pathway activation. To comprehensively identify the potential molecular targets of PP, we integrated predictions from three databases-SuperPred, STITCH, and Swiss Target Prediction, resulting in a total of 144 candidate target genes associated with PP (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eE). Next, differential expression analysis was performed using transcriptomic data from BC patients in the TCGA database. A volcano plot was generated to illustrate the overall distribution of DEGs (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA), identifying a total of 4,788 significant DEGs. By intersecting the 144 predicted PP target genes with the DEGs, 50 overlapping genes were identified (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eB). These genes were preliminarily considered potential key regulatory targets through which PP may influence BC progression.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eFurthermore, a heatmap was constructed to display the expression patterns of these 50 genes in BC tissues versus normal tissues (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eC), revealing distinct differential expression characteristics. Additionally, a compound\u0026ndash;target interaction network was constructed to map the interactions between PP and its potential target genes (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eD), providing a visual representation of how PP may exert its carcinogenic effects by modulating multiple key targets.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec14\" class=\"Section2\"\u003e\u003ch2\u003e3.2 GO Functional Annotation and KEGG Pathway Enrichment Analysis of Target Genes\u003c/h2\u003e\u003cp\u003eTo further investigate the biological functions of the potential PP target genes in BC, GO functional annotation and KEGG pathway enrichment analyses were performed on the 50 intersecting genes. The analyses were conducted using the R, with the species restricted to \u003cem\u003eHomo sapiens\u003c/em\u003e, and included both GO and KEGG enrichment assessments.\u003c/p\u003e\u003cp\u003eGO analysis was performed across three dimensions: BP, CC, and MF. The top ten most significantly enriched terms from each category were visualized (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eE). In the BP category, rhythmic process was significantly enriched. Disruption of circadian rhythms shapes the tumor microenvironment, promotes epithelial-mesenchymal transition (EMT), and thereby influences the progression and metastasis of BC(\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e). Closely associated with normal mammary gland development, its dysregulation is considered a hallmark of early pathological changes in BC(\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e). In the MF category, the regulation of nuclear steroid receptor activity pathway was significantly enriched, suggesting that PP may influence typical nuclear steroid receptors such as estrogen receptors (ERα/ERβ), thereby contributing to the development of ER\u003csup\u003e+\u003c/sup\u003e BC(\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e). Moreover, the enrichment of estrogen response element (ERE) binding activity further indicates that these genes may participate in the regulation of ER binding to promoter regions of target genes, playing a critical role in the progression of BC(\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eFurther KEGG pathway enrichment analysis (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eF) revealed that these target genes are significant enriched in multiple cancer-related signaling pathways, with the proteoglycans in cancer being the most prominent. Proteoglycans are complexes composed of a core protein and glycosaminoglycans (GAGs), broadly distributed on the cell membrane and within the extracellular matrix (ECM). They play a key role in regulating cell adhesion, migration, and interactions between ECM components and cells. The genetic signatures of proteoglycans have emerged as important biomarkers for the diagnosis, prognosis, and targeted therapy of BC(\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e). The estrogen signaling pathway is a central driving mechanism in ER⁺ BC(\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e). These findings further support the hypothesis that PP, as an estrogen-like compound, may promote breast cancer development by influencing proteoglycan expression and modulating estrogen signaling. Additionally, these genes were also significantly enriched in the MAPK signaling pathway, lipid and atherosclerosis, and other key signaling cascades that are closely involved in regulating cell proliferation, apoptosis, and the tumor microenvironment(\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec15\" class=\"Section2\"\u003e\u003ch2\u003e3.3 Construction of the PPI Protein\u0026ndash;Protein Interaction Network\u003c/h2\u003e\u003cp\u003eTo further explore the potential molecular mechanisms linking PP and BC, we constructed a PPI network based on the 50 intersecting genes identified earlier. The network was generated using the STRING database and visualized with Cytoscape software for topological analysis. In the PPI network, each node represents a candidate target gene, and edges indicate known or predicted protein\u0026ndash;protein interactions. We performed topological analysis by ranking nodes based on their degree values (i.e., the number of direct interactions with other proteins). The resulting network exhibited a clear hub structure, with several genes displaying high connectivity, suggesting their potential roles in mediating PP-related signaling pathways in BC (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eA). To identify biologically relevant hub regulators, we selected the top 20 genes ranked by degree as potential hub genes (Supplementary Table\u0026nbsp;2). These hub genes exhibited high connectivity in the PPI network and are likely to play pivotal roles in PP-induced breast cancer progression, providing a theoretical foundation for subsequent machine learning-based gene prioritization and survival analysis.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec16\" class=\"Section2\"\u003e\u003ch2\u003e3.4 Identification of Core Genes via Machine Learning Algorithms\u003c/h2\u003e\u003cp\u003eTo further identify key regulatory target genes potentially involved in PP-mediated BC development, we applied three commonly used machine learning algorithms-LASSO regression, Random Forest and SVM-RFE to the top 20 hub genes ranked by degree in the previously constructed PPI network.\u003c/p\u003e\u003cp\u003eIn the LASSO regression analysis, the model exhibited minimal fluctuation around λ\u0026thinsp;=\u0026thinsp;λ.min. Based on non-zero regression coefficients, 13 candidate genes were selected for further analysis (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eB). A Random Forest classification model was developed, and model performance was evaluated by plotting the error convergence curve as the number of trees increased (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eC). When the number of trees reached 500, the model achieved optimal convergence, balancing accuracy and computational efficiency. Genes with importance scores (Mean Decrease Gini) greater than 5 were retained, resulting in 8 candidate genes (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eD). SVM-RFE analysis identified 8 optimal genes with the lowest root mean squared error, as shown in the feature selection curve. These top-ranked genes were retained based on feature importance (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eE).\u003c/p\u003e\u003cp\u003eTo integrate the results from the three algorithms, an intersection analysis of their respective candidate gene sets was performed, visualized using a Venn diagram (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eF). A total of 6 genes were identified as common to all three methods and were preliminarily considered as the most promising regulatory targets in PP-induced BC development. To further characterize these 6 genes, boxplots were generated to display their expression patterns among different groups (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eG).\u003c/p\u003e\u003cp\u003eTo assess their diagnostic performance, ROC curves were constructed and calculated the AUC was calculated for each gene (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eH). Furthermore, to evaluate the overall diagnostic performance and robustness of the combined model based on the 6 candidate genes, an integrated ROC curve was constructed. The model demonstrated excellent discriminative ability (\u003cem\u003eAUC\u0026thinsp;=\u0026thinsp;0.996,95%CI:0.992\u0026ndash;0.999\u003c/em\u003e) (Supplementary Fig.\u0026nbsp;1). All 6 genes demonstrated strong diagnostic power for distinguishing between the disease and control groups, with AUC values of 0.988 for \u003cem\u003eCA4\u003c/em\u003e, 0.975 for \u003cem\u003eCDK1\u003c/em\u003e, 0.946 for \u003cem\u003eCDK5\u003c/em\u003e, 0.948 for \u003cem\u003eMADA\u003c/em\u003e, 0.965 for \u003cem\u003eNR3C1\u003c/em\u003e, and 0.904 for \u003cem\u003ePIK3R1.\u003c/em\u003e These results suggest that the identified genes may serve as promising biomarkers for BC diagnosis related to PP exposure and warrant further functional and clinical validation. Subsequent studies will focus on these genes for mechanistic investigation and functional validation.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec17\" class=\"Section2\"\u003e\u003ch2\u003e3.5 Mendelian Randomization Analysis\u003c/h2\u003e\u003cp\u003eTo validate the causal relationships between PP and key BC target genes, we performed MR analysis. Expression quantitative trait loci (eQTLs) for the candidate genes were retrieved from the IEU Open GWAS database, and suitable instrumental variables (IVs) were selected for MR based on the following criteria: P-value\u0026thinsp;\u0026lt;\u0026thinsp;1\u0026times;10⁻⁵; Linkage disequilibrium r\u0026sup2; \u0026lt; 0.01; Window size\u0026thinsp;=\u0026thinsp;10,000 kb; F-statistic\u0026thinsp;\u0026gt;\u0026thinsp;10. Using these standards, valid eQTLs were identified for two genes: PIK3R1 and CA4. The remaining four genes were excluded from MR analysis due to the lack of qualified SNPs (Supplementary Table\u0026nbsp;3).\u003c/p\u003e\u003cp\u003eBy integrating the selected eQTLs with the GWAS dataset for ER⁺ BC (ieu-a-1127), MR analysis was performed using the Inverse Variance Weighted (IVW) method as the primary approach. The results demonstrated a significant causal association between decreased PIK3R1 expression and increased BC risk (\u003cem\u003eOR\u003c/em\u003e\u003csub\u003e\u003cem\u003eIVW\u003c/em\u003e\u003c/sub\u003e \u003cem\u003e= 0.869, 95% CI: 0.774\u0026ndash;0.975, P\u0026thinsp;=\u0026thinsp;0.017\u003c/em\u003e) (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA). Collectively, these results suggest that PIK3R1 may exert a protective role in the development of PP-induced ER⁺ BC, and that reduced expression of PIK3R1 may significantly elevate the risk of disease onset. The Mendelian randomization analysis did not show a statistically significant causal association for CA4 (Supplementary Fig.\u0026nbsp;2).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eTo assess the robustness of the MR analysis, sensitivity analyses were performed. Cochran\u0026rsquo;s Q test revealed no significant heterogeneity in the IVW model (\u003cem\u003eQ\u0026thinsp;=\u0026thinsp;7.15, P\u0026thinsp;=\u0026thinsp;0.13\u003c/em\u003e), and residual heterogeneity in the MR-Egger model was also non-significant (\u003cem\u003eQ\u0026thinsp;=\u0026thinsp;3.35, P\u0026thinsp;=\u0026thinsp;0.34\u003c/em\u003e), indicating good model stability. The funnel plot showed a relatively symmetrical distribution of SNP effects, indicating no substantial directional bias and suggesting an absence of publication or selection bias (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eB). In addition, the intercept test from MR-Egger regression yielded a P-value of 0.16, suggesting no evidence of horizontal pleiotropy. In addition, we plotted a funnel plot. In the scatter plot (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eC), the slopes of fitted lines from multiple MR methods were largely consistent, indicating that the estimated causal effect of PIK3R1 expression on ER⁺ BC risk was directionally robust across different analytical approaches. Sensitivity analyses were also performed, the forest plot (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eD) displayed individual causal estimates for each SNP and their contribution to the overall MR result, allowing visual assessment of effect direction and confidence intervals. Leave-one-out sensitivity analysis (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eE) revealed that exclusion of any single SNP did not substantially alter the overall MR estimate, supporting the robustness of the results despite minor heterogeneity potentially introduced by specific variants. These results support the validity of the instrumental variables and enhance confidence in the causal inference.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec18\" class=\"Section2\"\u003e\u003ch2\u003e3.6 Differential Expression and Prognostic Analysis of PIK3R1 in Breast Cancer\u003c/h2\u003e\u003cp\u003eBased on RNA-seq data from the TCGA database, including 1086 BC tissues and 99 adjacent normal tissues, the expression difference of PIK3R1 was evaluated using the Mann-Whitney U test. The results showed that PIK3R1 expression was significantly lower in BC tissues compared to normal tissues (\u003cem\u003eP\u0026thinsp;\u0026lt;\u0026thinsp;0.05\u003c/em\u003e), suggesting possible downregulation in tumors (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA). To further explore the clinical relevance of PIK3R1 expression in breast cancer prognosis, the KM Plotter online platform (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://kmplot.com/\u003c/span\u003e\u003cspan address=\"https://kmplot.com/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) was used to analyze the association between PIK3R1 expression levels and OS. A total of 4929 BC cases were included, divided into a high-expression group (n\u0026thinsp;=\u0026thinsp;2462) and a low-expression group (n\u0026thinsp;=\u0026thinsp;2467) based on median expression. Kaplan\u0026ndash;Meier survival analysis indicated that patients in the high-expression group had significantly better OS compared to the low-expression group (\u003cem\u003eOR\u0026thinsp;=\u0026thinsp;0.70, 95% CI: 0.63\u0026ndash;0.78\u003c/em\u003e) (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eB). These findings suggest that PIK3R1 may function as a protective factor in BC, and its high expression is closely associated with improved survival outcomes.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec19\" class=\"Section2\"\u003e\u003ch2\u003e3.7 Molecular Docking Analysis of PIK3R1\u003c/h2\u003e\u003cp\u003eTo further evaluate the potential interaction between PP and the key BC target protein PIK3R1, molecular docking analysis was performed. The docking results showed that the lowest binding free energy between PP and PIK3R1 was \u0026minus;\u0026thinsp;6.37 kcal/mol, indicating a strong binding affinity (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eC). Subsequent interaction analysis using the PLIP Web Tool revealed that PP stably binds within the active pocket of PIK3R1 through multiple non-covalent interactions. Specifically, PP formed several stable hydrogen bonds with residues GLU650, GLN654, and ASN630, contributing to the complex\u0026rsquo;s stability. In addition, hydrophobic interactions were observed with SER628, SER629, and ARG649, and a π\u0026ndash;π stacking interaction occurred between PP and SER651, further enhancing the binding strength. Auxiliary polar interactions were also noted with residues ARG631 and ASN632. Three-dimensional structural visualization showed that PP is stably embedded in the hydrophobic pocket of PIK3R1, with a clear binding conformation and a dense distribution of surrounding residues, providing a structural basis for potential functional modulation. Collectively, these findings support the hypothesis that PP may exert pathogenic effects in BC by directly binding to and regulating the activity of PIK3R1.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec20\" class=\"Section2\"\u003e\u003ch2\u003e\u003cb\u003e3.8 Association of PIK3R1 Expression with Immune Infiltration and Tumor Microenvironment\u003c/b\u003e\u003c/h2\u003e\u003cp\u003eBased on the single-sample Gene Set Enrichment Analysis (ssGSEA), comparison between groups revealed that multiple immune cell infiltrations were significantly increased in the PIK3R1 high-expression group, including activated B cells, effector memory CD8⁺ T cells, eosinophils, immature B cells, mast cells, natural killer cells, neutrophils, regulatory T cells, follicular helper T cells, Th1 cells, and Th2 cells. In contrast, activated CD8⁺ T cells, CD56 dim natural killer cells, and monocytes showed higher infiltration levels in the low-expression group (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eA). Spearman correlation analysis further revealed significant associations between PIK3R1 expression and various immune cells (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eB). Specifically, PIK3R1 expression was positively correlated with mast cells (ρ\u0026thinsp;=\u0026thinsp;0.29), plasmacytoid dendritic cells (ρ\u0026thinsp;=\u0026thinsp;0.42), memory B cells (ρ\u0026thinsp;=\u0026thinsp;0.21), regulatory T cells (ρ\u0026thinsp;=\u0026thinsp;0.17), follicular helper T cells (ρ\u0026thinsp;=\u0026thinsp;0.24), Th1 cells (ρ\u0026thinsp;=\u0026thinsp;0.22), and Th2 cells (ρ\u0026thinsp;=\u0026thinsp;0.17), while showing negative correlations with CD56 dim natural killer cell (ρ = \u0026minus;0.20), activated CD8⁺ T cells (ρ = \u0026minus;0.16), and monocytes (ρ = \u0026minus;0.30) (all p\u0026thinsp;\u0026lt;\u0026thinsp;0.001).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eNotably, regulatory T cells, Th1 cells, Th2 cells, and mast cells exhibited consistent trends in both the differential and correlation analyses, with higher infiltration levels observed in the PIK3R1 high-expression group. These findings suggest that PIK3R1 may play a protective role in maintaining immune microenvironment homeostasis by promoting the enrichment of antitumor immune-related cells.\u003c/p\u003e\u003c/div\u003e"},{"header":"4. Discussion","content":"\u003cp\u003eAs a potential emerging environmental contaminant, PP has not been systematically investigated for its association with BC, particularly ER⁺ subtypes. In this study, we employed a comprehensive approach integrating network toxicology, bioinformatics, machine learning, Mendelian randomization, and molecular docking to systematically elucidate the potential molecular mechanisms by which PP may contribute to breast cancer development. Fiftyputative PP target genes were identified through toxicity prediction and integration with TCGA datasets. Prioritized through GO/KEGG enrichment analysis and PPI network construction, six core genes were selected by LASSO regression, Random Forest, and SVM-RFE. ROC analysis showed excellent diagnostic performance of the core genes (AUC\u0026thinsp;\u0026gt;\u0026thinsp;0.90). PIK3R1 was further confirmed as a causal risk gene for breast cancer using MR analysis. At the experimental level, differential expression, survival analysis, and molecular docking demonstrated that PP downregulates PIK3R1 expression and promotes BC cell proliferation. Immune infiltration analysis revealed PIK3R1\u0026rsquo;s close association with antitumor immunity. These findings consistently support PIK3R1 as a key molecular target mediating BC related to PP progression, uncovering a novel mechanism of environmental exposure in ER⁺ breast cancer. Parabens, including PP, are widely used as preservatives in daily consumer products. Humans are exposed to these compounds through various routes, such as the use of cosmetics, consumption of food and pharmaceuticals, and contact with industrial goods. Previous studies have suggested that parabens may possess carcinogenic potential(\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e). An epidemiological study comparing normal and malignant breast tissues in patients with BC revealed a significant association between paraben accumulation and BC risk, particularly in ER\u003csup\u003e+\u003c/sup\u003e and PR\u003csup\u003e+\u003c/sup\u003e cases(\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e). Among parabens, PP is the most used and has attracted extensive attention. PP has been reported to adversely affect multiple human physiological systems, posing broad health risks. For instance, PP may interfere with neurodevelopment in children, increasing the risk of attention deficit hyperactivity disorder, with notable sex-specific differences in effect(\u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e). Additionally, PP exposure has been associated with elevated levels of xanthine, hypoxanthine, uric acid, and acylcarnitines, indicating disruptions in purine metabolism and fatty acid β-oxidation, which may contribute to energy metabolic disorders and a heightened risk of diabetes(\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e). PP has also been implicated in reproductive toxicity, with evidence suggesting reduced fertility following exposure(\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e). As an EDC, PP structurally resembles estrogen and exhibits estrogenic activity, potentially playing a crucial role in the initiation and progression of BC(\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e). Our findings support the notion that PP exposure increases breast cancer risk. To further explore the underlying mechanisms, GO and KEGG enrichment analyses were conducted on 50 potential BCtarget genes. These genes were significantly enriched in the estrogen signaling pathway. The estrogen signaling pathway is closely linked to breast carcinogenesis. Its oncogenic mechanisms include the interaction of 17β-estradiol with estrogen receptors, the genotoxic effects of estrogen metabolites, and epigenetic modifications. As a xenoestrogen, PP can activate this pathway, thereby trigger carcinogenic processes and disrupt normal breast tissue growth, proliferation, and differentiation, ultimately contributing to malignancy development(\u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e52\u003c/span\u003e). Moreover, the enrichment analysis suggested that biological rhythms, and their disruption influences tumor angiogenesis and the immune microenvironment, thereby increasing the risk of BC(\u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e53\u003c/span\u003e). PP may also influence cancer invasiveness via proteoglycan pathways. Core proteoglycans have been shown to inhibit tumor growth and invasion in inflammatory breast cancer by destabilizing E-cadherin and suppressing EGFR/ERK signaling(\u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e54\u003c/span\u003e). Given these findings, it is necessary to adopt precautionary measures, such as reducing the use of PP in preservative systems and implementing more comprehensive and standardized assessments of its application. Furthermore, increasing public awareness of the potential health risks associated with preservative ingredients is essential to minimize adverse effects caused by exposure to PP on human health.\u003c/p\u003e\u003cp\u003eTo further identify the key pathways involved in PP-induced BC, we integrated multiple public databases with TCGA data to screen for potential molecular targets. By combining PPI analysis, machine learning algorithms, and MR, we progressively narrowed down key candidate genes. Multilevel validation, including prognostic analysis, and molecular docking, consistently highlighted PIK3R1 as the core gene mediating PP-induced breast carcinogenesis. The PIK3R1 gene, located on chromosome 5q13.1, encodes the regulatory subunit p85α of class I PI3K, a well-recognized tumor suppressor. Its mutation, amplification, or deletion has been widely reported as a molecular hallmark of metastatic BC(\u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e55\u003c/span\u003e). Together with the p110 catalytic subunit encoded by PIK3CA, p85α forms the PI3K complex, which plays a central role in estrogen signaling(\u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e56\u003c/span\u003e). While PIK3CA mutations are typically early events in breast tumorigenesis, PIK3R1 dysregulation is more often associated with tumor invasion(\u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e57\u003c/span\u003e). Abnormalities in either gene can lead to sustained activation of the PI3K/Akt signaling pathway, thereby promoting cell proliferation, angiogenesis, and apoptosis resistance. The PI3K/Akt pathway has been implicated in various malignancies, including breast, lung, and cervical cancers(\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e, \u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e58\u003c/span\u003e). In ER⁺ BC, estrogen or growth factors bind to tyrosine kinase receptors or G protein\u0026ndash;coupled estrogen receptors, initiating a cascade that activates PI3K. Activated PI3K catalyzes the generation of PIP3, which recruits and activates Akt, triggering downstream phosphorylation events that regulate cell proliferation, survival, and motility(\u003cspan citationid=\"CR59\" class=\"CitationRef\"\u003e59\u003c/span\u003e). Previous studies have shown that targeting PIK3R1 can inhibit PI3K/Akt pathway activation and reverse epithelial\u0026ndash;mesenchymal transition, a process critical for cancer invasion, metastasis, and stemness acquisition(\u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e60\u003c/span\u003e, \u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e61\u003c/span\u003e). Under physiological conditions, p85α binds tightly to p110 and suppresses its catalytic activity(\u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e62\u003c/span\u003e). However, under PP exposure, downregulation of PIK3R1 may lead to reduced levels of free p85α, relieving inhibition on p110, and consequently enhancing PI3K/Akt signaling(\u003cspan citationid=\"CR63\" class=\"CitationRef\"\u003e63\u003c/span\u003e). This aberrant activation may promote tumor cell growth, survival, and angiogenesis, ultimately accelerating breast cancer progression.\u003c/p\u003e\u003cp\u003eDespite the comprehensive integration of multi-omics data and the application of advanced analytical methods to elucidate the molecular mechanisms underlying breast cancer related to PP exposure, this study has several limitations. First, although network toxicology analyses identified multiple candidate targets and signaling pathways associated with PP, the subsequent analyses primarily focused on PIK3R1, which may have led to the oversight of other potentially important regulatory genes or pathways. Second, although molecular docking and in vitro assays preliminarily validated the interaction between PP and PIK3R1, the lack of in vivo functional validation limits the depth and completeness of mechanistic insight. Future research should incorporate animal models, multidimensional functional assays, and large-scale, multi-ethnic, and multicenter cohort data to strengthen the causal evidence and enhance the generalizability and robustness of the conclusions. Nevertheless, this study presents several important strengths. First, it is methodologically innovative and systematically designed, integrating diverse approaches, including network toxicology, transcriptomic analysis, protein\u0026ndash;protein interaction network construction, machine learning\u0026ndash;based gene prioritization, Mendelian randomization, molecular docking, and cellular validation-into a coherent research framework. This comprehensive strategy greatly enhances the reliability and scientific rigor of the findings. Second, beyond computational identification of candidate genes, the study incorporates clinical prognostic analysis and experimental validation, ensuring the biological and translational relevance of the results. Notably, this is the first study to clearly identify PIK3R1 as a core mediator of PP-induced breast carcinogenesis, expanding the theoretical understanding of environmental carcinogenesis and offering a novel molecular target for future preventive and therapeutic strategies.\u003c/p\u003e\u003cp\u003eIn conclusion, this study is the first to systematically reveal that exposure to PP may promote the initiation and progression of BC by downregulating PIK3R1 expression and activating the PI3K/Akt signaling pathway. These findings not only deepen our understanding of the molecular carcinogenic mechanisms of environmental chemicals, but also highlight PIK3R1 as a key tumor suppressor involved in PP-induced breast carcinogenesis. This provides a theoretical foundation for future risk assessment and the development of targeted prevention and therapeutic strategies for environmentally induced cancers.\u003c/p\u003e"},{"header":"5. Conclusion","content":"\u003cp\u003eThis study systematically elucidates the potential molecular mechanism by which PP exposure promotes BC initiation and progression via downregulation of PIK3R1 and activation of the PI3K/Akt signaling pathway. By integrating multiple approaches\u0026mdash;including network toxicology, transcriptomic differential expression analysis, PPI network construction, machine learning\u0026ndash;based gene prioritization, Mendelian randomization, molecular docking, and cellular validation-PIK3R1 was confirmed as a key regulatory gene in breast carcinogenesis related to PP exposure. These findings not only expand our molecular understanding of the carcinogenic mechanisms of environmental chemicals but also highlight the potential health risks posed by PP as an EDC. Moreover, they underscore the biological significance of PIK3R1 as a putative tumor suppressor. Collectively, this study provides a theoretical basis for the development of targeted prevention and therapeutic strategies and offers scientific insight for future research on environmentally induced cancers and the risk assessment and control of environmental carcinogens.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that they have no competing interests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor contributors\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eKaiwen Wang\u003c/strong\u003e: Conceptualization, Methodology, Writing-original draft, Writing-review \u0026amp; editing;\u0026nbsp;\u003cstrong\u003eGuang Yao\u003c/strong\u003e:\u0026nbsp;Formal Analysis, Resources, Writing-original draft;\u0026nbsp;\u003cstrong\u003eXiaobin Zhang\u003c/strong\u003e: Investigation, Software, Visualization,\u0026nbsp;Writing-original draft; \u003cstrong\u003eRuiting Ma\u003c/strong\u003e: Visualization, Writing-original draft; \u003cstrong\u003eXinyi Li\u003c/strong\u003e: Methodology, Writing-review \u0026amp; editing; \u003cstrong\u003eWeize Kong\u003c/strong\u003e: Supervision, Validation;\u0026nbsp;\u003cstrong\u003eXu Hui\u003c/strong\u003e: \u0026nbsp; Data Curation, Software; \u003cstrong\u003eQian Liu\u003c/strong\u003e: Formal Analysis, Investigation; \u003cstrong\u003eYanan Wu\u003c/strong\u003e: Software,\u0026nbsp;Visualization;\u0026nbsp;\u003cstrong\u003eYi Xiao\u003c/strong\u003e:\u0026nbsp;Methodology, Supervision; \u003cstrong\u003eJingru Yang\u003c/strong\u003e: Funding Acquisition, Resources, Writing-review \u0026amp; editing; \u003cstrong\u003eYongbin Lu\u003c/strong\u003e: Funding Acquisition, Resources, Writing-review \u0026amp; editing.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis work was supported by the Natural Science Foundation of Gansu Province (22JR5RA907); the Postdoctoral Fellowship Program of CPSF under Grant Number (No. GZC20240624); the Science and Technology Program of Gansu Province (No. 24JRRA508); the Fundamental Research Funds for The Central Universities (lzujbky-2024-pd01); the Wu Jieping Medical Foundation Clinical Research Special Fund (320.6750, 2022-21-61); and the United Research Foundation of Gansu Province (25JRRA1250).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and materials\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eData will be made available on request. Requests for specific analyses or data can be submitted by email to [email protected].\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eBray F, Laversanne M, Sung H, Ferlay J, Siegel RL, Soerjomataram I, et al. Global cancer statistics 2022: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 2024;74(3):229-63.\u003c/li\u003e\n\u003cli\u003eDeSantis CE, Ma J, Gaudet MM, Newman LA, Miller KD, Goding Sauer A, et al. Breast cancer statistics, 2019. CA Cancer J Clin. 2019;69(6):438-51.\u003c/li\u003e\n\u003cli\u003eHarbeck N, Penault-Llorca F, Cortes J, Gnant M, Houssami N, Poortmans P, et al. Breast cancer. Nat Rev Dis Primers. 2019;5(1):66.\u003c/li\u003e\n\u003cli\u003eXu H, Xu B. Breast cancer: Epidemiology, risk factors and screening. Chin J Cancer Res. 2023;35(6):565-83.\u003c/li\u003e\n\u003cli\u003eKatsura C, Ogunmwonyi I, Kankam HK, Saha S. Breast cancer: presentation, investigation and management. Br J Hosp Med (Lond). 2022;83(2):1-7.\u003c/li\u003e\n\u003cli\u003eProssnitz ER, Barton M. The G protein-coupled oestrogen receptor GPER in health and disease: an update. Nat Rev Endocrinol. 2023;19(7):407-24.\u003c/li\u003e\n\u003cli\u003eSang C, Song Y, Jin TW, Zhang S, Fu L, Zhao Y, et al. Bisphenol A induces ovarian cancer cell proliferation and metastasis through estrogen receptor-\u0026alpha; pathways. Environ Sci Pollut Res Int. 2021;28(27):36060-8.\u003c/li\u003e\n\u003cli\u003eGłaz P, Rosińska A, Woźniak S, Boguszewska-Czubara A, Biernasiuk A, Matosiuk D. Effect of Commonly Used Cosmetic Preservatives on Healthy Human Skin Cells. Cells. 2023;12(7).\u003c/li\u003e\n\u003cli\u003eVandenberg LN, Bugos J. Assessing the Public Health Implications of the Food Preservative Propylparaben: Has This Chemical Been Safely Used for Decades. Curr Environ Health Rep. 2021;8(1):54-70.\u003c/li\u003e\n\u003cli\u003ePetric Z, Ružić J, Žuntar I. The controversies of parabens - an overview nowadays. Acta Pharm. 2021;71(1):17-32.\u003c/li\u003e\n\u003cli\u003eR\u0026auml;s\u0026auml;nen A, Albrecht E, Annala M, Aro L, Laine AM, Maanavilja L, et al. After-use of peat extraction sites - A systematic review of biodiversity, climate, hydrological and social impacts. Sci Total Environ. 2023;882:163583.\u003c/li\u003e\n\u003cli\u003eGonzales I, Sengupta R, Mani S, Correa-Soto CE, Bejgum B, Alvarez-Nunez F, et al. Understanding the sorption of paraben on plastics using molecular dynamics simulations. Int J Pharm. 2025;671:125256.\u003c/li\u003e\n\u003cli\u003eG\u0026eacute;ni\u0026egrave;s C, Jacques-Jamin C, Duplan H, Rothe H, Ellison C, Cubberley R, et al. Comparison of the metabolism of 10 cosmetics-relevant chemicals in EpiSkin\u0026trade; S9 subcellular fractions and in vitro human skin explants. J Appl Toxicol. 2020;40(2):313-26.\u003c/li\u003e\n\u003cli\u003eMao W, Qu J, Zhong S, Wu X, Mao K, Liao K, et al. Associations between urinary parabens and lung cancer. Environmental science and pollution research international. 2023;30(24):66186-94.\u003c/li\u003e\n\u003cli\u003eFern\u0026aacute;ndez-Mart\u0026iacute;nez NF, Rodr\u0026iacute;guez-Barranco M, Zamora-Ros R, Guevara M, Colorado-Yohar SM, Jim\u0026eacute;nez-Zabala A, et al. Relationship between exposure to parabens and benzophenones and prostate cancer risk in the EPIC-Spain cohort. Environmental science and pollution research international. 2024;31(4):6186-99.\u003c/li\u003e\n\u003cli\u003eWr\u0026oacute;bel AM, Gregoraszczuk E. Action of methyl-, propyl- and butylparaben on GPR30 gene and protein expression, cAMP levels and activation of ERK1/2 and PI3K/Akt signaling pathways in MCF-7 breast cancer cells and MCF-10A non-transformed breast epithelial cells. Toxicology letters. 2015;238(2):110-6.\u003c/li\u003e\n\u003cli\u003eLiang J, Liu QS, Ren Z, Min K, Yang X, Hao F, et al. Studying paraben-induced estrogen receptor- and steroid hormone-related endocrine disruption effects via multi-level approaches. Sci Total Environ. 2023;869:161793.\u003c/li\u003e\n\u003cli\u003eOpinion on Parabens. 2010.\u003c/li\u003e\n\u003cli\u003eXiong X, Zheng LW, Ding Y, Chen YF, Cai YW, Wang LP, et al. Breast cancer: pathogenesis and treatments. Signal Transduct Target Ther. 2025;10(1):49.\u003c/li\u003e\n\u003cli\u003eOkubo T, Yokoyama Y, Kano K, Kano I. ER-dependent estrogenic activity of parabens assessed by proliferation of human breast cancer MCF-7 cells and expression of ERalpha and PR. Food Chem Toxicol. 2001;39(12):1225-32.\u003c/li\u003e\n\u003cli\u003eByford JR, Shaw LE, Drew MG, Pope GS, Sauer MJ, Darbre PD. Oestrogenic activity of parabens in MCF7 human breast cancer cells. J Steroid Biochem Mol Biol. 2002;80(1):49-60.\u003c/li\u003e\n\u003cli\u003eKim S, Chen J, Cheng T, Gindulyte A, He J, He S, et al. PubChem 2025 update. Nucleic Acids Res. 2025;53(D1):D1516-d25.\u003c/li\u003e\n\u003cli\u003eBanerjee P, Kemmler E, Dunkel M, Preissner R. ProTox 3.0: a webserver for the prediction of toxicity of chemicals. Nucleic Acids Res. 2024;52(W1):W513-w20.\u003c/li\u003e\n\u003cli\u003eGallo K, Goede A, Preissner R, Gohlke BO. SuperPred 3.0: drug classification and target prediction-a machine learning approach. Nucleic Acids Res. 2022;50(W1):W726-w31.\u003c/li\u003e\n\u003cli\u003eSzklarczyk D, Santos A, von Mering C, Jensen LJ, Bork P, Kuhn M. STITCH 5: augmenting protein-chemical interaction networks with tissue and affinity data. Nucleic Acids Res. 2016;44(D1):D380-4.\u003c/li\u003e\n\u003cli\u003eDaina A, Michielin O, Zoete V. SwissTargetPrediction: updated data and new features for efficient prediction of protein targets of small molecules. Nucleic Acids Res. 2019;47(W1):W357-w64.\u003c/li\u003e\n\u003cli\u003eTomczak K, Czerwińska P, Wiznerowicz M. The Cancer Genome Atlas (TCGA): an immeasurable source of knowledge. Contemp Oncol (Pozn). 2015;19(1a):A68-77.\u003c/li\u003e\n\u003cli\u003eSzklarczyk D, Nastou K, Koutrouli M, Kirsch R, Mehryary F, Hachilif R, et al. The STRING database in 2025: protein networks with directionality of regulation. Nucleic Acids Res. 2025;53(D1):D730-d7.\u003c/li\u003e\n\u003cli\u003eMajeed A, Mukhtar S. Protein-Protein Interaction Network Exploration Using Cytoscape. Methods Mol Biol. 2023;2690:419-27.\u003c/li\u003e\n\u003cli\u003eSvetnik V, Liaw A, Tong C, Culberson JC, Sheridan RP, Feuston BP. Random forest: a classification and regression tool for compound classification and QSAR modeling. J Chem Inf Comput Sci. 2003;43(6):1947-58.\u003c/li\u003e\n\u003cli\u003eFontanarosa JB, Dai Y. Using LASSO regression to detect predictive aggregate effects in genetic studies. BMC Proc. 2011;5 Suppl 9(Suppl 9):S69.\u003c/li\u003e\n\u003cli\u003eDuan KB, Rajapakse JC, Wang H, Azuaje F. Multiple SVM-RFE for gene selection in cancer classification with expression data. IEEE Trans Nanobioscience. 2005;4(3):228-34.\u003c/li\u003e\n\u003cli\u003eMichailidou K, Lindstr\u0026ouml;m S, Dennis J, Beesley J, Hui S, Kar S, et al. Association analysis identifies 65 new breast cancer risk loci. Nature. 2017;551(7678):92-4.\u003c/li\u003e\n\u003cli\u003eJia B, Zhao X, Wang Y, Wang J, Wang Y, Yang Y. Prognostic roles of MAGE family members in breast cancer based on KM-Plotter Data. Oncol Lett. 2019;18(4):3501-16.\u003c/li\u003e\n\u003cli\u003eUniProt: the Universal Protein Knowledgebase in 2025. Nucleic Acids Res. 2025;53(D1):D609-d17.\u003c/li\u003e\n\u003cli\u003eBerman HM, Burley SK. Protein Data Bank (PDB): Fifty-three years young and having a transformative impact on science and society. Q Rev Biophys. 2025;58:e9.\u003c/li\u003e\n\u003cli\u003eSeeliger D, de Groot BL. Ligand docking and binding site analysis with PyMOL and Autodock/Vina. J Comput Aided Mol Des. 2010;24(5):417-22.\u003c/li\u003e\n\u003cli\u003eEberhardt J, Santos-Martins D, Tillack AF, Forli S. AutoDock Vina 1.2.0: New Docking Methods, Expanded Force Field, and Python Bindings. J Chem Inf Model. 2021;61(8):3891-8.\u003c/li\u003e\n\u003cli\u003eCharoentong P, Finotello F, Angelova M, Mayer C, Efremova M, Rieder D, et al. Pan-cancer Immunogenomic Analyses Reveal Genotype-Immunophenotype Relationships and Predictors of Response to Checkpoint Blockade. Cell Rep. 2017;18(1):248-62.\u003c/li\u003e\n\u003cli\u003eLi DD, Zhou T, Gao J, Wu GL, Yang GR. Circadian rhythms and breast cancer: from molecular level to therapeutic advancements. J Cancer Res Clin Oncol. 2024;150(9):419.\u003c/li\u003e\n\u003cli\u003eIngthorsson S, Traustadottir GA, Gudjonsson T. Breast Morphogenesis: From Normal Development to Cancer. Adv Exp Med Biol. 2025;1464:29-44.\u003c/li\u003e\n\u003cli\u003eToska E. Epigenetic mechanisms of cancer progression and therapy resistance in estrogen-receptor (ER+) breast cancer. Biochim Biophys Acta Rev Cancer. 2024;1879(3):189097.\u003c/li\u003e\n\u003cli\u003eBuraschi S, Pascal G, Liberatore F, Iozzo RV. Comprehensive investigation of proteoglycan gene expression in breast cancer: Discovery of a unique proteoglycan gene signature linked to the malignant phenotype. Proteoglycan Res. 2025;3(1).\u003c/li\u003e\n\u003cli\u003eTreeck O, Haerteis S, Ortmann O. Non-Coding RNAs Modulating Estrogen Signaling and Response to Endocrine Therapy in Breast Cancer. Cancers (Basel). 2023;15(6).\u003c/li\u003e\n\u003cli\u003eWang D, Liu X, Hong W, Xiao T, Xu Y, Fang X, et al. Muscone abrogates breast cancer progression through tumor angiogenic suppression via VEGF/PI3K/Akt/MAPK signaling pathways. Cancer Cell Int. 2024;24(1):214.\u003c/li\u003e\n\u003cli\u003eWang L, Chen L, Schlenk D, Li F, Liu J. Parabens promotes invasive properties of multiple human cells: A potential cancer-associated adverse outcome pathway. Sci Total Environ. 2024;926:172015.\u003c/li\u003e\n\u003cli\u003eAmin MM, Tabatabaeian M, Chavoshani A, Amjadi E, Hashemi M, Ebrahimpour K, et al. Paraben Content in Adjacent Normal-malignant Breast Tissues from Women with Breast Cancer. Biomed Environ Sci. 2019;32(12):893-904.\u003c/li\u003e\n\u003cli\u003eTung CJ, Chen MH, Lin CC, Chen PC. Association between parabens exposure and neurodevelopment in children. Environ Int. 2024;188:108671.\u003c/li\u003e\n\u003cli\u003eZhao H, Zheng Y, Zhu L, Xiang L, Zhou Y, Li J, et al. Paraben Exposure Related To Purine Metabolism and Other Pathways Revealed by Mass Spectrometry-Based Metabolomics. Environ Sci Technol. 2020;54(6):3447-54.\u003c/li\u003e\n\u003cli\u003eAo J, Qiu W, Huo X, Wang Y, Wang W, Zhang Q, et al. Paraben exposure and couple fecundity: a preconception cohort study. Hum Reprod. 2023;38(4):726-38.\u003c/li\u003e\n\u003cli\u003eKhanna S, Dash PR, Darbre PD. Exposure to parabens at the concentration of maximal proliferative response increases migratory and invasive activity of human breast cancer cells in vitro. J Appl Toxicol. 2014;34(9):1051-9.\u003c/li\u003e\n\u003cli\u003eStarek-Świechowicz B, Budziszewska B, Starek A. Endogenous estrogens-breast cancer and chemoprevention. Pharmacol Rep. 2021;73(6):1497-512.\u003c/li\u003e\n\u003cli\u003eYan Y, Su L, Huang S, He Q, Lu J, Luo H, et al. Circadian rhythms and breast cancer: unraveling the biological clock\u0026apos;s role in tumor microenvironment and ageing. Front Immunol. 2024;15:1444426.\u003c/li\u003e\n\u003cli\u003eYe H, Shen X, Li Y, Zou W, Hassan SSU, Feng Y, et al. Proteomic and metabolomic characterization of bone, liver, and lung metastases in plasma of breast cancer patients. Proteomics Clin Appl. 2024;18(5):e2300136.\u003c/li\u003e\n\u003cli\u003eWheler JJ, Parker BA, Lee JJ, Atkins JT, Janku F, Tsimberidou AM, et al. Unique molecular signatures as a hallmark of patients with metastatic breast cancer: implications for current treatment paradigms. Oncotarget. 2014;5(9):2349-54.\u003c/li\u003e\n\u003cli\u003eVallejo-D\u0026iacute;az J, Chagoyen M, Olazabal-Mor\u0026aacute;n M, Gonz\u0026aacute;lez-Garc\u0026iacute;a A, Carrera AC. The Opposing Roles of PIK3R1/p85\u0026alpha; and PIK3R2/p85\u0026beta; in Cancer. Trends Cancer. 2019;5(4):233-44.\u003c/li\u003e\n\u003cli\u003eMiron A, Varadi M, Carrasco D, Li H, Luongo L, Kim HJ, et al. PIK3CA mutations in in situ and invasive breast carcinomas. Cancer Res. 2010;70(14):5674-8.\u003c/li\u003e\n\u003cli\u003eMoghbeli M. PI3K/AKT pathway as a pivotal regulator of epithelial-mesenchymal transition in lung tumor cells. Cancer Cell Int. 2024;24(1):165.\u003c/li\u003e\n\u003cli\u003eKhatpe AS, Adebayo AK, Herodotou CA, Kumar B, Nakshatri H. Nexus between PI3K/AKT and Estrogen Receptor Signaling in Breast Cancer. Cancers. 2021;13(3).\u003c/li\u003e\n\u003cli\u003eAnsieau S. EMT in breast cancer stem cell generation. Cancer Lett. 2013;338(1):63-8.\u003c/li\u003e\n\u003cli\u003eYan LX, Liu YH, Xiang JW, Wu QN, Xu LB, Luo XL, et al. PIK3R1 targeting by miR-21 suppresses tumor cell migration and invasion by reducing PI3K/AKT signaling and reversing EMT, and predicts clinical outcome of breast cancer. Int J Oncol. 2016;48(2):471-84.\u003c/li\u003e\n\u003cli\u003eThorpe LM, Spangle JM, Ohlson CE, Cheng H, Roberts TM, Cantley LC, et al. PI3K-p110\u0026alpha; mediates the oncogenic activity induced by loss of the novel tumor suppressor PI3K-p85\u0026alpha;. Proceedings of the National Academy of Sciences of the United States of America. 2017;114(27):7095-100.\u003c/li\u003e\n\u003cli\u003eZhang HP, Jiang RY, Zhu JY, Sun KN, Huang Y, Zhou HH, et al. PI3K/AKT/mTOR signaling pathway: an important driver and therapeutic target in triple-negative breast cancer. Breast Cancer. 2024;31(4):539-51.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"bmc-pharmacology-and-toxicology","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"phat","sideBox":"Learn more about [BMC Pharmacology and Toxicology](http://bmcpharmacoltoxicol.biomedcentral.com)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/phat/Default.aspx","title":"BMC Pharmacology and Toxicology","twitterHandle":"@BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Propylparaben, Breast cancer, PIK3R1, Machine learning, MR, Molecular docking","lastPublishedDoi":"10.21203/rs.3.rs-7400738/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7400738/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e\u003cp\u003ePropylparaben (PP), a preservative in cosmetics, food, and pharmaceuticals, is a potential endocrine disruptor. However, the molecular mechanisms linking PP exposure to estrogen receptor-positive (ER⁺) breast cancer (BC) remain unclear.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e\u003cp\u003eThis study integrated network toxicology with The Cancer Genome Atlas (TCGA) data to identify carcinogenic targets of PP. Protein-protein interaction (PPI) networks were constructed, and machine learning algorithms identified core genes. Diagnostic performance was evaluated by Receiver Operating Characteristic (ROC) analysis. Mendelian Randomization (MR) assessed causal links between core genes and BC risk. Molecular docking verified the binding affinity between PP and PIK3R1. Immune infiltration was analyzed using single-sample Gene Set Enrichment Analysis.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e\u003cp\u003eWe identified 50 candidate genes related to BC. PPI analysis revealed 20 key genes, and machine learning narrowed it to 6 core genes. ROC analysis showed excellent diagnostic performance (AUC\u0026thinsp;\u0026gt;\u0026thinsp;0.90). MR analysis showed that decreased PIK3R1 expression significantly increased BC risk (OR\u0026thinsp;=\u0026thinsp;0.869, p\u0026thinsp;=\u0026thinsp;0.017). Molecular docking confirmed strong binding between PP and PIK3R1. Immune analysis suggested a correlation between PIK3R1 expression and immune cell abundance.\u003c/p\u003e\u003ch2\u003eConclusion\u003c/h2\u003e\u003cp\u003ePP may promote ER⁺ BC progression by binding to and suppressing PIK3R1, suggesting a potential carcinogenic effect, warranting further investigation in cohort studies.\u003c/p\u003e","manuscriptTitle":"Propylparaben Promotes Estrogen Receptor-Positive Breast Cancer Risk through PIK3R1: A Comprehensive Integrative Study","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-10-31 10:18:41","doi":"10.21203/rs.3.rs-7400738/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2026-02-09T15:17:30+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-02-05T13:56:14+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-02-05T02:31:36+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"84481502888705047451374785970004450239","date":"2026-01-29T05:47:15+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"194228858267098232821266408186820971038","date":"2026-01-24T03:02:47+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-10-21T01:21:38+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-10-15T09:06:50+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-10-07T09:35:22+00:00","index":"","fulltext":""},{"type":"submitted","content":"BMC Pharmacology and Toxicology","date":"2025-10-07T09:30:50+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"bmc-pharmacology-and-toxicology","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"phat","sideBox":"Learn more about [BMC Pharmacology and Toxicology](http://bmcpharmacoltoxicol.biomedcentral.com)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/phat/Default.aspx","title":"BMC Pharmacology and Toxicology","twitterHandle":"@BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"e0a2e926-596d-4b95-84be-d7e464602b95","owner":[],"postedDate":"October 31st, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-02-23T02:53:37+00:00","versionOfRecord":[],"versionCreatedAt":"2025-10-31 10:18:41","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7400738","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7400738","identity":"rs-7400738","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}