Integrative analysis reveals SRD5A2 as a central mediator in PPCPs induced modulation of the prostate cancer immune microenvironment

Abstract

Background

Pharmaceuticals and Personal Care Products (PPCPs) are an emerging class of micropollutants that have attracted increasing attention in recent years. However, their potential role in prostate cancer (PCa) remains largely unexplored. This study aims to elucidate the mechanisms by which PPCPs influence PCa progression and identify key genes involved in this process.

Methods

A network toxicology approach was employed to identify PPCPs related genes associated with PCa. Bioinformatics analyses were conducted to pinpoint key genes and their underlying mechanisms. Additionally, molecular docking was performed to assess the binding affinity between PPCPs and the identified genes.

Results

The study analyzed the relationship between 49 types of PPCPs and PCa, ultimately identifying 24 types of PPCPs that may promote the development of PCa. Network toxicology analysis identified 156 PPCPs-PCa related genes. Further screening using univariate Cox regression analysis, clinical correlation analysis, and Lasso prognostic modeling revealed four prognostic genes closely linked to PCa outcomes. Molecular docking demonstrated strong binding affinities between these genes and the 24 PPCPs, with SRD5A2 exhibiting the highest binding affinity. Moreover, immune infiltration analysis indicated a significant correlation between SRD5A2 and the tumor immune microenvironment (TIME).

Conclusion

The genes F2, SRD5A2, NEK2, and HGFAC play critical roles in PPCPs induced PCa progression, with SRD5A2 being particularly significant. These findings suggest that PPCPs may disrupt TIME homeostasis by downregulating SRD5A2 expression, thereby promoting immune evasion of PCa cells and accelerating tumor progression.

Graphical Abstract

Supplementary Information

The online version contains supplementary material available at 10.1007/s12672-025-03995-x.

Introduction

Cancer remains one of the leading causes of mortality and disease burden worldwide. Its intricate molecular mechanisms and pronounced clinical heterogeneity pose significant challenges for both prevention and treatment. In recent years, several comprehensive reviews, such as Cancer Treatments: Past, Present, and Future (2024) and Overview of Perspectives on Cancer, Newer Therapies, and Future Directions [1, 2], have systematically summarized the advances and trends in cancer research, providing essential theoretical frameworks for elucidating tumorigenesis and developing novel therapeutic strategies. Meanwhile, an increasing body of evidence indicates that environmental pollutants play a critical role in the initiation and progression of urinary system cancers.

Pharmaceuticals and Personal Care Products (PPCPs) constitute a class of emerging micropollutants comprising hundreds of chemical substances widely used in personal healthcare, cosmetics, and agriculture [3]. With the continuous rise in global PPCPs consumption, concerns regarding their environmental persistence and ecological risks have intensified. Studies indicate that 90–95% of PPCPs ingested by humans are excreted unmetabolized into the environment, and current wastewater treatment processes are insufficient for their complete removal. Consequently, PPCPs enter natural water systems, posing potential threats to aquatic ecosystems and human health [4]. The frequent detection of PPCPs in diverse aquatic environments, including oceans, rivers, lakes, and groundwater, has highlighted their toxicological and ecological impacts, underscoring their status as a pressing global environmental concern [5].

Prostate cancer (PCa), one of the most prevalent malignancies affecting men, has shown a rising incidence worldwide, particularly among elderly populations [6]. The etiology and progression of this disease are closely linked to androgen signaling pathways, genetic susceptibility, and environmental factors [7]. Recently, the association between environmental pollution and PCa risk has garnered increasing scientific attention. Research suggests that environmental pollutants, including heavy metals, air pollutants, pesticide residues, and harmful waterborne and soil contaminants, may elevate PCa risk by disrupting hormone levels, inducing oxidative stress, or modulating gene expression [8, 9].

The tumor immune microenvironment (TIME) is a complex milieu composed of tumor cells, immune cells, stromal cells, and extracellular matrix components [10]. Within the TIME, various immune cells interact to regulate tumor growth, invasion, and metastasis. Among these, macrophages exhibit dual functionality: M1-type macrophages exert pro-inflammatory and anti-tumor effects, whereas M2-type macrophages facilitate tumor progression by secreting anti-inflammatory factors and promoting angiogenesis. Additionally, regulatory T cells contribute to immune evasion by suppressing effector T-cell activity. The dynamic equilibrium among these immune cells determines the immune status of the TIME and is closely associated with tumor progression and prognosis [11].

Notably, accumulating evidence suggests that certain PPCPs exhibit carcinogenic potential, with some pharmaceutical and personal care product components being implicated in PCa development [12, 13]. However, a systematic understanding of the common mechanisms underlying the association between multiple PPCPs compounds and PCa risk remains lacking.

To address this gap, network toxicology, bioinformatics analysis, and molecular docking have emerged as powerful tools in modern toxicology and environmental health research, enabling the investigation of complex chemical–biological interactions. Network toxicology systematically elucidates the toxic mechanisms and potential health risks of chemical exposures by constructing multi-level networks of chemical–target–disease relationships. Bioinformatics analysis, leveraging high-throughput omics data and computational biology, facilitates the identification of key molecular markers and signaling pathways associated with chemical exposure and disease progression. Molecular docking, by simulating three-dimensional molecular interactions, predicts binding affinities and structural compatibilities between small molecules and biological macromolecules, offering insights into the molecular mechanisms underlying chemical toxicity.

In this study, we integrated network toxicology, bioinformatics analysis, and molecular docking to systematically analyze 24 PPCPs and their potential roles in PCa. Our findings identified four prognostic genes implicated in PPCPs induced PCa progression, among which SRD5A2 emerged as a key regulator of the TIME. These findings provide novel insights into the environmental health risks posed by PPCPs and contribute to a deeper understanding of their mechanistic involvement in PCa pathogenesis.

Materials and methods

Collection of PPCPs structural information and carcinogenicity testing

Based on the study by Sheyla Ortiz de García et al. [14], 49 PPCPs were identified. Their SMILES structures and 3D structural files were retrieved from the PubChem database (https://pubchem.ncbi.nlm.nih.gov). The retrieved structural data were then input into the ADMETLAB 3.0 platform (https://admetlab3.scbdd.com) for carcinogenicity assessment [15]. According to the description provided on the website, a predicted probability of less than 0.5 is indicated as “(–)” for toxicity. Therefore, compounds with a carcinogenicity score greater than 0.5 were classified as potential carcinogens and included in subsequent analyses.

Collection of PPCPs target genes

Human target genes associated with PPCPs were obtained from the TargetNet database (http://targetnet.scbdd.com), the Swiss Target Prediction database (http://www.swisstargetprediction.ch/), and the SEA database (https://sea.bkslab.org/).

For the TargetNet database, the threshold was set at an area under the ROC curve (AUC) ≥ 0.7, with the fingerprint type specified as ECFP4 fingerprints, and only genes with a predicted probability greater than 0.6 were considered valid targets. After obtaining the UniProt IDs of these genes, we performed ID conversion to common gene names using the UniProt database (https://www.uniprot.org/) and restricted the selection to human-derived genes. In the SwissTargetPrediction database, the species was set to Homo sapiens, and to maximize the number of candidate genes, all predicted human genes provided by the database were included as valid targets. For the SEA database, we selected all human genes with a MaxTC value greater than zero as valid targets. Finally, valid targets obtained from the three databases were combined and deduplicated to generate the set of valid genes for each compound.

Based on their applications in daily life, the compounds were categorized as either “pharmaceuticals” or “personal care products.” The corresponding target genes were then classified and merged according to this categorization, with duplicates removed, yielding two distinct gene sets: pharmaceutical related target genes and personal care product related target genes.

Collection of PCa target genes

Based on prior research [9], PCa related target genes were identified through searches in the GeneCards database (https://www.genecards.org/) using the keyword "prostate cancer" with the top 50% of genes ranked by relevance score selected. Additionally, potential target genes were retrieved from the OMIM database (https://omim.org/) using the same keyword, with the top 25% of genes selected. The gene lists from both databases were merged, and duplicates were removed to obtain the final set of PCa related target genes. This strategy of excluding genes with lower relevance helps to substantially reduce false-positive results arising from weakly correlated genes, thereby enhancing the reliability and feasibility of subsequent analyses.

Identification of PPCPs-PCa related genes and enrichment analysis

To identify PPCPs-PCa related genes, an intersection analysis was performed among pharmaceutical target genes, personal care product target genes, and PCa target genes. Functional enrichment analyses, including Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway and Gene Ontology (GO) analyses, were conducted using the "ClusterProfiler" package in R. The results were visualized using the online platform https://www.bioinformatics.com.cn.

PPI network visualization

To explore the potential molecular interactions among PPCPs–PCa related genes, the identified candidate genes were uploaded to the STRING database (https://cn.string-db.org/). A confidence score threshold of ≥ 0.4 was applied to ensure medium-to-high reliability of the protein–protein interaction (PPI) data, thereby minimizing potential false-positive associations. The resulting PPI network was subsequently imported into Cytoscape software (https://cytoscape.org/) for further topological analysis and visualization. Within Cytoscape, the cytoHubba plugin was employed to calculate the maximal clique centrality (MCC) score for each node. This approach enabled the identification of hub genes with the highest degree of connectivity and potential biological significance, thereby highlighting the intricate interactions among candidate genes. Furthermore, this analysis facilitated the recognition of key genes that may play pivotal roles in the pathogenesis of prostate cancer while eliminating isolated nodes.

Univariate Cox regression analysis and clinical feature analysis

Univariate Cox regression analysis was performed on PPCPs-PCa related genes using the TCGA-PRAD dataset to identify genes significantly associated with prognosis. A P value < 0.05 was considered statistically significant, and genes meeting this threshold were included in subsequent analyses.

Subsequently, clinical relevance validation was performed using the online database Xiantao Platform (https://www.xiantaozi.com/). Expression differences of these genes between tumor and non-tumor samples in the TCGA-PRAD dataset were examined. Additionally, their expression levels across different clinical T and M stages, as well as pathological T and N stages, were analyzed to identify genes significantly associated with both tumor prognosis and clinical features. Genes unrelated to both prognosis and clinical characteristics were excluded from further analysis.

Prognostic model construction

LASSO regression analysis was conducted on the filtered genes in the TCGA-PRAD dataset to construct a prognostic model. The risk score for each patient was calculated using the formula: RS = EXPgene1*β₁ + EXPgene2*β₂ + EXPgene3*β₃ + … + EXPgenen*β_n where EXP represents gene expression levels and β denotes the regression coefficients from multivariate LASSO regression. Patients were stratified into high-risk and low-risk groups based on the median risk score. Kaplan–Meier survival analysis was employed to evaluate the predictive performance of the model, with P < 0.05 considered statistically significant for prognostic prediction. Additionally, time-dependent receiver operating characteristic (ROC) curve analyses were conducted at 1-, 3-, and 5-year follow-up intervals. The AUC was calculated to quantify the discriminative ability of the model, with higher AUC values indicating better predictive performance.

Prognostic model validation and nomogram construction

The stability of the prognostic model was validated using external datasets, including GSE46602 and the Cambridge dataset. Based on the expression levels of the four key genes, a risk score for each patient was calculated using the formula derived from the predictive model. Patients were then stratified into high-risk and low-risk groups according to the median risk score. To further evaluate the predictive performance of the model, Kaplan–Meier survival analysis and ROC curve analysis were conducted. Subsequently, a nomogram was constructed integrating age, pathological T and N stages, prostate specific antigen (PSA) levels, Gleason score, and risk score to predict individual survival probabilities. Calibration curves were used to assess the accuracy of 1-year, 3-year, and 5-year survival rate predictions.

Molecular docking

Protein Data Bank (PDB) files for prognosis related genes were obtained from the RCSB database (https://www.rcsb.org/). Molecular docking analysis was conducted using the CB-Dock2 platform (https://cadd.labshare.cn/cb-dock2/php/blinddock.php#job_list_load). Binding energies were used to assess ligand-receptor interactions, with binding energies below − 5 kcal/mol considered indicative of strong binding, and values below − 7 kcal/mol suggesting highly stable interactions. Molecular docking results were visualized using R software, and a heatmap was generated to depict the interaction strength between prognosis related genes and the 24 identified PPCPs.

Immune cell infiltration analysis and correlation with core genes

The correlation between prognosis related genes and 24 immune cell types was analyzed using the online platform https://www.xiantaozi.com/, and a correlation heatmap was generated. The "ESTIMATE" R package was used to calculate Stromal scores, Immune scores, and Estimate scores based on gene expression data. Key genes were identified based on their correlation with immune scores, and TCGA-PRAD patients were stratified into high and low expression groups using the median gene expression level. The proportions of 22 immune cell types in each patient were calculated, and immune infiltration mechanisms were further explored based on gene expression stratification.

Statistical analysis

All statistical analyses were performed using R software (version 4.3.2) and corresponding R packages. Continuous variables were presented as mean ± standard deviation (SD), and categorical variables as counts and percentages. Group comparisons were conducted using the t-test or Wilcoxon rank-sum test for continuous variables, and the Chi-square test or Fisher’s exact test for categorical variables. Survival outcomes were analyzed using Kaplan–Meier curves with log-rank tests to assess differences between groups. Univariate Cox proportional hazards regression analysis was performed to identify prognosis related genes, and significant variables were subsequently incorporated into the LASSO regression model for multivariate prognostic modeling. The predictive accuracy of the model was evaluated using time-dependent ROC curves and the AUC. Calibration curves were constructed to assess the concordance between predicted and observed survival probabilities. Spearman’s correlation analysis was applied to evaluate associations between gene expression and immune cell infiltration. A two-tailed P value < 0.05 was considered statistically significant.

Results

Collection of structural information and carcinogenicity testing of PPCPs

We queried the PubChem database to obtain the structural formulas of 49 PPCPs compounds. (Supplementary Table 1). Next, we assessed the carcinogenic potential of these PPCPs using the ADMETlab 3.0 platform. Following carcinogenicity testing, 24 PPCPs, including 19 drugs and 5 personal care products, were retained for further analysis (Supplementary Table 1).

Collection of PPCPs target genes and PCa related genes

To identify human target genes of these 24 PPCPs, we queried the TargetNet, SwissTargetPrediction, and SEA databases. After merging and deduplication, we identified 415 target genes associated with the personal care compounds and 819 target genes related to drug compounds (Supplementary Table 2).

We then obtained PCa related genes from the GeneCards and OMIM databases. After merging and removing duplicates, we identified 5,358 PCa related genes (Supplementary Table 2).

Identification of PPCPs-PCa related genes and PPI network visualization

By intersecting the gene sets associated with personal care compounds, drug compounds, and PCa, we identified 156 PPCPs-PCa related genes (Fig. 1A). These genes were then analyzed using the STRING database, and isolated genes were removed, leaving 155 genes. The genes were ranked by MCC scores in Cytoscape, with darker colors indicating stronger interactions (Supplementary Fig. 1).

Fig. 1.

Identification and Functional Analysis of PPCPs-PCa Related Genes. A The Venn diagram illustrates the intersection of target genes associated with PPCPs-PCa related genes. B GO enrichment analysis of PPCPs-PCa-related genes. C KEGG pathway enrichment analysis of PPCPs-PCa-related genes. D Forest plot of univariate cox regression analysis in the TCGA-PRAD cohort

Enrichment analysis

GO enrichment analysis revealed that these PPCPs-PCa related genes are primarily involved in steroid hormone signaling pathways, stress responses, cell membrane and organelle functions, and transcriptional regulation (Fig. 1B). KEGG analysis indicated that these genes are significantly associated with cancer development, hormone signaling regulation, signal transduction, metabolic pathways, and cardiovascular diseases (Fig. 1C). These findings suggest that disruptions in sex hormone balance and abnormal activation of multiple signaling pathways may play a critical role in PPCPs induced PCa.

Univariate Cox regression analysis and gene-clinical feature analysis

To identify genes significantly associated with PCa prognosis, we conducted univariate Cox regression analysis on the 156 PPCPs-PCa related genes. Eight genes were found to significantly impact patient survival: PTGS2, PGR, SRD5A2, KCNMA1 and MYLK were identified as protective factors, whereas HSD17B3, NEK2, F2, and HGFAC were associated with poorer prognosis (Fig. 1D).

We then examined the expression and clinical relevance of these eight genes in the TCGA-PRAD dataset (Fig. 2). Seven genes demonstrated significant clinical correlations: SRD5A2, KCNMA1, PGR, MYLK and PTGS2 decreased with tumor progression, suggesting a protective role, whereas F2, HGFAC, and NEK2 increased with tumor progression, indicating a higher risk for PCa. HSD17B3 exhibited no significant variation across different clinical statuses and was therefore excluded from further analysis.

Fig. 2.

Boxplots and violin plots depicting expression differences and clinical characteristics of nine prognosis related genes

Prognostic model establishment and validation

Lasso regression analysis of the seven clinically relevant genes identified an optimal prognostic model incorporating SRD5A2, NEK2, F2, and HGFAC (Fig. 3A, B). The corresponding regression coefficients are presented in Supplementary Table 3.

Fig. 3.

Construction and Validation of a Prognostic Model for Prostate Cancer. A LASSO coefficient profiles. B Ten-time cross-validation for tuning parameter selection in the LASSO model. C Kaplan–Meier survival curves comparing high-risk and low-risk groups based on the prognostic model in the TCGA training set. D Time-dependent ROC curves demonstrating the predictive accuracy of the model at 1, 3, and 5 years in the TCGA training set. E Kaplan–Meier survival curves comparing high-risk and low-risk groups based on the prognostic model in the Cambridge testing set. F The ROC curves demonstrating the predictive accuracy of the model in the Cambridge testing set. G Kaplan–Meier survival curves comparing high-risk and low-risk groups based on the prognostic model in the GSE46602 testing set. H The ROC curves demonstrating the predictive accuracy of the model in the GSE46602 testing set. I Nomogram for predicting prostate cancer survival probability. J Calibration curve for predicting prostate cancer survival probability

Kaplan–Meier survival analysis demonstrated that the predictive model effectively stratified patient survival outcomes across the TCGA-PRAD, Cambridge, and GSE46602 cohorts. Meanwhile, ROC curve analysis revealed consistently high AUC values in these datasets, further supporting the robust predictive performance of the model (Figs. 3C, D).

To validate the model's robustness, we performed external validation using the GSE46602 and Cambridge datasets. Both datasets confirmed the model's predictive stability (Fig. 3E–H). Furthermore, we developed a nomogram integrating age, pN, pT, PSA, Gleason score, and risk score to estimate patient survival at 1, 3, and 5 years (Fig. 3I, J). The results indicated that the prognostic model provides stable and reliable predictions.

Molecular docking

Molecular docking analysis of the four prognostic genes with 24 PPCPs revealed that all four genes exhibited binding affinity to PPCPs, suggesting their potential roles in PPCPs mediated toxicity (Fig. 4A). Notably, SRD5A2 demonstrated the most stable binding interactions (Fig. 4C–F). Statistical analysis confirmed that SRD5A2 exhibited significantly stronger binding affinity compared to the other three genes (Fig. 4B), indicating that SRD5A2 may be a key mediator of PPCPs toxicity.

Fig. 4.

Molecular Docking of PPCPs with Prognosis Related Genes. A The heatmap illustrates the binding energies (kcal/mol) derived from molecular docking analysis between 4 key molecules and 24 PPCPs. B Comparison of binding affinities between four prognosis related genes and PPCPs. C Molecular docking visualization of SRD5A2 with levofloxacin. D Molecular docking visualization of SRD5A2 with paroxetine. E Molecular docking visualization of SRD5A2 with irbesartan. F Molecular docking visualization of SRD5A2 with tamoxifen

Immune infiltration analysis and SRD5A2 correlation with immune cells

To explore the potential mechanisms underlying PPCPs induced PCa, we analyzed the associations between the four prognostic genes and 24 immune cell types (Fig. 5A). SRD5A2 displayed significant correlations with 22 of these immune cells. Moreover, analysis of stromal, immune, and ESTIMATE scores demonstrated that SRD5A2 was significantly associated with all three indices (Fig. 5B), suggesting its role in modulating the TIME.

Fig. 5.

Correlation of Prognosis Related Genes and SRD5A2 Expression with Immune Cell Infiltration in Prostate Cancer. A Correlation analysis between four prognosis related genes and 24 immune cell types. B Relationship between four prognosis related genes and Stromal, Immune, and Estimate scores. C Immune cell distribution in TCGA-PRAD patients stratified by SRD5A2 expression levels. D Differential immune cell infiltration in TCGA-PRAD patients stratified by SRD5A2 expression levels (*p < 0.05, **p < 0.01, ***p < 0.001)

Further subgroup analysis stratified PCa patients into SRD5A2 high and low expression groups (Fig. 5C). In the SRD5A2 low-expression group, M0, M1, M2 macrophages, and regulatory T cells were significantly increased, whereas resting and activated mast cells and naive B cells were significantly reduced (Fig. 5D). These findings suggest that SRD5A2 downregulation may promote immune suppression and facilitate tumor immune escape in PCa.

Discussion

PPCPs are a class of trace organic pollutants widely present in the environment, encompassing various chemical substances such as prescription drugs, over the counter medications, cosmetics, skincare products, and fragrances [16]. These substances primarily enter water bodies, soil, and the atmosphere through human and animal excretion, domestic wastewater discharge, and improper disposal. Although their environmental concentrations are generally low, their persistence and bioactivity may pose potential threats to ecosystems and human health [17]. Previous studies have indicated that certain components of PPCPs may disrupt the reproductive systems of aquatic organisms, promote the spread of antibiotic resistance, and even affect human health through bioaccumulation in the food chain [5]. In recent years, research on the environmental behavior, ecological toxicity, and removal technologies of PPCPs has gained significant attention, aiming to mitigate their potential hazards to ecological environments and public health.

PCa is one of the most common malignant tumors in men, with its occurrence and progression closely linked to androgen signaling pathways [18]. Recently, the role of environmental factors in PCa has received increasing attention, with PPCPs, as emerging environmental pollutants, potentially influencing PCa risk by disrupting the endocrine system [19, 20]. Some PPCPs compounds can mimic or antagonize androgens, thereby disrupting androgen signaling pathways and promoting the occurrence and progression of PCa [21]. Additionally, PPCPs may indirectly affect PCa development through mechanisms such as inducing oxidative stress, inflammatory responses, or epigenetic modifications [22]. Therefore, studying the association between PPCPs and PCa not only helps elucidate the mechanisms by which environmental factors contribute to cancer but also provides new insights for PCa prevention and treatment.

This study identified four PPCPs-PCa prognosis related genes—SRD5A2, NEK2, F2, and HGFAC—that can effectively predict PCa survival prognosis. Molecular docking analysis demonstrated that these four genes tightly bind to 24 PPCPs, suggesting they may be key molecules mediating PPCPs-induced toxicity. Among them, SRD5A2 plays a particularly critical role. SRD5A2 exhibits significantly higher binding affinity with PPCPs compared to the other three genes and also shows strong clinical relevance and immune cell correlation, suggesting it may mediate PPCPs induced TIME imbalances, thereby accelerating PCa progression.

SRD5A2 is a key enzyme highly expressed in prostate tissue, primarily responsible for catalyzing the conversion of testosterone to the more biologically active dihydrotestosterone (DHT) [23]. DHT, as the primary ligand of the androgen receptor (AR), activates downstream signaling pathways upon AR binding, regulating prostate cell growth, differentiation, and survival [24]. Reduced SRD5A2 expression may lead to decreased DHT production, thereby weakening AR-dependent gene regulation, affecting cell differentiation, apoptosis, and proliferation, ultimately disrupting prostate cell homeostasis and increasing the risk of malignant transformation [25, 26]. Additionally, decreased DHT levels may upregulate pro-inflammatory factors, further activating inflammation related pathways such as NF-κB, promoting PCa cell proliferation and survival [27]. Furthermore, reduced SRD5A2 expression may disrupt redox balance, leading to increased reactive oxygen species levels, thereby inducing DNA damage and genomic instability, accelerating PCa progression [28].

Notably, the phenotypic plasticity of cancer cells plays a critical role in tumor initiation, progression, and therapeutic resistance. As highlighted in Unraveling the Dangerous Duet between Cancer Cell Plasticity and Drug Resistance (2023), tumor cells can dynamically undergo phenotypic transitions that enable them to enter drug tolerant states, thereby evading immune surveillance and escaping the inhibitory effects of targeted therapies [29]. This study found that PPCPs significantly reduce SRD5A2 expression, accompanied by a marked increase in M0, M1, and M2 macrophages and regulatory T cells, while leading to a significant decrease in resting mast cells, activated mast cells, and naive B cells. These changes suggest that SRD5A2 may promote PCa progression by reshaping the immune microenvironment. Specifically, the increase in M0, M1, and M2 macrophages indicates enhanced macrophage recruitment and polarization within the tumor microenvironment, potentially promoting inflammatory responses and supporting tumor growth and invasion [30]. Meanwhile, increased regulatory T cells (Tregs) may enhance immune suppression, inhibiting effector T cell-mediated anti-tumor immunity, thereby facilitating tumor immune escape [31]. Simultaneously, the decrease in resting and activated mast cells may weaken local immune regulation and inflammation-mediated anti-tumor effects, while the decrease in naive B cells may impair humoral immune responses, reducing antibody-mediated anti-tumor effects [32]. Overall, PPCPs-induced SRD5A2 downregulation may enhance immune suppression and macrophage infiltration, exacerbating adaptive immunity weakening, promoting tumor immune escape, and ultimately accelerating PCa progression.

This study has several strengths. First, it comprehensively integrated network pharmacology, bioinformatics analyses, and molecular docking to systematically investigate the roles of PPCPs–PCa related genes in prostate cancer prognosis, clinical staging, and immune microenvironment regulation. This multidimensional approach enhances the scientific rigor and reliability of the study. Second, compared with a previously reported four-gene prognostic model that also included SRD5A2 [33], our model demonstrated higher AUC values, indicating superior predictive performance. Finally, this study, for the first time, uncovers the potential impact of PPCPs on the prostate cancer immune microenvironment, thereby providing a novel theoretical foundation for the development of immune checkpoint inhibitors and precision gene-targeted therapies.

However, several limitations should also be acknowledged. First, the data used in this study were primarily derived from public databases, and potential heterogeneity and variable quality of these datasets may introduce bias. Second, unlike previous prognostic models incorporating SRD5A2 [34], this study lacks functional validation of SRD5A2 knockdown or overexpression at the animal and cellular levels. Furthermore, while we preliminarily identified a potential association between SRD5A2 and the immune microenvironment, further validation through single-cell sequencing and in vitro experiments is warranted to elucidate the underlying mechanisms. In addition, although molecular docking analysis provided preliminary clues to potential mechanisms of action, more refined and dynamic approaches, such as molecular dynamics simulations, are required to enhance predictive reliability. Finally, variations in PPCPs exposure levels and potential population biases may contribute to uncertainties in the interpretation of results. Therefore, this study should be regarded as an initial exploration of the relationship between PPCPs and prostate cancer, which requires further validation and refinement through large-scale, multicenter cohorts and functional experiments in the future.

Conclusion

This study systematically employed network toxicology, bioinformatics analysis, and molecular docking to reveal the potential mechanisms by which PPCPs influence PCa. The findings suggest that SRD5A2, NEK2, F2, and HGFAC play key roles in this process, with SRD5A2 expression changes being particularly significant in PCa progression, likely contributing to TIME imbalances. This research provides new theoretical foundations for developing immune inhibitors and gene-targeted therapy while also emphasizing the need for environmental protection measures to mitigate PPCPs-associated health risks.

Author contributions

TL and YY participated in the design of the study, YL and ZT conducted the data acquisition, and drafted the manuscript. YL, TZ and PJ interpreted and analyzed the data. ZL, ZZ and CM performed the statistical analysis. All authors read and approved the final version of the manuscript.

Funding

Our work was supported by grants from the Mianyang Municipal Health Commission (2024128).

Data availability

The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethical approval

The data used in this study were obtained from publicly available databases, including The Cancer Genome Atlas (TCGA) and the Gene Expression Omnibus (GEO). Since these databases consist of anonymized data that is freely accessible to the public, no ethical approval or informed consent was required for this study. The data were used solely for research purposes and were in compliance with the relevant ethical guidelines for the use of publicly available data.

Consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.