Bisphenol A and breast cancer: Can mechanistic plausibility be reconciled with epidemiological inconsistency?

Introduction
Breast cancer constitutes a group of malignant tumors originating from the epithelial cells of the mammary gland. Currently, it remains the most prevalent malignancy and a leading cause of cancer-related mortality among women worldwide.1,2 Despite the implementation of comprehensive screening programs and advances in multimodal therapeutic strategies, including surgery, chemotherapy, radiotherapy, and endocrine therapy, the global incidence of breast cancer continues to rise.3 Furthermore, clinical management faces persistent challenges, particularly therapeutic resistance and disease recurrence, which significantly compromise long-term survival.4,5,6 While genetic predispositions, such as BRCA1/2 mutations, are well-characterized, they account for only a minority of cases, underscoring the critical need to investigate modifiable environmental factors in the disease’s etiology.7,8
Bisphenol A (BPA) is a high-production-volume industrial chemical extensively utilized in the manufacture of polycarbonate plastics and epoxy resins.9 Due to its pervasive use in food packaging, thermal paper, and medical devices, human exposure to BPA is ubiquitous, occurring primarily through dietary ingestion, dermal contact, and inhalation.10,11,12 As a lipophilic endocrine-disrupting chemical (EDC), BPA does not merely circulate transiently; it possesses the capacity to bioaccumulate within human adipose tissue. Notably, studies have detected significant concentrations of BPA in the breast adipose tissue of patients undergoing mastectomy, confirming its direct localization within the mammary microenvironment.13,14 Beyond its local accumulation, BPA exposure has been implicated in a spectrum of systemic pathologies, ranging from metabolic disorders to reproductive dysfunction, which are attributed to its ability to interfere with hormonal signaling.15
However, the specific association between BPA exposure and breast cancer risk remains a subject of scientific complexity and controversy. While mechanistic studies have provided robust evidence that BPA can induce proliferation, metastasis, and drug resistance in breast cancer models,16,17 epidemiological evidence presents a conflicting landscape. Several case-control studies have reported a positive correlation between BPA exposure and increased breast cancer risk or mammographic density.18,19,20 In contrast, large-scale prospective cohort studies and recent meta-analyses have frequently failed to establish a statistically significant causal link, reporting null associations.21,22,23 Given this discrepancy between biological plausibility and population-level observations, this review aims to critically synthesize current evidence. We will explore the nuanced relationship between BPA and breast cancer, analyzing the divergence between laboratory and epidemiological findings, to provide a more comprehensive perspective for clinical prevention and therapeutic strategies. An integrated conceptual framework encompassing BPA exposure sources, mechanistic pathways, and public health implications is presented in Figure 1.

Environmental pervasiveness and toxicokinetic disposition of bisphenol A

Chemical attributes and multi-vector exposure
BPA, chemically identified as 2,2-bis(4-hydroxyphenyl)propane, is a diphenylmethane derivative characterized by two phenol functional groups connected by a methyl bridge.9,24 As a high-production-volume industrial chemical, it serves as the fundamental monomer for polycarbonate plastics and a precursor for epoxy resins.9,25 As one of the most ubiquitous xenoestrogens in the global environment, its chemical nature allows it to permeate daily life through multiple vectors.26,27
Human exposure to BPA is cumulative and stems from a complex profile of sources rather than a single origin.24,28 Dietary ingestion remains the predominant route of exposure, primarily driven by the leaching of BPA from the linings of food and beverage containers, jar caps, and internal coatings.24,26 Furthermore, processed meat products, both canned and non-canned, have been identified as significant dietary reservoirs, with contamination levels heavily influenced by industrial processing and packaging materials.29
Beyond dietary sources, humans are subjected to a “cocktail” of non-dietary exposures.30 It has been systematically evaluated in an umbrella review, which highlights that every major class of plastic-associated chemicals is linked to at least one adverse health outcome, with BPA specifically identified as a major driver of endocrine and circulatory disorders.31 Dermal absorption constitutes a critical pathway; thermal paper receipts serve as a high-concentration source of free BPA, posing occupational hazards to individuals such as cashiers who may accumulate the compound through frequent skin contact.10 Environmental monitoring has also confirmed the presence of BPA in ambient air particulates, particularly in urban settings, suggesting that inhalation is a relevant entry route.11
Significantly, iatrogenic contamination represents a direct and often high-dose route that is clinically overlooked. Human exposure occurs through dental materials and various healthcare equipment.24,32 Systematic reviews indicate that patients utilizing medical-hospital devices exhibit significantly elevated plasma BPA levels.33 For chronic renal patients, a notable association has been found between plasma BPA concentrations and disease severity, directly linking healthcare interventions to toxicant burden.33

Biological fate: The paradox of metabolism and tissue bioaccumulation
Understanding the potential pathogenicity of BPA requires examining its toxicokinetics, which presents a paradox between rapid clearance and persistent tissue burden. BPA has a short biological half-life, with blood clearance reported between 44 and 75 min in animal models.34,35 However, this pharmacokinetic parameter can be misleading. Due to ubiquitous environmental sources, individuals experience chronic, low-dose exposure via multiple portals, leading to a state of continuous bodily burden.24,36
Crucially, BPA is a lipophilic compound, a property that facilitates its bioaccumulation in lipid-rich tissues.28,37 While urinary measurements fluctuate with recent dietary habits and reflect rapid excretion, they often fail to capture the long-term cumulative burden stored in tissues.24 Clinical data indicate that BPA bioaccumulates in adipose tissue, which can disrupt metabolic and inflammatory functions.28,38 Studies comparing biological matrices have demonstrated that BPA concentrations in breast adipose tissue are significantly higher in patients with breast cancer compared to healthy controls, suggesting that the mammary gland acts as a reservoir for this toxicant.39,40 The observed concentration differences may be associated with polymorphisms in metabolic enzymes locally within the breast.41
This bioaccumulation is particularly concerning during critical developmental windows. Maternal-fetal transfer is well-documented, with the fetus remaining continuously exposed during the sensitive period of organogenesis.25,42 Prenatal exposure has been linked to adverse neonatal health outcomes and intestinal or placental oxidative stress in gestational models.25,43 This evidence suggests that assessing environmental exposure history, particularly in relevant tissue matrices rather than spot urine samples, is essential for accurately estimating the carcinogenic and metabolic risk.24,28

Heterogeneity of epidemiological evidence
Although laboratory studies have provided a solid biological rationale for the carcinogenicity of BPA, translating these findings into consistent evidence at the population level remains challenging. The current epidemiological landscape exhibits significant heterogeneity. Large-scale prospective studies often report null associations, while studies employing refined exposure assessment, focusing on specific genetic backgrounds, or targeting critical developmental windows frequently observe significant risks. This section aims to systematically review the existing epidemiological evidence, objectively presenting the complexities and contradictions within this field.

Overall null associations in large-scale population studies
Within the evidence-based medicine hierarchy, large-scale prospective cohort studies and meta-analyses are generally considered high-level evidence for evaluating associations between environmental exposures and disease. However, multiple such studies investigating BPA and breast cancer risk have commonly failed to detect robust positive associations.
A recent meta-analysis incorporating nine studies involving a total of 10,695 subjects found a pooled odds ratio of 1.00 (95% CI: 0.92–1.08) for the association between BPA exposure and breast cancer risk, which was not statistically significant.21 This finding aligns with an earlier meta-analysis of 7,820 subjects, which also reported a non-significant pooled odds ratio of 0.85.44 These results suggest that routinely assessed BPA exposure levels do not demonstrate a strong association with breast cancer incidence in the general population when individual differences and specific exposure windows are not considered.
Longitudinal cohort studies, which collect biospecimens pre-diagnostically, help reduce recall bias. Several influential large-scale cohort studies reinforce the observation of null associations. For instance, the EPIC-Spain prospective study, which followed over 4,000 participants, found no significant association between serum BPA levels at recruitment and subsequent breast cancer risk.23 Similarly, a nested case-control study within the multiethnic cohort reported no correlation between pre-diagnostic urinary BPA or its metabolites and breast cancer risk.22 A large cross-sectional study published in 2025 involving 4,455 subjects identified associations for other phenols including triclosan but found no significant link for BPA after adjusting for confounders.45

Risk signals from refined exposure assessment and specific subgroups
In contrast to the overall conclusions from large-scale studies, associations between BPA and breast cancer risk present significant positive signals when investigations employ more biologically targeted exposure assessment methods or focus on genetically susceptible subgroups. These findings suggest that risks diluted in the general population may be revealed under specific conditions or with more precise measurements.
Regarding exposure assessment, differences in measurement strategies directly impact outcomes. Multiple case-control studies illustrate this point through comparisons of different biomarkers. For example, a study in Poland involving 1,150 postmenopausal women measured the primary urinary metabolite BPA-glucuronide and found no association with breast cancer risk.46 Conversely, a study in Mexico involving 798 women, while also using urine samples, specifically quantified the estrogenically active free BPA prototype. This study provides quantitative evidence for a dose-response relationship, demonstrating that women in the highest quartile of exposure to free BPA, the estrogenically active prototype, had a 2.31-fold higher risk (OR = 2.31, 95% CI: 1.43–3.74) of breast cancer compared to those in the lowest quartile.19 The direct comparison of these two studies highlights the importance of the biological activity of the target analyte. Furthermore, studies directly assessing target organ burden provide evidence from another dimension. An analysis of breast adipose tissue from 52 women showed that the average BPA concentration in tissue from patients with cancer (4.20 ng/g) was significantly higher than in controls (1.80 ng/g, p < 0.01).13 Another smaller tissue study reported detectable BPA levels but found no concentration difference between cancerous and benign tissues, suggesting factors such as tissue sampling site and sample size may also influence results.14
In identifying susceptible subgroups, genetic epidemiological studies offer key insights. A case-control study in China involving 604 women not only found higher median urinary BPA concentrations in cases (1.36 μg/g creatinine) versus controls (0.89 μg/g creatinine, p < 0.001) but also delved into gene-environment interactions.18 The study showed that among women carrying the high-risk allele A of the CYP17A1 gene rs743572, those with high BPA exposure had a 2.49-fold increased breast cancer risk (95% CI: 1.52–4.13) compared to those with low exposure carrying the wild-type GG genotype. This strong interaction effect was masked in the overall population analysis. An exploratory study from Pakistan reported significantly higher serum BPA levels in patients with breast cancer compared to controls, correlating with increased oxidative stress markers and upregulation of genes, including p53, in tumor tissue, providing preliminary clues linking exposure to tumor biology phenotypes.47

Lifecycle-dependent evidence using mammographic density as an intermediate phenotype
Mammographic density, one of the strongest imaging-based risk factors for breast cancer, serves as a valuable intermediate endpoint for investigating the early biological effects of BPA. Existing evidence indicates that the association between BPA and mammographic density is highly dependent on the life stage assessed and exhibits complex patterns.
Research in adolescence reveals the potential for non-monotonic associations. A key piece of evidence comes from a longitudinal study of 200 Chilean girls.48 This research found a significant U-shaped association between urinary BPA concentration and fibroglandular volume measured at Tanner stage IV. Compared to the lowest tertile, girls in the middle tertile had an approximately 10% lower average FGV, while those in the highest tertile rebounded to levels similar to the low-exposure group. This U-shaped, non-monotonic dose-response is a hallmark of EDCs such as BPA. Unlike classical toxicants, which typically exhibit linear dose-response curves, EDCs can exert effects at low doses that are not predicted by high-dose studies. The observed pattern—decreased fibroglandular volume at moderate exposure followed by recovery at higher exposure—likely reflects the biphasic nature of hormone receptor signaling. At low to moderate concentrations, BPA may activate estrogen-related pathways that transiently influence mammary gland development. At higher concentrations, compensatory homeostatic mechanisms, receptor desensitization, or mild cytotoxicity may counteract this initial effect. This highlights the importance of assessing EDC effects across environmentally relevant dose ranges, as risk assessments based solely on high-dose studies may underestimate real-world hazards.
Among adult women, the evidence is inconsistent and may be influenced by menopausal status. A cross-sectional study of 264 postmenopausal women in the United States found that those with detectable serum BPA levels above the median (>0.55 ng/mL) had a significantly higher average percent breast density (17.6%) compared to women with undetectable BPA (12.6%), representing a relative increase of approximately 40%.20 In contrast, a study of 97 young American women with an average age of about 25, utilizing more precise pooled 24-h urine samples, found no significant association between any measured phenols or phthalates and breast density.49 Another large cross-sectional study from 2025 similarly found no association between urinary BPA and breast cancer risk in adult women overall but suggested a risk for triclosan.45 These contradictory outcomes may reflect differences in study population age, menopausal status, exposure assessment methods, and sample size.

Evidence from genomics and novel causal inference methods
To overcome residual confounding and reverse causation in observational studies, researchers have begun employing genetic tools such as Mendelian randomization to infer potential causal relationships between BPA-related exposures and breast cancer. This method uses genetic variants strongly associated with an exposure factor as instrumental variables. Since genotypes are randomly assigned at conception and generally precede disease onset, they can provide stronger causal evidence.
The most direct evidence currently comes from a large-scale Mendelian randomization study published in 2025.50 This study utilized public genomics databases, including eQTLGen, to use the genetically predicted expression levels of BPA-responsive genes such as TET2 as instrumental variables. The analysis found a statistically significant association between genetically predicted higher expression of these BPA-targeted genes and an increased risk of estrogen receptor-positive breast cancer, suggesting a potential causal relationship. This finding links, for the first time at a population genetics level, molecular targets of BPA action with the risk of a specific breast cancer subtype. Although no Mendelian randomization study has yet directly used circulating BPA concentration as the exposure, this research provides indirect support from human genetic data for the carcinogenic potential of biological pathways known to be disrupted by BPA. Furthermore, some cutting-edge exploratory work has utilized long-term in vitro exposure models to identify and establish molecular signatures associated with induction by pollutants including BPA, such as signatures involving STAT3 and VEGFA.51 Bioinformatics analyses have confirmed significant associations between these signatures and poorer prognosis in patients with breast cancer, providing an important direction for future integration of exposomics and genomics to precisely identify high-risk individuals.
A synopsis of the key epidemiological and genomic studies discussed is provided in Table 1.

Integrated carcinogenic mechanisms: the cascade from molecular signaling to histopathological alterations
Experimental evidence indicates that the promotion of breast cancer initiation and progression by BPA involves a complex, multi-layered pathophysiological process. This process begins with the interaction of BPA with diverse cellular targets, triggering alterations in a series of intracellular signal transduction events. These initial molecular disturbances can be perpetuated long-term through epigenetic mechanisms, concurrently accompanied by adaptive remodeling of the local microenvironment. Exposure to BPA during critical developmental windows may induce irreversible programming changes in mammary tissue, thereby establishing the foundation for long-term tumorigenesis.

Activation and interaction of multiple receptor signaling pathways
As an exogenous compound with promiscuous affinity, BPA can disrupt intracellular homeostasis at low concentrations. Its tumor-promoting effects are not solely dependent on classical nuclear estrogen receptors but are achieved through the activation or modulation of multiple signaling pathways. For instance, in models such as triple-negative breast cancer, which lack functional estrogen receptor α (ERα), BPA plays a key role by binding to and activating the G protein-coupled estrogen receptor (GPER, also known as GPR30). Studies confirm that BPA binding to GPER triggers the transactivation of the epidermal growth factor receptor (EGFR), subsequently activating focal adhesion kinase (FAK), Src kinase, and extracellular signal-regulated kinase 2 (ERK2) in sequence.56 This signaling cascade directly regulates cytoskeletal reorganization, focal adhesion dynamics, and cell migratory capacity. Beyond these receptor kinase pathways, recent findings highlight a novel ion-dependent mechanism driving malignancy, where BPA upregulates the voltage-gated potassium channel Kv3.4, which subsequently activates Integrin-β/FAK signaling to further facilitate cell migration.57 Furthermore, activated ERK can phosphorylate downstream transcription factors, such as c-Fos; upon entering the nucleus, these may synergize with nuclear receptor signaling pathways to co-regulate the expression profile of pro-invasive genes.
Simultaneously, BPA activation of estrogen-related receptor gamma (ERRγ) is a central link in its impact on cancer cell energy metabolism and proliferation. BPA can upregulate and activate ERRγ via the ERK1/2 signaling pathway.58 As an orphan nuclear receptor, ERRγ is a master regulator of genes involved in mitochondrial biogenesis and glycolysis. Sustained ERRγ activation induced by BPA drives metabolic reprogramming in tumor cells, providing the necessary bioenergy and biosynthetic precursors for uncontrolled proliferation. Additionally, BPA disrupts normal cell cycle checkpoint function through multiple parallel mechanisms. For example, BPA inhibits microRNA miR-381-3p, relieving its post-transcriptional repression of the downstream target gene PTTG1, leading to aberrantly elevated PTTG1 protein levels and thereby driving G1/S phase transition.59 This mechanism may have additive or synergistic effects with the upregulation of cyclin D1 (cyclin D1) expression via other pathways such as ERα, collectively impairing negative cell cycle regulation.
Crucially, these transient signaling events and functional alterations are often translated into stable genomic alterations via epigenetic reprogramming, explaining the latency between exposure and disease manifestation. As illustrated in Figure 2, these transient signaling events are translated into stable genomic alterations via epigenetic reprogramming. BPA-bound ERα recruits histone methyltransferase EZH2 to target promoters, increasing the repressive H3K27me3 mark.60,61 Furthermore, a novel feedback circuit has been identified wherein BPA-induced ERα activation recruits DNA methyltransferases (DNMTs) to suppress the TET2 promoter. This suppression leads to a reduction in global DNA hydroxymethylation (5-hmC), effectively silencing tumor suppressor genes.62

Molecular mechanisms of BPA-induced epigenetic reprogramming
Transient molecular signals induced by BPA exposure can be long-term “memorized” through changes in the chemical modification state of chromatin, which is a core mechanism providing a mechanistic basis for its delayed biological effects. These epigenetic alterations involve multi-layered modifications of DNA and histones. Receptor complexes activated by BPA can directly or indirectly alter the localization and activity of chromatin-modifying enzymes. Research indicates that BPA can induce the enrichment of histone methyltransferase EZH2 and histone acetyltransferase CBP/p300 at specific genomic loci, including tumor suppressor gene promoters.61 This leads to increased repressive histone mark H3K27me3 and aberrant deposition of the active mark H3K27ac, thereby systematically altering chromatin accessibility. The long non-coding RNA HOTAIR acts as a molecular scaffold, guiding the polycomb repressive complex 2 (PRC2) to perform locus-specific gene silencing, further amplifying the effects of epigenetic regulation.63 Another crucial molecular axis involves the DNA hydroxymethylation pathway. By activating ERα, BPA upregulates the expression or activity of DNMTs, leading to aberrant hypermethylation of the promoter region of the TET dioxygenase 2 (TET2) gene and consequent transcriptional suppression.62 TET2 is a key enzyme catalyzing the conversion of 5-methylcytosine (5 mC) to 5-hydroxymethylcytosine (5hmC). Downregulation of TET2 expression causes a significant decrease in global 5hmC levels. Since 5hmC is not only an intermediate in active DNA demethylation but also frequently associated with actively transcribed chromatin regions, its loss is closely linked to genomic instability and impaired cellular differentiation potential, constituting a critical epigenetic node in BPA-mediated carcinogenesis. Subsequent research has further found that BPA exposure can upregulate the expression of the histone demethylase KDM2A. By removing the activating histone mark H3K36me2, KDM2A can further inhibit the binding and function of TET2 protein on chromatin, exacerbating the loss of DNA hydroxymethylation and forming a self-reinforcing regulatory circuit involving “KDM2A-ERα-DNMT-TET2,” which perpetuates the state of epigenetic dysregulation.64
Together, these interconnected mechanisms establish a molecular framework through which BPA can reprogram the epigenetic landscape of mammary cells. The persistence and long-term consequences of such reprogramming, however, are critically dependent on the developmental timing of exposure, as discussed in Section 4.4.

Systemic remodeling of the tumor microenvironment
The tumor-promoting effects of BPA are not confined to epithelial cells themselves; it can induce extensive and profound remodeling of the local mammary microenvironment, thereby creating a local milieu more conducive to tumor growth, invasion, and metastasis,65 the systemic remodeling process is summarized in Figure 3. In terms of energy metabolism, long-term, low-dose BPA exposure upregulates the expression of peroxisome proliferator-activated receptor gamma coactivator 1-alpha (PGC-1α), thereby promoting mitochondrial biogenesis.66 This metabolic adaptation not only meets the high energy demands of biological processes such as epithelial-mesenchymal transition (EMT), but the concomitant increase in reactive oxygen species (ROS) levels may also exacerbate DNA damage and chronic inflammatory responses. The function of stromal cells also undergoes pathological transformation via multiple signaling axes. BPA can induce the transformation of normal mammary adipocytes into “cancer-associated adipocytes” (CAAs). These transformed adipocytes secrete large amounts of the chemokine CXCL12, which binds to the CXCR4 receptor on breast cancer cells to activate the PI3K/AKT signaling pathway, thereby promoting EMT and cell migration.67 Furthermore, BPA targets cancer-associated fibroblasts (CAFs) by interacting with the GPER.16 This GPER-mediated signaling in CAFs induces the expression of target genes such as c-FOS and CTGF, sustaining fibroblast activation and tumor cell proliferation.16 Simultaneously, BPA skews the immune landscape toward an immunosuppressive and pro-angiogenic state. BPA exposure promotes the polarization of macrophages toward an M2-type (CD206+) phenotype both in vitro and in vivo.68 While inducing M2 polarization, BPA exposure also leads to the upregulation of pro-inflammatory and pro-metastatic cytokines, such as IL-6 and TNF-α, creating a microenvironment that facilitates motility and vascularization.17 BPA exposure can also increase the infiltration of mast cells in mammary tissue; factors released by these cells, such as transforming growth factor β (TGF-β), can directly stimulate tumor progression, fibrosis, and EMT.69At the extracellular matrix (ECM) level, in utero exposure to BPA can lead to a significant increase in collagen deposition, disorganized fiber alignment, and ultimately increased tissue stiffness in the adult mammary stroma.70,71 This alteration in physical properties can, via mechanosensory signaling pathways such as those involving integrins, activate transcriptional co-activators such as YAP/TAZ within tumor cells, directly driving the expression of genes related to proliferation and stemness. Finally, recent evidence indicates that extracellular vesicles (EVs) secreted by BPA-treated tumor cells undergo changes in composition and function.72 These EVs are enriched with specific microRNAs and matrix metalloproteinases, including MMP-9, which can be taken up by recipient cells to alter their behavior, promoting tumor growth in vivo and fostering a microenvironment conducive to metastatic colonization in distant organs.72

Long-term consequences of exposure during critical developmental windows
The mammary gland is particularly vulnerable to disruption by BPA during sensitive developmental stages such as the embryonic, perinatal, and pubertal periods. Exposure during these times can lead to fundamental deviations in mammary gland developmental programming. Exposure to low-dose BPA during fetal life can interfere with key morphogenic signaling pathways such as bone morphogenetic proteins (BMPs) and Wnt, resulting in abnormal postnatal mammary ductal tree morphology, such as increased numbers of terminal end buds (TEBs) and excessive lateral branching. These structural anomalies are themselves recognized biomarkers of breast cancer risk.73 Developmental BPA exposure also directly targets mammary stem cells (MaSCs). Studies demonstrate that pubertal exposure to BPA not only increases the number of MaSCs but also persistently alters their functional properties and gene expression profiles, enabling these cells to spontaneously form early neoplastic lesions upon transplantation.74 Even at physiologically relevant concentrations, BPA can specifically enhance the self-renewal capacity of estrogen receptor-positive breast cancer stem-like cells by inducing the expression of the transcription factor SOX2, which may represent a cellular source for tumor recurrence and drug resistance.75
As detailed in Section 4.2, these effects are driven by stable epigenetic modifications, such as DNA methylation at the TET2 promoter and histone changes via EZH2. The resulting cancer risk depends heavily on the timing of exposure. During “windows of plasticity” such as fetal development and puberty, epigenetic marks are established and stem cell pools expand. BPA exposure can “imprint” abnormal patterns that are passed down through cell divisions. This phenomenon, known as epigenetic memory, persists even after the exposure ends. It creates a latent vulnerability that may only develop into malignancy decades later.
Animal model studies show that gestational exposure to BPA can induce persistent epigenetic reprogramming in the mammary tissue of offspring. This reprogramming involves not only genome-wide alterations in DNA methylation patterns but also exhibits significant gene specificity (e.g., preferential targeting of estrogen-responsive genes), directly contributing to increased susceptibility to mammary tumors in adulthood.76,77 This epigenetic “memory,” induced early in development and stably maintained in somatic cells, constitutes a core mechanism for the perpetuation of BPA’s carcinogenic hazard across generations.60,61 Additionally, based on the Developmental Origins of Health and Disease theory, it is postulated within the scientific community that such epigenetic dysregulation possesses the potential for transmission through the germline and impact on more distant generations (transgenerational inheritance), although direct evidence in the context of breast cancer remains an area for further in-depth investigation.78,79
In summary, the promotion of breast cancer initiation and progression by BPA is a multi-stage process involving multiple systems. It begins with acute disruption of diverse cellular signaling pathways, perpetuates these disturbances through the induction of stable epigenetic changes, and concurrently drives supportive remodeling of the tumor microenvironment. Exposure during critical individual developmental periods can lead to permanent “misprogramming” of mammary tissue, thereby establishing a high risk of tumorigenesis that persists for decades. This integrative mechanistic framework proposes a model for how environmental chemicals could interact with the host through complex biological networks to potentially contribute to malignant disease.

Translational research on BPA exposure and public health risk management
Although epidemiological studies have not yet established a definitive link between BPA exposure and breast cancer incidence, substantial laboratory evidence reveals its potential to interfere with disease progression and clinical management. Experimental models indicate that environmentally relevant concentrations of BPA may compromise the efficacy of standard therapeutic regimens. Therefore, a deeper understanding of these potential interactions is crucial for optimizing clinical treatment strategies and formulating preventive public health policies, even in the absence of conclusive etiological evidence.

Potential mechanisms of therapy resistance in experimental models
Despite heterogeneity in clinical data regarding BPA exposure and patient outcomes, extensive preclinical evidence demonstrates that BPA can diminish the effectiveness of anticancer drugs in model systems by activating non-canonical signaling pathways. These findings primarily focus on two key areas: resistance to endocrine therapy and reduced chemosensitivity.
Firstly, the potential interference of BPA with anti-estrogen therapies, particularly tamoxifen, is a major research focus at the intersection of environmental toxicology and oncology. In vitro studies using estrogen receptor-positive breast cancer cell lines confirm that co-exposure to BPA reduces the anti-proliferative effect of tamoxifen, revealing a potential mechanism of therapy resistance. This action involves cellular reprogramming via the ERRγ axis. Research by Huang et al. showed that BPA exposure upregulates the expression of the orphan nuclear receptor ERRγ and its coactivators PGC-1α/β, thereby establishing an alternative, ERα-independent pathway that sustains cell proliferation and circumvents hormonal blockade.4 Mechanistically, Heckler et al. identified the activation of the MAPK/ERK pathway as a key upstream regulatory event for BPA-induced ERRγ expression, providing important insights into how environmental stressors can alter cellular states to evade drug inhibition.80 Furthermore, BPA-induced activation of human telomerase reverse transcriptase may provide an additional compensatory survival mechanism, further entrenching the resistant phenotype.
Moreover, BPA’s impact on chemotherapy efficacy extends beyond estrogen signaling. In ER-negative models, BPA also demonstrates the potential to weaken the effects of cytotoxic chemotherapeutic agents, indicating a broader biological basis for its actions. One key mechanism involves the rapid activation of pro-survival signaling cascades. LaPensee et al. reported that nanomolar concentrations of BPA rapidly trigger the GPR30/PI3K/AKT axis, leading to the upregulation of anti-apoptotic proteins such as Bcl-2 and Bcl-xL. This mechanism has been validated in cells exposed to DNA-damaging agents such as doxorubicin and cisplatin, providing experimental support for BPA’s role in reducing chemosensitivity in preclinical models.5,81 BPA may also interfere with cytoskeletal dynamics. Lagunas-Rangel et al. observed that even brief pre-treatment with BPA can stabilize microtubule structures.82 This finding suggests that BPA could theoretically diminish the efficacy of antimitotic drugs such as vincristine by antagonizing their microtubule-depolymerizing action. The integrated mechanisms of endocrine and chemoresistance are depicted in Figure 4, Although this mechanism requires further in vivo validation.
Collectively, these preclinical findings provide a compelling mechanistic rationale for BPA as a potential modifier of therapy response. However, it is crucial to emphasize that this evidence is currently derived primarily from in vitro and animal models. While these data are invaluable for hypothesis generation and underscore the need for clinical vigilance, they are not yet sufficient to directly guide changes in therapeutic guidelines. Prospective clinical studies are urgently needed to validate whether real-world BPA exposure measurably impacts treatment outcomes in patients with breast cancer.

Exploratory biomarkers and risk stratification
The mechanistic research raises a core translational medicine question: can specific molecular signatures of BPA exposure be identified in humans for clinical risk prediction? Currently, this field remains in a hypothesis-generating stage. Some exploratory studies aim to correlate long-term pollutant exposure with specific molecular features. For instance, Confortin et al. experimentally demonstrated that long-term exposure to environmentally relevant concentrations of BPA can induce the upregulation of STAT3, VEGFA, and ESR2 in breast cancer cells.51 Notably, the same study found that perfluorooctanoic acid (PFOA), another environmental pollutant, could upregulate the drug efflux transporter ABCG2. These environmentally modulated gene signatures are often associated with more aggressive tumor phenotypes and poor prognosis in pan-cancer analyses. Epigenetic alterations are another research direction. Studies by Romagnolo and Seki indicate that in some retrospective clinical samples, molecular features potentially linked to BPA exposure are associated with larger tumor volume and higher pathological grade. It must be emphasized that these correlative findings do not establish causality.83,84 Their primary value lies in generating scientific hypotheses to guide the design of future prospective cohort studies aimed at directly verifying the “environmental exposure-specific molecular footprint-clinical outcome” chain.

Public health perspective: preventive strategies amid uncertainty
Evolving scientific understanding and evidence regarding the effects of chemical mixtures are driving public health strategies toward a more prevention-oriented approach. This approach acknowledges the potential health risks highlighted by experimental mechanistic research, even while their actual contribution to breast cancer incidence remains epidemiologically complex. Despite these shifts, some regulatory guidelines still rely on monotonic dose-response models.85
Recent updates to regulatory standards exemplify this shift. In 2023, the European Food Safety Authority drastically reduced the tolerable daily intake (TDI) for BPA to 0.2 ng/kg body weight/day, a 20,000-fold decrease from the 2015 standard.86 This revision was based primarily on concerns about potential toxicity to non-carcinogenic endpoints, such as the immune system, rather than traditionally confirmed carcinogenicity data. It reflects a modern risk assessment methodology designed to more comprehensively protect multiple health endpoints. Regulatory action has translated into concrete bans. France serves as a leading example, having prohibited BPA in all food packaging in 2015 and later classifying it as a “substance of very high concern.” This policy, enacted under the precautionary principle despite lingering epidemiological uncertainties, directly targets a major exposure source and provides a real-world model for preventive environmental health policy.
The reality of mixture exposure further complicates risk assessment. Research indicates that when BPA acts in combination with other common pollutants (e.g., parabens and PFOA), it can induce biological effects, such as promoting apoptosis evasion, even when each component is present at a concentration below its individual no-observed-effect level. This “co-activation” phenomenon supports the view that public health policy must increasingly shift toward assessing the “total environmental exposure burden” rather than relying solely on single-chemical risk assessments.
From a strict evidence grading perspective, the current human evidence for a causal role of BPA in breast cancer etiology remains limited or insufficient. However, within the context of widespread exposure, potential irreversible effects during critical developmental windows, and the reality of mixture exposures, public health decision-making can justifiably adopt a more conservative precautionary principle.87
Consequently, in clinical practice and health promotion, it is reasonable and low-risk to advise patients with breast cancer and high-risk individuals to adopt practical measures to reduce BPA exposure (e.g., minimizing use of polycarbonate plastics and thermal paper receipts). This application of the precautionary principle is justified while further evidence accumulates. Additionally, research suggests that certain natural bioactive compounds may counteract the adverse effects of BPA. For example, Jeong reported that the fungal metabolite beauvericin could inhibit BPA-induced breast cancer cell proliferation by modulating the ERα/p38 pathway.88 Li and Wang indicated that curcumin and melatonin, respectively, could reverse the proliferative effects of BPA on breast cancer cells by modulating related signaling axes.89,90 These findings offer potential new avenues for mitigating environmental pollutant-related health risks through dietary or interventional strategies.

Discussion: reconciling mechanistic plausibility with epidemiological inconsistency
The discordance between the well-characterized oncogenic mechanisms of BPA in experimental models and the inconsistent associations reported in human epidemiological studies represents a central challenge in environmental oncology. While mechanistic data provide a robust biological rationale for BPA-induced mammary transformation, translating these findings into population-level evidence requires a critical evaluation of methodological limitations. This section examines three key factors that explain the apparent discordance: (1) fundamental differences in how experimental and epidemiological research are designed, (2) methodological variations in exposure assessment and population heterogeneity, and (3) the critical role of exposure timing during developmental windows.

Bridging paradigm disparities: challenges in translating experimental models to human populations
Experimental studies establish mechanistic pathways under controlled conditions using fixed-dose, single-compound exposures. While this approach has high internal validity, it fails to reflect the reality of chronic, low-level, and mixed exposures in human populations. Humans are simultaneously exposed to a “cocktail” of endocrine disruptors, including BPA, phthalates, and parabens. Studies have confirmed that when BPA is combined with PFOA and preservatives, the mixture can synergistically induce apoptosis evasion and ERα phosphorylation in mammary cells, even when each component is present at concentrations below its individual no-observed-effect level.91 This indicates that risk assessments based on single chemicals may significantly underestimate real-world carcinogenic risk.
Furthermore, although cell-based assays are fundamental for mechanistic exploration, traditional 2D culture models cannot fully replicate the intricate epithelial-stromal-immune interactions inherent in human physiology. BPA’s pro-carcinogenic effects extend beyond direct action on mammary epithelial cells to comprehensive remodeling of the tumor microenvironment, including activating CAFs, promoting M2 macrophage polarization, and driving adipocyte conversion.56,67,68 These systemic responses are difficult to simulate in simplified experimental systems.
In summary, experimental research emphasizes in-depth mechanistic dissection under idealized conditions, while epidemiological research assesses real-world associations subject to confounding and measurement error. Understanding these inherent differences is essential for interpreting the discordance discussed in subsequent sections.

Factors driving heterogeneity in epidemiological findings
The inconsistency in epidemiological findings, spanning from significant positive associations to null and even inverse correlations, is largely attributable to methodological variations in exposure assessment, diversity in population characteristics, insufficient consideration of factors such as genetic susceptibility, and inherent limitations in study design itself.
First, the significant divergence in conclusions among epidemiological studies primarily stems from the choice of biomarkers in exposure assessment, which directly determines the biological relevance of the measured exposure. Upon entering the human body, BPA is rapidly metabolized. Its primary urinary metabolite, BPA-glucuronide, is a hormonally inactive product of hepatic detoxification, and its concentration mainly reflects recent exposure and an individual’s capacity for detoxification and excretion. In contrast, free BPA is the active parent compound capable of interacting with cellular targets such as estrogen receptors, thereby possessing the potential to disrupt endocrine function. Consequently, measuring these different forms tracks fundamentally distinct biological processes. Studies quantifying metabolites, as exemplified by Trabert and colleagues in 2014, assess detoxification and clearance pathways.46 Conversely, research that specifically quantifies the parent compound, such as the study by López-Carrillo et al. in 2021, captures the potential biologically effective dose.19 This distinction is crucial for understanding why investigations using the same biological matrix such as urine yield contradictory results: analyses of metabolites typically report null associations, whereas measurements of the free compound can reveal clear dose-response relationships, such as an odds ratio of 2.31 for the highest versus lowest exposure quartile. This discrepancy underscores that the misclassification of inactive metabolites as relevant exposure markers is a key methodological factor leading to the attenuation or loss of true epidemiological signals.
Selecting appropriate biomarkers requires an integrated understanding of compound toxicokinetics and toxicodynamics. An ideal biomarker should reflect the long-term, cumulative burden within the target organ system, or at least represent the level of exposure to the biologically active form. When direct measurement in target tissues such as breast adipose tissue is not feasible, analyzing free BPA in urine provides a closer approximation of the biologically effective dose compared to its metabolites. However, this approach remains limited by reflecting primarily recent exposure due to the compound’s short half-life. A deeper layer of complexity arises from inter-individual genetic variation in metabolizing enzymes, including those from the UGT and CYP450 families. These polymorphisms can lead to significant differences in internal concentrations of free BPA resulting from identical external exposures. The work by He et al. in 2022 demonstrated this clearly, showing that among individuals with a specific CYP17A1 genotype, high BPA exposure was associated with a significantly elevated breast cancer risk, with an odds ratio of 2.49. Failure to account for such gene-environment interactions in population-level analyses dilutes the observable risk estimate, thereby masking identifiable high-risk subgroups.
Finally, the methodological implications of null and “inverse” findings must be interpreted with caution. Some studies reporting null associations may be limited by small sample sizes. More notably, several large studies not only found no association for BPA but also reported weak inverse or seemingly “protective” associations for other environmental phenols with breast cancer risk. For instance, Wu found a weak inverse association for total paraben exposure,22 and a meta-analysis by Liu showed inverse associations for specific phthalate metabolites MBzP and MiBP.44 Such seemingly paradoxical findings likely stem from reverse causality, inherent in cross-sectional or case-control designs, or from unmeasured confounding factors. Rather than proving the harmlessness of these chemicals, this further underscores the paramount importance of prospective longitudinal cohort studies with repeated biospecimen collection prior to diagnosis. Such designs are essential to overcome the inherent limitations in establishing temporal sequence present in much of the existing research. The role of exposure timing, which is briefly noted above, is examined in detail in Section 6.3 as a primary explanation for the discordance.

Critical exposure windows: explaining the temporal mismatch
The most important explanation for the discordance between laboratory and epidemiological evidence is a temporal mismatch: BPA’s carcinogenic effects depend critically on exposure during developmental windows, but most epidemiological studies measure exposure only in adulthood. The fundamental reason for the inconsistency between epidemiological findings and laboratory evidence lies in the fact that existing epidemiological designs often fail to effectively assess exposure levels during critical windows of susceptibility. Basic research and animal models collectively demonstrate that BPA acts not only as an immediate endocrine disruptor but also through long-term developmental programming effects induced during specific stages of growth.
As discussed in Sections 4.2 and 4.4, BPA exposure during fetal development or puberty induces persistent epigenetic alterations—including DNA methylation changes at the TET2 promoter and sustained upregulation of EZH2 and HOTAIR—that reprogram mammary tissue and create latent cancer risk. These alterations persist for decades after exposure ceases.
However, most epidemiological studies measure BPA exposure in adulthood using single spot urine or blood samples. BPA has a half-life of less than 6 h, so these measurements reflect recent exposure only. They cannot capture exposures during critical developmental windows that may have initiated carcinogenic processes years or decades earlier. This temporal disconnect biases results toward the null.
Studies that assess exposure during developmental windows support this explanation. Binder found a U-shaped association between urinary BPA and fibroglandular volume in adolescent girls—a pattern consistent with endocrine disruption that would be missed in adult studies.48 In contrast, large adult cohorts such as EPIC-Spain and the multiethnic cohort reported null associations using adult spot samples.22,23 This pattern across studies suggests that when exposure measurement aligns with biologically relevant windows, associations emerge.
Second, developmental BPA exposure disrupts MaSC homeostasis. Pubertal exposure increases the number of MaSCs and causes delayed expansion of luminal progenitors in the adult gland.74 This stem cell reprogramming increases susceptibility to subsequent carcinogenic challenges. Epidemiological studies focused on postmenopausal women22,46 cannot capture these early developmental alterations.22,46
Furthermore, the long-term remodeling of the mammary stromal microenvironment by BPA cannot be overlooked. Fetal exposure to BPA affects the reciprocal interaction between stromal and epithelial cells, leading to a significant increase in tissue stiffness by altering the composition of the ECM, such as increasing collagen deposition.70 Increased tissue stiffness is a well-established physical risk factor for breast cancer. Although Sprague et al. found a positive association between serum BPA and mammographic breast density in postmenopausal women.20 In contrast to the association observed in adults, a longitudinal study of adolescent girls revealed a non-monotonic U-shaped curve: Compared to the lowest tertile, girls in the middle tertile of urinary BPA exposure showed an approximately 10% reduction in fibroglandular volume, while the highest tertile rebounded to levels similar to the low-exposure group.48 This pattern underscores the complex, life-stage-dependent effects of endocrine disruptors. Together, these findings demonstrate that the timing of exposure measurement is a critical determinant of whether epidemiological studies detect associations.
The prevalent null associations in epidemiological studies do not imply a lack of biological effect for BPA but rather reflect a lack of sufficient resolution in the temporal dimension of current research. Since carcinogenic initiation may occur in early life while clinical onset is often delayed by decades, relying solely on a single sample measurement in adulthood makes it extremely difficult to accurately reconstruct the complex carcinogenic process driven by developmental programming. Future epidemiological research urgently requires prospective designs based on a life-course approach.

Genomic evidence: Validating pathways, not chemical exposure
Recent genomic studies offer a complementary approach to traditional observational designs, though careful interpretation is required. The 2025 Mendelian randomization study by Hong et al. utilized genetically predicted gene expression to suggest a potential causal association between the expression of BPA-responsive genes, such as TET2, and the risk of ER-positive breast cancer.50
It is important to clarify what this study demonstrates. Hong et al. did not use circulating BPA levels as the exposure; they used genetic variants as instruments for the expression of BPA-targeted genes. Therefore, this evidence supports a causal role for BPA-responsive biological pathways in breast cancer, not direct proof that BPA exposure itself causes the disease.
Nevertheless, this finding contributes to reconciling the discordance. By demonstrating at the population level that genes disrupted by BPA in experimental models are causally linked to breast cancer risk, Hong et al. provide genetic validation of the mechanistic framework developed in Section 4.50 This suggests that individuals with genetic variants affecting these pathways may be particularly susceptible to BPA’s effects, offering one explanation for population heterogeneity.
In summary, the discordance between mechanistic and epidemiological evidence underscores that evaluating BPA’s risk requires moving beyond simple, single-time-point exposure metrics. To address the methodological gaps identified here, future research must prioritize longitudinal studies with biospecimen collection spanning susceptible life stages. Studies should quantify free BPA rather than total BPA or metabolites, ideally in matrices reflecting long-term tissue burden. Genetic data should be integrated to identify susceptible subgroups and account for gene-environment interactions. From a public health perspective, null findings from adult studies do not indicate that BPA is safe; they reflect the profound methodological challenges of studying a chemical whose effects depend critically on exposure timing.

Conclusion
BPA exemplifies the complexity of contemporary environmental carcinogens. It does not function as a classic toxicant with a linear dose-response relationship, but rather as a conditional risk modifier whose effects depend on the timing of exposure, individual genetic susceptibility, and the context of coexposure with other pollutants. The apparent discordance in evidence can be reconciled: rigorous laboratory investigations have elucidated the biological plausibility for BPA to promote breast cancer pathogenesis at molecular, cellular, and animal model levels, whereas inconsistencies in human epidemiological findings primarily stem from methodological limitations inherent to population studies, particularly the challenge of accurately assessing long-term exposures, especially during critical early-life windows.
The consistent identification of oncogenic pathways in vitro and in vivo—including receptor-mediated signaling, induction of epigenetic alterations, and remodeling of the tumor microenvironment—collectively provides an experimental foundation for understanding BPA’s potential pathogenic role. The frequent null associations observed in population studies may partly reflect limitations in current exposure assessment methodologies and population heterogeneity, rather than an absence of biological effect. Emerging genetic epidemiological approaches, such as Mendelian randomization analyses linking BPA-target gene expression to breast cancer risk, are beginning to bridge this gap by providing genetic evidence that indirectly supports the carcinogenic potential of BPA-perturbed pathways.
Consequently, advancing the field requires progress along two critical fronts. In research, the priority must be establishing longitudinal cohorts that span from early life to adulthood, incorporating biospecimens reflective of long-term tissue burden alongside genetic data for integrated analysis. In public health and risk assessment, frameworks must evolve to evaluate the combined effects of pollutant mixtures and non-monotonic dose responses, guiding precautionary actions to reduce exposure. Given the elucidated biological mechanisms of BPA in experimental models and its implication in potentially compromising therapeutic efficacy, prudent measures to minimize exposure in the population, particularly during sensitive periods, are warranted. Ultimately, elucidating the relationship between BPA and breast cancer is not only a scientific necessity but also a foundational step toward developing more precise and effective prevention strategies for environmentally influenced cancers.

Data availability
All data used are from public datasets cited in the references.

Acknowledgments
The authors acknowledge financial support from the Jilin Provincial Department of Science and Technology (Grant No. 20230401091YY). We would also like to thank Home for Researchers (www.home-for-researchers.com) and Figdraw (https://www.figdraw.com/) for their support, as we used Figdraw, which belongs to Home for Researchers, to draw figures in this review.

Author contributions
Ying Liu: writing – original draft, visualization, data curation. hongliang cao: writing – original draft, visualization, and conceptualization. Shuaiyang Zhang: visualization and data curation. Qianhui Li: supervision. Xin Wang: supervision. Mingyue Ma: writing – review and editing. Yang Zhou: visualization. Bing Han: writing – review and editing, visualization, supervision, and conceptualization.

Declaration of interests
The authors declare no competing interests.

Declaration of generative AI and AI-assisted technologies in the writing process
During the preparation of this work the authors used Google Gemini in order to assist with language translation and polishing to improve readability. After using this tool, the authors reviewed and edited the content as needed and take full responsibility for the content of the published article.