Endocrine‐disrupting compounds and metabolomic reprogramming in breast cancer

Abstract

Endocrine‐disrupting chemicals pose a growing threat to human health through their increasing presence in the environment and their potential interactions with the mammalian endocrine systems. Due to their structural similarity to hormones like estrogen, these chemicals can interfere with endocrine signaling, leading to many deleterious effects. Exposure to estrogenic endocrine‐disrupting compounds (EDC) is a suggested risk factor for the development of breast cancer, one of the most frequently diagnosed cancers in women. However, the mechanisms through which EDCs contribute to breast cancer development remain elusive. To rapidly proliferate, cancer cells undertake distinct metabolic programs to utilize existing nutrients in the tumor microenvironment and synthesize macromolecules de novo. EDCs are known to dysregulate cell signaling pathways related to cellular metabolism, which may be an important mechanism through which they exert their cancer‐promoting effects. These altered pathways can be studied via metabolomic analysis, a new advancement in ‐omics technologies that can interrogate molecular pathways that favor cancer development and progression. This review will summarize recent discoveries regarding EDCs and the metabolic reprogramming that they may induce to facilitate the development of breast cancer.

Abbreviations

AhR

aryl‐hydrocarbon receptor

BBP

benzyl butyl phthalate

BPA

bisphenol A

BPF

bisphenol F

BPS

bisphenol S

CDP

cytidine diphosphate

CTP

cytidine triphosphate

DBP

dibutyl phthalate

DDT

dichlorodiphenyltrichloroethane

DM

diabetes mellitus

EDCs

endocrine‐disrupting compounds

ER

estrogen receptor

MS

mass spectrometry

NHANES

National Health and Nutrition Examination Survey

NMR

nuclear magnetic resonance

PCBs

polychlorinated biphenyls

PFAA

plasma‐free amino acid

PFAS

per‐ and polyfluoroalkyl substances

RALs

resorcyclic acid lactones

ROS

reactive oxygen species

TB‐BPA

tetrabromo‐BPA

TC‐BPA

tetrachloro‐BPA

ZEN

zearalenone

1. INTRODUCTION

Outside of skin cancers, breast cancer is the most common cancer diagnosed in women.^{[ 1} , ^{2 ]} Only a small portion of breast cancers are linked to known hereditary mutations; most breast cancers are the result of unknown factors.^{[ 3} , ^{4 ]} It is also clear that the etiology of breast cancer is multifactorial. Both the incidence of breast cancers in the United States and the use of endocrine‐disrupting compounds (EDCs) in consumer‐grade products, as well as their prevalence in the environment, rose drastically in the last half‐century.^{[ 5 ]} The increase in breast cancer diagnoses could at least in part be due to the increased exposure to various EDCs.

It is the estrogenic effect of EDCs that most likely facilitates the development of breast cancer. The most notable and well‐studied estrogenic EDCs include phthalates, polychlorinated biphenyls (PCBs), and bisphenol A (BPA).^{[ 6 ]} However, there are many EDCs that have not garnered attention until recently, such as the mycotoxins produced by crop‐contaminating fungi^{[ 7 ]} and phytoestrogens produced by plants.^{[ 8 ]} These EDCs have demonstrated cause for concern due to their binding affinity and activation of the estrogen receptors (ERs).^{[ 9} , ^{10 ]} Mechanistically, the downstream estrogen signaling induced by estrogenic EDCs and the related increase in cancer cell proliferation, invasion, and metastasis have been well‐established.^{[ 11} , ¹² , ¹³ , ^{14 ]} These chemicals may ultimately result in elevated mammary proliferation and breast cancer.

Metabolomics utilizes computational biochemistry and bioinformatics to map the ratios of chemicals present in a sample. In biological applications, this can uncover novel mechanisms of complex diseases such as cancer. The metabolome has been a subject of cancer research for decades. In the late 1920s, Warburg described a metabolic shift in cancer cells from complete oxidation of glucose to carbon dioxide to aerobic glycolysis. This pathway results in the conversion of glucose into lactate, instead of full oxidation.^{[ 15 ]} This phenomenon was later titled “the Warburg effect.” Since Warburg's initial discovery, countless researchers have shown that this transition, key to rapid cell growth and proliferation, is not restricted to cancer cells. In fact, budding yeast,^{[ 16 ]} rapidly developing tissues,^{[ 17 ]} and even replicating pathogens^{[ 18 ]} use this metabolic reorientation to sustain rapid growth. Many distinct rationales have been proposed for why cancer cells use Warburg's metabolism to grow, including speed of energy retrieval, involvement of the tumor microenvironment, and the production of carbon skeletons for biomolecule synthesis.^{[ 19} , ²⁰ , ^{21 ]} Each explanation fails to provide sufficient conceptual evidence for a complete explanation of why cancer cells invariably use Warburg's metabolism to grow. Since the advent of ‐omics technologies, newfound interest has been placed on the subject of metabolomic pathways in cancer. In addition to Warburg's phenomenon, scientists have uncovered many more hallmarks of cancer cell metabolism, such as alterations to mitochondrial metabolic pathways, lipid turnover, and nucleotide biosynthesis.^{[ 22} , ²³ , ^{24 ]} Even still, a consensus on how these distinct metabolomic programs aid cancer proliferation, survival, and resistance to treatment remains elusive.^{[ 25} , ^{26 ]}

To interpret how cancer cells shift their nutrient utilization, it is paramount to also consider the metabolites that are present within the tumor microenvironment.^{[ 27 ]} This extrinsic factor can influence the nutrients available to cancer cells, leading to an increase or decrease in overall survival. Heterogeneity in the tumor microenvironment among patients with cancer is largely dictated by the metabolomic state of the individual, in addition to the organ in which the cancer resides. Thus, metabolomic considerations must also use a whole‐body, organ‐system‐encompassing perspective. It is well established that pathological metabolic dysregulation is closely associated with cancer risk.^{[ 28 ]} Type II diabetes mellitus (DM) has been epidemiologically associated with the lifetime diagnosis of many cancers, including liver, pancreatic, and breast cancers.^{[ 29 ]} Mechanistically, this association is thought to be related to the metabolic shifts that occur as a result of type II DM, such as hyperlipidemia, hyperinsulinemia, and hyperglycemia.^{[ 30 ]} Higher levels of lipids, insulin, and glucose are all strongly associated with cancer promotion.^{[ 31 ]} Another hypothesis, which is not mutually exclusive from the former, proposes that the reactive oxygen species (ROS) and disturbance to antioxidant protection well‐associated with metabolic diseases may also favor cancer progression.^{[ 32 ]} Thus, the global metabolome may also be a key factor in facilitating cancer proliferation and progression, in addition to intrinsic cancer cell metabolism pathways.

2. METABOLOMICS AS A TECHNIQUE TO IDENTIFY CELLULAR CHANGES

Metabolomics offers an accurate and physiologically relevant perspective on how the body is responding to a given stimulus.^{[ 33 ]} It offers insight into the molecular pathways that are followed after the cellular “central dogma” of DNA, RNA, and protein signaling is altered. Integration of metabolomics with other ‐omics techniques, such as transcriptomics, can reveal novel pathways in response to perturbations. To validate metabolomics as a tool to assess downstream molecular pathways, studies have demonstrated that the changes to the metabolomic signature correlate to transcriptomic changes in biological systems. This has been demonstrated in breast cancer,^{[ 34} , ^{35 ]} cervical cancer,^{[ 36 ]} prostate cancer,^{[ 37 ]} and pancreatic cancer.^{[ 38 ]} Thus, metabolomic analysis appears to be a valid and consistent approach to measuring the biological mechanisms that contribute to the initiation and promotion of cancer.

Metabolomic analysis can be performed on many sample types. In human studies, this approach can be used to assess the ratios of metabolites in serum or in tumor explants. Metabolomics can also be applied to samples collected from in vivo animal models. Metabolites can also be extracted from in vitro cell models to identify intrinsic cellular metabolic pathways, or from the extracellular media to assess what compounds are being produced by these colonies. However, using in vitro cell culture for metabolomic analysis has severe limitations. Cell metabolism is highly sensitive and subject to rapid changes following any type of perturbation. Even the process of extracting metabolites from cell culture can produce artifacts by introducing stress‐based metabolomic responses.^{[ 39 ]} Additionally, there are stark differences between how cells perform in cell culture conditions versus in physiological conditions. While they have their limitations, in vitro studies are increasingly popular, especially after the growing sentiment within the scientific community to abide by the 3Rs and refrain from unnecessary animal use. In vitro studies are also quick, inexpensive, and excellent for generating hypotheses and screening compounds. However, results from cell culture studies must be validated in animal and/or human studies.

There are several detection techniques that can be utilized to identify the ratios of chemicals within a sample. Nuclear magnetic resonance (NMR) spectroscopy and mass spectrometry (MS) are among the most common approaches utilized to analyze the metabolites within a given sample.^{[ 40 ]} NMR spectroscopy utilizes the principles of nuclear spin and magnetism. When an external magnetic field is applied to a chemical, the electrically charged nuclei of each atom will respond by spinning in a direction that correlates with its chemical environment.^{[ 41 ]} The most commonly used nucleus in NMR analysis is the hydrogen proton (¹H). The different resonant frequencies of each hydrogen atom in a compound after the application of a magnetic field are recorded and plotted in spectra.^{[ 41 ]} The peaks within an NMR spectrum can be analyzed to determine the structure of the original compound.^{[ 41 ]} NMR analysis is very reproducible, but not as sensitive as MS analysis. MS analysis utilizes ionization to add electrical charges to the sample material. Then, an ion detector separates these ions based on their mass‐to‐charge ratio (m/z) and quantifies their abundance within the sample.^{[ 42 ]} The resulting spectra can be analyzed to determine the chemical structure. MS can detect a greater quantity of individual metabolites, though more complex sample preparation is required.^{[ 43 ]} The metabolomic analysis generated from both NMR and MS can be targeted or untargeted. The former involves assessing known biomarkers for which identified NMR or MS peaks have been established. The latter involves a more thorough approach by looking at all analytes in a sample, even if unknown, yielding a vast data set and potential for novel discoveries.^{[ 44 ]} After data have been collected, a set of intricate statistical tools is available to interpret the data set. There are two broad classes of analysis that researchers use when assessing data: supervised analysis and unsupervised analysis.^{[ 45 ]} The supervised analysis uses preinputted patterns, such as known cellular molecular pathways, to uncover correlations from the data set. The unsupervised analysis draws patterns from the data set without any preinput knowledge, with all of the data points treated as equal variables. This form of analysis is often used on preliminary exploratory data sets and is more useful on smaller data‐sets.^{[ 46 ]}

3. METABOLOMICS AND BREAST CANCER

Many metabolomic shifts have been identified in breast cancer.^{[ 47} , ⁴⁸ , ^{49 ]} One study found that breast cancer cells have marked changes in choline and fatty acid metabolism in vitro.^{[ 50 ]} It has also been shown that residual breast cancer cells from primary tumors of human patients undergo distinct changes to fatty acid synthesis to survive with limited nutrients.^{[ 51 ]} By increasing free fatty acid stores, cancer cells can survive without nutrients from the vasculature, especially before angiogenesis occurs. These changes are thought to reflect the high energy requirement of cancer cells. Of additional importance are the metabolic trends that occur throughout disease progression.^{[ 52 ]} As breast cancer develops, cancer cells face numerous diverse challenges and must adapt to survive in their ever‐changing environment. Thus, metabolism changes reflect how these cells are altering their energy sources and utilization. In a retrospective study of patients after surgery, the metabolic spectra from early‐ and late‐stage breast cancer were clearly distinguishable. Serums from patients with metastatic breast cancer had lower levels of histidine and higher levels of glucose and lipids. A model developed based on these changes was able to identify early, operable breast cancers with a histological grade of 3 versus metastatic breast cancer 83.7% of the time.^{[ 53 ]} It is important to note that this study investigated the global metabolome of study participants by assessing their blood serum, rather than the metabolome of individual breast cancer cells. The cause of the observed serum metabolic profiles is unknown; the increased metabolite concentrations may be a result of metastatic breast cancer, or a reflection of a global metabolic state that may have influenced the metastasis of the breast cancer. Sufficient data is lacking to confirm either hypothesis. Future studies should also include metabolic disorders like obesity and type II DM as co‐variables to assess how these pathologies may influence the global metabolome and risk of breast cancer metastasis. However, this study does highlight the value of metabolomics not only as a mechanistic tool but also as a potential diagnostic one. Biomarkers may be able to distinguish the disease prognosis of breast cancer patients. In fact, a well‐documented metabolomic biomarker used to screen cancer patients is plasma‐free amino acid (PFAA) profiling.^{[ 54 ]} A study that assessed the PFAA profile of breast cancer patients found that amino acid concentrations in serum differed depending on their tumor node metastasis clinical stage.^{[ 55 ]} These studies largely focused on outcome prediction in diagnosed breast cancer cases based on serum metabolites. These alterations may be useful to diagnose breast cancer and determine prognosis in a less invasive manner. However, they do not yet point toward any identifiable metabolomic mechanism for the initiation and progression of breast cancer. Many more human and in vivo studies are required to elucidate the role metabolism plays in breast cancer, and how this can be influenced by exposure to toxic chemicals.

4. EDCs AND THE BREAST CANCER METABOLOME

“Endocrine disruptors or EDCs” are defined as any chemical, naturally occurring or manmade, that mimics endogenous hormones involved in the endocrine system. They most commonly disrupt endocrine functions by interacting with hormone receptors. In recent years, EDCs have become increasingly prevalent due to environmental contamination and industrial applications.^{[ 56 ]} The most common EDCs agonize the ER due to their chemical similarity to the hormone β‐estradiol^{[ 57 ]} (Figure 1). This structural similarity confers estrogenic EDCs with the ability to signal through estrogen pathways as both agonists and antagonists.^{[ 58 ]} Direct and indirect interaction with hormone signaling pathways via exposure to EDCs has been linked to many human diseases, including breast cancer, thyroid disease, metabolomic disorders, and reproductive pathologies.^{[ 59} , ⁶⁰ , ⁶¹ , ⁶² , ⁶³ , ⁶⁴ , ⁶⁵ , ⁶⁶ , ⁶⁷ , ^{68 ]} EDCs can also disrupt the metabolome and have been associated with metabolomic disorders like type II DM, obesity, and fatty liver disease.^{[ 69 ]} However, the effects of EDCs on the metabolome of breast cancer are less clear (Table 1). It is important to identify how EDCs may facilitate breast cancer development so that proper risk assessment can be performed. Additionally, this research could be used to develop therapeutic intervention techniques and/or chemoprevention measures for a deadly disease. This review will describe the current state of knowledge of how various EDCs alter the global metabolome and the metabolome of breast cancer.

Figure 1.

Chemical structures of environmentally relevant estrogenic endocrine disrupting compounds. The structures of estrogen (beta‐estradiol), bisphenol‐A, di(2‐ethylhexyl)phthalate, alpha‐zeranol, and polychlorinated biphenyls are shown.

Table 1.

Metabolomic alterations induced by relevant environmental EDCs.

Type of study	Principle findings	Reference
PCBs
Human	PCBs were paramount in altering fatty acid biosynthesis pathways	[70]
In vitro/in vivo	The PCB metabolite PCB29‐pQ induced aerobic glycolysis and GLUT1 upregulation in cell culture, resulting in metastasis in a xenograft model	[71]
In vivo	PCB exposure increased the metastatic potential of xenografts	[72]
Bisphenols
In vitro	BPA altered apolipoproteins, and urea cycle, and Krebs cycle intermediates, resulting in amino acid concentration changes in HepG2 cells.	[73]
In vitro	BPA impacted nucleotide synthesis in MCF‐7 cells	[74]
In vitro	In a “leave one out” approach, BPA was shown to be necessary for inducing changes in purine and pyrimidine metabolism in MCF‐7	[75]
In vitro	BPS dysregulated the citric acid cycle, purine metabolism, and lipid metabolism in MCF‐10A cells.	[76]
In vitro	TB‐BPA and TC‐BPA induced glycolytic intermediates and altered glutathione metabolism in MCF10A cells.	[77]
In vivo	BPF‐induced changes in glutathione biosynthesis, glycerophospholipid, and glycerolipid degradation, and glycolysis in MDA‐MB‐231 xenografted mice	[78]
In vivo	BPA changed the global metabolome in the whole body of CD‐1 mouse pups exposed perinatally.	[79]
Phthalates
Human	Phthalate metabolites in the serum of pregnant women were correlated with lipid dysregulation	[80]
In vivo	DBP and BBP increased xenografted MDA‐MB‐231 tumor size	[14]
Zeranols
Humans	Exposure to ZEN and its metabolites was associated with an increased incidence of breast cancer in Tunisian women[ 81 ]	[81]
In vitro	α‐ZAL induced changes to protein biosynthesis, the urea cycle, methionine metabolism, and arginine/proline metabolism in MCF‐7 cells	[82]
In vivo	α‐ZAL increased mammary proliferation in a rat model	[83]

Abbreviations: BBP, benzyl butyl phthalate; BPA, bisphenol A; BPF, bisphenol F; DBP, dibutyl phthalate; EDC, endocrine‐disrupting compounds; PCB, polychlorinated biphenyl; TB‐BPA, tetrabromo‐BPA; TC‐BPA, tetrachloro‐BPA; ZAL, zearalanol; ZEN, zearalenone.

4.1. PCBs

PCBs are a family of over 200 congener chemicals with similar chemical structures, featuring biphenyl moieties with chlorine groups variably attached to the carbon ring.^{[ 84 ]} They were used in the manufacture of electrical conductors and fluid coolants until 1979, when their use was banned in the United States due to their well‐known carcinogenic properties^{[ 85 ]} (Figure 2). However, human exposure still occurs due to environmental persistence and lipid bioaccumulation,^{[ 86 ]} and they are still found consistently in National Health and Nutrition Examination Survey (NHANES) biomonitoring studies.^{[ 87 ]} Many PCB congeners have been linked with breast cancer in both retrospective and prospective epidemiology studies.^{[ 88} , ⁸⁹ , ^{90 ]} In vivo studies, though few in quantity, also corroborate the role of PCB exposure in breast cancer development (Figure 3). Exposure to PCBs increased the metastatic potential of MDA‐MB‐231 xenografts in nu/nu mice, evidenced by an increase in the number and size of detectable metastases in distal organs.^{[ 72 ]} Interestingly, MDA‐MB‐231 is derived from a triple‐negative breast adenocarcinoma, meaning it does not express the classical ER alpha. This observation may be an indirect effect of general immunosuppression, as PCB is suggested to impair immune surveillance and response.^{[ 91 ]} It is unclear what specific mechanisms may be responsible for the observed increase in metastatic potential seen in this study. In vitro studies have demonstrated that some PCBs have estrogenic effects in ER‐positive breast cancer cell lines. A study found that of 17 PCB congeners tested, only one increased estrogen signaling activity in an MCF‐7 reporter assay, and that one congener actually antagonized estrogen signaling. This study highlights the importance of studying PCB congeners individually, instead of grouping structurally different compounds together. In another study, four structurally diverse PCB congeners increased MCF‐7 cell growth. Additionally, co‐treatment with the estrogen antagonist hydroxytamoxifen reversed the observed cell proliferation, suggesting the effect was caused by estrogen signaling.^{[ 92 ]} PCB metabolites have been demonstrated as mutagens, though this requires biotransformation in the liver, where the DNA damage has been predominantly identified.^{[ 93 ]} Taken together, these data do suggest that PCBs may play a role in breast cancer development in human, animal, and cell culture studies. Both estrogen‐dependent and ‐independent mechanisms of cancer promotion have been suggested.

Figure 2.

Environmental sources of endocrine disrupting compounds (EDCs). Humans are exposed to environmental EDCs from various sources. Historically, polychlorinated biphenyls (PCBs) were added to electrical conductors and fluid coolants. Phthalates are included in polyvinyl chloride (PVC) pipes and soft, malleable plastics, like food bagging. Bisphenols are incorporated into hard, stiff plastics, like water bottles. Zeranols are produced by the Fusarium fungus, a common contaminant of agricultural crops like corn and maize.^{[ 94} , ⁹⁵ , ⁹⁶ , ^{97 ]} Human EDC exposure levels are shown.

Figure 3.

The role of endocrine disrupting compounds (EDCs) throughout breast cancer initiation, progression, and metastasis. EDCs likely have diverse mechanisms throughout carcinogenesis.^{[ 14} , ⁷² , ⁹⁸ , ⁹⁹ , ¹⁰⁰ , ¹⁰¹ , ¹⁰² , ¹⁰³ , ¹⁰⁴ , ¹⁰⁵ , ^{106 ]} AhR, aryl‐hydrocarbon receptor; AKT, protein kinase B; ERK 1/2, extracellular signal‐regulated kinase 1/2; IL‐19, interleukin 19; PCB, polychlorinated biphenyls; PI3K, phosphoinositide 3‐kinases; ROS, reactive oxygen species.

Less is known about how PCBs may affect cancer cell metabolism. A recent study found that a PCB quinone‐type metabolite, PCB29‐pQ, significantly increased aerobic glycolysis in MDA‐MB‐231 cells, evidenced by an increase in the mRNA expression of aerobic glycolysis enzymes.^{[ 71 ]} Additionally, the expression of GLUT1, a key transporter involved in glucose uptake known to be overexpressed in breast cancer, was significantly induced by treatment with PCB29‐pQ. This overexpression was suggested to activate the GLUT1/integrin B1/Src/FAK pathway, which is crucial for the epithelial–mesenchymal transition, extravasation, and ultimate metastasis.^{[ 107} , ^{108 ]} In an in vivo LUC‐1 nu/nu mice xenograft study, inhibition of GLUT1 ameliorated the metastatic effect of PCB29‐pQ, demonstrating a clear role for aerobic glycolysis and GLUT1 in the increased progression and invasive potential of breast cancer.^{[ 71 ]} Overall, these results demonstrate that PCBs may affect the metabolism of breast cancer cells, specifically by increasing energy production via glycolysis (Figure 4). The role of the ER is unclear, as no studies to date have been performed on the effects of PCBs on ER‐positive cell lines.

Figure 4.

Major sites of action and molecular effects of endocrine disruptor exposure. EDCs can work through diverse molecular mechanisms upon arrival at a target site, though some pathways are conserved across various types of compounds. Bisphenols and PCBs both initiate aerobic glycolysis and lipid biosynthesis within the mammary gland. Phthalates and PCBs can both activate AhR in different target tissues. Other pathways, like nucleotide synthesis, are specific to certain compounds. AhR, aryl‐hydrocarbon receptor; EDC, endocrine‐disrupting compounds; ER, estrogen receptor; HDAC, histone deacetylases; PCB, polychlorinated biphenyls.

PCBs may also interact with global metabolism, in addition to directly affecting the metabolome of breast cancer cells. PCBs have been shown to induce lipid metabolism dysregulation, leading to chronic metabolic disorders like obesity and type II DM.^{[ 109 ]} One study used detailed community mapping techniques to connect the metabolome of the serum samples taken from study participants in the 1960s. The serum was used recently to assess exposure to many different environmental chemicals.^{[ 70 ]} They found that pathways involving fatty acid biosynthesis and fatty acid activation were enriched in direct association with PCB concentration in blood serum. Additionally, the study found that these pathways were significantly altered when the exposome of these individuals involved other EDCs, namely dichlorodiphenyltrichloroethane (DDT) and per‐ and polyfluoroalkyl substances (PFAS).^{[ 70 ]} This suggests that EDCs may have a synergistic effect on the global metabolome. Likely, these EDCs affect fatty acid biosynthesis and activation through the aryl‐hydrocarbon receptor (AhR), pregnane X receptor, and peroxisome proliferator‐activated receptor nuclear transcription signaling.^{[ 109 ]} Lipophilic EDCs like PCBs, DDT, and PFAS can interact with these nuclear transcription factors, resulting in the activation of several lipid metabolizing enzymes, fatty acid uptake proteins, and lipid biosynthesis enzymes.^{[ 109} , ¹¹⁰ , ^{111 ]} Thus, EDCs may initiate common downstream metabolic pathways that facilitate the development and progression of metabolic disease, a known risk factor for breast cancer. Lipid dysregulation is a risk factor for the development of several cancers, such as prostate cancer,^{[ 112 ]} colon cancer,^{[ 113 ]} and pancreatic cancer.^{[ 114 ]} Additionally, altered lipid metabolism has been associated with breast cancer incidence.^{[ 115 ]} Several studies have found that circulating free fatty acids are associated with breast cancer incidence epidemiologically, and can increase breast cancer proliferation and tumorigenicity in vitro.^{[ 116} , ^{117 ]} Thus, the properties of PCBs as a global lipid metabolic disruptor and a local promoter of cancer metabolism may both be responsible for facilitating their carcinogenic properties.

Understanding how PCBs increase breast cancer risk is still incomplete, and more information is needed on how PCBs interact with both cancer cell and whole‐body metabolism. It is difficult to take into account all of the exposure parameters that may affect toxicological endpoints in epidemiology studies. Exposure concentration and frequency, age at exposure, secondary health factors, and genetic polymorphisms all may affect susceptibility to PCB‐induced metabolomic dysfunction and/or breast cancer. It is highly likely that PCBs, like most EDCs, have more potent effects during the windows of susceptibility, such as early in development.^{[ 118 ]} Late‐life sampling of PCBs has been shown to poorly recapitulate exposures during early life.^{[ 119 ]} Early‐life exposures to EDCs, mainly in utero and postnatal via lactation, may reprogram both the metabolome and breast development to prime the breast for developing cancer during adulthood.^{[ 120 ]} This could also interact with other known risk factors of breast cancer, like genetic polymorphisms, diet, and lifestyle in a synergistic fashion. Thus, more in vivo studies focusing on early‐life exposures to PCBs are needed, in addition to studies assessing the metabolic axes as a potential mechanism for carcinogenesis and promotion.

4.2. Bisphenols

Bisphenols are a group of industrial chemicals used as plasticizers.^{[ 121 ]} They are incorporated into almost all plastic consumer products and are consequently environmentally ubiquitous ^{[ 122 ]} (Figure 1). The most notorious of this class is BPA. While BPA has been banned in most countries due to its well‐known toxicity as an endocrine disruptor, it is being replaced with other bisphenols, like bisphenol F (BPF), bisphenol S (BPS), tetrabromo‐BPA (TB‐BPA), and tetrachloro‐BPA (TC‐BPA).^{[ 123 ]} However, these replacement bisphenols are all structurally similar to BPA; in some cases, they differ by only one functional group. It is thus no surprise that these new bisphenols share many of the endocrine‐disrupting properties of BPA.^{[ 124 ]}

Many studies have suggested the carcinogenic potential of BPA, but the mechanisms are unclear. In Sprague–Dawley rats, in utero exposure to low doses of BPA resulted in neoplastic lesions and carcinomas in offspring.^{[ 125 ]} In vivo studies suggest that BPA plays a role in increasing susceptibility to breast cancer (Figure 3). Since breast cancer is multifactorial, many factors may play a role in the initiation and promotion of the disease. Thus, BPA may act synergistically with other initiating events to promote cancer growth. One study found that in utero exposure to BPA increased the risk of developing 7,12‐dimethylbenz(a)anthracene‐induced tumors in FVB/N mice. This same study found that BPA also acts as a growth promoter when administered to nu/nu mice xenografted with MCF‐7 cells.^{[ 126 ]} This xenograft model is limited by the fact that the cells, which were derived from an already transformed cancer line, were injected into the flank of the mice, and not within the mammary gland itself. Direct extrapolation of these data to humans is difficult, as there are numerous species differences between the human breast and the rodent mammary gland. Rodent mammary glands are simpler in structure and do not include the heterogeneous collagen‐rich stroma that the human breast contains.^{[ 127 ]} This can complicate the study of in vivo mammary cancers in rodent models. BPA has been shown to alter the mammary gland development of rhesus monkeys when exposed perinatally. Ultimately, this exposure resulted in significant increases in mammary bud density and complexity but did not cause overt carcinogenic transformation.^{[ 128 ]} This effect was comparable to the effects of BPA on the mammary gland seen in rodents, supporting the validity of studies performed in rodents.

Metabolomic analysis has been employed to assess bisphenol‐induced breast cancer both in vitro and in vivo. In vitro studies have severe limitations, as they often poorly recapitulate the conditions found in the target organ the cell line was derived from. They are also clonal, meaning they have very limited genetic diversity, which hinders the ability to draw conclusions. Though they have disadvantages, in vitro studies have suggested that BPA alters global metabolism throughout the body by targeting the liver, which is crucial for regulating systemic energy homeostasis (Figure 4). As described previously, dysregulation of global metabolism has been associated with the incidence of several cancers, including breast cancer. Human hepatic cell line HepG2 exposed to low doses of BPA in the micromolar to picomolar range displayed activation of similar metabolic pathways as estrogen. Both BPA and β‐estradiol treatment modulated the urea cycle and the Krebs cycle, resulting in increases in concentrations of arginine, creatine, and glutamine within the cell.^{[ 73 ]} The uptake and metabolism of amino acids, most notably including arginine and glutamine, is characteristically upregulated in many cancer types.^{[ 129 ]} Additionally, BPA and β‐estradiol dysregulated lipid metabolism by altering levels of apolipoproteins.^{[ 73 ]} Apolipoproteins, which are responsible for binding to and trafficking lipids, have also been associated with poor cancer prognosis and are vital for cancer proliferation and invasion.^{[ 130 ]} Prior studies have also linked apolipoprotein dysregulation, specifically the upregulation of Apob and downregulation of Apoc3, to BPA exposure in the liver of CD‐1 mice, reinforcing this in vitro finding.^{[ 131 ]} Taken together, these findings suggest a mechanism through which BPA facilitates cancer incidence by dysregulating lipid and amino acid metabolism throughout the body.

BPA also has a direct, local effect on the metabolome of breast cancer. BPA exposure altered nucleotide metabolism in the ER‐positive cell line, MCF‐7, evidenced by increases in the nucleotide precursors cytidine triphosphate (CTP) and uridine diphosphate (UDP) glucuronic acid.^{[ 74 ]} Further supporting this result, a study that tested cocktails of exposures with a “leave one out” approach found that BPA was crucial for altering purine and pyrimidine metabolism in MCF‐7 cells, evidenced by increases in UDP, uridine triphosphate, cytidine diphosphate, and CTP.^{[ 75 ]} Nucleotides are essential for cellular viability and proliferation. In rapidly dividing cells, like cancer, nucleotides are used quickly, so shifting to a nucleotide salvage pathway is more efficient for DNA synthesis.^{[ 23 ]} However, when interpreting the results of these studies, it is important to note that the MCF‐7 cell line was immortalized from a fully transformed breast cancer. While its ER positivity makes it an ideal candidate for EDC testing, it is limited in that it cannot recapitulate the unique features of cancer initiation. Distinguishing studies that utilize models that reflect early‐timepoint initiation of breast cancer and models that reflect promotion and/or progression of already existing breast cancer is necessary when drawing conclusions about the role EDCs may play in facilitating the multi‐step model of carcinogenesis.

Using the ER‐negative cell line, MDA‐MB‐231, xenografted into mice, Zhao et al. demonstrated that BPF, a BPA analog, induced changes in glutathione biosynthesis, glycerophospholipid and glycerolipids degradation, and glycolysis, resulting in larger tumors.^{[ 78 ]} Decreases in glutathione biosynthesis can raise the concentration of ROS, leading to DNA damage and even carcinogenesis.^{[ 132 ]} The study also found that BPF promoted glycolysis, as assessed via the increase in the intermediates lactate, pyruvate, fructose‐6‐phosphate, and glucose‐6‐phosphate.^{[ 78 ]} Whether these glycolytic changes are a result of the proliferation induced by BPA, or if they are the drivers of the increased tumor size remains to be determined.

BPS altered a number of metabolic pathways, including the citric acid cycle, purine metabolism, and lipid metabolism in MCF‐10A cells, ultimately resulting in increased cellular proliferation.^{[ 76 ]} The same group found similar results in MCF‐7 cells exposed to TB‐BPA and TC‐BPA. This study revealed that TB‐BPA and TC‐BPA also induced glycolytic intermediates and decreased glutathione concentrations, leading to the induction of ROS.^{[ 77 ]} Exposure to TB‐BPA and TC‐BPA also induced energy metabolism from amino acids via glycolytic shunt pathways.^{[ 77 ]} These results suggest that bisphenol analogs behave similarly in altering the metabolome of breast cancer cells, highlighting the danger of continuing the use of these alternatives commercially.

4.3. Phthalates

Phthalates are common additives found in plastics used to make softer materials like vinyl, PVC, and food packaging (Figure 2). Phthalates share chemical similarities with β‐estradiol, lending them endocrine‐disrupting properties (Figure 1). However, the connection between phthalate exposure and breast cancer is still debated. A prospective study of more than 150,000 postmenopausal women failed to demonstrate a significant link between urinary phthalate levels and breast cancer incidence.^{[ 133 ]} However, this study did not take early‐life exposures into account, which may be most important to EDC‐induced breast cancer. In contrast, a multi‐ethnic prospective cohort study did find significantly increased risk of developing breast cancer in proportion to urinary levels of phthalate metabolites.^{[ 134 ]} This study had a much lower sample size, and disclosed that their analyses were based on a single urinary sample measurement, which complicates its interpretation. While epidemiological studies have failed to demonstrate a clear and reproducible association or lack thereof between breast cancer and phthalate exposure, mechanistic studies in vitro and in vivo have been able to provide evidence that phthalates may facilitate breast cancer development (Figures 3 and 4).

The most estrogenic phthalates are di(2‐ethylhexyl)phthalate, dibutyl phthalate (DBP), and benzyl butyl phthalate (BBP). These phthalates increase cell viability and proliferation as well as ER‐alpha activation at low doses in the cell line MCF‐7.^{[ 106 ]} The phthalates DBP and BBP also increased the proliferation of the ER‐negative cell line MDA‐MB‐231. This was shown to be a result of AhR activation, resulting in downstream signaling through a histone deacetylase enzyme that increased the expression of the oncogene and tumor promoter gene c‐myc.^{[ 14 ]} Additionally, when xenografted into nu/nu mice, MDA‐MB‐231 tumor size was increased following exposure to DBP and BBP.^{[ 14 ]} This study provided mechanistic insight toward how phthalates may increase the progression of ER‐negative breast cancer, but it does not reveal how phthalates may play a role in initiating breast cancer.

The effects of phthalates on the metabolome in human exposures have been evaluated. One study assessed metabolomic changes in the serum of pregnant women with high urinary levels of phthalate metabolites. Phthalate exposure was positively correlated with lipid dysregulation, including altered levels of lipid membrane constituents, cholesterol, and sphingomyelin.^{[ 78 ]} This also resulted in increased levels of triacylglycerols in serum, and intermediates involved in sphingosine biosynthesis.^{[ 78 ]} These intermediates are involved in inflammatory signaling pathways. Zhao et al. also found a positive correlation between body mass index and urinary phthalate metabolite levels, implying that phthalates influence metabolic function.^{[ 78 ]} The hyperlipidemia induced in these pregnant women could facilitate and favor the development of breast cancer, especially in the presence of pseudo‐hormones like phthalates. It is well established that metabolic dysregulation‐related pathologies, namely type II DM, increase the risk of developing breast cancer.^{[ 135 ]} Epidemiology studies have found an association between metabolic disorder and phthalate exposure. Indeed, urinary levels of phthalate metabolites from an NHANES 2001‐2010 study were linearly correlated with metabolic syndrome.^{[ 136 ]} The metabolic changes induced by phthalate exposure may increase the risk of breast cancer by systemically transitioning the metabolome to cancer‐favoring conditions.

4.4. Zeranols

Zearalenones (ZEN) are mycotoxins produced by the Fusarium fungal genus, common contaminates of corn and grains (Figure 2). Humans consume zearalenone‐containing grains, resulting in oral exposure. Zearalenone is metabolized to other beta‐resorcyclic acid lactones (RALs) in the liver, some of which possess more potent endocrine‐disrupting properties than the parent compound.^{[ 137 ]} Among the most estrogenic metabolites of ZEN is alpha‐zearalenol. Alpha‐zearalenol has been manufactured into the synthetic hormone Zeranol™. In the cattle industry, this compound is routinely administered to beef cattle to boost growth and meat production rates. This practice has led to concerns in recent years of additional human exposure to this potentially damaging class of EDCs. As a result of these multiple exposure routes, zearalenone, and its derivatives have been found in human serum ^{[ 138 ]} and human urine.^{[ 139 ]}

Zeranol, ZEN, and other RALs share chemical similarities with endogenous beta‐estradiol (Figure 1). ZEN and its related metabolites have been shown to competitively bind to the mammalian ER.^{[ 140 ]} This is perhaps attributable to the flexibility of the RAL ring structure.^{[ 141 ]} Additionally, after binding to the receptor, zeranol activates downstream estrogenic signaling pathways, indicating that the chemical is not only capable of fitting into the binding pocket of the ER but also is able to induce the downstream events activating an estrogenic signaling cascade.^{[ 142 ]}

Epidemiological studies have demonstrated that exposure to ZEN and its metabolites is associated with developmental defects consistent with disruption of the endocrine system. ZEN levels were associated with developmental delays in height and breast development.^{[ 143 ]} In a 10‐year longitudinal study, high serum levels of ZEN were associated with amenorrhea and pubertal delays.^{[ 96 ]} Most notably, ZEN levels in urine were associated with an increased incidence of breast cancer in a Tunisia population of women.^{[ 81 ]} Thus, more studies are required to elucidate the role of ZEN and its related metabolites in initiating and promoting breast cancer.

Both ZEN and alpha‐zearalenol have been shown to stimulate the proliferation of various cancer cell lines in vitro.^{[ 144 ]} Investigators have provided evidence of cancer promotion during several distinct stages of breast cancer development (Figure 3). Zeranol decreases doubling time in MCF‐10A cells and induces a neoplastic transformation when dosed repeatedly in this cell model.^{[ 145 ]} MCF‐10A is an immortalized normal breast cell line that models a very early proliferative stage of breast tissue. Zeranol at low concentrations stimulates the growth of the ER‐positive cancer cell lines MCF7 and KPL‐1, perhaps mediated by the downregulation of the cell cycle inhibitor p21Cip1.^{[ 146 ]} These cell lines represent a later stage of breast cancer development when the tumor already displays invasive characteristics. In the ER‐positive breast adenocarcinoma cell line MCF‐7, the obesity hormone leptin exhibited a synergistic effect on cell proliferation when co‐treated with zeranol. This suggested that metabolomic dysfunction was involved in zeranol's tumor‐promoting effect on these breast cells.^{[ 147 ]} These changes also occur in primary breast cancer epithelial cells, strengthening the conclusion that zeranol plays a role in metabolomic dysfunction and cancer.^{[ 148 ]}

In vivo studies modeling zeranol exposures with breast cancer‐related endpoints are few. However, existing studies suggest zeranol behaves similarly to estrogen, which is known to cause growth proliferation and carcinogenesis in animal models. In rats, zeranol has been shown to increase mammary growth.^{[ 83 ]} This proliferative response mimics the characteristic cancer‐promoting pattern of estrogen. Zeranol may also cause overt carcinogenesis in the august copenhagen irish (ACI) rat model, though the results of this study are very unclear due to the small sample size.^{[ 149 ]}

It is important to distinguish between hyperplasia and carcinogenesis when considering animal studies that model mammary cancer through hormonal stimulation. Estrogen behaves as a growth‐promoting agent through its activation of downstream estrogenic targets involved in cell cycle control. Additionally, estrogen can act as a genotoxic agent by producing ROS, though the exact mechanisms are unclear.^{[ 150 ]} However, stimulation of mitochondrial metabolism and excessive production of hydrogen peroxide have been proposed as potential mechanisms.^{[ 151 ]} Taken together, estrogen induces mammary hyperplasia in the short‐term and carcinogenesis in the long term in animal models, including the ACI rat.^{[ 152} , ^{153 ]} Zeranol has been demonstrated as a growth‐promoter through activation of the ER, but it remains to be seen whether it can initiate the first step of the carcinogenic process, the induction of DNA damage that ultimately progresses to transformed cells. ZENs may cause oxidative stress through the production of ROS, ultimately resulting in DNA damage. In CHO‐K1 ovarian cells, an increase in ROS and a resulting induction of DNA damage were identified after exposure to ZEN and its metabolites.^{[ 154 ]} Mechanistically, it is plausible that zeranol may be able to initiate breast cancers by causing DNA damage and estrogenic growth promotion. However, more robust animal studies are required to test this hypothesis.

To date, only one study has assessed metabolomic changes induced by zeranols in a cancer model. This study, which used MCF‐7 as an estrogen sensitive in vitro model, found that zeranol changed the metabolomic signature of cancer cells to facilitate cancer progression and invasion.^{[ 82 ]} Protein biosynthesis, the urea cycle, methionine metabolism, and arginine/proline metabolism were found to be differentially regulated by pathway analysis. Specifically, the study found that cysteine and glutathione levels were elevated in the zeranol treated MCF‐7 cells. This supports the finding that zeranol increases ROS, as cysteine and glutathione are stimulated by higher levels of ROS. Additionally, the study found that choline levels were elevated following zeranol exposure. Choline is vital for maintaining membrane stability and is a biomarker that has been suggested for clinical use to identify invasive tumors.^{[ 155} , ^{156 ]} Overall, there are many knowledge gaps in how zeranols may promote breast cancer, but the studies reported thus far demonstrate cause for reasonable concern that this endocrine disruptor may be able to initiate, promote, and progress breast cancer.

5. CONCLUSION

Endocrine disruptor exposure has been linked to breast cancer incidence, prognosis, and metastasis. The scientific community has not yet reached a consensus on the risk endocrine‐disrupting chemicals pose, and how they may cause hormonal cancers. A clearer mechanistic understanding of how these chemicals induce their putative carcinogenic action(s) is needed. Metabolomic analysis has demonstrated that various well‐known EDCs alter the metabolic signature of breast cancer cells to better suit survival needs, most commonly by increasing lipid metabolism. Additionally, EDCs may dysregulate global metabolism, leading to higher nutrient availability that could ultimately aid in cancer initiation and progression. These trends can be used to identify and characterize the effects of other EDCs and cancer‐causing toxicants more accurately and efficiently. Additionally, further research should investigate potential mechanisms, both ER‐dependent and independent, through which EDCs may alter metabolism in breast cancer.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflict of interest.

DATA AVAILABILITY STATEMENT

Data sharing is not applicable to this article as no new data were created or analyzed in this study.