Estrogens are critical for sexual and reproductive development especially in females and come in several similar chemical forms. Estradiol (E2) is the most abundant in cycling females, while estrone (E1) predominates in the postmenopausal period, and estriol (E3) and estrane (E4) predominate during pregnancy. Activity of endogenous estrogens, predominantly estradiol (E2) is primarily potentiated by the estrogen receptor-α (ERα) and ERβ, although an E2-specific G-protein coupled estrogen receptor-1 (GPER-1) has also been described. Hormone diffusion through the plasma membrane followed by receptor binding in the cytoplasm initiates canonical estrogen-to-ER signaling. Ligand-induced receptor dimerization and nuclear translocation lead to receptor-complexes that either directly bind estrogen responsive DNA elements (EREs) or complex with co-activators and co-repressors to regulate transcription of target genes (Levin and Hammes, 2016; Fuentes and Silveyra, 2019). ERα and ERβ can also heterodimerize, however their function is unclear, especially in contrast with the role of homodimers, such as proliferative ERα-ERα and suppressive/apoptotic ERβ-ERβ (Cowley et al., 1997; Papoutsi et al., 2009; Thomas and Gustafsson, 2011). ER signaling is complex and further diversified by the ability of receptors to localize at the plasma membrane in either truncated or full-length form, where they potentiate rapid non-genomic (non-translational) estrogen signaling (Figure 1) (Moss et al., 1997; Flouriot et al., 1998; Chambliss et al., 2000; Govind and Thampan, 2003; Hammes and Levin, 2007; Zhang et al., 2011). Other binding events at the cell membrane may also occur, such as interaction of full length ERα with G-proteins or with truncated isoforms of either ERα or ERβ (Kumar et al., 2007). Isoforms of ERβ, in particular, may further dimerize. Both homo- and heterodimers of ERβ isoforms have been characterized, each with its own distinct signaling modality. For example, homodimers of ERβ isoforms initiate estrogen-independent transcription whereas only heterodimers can initiate ligand-dependent transcription (Moore et al., 1998; Leung et al., 2006). Heterodimers of ERβ isoforms with full length ERβ can also form, and this increases the estrogen-independent transcriptional activity by four-fold. Activity of some, but not all, heterodimers of ERβ isoforms can be inhibited by ERα (Poola et al., 2005). Evidently, the complexity of estrogen signaling, and the diversity of ER modalities are critical for properly evaluating the effects of EDCs on individual immune populations. Although E2 generally has a higher affinity for its receptors compared to most EDCs, xenoestrogens can activate alternatively available ERs when estrogen is otherwise engaged. These EDC complexes may then act in retrograde upon complexed E2-ERs, ER-genes, or EREs to disrupt their activity.

Alternative targets to the ERs have been extensively described. Estrogen-related receptors (ERRs), notably ERRγ, constitutively (sans-ligand) bind either estrogen-related and EREs or naked DNA and non-E2-related co-activators (Horard and Vanacker, 2003; Audet-Walsh and Giguere, 2015). GPER-1, a G-protein coupled receptor (GPCR) with preferential specificity for 17-β estradiol (E2β), potentiates the rapid, but transient, modulation of several signaling pathways (Filardo et al., 2000; Revankar et al., 2005; Thomas et al., 2005; Ge et al., 2013). Distribution of GPER-1 and its functionality have been comprehensively reviewed elsewhere (Cheng et al., 2011; Xu et al., 2019). In line with its structural classification as a seven transmembrane GPCR (7TMR), GPER-1 undergoes endocytosis, often in the absence of ligand, and has predominantly been localized in the perinuclear space. Constitutive internalization and localization of GPER-1 indicate its potential for nuclear activity, which is corroborated by identification of a putative nuclear localization sequence. This proposes that GPER-1 can initiate gene expression both by interaction with other DNA-binding nuclear receptors, or by direct E2 mediated promoter binding (Madeo and Maggiolini, 2010; Pupo et al., 2013; Rudelius et al., 2015). The way by which EDCs interfere with normal GPER-1 signaling is not well defined. Many EDCs can diffuse through the cell membrane and localize within the same cytoplasmic and/or perinuclear space as occupied by GPER-1. In this vicinity, EDCs can perturb local concentration dynamics or act directly on the GPCR binding site, and thereby interfere with GPER-1 target activation.

Competitive binding of EDCs to estrogen receptor targets has preceded interference on cell metabolism, nuclear receptor turnover, and hormone sensitization. As a result of binding to ERs, transcriptional enhancement along with modulation of other signaling pathways have been increasingly reported. The diversity of EDC categories also lends diversity to EDC effects. For example, certain xenoestrogens may disrupt activation of nuclear pathways by inherent ability to diffuse and translocate through the lipid membrane, where they promote rapid and direct gene expression. Other EDCs act more distally, on G-protein coupled receptors, to disrupt homeostatic expression of transcription factors, and latently affect gene expression (Yoon et al., 2014). Understanding the interplay of individual exposure, estrogen activity and immunomodulation, along with how this contributes to the collective phenomenon of autoimmunity, will lead to improved healthcare guidelines, reduced exposure burdens and novel therapeutics for the treatment of immune related disease.

The ability of phenols to bind endogenous estrogen receptors, compete with endogenous ligand, and initiate activation of downstream pathways is central to understanding how they may modulate immune cells (Figure 1). In contrast with studies on non-genomic ER activation, phenols have been more extensively characterized in the context of genomic signaling. Relative to E2, BPA has reduced affinity for ERα (60–70% transactivation of E2) and exhibits preferential binding for ERβ (80%) over ERα. In consequence, phenols behave as partial antagonists in the presence of E2, in that binding of the EDC is less efficacious. Related factors such as bioaccumulation and tissue distribution of phenols, as well as generation of metabolites, render these EDC compounds with a complexity that exceeds E2 binding affinity alone (Matthews et al., 2001; Takayanagi et al., 2006). Moreover, variations in steric complexity seem to direct binding orientation of different phenols, which controls whether transduced effects will be agonistic (BPA) or antagonistic (BPC). This is also true of co-activator recruitment and/or binding, whereby BPA is a weak agonist, but bisphenol C (BPC) serves as an antagonist, and binding of both is typically favored in the presence of high co-activator concentrations (Delfosse et al., 2012). Despite this, bisphenols do act as full ERα agonists on cell growth, suggesting that this may be induced by direct ER-bisphenol binding to ERE DNA regions. Similar to ER-binding, phenol-GPER-1 agonist binding has been demonstrated, with relatively high affinity reported for BPA and NP (Thomas and Dong, 2006). In this context, GPER-1 on the rough endoplasmic reticulum can prevent phenol nuclear accessibility and allow E2-ER nuclear complex formation. Alternatively, pre-existing GPER-1 occupancy by E2, such as in the endoplasmic reticulum, can free up cytosolic or nuclear receptor targets for EDC binding and vice-versa. Thus, the physiological state of the organism at the time of exposure is critical for transduction of xenoestrogen effects and may help explain how EDCs function as triggers for disease in some individuals, but not in others. As well, synergistic effects of multiple pathways should similarly be considered. Phenols can further act as inverse antagonists on other receptor targets. BPA, in particular, has been shown to increase ERRγ constitutive activity, and could thereby lead to a deregulated cell metabolism with implications for both cancer and autoreactivity (Liu et al., 2007). Subsequent to their direct binding activity, phenols can augment the rate of receptor turnover. BPA decreases ubiquitination and degradation of ERβ, as well as intracellular ERα protein levels and mRNA transcripts (Masuyama and Hiramatsu, 2004; La Rosa et al., 2014). Similarly, BPF increases ERβ, however, other EDCs, such as NP, have yielded more complex, potentially sex and species specific effects upon receptor turnover (Sakimura et al., 2002; Seo et al., 2006; Okazaki et al., 2017). Collective observations have thereby surmised that many phenols display characteristics of selective ER modulators (SERMs), most notably in their ability to behave in a discriminatory manner (Delfosse et al., 2012). This functional classification is an important consideration for study design and should be deliberated when inferring results to human health.

The turnover rate of ERs and endogenous ligand are precisely balanced to control the downstream signaling necessary for proper function of immune cells. Docking of estrogen to ERs does not produce simple, all-or-nothing target activation. Instead, this binding event appears to induce a state that depends on the equilibrium between receptor and ligand. EDC-induced changes to such a system may impact either the perceived ligand concentration or the receptor availability, leading to complex and sometimes dichotomous results on immune function. Interpretation of immunological findings is further complicated by contrasting pre-clinical and clinical results, likely both influenced by study design and experiment dose.

Dichotomy of immune findings is best illustrated by pre-clinical, innate immune findings. For example, phenols decrease downstream signaling and key regulators of inflammation, such as TNF-α, nitric oxide (NO), and NF-κB in both murine macrophages (BPA 228–22,800 μg/L) and immortalized macrophages (BPA 5,700–22,800 μg/L; NP 2,200 μg/L; OP 2,000 μg/L) (Byun et al., 2005; Yoshitake et al., 2008). Alternatively, BPA at similar doses can, instead, increase levels of NO, reactive oxygen species (ROS), NF-κB activation and mRNA transcripts of IL-1β and IL-10 pro-inflammatory cytokines. Interestingly, macrophage bactericidal function was improved following low-dose BPA exposure (0.10–10 μg/L), whereas administration of high dose BPA (100–10,000 μg/L) induced apoptosis, presumably in an oxidative stress-dependent manner (Yang et al., 2015). Dose-independent effects on cytokine secretion have been further observed for both BPA (0.10–1,000 μg/L) and NP (0.10–100 μg/L), such as increased transcripts for IL-1β, IL-10, TNF-α, IFN-γ, and MyD88. Consistently, NO metabolism was affected only by the high dose of BPA or NP (Xu et al., 2013). This dichotomy of results is not only biphasic but may reflect a species specificity for estrogen regulation that results in varying outcomes on inflammatory factors (Table 1).

Paradoxical results from clinical studies have likewise been reported. Human macrophages and T lymphocytes responding to BPA (2,850–45,700 μg/L) showed a dose-dependent decrease in neopterin and IFN-γ, respectively (Gostner et al., 2015). In contrast, production of IL-6, TNF-α and IFN-γ from human adipocytes was increased after exposure to BPA (0.02–22.8 μg/L Ben-Jonathan et al., 2009; 0.228 μg/L Valentino et al., 2013). This corresponded with a lowered production of the anti-inflammatory hormone, adiponectin, and an impaired glucose metabolism (Ben-Jonathan et al., 2009; Valentino et al., 2013). In the context of clinical obesity and chronic inflammation, low-dose EDC exposure effects suggest that EDCs may exacerbate or even cause autoimmunity, including post-infection autoreactivity of metabolic etiology. Such potential has been identified for BPA, which increased activation of JNK, STAT3 and NF-kB signaling pathways. Conversely, adipogenesis was unaffected despite previous pre-clinical reported effects (BPA 0.228 μg/L; Valentino, D'Esposito et al., 2013; Ariemma, D'Esposito et al., 2016).

In other clinical findings, phenol effects were transduced by GPER-1 activation to induce human polymorphonuclear neutrophils (PMNs) toward a pro-inflammatory phenotype (Rodenas, Tamassia et al., 2017). Despite limited research into EDC effects upon PMNs, administration of BPA (228–22,800 μg/L) did increase ROS production in these cells, in a dose-dependent manner. ROS activity was also associated with diminished bactericidal function and inhibition of PMN chemotaxis, whereas phagocytic function remained unaffected (Balistrieri, Hobohm et al., 2018).

Phenols have additionally been characterized for their effects on adaptive immunity. Gestational BPA treatment (300–3,000 μg/kg) alters the humoral response to immunization, leading to higher levels of antigen specific immunoglobulin (Ig) Gs and increased lymphoproliferation (Yoshino et al., 2004). Increased IFN-γ and IL-4 were also noted as a result of BPA effects on T cell helper phenotypes, particularly TH1 differentiation. Furthermore, despite an overall increase in T cells, the ratio of CD8/CD4 T cells was skewed in favor of CD8 T cells. Collectively, these results suggest that phenols may prime the humoral response toward sustained inflammation (Yoshino et al., 2004). In the context of pathogen response or autoimmunity, sustained humoral activation as induced by BPA can degenerate into chronic inflammation and the onset or exacerbation of autoreactivity. This is further substantiated by evidence of BPA-induced (2.28–228 μg/L) increased differentiation of TH17 cells, including increased secretion of IL-17 and IL-21, as well as TH17 differentiating cytokines, IL-6 and IL-23. Such changes were altogether dose-dependent and sex-specific (female bias), further suggesting that BPA alters dynamics of adaptive immunity and leads toward a proinflammatory phenotype that may cause adverse health outcomes (Luo, Li et al., 2016).

In contrast, in vitro treatment with BPA (2,280–22,800 μg/L) or NP (220–22,000 μg/L) decreased lymphoproliferation, with BPA demonstrating preferential inhibition of B cells over T cells (Sakazaki et al., 2002). Comparable with BPA (22,800 μg/L), analog compounds, BPF (20,000 μg/L) and BPS (7,500 and 25,000 μg/L), decreased splenocyte proliferation and augmented human B cell cycle transitions. The reduced viability of human B cells, likely as a result of enhanced ROS production by all three bisphenols, was comparable with effects previously described for innate cells (Yang et al., 2015; Jang et al., 2020). This ability of EDCs to downregulate B cell responses may be critical in the context of viral or other chronic infections, which have been linked to the onset of several autoimmune phenotypes (Box 1). By affecting TH differentiation, altering the distribution of CD8 to CD4 T cells, and diminishing B cell responses, EDCs may directly contribute toward an ineffective adaptive response, one which may allow pathogen escape and influence the development of chronic inflammation and self-reactivity.

Exposure to phenols is a potential risk factor for autoimmunity, however studies are similarly limited by complex, often biphasic results. In animal models of autoimmune disease, such as systemic lupus erythematosus (SLE), long-term BPA exposure (300–350 mg/kg) enhanced production of anti-Br-RBC autoantibodies. Elevated levels of IgM secretion from B1 cells have also been reported following exposure to both BPA (228 μg/L) and NP (220 μg/L) (Yurino et al., 2004). Gestational BPA (10,000 μg/L) followed by induction of autoimmune encephalomyelitis (mild EAE; MOG_35–55/CFA-only) did not affect the clinical course of disease. Conversely, trends in severity of EAE were observed, such as an increased disease score during the first EAE episode in female mice and a higher relapse incidence in male mice (Krementsov et al., 2013). Similarly, an earlier onset of MS symptoms and worsened disease severity were shown in a virus-induced demyelination model of MS (BPA 10 μg/kg) (Brinkmeyer-Langford et al., 2014). Interestingly, colitis was more severe in MS mice exposed to BPA, which corroborates clinical data on the co-occurrence of MS and inflammatory bowel diseases (IBDs) in humans (Kimura et al., 2000; Alkhawajah et al., 2013; Brinkmeyer-Langford et al., 2014).

Consideration of phenols in the context of other autoimmune conditions provides additional evidence for their harmful effects. For example, EDC levels (BPA 1,000 μg/L) relevant to human exposure increased diabetic incidence and corresponded with higher grade insulitis in pancreatic islets, concurrent with decreased macrophage infiltrates of lowered phagocytic potential (Bodin et al., 2015). Similar findings on type-1 diabetes identify BPA as pro-apoptotic for pancreatic islet cells, which suggests a mechanism by which BPA promotes self-antigen auto-activation and diabetic onset (Bodin et al., 2014; Bodin et al., 2015). Furthermore, high-fat diet and BPA exposure (50 ug/kg) were associated with immune-metabolic dysfunction, characterized by increased transcripts for toll-like-receptor-4 and related signaling pathways, such as NF-kB and NLRP3 inflammasome, which amplified proinflammatory cytokine production to promote liver damage (Pirozzi et al., 2020). More specifically, in diabetic animals, decreased levels of secreted IL-10, TNF-α, IFN-γ and IL-4 have also been reported. Animal findings thus suggest that BPA and other phenols can alter the severity of autoimmune disease, likely through a mélange of effects, including increased production of autoantibodies, altered cytokine secretion and decreased macrophage scavenging. Altogether, these changes contribute to the persistence of self-antigen and may exacerbate autoimmunity (Bodin et al., 2015). There is also evidence that phenol exposure could contribute to sex-bias in autoimmune disease outcomes. In non-obese diabetic mice, BPA affected both development and pathology of T1D in a sex-specific manner. In females, development of T1D was accelerated and accompanied by a shift in proinflammatory immune factors, whereas disease development in males was delayed, associating with an increase in anti-inflammatory factors (Xu et al., 2019). These results highlight the potential of phenol EDCs to induce sex-specific changes in the development and course of autoimmune diseases.

In humans, the role of phenol EDCs in autoimmunity is less defined. Nevertheless, clinical studies have provided compelling evidence on the detriment of EDCs to human autoimmunity (Table 2). When isolated DNA from healthy individuals was exposed to BPA (228,000 μg/L), formation of DNA-BPA complexes and BPA-induced conformational changes (48.2% hypochromicity, 260 nm) were observed. Relative to native DNA, DNA-BPA complexes had increased stability, which suggests a novel process for how environmental factors may influence the development of self-reactive autoantibodies. In conformity, serum IgG antibodies from SLE patients showed affinity and recognition for such DNA-BPA complexes, but were otherwise undetected by IgG antibodies purified from healthy individuals, presumably because the latter have no affinity for self-antigens, such as anti-double stranded DNA (dsDNA) recognition. Interestingly, the SLE IgG antibodies showed preferential recognition of BPA-altered DNA over native DNA, which could help explain how an otherwise weak immunogen (dsDNA) is enhanced to autoimmunity (Rekvig and Mortensen, 2012; Alhomaidan et al., 2019). An impact of phenols has also been reported in autoimmune thyroid disease, where serum BPA showed a positive trend with antibody levels for thyroglobulin (TgAb), thyroid peroxidase (TPOAb) and thyroid receptor (TRAb). BPA was also an independent predictor of TgAb and TPOAb, but not TRAb (Chailurkit et al., 2016). Whether modulation of immune populations by phenols has the capacity to influence the onset or progression of autoimmune disease remains controversial, however, animal findings implicate immune effects that are dependent on several factors including study design (Table 1).

Paraben-ER activity and binding mechanism depends on the octanol-water partition coefficient (K _ow), an index of cell membrane transfer. Parabens with a relative high K _ow, such as BP, can diffuse across the cell membrane and directly bind DNA in silico (Manzetti, 2018). Ability to bind DNA directly is not exclusive to BP. Transcriptional and translational products of ERE-genes have also been reported following high dose (1,000 mg/kg/d) exposure to isopropyl and isobutyl parabens. The concurrent decrease in mRNA and protein expression of ERα further suggests that parabens may not always follow an ER-to-ERE DNA activation pattern, and can instead directly bind to the ERE (Vo and Jeung, 2009). Alternatively, PP induces morphogenic changes in glandular structures via activation of both ERα and ERβ and GPER-1 (Marchese and Silva, 2012). Exposure to MP (3 μg/L), PP (3.6 μg/L) or BP (3.9 μg/L), can also increase GPER-1 gene and protein expression, which indicates that parabens can activate a GPER-1 autoregulatory feedback loop (Wrobel and Gregoraszczuk, 2015). Additional activity of parabens on other estrogen targets involves the inverse antagonism of ERRγ. Parabens (methyl-to-benzyl parabens) are recruited in silico to the ERRγ active site, and form bonds with active site residues (Zhang et al., 2013). Binding of parabens to ERα or ERβ conformations that transduce non-genomic signaling has not yet been described, likely due to the size limitation on membrane permeability. However, in the intercellular space, size-excluded parabens could preferentially target surface ERα or ERβ to activate non-genomic signaling. Alternatively, the alkyl bulk of parabens may impede ligation to the active site of surface ERs. Nevertheless, activation of estrogen targets by parabens with a high K_ow offers sufficient insight to raise concern regarding the interference of paraben EDCs with normal endocrine signaling in immune populations.

Effects of paraben exposure on innate immunity are often derived from cosmetic allergology studies assessing the safety of personal care products. This restricts findings within a narrow niche of innate immune cells, specifically those that relate to allergic sensitivities. Nevertheless, findings can help inform on effective dosages and exposure schedules. Notably, dose ranges often fall within or surpass the doses at which activation of ERs has been described. For example, straight-chain parabens spanning from heptyl-to-decyl (low dose HP-DP, 2,360–2,780 μg/L) or pentyl-to-dodecyl (high dose PP-DDP, 20,800–30,600 μg/L) cause a significant release of histamine from degranulating mast cells (Uramaru et al., 2014). This is also sustained by exposure to the branched isopentyl paraben (IPP, 20,800 μg/L). MP causes immediate hypersensitivity, including contact urticaria, although it is unclear whether paraben-reactivity is a result of innate or adaptive activation, the latter of which would be an anti-paraben allergic recall (Henry et al., 1979; Macy et al., 2002). To date, discerning these effects has been limited by unsuccessful detection of anti-paraben IgE antibodies in the serum. However, this does not rule out a paraben-induced adaptive recall. IgE antibodies have a relative short-half life, which temporally limits their detection, and may lead to the incorrect assumption that paraben hypersensitivity is a localized innate response. Alternatively, other immunoglobulin classes, such as IgG antibodies, may be involved in activation of innate cells, transducing effects otherwise perceived as immediate reactivity to parabens (Kokubu et al., 1989). A more detailed discussion is included in the adaptive immunity subsection on paraben effects.

Observations independent of allergic innate responses have been modestly reported. In the brain, BP (200 mg/kg/d) increases the levels of NO, IL-6 and TNF-α but downregulates IL-1β (Hegazy et al., 2015). Peripherally, it can also enhance dendritic cell (DC) trafficking to draining lymph nodes (dLN) and increases the secretion of IL-4 and IFN-γ from dLNs (Matsuoka et al., 2018). During pregnancy, exposure to BP and PP associate with higher measures of oxidative stress, lower expression of c-reactive protein, and changes in cytokine secretion. BP for example, increases IL-1β, IL-6, IL-10, and TNF-α levels, whereas MP increases only IL-6 and IL-10 and decreases levels of IL-1β (Watkins et al., 2015). The ability of parabens to influence cytokine dynamics implicates not only innate but also adaptive responses, as the latter depend upon the precise balance of inflammatory factors to mount the appropriate humoral defense.

Effects of parabens on adaptive immunity have primarily remained unexplored, perhaps because of the restrictive interest in parabens outside of cosmetics and personal care products. Alternatively, parabens can be erroneously overlooked since they have minimal bioaccumulation potential and relatively lower toxicity compared to other EDC categories. Nevertheless, clinical case reports identify anomalous serology characterized by both anti-paraben antibodies and paraben-associated auto anti-Jk^a antibodies (Judd et al., 1982; Judd et al., 2001; Gniadek et al., 2018). Although characterization of paraben-specific IgE antibodies is limited by their half-life, sensitization with BP and HP followed by challenge produces a significant allergic reaction (Uramaru et al., 2014). This suggests that prior paraben priming is sufficient to generate recall upon antigen re-presentation and is a hallmark feature of adaptive immunity. Interestingly, patients presenting with intractable dermatitis resistant to typical corticosteroid creams, achieve rapid and complete recovery when administered paraben-free corticosteroid formulations. This indirectly confirms the existence of paraben-specific antibodies, although it does not offer an explanation as to how parabens may trigger an adaptive response (Guyton et al., 2005). One possibility involves a combination of effects that involves both innate and adaptive mechanisms. Parabens acting on innate cells result in an unbalanced inflammatory milieu that recruits adaptive immune cells. In turn, recruited cells activate in the presence of cytokines and are then primed by exogenous parabens, generating antibodies with paraben affinity. In the context of dermatitis, this presents an easily remediable effect of exposure. However, such immunomodulation may be detrimental when paraben exposure occurs simultaneous with an ongoing innate or adaptive response, and could trigger hyperinflammatory syndrome and autoreactivity, respectively.

The role of parabens in autoimmunity is similarly limited to a few exploratory studies. To the best of our knowledge, no findings have been reported for paraben effects on immune cells that drive autoimmunity, such as B cells or T cells. However, an association of EP urinary levels and gestational diabetes has been reported, which suggests an increased risk of diabetes in later life (Liu et al., 2019). In contrast, high concentrations of methyl-to-butyl parabens in urine were associated with lower odds of diabetes overall (Ward et al., 2020). This suggests that exposure to parabens in the context of diabetic autoimmunity is more complex and requires a molecular approach for validation.

In children, sex-biased associations have been determined between urinary concentrations of methyl and propylparaben, whereby boys have increased asthma morbidity, as modeled by emergency department visits. A similar male-bias has also been reported in paraben-related triclosan exposure and food sensitization among children, further suggesting that EDCs influence human immune disease outcomes in a sex-specific manner (Savage et al., 2012; Quiros-Alcala et al., 2019). Studies on central nervous autoreactivity have shown that paraben exposure, particularly BP (1.17 × 10^–2 μg/L), inhibits lysosomal enzyme secretion, which is otherwise elevated extracellularly in MS patients (Bairati et al., 1994; Guyton et al., 2005). Nevertheless, combined findings illustrate the diversity of effects with which paraben EDCs influence immunity.

Like members of other EDC categories, phthalates both activate and disrupt endocrine signaling, particularly pathways regulated by estrogen and its receptors (Figure 1). Although non-genomic activation has been defined for diethyl phthalate (DEP), it has remained unexplored for other compounds (Kumar et al., 2014). Nuclear effects, however, are generally more common. For example, butyl benzyl phthalate (BBP; 3,120 μg/L) caused demethylation of ERα promoter-associated CpG islands and increased ERα gene expression (Kang and Lee, 2005). In GPER-1⁺ cells, BBP (31,200 μg/L), DBP (2,800 μg/L), and DEHP (3,900 μg/L) all led to increased proliferation. In the absence of GPER-1, this effect was undetected for all phthalates, which suggests that they initiate proliferation via GPER-1 (Kim et al., 2004). Additionally, both DBP and diisopentyl phthalate (DiPeP) (500 mg/kg/d) increase GPER-1 gene expression, whereas DiPeP at a lower dose (125–250 mg/kg/d) decreases ERα without affecting levels of ERβ (Curi et al., 2019). Interestingly, phthalates may act simultaneously as agonists and/or antagonists, primarily via ERα; only BBP has been shown to have activity on ERβ. Binding of phthalates to other estrogen targets, including ERRs, has not been extensively characterized. Only one study found that DEHP (50,000 μg/L) increased ERR gene expression (Park and Kwak, 2010). Nevertheless, collective data (Figure 1) suggests that phthalates can exert effects via both ERα and GPER-1 and have the potential to disrupt cells that depend on estrogen signaling, which includes cells of the immune system.

Phthalate compounds and their metabolites have been identified as adverse modulators of innate immunity (Table 4). Perinatal DEHP (60 and 600 μg/kg/d) alters inflammatory functions of peritoneal macrophages and results in suppression of TNF-α and IL-1β gene expression. In contrast, gene expression for IL-6 was significantly upregulated, suggesting that phthalate exposure contributes to the dysregulated continual production of IL-6, which has been associated with the onset of various diseases. Interestingly, both DEHP and its metabolite, mono-(2-ethylhexyl) phthalate (MEHP, 20 μg/L), reduced macrophage phagocytosis (Li et al., 2018). Because phthalates are known to leak from plastic goods and accumulate as contaminants in soil and water sources, there is concern that metabolite exposure may lead to inadvertently persistent exposure. Similar cytokine dysregulation in human innate immune cells has also been reported. DEP (22,200 μg/L) and DBP (27,800 μg/L) upregulate secretion of IL-6, CXCL8 and IL-10 secretion and inhibit TNF-α secretion in macrophages (Hansen et al., 2015). DEHP (3,900 μg/L) further modulates the differentiation and maturation of PBMC-derived DCs (Ito et al., 2012). Even at low doses, both DEHP (39 μg/L) and BBP (31 μg/L) suppress expression of type I interferons, such as IFN-α and IFN-β, from plasmacytoid DCs (Kuo et al., 2013). Collectively, animal and human studies are congruent in implicating phthalates in cytokine dysfunction that can drive aberrant innate immune responses.

Phthalate-induced dysregulation of adaptive response has unsurprisingly resulted from both parent esters and their metabolites. DEP (22,200 μg/L) and DBP (27,800 μg/L) impaired activation-induced T cell secretion of IL-2, IL-4, TNF-α and IFN-γ, but not levels of IL-6 or IL-10. In contrast, the monoester metabolite of DBP, monobutyl phthalate (MBP, 22,200 μg/L) upregulated expression of IL-6, suggesting that phthalate compounds are twice as immunomodulatory (Hansen et al., 2015). In their intact form, phthalates may disrupt signaling pathways that control normal cytokine expression, such as E2-mediated transduction. However, once inside the cell, they can also be processed into various metabolites. This extends their activity and sometimes produces effects that are independent of the parent compound. Effects induced by monoester metabolites have been reported, including an associated reduction of cord blood T _reg cells by both monoethyl (MEP) and monoisobutyl phthalate (MiBP) (Herberth et al., 2017). Nevertheless, not all studies report metabolite activities but they do corroborate adaptive immune dysfunction. For example, DEHP (39 μg/L) and BBP (31 μg/L) suppress IFN-γ and can increase IL-13 secretion from CD4 T cells (Kuo et al., 2013). In B cells, DEHP inhibits proliferation (≥1,560 μg/L), which causes a significant reduction in IgM-secreting plasmablasts and plasma cells (≥6,250 μg/L) (Martins et al., 2015). In contrast, DEHP (30, 300, 3,000 μg/kg) can also increase expression of SLAMF1 and SAP in T_FH cells, thereby acting as an immunoadjuvant to promote co-stimulatory activity. DEHP also leads to enhanced formation of germinal centers by elevating expression of transcription factors, Bcl-6 and c-MAF in T_FH cells, as well as IL-21 and IL-4 cytokines (Han et al., 2014; Han et al., 2019). Interestingly, early-life DEHP exposure elevates secretion of virus-specific IgM following post-natal hepatitis B immunization. This, however, alters the gut microbiota in newborns and could have lasting repercussions on immune responses in later-life, including onset of autoimmune disease (Yang et al., 2019).

Autoimmunity. Contributions of phthalates to autoimmunity have been thoroughly characterized, especially their potential to exacerbate, or perhaps, induce autoreactivity. In testicular autoimmunity, DEHP (0.01% or 11.3–13.3 mg/kg/d) increases severity of mild experimental autoimmune orchitis (EAO) and is accompanied by an increase in IFN-γ⁺ cells and overall macrophages. Elevated lymphocytic infiltrates and anti-testicular germ cell autoantibodies further suggest that phthalates can exacerbate autoimmune disease (Hirai et al., 2015). In rheumatoid arthritis (RA), maternal BBP administration (3,000 μg/L) similarly increases the prevalence and severity of RA in the progeny, illustrating a transgenerational effect of phthalate exposure. This is associated with elevated levels of serum IgG1 and IgG2a, as well as enhanced secretion of IFN-γ and IL-17 from splenocytes (Elter et al., 2020). DEHP (7.5 mg/kg) can also increase levels of autoantibodies and associated proteinuria-induced renal dysfunction (Lim and Ghosh, 2005). Effects of phthalates on other autoimmune conditions have also been identified. In children with new onset of type 1 diabetes, higher levels of DiBP metabolites were detected by comparison with healthy controls (Castro-Correia et al., 2018). Furthermore, although the temporal context of immune deregulation is yet debated, inappropriate inflammatory responses and altered functions of lymphocytes have been proposed to be causative for and/or correlative with type 2 diabetes (de Candia et al., 2019). At levels relevant to human exposure, phthalates induce both transient and life-long metabolic dysfunction, including an increased risk of type 2 diabetes and insulin resistance (Radke et al., 2019). Combined data thus illustrates that phthalates directly and indirectly (transgenerationally) disrupt the normal function of both innate and adaptive immune processes, which may trigger onset of autoimmunity, especially in predisposed individuals. In addition, phthalate immunomodulation may lead to somewhat milder, but chronic, effects, such as persistent inflammation. This disruption in cytokines may, over time, provide a favorable context for self-reactivity to develop, even in individuals otherwise disinclined to autoimmune disease.