Bisphenols as Environmental Triggers of Thyroid Dysfunction: Clues and Evidence.

Bisphenols (BPs), and especially bisphenol A (BPA), are known endocrine disruptors (EDCs), capable of interfering with estrogen and androgen activities, as well as being suspected of other health outcomes. Given the crucial role of thyroid hormones and the increasing incidence of thyroid carcinoma in the last few decades, this review analyzes the effects of BPS on the thyroid, considering original research in vitro, in vivo, and in humans published from January 2000 to October 2019. Both in vitro and in vivo studies reported the ability of BPs to disrupt thyroid function through multiple mechanisms. The antagonism with thyroid receptors (TRs), which affects TR-mediated transcriptional activity, the direct action of BPs on gene expression at the thyroid and the pituitary level, the competitive binding with thyroid transport proteins, and the induction of toxicity in several cell lines are likely the main mechanisms leading to thyroid dysfunction. In humans, results are more contradictory, though some evidence suggests the potential of BPs in increasing the risk of thyroid nodules. A standardized methodology in toxicological studies and prospective epidemiological studies with individual exposure assessments are warranted to evaluate the pathophysiology resulting in the damage and to establish the temporal relationship between markers of exposure and long-term effects.


Introduction
Thyroid hormones (THs) play a critical role in the regulation of physical development, somatic growth, metabolism, and energy provision and are essential for normal brain development in humans [1]. Thus, any interference with THs status and signaling during development may have an impact on physical health, and can be associated with neurological deficits and even irreversible mental retardation in the case of severe maternal TH deficiency [2]. Meanwhile, thyroid cancer (TC) incidence rates have been rising in many western countries, including the United States where the incidence increased 3.6% per year during 1974 to 2013 [3]. TC is the most common endocrine malignancy, and by 2030, it is estimated to become the fourth leading cancer diagnosis in the United States [4]. Papillary thyroid cancer (PTC), in particular, is the most frequent histotype with a typically excellent prognosis, accounting for 70% to 90% of well-differentiated thyroid malignancies, and though over diagnosis of small tumors is thought to contribute significantly to the increase in incidence, PTC incidence has significantly increased for every stage and tumor size category [3]. The etiology of TC is multifactorial and the proposed risk factors in the literature include sex, family history of TC, radiation exposure, excess weight, iodine intake, and dietary habits [5]. Although the thyroid is characterized by a low proliferation index, it is particularly susceptible to environmental chemicals that may contribute to the increasing incidence of TC [6]. Table 1. Concentration of bisphenols in the environment and human body, and estimated exposure by age groups to bisphenol A (a), principal bisphenol A substitutes (b), and halogenated derivatives of bisphenol A (c).

Thyroid Disrupting Properties of BPs: in Vitro Studies
Biological function of thyroid hormone triiodothyronine (T3) is generally mediated by the nuclear receptors TRα1, TRβ1, and TRβ2 that are conserved in all vertebrates. T3 binds to the TRs with similar affinities mediating TH-regulated transcription with different levels in different tissues [60]. TRα1 is the predominant subtype in cardiac muscle and bone, TRβ1 is the predominant subtype in kidney and liver, while TRβ2 is more abundantly expressed in the hypothalamus and in the pituitary gland, and has a critical role in the regulation of the hypothalamic-pituitary-thyroid (HPT) axis [61]. TRs bind at DNA as homodimers or forms heterodimers with the retinoid X receptor to T3 response elements and they can regulate transcription both in the absence and in the presence of ligands [62]. On positively regulated genes, the unliganded TRs bind to corepressor proteins such as the silencing mediator of retinoid and thyroid hormone receptor (SMRT) or the nuclear receptor corepressor (N-CoR), resulting in the suppression of transcription [63]. The binding of T3 to TRs leads to a dissociation of the corepressors, and the subsequent recruitment of coactivator proteins, such as those of the p160/SRC (steroid receptor coactivator) family, including SRC1, SRC2, and SRC3, thus promoting activation of transcription [63].
In vitro models have tested and verified the ability of BPs to disturb thyroid function through multiple mechanisms that may produce different consequences depending on the heterogeneity of experimental conditions among studies such as the chemical tested, the concentrations used, and the presence/absence of T3 or T3 antagonists. BPs were reported to exert numerous effects on the thyroid, and each affected pathway may lead to perturbations of thyroid hormone levels, leading to a dysregulation of thyroid function. The pathways are not necessarily inter-connected, but there is some evidence that BPs may lead to an impact on the gland and its function at multiple levels, as reported in the following paragraphs.

Interference with T3 Transcriptional Activity
Numerous studies have evaluated the ability of BPs to suppress hormonal transcriptional activities mediated by TRα1 and TRβ1 in competitive binding and transient expression assays (Table 2).
Whereas BPA alone did not induce visible effects on T3-induced transcription [2,73,75,76], in the presence of physiological concentrations of T3, low-dose BPA enhanced the interaction of TR with N-CoR by directly binding to TR [2].

Cell Proliferation
The rat tumor pituitary cell line GH3 has been frequently employed as a standard pituitary cell model for assessing TH effects [81]. Indeed, cell proliferation and growth hormone (GH) secretion primarily depend on THs [81] and involve TR-mediated mechanisms, specifically the induction of gene expression [82].
A series of investigations assessed the agonistic and antagonistic properties of BPs in GH3 cell growth both in absence and in presence of T3 (Table 2). BPA, and in particular BPA derivatives, generally promoted GH3 cell proliferation and GH release in the concentration range of 10 −6 -10 −4 M [70,81,83]. In some studies, the agonistic activity was detected exclusively in the presence of T3 [82,84], whereas in others BPA and its substitutes inhibited cell growth with T3, and TH-antagonistic effects appeared to depend on the tested dose and the time of exposure [80,85].
Effects of BPs on cell growth were antagonized by amiodarone, a known TR antagonist [80]. Nonetheless, amiodarone was also reported to act as a slight agonist at low concentrations and antagonist at increasing doses, and BPA and its halogenated derivatives exhibited comparable doseresponse curves [76]. In PTC cells, BPA had similar proliferative effects as E2 [86], and consistent with this finding, co-exposure to E2 potentiated the increased GH3 cell proliferation (from 190% to 252% after 96 h) by BPA and BPAF [85]. In contrast, TBBPA could not counteract the inhibitory effect of fulvestrant, a strong antiestrogen, on cell growth [81].
Cell growth was further antagonized by U0126, an inhibitor of MEK, the kinase responsible for the activation of ERK in the Raf-MEK-ERK pathway in mammalian cells [87]. Similarly, TBBPA at concentrations in the lower micromolar range caused arrest of cells growth in the G1 or G2 phase, depending on the duration and intensity of the treatment and on cell specific and dose dependent modulations of the Raf-MEK-ERK pathway [87]. BPA may exert disrupting effects on TH-mediated transcription interfering with a different non-genomic mechanism mediated by integrin αvβ3, a heterodimeric transmembrane glycoprotein [77]. In normal conditions, T3 and thyroxine (T4) induce serine phosphorylation of TR-β1 by binding to αvβ3 and activating mitogen-activated protein kinases (MAPK) and/or c-Src/PI3K pathways [78], which determines the dissociation of N-CoR or SMRT from TR-β1 and consequent activation of transcription. The competitive binding of BPA to αvβ3 antagonizes the serine phosphorylation of TR-β1 leading to the recruitment of N-CoR/SMRT to TR-β1 and suppression of transcription [79].

Cell Proliferation
The rat tumor pituitary cell line GH3 has been frequently employed as a standard pituitary cell model for assessing TH effects [81]. Indeed, cell proliferation and growth hormone (GH) secretion primarily depend on THs [81] and involve TR-mediated mechanisms, specifically the induction of gene expression [82].
A series of investigations assessed the agonistic and antagonistic properties of BPs in GH3 cell growth both in absence and in presence of T3 (Table 2). BPA, and in particular BPA derivatives, generally promoted GH3 cell proliferation and GH release in the concentration range of 10 −6 -10 −4 M [70,81,83]. In some studies, the agonistic activity was detected exclusively in the presence of T3 [82,84], whereas in others BPA and its substitutes inhibited cell growth with T3, and TH-antagonistic effects appeared to depend on the tested dose and the time of exposure [80,85].
Effects of BPs on cell growth were antagonized by amiodarone, a known TR antagonist [80]. Nonetheless, amiodarone was also reported to act as a slight agonist at low concentrations and antagonist at increasing doses, and BPA and its halogenated derivatives exhibited comparable dose-response curves [76]. In PTC cells, BPA had similar proliferative effects as E 2 [86], and consistent with this finding, co-exposure to E 2 potentiated the increased GH3 cell proliferation (from 190% to 252% after 96 h) by BPA and BPAF [85]. In contrast, TBBPA could not counteract the inhibitory effect of fulvestrant, a strong antiestrogen, on cell growth [81].
Cell growth was further antagonized by U0126, an inhibitor of MEK, the kinase responsible for the activation of ERK in the Raf-MEK-ERK pathway in mammalian cells [87]. Similarly, TBBPA at concentrations in the lower micromolar range caused arrest of cells growth in the G1 or G2 phase, depending on the duration and intensity of the treatment and on cell specific and dose dependent modulations of the Raf-MEK-ERK pathway [87].

Cytotoxicity
MAPKs have an important role in cellular signaling pathways, and the kinases JNKs/SAPKs and p38 MAPKs are often activated by cellular stresses and thus primarily linked to cytokine biosynthesis and induction of apoptosis [88]. Thus, any interference of exogenous chemicals with kinases and phosphatases involved in cellular signaling processes can result in possible cytotoxic effects, including cell death [87].
Similar to cell proliferation, cell viability has been evaluated in cell lines exposed to BPs (Table 2). Cytotoxicity was observed after exposure to BPA and its halogenated derivatives at a concentration range of 10 −5 -10 −4 M; alone and/or with T3 [10,73,76,82,87]. TBBPA was found to produce cytotoxicity 100 times higher than BPA [75] although in other cell models comparable doses of BPA, TBBPA, and TCBPA did not cause changes in cell viability [65,70,89].

Competitive Binding with Thyroid Hormone Binding Proteins
One of the possible mechanisms of BPs for disrupting TH homeostasis is the competitive binding with serum transport proteins due to the structural similarity to T4 and T3. THs mainly bind to three transport proteins in human serum, namely thyroxine-binding globulin (TBG), which is responsible for 75% of the specific T4 binding activity, transthyretin (TTR), and human serum albumin [90]. A few studies tested the capability of these chemicals to compete with THs for binding to TTR, which in non-mammalian vertebrates exhibit a higher affinity for T3 than T4, whereas in human plasma is responsible for only 10% to 15% of the TH transport [91] (Table 2).
Meertz et al. [92] found no TTR binding for 17 polybrominated diphenyl ethers at maximum concentrations confirming that hydroxylation at the para position with at least one adjacent halogen substituent could represent a prerequisite for TTR binding. Indeed, TBBPA was the most potent competitor among the phenolic compounds tested, binding to TTR in a range from 1.6 [84] to 10.6-fold [92] stronger than the natural ligand T4. Moreover, the affinity of TBBPA for TTR was three times greater than that of BPA [70], and this is in line with the higher binding affinity of halogenated derivatives for TRs compared with BPA [10,71]. Actually, the hydroxylated derivatives of BPA also exhibited a strong affinity for TBG, as elucidated in a transport protein-based biosensor assay [93]. Using a fluorescent probe, Cao et al. observed that BPA affinity for TTR and TBG was weaker than T4 by 300 to 2666 fold; hence the current levels of BPA in humans are unable to interfere with T4 serum transport [90].

Perturbation of Thyroid Hormone Uptake
Thyroid hormone uptake into target cells is controlled by membrane bound transporters, such as monocarboxylate transporter (MCT) 8, MCT10, and multiple members of the Na-independent organic anion transport protein (OATP) family [94]. OATP1C1, in particular, shows a high degree of tissue selectivity, being expressed predominantly in brain and testis, and high preference for T4 and reverse T3 as the ligand [95], and it facilitates the transport of T4 across the blood-brain barrier [96]. In different species, MCT expression has been detected in numerous tissues including the brain wherein MCT8 is responsible for the neuronal uptake of T3 [95]. Mutations in the MCT8 gene cause a severe X-linked psychomotor retardation associated with highly elevated serum T3 levels and decreased T4 concentrations whereas the thyroid-stimulating hormone (TSH) values remain in the normal or slightly elevated range levels [97].     Stimulation of growth by all tested chemicals. In presence of T3, potentiating effect on T3-induced growth.
[81] 0.5/1 nM T3; 5*10 −7 -5*10 −6 BPA/BPA-DM; 1*10 −5 /2.5*10 −5 M TBBPA ±1 nM ICI Suppression of induced cell proliferation by the antiestrogen ICI. None of the compounds able to counteract the inhibitory effects of ICI.    In a recent study performed in cells overexpressing the human MCT8 gene, among the several common environmental contaminants classified as flame retardants, pesticides, plasticizers, and others that are suspected to disrupt TH signaling, only BPA was observed to reduce T3 uptakes to around 60% and 40% of the control at concentrations (125 µM) below those that reduced cell viability <80% [94]. This finding is consistent with an earlier study that detected a slight inhibition of T3 transport capabilities of MCT8 by BPA, though at concentrations likely higher than those occurring in vivo [98].

Dysregulation of Gene Expression
In addition to the ability to interfere with TR signaling throughout a direct binding to the receptor, a number of studies observed that BPs may directly affect thyroid gene expression ( Table 2).
At doses as low as 10 −6 M, BPA and its analogues induced expression of transcripts of genes implicated in thyroid cell activity and proliferation (e.g., the Thyroid stimulating hormone-receptor (Tsh-r)), TH biosynthesis (e.g., Tg, Sodium iodide symporter (Slc5a5 encoding NIS), Thyroid-peroxidase (Tpo) and their transcription regulators (e.g., Paired box 8 (Pax8), NK2 homeobox 1 (Nkx2-1), and Forkhead box E1 (Foxe1)) by over 1.5 fold [89,99,100]. Conversely, BPA did not markedly affect transcriptional expression of Slc5a5, Nkx2-1, and Tpo but inhibited NIS-mediated iodide uptake [100]. BPA increased the expression of Tg gene in the presence of increasing TSH amounts, suggesting a potency similar to that of TSH in enhancing Tg-promoter activity [94]. The authors also reported that two anti-estrogens, which alone induced the activity of the Tg promoter, were not able to enhance BPA activity on the Tg promoter, indicating that the effects triggered by BPA do not necessarily involve ER signaling [89]. BPA at the nanomolar range significantly impaired the transcriptome of thyroid cells in a time dependent manner [101]. In fact, whereas short-term exposure to BPA did not cause any relevant transcriptomic changes, long-term exposure, though unable to exert visible damage on cells, determined a slight deregulation of many genes involved in cell proliferation/death, cancer, and DNA repair [101].
BPA inhibited the activities of DIO1 and DIO2 [102], and both BPA and TBBPA markedly dysregulated transcription of Dio3, which is responsible for protection of tissues from TH excess and is the predominant deiodinase expressed in human placenta [103], and hepatic phase II metabolizing genes (Sulfotransferases (Sult1) and UDP-glucuronosyltransferases (Ugt)) [75]. TBBPA, but not BPA, increased expression of the Ttr gene [75], an action at mRNA level that corroborates the competing binding capabilities of TBBPA with TTR [70,84].

Thyroid Disrupting Properties of BPs: in Vivo Studies
In vivo effects of BP exposure on thyroid function/action are contradictory and difficult to compare, as a consequence of the scarce number of studies performed especially in mammals, the different models, and diversity of the experimental conditions, i.e., the chemical used, the time and dose of treatments, the outcomes assessed. In regards the risk of thyroid cancer associated to BPs exposure, the subject remains almost entirely unexplored.

Rodents
In rodents, most of the studies have been performed in pregnant females, and have evaluated the variations of TH levels in mothers and pups following prenatal and/or lactational exposures (Table 3).
In accordance with numerous in vitro studies, BPA can act as a selective TH antagonist on TRβ, inhibiting TH-negative feedback. Indeed, Zoeller et al. [104] and Zhang et al. [105] observed a significant increase of serum T4 levels in pups of both sexes and in female adults, respectively, without any apparent interference on TSH release. In male adult rats, treatment with BPA led to an increase of T4 levels and a reduction of the T3/T4 ratio, suggesting that in exposed animals BPA may impair the peripheral conversion of T4 to T3 [102].
In other experiments, BPA exposure did not produce significant variations in plasma T4 levels [106][107][108][109] or, alternatively, the effects may not endure after BPA removal and metabolism [110][111][112]. It is still unclear whether exposure to BPA and its derivatives can cause hypothyroidism due to limited evidence. A decrease in T4 levels was found in male and female adult rodents [110,113,114] and in rat pups of both sexes [114] or with a sex-specific effect [112,115]. The competition of BPs with TTR, as observed in vitro [70,84,92], resulting in a portion of serum T4 displaced from TTR, could determine an increased rate of T4 metabolism and elimination and the consequent reduction of T4 circulating levels [110].
Perinatal or neonatal exposure to BPA was associated with a significant increase of TSH levels in juvenile males [116] and in females in estrus [117,118], accompanied by a significant increase of GH levels and an impaired sensitivity of the thyroid gland to TSH stimulation, respectively, both of which indicate an alteration of the HPT axis [117]. On the other hand, the reduction of serum T3 or T4 after TBBPA treatment induced feedback stimulation, as suggested by the increased pituitary weight [114], whereas in other studies this was insufficient to affect serum TSH or TH levels, thyroid histopathology, and thyroid weight [107,110].
Adult males treated with BPA showed a decrease activity of hepatic DIO1, coherently with what has been reported in vitro [102]. Moreover, in female adult rats BPA lowered thyroid iodide uptake and thyroid peroxidase (TPO) activity, which are two essential steps in TH biosynthesis, probably due to an elevation of reactive oxygen species (ROS) production. Both NIS and TPO have been found to be sensitive to ROS [119,120], and in particular the decrease in TPO activity could be attributable to the oxidation of this enzyme [106]. Increased expression of pituitary Tshβ was reported in female rat neonates exposed to BPA [117], whereas Silva et al. did not find any significant reduction of Tshβ mRNA levels in treated female rats [106].
To date, it remains unclear whether BPA plays a role in the pathogenesis of thyroid carcinoma. Zhang et al. recently demonstrated that BPA could enhance the susceptibility to TC [105]. Rats pre-treated with N-bis (2-hydroxypropyl) nitrosamine, a drug stimulating thyroid proliferation and promoting a cancerous phenotype [121] and then exposed to BPA and excess iodine for 64 weeks, exhibited a significant increase in incidence of TC and thyroid hyperplasia lesions as well as the up-regulation of ERα in the hyperplasia lesions. The authors speculated that BPA could increase ERα expression in the thyroid, which possibly participated in the proliferation process [105].

Sheep
Sheep are considered a more relevant model to humans than rodents to evaluate fetal exposure to thyroid disruptors and their effects on the mother/newborn thyroid functions because of a similarity in the timing of the ontogenesis of thyroid [122]. In both species thyroxine binding globulin is the main blood transport protein for THs [123], and thyroid system maturation is qualitatively similar in the sheep and human fetuses, although the total maturation time is different (165 days vs. 300 days) [122,124].
Two studies have investigated the relationship of BPA exposure with thyroid function (Table 3). Viguié et al. [122] reported that BPA exposure of pregnant ewes was associated with a transitory hypothyroxinemia of both mothers and their newborn lambs, with a significant reduction of both circulating total T4 (TT4) and free T4 (FT4), findings in agreement with rodent studies [111,112]. In a following study, the authors confirmed alterations of gestational thyroid function, observing a significant reduction of FT4 and total T3 (TT3), but not TT4, in pregnant ewes treated with environmentally-relevant BPA concentrations via subcutaneous and dietary routes of administration [123]. After subcutaneous administration, the maximum serum concentration of BPA obtained was significantly higher (0.4 nmol/mL vs. 0.1 nmol/mL) and more prolonged than after dietary administration [123].

Zebrafish
Numerous studies have been published on the use of zebrafish (Danio rerio) to explore the effects of EDCs on the thyroid, due to several advantages: a short life cycle, high rates of production, real-time observations during the entire embryonic development, and high conservation of the molecular mechanisms regulating thyroid development with those of mammals [125,126]. The early life of fish, in particular, is acknowledged as highly sensitive to the effects of EDCs [127].
Coherently with results observed in vitro and in rodents, BPs may disturb TH homeostasis and gene expression in zebrafish embryos/larvae (Table 3).
Positive [37,38,128] and negative [37,38,[128][129][130][131] associations between exposure to BPs and T3 and/or T4 levels have been reported, depending on the chemical used, the dose tested, and the time of exposure. Reductions in T4 concentrations, when accompanied with higher TSH contents, may compensate hypothyroidism in zebrafish larvae and stimulate TH synthesis [38,129]. Some experiments observed an interaction between TH levels and sex [129,131]. Tang et al. showed a reduction in whole-body TT4 and TT3 levels but not a significant variation of ratio TT3/TT4, which indicates the relative normal TH homeostasis [130].
Similarly, BPs disrupting effects on thyroid gene expression vary according to the different experimental conditions, especially the duration of exposure [126]. Hence, transcription levels of genes implicated in thyroid cell function and proliferation (Tsh-r), TH activity (Trα, Trβ), and transport (Ttr) can be up- [37,68,126,130,132,133] or down-regulated [38,68,126,128,130,133]. The transcription of Hematopoietically expressed homeobox (Hhex) was up-regulated in larval fish following exposure to BPA or BPF, although it is important to note that the Hhex gene is expressed in early life, contributing to differentiation and development of the thyroid gland, as well as of other organs, such as the pancreas and liver [37].
Increased [89,128,129] or decreased [130,132,134,135] expression of Slc5a5, Tpo, Pax8, and Tg transcripts was dependent on the dose and window of exposure. Additionally, BPA and BPS appeared to interact with PAX8 and thyroid transcription factor 1 (TTF1) in silico [135]. Genes such as Tpo, Tg, and Slc5a5 have binding sites for PAX8 or TTF1 on their enhancer or promoting regions. Differences of interactions between BPs and the transcription factors could be attributable to stimulation or inhibition with varying BPs doses, and produced as final effect altered expression of the genes controlled by PAX and TTF1 [135].
Exposure to BPA and BPA analogues further induced transcription of genes involved in TH metabolism, i.e., Dio1 and/or Dio2 [37,129,132,134], which are implicated in activation/inactivation of T4 and in conversion of T4 to T3 in peripheral tissues, respectively [136,137], and Ugt1ab [37,38,129].
Notably, co-treatment with T3 appeared to reverse or eliminate thyroid disrupting effects of TBBPA on THs levels and gene transcription in zebrafish larvae [128], whereas a combined exposure of BPAF and sulfamethoxazole, an antibiotic used especially in aquaculture, produced more pronounced changes in transcription levels [134].   Higher T4 levels in BPA-exposed animals, whereas no variations of T3 levels. [106] Thyroid iodine uptake/TPO activity Significant reduction of TPO and NIS activity.

ROS generation
In BPA-treated animals, significant generation of H 2 O 2 generation in the thyroid. In the reproduction study, decrease of T4 levels in pups of both sexes, and increase of T3 levels only in females.
In the subacute toxicity study, significant decrease of T4 and increase of T3 levels in males, whereas in females parallel though not significant trends. [114] Necropsy Dose-dependently increase of pituitary weight in male pups in the reproduction study. No effects in the subacute toxicity study.

Histology
No changes observed in the histology of the pituitary gland both in the reproduction and in the subacute toxicity studies. In the group exposed to DHPN, statistical significances among all groups, and the KI group had the heaviest thyroid weights. In the group not exposed, no significant differences were found among groups. [105]

Histology
In the group exposed to DHPN + KI + 1000BPA, all thyroids had a tumor or focal hyperplasia. Significant difference in the total number of hyperplasia lesions among all groups of animals exposed to DPNA.

Chemiluminescence immunoassay/ELISA
In the groups exposed to DHPN, TSH was significantly higher in the KI group than in the controls and the highest FT4 concentration was in the BPA1000 group. In the groups not exposed to DHPN, the highest concentration of TSH was in the controls whereas FT4 increased with increasing doses of BPA.

Western blotting detection
In the groups exposed to DPNA, increased protein levels of ERα in the BPA250 and BPA1000 groups compared to the control.

Thyroid Disrupting Properties of BPs: Human Studies
Perturbations in THs parameters consequent to exposure to BPA have been documented in humans, i.e., the general population, pregnant women, or occupational settings, although the study design, predominantly cross-sectional, does not allow establishment of any causal relationship. Research has highlighted positive, negative, or null associations with T4 levels, whereas a few prospective birth cohort studies suggest that prenatal BPA exposure may modify THs normal serum concentration in a sex-specific manner. Several investigations have demonstrated BPA-induced disruption of thyroid function by altering serum TSH levels. This effect could occur from a direct action of BPA on the pituitary gland via the estrogen signaling pathway or, alternatively, from a transient increase of T3 or T4 production that could lead to a feedback mechanism and the subsequent release inhibition of TSH. Overall, discrepant results among studies may be attributable to BPA levels, time of exposure, iodine intake, differences in age, ethnicity, diet, socioeconomic status, and the determination methods of THs, while at present the potential role of BPs in thyroid carcinogenesis in humans remains to be deeply explored (Table 4). It is noteworthy that the absence of adjustment for other confounding factors such as co-exposure to other EDCs makes the overall evaluation of thyroid dysfunction related to BPs exposure complex. Furthermore, a comparison between effects observed in animal models with those reported in epidemiological studies is complex given different serum T4 half-lives (12-24 h in rats vs. 5-9 days in humans), metabolic pathways of BPs, and doses of exposure, which in humans is more likely to be chronic and low level [138].
In agreement with results from studies in animals [104,105,108], exposure to BPA led to TSH release/suppression independent of alterations in circulating THs levels [143,145,147,148,150,152] or, less frequently, was associated with variations of serum T4 levels [139,149].
There was no association of BPA with hypothyroidism in Japanese women with a history of recurrent miscarriages [151], nor any significant relationship between serum TBBPA in Korean infants with congenital hypothyroidism and THs levels [140]. Conversely, middle-aged and elderly Chinese with overt or subclinical hyperthyroidism had higher urinary BPA than euthyroid subjects [151], and an increased content of urinary BPA was also observed in obese adults undergoing a diet program or bariatric surgery compared to lean controls, probably due to differences in food intake [152].
Sex-related differences in the relationship between BPA and THs were reported both in the general population [142] and in newborns [143,148], coherently with studies performed in rats [112,115], and possibly attributable to a less efficient ability to metabolize BPA, i.e., a reduced expression of uridinediphosphate-glucuronosyltransferase 2B1 in male compared to female livers [153], or to a different androgen-related metabolism of BPA [154].
The interactions between BPA and THs during pregnancy and fetal development have been recently studied. The association between BPA and TSH levels in newborns was stronger when the time elapsed between the two measurements was shorter [143,148], suggesting that specific windows of exposure may influence susceptibility to BPA or, alternatively, that a transient effect on the HPT axis may occur, as shown in rodents [111,112]. However, the inverse association of BPA-TSH in pregnant women detected through repeated measures as well as stratified analyses by visit could indicate the absence of a specific window of vulnerability [139].

Association with Thyroid Diseases
The influence of BPA on thyroid autoimmunity is controversial (Table 4). Whereas urinary BPA concentration was associated with variations of TH levels both in children and adults of both sexes, independent of serum thyroglobulin antibodies (TgAb) and thyroid peroxidase antibodies (TPOAb) [149,155], another study found a positive relationship between serum BPA and TPOAb in men and women [156]. Moreover, a significant negative correlation of serum BPA with FT4 in male subjects was found only after exclusion of subjects with positive thyroid antibodies, suggesting that TgAb might be a mediator of the relationship between BPA and FT4 [144].
Kim and Oh reported a slight positive correlation between serum TBBPA and thyroid-stimulating hormone receptor antibodies, indicative of metabolic diseases, in mothers of infants with congenital hypothyroidism, suggesting that brominated derivatives of BPA might affect thyroid function status [140].
Recent investigations have explored the role of BPs as risk factors of occurrence of thyroid nodules (TNs), palpably and/or ultrasonographically discrete lesions, distinct from the surrounding parenchyma of the thyroid gland, which are either benign or malignant [157]. A study reported no association between BPs and higher risk of TNs in adult females [147], whereas Wang et al. observed an inverse correlation of urinary BPA and the risk of multiple TNs but not of solitary TNs in schoolchildren [158]. On the other hand, a significant near linear association between BPA and higher risk of TNs was observed exclusively among participants positive for TgAb and TPOAb [159]. Both urinary BPA and creatinine-adjusted BPA levels were higher in Chinese women with TNs than those without TNs [159], which is consistent with the increased urinary BPA contents in patients with nodular goiter and PTC (160), while median urinary BPA levels were lower in the cases compared to controls among women from Cyprus and Romania [147]. In the study by Zhou et al. [160], which was aimed to investigate the relationship between BPA and iodine exposure with nodular goiter and PTC, sex-specific associations were shown, with higher concentrations of BPA in women than in men affected by PTC and nodular goiter, and a lower urinary BPA content in the female PTC group than the female nodular goiter group, probably due to differences in BPA elimination rates. Marotta et al. recently found a significant dose-independent correlation between BPAF and the risk of differentiated TC in subjects with TNs. Of note, this association was not related to an increase of TSH levels, indicating a potential direct mutagenic action of BPAF on thyroid cells [161]. Table 4. Summary of human studies on the association between bisphenols exposure and thyroid parameters.

Study Design
Country Study Sample Sample Size (N) Age Principal Results

≥18
The obese group had higher urinary levels of BPA. Positive relationship of urinary BPA with serum TSH in lean subjects.

Discussion
The thyroid is highly susceptible to environmental pollutants, which may act as either genotoxic or non-genotoxic carcinogens [147]. BPA is a widespread chemical detected in urine specimens of the majority of adult populations. BPA analogues and derivatives are ubiquitous contaminants, measured in environmental and biological matrices, exhibiting a thyroid disrupting potential comparable and even stronger than BPA. The mechanisms of BPs action on THs are complex and need to be still elucidated.
Overall, the in vitro studies demonstrate that BPs may bind to TRs, acting mainly as TR antagonists, but also as agonists or without exerting any effect on TH signaling. Similarly, different patterns of Trβ expression following BPs exposure were observed in in vivo models. THs and their receptors regulate many important processes such as proliferation, differentiation, and apoptosis, and since TRβ is the major isoform in the thyroid, it can be hypothesized that disruption of its expression, leading to abnormalities in T3-induced transcriptional activity, could be involved in tumorigenesis [72,162].
In vivo experiments, supported by in vitro evidence, highlighted the ability of BPA and its substituting chemicals to affect thyroid follicular cell gene expression, particularly transcriptional levels of those genes encoding for factors involved in THs synthesis (TPO, NIS, Tg, PAX8). Up-regulation of Tg and Slc5a5 transcript levels may promote thyroid development to compensate for the depressed T4 concentration, as also reported for polybrominated diphenyl ethers [163]. Transcriptional levels of deiodinases were more elevated in exposed zebrafish, in accordance with a study reporting that hypothyroidism caused by EDCs is associated with higher activity and expression of Dio2 [164]. On the other hand, a recent study reported a reduction of liver DIO1 activity in BPA-treated adult rats [102], which is a finding worth of note as decreased expression of DIO1 was observed in nearly all PTCs and is likely an early event in malignant TC [165].
In rodents and in two different types of cell lines, BPA up-regulated Pax8 transcripts, suggesting a role of BPA in increasing Pax8 expression independent from the cellular context [89]. PAX8 is a cell-lineage-specific transcription factor that has been mainly characterized in the thyroid gland for its role in thyrocyte differentiation, and it has been revealed as a potential diagnostic marker for several cancer sites including TC [166].
TSH should represent an effective index of activation of the HPT axis to evaluate central effects of xenobiotics on thyroid function through measurement of TSH secretion or expression as a compensatory mechanism for maintaining TH homeostasis. Moreover, TSH levels are an independent predictor of thyroid nodule malignancy regardless of age, sex or family history [6]. Increased expression of TSH and TSHβ observed in vivo was also reported after exposure to pesticides and halogenated chemicals in fish [163,167] suggesting that elevated production of TSH could represent one of the mechanisms of action of BPs, as already hypothesized for other EDCs (9). In pituitary cells, BPA and E 2 could further induce release of TSH desensitizing the response to thyrotropin releasing hormone from hypothalamus [118]. In contrast, humans and pregnant ewes exhibited hypothyroxinemia after BPA exposure without significant modifications of TSH, whilst other epidemiological studies reported a decreased TSH production, probably consequent of a direct action of BPA on pituitary gland through estrogen receptor signaling or of a feedback mechanism triggered by BPA-mediated perturbations on circulating T3 and T4 [139].
The frequency of chronic autoimmune Hashimoto's thyroiditis, the most common cause of primary hypothyroidism in western countries, has increased in the last two decades, and a variety of factors such as tobacco smoking, iodine and selenium intake, and exposure to EDCs, may contribute to the elevated incidence by interacting with susceptibility genes (6). Autoimmune thyroiditis may coexist with TC [168], and a recent meta-analysis demonstrated that this condition predisposes patients to the development of the papillary histotype [169]. Thyroid autoantibodies were reported to be positively associated with the level of urinary BPA, therefore subjects with thyroid autoantibodies positivity, as characterized by immune dysfunction and a lower ability to eliminate damaged cells, are probably more vulnerable to the effects of BPs on TNs [159].
It cannot be excluded that exposure of thyrocytes to BPA involves hydrogen peroxide generation due to an elevated activity of a calcium-dependent NAPDH oxidase (DUOX) [106]. TPO is a key enzyme in the synthesis of THs, catalyzing, through the cofactor H 2 O 2 , the iodination of tyrosyl residues in Tg [106]. Thus, the increased oxidative stress in the thyroid gland, which is related to a reduction of TPO activity, corroborates the negative correlation between TPO and DUOX2 in thyroid nodular lesions [170]. Furthermore, oxidant/antioxidant balance was recently reported to be impaired in children affected by autoimmune thyroiditis, though it is unclear whether oxidative stress is the real cause of the disease or the likely consequence of exposure to EDCs, including BPA [155].
Finally, BPA is potentially linked to excess iodine in the pathogenesis of the nodular goiter and TC in animals [105] and humans [160]. High urinary iodine is a risk factor for the development of benign TNs and PTC [171], being associated with reduced expression of NIS, an early abnormality in the pathway of thyroid cell transformation, and increased occurrence of BRAF mutations [172], which are both hallmarks of differentiated TC [173].

Conclusions
This review aims at evaluating the extensive body of experimental and human studies that in the last two decades have attempted to explore the effects of BPA, its substitutes, and its halogenated derivatives on the thyroid at different levels. Despite the variety of approaches applied and the heterogeneous and sometimes even conflicting results from the examined studies, a series of interesting indications supports the hypothesis of a role of BPs in interfering with the normal thyroid function. Although the toxicity pathways of BPs on the thyroid need to be further elucidated, BPA analogues and halogenated derivatives do not emerge as safer alternatives to BPA in term of TH disruption. There is evidence that BPs alters THs circulating levels, inhibiting TH-negative feedback, act as selective TR antagonists, and interfere with expression of genes involved in thyroid stimulation, TH synthesis, TH activity, and TH transport and metabolism. Several reported findings, mainly from experimental studies, are, however, rather inconsistent, while the association of BPs exposure with thyroid cancer is so far almost unexplored. The lack of uniformity in experimental methodology, as well as substantial differences in populations investigated in epidemiological studies, do not allow definitive conclusions to be drawn. Standardized in vivo, in vitro, and in silico studies are recommended to evaluate the physiopathology of the damage associated with exposure to environmentally relevant levels of BPs, identify other potential molecular targets, and clarify the structure−activity relationship of BPs. At the same time, large population-based human studies with prospective designs and repeated measures of urine BPs concentrations and thyroid volume over time, as well as an accurate control of confounders, should be performed for the assessment of the temporal relationship between markers of exposure and long-term effects.  Thyrotropin-releasing hormone  Trh-R  Thyrotropin-releasing hormone receptor  TSH  Thyroid stimulating hormone  Tshβ  TSH-specific β subunit  Tsh-r  Thyroid stimulating hormone-receptor  TTR  Transthyretin  TT3  Total triiodothyronine  TT4  Total thyroxine  Ugt Uridine diphosphate -glucuronosyltransferase