Impact of Antioxidant Natural Compounds on the Thyroid Gland and Implication of the Keap1/Nrf2 Signaling Pathway.

BACKGROUND
Natural compounds with potential antioxidant properties have been used in the form of food supplements or extracts with the intent tfor prevention and or treatment of various diseases. Many of these compounds The cytoprotective Nrf2 pathway can be activated the cytoprotective Nrf2 pathwayby many of these compounds. Besides, Ssome of themse compounds are known to impact the thyroid gland, often with potential for side-effects, but in other instances with potential utility for in the treatment of thyroid disorders.


OBJECTIVE
In view of recent data regarding the multiple roles of Nrf2 in the thyroid, this review summarizes the current bibliography on natural compounds that can have an effect on thyroid gland physiology and pathophysiology, and it discusses the respective potential implication of the Nrf2 system in the respective mechanisms.


METHOD & RESULTS
Literature searches for articles from 1950 to 2018 wereas performed in PubMed andor Google Scholar using relevant keywords about phytochemicals, Nrf2 and thyroid. Natural substances were categorized into phenolic compounds, sulfur-containing compounds, quinones, terpenoids, andor under the general category of plant extracts. For individual compounds in each category, respective data were summarized, as derived from in vitro (cell lines), preclinical (animal models) and clinical studies. The Mmain emerging themes were the following: phenolic compounds often showed potential to affect the production of thyroid hormones; sulfur-containing compounds impacted the pathogenesis of goiter and the proliferation of thyroid cancer cells; while quinones and terpenoids modified Nrf 2 signaling in thyroid cell lines.


CONCLUSION
Natural compounds that modify the activity of the Nrf2 pathway should be evaluated carefully, not only for their potential to be used as therapeutic agents for thyroid disorders, but also for their thyroidal safety when used for the prevention and treatment of non-thyroidal diseases.


INTRODUCTION
Food contains chemical compounds with potential ability to impact human health, either positively or negatively, by altering enzymatic reactions, cell signaling, and other biological processes [1]. Consumption of vegetables, fruits and whole grains can provide non-nutrient bioactive compounds, also known as phytochemicals, which can be involved in stimulation of the immune system, hormone metabolism, scavenging of oxidants, regulation of genes implicated in cell proliferation or apoptosis, antiviral and antibacterial processes, etc. Such compounds can potentially reduce the risk of diabetes, cardiovascular diseases, Alzheimer's disease and other aging-related disorders [2][3][4]. Changing dietary habits in favor of plant-based foods may thus provide health benefits beyond the provision of basic nutrients. Adding fruits and vegetables rich in phytochemicals into a diet is therefore promoted as a strategy for the prevention of chronic diseases. One plausible contributing mechanism is that phytochemicals can provide protection against reactive oxygen species (ROS), thereby reducing the risk of various pathological processes and diseases related to oxidative stress [5]. According to their mechanisms of action that confer protection against ROS, phytochemicals with antioxidant activity can be classified into (i) direct antioxidants, which are redox-active and have short half-lives; (ii) indirect antioxidants, which induce intracellular pathways that can ultimately protect against oxidative stress; and (iii) bifunctional antioxidants, which have both properties [6]. However, the beneficial effect of natural compounds is often overemphasized, whereas data about potential toxicities associated with excess intake are in general scarce. Taking into account the classic principle of toxicology that "the dose makes the poison", careful consideration should be given to the effects of the various phytochemicals that are mainly consumed in the form of food supplements, often without having undergone rigorous toxicological or clinical safety testing. Consequently, low levels of phytochemicals present in fruits and vegetables may offer health benefits, but these same compounds may be ineffective or even unsafe when consumed at higher doses [7].
The transcription factor Nrf2 (nuclear factor erythroid 2-related factor 2) is a major effector of antioxidant cellular responses and a target of phytochemicals, with protective effects against oxidative stress-related diseases [8,9]. These phytochemicals are thus considered indirect antioxidants. Under basal conditions, Nrf2 is sequestered in the cytoplasm by the cysteine-rich protein Keap1 (Kelch-like ECH-associated protein 1) that facilitates its ubiquitination and its subsequent degradation by the proteasome [10]. The sulfhydryl (-SH) groups of Keap1 cysteines act as "sensors" [11] that react with oxidative or electrophilic stressors or pharmacological inducers of the Keap1/Nrf2 pathway to provoke conformational changes in Keap1 that render it incapable of facilitating the degradation of Nrf2 molecules. Thus, de novo transcribed and translated Nrf2 can accumulate and, after various post-translational modifications, can enter the nucleus and induce the expression of its target genes by binding to conserved ARE (Antioxidant Response Element) sequences in their regulatory regions. Classic target genes of Nrf2 comprise cytoprotective and antioxidant enzymes such as heme oxygenase-1, NAD(P)H-quinone oxidoreductase-1, glutamate cysteine ligase, glutathione peroxidase, glutathione S-transferases, thioredoxin, thioredoxin reductase, and many others, including mediators of secondary antioxidant mechanisms, such as several proteasome subunits [12]. Pharmacological inducers of Nrf2 interact with specific Keap1 cysteines and activate the Nrf2 pathway [13]. A very well characterized Nrf2 activator that can be found in its precursor form in natural products (cruciferous vegetables and especially broccoli) is the isothiocyanate sulforaphane [14][15][16]. Sulforaphane-rich broccoli sprouts extracts have been extensively used in series of preclinical and clinical studies, with the majority of them addressing the field of cancer chemoprevention, where sulforaphane has been proven effective against both genetic and chemical models of carcinogenesis in rodents [17]. As the use of sulforaphane in clinical trials has been expanding, it has been shown to enhance the detoxification of air pollutants [18], repress hepatic glucose production [19], and even ameliorate symptoms of autism spectrum disorder [20].
Nrf2 is activated not only by sulforaphane but by many other phytochemicals. Broadly, natural compounds that can directly react with Keap1 cysteines or can induce low-level oxidative stress, may have the potential to induce the Nrf2 pathway, possibly along with other pathways [21]. Because very few studies have specifically focused on the potential impact of phytochemical-mediated Nrf2 modulation on the thyroid gland and its pathologies, this minireview focuses on redox-active natural compounds that may affect the thyroid gland and may implicate the Nrf2 pathway as a mechanism of action.

Thyroid Gland Physiology and Hormone Synthesis
Normal anatomy and function of the thyroid gland, and normal concentration of thyroid hormones, are crucial prerequisites for optimal growth and development. Disorders of thyroid gland function, namely hyperthyroidism (hyperfunction) and hypothyroidism (hypofunction), as well as of thyroid anatomy, namely goiter (thyroid enlargement), are among the most common endocrine diseases. Though they can affect subjects of all ages, it is well documented that their prevalence increases with age [22][23][24][25][26][27]. Many substances and conditions are able to affect thyroid hormone biosynthesis and metabolism, including environmental and dietary factors, particularly in areas characterized by nutritional iodine deficiency [28,29].
The thyroid gland synthesizes and secretes two main thyroid hormones, the highly active triiodothyronine (T3) and its much less active precursor, thyroxine (T4). For thyroid hormone synthesis to take place, it is necessary that the pituitary hormone thyrotropin (thyroid-stimulating hormone, TSH) binds to its receptor on the basolateral membrane of thyrocytes. The synthesis and secretion of TSH by the pituitary are stimulated by thyrotropin-releasing hormone (TRH), which is secreted by the hypothalamus into the hypothalamic-pituitary portal vein system. In turn, optimal serum levels of thyroid hormone are ensured principally due to the inhibitory effects of circulating T4 and of T3 on the secretion of TSH and TRH (the so-called negative feedback loop) [30].
Another essential aspect for thyroid hormone synthesis is the particular organization of thyroid tissue and especially the polarity of thyroid follicular cells. The thyroid is comprised of follicles, spherical structures that represent the gland's secretory and functional units. Every thyroid follicle consists of grouped thyroid epithelial cells forming a spherical monolayer around a lumen filled with a viscous substance called colloid. Thyroid hormone synthesis takes place in the colloid, facing the apical plasma of thyrocytes. The opposite, basolateral side, is in contact with a network of blood capillaries, into which, the final synthesized thyroid hormones are secreted.
From the same capillaries, iodide (I -, the negatively charged anionic form of iodine) is taken up into the cells through the sodium (Na + )/iodide(I -) symporter (NIS), in order to be used for thyroid hormone synthesis [31]. Iodide is then transported towards the follicular lumen ("iodine efflux"), in a process involving apical plasma membrane channels such as pendrin and anoctamin [32][33][34].
In addition to iodide, the other main structural component required for thyroid hormone synthesis is thyroglobulin (Tg), a high molecular weight protein synthesized by the endoplasmic reticulum of thyrocytes. Iodide and Tg are brought together via the actions of thyroid peroxidase (TPO), an oxidoreductase enzyme localized at the apical plasma membrane that catalyzes a number of consecutive reactions. In the first step, iodide is oxidized by TPO in a reaction that requires hydrogen peroxide (H 2 O 2 ), actively generated in thyrocytes by the enzyme, dual oxidase (DUOX, a NADPH oxidase) [35]. Oxidized iodide then binds to tyrosyl residues of Tg in a process called Tg iodination or iodine organification; this process results in the formation of monoiodotyrosines (MIT) and diiodotyrosines (DIT) on the Tg molecule. In the next step, T3 and T4 are formed by the coupling of two iodotyrosines, one MIT and one DIT for T3, or two DIT for T4 [36]. When demand for thyroid hormone secretion is low or absent, Tg is stored in the follicular lumen. Once needed, Tg enters the thyrocytes by micropinocytosis, whereby it is packed in vesicles. Inside the thyrocytes, these vesicles fuse with lysosomes, facilitating a proteolytic breakdown of Tg mediated by enzymes like cathepsin B, D and L. This is followed by degradation reactions catalyzed by lysosomal dipeptidase I, dipeptidylpeptidases I and II and N-acetyl-l-phenylalanyl-L-tyrosine hydrolase [37,38]. Together, these reactions lead to the release of T3 and T4 from Tg. MIT and DIT undergo deiodination by the enzyme dehalogenase 1 (DEHAL1) that recycles iodide to be reused by thyrocytes, thus economizing on this trace element [39].

Role of Nrf2 in the Thyroidal Antioxidant Response
In general, tissues that generate high amounts of free radicals (e.g., muscle) or have functions in detoxification (e.g., liver) are more susceptible to disorders caused by oxidative stress. As mentioned, in order to synthesize thyroid hormones, follicular cells require a constant physiological generation of H 2 O 2 as the initiator of a series of oxidation reactions that result in the organification of iodide by the iodination of Tg. It is therefore not surprising that defense against oxidative stress is particularly vital for thyrocytes. Yet, in comparison to other tissues, relatively little is known about the role of proteostatic and antioxidant systems in the thyroid [24]. While a minimal amount of oxidative load is necessary in order for thyrocytes to proliferate and function normally [40,41], intrathyroidal oxidative stress occurs during pathological conditions like goitrogenesis or iodine deficiency, and this in turn activates the thyroid's antioxidant defenses [42][43][44][45].
Recent work has shown that Nrf2 signaling has a central role in the antioxidant response of the thyroid gland [43]. Nrf2 regulates the expression levels of genes like thioredoxin reductase 1 and glutathione peroxidase 2 [43], which are known to maintain homeostasis in thyroid cells by ameliorating oxidative insults [46,47]. In addition, Nrf2 protects the thyroid from oxidative damage induced by iodide overload [43]. Importantly, Nrf2 coordinates antioxidant defense in the thyroid gland with Tg production and iodination, and it can actually regulate the transcription of the gene encoding Tg by binding to two ARE sequences in a conserved enhancer [43].
Regulation of Tg expression by Nrf2 makes biological and physiological sense, because both the Tg folding process and the iodination of Tg are mediated by oxidation reactions, as discussed above. For normal follicular function with optimal Tg economy, it is necessary to coordinate the regulation of antioxidant defense by Nrf2 with Tg production and iodination. Under normal conditions, Nrf2 prevents excessive Tg iodination by reducing ROS levels via a positive control of the expression of antioxidant genes. The role of Nrf2 in protection from oxidative stress is even more important during iodide overload: as a response to high iodide intake, Nrf2 activates a transcriptional program that limits ROS levels, preventing oxidative damage to biomolecules in follicular cells [43,48,49]. Other genes involved in the specification, differentiation or function of thyroid follicular cells may also be impacted by Nrf2. Yet, in contrast to Tg, such regulation is apparently indirect and could result from the general redox status of the cell and/or from the crosstalk of Nrf2 with other pathways, as has been described also in other tissues [50][51][52]. Interestingly, when high doses of iodide were provided to mice lacking NIS, Tg iodination and thyroid hormone synthesis proceeded successfully to a certain degree [53]. In these conditions, a decrease in the mRNA levels of Nrf2 and of various antioxidant genes was also observed. Together, these data suggest that elevated TSH levels lead to a repression of Nrf2 signaling in order to increase ROS levels, which in turn favors the oxidative reactions necessary for Tg iodination and thyroid hormone synthesis [43,54].
Beyond its protective and physiological roles in normal thyroid, Nrf2 signaling has also been found to be activated in thyroid carcinomas, where it confers protection against oxidative stress and resistance to antineoplastic agents [55,56].

Research aims and Strategy
The aim of this paper was to review the literature regarding natural substances with the potential to affect the physiology or pathophysiology of the thyroid gland, with a focus on those that are known to modulate the activity of Nrf2 as a mechanism of antioxidant response. The ultimate goal was to assess the need for future studies on such compounds regarding their thyroidal effects and the estimation of doses that can be beneficial or at least safe for the thyroid gland.
The literature search and selection of articles for this minireview was conducted using Pubmed and Google Scholar for articles published in English from January 1950, through July, 2018. The following search terms were used: "natural compounds" or "natural substances" or "dietary flavonoids" or "phytochemicals" or "extracts" or "nutraceuticals" or "natural products", in combination with "Nrf2" or "Nrf2 activation" or "Nrf2 inhibition" or "Nrf2 modulation" or "Nrf2 pathway", and "thyroid" or "antithyroid" or "goiter" or "goitre" or "goitrogenic". To determine the relevance of publications, abstracts and/or titles were screened. Next, the full texts of selected studies were reviewed and all relevant references cited in these studies were also screened and included when pertinent. For compounds shown to have an effect on Nrf2, the literature was checked for any known activities in the thyroid, and vice versa.
In this manner, we identified, in particular, studies providing evidence regarding different effects of natural compounds on thyroid gland morphology and/or function. Regarding the involvement of Nrf2 as a mechanism of action of these compounds, we included only studies providing experimental evidence; papers with speculative but experimentally unsupported mechanisms were excluded.

PHENOLIC COMPOUNDS
Plants have the ability to synthesize secondary metabolites, including phenolic compounds [57]. In plants, these compounds play various roles during interactions with the environment, such as in the protection against ultraviolet radiation, in the attraction of seed-dispersing animals or pollinators, as well as in the defense against predators, parasites and pathogens [58]. Aside from providing plants with adaptive or survival strategies, phenolic compounds have commercial significance, because they are being used in the production of fibers, dyes, oils, flavoring agents and perfumes, and also as a source of new antibiotics, natural drugs, herbicides and insecticides [59,60]. This makes phenolic compounds a main research focus for many scientists globally. Currently, around 8000 phenolic compounds are known; they demonstrate wide structural diversity, ranging from simple phenolic acids to polyphenols (like flavonoids) to polymeric compounds. Flavonoids comprise the largest and most studied group among phenolic compounds; they are present in vegetables, fruits, spices, seeds, nuts, red wine and tea, and they are usually classified into isoflavonoids, flavanones, flavonols, flavanols (catechins), flavones, anthocyanidins and chalcones [61].
Phenolic compounds originating from medicinal herbs and dietary plants have various biological activities, including modulation of enzyme activities involved in oxidation and detoxification, as well as antiviral, antibacterial, anticancer, and anti-inflammatory effects[62].

Effects of Phenolic Compounds on Thyroid
Numerous conditions and chemical substances can alter thyroid function and affect the secretion of thyroid hormones and/or their availability to target tissues. Flavonoids, the largest class of phenolic compounds, present in dietary plants and medical herbs, can interfere with thyroid hormone economy [29]. For example, in one of their first reported antithyroid effects, rats fed arachidoside and anacardioside pigments isolated from nuts developed goiter [63].
Exposure to flavones apigenin, chrysin, vitexin and baicalein, present in parsley, cherries, thyme, olives, tea and broccoli, can inhibit TPO, the essential enzyme for thyroid hormone synthesis, which may lead to hypothyroidism ( Table 1). As mentioned above, TPO catalyzes the oxidation reactions involving the iodide ion, the iodination of Tg, as well as the coupling of iodotyrosine residues on the Tg molecule. When TPO activity is inhibited by phenolic compounds, reducing thyroid hormone synthesis, a compensatory increase in TSH may be observed; this may lead to goiter, especially when these compounds are consumed in high quantities [28].
The flavanols quercetin, fisetin, kaempferol, morin, myricetin and rutin, present in a wide range of food sources such as kale, onions, tomatoes, cherries, apples and red wine [5,64], together with the flavanones naringenin and naringin, can also inhibit tyrosine iodination by TPO with varying potencies. TPO inactivation by kaempferol, quercetin, naringenin and fisetin is achieved by covalent binding of reactive resorcinol radicals to catalytic amino acid radicals on TPO compound II, one of the intermediate oxidized states of the TPO enzyme formed during the coupling reactions [65]. Myricetin and naringin inhibit TPO non-competitively by different mechanisms, whereby inhibition of tyrosine iodination results from the binding of a substrate and inhibitor to different enzymatic forms or sites. They interact with TPO compound I (another oxidized intermediate, formed during both the iodination and coupling reactions), as well as with compound II, but not with native TPO or with the enzymatic iodinating TPO species (another intermediate, resulting from the reaction of compound I with iodide). On the other hand, the isoflavone biochanin A can serve as an alternative substrate for iodination; the competition between substrate and tyrosine for the enzymatic iodinating TPO species initially leads to incomplete blockade of tyrosine iodination, because the affinity for biochanin A is higher [65,66].
The soybean-derived isoflavones genistein and daidzein, in concentrations similar to those present in the serum of subjects consuming soy derivatives, are able to inactivate TPO and to inhibit iodine organification in vitro [77]. Isoflavones can also interfere with thyroid hormone transport proteins and with 5´-deiodinase type I (5DI) in peripheral tissues, leading to changes in thyroid hormone activity at cellular level. The enzyme 5DI catalyzes the conversion of T4 to T3, the biologically active hormone [91]. In orchidectomized middle-age rats, total serum T4 and T3 levels decreased after treatment with daidzein and genistein [74]. Besides TPO activity, discussed above, deiodination via 5DI can also be affected by baicalein, quercetin, kaempferol and rutin [69]. Xanthohumol FRTL-5 cells [89]; HEK293 cells [90] ↑IU [89]; iodothyronine deiodinases I [90] Effects listed in the right column and separated by semicolons have been derived from corresponding models shown in the middle column and separated by respective semicolons. 5DI, type I 5´-deiodinase; AC, anticancer; R, rat; ↑, increase; ↓, decrease; P, porcine; I , inhibit; G, goitrogenic; AT, antithyroid; HypoT, hypothyroidism; IU, iodide uptake; TH, thyroid hormone; M, mice; ATC, anaplastic thyroid carcinoma; FTC, follicular thyroid carcinoma; PTC, papillary thyroid carcinoma; MTC, medullary thyroid carcinoma.
Hesperetin and naringenin are flavones from citrus fruits. Being potent antioxidants, they can improve the cardiovascular and metabolic status; however, they may also interfere with thyroid hormone economy. When administered to aged rats at high doses, serum TSH increased in response to naringenin treatment, while T4 remained unaltered in response to either naringenin or hesperetin treatment [78].
Quercetin is a flavonoid that is naturally abundant in a broad range of vegetables and fruits; it is also available as a dietary supplement with purported anti-inflammatory, antioxidant and antiproliferative properties. However, quercetin might carry a potential toxicity risk when consumed in excessive quantities, and it can interfere with thyroid function by inhibiting thyroid cell growth and iodide uptake. Iodide uptake inhibition is the result of downregulation of NIS gene expression [84]. Quercetin may also downregulate the expression of Tg and of the TSH receptor [83].
Besides inducing changes in thyroid function upon excessive intake, flavonoids can also have a crucial role in carcinogenesisassociated events, including invasion, kinesis modulation, apoptosis induction and cell cycle regulation. Cell culture studies have shown that flavonoids have an inhibitory effect on the proliferation of thyroid cancer cells by interfering with key enzymes in cell signaling and cell proliferation pathways, such as protein tyrosine kinase (PTK), protein kinase C (PKC), and DNA topoisomerases I and II [92,93].
Apigenin, biochanin A, chrysin, genistein, luteolin and kaempferol can effectively inhibit the proliferation of cell lines derived from follicular, papillary and anaplastic thyroid carcinomas [66]. Besides inhibiting cancer initiation and progression, resveratrol, quercetin and genistein can also induce redifferentiation of thyroid cancer cells [76]. Many flavonoids inhibit thyroid cancer growth, but they also inhibit iodine uptake. Iodine uptake is not only an important step in thyroid hormone synthesis, but also a crucial exploitable mechanism for the treatment of differentiated (papillary and follicular) thyroid carcinoma with radioactive iodine. In this regard, among flavonoids thus far tested, myricetin would be an interesting modality for the treatment of thyroid carcinoma due to its capacity to increase iodide uptake and retention [67].

Modulation of Nrf2 Activity by Phenolic Compounds
Using in vitro and in vivo approaches, several studies have demonstrated the potential role of Nrf2 activation by phenolic compounds in the protection against oxidative stress, inflammation, cancer, and various other specific cellular toxicities, such as hepatic, cardiac and neuronal toxicities ( Table 2). Of note, these studies did not focus specifically on thyroid tissue or thyroid cell lines, and therefore further research is warranted to evaluate if the described effects on the Nrf2 pathway are also applicable to the thyroid gland.
Many phenolic compounds, such as epigallocatechin-3-gallate, curcumin and catechins exert an antioxidative and antiinflammatory effect through antioxidant enzymes that are regulated by Nrf2 [159][160][161]. Curcumin, found in turmeric, stimulates Nrf2 transcriptional activity in a time-and dose-dependent manner, resulting in the induction of heme oxygenase 1 (HO-1) protein expression [107]. Luteolin and apigenin protect against oxidative stress through up-regulation of HO-1 and of glutamate cysteine ligase catalytic subunit via Nrf2 pathway induction [96]. The isoflavone biochanin A from cabbage, red clover and alfalfa showed antioxidative [162] and hepatoprotective effects by inducing Nrf2 nuclear translocation and up-regulation of HO-1 expression, resulting in the inhibition of oxidative liver injury induced by lipopolysaccharide in combination with D-galactosamine N (LPS/GalN) [101]. In a study with aged rats, treatment with hesperidin elevated the myocardial antioxidative status by lowering lipid peroxidation and protein oxidation, increasing glutathione (GSH) levels and inducing the Nrf2 pathway, which led to higher expression levels of antioxidant enzymes [122]. Morin exerts cytoprotective and genoprotective effects against oxidative stress. Morin treatment prior to H 2 O 2 exposure significantly prevented the generation of ROS, increased cell viability and prevented DNA damage by up-regulating Nrf2-dependent HO-1 expression through extracellular signal-regulated kinases (ERK) in C2C12 myoblasts [133]. In the case of myricetin, production of pro-inflammatory mediators through the suppression of signal transducer and activator of transcription 1 (STAT1) and Nuclear factor kappa-lightchain-enhancer of B cells (NF-kB) activation was inhibited by induction of Nrf2-mediated HO-1 expression in LPS-stimulated macrophages [134].
Interestingly, besides activating the Nrf2 pathway, some phenolic compounds can also inhibit the same pathway, with the potential therapeutic application against chemotherapy resistance. Apigenin significantly lowered Nrf2 expression at the protein and mRNA levels, at least partially through down-regulation of the PI3K/Akt pathway in a human papillomavirus-related endocervical adenocarcinoma cell line. Inhibition of Nrf2 by apigenin increased the intracellular concentration of doxorubicin and led to resensitization of these cells that were formerly resistant to the drug [163]. Reduction of protein and mRNA levels of Nrf2 after luteolin treatment sensitized A549 lung cancer cells to doxorubicin, potentially due to depletion of reduced glutathione resulting from down-regulation of ARE-driven genes [131]. Luteolin has also been shown to repress Nrf2 signaling in mouse livers and in A549 xenografts [132]. Hence, there is substantial mechanistic evidence that apigenin and luteolin can downregulate Nrf2 signaling by affecting Nrf2 mRNA stability [131] or modulating pathways such as PI3K [98]. This apparent discrepancy in the effects of apigenin and luteolin on the activity of the Nrf2 pathway might be due to the different models used in the various studies, or/and to some unappreciated toxic effect of these compounds on some cell lines, which could lead to Isoorientin HepG2 cells [124] induces Nrf2 PtW-driven AO response [124] Kaempferol HepG2-C8 [125]; HEI-OC1 [126]; RBL-2H3 [127] ↑mRNA and protein expression of Nrf2-regulated genes, HP [125]; via JNK PtW and Nrf2-translocation A , CyP [126]; Nrf2 mediated HO-1 induction, AAl [127] (  secondary induction of the Nrf2 pathway. In that regard, it should be noted that the studies showing inhibition of the Nrf2 pathway are better supported by mechanistic analyses [98,131], whereas more research is warranted to determine how it may be mechanistically possible for these compounds to also activate the Nrf2 pathway in certain settings.

EFFECTS OF SULFUR-CONTAINING COMPOUNDS ON THE NRF2 PATHWAY AND ON THE THYROID
Isothiocyanates including allyl isothiocyanate (AITC), benzyl isothiocyanate (BITC), phenethyl isothiocyanate (PEITC) and sulforaphane, from cruciferous vegetables such as brussels sprouts, broccoli, cauliflower, cabbage and watercress, are strong chemo-protective and anticancer agents [164]. Isothiocyanates are present in these vegetables as inactive precursors (glucosinolates), from which they are released when plant cells are damaged. Glucosinolates are converted into isothiocyanates by the enzyme myrosinase, which is also present in the human gut microflora. Exposure to antibiotics that alter the microflora can largely reduce the conversion of glucosinolates into isothiocyanates [165]. The bioavailability of isothiocyanates is also affected by diurnal variations, with higher conversion of glucosinolates to dithiocarbamates during the day, and higher conversion of isothiocyanates to dithiocarbamates during night [166]. Isothiocyanates induce cytoprotective proteins via the Nrf2/Keap1/ARE pathway, thus conferring protection against oxidants and electrophiles and thereby potentially exerting chemopreventive and/or therapeutic actions against a wide range of diseases [167].
Sulforaphane and AITC treatment in hepatocellular carcinoma HepG2 cells potently induced HO-1 protein expression via Nrf2 pathway induction ( Table 3). PEITC can also induce the protein expression levels of Nrf2 and HO-1 by stimulating ARE-mediated transcription in Hela cells [168,169].
Sulfur-containing compounds can modulate thyroid function and might be efficacious in inhibiting thyroid cancer progression. Through an ROS-dependent pathway, sulforaphane inhibits thyroid cancer cell proliferation, invasion and migration, and it induces cell cycle arrest without significant toxicities such as liver injury, which makes it a potentially effective and safe modality for cancer treatment [182]. In many animal models, DAS inhibits chemicallyinduced carcinogenesis. For example, treatment of the anaplastic thyroid cancer cell line ARO with DAS inhibited proliferation by increasing in a dose-dependent manner, the accumulation of sub-G1 DNA and the associated accumulation of cells in the G2/M phase via mitochondrial signaling pathways [175].
BITC, which has been extensively used as a chemopreventive agent, has been shown to decrease total T4 and T3 levels in rats after two weeks of daily treatment at doses known to have anticarcinogenic effects [172].
In another study, rats were fed with a combination of goitrin, a sulfur-containing oxazolidine, thiocyanate and AITC, in concentrations comparable to those present in the daily amount of cabbage that these rats can consume, and this treatment resulted in increased thyroid size. This effect could not be reproduced by administering individually any of these compounds, meaning that only their com- AC [182] Thiocyanates Goitrin -R -↑T size, AT [171] Effects listed in the fourth and fifth column and separated by semicolons have been derived from the corresponding models shown in the second and third columns, respectively, and separated by respective semicolons. R, rat; T, thyroid; AT, antithyroid; DME, drug-metabolizing enzyme; JNK1, c-Jun N-terminal kinases; PtW, pathway; ChP, chemoprotective; HO-1, heme oxygenase 1; ↑, increase; NP, neuroprotective; A , activate; AI, anti-inflammatory; , up-regulate; CarP, cardioprotective; NephP, nephroprotective; AC, anticancer; MAPK, mitogen-activated protein kinase; HP, hepatoprotective; AO, antioxidative; NQO1, NAD(P)H quinone dehydrogenase 1; CyP, cytoprotective.
bination, as found in cabbage, can lead to increased thyroid size [186]. Of note, in a recent randomized clinical trial, we showed that daily administration of a sulforaphane-rich broccoli sprout extract to healthy volunteers did not have any detrimental effect on thyroid function or the thyroid autoimmunity status [187]. The most abundant sulfur-containing compound from garlic, Sallylcysteine (SAC), has a potential beneficial effect in diabetes. In streptozotocin-induced diabetic rats, TSH and thyroid hormone levels decreased, along with circulating catalase (CAT), GSH, glutathione peroxidase (GPx), superoxide dismutase (SOD) and insulin levels. Treatment with SAC reversed these effects and could be used to restore antioxidant capacity, reduce lipid peroxidation and preserve thyroid function in diabetic subjects [177].

QUINONES AND TERPENOIDS
Quinones and terpenoids, including mono-, di-, tri-, and sesquiterpenoids, are organic compounds with high diversity in structure and bioactivity ( Table 4). Even though many reports demonstrate that these compounds can activate the Nrf2 pathway, their potential effects on thyroid physiology and disease are rarely examined. In view of the highly variable chemical structures and biological activities of quinones and terpenoids, further research of their actions on the thyroid could give interesting insights into potential beneficial and/or detrimental effects.

PLANT EXTRACTS AND THYROID
Extracts from plants and medical herbs, even though they may contain several compounds in variable and unstandardized concentrations, comprise a rich source of candidate drugs that may have potential effects on thyroid gland function ( Table 5).
Medical herbs such as Lycopus americanus, Lycopus europaeus, Lycopus virginicus arious, Lithospermum officinale and Lithospermum ruderale show promising potential for the treatment of hyperthyroidism, while Plectranthus barbatus, Withania somnifera, Centella asiatica and Commiphora mukul might be useful to treat hypothyroidism. Interestingly, withaferin A, which is contained in Withania somnifera, has recently been shown to be a potent inducer of the Nrf2 pathway [215]. Treatment for hyperthyroidism and hypothyroidism with herb extracts can have variable effects and may be sufficient for treating symptoms in some cases, but at the present state of knowledge, it should not be used in the place of established conventional treatment, especially for serious thyroid disorders [216].
Boiled extracts of cassava, bamboo shoot, cauliflower and cabbage can inhibit TPO activity [217]. Kalanchoe brasiliensis aqueous extract can also inhibit TPO, and experiments with this extract showed that chronic consumption combined with low iodine intake can cause hypothyroidism and goiter [218]. In mice, leaf extracts of Aegle marmelos, Bacopa monnieri and Aloe vera can alter thyroid hormone concentrations. Aloe vera and Aegle marmelos decreased T3 concentration and could thus provide a beneficial effect in the treatment of hyperthyroidism, whereas Bacopa monnieri increased T4 concentration and could be useful for the treatment of hypothyroidism [219].
Rats treated with green and black tea extract rich in catechins showed decreased T3 and T4 serum levels as well as decreased 5DI and TPO activity, while Na+/K+-ATPase activity was increased along with TSH levels. Treatment also led to hypertrophy and/or hyperplasia of thyroid follicles, with reduced colloid content [220].

OTHER COMPOUNDS MODULATING NRF2 PATHWAY AND THYROID FUNCTION
Apart from the aforementioned categories of phytochemicals, other naturally occurring compounds can also activate the Nrf2 pathway and affect thyroid function; these are listed in Table 6.

CONCLUSION
The use of phytochemicals from edible plants may provide a sustainable and effective solution for the treatment and prevention of diseases, with potential clinical utility both in underdeveloped countries that still do not have access to expensive medications, as well as in developed countries challenged by increasing healthcare costs and/or endorsing complementary medicine. Food-derived phytochemicals with cytoprotective and antioxidant properties have already been tested for potential use in a wide range of diseases, including cardiac and psychiatric disorders, diabetes, viral, microbial and parasitic infections, inflammation and cancer. However, little is known about the impact of phytochemicals on thyroid function and the potential underlying modulation of the Nrf2 pathway, an important mediator of antioxidant, cytoprotective and other effects in cells. We and others have recently shown the importance of the Nrf2 pathway in the thyroid gland; Nrf2 not only coordinates the antioxidant response under basal and iodide excess conditions, but it also regulates the transcription of the Tg gene itself, and it further affects the iodination of Tg under various conditions [43]. Thus, the Nrf2 pathway can potentially integrate the protection of thyroid cells from excessive ROS production with the regulation of thyroid function through Tg gene transcription and Tg iodination [30,48].
The present literature review identified several categories of phytochemicals that may have an impact on thyroid physiology and disease, and discussed their potential effects on the Nrf2 pathway. Reports that address the effects of specific phytochemicals on both the thyroid and the activity of Nrf2 are rather few; therefore, we also included studies using other tissues/cell lines to provide insights into whether a certain phytochemical's effects on the thyroid may be derived from its potential to modulate the Nrf2 pathway. Figure 1 briefly summarizes the potential effects of these compounds on thyroid gland function and physiology.
In summary, there are phytochemicals with known antioxidant or other cytoprotective properties that have shown promising results in the field of cancer prevention, but may have a detrimental effect on thyroid function at certain doses or in certain combinations. Thus, caution should be exercised when drawing conclusions about the potential beneficial or detrimental effects of a phytochemical on thyroid function, and this issue should be carefully characterized, including via classical toxicology approaches. While the health benefits associated with natural compounds found in plants in low amounts are generally well accepted, attention should be drawn to potential toxicities arising from the use of concentrated extracts that may contain excess amounts of phytochemicals. For instance, a daily dose of 600 µmol glucoraphanin and 40 µmol sulforaphane in the form of broccoli sprout extracts over 12 weeks has been shown to be effective against the detoxification of airborne pollutants [234], and this dosing scheme was proven to be safe for the thyroid [187]. Table 4. Modulation of Nrf2 pathway and thyroid function by quinones and terpenoids.
To conclude, this review catalogues the phytochemicals that may have an effect on the thyroid gland and on the Nrf2 pathway, and it conveys the message that each dietary supplement intended for human use that contains such phytochemicals, should be carefully evaluated in this aspect, first by analyzing its contents, and then by assessing its potential effects on thyroid function. Such diligence will contribute to realizing the promise of the old yet highly active field of "chemoprevention" against diseases via natural compounds, aiming to propose novel treatments by identifying new, cost-efficient and safe therapeutic agents to promote public health.

CONSENT FOR PUBLICATION
Not applicable.

FUNDING
This work was supported by the Swiss National Science Foundation Research Grants 31003A_153062 and 31003A_182105; Leenaards Foundation 2016 Fellowship for Academic Promotion in Clinical Medicine; Swiss National Science Foundation -State Secretariat for Education, Research and Innovation (SNF-COST) Project 174626 -C15.0045; and SNF COST Project IZ-COZ0_177070 (all to GPS).

CONFLICT OF INTEREST
The authors declare no conflict of interest, financial or otherwise. Fig. (1). Simplified model summarizing the potential effects that antioxidant natural compounds can have on thyroid physiology through Nrf2-dependent and Nrf2-independent pathways. Antioxidant natural compounds (phenolic compounds, sulfur-containing compounds, quinones, terpenoids and diverse plant extracts) have the potential to upregulate Nrf2 signaling through various mechanisms. The induction of Nrf2 signaling exerts antioxidant cytoprotective functions, upregulates the expression of thyroglobulin (Tg) and has potential cancer-preventing effects (indicated by solid arrows). These natural compounds have also been described to impact iodide uptake, thyroid hormone synthesis and thyroid size (goiter); the mechanisms of these latter effects are less clearly specified (indicated by dashed arrows), and the potential involvement of Nrf2 has not been well documented. Research is thus warranted especially in animal models and humans to further elucidate these mechanisms.