Cross-species physiological interactions of endocrine disrupting chemicals with the circadian clock.

Extensive metabolic dysregulation has been documented as a result of EDC exposure in humans and model organisms, with epidemiological and mechanistic links to obesity, diabetes and metabolic syndrome ( Heindel et al., 2017 ; Veiga-Lopez et al., 2018 ). Nuclear receptors are targets of EDC binding, and perturbed NR signaling plays a central role in mediating diabetogenic and obesogenic action of EDCs ( Casals-Casas and Desvergne, 2011 ). EDC exposure and circadian disruption induce similar metabolic pathologies ( Bass and Lazar, 2016 ; Heindel et al., 2017 ), but the extent to which circadian clocks mediate EDC action is not well characterized. In the following sections, we will explore crosstalk among circadian and endocrine systems in controlling metabolism and evidence of parallel metabolic and circadian disruption following EDC exposure across organisms. The circadian system is a critical regulator of metabolic and energetic homeostasis. Temporal separation of incompatible metabolic processes such as catabolism vs. anabolism and redox processes helps prevent futile cycling and contributes to energetic homeostasis ( Bass and Takahashi, 2010 ). Central and peripheral clocks modulate lipid metabolism, glucose homeostasis, body weight and energy metabolism ( Froy and Garaulet, 2018 ; Gooley, 2016 ; Lamia et al., 2008 ; O’Neill and Feeney, 2013 ). In humans, circadian disruption such as seen in shift work impacts energy homeostasis, leading to metabolic pathologies such as obesity, diabetes and metabolic syndrome ( Table 1 ). In rodent models, core clock gene ablation causes extensive perturbation of metabolic health and energy homeostasis, leading to a range of diabetogenic and obesogenic phenotypes following systemic as well as liver, pancreas or adipose-specific gene ablation ( Table 2 ). Environmental circadian disruption in rodent models via light at night or simulated shift work impairs pancreatic clock function, leading to islet failure and type 2 diabetes mellitus ( Gale et al., 2011 ; Qian et al., 2013 ). Pancreatic islets from type 2 diabetic patients grown in vitro were found to have dampened circadian oscillation and deficits in insulin and glucagon exocytosis compared with controls ( Petrenko et al., 2020 ). Obesogenic diet as well as altered feeding rhythms have been extensively characterized to perturb clock gene expression and dysregulate metabolism in metabolically active tissues ( Cunningham et al., 2016 ; Engin, 2017 ; Fonken et al., 2013 ; Guan et al., 2018 ; Qian et al., 2015 ; Salgado-Delgado et al., 2013 ). Nuclear receptors exhibit rhythmic expression patterns in metabolic tissues, and can directly modulate clock gene expression ( Teboul et al., 2008 ; Yang et al., 2006 ) ( Table 3 and Figure 2 ). Nuclear receptors Reverb-α/β, and RORα/β/γ form secondary regulatory feedback loops to the molecular clock and function in regulation of circadian rhythm and metabolism ( Guillaumond et al., 2005 ). Reverb-α regulates glucagon secretion in pancreatic alpha cells and mediates rhythmic cholesterol, lipid and bile acid metabolism in liver ( Martelot et al., 2009 ; Vieira et al., 2013 ). Adrenal glucocorticoids regulate clock gene expression in brain and a range of peripheral tissues, and modulate rhythmic metabolic processes in liver ( Table 3 ) ( Quagliarini et al., 2019 ; So et al., 2009 ). Nutrient and metabolite sensing nuclear receptors such as PPARs, constitutive androstane receptor (CAR), liver X receptor (LXR) and farnesoid X receptor (FXR) mediate nutrient, lipid and energy metabolic pathways (Preidis et al., n.d.; Xu et al., 2018 ). PPARα/γ directly modulate core clock gene expression and are critical molecular links between peripheral clocks and energetic homeostasis ( Table 3 and Figure 2 ) ( Chen and Yang, 2014 ; Yang et al., 2012 ). Some metabolite sensing receptors such as LXR are expressed in multiple endocrine organs and can modulate both metabolic and endocrine processes such as glucose homeostasis and insulin sensitivity, and gonadal steroidogenesis, among others ( Maqdasy et al., 2016 ). For a detailed review of endocrine / neuroendocrine control of metabolism in the context of EDC action, see ( Heindel et al., 2017 ). Two studies explored transcriptional and physiological responses following exposure to EDCs or an obesogenic diet. Kopp et al. (2017) assessed lipid accumulation and circadian rhythm in zebrafish following developmental exposure to a range of obesogenic EDCs or to high calorie diet (HCD). The study was conducted in a transgenic zebrafish line containing a luciferase reporter driven by 4 E-boxes, representing binding sites for the core clock transcriptional activator and heterodimer Clock / Bmal1. Two compounds, benzophenone-3 and tetrabrominated bisphenol A, abolished rhythmicity of luciferase expression in the transgenic zebrafish model, as did HCD control. Exposure to tributylin and tris (1,3-dichloroisopropyl) phosphate resulted in altered amplitude and periodicity of the luciferase reporter. Dietary EDC exposure resulted in a greater extent of lipid accumulation than HCD in zebrafish larvae ( Kopp et al., 2017 ). Labaronne et al. (2017) examined transcriptional responses in mouse liver following lifelong exposure to low dose pollutant mixture vs. a high fat, high sucrose (HFHS) diet. Circadian rhythm signaling was significantly perturbed by low dose pollutant exposure, with downregulation of positive elements of the clock (c lock, nr1d1 and bmal1 ) and upregulation of negative elements of the clock ( per1, per2, per3, cry2, nr1d2 and rorc ). A similar pattern of clock gene perturbation was observed in the HFHS diet fed mice, but with fewer genes significantly impacted. Other pathways dysregulated by low dose pollutant exposure or HFHS diet include drug and xenobiotic metabolism, steroid biosynthesis, and fatty acid metabolism, with overlapping but not identical gene sets contributing to these pathway impacts ( Labaronne et al., 2017 ). Fader et al. (2019) conducted a circadian time-course sample collection to examine impact of oral 2,3,7,8-Tetrachlorodibenzo- p -dioxin (TCDD) exposure on hepatic transcript and metabolite rhythmicity in male C57BL/6J mice. Transcript and protein-level cycling of several core clock genes including bmal1, clock, nr1d1, per1, cry1 , and nfil3 was impacted by TCDD exposure in a dose-dependent manner, exhibiting reduced amplitude or loss of cycling. TCDD exposure also resulted in an almost complete collapse of clock-controlled hepatic transcript and metabolite cycling. Pathways impacted by TCDD exposure in liver included lipid metabolism, glucose / glycogen metabolism and heme metabolism. The results indicated diminished metabolic efficiency and energy storage following TCDD exposure in liver ( Fader et al., 2019 ). As discussed in the previous section, exposure to the fungicide tolyfluanid in C57BL/6J mice altered diurnal rhythm of energy expenditure and food intake and led to increased body weight, impaired metabolic flexibility, glucose intolerance and insulin resistance ( Regnier et al., 2015 ). Circadian rhythm signaling was the most significantly dysregulated pathway following exposure to a hydroxylated PCB congener in liver of adult male Wistar rats ( Ochiai et al., 2018 ), and core clock gene expression was perturbed in liver of B6C3F1 male mice following oral exposure to Di-2-ethylhexyl phthalate (DEHP), with a corresponding increase in liver weight following 48 and 72 hour exposures. ( Currie et al., 2005 ). In vitro , bisphenol A exposure altered core clock gene expression and dysregulated neuropeptide expression in feeding-related neurons of the hypothalamus; altered neuropeptide Y expression following BPA exposure was dependent on the presence of bmal1 ( Loganathan et al., 2019 )

EDCs have been extensively characterized for their impact on reproductive function ( Gore et al., 2015 ; Hachfi et al., 2012 ). Reproductive function is driven by hourly, daily, weekly, monthly and seasonal rhythms of hormone production and action in the hypothalamus – pituitary – gonad (HPG) axis of mammals, and the hypothalamus – pituitary – gonad – liver (HPGL) axis of oviparous organisms, with the liver contributing proteins necessary for oocyte growth and development ( de la Iglesia and Schwartz, 2006 ; Rosa et al., 2016 ). In the reproductive axis, the pituitary clock and clocks inherent to gonadotropic cells drive rhythmic hormone production and release in the reproductive axis ( Hickok and Tischkau, 2010 ; Lin et al., 2015 ). Gonadotropins themselves function as zeitgebers in gonadal clocks. Luteinizing hormone (LH) induces clock gene expression during follicle development in rat ovary and on the day of pro-estrous; in males, LH synchronizes the leydig cell clock and contributes to rhythmicity of testosterone production ( Baburski et al., 2019 ; Gräs et al., 2012 ; Karman and Tischkau, 2006 ). Gonadal steroids in turn exert feedback on brain, liver and reproductive organ clocks ( Karatsoreos et al., 2011 ; Nakamura et al., 2010 , 2005 ; Y. Zhang et al., 2019 ) and Table 3 . In zebrafish, rhythmic expression patterns of genes involved in steroidogenesis, gonadal function and sex ratio control were observed in brain and gonad ( Rosa et al., 2016 ). Rhythmic reproductive physiology can be driven by endogenous circadian clocks as well as external zeitgebers such as feeding rhythm and thermocycles ( Hontela and Peter, 1983 ; Weber and Spieler, 1987 ). Mouse genetics studies reveal a range of impairments in reproductive physiology and fertility following core clock gene ablation ( Table 2 ). Physiological perturbations include altered estrous cycle, impaired timing of LH and FSH pro-estrous surge and disrupted ovulation in females, and impaired gonadal steroidogenesis in both males and females ( Table 2 ). Both male and female infertility are observed following core clock gene ablation as well as impaired oocyte fertilization rate, embryo implantation failure, and increased rate of full term pregnancy failure ( Table 2 ). Ablation of Clock was additionally found to disrupt nursing and maternal behavior in dams, as well as growth and survival rate of pups ( Table 2 ). Environmental perturbation of circadian function via shift work in humans is associated with altered ovarian cycle pattern, endometriosis risk and prolonged wait time to pregnancy ( Table 1 ). Impaired LH surge, reduced fertility and an increase in adverse pregnancy outcomes has been reported in rodent models of simulated shift work ( Bahougne et al., 2020 ; Summa et al., 2012 ). Short photoperiod disrupted FSH and prolactin secretion in male hamsters, and led to impaired testicular steroidogenesis, and sleep fragmentation disrupted nocturnal testosterone rhythm in normal men ( Chandrashekar and Bartke, 1989 ; Luboshitzky et al., 2001 ). Reproduction is a seasonal event in many vertebrate species, and seasonal reproductive cycles are modulated in part by the circadian clock sensing and responding to changing photoperiod ( Nakane and Yoshimura, 2019 ; D. Zhang et al., 2009 ). In addition to photoperiod, annual fluctuation in water temperature and lunar periodicity contribute to timing of seasonal reproductive cycles in fish ( Ikegami et al., 2014 ; Oliveira et al., 2009 ). Endogenous reproductive behaviors are a critical component of reproductive success. EDCs have been characterized to disrupt social dominance hierarchies, sexual selection and courtship behaviors in fish ( Coe et al., 2008 ; Colman et al., 2009 ; Söffker and Tyler, 2012 ). Some evidence exists for circadian control of reproductive behaviors. In midshipman fish, circadian rhythm and melatonin control timing of nocturnal courtship vocalization ( Feng and Bass, 2016 ; Rubow and Bass, 2009 ) and egg laying and courtship behaviors were found to entrain to the light-dark cycle ( Weber and Spieler, 1987 ). Short-term (pulsatile) through longer-term (seasonal) rhythmicity is inherent to reproductive function across species. Central and peripheral timing signals are critical for rhythms of gonadotropin synthesis and release, gonadal steroidogenesis and estrous cycle patterns. Environmental perturbation of circadian clock function via shift work or altered photoperiod disrupts rhythmic reproductive physiology and leads to fertility impairments. In parallel, EDC exposure is extensively linked to adverse reproductive outcomes ( Gore et al., 2015 ; Hachfi et al., 2012 ). Given the interconnected nature of circadian and hormonal timing signals in the reproductive axis, it is interesting to explore whether reproductive impairments induced by EDC exposure are due in part to altered clock function or perturbed physiological rhythms. In this section, we will explore evidence for parallel circadian and endocrine impacts following EDC exposure in the reproductive axis. Altered core clock and circadian rhythm network gene expression has been demonstrated in zebrafish models following exposure to a range of environmental steroid hormones. Natural and synthetic progestins disrupted clock gene expression in brain, and to a lesser extent ovary, of adult zebrafish, with corresponding impacts on gonadotropin expression and gonadotropin regulation pathways ( Zhao et al., 2015a , 2015b ; Zucchi et al., 2014 , 2013 ). High dose progestin exposure decreased fecundity in adult zebrafish, and a network correlation analysis indicated a significant relationship between the circadian rhythm gene network and regulation of genes in the HPGL axis ( Zhao et al., 2015a ). Effects were observed across generations, with perturbations of clock gene expression in F1 generation offspring as well as directly exposed eleuthero embryos ( Liang et al., 2019 ; Zhao et al., 2015a , 2015b ). Progestins additionally altered circadian rhythm network gene transcription and increased photo-transduction signals in eyes of adult male and female zebrafish ( Zhao and Fent, 2016 ). Across generations, exposure to natural and synthetic steroid hormone mixtures (progesterone and drospirenone; medroxyprogesterone acetate and dydrogesterone; ethinylestradiol and norgestrel; and to a lesser extent progesterone and estradiol) impacted clock gene expression more strongly than exposure to the individual compounds ( Liang et al., 2019 , 2017 ; Zhao et al., 2015b ; Zucchi et al., 2014 ). Exposure to the corticosteroid fludrocortisone acetate altered circadian rhythm, glucose homeostasis and immune response gene networks across generations in zebrafish ( Zhao et al., 2016 ). The circadian clock gene network was among the most significantly altered gene networks in pituitary of juvenile female coho salmon following exposure to 17α-ethynylestradiol, with corresponding pathway impacts observed on gonadotropin regulation, calcium signaling and lipid metabolism ( Harding et al., 2013 ). In a circadian time course analysis, Zhao et al. (2018) examined the effect of a range of environmental steroid hormones on clock gene rhythmicity in zebrafish eleuthero embryos under constant darkness conditions. Progestins and corticosteroids induced similar transcriptional alterations of clock gene expression, which differed from that of estrogens. Exposure did not abolish transcript cycling, but revealed clock time (CT) 18 to be a sensitive time point for altered clock gene expression. Among the 18 circadian rhythm network genes analyzed by quantitative real time PCR, per1a and nr1d2a were the most strongly dysregulated genes following a range of steroid hormone exposures. Progestin and corticosteroid-induced impacts on per1a and nr1d2a expression were recovered using the progesterone receptor (PR) / glucocorticoid receptor (GR) antagonist mifepristone, indicating a role for activated PR / GR in regulating clock gene expression ( Zhao et al., 2018 ). Estrogen-mediated impacts on clock gene cycling were not recovered with tamoxifen, but as this is a selective estrogen receptor modulator, rather than a pure antagonist, the role of estrogen receptor (ER) signaling in altered clock gene expression in this model remains unclear ( Xia et al., 2016 ). Perturbed circadian transcriptional rhythm has been observed in the HPGL axis in teleost fish following exposure to other EDCs. Bisphenol A perturbed transcriptional rhythm of core clock genes in liver of juvenile mangrove killfish. 17β-estradiol, 4-tert-octylphenol and bisphenol A downregulated Km-clock expression in pituitary / brain, muscle and skin of adult hermaphrodite and secondary male fish, with no effect observed in gonad or liver ( Rhee et al., 2014 ). Altered clock gene expression, analyzed at a single time point, and reproductive impairment was observed in fathead minnow following exposure to pulp and paper mill effluents from five different sources ( Popesku et al., 2010 ). Additional impacts on circadian gene transcription in the HPGL axis of teleost fish have been reported following exposure to bisphenol A and cyanotoxins ( Choi et al., 2018 ; Faltermann et al., 2014 ; Qiao et al., 2016 ), and lead was found to alter circadian variation of brain neurotransmitters in fathead minnow ( Spieler et al., 1995 ). In rodent models, EDCs similarly perturbed clock gene expression and rhythmic HPG axis physiology. Walker et al. (2014) examined gene expression in two regions of the hypothalamus known to regulate reproductive function: the anteroventral periventricular nucleus (AVPV) and the arcuate nucleus (ARC) following gestational exposure to estrogenic EDCs in Sprague-Dawley rats. Exposure to the industrial PCB mix Aroclor 1221 resulted in delayed timing of eye opening in females and delayed timing of puberty in males, and estradiol benzoate exposure disrupted the estrous cycle in females. Sexually dimorphic impacts on AVPV and ARC gene expression were observed following gestational exposure to estrogenic EDCs. Females exhibited a masculinized expression profile in the AVPV region following A1221 and EB exposure, and exposure resulted in upregulation of two clock genes – bmal1 and per2 . Gene expression changes in the ARC were observed to a greater extent in males, with androgen receptor and leptin receptor identified as two candidate genes possibly related to the delayed onset of puberty in males following A1221 exposure ( Walker et al., 2014 ). In golden hamster, exposure to the pharmaceutical triazolam altered circadian rhythm of pituitary LH release ( Turek, 1988 ). The heavy metal cadmium altered rhythmic transcription and physiology in rodent pituitary. In male Wistar rats, low dose cadmium exposure disrupted rhythmic expression pattern of core clock genes in anterior pituitary, altered transcriptional rhythm of redox enzyme expression, and altered daily pattern of secreted prolactin, luteinizing hormone, thyrotropin and corticosterone ( Jiménez-Ortega et al., 2012 ). In two additional reports, cadmium exposure resulted in a disrupted 24-hour pituitary secretion pattern of adrenocorticotropic hormone, growth hormone and thyrotropin, and disrupted the daily pattern of pituitary glutamine, glutamate and aspartate content ( Caride et al., 2010b , 2010a ). In summary, a variety of EDCs perturb clock gene expression or rhythmic physiology in the HPG(L) axis across organisms and across life stages, with circadian rhythm network frequently reported to be among the most significantly perturbed pathways following exposure. Environmental steroid hormone exposure is particularly important to investigate in aquatic organisms, and provides a useful model for investigating cross talk of activated nuclear receptors with clock proteins ( Rubel et al., 2012 ; Zhao et al., 2018 ). Evidence presented demonstrates circadian redox oscillations to be a target of exposure to the heavy metal cadmium ( Jiménez-Ortega et al., 2012 ), and analysis of specific brain regions of the hypothalamus shed light on sexually dimorphic gene expression following EDC exposure ( Walker et al., 2014 ).

EDC exposure has been demonstrated to impact locomotor activity as well as social or reproductive behaviors in a range of species ( Colman et al., 2009 ; Patisaul and Adewale, 2009 ; Saili et al., 2012 ). Critical components of behavioral regulation by the circadian system include sleep / wake patterns, locomotor activity patterns and timing of food intake. The molecular clock drives behavioral rhythmicity and influences overall activity level. In mice lacking bmal1 , locomotor activity pattern was abolished under constant darkness and overall activity level was reduced ( Bunger et al., 2000 ). Locomotor activity pattern and rhythmic feeding behavior have been described to persist under constant conditions in fish, indicating that these processes can be endogenously circadian, rather than occurring only in response to light – dark cycle or feeding time ( Iigo and Tabata, 1996 ; Mata-Sotres et al., 2015 ). Sex steroids play an integral role in modulating behavioral patterns. Circadian activity patterns are strongly sexually dimorphic, with phase, amplitude and period modulated by both estrogens and androgens ( Krizo and Mintz, 2015 ). Estrogen signaling during development programs sexually dimorphic behaviors in adult mice ( Royston et al., 2016 ), reviewed in ( Hatcher et al., 2018 ). Estrogen signaling modulates behavioral rhythmicity in adulthood as well and estrogen action via ER directly modulates clock gene expression in the SCN and in peripheral tissues ( Hatcher et al., 2018 ; Royston et al., 2014 ) and Table 3 . In male and female rodents, gonadal steroids modulate locomotor activity rhythm, time of activity onset and overall activity levels ( Daan et al., 1975 ; Juárez-Tapia and Miranda-Anaya, 2017 ; Karatsoreos et al., 2007 ; Morin et al., 1977 ). The studies presented in this section demonstrate perturbed locomotor and behavioral rhythmicity following EDC exposure. Progesterone exposure resulted in dose-dependent dampening of circadian rhythm of locomotor activity in zebrafish eleuthero embryos, as measured under constant darkness conditions ( Zhao et al., 2018 ). In adult zebrafish, bisphenol A (BPA) and estradiol (E2) altered circadian rhythm of light / dark preference, and E2 dampened locomotor activity pattern ( Wang et al., 2015 ). Developmental BPA exposure reduced total activity level and altered circadian activity pattern in adult male zebrafish, with no comparable impact observed in females ( Weber et al., 2015 ). Dietary polychlorinated biphenyls (PCBs) altered swimming activity pattern in adult zebrafish and their offspring ( Péan et al., 2013 ) and in goldfish, lithium chloride perturbed circadian locomotor activity rhythm ( Kavaliers, 1981 ). Corresponding behavioral phenotypes were reported following EDC exposure, for example, reduced aggressive behavior and decreased group preference ( Wang et al., 2015 ), increased anxiety ( Porseryd et al., 2017 ), sexually dimorphic alteration to social behavior ( Weber et al., 2015 ) and altered shoaling behaviors ( Kavaliers, 1981 ; Porseryd et al., 2017 ). Whether there is a circadian component to these social behaviors is less clear due to limitations of experimental design in the studies described and certainly warrants future exploration. Some evidence exists for circadian control of social behaviors; an endogenous social rhythm was found to persist under constant conditions in C57BL/6J mice and entrain to the light-dark cycle ( Panksepp et al., 2008 ). In rodent models, exposure to a hydroxylated PCB metabolite in male Wistar rats increased locomotor activity during both light and dark phases, but it is unclear whether this was a circadian impact or general hyperactivity ( Lesmana et al., 2014 ). The pharmaceutical triazolam induced a phase advance in circadian locomotor activity rhythm in golden hamsters, which corresponded with dysregulating timing of LH surge ( Turek, 1988 ). In male C57BL/6J mice, dietary exposure to the fungicide tolyfluanid impacted diurnal circadian activity pattern ( Regnier et al., 2015 ). Altered diurnal rhythm of energy expenditure and food intake was reported, with exposed mice exhibiting increased food consumption, activity and energy expenditure during the light phase. The study revealed a link between activity rhythm and feeding pattern and perturbed energy homeostasis. Exposed mice exhibited increased body weight and adiposity, glucose intolerance, insulin resistance and impaired metabolic flexibility. Evidence of parallel circadian and metabolic disruption following EDC exposure will be further explored in the following section.

Per-Arnt-Sim (PAS)-domain containing proteins function as environmental sensors. Members of this protein family include elements of the molecular clock as well as aryl hydrocarbon receptor (AhR) and hypoxia-sensing proteins, such as the hypoxia-inducible factors (HIFs) ( Hahn et al., 2017 ; McIntosh et al., 2010 ; Pelster and Egg, 2018 ). Crosstalk between hypoxia response and circadian rhythm transcription factors has been documented at the molecular level ( Pelster and Egg, 2018 ), and crosstalk between AhR and clock proteins has been described in mammalian models ( Mukai and Tischkau, 2007 ; Xu et al., 2010 ). AhR functions as a xenobiotic sensor, directing cellular response to xenobiotic exposure. Known agonists of AhR include dioxin, polycyclic aromatic hydrocarbons, dioxin-like polychlorinated biphenyl congeners, some pharmaceuticals and endogenous and dietary ligands such as plant flavonoids and photo-oxidation products of tryptophan ( Denison and Nagy, 2003 ). Circadian control of xenobiotic sensor and drug metabolizing gene expression patterns imparts a temporal component to detoxification processes. Expression patterns of the drug processing gene battery as well as xenobiotic and nutrient sensing nuclear receptors CAR, PXR, AhR and its binding partner ARNT and PPARα are rhythmic in mouse liver, with circadian expression patterns found to vary based on sex ( Huang et al., 2002 ; Lu et al., 2013 ; Y.-K. J. Zhang et al., 2009 ). Rhythmic expression patterns of drug metabolizing genes and AhR are also observed in zebrafish liver ( Carmona-Antoñanzas et al., 2017 ). Crosstalk between core clock proteins and AhR modulates the AhR-induced detoxification response. In mouse lungs, Clock modulated transcriptional response to AhR ligand benzo[a]pyrene and induction of detoxification enzymes, and per1 and per2 modulated transcriptional response to dioxin in mouse mammary gland and liver ( Qu et al., 2010 ; Tanimura et al., 2011 ). AhR additionally intersects with hormone and nutrient-sensing receptors such as ER and PPARs in mediating xenobiotic response ( Casals-Casas and Desvergne, 2011 ; La Merrill Michele et al., 2013 ; Swedenborg and Pongratz, 2010 ). The studies described in this section explored physiological and circadian impacts of AhR activation. Altered circadian transcriptional or behavioral rhythms have been observed in rodent models following AhR activation by TCDD. TCDD altered core clock gene expression or transcriptional rhythm in mouse liver, SCN and ovary of C57BL/6J mice ( Mukai et al., 2008 ; Tischkau et al., 2011 ; Xu et al., 2010 ). TCDD-exposed mice exhibited decreased phase shifts in response to light ( Mukai et al., 2008 ). In deer mice, TCDD disrupted rest-activity rhythm and altered transcriptional rhythm of core clock genes in the SCN ( Miller et al., 1999 ). In female C57BL/6J mice, TCDD exposure disrupted 24-hour rhythm of progenitor hematopoietic stem cell numbers, and perturbed transcriptional rhythm of per1 and per2 in progenitor stem cell populations ( Garrett and Gasiewicz, 2006 ). As discussed in the previous section, TCDD exposure perturbed core clock gene cycling in liver as well as hepatic transcriptome and metabolome rhythmicity in male C57BL/6J mice ( Fader et al., 2019 ). The PAH β-naphthoflavone (BNF) reduced light-induced phase shifts in C57BL/6J mice, and response was dependent on the presence of AhR ( Xu et al., 2013 ). In Rainbow trout, PAH exposure altered diurnal variation of melatonin, serotonin and other neurotransmitters in the pineal gland ( Gesto et al., 2009 ). While PAHs are characterized as AhR agonists, it is not clear whether findings reported in this study involved AhR activation. Interactions of xenobiotic and hypoxia responses were explored following exposure to the pharmaceutical diclofenac in wild-caught three-spined stickleback ( Prokkola et al., 2015 ). Exposure to both hypoxia and diclofenac, but not combination, resulted in altered expression level and dampened oscillation of per1 and clock. Both individual and combined exposures led to AhR activation, and diclofenac exposure in combination with hypoxia interfered with the expected hypoxia response ( Prokkola et al., 2015 ). The nature of the interaction between activated AhR and other PAS-domain containing proteins across organisms and in various xenobiotic exposure contexts warrants future exploration. Additional circadian impacts of xenobiotic exposure have been demonstrated following exposure to pharmaceuticals and anxiolytic drugs ( Akiyama et al., 1999 ; Kavaliers, 1981 ; Oggier et al., 2010 ; Xiao et al., 2017 ) and to copper ( Doria et al., 2018 ; Kim et al., 2017 ; Vicario-Parés et al., 2018 ). While pharmaceuticals, PAHs and some heavy metals are not primarily classified as EDCs, some pollutants within these classes have known endocrine activity ( Handy, 2003 ; Overturf et al., 2016 ; Sabir et al., 2019 ; Toppari and Juul, 2010 ; Y. Zhang et al., 2016 ); studies summarized in Table 4 indicate pollutants from these classes that may additionally interact with circadian systems.

The hypothalamic – pituitary – thyroid axis is important for energy balance and homeostatic systems such as thermoregulation. Core body temperature exhibits circadian oscillation, and temperature rhythms contribute to entrainment of peripheral tissue clocks ( Brown et al., 2002 ; Morf and Schibler, 2013 ). A range of EDCs have been demonstrated to impact thyroid axis physiology ( Duntas, 2015 ). Evidence of perturbations in rhythmic thyroid physiology is scarce, but some indication of disrupted circadian temperature rhythm is present in rodent models following EDC exposure. In Long-Evans rats, perinatal exposure to the anti-thyroid drug propylthiouracil (PTU) resulted in a permanently altered core temperature in adult male offspring. Non-monotonic impacts on daily temperature rhythm were present over a range of doses, with a reduction in mean core temperature of 0.4 degrees C at the highest dose ( Johnstone et al., 2013 ). Additionally, dioxin perturbed circadian temperature rhythm in Long Evans rats and in golden hamsters ( Gordon et al., 1996 ; Gordon and Miller, 1998 ).

Endocrine disrupting chemicals (EDCs) are endocrine-active chemical pollutants that disrupt reproductive, neuroendocrine, cardiovascular and metabolic health across species ( Gore et al., 2015 ). EDC exposure during embryonic and fetal development can result in severe impacts on growth and physiology; responses are frequently sexually dimorphic and may be mediated via epigenetic mechanisms ( Xin et al., 2015 ). EDC exposure has additionally been linked to development of hormonal cancers ( Gore et al., 2015 ). Classes of EDCs include polyhalogenated and phenolic compounds, phthalates, pesticides, steroid hormones, pharmaceuticals and personal care products ( Wee and Aris, 2017 ). EDCs enter the environment through industrial, municipal and agricultural runoff as well as consumer product use and subsequent entry into wastewater, and environmental exposures present health threats to humans, wildlife and aquatic organisms ( Kabir et al., 2015 ). The circadian clock is a light-driven transcriptional oscillator responsible for entraining 24-hour rhythms of physiology, behavior and metabolism ( Bass and Lazar, 2016 ; Malik et al., 2020 ). The molecular architecture of the circadian clock was first discovered in Drosophila , and was subsequently discovered to be highly conserved across species ( Wager-Smith and Kay, 2000 ). Across species, the core clock consists of an activator arm driving transcription of negative clock elements and clock-controlled genes, and repressor arms, which confer a time delay to clock-driven transcription and physiology ( Bhadra et al., 2017 ). In mammals, the suprachiasmatic nucleus (SCN) of the hypothalamus is the central pacemaker for the body, which directly responds to light input from the retina ( Takahashi, 2016 ). The SCN generates physiological rhythms of sleep, rest / activity, core body temperature, metabolism and neuroendocrine function, among other systemic processes. The molecular clock is present in most cells of the body, and organ-level peripheral oscillators are entrained by autonomic innervation, hormonal action (eg. glucocorticoids) and other synchronizing signals (zeitgebers) such as feeding / fasting rhythms ( Dibner et al., 2010 ). Clocks present across organisms orchestrate hourly, daily, weekly, monthly and seasonal rhythms in physiology ( Lincoln et al., 2003 ). In teleost fish, peripheral tissues are directly entrainable by light, though some regions of the hypothalamus contain pacemaker neurons as well ( Frøland Steindal and Whitmore, 2019 ; Moore and Whitmore, 2014 ; Watanabe et al., 2012 ). Genome duplication events during evolution led to multiple copies of many of the core clock proteins, lending a greater degree of biological redundancy to the teleost molecular clock ( Toloza-Villalobos et al., 2015 ). The circadian system is a critical mediator of physiological processes at all life stages, and disrupted clock function leads to a variety of pathologies such as sleep disorders, depression, chronic stress, cancer, and metabolic disease ( Bass and Lazar, 2016 ). Extensive bidirectional crosstalk exists between circadian and endocrine systems, and circadian regulation of the major endocrine axes has been recently reviewed. ( Neumann et al., 2019 ). Circadian rhythmicity is present at all levels of endocrine control, from synthesis and release of hormones, to sensitivity of target tissues to hormone action ( Neumann et al., 2019 ). Rhythms of endocrine factors can be short term (hourly and daily) or longer term (monthly and seasonal) and are strongly sexually dimorphic ( Bailey and Silver, 2014 ). Master pacemaker as well as endocrine organ clocks drive rhythmic endocrine physiology and endocrine organ clocks have been extensively characterized in mammalian systems, for example in pituitary ( Lin et al., 2015 ), pineal gland ( Borjigin et al., 2012 ), pancreas ( Perelis et al., 2016 ), adrenal ( Leliavski et al., 2015 ), liver ( Zwighaft et al., 2016 ), adipose ( Froy and Garaulet, 2018 ) and gonad ( Baburski et al., 2019 ; Sellix, 2015 ). Rhythmic endocrine physiology is observed across species and has a seasonal component as well, with seasonal variation observed in reproduction, growth and metabolism and locomotor activity rhythms ( Cowan et al., 2017 ; Gómez‐Milán and Lozano, 2007 ; Marchant and Peter, 1986 ; Nakane and Yoshimura, 2019 ; Yokota and Oishi, 1992 ). Perturbation of circadian systems can lead to a range of endocrine pathologies as summarized for human populations in Table 1 . Perturbed reproductive physiology as well as impaired fertility has been documented in female shift workers, and genetic variation in circadian genes bmal1 and npas2 was found to be associated with fertility outcomes (number of pregnancies and miscarriage rate) in a cohort study conducted in Finland ( Table 1 ). Shift work is associated with breast cancer incidence in human epidemiology studies, and similarly, circadian gene variance has been found to modify breast cancer risk, as well as breast cancer prognosis and survival in human populations ( Table 1 ). Epidemiological evidence links shift work with a range of metabolic pathologies including diabetes mellitus, impaired energy handling, body mass index and lipid alterations, and metabolic syndrome ( Table 1 ). Further insight into endocrine and metabolic dysregulation following circadian disruption has been demonstrated in core clock gene ablation studies conducted in mice, as outlined in Table 2 . Parallel to human epidemiological evidence, altered reproductive physiology and fertility are observed in mouse models following core clock gene ablation systemically or specifically in pituitary or gonad ( Table 2 ). Reproductive physiological perturbations include estrous cycle alterations, impaired fertility and fecundity and impaired gonadal steroidogenesis ( Table 2 ). In the hypothalamic – pituitary – adrenal axis (HPA), systemic or adrenal-specific core clock gene ablation alters rhythmicity of steroid production, impairs adrenal responsiveness to ACTH, impacts behavior and perturbs stress response in a sexually-dimorphic manner ( Table 2 ). Core clock gene ablation similarly causes extensive perturbation of metabolic health and energy homeostasis, leading to alterations in glucose homeostasis and insulin sensitivity, pancreatic dysfunction, altered lipid profiles, obesity and altered feeding rhythms following systemic as well as liver, pancreas or adipose-specific gene ablation ( Table 2 ). EDCs perturb endocrine physiology in part through binding a range of nuclear receptors ( Bainy, 2007 ; Casals-Casas and Desvergne, 2011 ). Nuclear receptor expression is clock-driven and rhythmic, therefore EDC action through nuclear receptor binding has a temporal component which is not well-characterized ( Teboul et al., 2008 ; Yang et al., 2006 ). Molecular crosstalk between ligand-occupied nuclear receptors and core clock proteins has been observed in mammalian systems, and are summarized in Table 3 . REV-ERB α/β and retinoic acid-related orphan receptor (ROR) α/β/γ are nuclear receptors that form a secondary feedback loop with the core molecular clock and directly modulate bmal1 transcription ( Guillaumond et al., 2005 ). Core clock repressors Cry1 and Cry2 interact with a range of nuclear receptors, including steroid hormone receptors, peroxisome proliferator-activated receptors (PPARs) and xenobiotic sensing receptors PXR and CAR ( Kriebs et al., 2017 ). Glucocorticoids are well-known zeitgebers, capable of synchronizing the molecular clock in the brain and in peripheral tissues, and function to modulate a range of physiological processes such as stress response, energy metabolism, cardiovascular and immune function and cognition ( Oster et al., 2016 ). Direct interaction and transcriptional feedback between the glucocorticoid receptor (GR) and components of the circadian clock has been documented in mammalian systems, and is outlined in Table 3 . Peroxisome proliferator-activated receptors (PPARs) are nutrient and metabolite sensing nuclear receptors which function to integrate circadian rhythms with metabolic outputs ( Charoensuksai and Xu, 2010 ). PPARα and γ directly modulate bmal1 transcription and the PPARα/RXRα heterodimer was found to modulate Clock / Bmal1- mediated transcription of period and cryptochrome genes ( Canaple et al., 2006 ; K. Nakamura et al., 2008 ; Wang et al., 2008 ). In vascular cells, retinoic acid-bound RARα and RXRα interacted with Clock and Mop4 and functioned as a negative regulator of Clock / Mop4 : Bmal1 transcriptional activity ( McNamara et al., 2001 ). Retinoic acids were also found to upregulate per1 , per2 and PPARα expression in an E-box dependent manner ( Shirai et al., 2006 ). A range of interactions between sex steroid receptors and the molecular clock have been documented. Estrogen and androgen signaling in the SCN modulates responses to photic input ( Abizaid et al., 2004 ; Butler et al., 2012 ; Karatsoreos et al., 2011 ) and programs circadian behavioral patterns ( Butler et al., 2012 ; Juárez-Tapia and Miranda-Anaya, 2017 ; Model et al., 2015 ; Royston et al., 2016 ). Estrogens modulate core clock gene expression in the SCN and in peripheral organs ( Nakamura et al., 2005 , 2001 ; T. J. Nakamura et al., 2008 ) and androgens modulate light-induced Per2 expression in the SCN ( Karatsoreos et al., 2011 ). Progesterone, via the progesterone receptor (PR) modulates the circadian clock in the mammalian uterus ( Rubel et al., 2012 ; Y. Zhang et al., 2019 ) and estrogen-related receptor α (ERRα) was found to modulate diurnal expression pattern of several core clock genes in mouse liver ( Dufour et al., 2011 ). Figure 1 depicts an overview of circadian aspects of nuclear receptor biology, and Table 3 and Figure 2 summarize documented molecular crosstalk between ligand-occupied nuclear receptors and core clock proteins in mammalian systems. EDCs are defined by the Endocrine Society as: “an exogenous chemical, or mixture of chemicals, that interferes with any aspect of hormone action” ( Zoeller et al., 2012 ). As documented above, circadian rhythmicity is innate to endocrine function, bidirectional crosstalk occurs between circadian and endocrine systems, and hormones exert feedback on both peripheral as well as central oscillators through nuclear receptor binding ( Table 3 ). It is plausible that in mimicking or disrupting hormone action, EDCs could similarly alter circadian oscillators themselves or rhythmic physiology. Clear mechanistic evidence demonstrating clock disruption by EDCs does not exist to date, but there are a range of studies demonstrating alterations in core clock gene expression, circadian transcriptional rhythms and rhythmic endocrine physiology following EDC exposure. This review will explore the evidence for parallel endocrine and circadian disruption following EDC exposure across species. Following a survey of the literature, parallel circadian and endocrine perturbations were identified in a range of species of teleost fish and in rodents. No studies were identified demonstrating parallel endocrine and circadian impacts in amphibian or reptilian species, but adverse impacts of EDCs on reproduction and development have been documented in these organisms as well as invertebrates ( Kanda, 2019 ). In humans, circadian disruption via shift work is associated with a range of endocrine and metabolic pathologies, as outlined in Table 1 , and one study explored the association between clock gene polymorphism and cadmium exposure in the context of thyroid cancer risk and severity ( Q. Zhang et al., 2019 ) ( Table 1 ). The studies identified fall into three major categories: circadian gene transcription and rhythmic physiology in the reproductive axis, altered locomotor or behavioral rhythms and altered metabolic homeostasis. In exploring the evidence, we will indicate when studies were circadian in design (conducted under constant conditions) and thus more adequately designed to assess EDC impacts on rhythmic physiology. Direct evidence presented in this manuscript is additionally summarized in Table 4 , sorted by EDC class.

There is a small but growing body of evidence of the potential for EDCs to disrupt endocrine and circadian systems in parallel. The studies discussed in this review demonstrate circadian impacts of EDCs in three major contexts: rhythmic physiology in the HPG(L) axis, locomotor and behavioral rhythms, and metabolic and energetic homeostasis. Figure 3 summarizes physiological perturbations following EDC exposure in the reviewed studies. EDC exposure alters expression level of core clock and circadian rhythm network genes, perturbs transcriptional and metabolite rhythmicity in brain and / or peripheral organs and alters locomotor and behavioral rhythms ( Table 4 ). The extent to which circadian systems modulate reproductive, metabolic/energetic and behavioral impacts of EDCs warrants further exploration. While extensive bidirectional molecular crosstalk between circadian and endocrine systems has been described, mechanistic links between EDC exposure and perturbations to the core clock and clock-driven physiology remain to be elucidated. In this review, we aimed to highlight parallel circadian and endocrine impacts following EDC exposure across species and propose a call for future research to explore mechanisms driving these observations. Across species, EDCs strongly perturbed core clock and circadian rhythm network gene transcription in the reproductive axis, and rhythmic physiology, such as gonadotropin release from pituitary, was impacted as well. EDC exposure characteristically reduced fecundity, and circadian transcriptional impacts were observed to a greater extent following exposure to EDC mixtures. Endocrine rhythms in the reproductive axis range from short (hourly) to long-term (seasonal). Across species, there is a balance of growth and energetics vs. reproduction ( Fernandez-Fernandez et al., 2006 ; Wade et al., 1996 ; D. Zhang et al., 2009 ). Timing of reproduction is critical, particularly in species exhibiting seasonal reproductive cycles, and variation in susceptibility to EDC action based on timing of exposure remains to be further clarified. Natural rhythms of behavior and group social behaviors are crucial for safety, predator avoidance, survival and reproductive success across species. The circadian clock drives behavioral rhythms to adapt physiology to the 24-hour light / dark cycle. A limited number of studies demonstrated EDC-induced impacts on circadian rhythm of locomotor activity, energy expenditure or food intake across organisms. EDCs altered social behaviors in teleost fish, including group preference, cohesive shoal formation, aggression and ability to adapt to a new environment. Whether there is a circadian component to these social behaviors is less clear due to limitations of experimental design in the studies described and certainly merits future exploration. Circadian regulation of social and reproductive behaviors has been described, and EDC exposure can alter sexual behaviors in fish ( Feng and Bass, 2016 ; Panksepp et al., 2008 ; Rubow and Bass, 2009 ; Söffker and Tyler, 2012 ). Further, circadian patterns of locomotor activity and feeding are interconnected with metabolic health ( Engin, 2017 ; Regnier et al., 2015 ). The circadian clock is intimately connected with energetic and metabolic homeostasis. Steroidogenic, xenobiotic, redox and lipid metabolic pathways, among others, are subject to circadian control ( Carmona-Antoñanzas et al., 2017 ; Gooley, 2016 ; Johnson et al., 2014 ; O’Neill and Feeney, 2013 ; Son et al., 2008 ). The tight coupling between circadian and metabolic oscillations is mediated through a variety of mechanisms, including NAD-dependent enzymes as well as nutrient, redox and temperature-sensing factors ( Asher and Schibler, 2011 ). Signaling through nuclear receptors modulates energetic homeostasis and exerts feedback on the clock in metabolically active tissues (Preidis et al., n.d.; Yang et al., 2012 , 2006 ). Comparing organismal responses to EDCs and obesogenic diet provides insight into overlapping transcriptional responses and phenotypes of metabolic and circadian disruption ( Kopp et al., 2017 ; Labaronne et al., 2017 ). Altered lipid metabolism in particular is implicated in obesogenic action of EDCs across species ( Carnevali et al., 2017 ; Maradonna and Carnevali, 2018 ). Given the critical role of nuclear receptors in orchestrating metabolic processes and the ability of EDCs to signal through NRs, experimental design aimed at characterizing rhythmic physiology of EDC / NR interactions will be helpful in delineating obesogenic mode of action of EDCs. Additional physiological impacts of EDCs include perturbation of circadian temperature rhythm ( Johnstone et al., 2013 ), redox metabolism ( Jiménez-Ortega et al., 2012 ), and immune response ( Faltermann et al., 2014 ; Xiao et al., 2017 ; Zhao et al., 2016 ). Circadian impacts of EDC exposure were sexually dimorphic as well as transgenerational ( Walker et al., 2014 ; Zhao et al., 2015a ). Future research is needed to investigate EDC action in the context of endogenous biological rhythms and whether parallel circadian disruption plays a role in exacerbating or driving endocrine pathologies. Extensive bidirectional crosstalk exists between circadian clocks and endocrine signaling. Gonadotropins, adrenal and gonadal steroids, PPAR ligands and retinoids via nuclear receptor signaling directly regulate core clock gene expression and modulate circadian rhythmicity in brain and peripheral organs ( Table 3 and Figure 2 ). Whether EDCs activate similar signaling pathways and directly modulate clock function warrants future exploration. Characterizing NR and AhR crosstalk with the molecular clock in the context of EDC exposure may prove crucial to understanding the nature of interaction between circadian and endocrine systems in mediating EDC effects. Characterizing circadian physiology in the context of life-stage dependent, multigenerational and seasonal effects of EDCs will also aid in understanding windows of vulnerability to EDC action across species.