MECHANISMS OF CIRCADIAN CLOCK INTERACTIONS WITH ARYL HYDROCARBON RECEPTOR SIGNALING

Abstract

Per-Arnt-Sim (PAS) domain-containing proteins are critical to homeostatic regulatory networks that mediate responsiveness to environmental change. PAS domains are multifunctional structural motifs that allow protein-protein interactions among family members, typically forming heterodimeric transcription factors to affect the transcription of target genes. Prototypical PAS domain-dependent pathways include the circadian clock network and metabolic regulation of the xenobiotic response through the aryl hydrocarbon receptor (AhR). Both pathways are increasingly linked to health, and alteration in their function contributes to development of disease. The aryl hydrocarbon receptor (AhR) demonstrates promiscuity in ligand binding and selectivity during heterodimer formation, which allows varied combinations of protein/protein interactions with other Per-Arnt-Sim (PAS) domain containing proteins and crosstalk among signaling pathways, including the molecular clockworks. AhR and the circadian signaling pathways are highly integrated and reciprocally regulated. AhR exhibits a rhythmic expression and time-dependent sensitivity to activation by AhR agonists. Conversely, AhR influences amplitude and phase of rhythms in circadian clock genes, hormones, and behavior. Understanding the molecular interactions between AhR and the clock provides insight into physiological regulation of rhythmic processes and provides an innovative approach to development of therapeutics.

Introduction

The revolution of the Earth about its own axis, which divides the solar day into a pattern of light and darkness, may be the most persistent environmental factor subserving the evolution of modern species. The sense of time endowed by this environment is accentuated by the presence of an internal mechanism for keeping time on a 24 h (circadian) scale, inherent in the genetic framework of living organisms ranging from cyanobacteria to humans. The internal, molecular program drives circadian oscillations within the organism that manifest at the molecular, biochemical, physiological and behavioral levels. The internal timing mechanism allows persistence of timekeeping in the absence of environmental cues, and anticipatory responses to changes in the environment to promote survival (Takahashi et al., 2008; Moore, 2013). By programming internal events to occur consistently at specific times across the 24 h cycle, physiological activity coordinates with the light/dark environment (Gillette & Tischkau, 1999; Gillette & Mitchell, 2002). For example, animals typically have bouts of rest and activity that correspond to specific times of day, allowing them to fill a particular ecological niche, to find food and avoid predators. Other examples of processes whose timing is controlled by a circadian clock include Neurospora crassa asexual spore production, leaf movements in plants, photosynthesis and nitrogen fixation in cyanobacterium (Bhadra et al., 2017), the firing rate of the suprachiasmatic nucleus (SCN) in mammals (Kuhlman & McMahon, 2006), clock gene expression levels and numerous aspects of metabolism (Takahashi, 2015, 2017).

Coherent circadian rhythms in mammals may be viewed as an organismic network, with a central hub located in the SCN of the ventral hypothalamus, the so-called “master clock”. Although individual neurons within the SCN are independent oscillators, coordinated function of specific SCN cell types is necessary for managing properly phased rhythms throughout the body (Yamaguchi et al., 2003; Liu et al., 2007; Ko et al., 2010). The SCN detects environmental signals and receives homeostatic feedback, which it uses to convey timing information to the rest of the organism through a variety of neural and endocrine mechanisms. Disruption of coordinated timing between the SCN and peripheral oscillators located in every organ system may upset the intricate homeostatic balance necessary for health and contribute to numerous pathologies, including cardiovascular disease, metabolic disease, cancer and others (Arble et al., 2015; Bass, 2017).

An important feature of the circadian timing system is its ability to adjust to resetting stimuli. Although the mammalian circadian timing system is sensitive to a variety of external and internal stimuli, the signals most relevant to this review are light, food and environmental chemicals. Nestled above the optic chiasm, the SCN receives monosynaptic neuronal input from the retina via specialized intrinsically photoreceptive retinal ganglion cells to function as the light-entrainable oscillator for the systemic circadian timing system (Hattar et al., 2003, 2006; Do & Yau, 2010; Hughes et al., 2015; Fernandez et al., 2016). N methyl d-aspartate (NMDA) receptor-mediated, glutamatergic neurotransmission activates signal transduction processes within the SCN to align the nucleus, and subsequently, the organism to light/dark cycles (Ebling, 1996; Gillette & Tischkau, 1999; Golombek & Rosenstein, 2010). Food can act independent of the SCN to entrain the circadian system. Although an anatomical site for the food-entrainable oscillator is yet to be determined, rhythms can be entrained by presentation of food at specific times of day. In response to food, animals demonstrate food anticipatory behavior, which is accompanied by an anticipatory rise in body temperature, increased corticosterone, gut motility, digestive enzyme activity and insulin secretion. Furthermore, rhythms can be shifted by altering the time when food is present. Importantly, this oscillator is independent of the SCN. The timing of food presentation alters clock gene transcripts in peripheral tissues, but not in the SCN; SCN ablation does not alter the timing of food anticipatory behavior. Thus, the system circadian timing system can be differentially entrained by light and food (Olivo et al., 2014; Hatori & Panda, 2015; Refinetti, 2015). Normally, these two signals harmonize with each other and the SCN coordinates timing of peripheral clocks. However, when the timing of light and food are presented out of their usual balance, rest/activity cycles remain controlled by the SCN, whereas peripheral systems may respond to food. Systemic disruption caused by discordance between the SCN and peripheral clocks may contribute to subsequent pathologies.

Rapid industrialization over the past 100 years has transformed the environment to trigger sweeping change in modern life-styles. Artificial lighting, computer and television screens, social demands incurred by the internet, work obligations, and weekend schedule changes may be considered “social zeitgebers” that act as environmental signals to affect circadian fidelity, and lead to chronic circadian disruption (Vetter, 2018). In addition, the industrial revolution and urbanization has increased exposure to chemical pollutants, whose impact on the circadian system in only beginning to be appreciated. The biological effects of many environmental xenobiotics are mediated through a common receptor mechanism, the aryl hydrocarbon receptor (AhR) signaling pathway. The AhR, similar to circadian clock components, is an ancient, evolutionarily conserved protein and member of the PAS domain family of proteins. AhR may function as a physiological bridging molecule between external and internal chemical signals; the exact physiological roles of AhR remain unclear, but seem to be multifactorial and certainly not limited to mediation of xenobiotic toxicity (Tian et al., 2015). Genomic organization, intron/exon splice patterns and conservation of the PAS domain suggests that brain-muscle ARNT-like 1 (BMAL1) and AhR are common derivatives of the same ancestral gene (Yu et al., 1999). AhR and components of its signaling pathways are regulated by the circadian clock. Conversely, anthropogenic and natural AhR ligands may impact clock function though AhR. The physiological interactions of AhR and the clock are further explored in this review.

SCN Communication with Peripheral Clocks

The SCN communicates timing information to the body through regulation of autonomic and neuroendocrine systems (reviewed in Dibner et al., 2010). Proximal SCN signals impact endocrine, pre-autonomic and intermediate neurons primarily in the dorsomedial hypothalamus, the medial preoptic area, and the subparaventricular zone of the hypothalamus (Kalsbeek et al., 2006). SCN neurons also project outside the hypothalamus to the paraventricular nucleus of the thalamus (PVT), the intergeniculate leaflet and the ventrolateral preoptic area (VLPO) (Hermes et al., 1996; Sun et al., 2001). Signals are relayed to pre-autonomic neurons in the brainstem and spinal cord, as well as endocrine neurons associated with secretion of corticotropin-releasing hormone, thyrotropin-releasing hormone and gonadotropin-releasing hormone (Kalsbeek & Buijs, 2002). Glutamate and gamma amino-butyric acid (GABA) (Buijs et al., 1995; Hermes et al., 1996) are primary neurotransmitters associated with SCN output, although a number of other molecules, including arginine vasopressin (Kalsbeek et al., 2006), cardiotrophin-like cytokine (Kraves and Weitz, 2006), prokineticin 2 (Cheng et al., 2002), vasoactive intestinal peptide (Kalsbeek & Buijs, 2002) and transforming growth factor α (Kramer et al., 2001) have also been identified.

The SCN has multiple outputs that influence hormone release. Although SCN-derived paracrine factors are sufficient to drive locomotor activity rhythms, direct neuronal connections are required for certain neuroendocrine rhythms (Lehman et al., 1987; Silver et al., 1996; de la Iglesia et al., 2003; Kalsbeek et al., 2008). A well-studied example of circadian control of peripheral endocrine rhythms is SCN regulation of the hypothalamic-pituitary-adrenal axis (HPA). The SCN provides stimulatory and inhibitory signals to influence the HPA axis (Tsang et al., 2014; Oster et al., 2017). SCN signaling to the PVN of the hypothalamus influences activity of corticotrophin-releasing neurons that regulate the secretion of adrenocorticotropic hormone (ACTH) from the pituitary gland (Vujovic et al., 2015). Furthermore, autonomic projections from the SCN through the PVN and intermediolateral column (IML) regulate diurnal changes in adrenal gland sensitivity to ACTH (Buijs et al., 1999). Retrograde labeling studies indicate that peripheral tissues receive both sympathetic and parasympathetic input from the SCN, suggesting that the clock regulates “rest” and “activity” phases (Buijs et al., 2003).

SCN-driven autonomic modulation of peripheral tissues is not limited to the adrenal gland; the pineal gland, heart, pancreas, intestines, ovaries and liver may also receive timing information from an SCN-derived autonomic signals (Buijs et al., 1999). Autonomic input from the SCN to the liver influences expression of hepatic clock and metabolic genes to regulate the daily plasma glucose rhythm (Kalsbeek et al., 2006, 2014). Inappropriately timed light exposure immediately resets liver rhythms, unless autonomic innervation to the liver is disrupted (Cailotto et al., 2005). Although SCN lesion does not always eliminate peripheral clock function, phase desynchrony among tissues suggests that the SCN coordinates rhythmicity among organ systems (Yoo et al., 2004; Guo et al., 2006). Over 100 transcripts lose circadian rhythmicity in the liver after adrenalectomy, which highlights the interaction of the SCN and the HPA axis in physiological regulation of clock function (Oishi et al., 2005).

Genome-wide transcriptional profiling indicates that between 4 (muscle, adipose tissue) and 16 (liver) percent of transcripts show circadian oscillations; moreover, over 50% of protein-coding transcripts are predicted to oscillate somewhere in the body (Hughes et al., 2009; Zhang et al., 2014). These oscillations subserve key organ-specific physiological functions. Most relevant to this review, xenobiotic detoxification in the liver, kidney and small intestine (Gachon et al., 2004), as well as carbohydrate (Lamia et al., 2008) and lipid (Le Martelot et al., 2009) metabolism in the liver and adipose tissue, are all clock-regulated events. The SCN likely coordinates inter-organ timing through autonomic and endocrine signals, to allow for systemic regulation of metabolic events.

The Molecular Clockworks

The molecular circadian clock, in mammals, consists of transcriptional activators and repressors that coordinate to function as interconnected positive and negative feedback loops (Reviewed in Takahashi, 2017). In the primary loop, transcriptional activators, CLOCK (Circadian locomotor output cycles kaput) and BMAL1, heterodimerize through their PAS domains bind at enhancer-box (Ebox) regions to drive transcription of PAS domain-containing clock genes, Period 1 (Per1), Per2 and Per3 and two Cryptochrome (Cry) genes Cry1 and Cry2 (Gekakis et al., 1998; Kume et al., 1999; Shearman et al., 2000). PER and CRY accumulate in the cytoplasm and form heterodimers that translocate to the nucleus, where they repress activity of CLOCK:BMAL1, effectively inhibiting their own expression (Lee et al., 2001; Lowrey & Takahashi, 2011). Posttranslational phosphorylation of PER by the casein kinase I family of enzymes (CKIδ/ε) regulate formation of dimers with CRY, nuclear import of the PER:CRY complex and degradation of PER (St John et al., 2014). CRY degradation is regulated by the F-box and leucine rich repeat proteins (FBLX3) and FBLX21 ubiquitin ligase complexes (Xing et al., 2013; Yoo et al., 2013). As PER and CRY are degraded, transcriptional repression is released and CLOCK:BMAL1 can begin a new cycle of transcription. Thus, PER and CRY stability are important for determining circadian period length.

The nuclear hormone receptors, Rev-erbα and Rev-erbβ are also rhythmically transcribed by CLOCK:BMAL1. REV-ERBα and REV-ERBβ compete with retinoic acid-related orphan receptors (RORα, β, γ) for binding at ROR elements (ROREs) in the promoter of the Bmal1 gene. REV-ERBs repress, whereas RORs activate, Bmal1 transcription (Preitner et al., 2002; Sato et al., 2004; Cho et al., 2012; Zhang et al., 2015), producing a rhythm that is anti-phase to that of Per and Cry. Finally, CLOCK:BMAL1 drive transcription of DBP (D-box binding protein) that can act together with the nuclear repressor NFIL3 (Nuclear Factor, Interleukin 3 regulated), a product of the REV-ERB/ROR loop, on D-boxes to affect rhythm transcriptional processes within the molecular clock (Mitsui et al., 2001; Gachon et al., 2004).

Dynamic epigenetic regulation of chromatin provides the appropriate template for activation and repression of transcription across the circadian cycle (reviewed in (Takahashi, 2017). CLOCK and BMAL1 associate with histone acetyltransferases to shape chromatin in preparation for transcription; CLOCK-derived histone acetyl transferase (HAT) activity acetylates BMAL1 (Etchegaray et al., 2003; Curtis et al., 2004; Doi et al., 2006; Hirayama et al., 2007). Interactions with the methyltransferase, MLL1 (Mixed lineage leukemia 1; Katada & Sassone-Corsi, 2010), and histone lysine demethylases, promote recruitment of CLOCK:BMAL1 to the target gene promoters, where co-activators are also recruited to enhance transcription (DiTacchio et al., 2011; Lande-Diner et al., 2013). As PER and CRY accumulate, they form a large repressor complex that interacts with CLOCK:BMAL1. The dbp-cullin-4 ubitquitin ligase complex covalently marks Ebox sites to recruit the PER complex. Subsequently, other repressor complexes get involved to deacetylate and methylate histones. RNA helicases are recruited to terminate transcription. Thus, chromatin remodeling facilitates transcriptional activation and repression across the circadian cycle to regulate rhythmic gene expression. Collectively, epigenetic regulation of chromatin, transcriptional activation and repression, and posttranslational modifications of repressor proteins establish period and amplitude of the self-sustained core clock oscillation. In addition, the circadian transcriptional drivers CLOCK:BMAL1 induce expression of other clock controlled genes, including numerous transcription factors and tissue specific output genes. This regulatory network allows for nycthemeral regulation of tissue function (Bozek et al., 2009; Sahar & Sassone-Corsi, 2013).

AhR Signaling

AhR is ligand-activated transcription factor of the basic helix-loop-helix-PAS (bHLH-PAS) family (Burbach et al., 1992). A large number of environmental pollutants, including polychlorinated dibenzo-p-dioxins, polychlorinated dibenzofurans, polychlorinated biphenyls and polycyclic aromatic hydrocarbons act through AhR to elicit their toxic effects (Denison & Nagy, 2003). Among those anthropogenic chemicals, 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) and β-Naphthoflavone (BNF) are prototypic AhR agonists with the highest affinity (Lindén et al., 2010; Tian et al., 2015). Dietary AhR ligands include, but may not be limited to, indole metabolites from cruciferous plants and flavonoids found in fruits and vegetables (Denison & Nagy, 2003). Arachidonic acid metabolites, equilenin, heme metabolites, indigoids and photoproducts of tryptophan metabolism are also endogenous ligands (Heath-Pagliuso et al., 1998; Nguyen & Bradfield, 2008). Finally, AhR may be activated in a ligand-independent manner by cyclic 3’5’ adenosine monophosphate (cAMP)-dependent activation of protein kinase A. AhR translocation into the nucleus in response to cAMP-signaling directs transcriptional activation that deviates significantly from xenobiotic-dependent events; cAMP-dependent activation may inhibit heterodimerization with ARNT (Oesch-Bartlomowicz et al., 2005). AhR responsiveness to a variety of man-made, natural and endogenous compounds, as well as its evolutionary conservation, raises important questions regarding its function in health and disease.

Beyond its role in xenobiotic metabolism, the AhR has physiologically relevant functions. Through multiple mechanisms, the versatile nuclear receptor AhR affects liver and vascular development, cell cycle progression, immune function and circadian rhythm (Jaeger, Khazaal, et al., 2017; Larigot et al., 2018). Genomic AhR signaling was elucidated experimentally using xenobiotic ligands (Figure 1A). In the absence of ligand, AhR partitions to the cytoplasm, complexed with two heat shock chaperones (hsp90), immunophilin-like protein X-associated protein 2 (XAP2) and the co-chaperone, p23 (Petrulis & Perdew, 2002). Hsp90 retains the structure of AhR needed for ligand binding and blocks a nuclear localization signal. XAP2 and p23 maintain the stability of the hsp90 complex (Guyot et al., 2013). Conformational change following ligand binding exposes a nuclear localization signal, which allows translocation of the AhR complex into the nucleus (Larigot et al., 2018). After the ligand is released, AhR heterodimerizes with aryl hydrocarbon receptor nuclear translocator (ARNT) in the nucleus, which subsequently targets AhR:ARNT to specific xenobiotic response elements (XRE) in target gene promoters. AhR:ARNT dimerization occurs through PAS domains, a feature shared with several circadian clock proteins. PAS-A and PAS-B regions act as an interactive surface for heterodimer and homodimer formation, interactions with hsp90, and ligand binding (Gu et al., 2000; McGuire et al., 2001). XRE binding and transcriptional initiation is facilitated by AhR:ARNT interactions with histone acetyltransferases and chromatin remodeling factors (Ko et al., 1996). A canonical AhR target gene is the Cytochrome P4501a1 (Cyp1a1), a phase I metabolizing enzyme that metabolizes polycyclic aromatic compounds and other P450 enzymes are canonical AhR:ARNT targets. In addition, AhR controls non-P450 transcripts that are involved in regulation of gene expression, drug metabolism, lipid metabolism, nucleic acid metabolism and small molecule biochemistry (Lo & Matthews, 2012). After DNA binding, a nuclear export signal signals AhR to exit the nucleus, where it is degraded by the 26S proteasome (Davarinos & Pollenz, 1999). AhR expression is attenuated by a negative feedback loop with the AhR repressor gene (AhRR), which is a target gene of AhR:ARNT (Baba et al., 2001; Haarmann-Stemmann & Abel, 2006). Although unable to bind ligand, AhRR inhibits AhR signal transduction by competing with AhR for ARNT dimerization and XRE binding, thereby acting as a transcriptional repressor (Figure 1A) (Gu et al., 2000; Haarmann-Stemmann & Abel, 2006).

Figure 1. AhR signaling pathways.

A. In genomic signaling, unbound aryl hydrocarbon receptor (AhR) typically partitions to the cytoplasm in a complex with heat shock protein 90 (hsp90), p23 and XAP. Ligand binding releases XAP and uncovers a nuclear localization signal. After translocation into the nucleus, AhR binds aryl hydrocarbon nuclear transporter (ARNT) and releases the hsp90/p23 complex for recycling into the cytoplasm. AhR:ARNT binds xenobiotic response elements (XREs) to drive transcription of target genes. Cytochrome P450 enzymes, particularly Cyp1A1 and Cyp1B1 are prototypical target genes that feed back to degrade xenobiotic ligands. AhR also induces expression of the aryl hydrocarbon receptor repressor (AHRR), which feeds back to inhibit AhR:ARNT-induced transcription. In addition AhR:ARNT activates other target genes involved in cell division and differentiation, and metabolism. B. As physiological roles for AhR are revealed, the importance of non-genomic signaling through AhR are becoming apparent. AhR can inhibit activity of other nuclear receptors, including estrogen receptor (ER), androgen receptor (AR), thyroid hormone receptor (TR), Wnt/β-catenin and nuclear factor kappa B (NF_κB) by either binding directly to receptor proteins or by binding and diverting the action of coactivators. AhR promotes increases in intracellular Ca²⁺ by actions on membrane Ca²⁺ channels or through release from the endoplasmic reticulum, leading to increased arachidonic acid and inflammation. AhR can also increase MAPK activity, which can synergize with Ca²⁺ to promote inflammation through production of arachidonic acid. MAPK activity also mediates AhR’s regulation of cellular senescence, apoptosis and proliferation.

Although genomic AhR signaling is essential to the toxicological effects of dioxins, more recent studies highlight the physiological relevance of AhR, which may be arbitrated through a combination of genomic and non-genomic mechanisms. Pathways downstream of AhR activation may be influenced by ligand affinity, cell-type and other environmental factors. Microarray and genome-wide analysis of XRE sequences have identified numerous genes regulated by AhR (Tijet et al., 2006; Boutros et al., 2009; Dere, Lee, et al., 2011; Dere, Lo, et al., 2011; Lo & Matthews, 2012). Formation of heterodimers with alternative partners allows AhR to target diverse signaling pathways. Alternative signaling through AhR includes genomic cross-talk with other nuclear receptors, and non-genomic regulation of cell cycle and mitogen-activated protein kinase (MAPK) cascades, modulation of the immune system and activation of immediate early genes (Matsumura, 2009; Patel et al., 2009; Tanos et al., 2012).

AhR cross-talk with other nuclear receptors, including those for steroid hormones, thyroid hormones and retinoids involves multiple mechanisms (Figure 1B) (Widerak et al., 2006; Swedenborg & Pongratz, 2010). Perhaps the best explored is the interaction with the estrogen receptor (ER), where both classical and non-classical mechanisms contribute to inhibition of the estrogen response by AhR activation (Beischlag & Perdew, 2005; Khan et al., 2006; Swedenborg & Pongratz, 2010). AhR:ARNT act as transcriptional repressors at inhibitory dioxin response elements (iDRE) to block activation of certain genes by ER, at least in part by blocking binding of ER to its DNA recognition site (Safe et al., 1998; Wang et al., 2001). In addition, AhR may directly bind co-activator proteins, including p300 and cAMP response element binding protein (CBP), rendering them unavailable as co-activators in ER-mediated transcription (Beischlag & Perdew, 2005; Ma et al., 2009). Formation of AhR:ARNT complexes decreases availability for ARNT to complex with ER, which inhibits ER beta transcriptional activation (Rüegg et al., 2008). Direct binding of AhR with ER diverts the complex away from estrogen-responsive elements (EREs) and towards AhR-responsive genes (Ahmed et al., 2009). Thus, AhR interacts at multiple levels to disrupt ER signaling.

AhR also interacts with various intracellular signaling pathways, including protein kinases, calcium signaling, and others, in a non-genomic manner (Figure 1B) (Haarmann-Stemmann & Abel, 2006; Matsumura, 2009; Larigot et al., 2018). Activation of AhR promotes increased intracellular calcium and subsequent activation of arachidonic acid production and MAPK signaling (Tsai et al., 2014). In breast cancer cells, AhR activation promotes phosphorylation of MAPK and inhibits phosphorylation of the phosphotidyl inositol 3 kinase/protein kinase B (PI3K/AKT) pathway leading to cell cycle arrest and cell senescence (Wang, Xu, et al., 2014). Similarly, AhR inhibits proliferation and differentiation of osteoblasts through MAPK phosphorylation (Yu et al., 2014). Finally, AhR promotes germ cell apoptosis through activation of extracellular signal-regulated kinase (ERK1/2), p38 and c-jun N-terminal kinase (JNK) (Banerjee et al., 2016). In addition, activated AhR modifies cellular adhesion by disrupting focal adhesion points (Tomkiewicz et al., 2013). Thus, AhR acts as a multifaceted hub to regulate intracellular signaling and numerous physiological processes, including inflammation, cell adhesion, proliferation, cell senescence and death, as well as modifying the transcriptional regulation of other nuclear receptors. Physiological functions of AhR may include interactions with the circadian clock.

Regulation of AhR Signaling by the Circadian Clock

AhR, Arnt and Cyp1a1 display variation with 24-h periodicity (Shimba & Watabe, 2009; Wang, Zhang, et al., 2014). AhR transcripts peak at the time of lights-off in the SCN and 4 h earlier in liver. Oscillations persist under constant darkness conditions, albeit with slightly reduced amplitude in the SCN; the similarity to changes in the pattern of Bmal1 suggest the potential for co-regulation of these transcripts (Mukai et al., 2008). The AhR target gene, Cyp1a1, also shows circadian variation, albeit out of phase with AhR (4–8 h delay) (Mukai et al., 2008). Nycethemeral expression of AhR protein demonstrates diurnal changes in the availability of AhR for activation (Richardson et al., 1998; Shimba & Watabe, 2009), which implies that sequelae associated with AhR activation may also be regulated by the circadian clock. Transcripts of AhR target genes, Cyp1a1, Cyp1b1, and Arnt, although expressed at low levels without stimulation, are diurnally rhythmic with increased expression during the day (Richardson et al., 1998; Huang et al., 2002; Mukai et al., 2008). Induction of the AhR target genes Cyp1a1 and Cyp1b1 in response to a prototypical AhR ligand provides evidence for differences in AhR responsiveness that are driven by the clock (Qu et al., 2007, 2009, 2010). Cyp1a1 and Cyb1b1 are induced by AhR activation regardless of the time of day, however, responsiveness is significantly enhanced when exposure occurs at night (Qu et al., 2007, 2010). Variation in induction of Cyp1a1 is abolished in Per1 and/or Per2 mutant mice (Qu et al., 2009, 2010). Furthermore, oscillations of AhR mRNA and time-dependent sensitivity to agonists are absent in CLOCK mutant mice (Tanimura et al., 2011). Although these data suggest that the clock is an important regulator of AhR, molecular mechanisms underlying these effects remain largely unknown.

Molecular Interactions between AhR and Clock Components

PAS domains are structural motifs that govern protein-protein interactions that ultimately target gene transcription in response to environmental change (McIntosh et al., 2010). PAS domains in AhR bind ligands and regulate interaction with partners that are required to direct it to specific sequences on the promoters of target genes (Reisz-Porszasz et al., 1994; McGuire et al., 2001). PAS domains in circadian clock proteins, including CLOCK, BMAL1 and PER, likewise facilitate protein-protein interactions that determine transcriptional activity to regulate molecular rhythmicity. PAS domain-containing proteins are notably promiscuous in both their ligand binding and selectivity of dimerization partners, which allows for multivariate protein-protein interactions and significant crosstalk among intracellular signaling pathways. The classic AhR partner, ARNT shares significant sequence similarity to BMAL1, including similar intron/exon splice patterns and conservation of five exons that compose the PAS domain (Yu et al., 1999). Known promiscuity of PAS interactions, coupled with conserved similarities to clock-relevant proteins, suggests that AhR may interact with clock proteins to influence clock function. The aggregate data suggest that interactions between the clock and AhR signaling are reciprocal and significant.

Activation of AhR alters circadian clock gene transcripts and suppresses molecular rhythms (Garrett & Gasiewicz, 2006; Mukai et al., 2008; Xu et al., 2013). Reciprocally, genetic alteration of the circadian clock influences AhR signaling and sensitivity to agonist-induced activation of AhR and AhR target genes (Qu et al., 2007, 2009, 2010). Nycthemeral expression of AhR mRNA and protein levels is regulated by CLOCK:BMAL1 interactions at Ebox elements within the AhR promoter (Garrison & Denison, 2000; Huang et al., 2002; Zhang et al., 2009). Molecular mechanisms by which AhR is regulated by the PER complex remain unknown.

AhR may directly interact with core clock proteins to affect the molecular clockworks. Subsequent to ligand activation, AhR can form multiple heterodimers. The dimerization partner determines the target site for DNA binding and transcriptional activity. AhR:ARNT heterodimers bind XRE elements to drive transcription of target genes. In contrast, AhR may also interact with BMAL1, which has high homology with ARNT (Tischkau et al., 2011; Xu et al., 2013). The AhR:BMAL1 heterodimer disrupts CLOCK:BMAL1 activation at Eboxes on the Per1 promoter. The AhR interaction with BMAL1 displaces CLOCK from the complex and acts as a repressor of Ebox-dependent Per1 transcription, thereby dampening and delaying the Per1 rhythm (Xu et al., 2013). Inhibition of Ebox function provides a mechanism for dampening the amplitude and perhaps delaying the phase of the endogenous Per1 rhythm (Figure 2A).

Figure 2. Mechanisms of crosstalk between AhR and the molecular clock.

A. Activated AhR can physically interact with BMAL1, thereby reducing CLK:BMAL1 interactions. AHR:BMAL1 heterodimers bind to Ebox elements in the Per1 promoter, where they act to repress Per1 transcription. In addition, AhR inhibits light-induced Per1 expression through activation of JNK, which blocks CREB:CBP activity at Cre elements in the Per1 promoter. B. Rhythmic transcription of AhR is driven by CLK:BMAL1 at EBox promoter elements. PER1 and PER2 inhibit AhR transcript levels. Although the molecular mechanism has not been determined, PERs can inhibit CLK:BMAL1 function at the EBox.

AhR activation may also affect Per levels through an Ebox-independent mechanism. CRE elements mediate acute induction of Per1 transcription in the SCN in response to light (Travnickova-Bendova et al., 2002; Tischkau et al., 2003). Light-induced glutamatergic neurotransmission increases intracellular calcium and activation of MAPK, phosphorylation of cAMP regulatory element binding protein (CREB) and CRE-mediated transcription of Per1 (Ginty et al., 1993; Obrietan et al., 1998). Similarly, AhR activation has extensive effects on MAPK signaling (Tan et al., 2002; Ma et al., 2009). AhR agonists can activate JNK through AhR (Weiss et al., 2005, 2008; Diry et al., 2006). AhR activation blocks CRE-dependent Per1 transcription without affecting phosphorylation of CREB. Rather, AhR phosphorylates and activates JNK, which represses CRE-mediated transcriptional activity (Zhang et al., 2004; Xu et al., 2013). Suppression of CRE-mediated Per1 transcription provides a non-genomic mechanism for AhR to influence responsiveness of the molecular clockworks to phase resetting signals (Figure 2A).

AhR Regulation of Circadian Rhythmicity

Mice bearing a germ-line deletion of the AhR (AhRKO) provide a model for investigation of interactions between AhR and the circadian clock. AhRKO mice readily entrain to a light/dark cycle, re-entrain their clock in response to changes in the timing of light/dark cycles, and maintain a free-running period similar to wild-type littermates (Mukai et al., 2008). However, increased variability in circadian parameters, which is commonly observed after deletion of genes associated with clock stability, has been observed. AhRKO mice display increased daytime activity, variability in activity onset and altered patterns of nocturnal activity compared to wild-type mice (S. Tischkau, unpublished observations). Common to the C57BL/6 strain, AhRKO mice have a period of decreased activity the night, which is often referred to as a “nap”. The timing of nocturnal inactivity occurs 1–2 h earlier in the AhRKO mice, similar to B6.D2NAahrd/j mice, which are genetically altered in the region that includes the AhR locus (Hofstetter et al., 2007). The aggregate behavioral data suggest a modulatory effect of AhR on clock function, which is similar to the type of effect of other nuclear receptors, such as REV-ERBα or RORa (Sato et al., 2004; Akashi & Takumi, 2005).

Activation of AhR also impacts circadian timing. TCDD exposure alters circadian rhythms in feeding behavior, locomotor activity, hormone levels and gene expression (Pohjanvirta et al., 1996; Miller et al., 1999; Garrett & Gasiewicz, 2006; Xu et al., 2010). Locomotor activity is phase-advanced by low dose TCDD treatment, whereas high doses can cause rhythm splitting and arrythmicity (Miller et al., 1999). Rhythms of clock gene transcripts, specifically Per1 and Bmal1, and the PER1 proteins are also perturbed in the SCN of TCDD-exposed mice (Mukai et al., 2008). Similarly, large doses of TCDD alter the timing of food intake, such that 50% of daily food consumption occurs during the early morning in nocturnal rats (Pohjanvirta & Tuomisto, 1990). Unfortunately, that study did not explore locomotor activity, which may account for the changes in feeding. However, these results definitively demonstrate that AhR activation alters the timing of food intake; treated animals seemed to be unable to identify the appropriate time to eat. Effects of AhR activation on plasma hormones indicate a similar effect on clock-regulated function. Corticosterone, prolactin, thyroid hormone and melatonin rhythms are attenuated after AhR activation (Jones et al., 1987; Pohjanvirta et al., 1989). The common denominator is that rhythmicity of each of these factors is controlled by the circadian clock in the SCN. These studies indicate that acute and chronic activation of AhR can affect physiological rhythms.

Perhaps more important than the consequences of AhR activation on central clock function of the SCN are the effects of AhR on circadian rhythmicity in peripheral tissues, where AhR is more highly expressed. Genomic depletion of AhR has a greater effect on rhythmicity in peripheral tissues compared to the SCN. AhRKO mice have increased amplitude of clock gene rhythms, which may protect from environmentally induced circadian disruptions incurred by jet lag or high fat diet (Jaeger, Xu, et al., 2017). Conversely, AhR activation suppresses amplitude and delays phase of Per1 and Bmal1 rhythms in the liver (Mukai et al., 2008; Xu et al., 2010, 2013). Similarly, AhR activation subsequent to TCDD administration significantly altered the phase, period and amplitude of circadian rhythms of Per1, Per2, as well as the numbers of antigenically and functionally defined hematopoietic progenitor cell classes in bone marrow (Garrett & Gasiewicz, 2006). Because cyclic expression of Per is required for rhythmicity, and constitutive or persistently amplified expression of core clock genes can alter period or completely disrupt rhythms, the effects of AhR on cycling of core clock gene transcripts may be important (Zeng et al., 1994; Wang & Tobin, 1998; Kadener et al., 2008).

Physiologically, AhR interactions with the clock may provide a fundamental integration of cellular metabolism. AhR activation alters the expression of genes associated with cholesterol and fatty acid synthesis, glucose metabolism and circadian rhythm in the liver (Sato et al., 2008). Many of the same genes affected by AhR activation are controlled physiologically by the circadian clock. For example, 3-hydroxy-3-methylglutaryl-coenzyme A reductase (HMGCR), the rate limiting enzyme in the synthesis of cholesterol, as well as the sterol element binding protein (SREBP), which is important in the regulation of sterol synthesis, are expressed with circadian periodicity under physiological conditions and downregulated by AhR activation (Brewer et al., 2005; Acimovic et al., 2008; Sato et al., 2008; Le Martelot et al., 2009; Matsumoto et al., 2010). These data make a compelling case that AhR interacts with the circadian system to regulate metabolic outputs. AhR activation weakens rhythm amplitude, whereas its inhibition bolsters rhythm amplitude (Figure 3). Although the physiological consequences of amplitude regulation remain to be determined, dampened amplitudes are associated with poor metabolic outcomes. Depletion of AhR increases healthy metabolic function and protects against the detrimental effects of poor diet (Xu et al., 2015; Jaeger, Xu, et al., 2017). It seems entirely plausible that these events are intricately connected.

Figure 3. Physiological relevance of AhR interactions with the Per1 promoter.

CLK:BMAL1 drive the endogenous rhythm of many transcripts in WT mice, including the Pers, AhR and Cytochrome P450’s (Cyps) and metabolically important genes such as 3-Hydroxy-3-Methylglutaryl-CoA Reductase (HMGCR) and sterol-regulatory element binding protein (SREBP). When AhR is deleted, the amplitude of circadian oscillations is increased. AhR activation inhibits CLK:BMAL1-mediated transcription, thereby suppressing amplitude and delays phase of target gene transcripts. Suppression of and changes of phase are associated with disease process including sleep disorders, metabolic syndrome, cancers and cardiovascular disease.

Cross-talk between AhR Signaling and Light Signaling in the Circadian Clock

Perception of the environmental presence of light and darkness is the perhaps most important function of the SCN. The SCN recognizes light at night as an error in timing and sets in motion a series of events to realign the clock with the environment. Light in the early night is perceived the same as an extension of daytime; thus, the SCN clock will respond by delaying the rhythm the next day. Alternatively, light in the late night is interpreted by the SCN as an early beginning to the forthcoming day. The clock responds to late night light by moving ahead, or phase advancing, on the ensuing day. Mechanisms that underlie clock resetting in response to changes in environmental light have been intensely investigated (reviewed in Antle et al., 2009; Golombek & Rosenstein, 2010). Essentially, phase resetting of the SCN clock in response to light proceeds through activation of NMDA-type glutamate receptors on retinorecipient SCN neurons, leading to an influx of calcium, activation of MAPKs and CREB, and ultimately to induction of the clock gene Per1. Although glutamate is the primary neurotransmitter that mediates the effect of light on the SCN, numerous other signals can interact with the glutamatergic input to modulate the light signal. For example, pituitary adenylate cyclase activating peptide (PACAP) interacts with glutamate to increase the magnitude of phase delays and to attenuate phase advances (Chen et al., 1999). Serotonin, or its precursor, tryptophan also acts to temper the magnitude of light and glutamate-induced responses in the SCN (Glass et al., 1995; Quintero & McMahon, 1999; Sterniczuk et al., 2008).

6-formylindolo[3,2-b]carbazole (FICZ), a high affinity AhR agonist physiologically derived from tryptophan metabolism (Helferich & Denison, 1991; Heath-Pagliuso et al., 1998; Bergander et al., 2004) blocks the glutamate-induced phase resetting of the SCN clock, in an AhR-dependent manner, similar to the effects of serotonin (Mukai & Tischkau, 2007). Activation of AhR with other ligands also attenuates light/glutamate-induced clock resetting (Xu et al., 2013). Furthermore, AhRKO mice have an increased responsiveness to light-induced phase resetting. These data are consistent with AhR acting as a modulatory brake on the circadian clockworks. Activation attenuates circadian clock responses, whereas removal of AhR activity from the system accentuates circadian responses to external signals.

Collectively, this provides compelling evidence for crosstalk between AhR signaling and the circadian clock.

Conclusions

Adaptation to environmental change is critical to survival. On a planet that rotates about its own axis creating a day that is consistently 24 hours in length, the value of measuring time on a daily scale is reflected in the incorporation of an internal clock into the genetic architecture of nearly all organisms. A robust defense in the face of exposure to environmental toxins, especially those present in food, is similarly adaptive and critical to survival. The bHLH-PAS domain family of proteins is central to regulation of each of these critical biological processes, which intersect in a complex and reciprocal manner. Disruption of clock function and alteration of AhR signaling contribute to a vast array of disease processes that include metabolic dysfunction and cancers. Components of the AhR signaling pathway are expressed in a diurnal pattern. Moreover, activation of AhR pathway is influenced by the circadian clock; sensitivity of the pathway to activation by high affinity agonists varies across the day. Conversely, both activation and downregulation of the AhR pathway alters the normal rhythmicity (Figure 3). How AhR activation impacts physiological rhythmicity remains to be determined. Interestingly, AhR activation has greater effects on clock processes in peripheral tissues, which may create discordance between the SCN master clock and clocks in other organs. The clock gene, Per1, appears to be central to the intersection of AhR and circadian signaling, although further studies are required to dissect molecular interactions. A better understanding of the crosstalk will provide insight into the physiological role of the AhR, as well as provide novel insight into the pathological events associated with clock disruption.

List of Abbreviations

ACTH

Adrenocorticotropic hormone

AhR

Aryl hydrocarbon receptor

AhRKO

aryl hydrocarbon receptor knockout

AhRR

Aryl hydrocarbon receptor repressor

AKT

protein kinase B

ARNT

aryl hydrocarbon receptor nuclear transporter

bHLH

basic helix-loop-helix

BMAL1

Brain muscle ARNT-like 1

BNF

beta-napthoflavone

Ca²⁺

calcium

cAMP

cyclic 3’5’ -adenosine monophosphate

CBP

CREB binding proten

CLOCK/CLK

Circadian locomotor output cycles kaput

CRE

cAMP response element

CREB

cAMP response element binding protein

Cry

Cryptochrome genes

Cyp1a1/1b1

cytochrome P450 enzyme 1a1, 1b1

DBP

D site of albumin promoter-binding protein

ER

estrogen receptor

ERK1/2

extracellular signal-regulated kinase 1/2

FBLX3 or 21

F-Box and Leucine Rich Repeat Protein 3 or 21

FICZ

6-Formylindolo[3,2-b]carbazole

GABA

Gamma amino-butyric acid

h

Hour

HAT

histone acetyl transferase

HMGCR

3-Hydroxy-3-Methylglutaryl-CoA Reductase

Hsp90

heat shock protein 90

JNK

c-jun N-terminal kinase

MAPK

mitogen activated protein kinase

MLL1

mixed lineage leukemia 1

NFIL

Nuclear factor interleukin 3 regulated

NF_κB

Nuclear factor kappa B

NMDA

N methyl d-aspartate

PACAP

pituitary adenylate cyclase activating peptide

PAS

Period-Aryl hydrocarbon receptor nuclear transporter-simpleminded

Per

Period genes

PI3K

Phosphatidylinositide 3-kinase

PVN

Hypothalamic paraventricular nucleus

PVT

Paraventricular nucleus of the thalamus

RNA

Ribonucleic acid

ROR

Retinoid-related orphan receptor

SCN

Suprachiasmatic nucleus

SREBP

sterol regulatory element-binding protein

TCDD

2,3,7,8 Tetrachlorodibenzo-p-dioxin

VLPO

Ventrolateral preoptic area

XAP2

X associated protein 2

XRE

xenobiotic response element