The tryptophan derivative 6-formylindolo[3,2-b]carbazole, FICZ, a dynamic mediator of endogenous aryl hydrocarbon receptor signaling, balances cell growth and differentiation

Abstract

The aryl hydrocarbon receptor (AHR) is not essential to survival, but does act as a key regulator of many normal physiological events. The role of this receptor in toxicological processes has been studied extensively, primarily employing the high-affinity ligand 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD). However, regulation of physiological responses by endogenous AHR ligands remains to be elucidated. Here, we review developments in this field, with a focus on 6-formylindolo[3,2-b]carbazole (FICZ), the endogenous ligand with the highest affinity to the receptor reported to date. The binding of FICZ to different isoforms of the AHR seems to be evolutionarily well conserved and there is a feedback loop that controls AHR activity through metabolic degradation of FICZ via the highly inducible cytochrome P450 1A1. Several investigations provide strong evidence that FICZ plays a critical role in normal physiological processes and can ameliorate immune diseases with remarkable efficiency. Low levels of FICZ are pro-inflammatory, providing resistance to pathogenic bacteria, stimulating the anti-tumor functions, and promoting the differentiation of cancer cells by repressing genes in cancer stem cells. In contrast, at high concentrations FICZ behaves in a manner similar to TCDD, exhibiting toxicity toward fish and bird embryos, immune suppression, and activation of cancer progression. The findings are indicative of a dual role for endogenously activated AHR in barrier tissues, aiding clearance of infections and suppressing immunity to terminate a vicious cycle that might otherwise lead to disease. There is not much support for the AHR ligand-specific immune responses proposed, the differences between FICZ and TCDD in this context appear to be explained by the rapid metabolism of FICZ.

Keywords:

• 6-Formylindolo[3,2-b]carbazole
• aryl hydrocarbon receptor
• tryptophan
• endogenous ligand
• immunity
• cancer
• toxicity
• feedback loop
• cytochrome P4501A1
• CYP1A1
• TCDD
• metabolism

1. Introduction

When first discovered, the aryl hydrocarbon receptor (AHR or dioxin receptor) was assumed to mediate metabolism of xenobiotics (Poland et al. 1976; Poland and Knutson 1982) since it bound 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) and similar halogenated hydrocarbons with high affinities. However, in 1991 Nebert proposed that similar to nuclear receptors that bind hormones and regulate both cellular growth and differentiation, and the expression of particular cytochrome P450 (CYP) isozymes, the dioxin receptor plays a key role in regulating growth and the CYP1 family of enzymes metabolizes its ligands (Nebert 1991). In addition, he proposed that dioxins and other xenobiotic ligands of AHR mimic natural ligands.

Later, the extensive evolutionary conservation of this receptor (Hahn et al. 1997, 2017) and its role in regulating the constitutive expression of hundreds of genes (Tijet et al. 2006; Sartor et al. 2009; Dere et al. 2011; Salisbury et al. 2014) provided strong indications for important endogenous functions. The AHR belongs to the basic helix-loop-helix/Per-Arnt-Sim (bHLH/PAS) family of transcription factors, some of whose members perform ancestral sensory functions and regulate circadian rhythms (McIntosh et al. 2010). However, the physiological roles of the AHR in vertebrates remain to be elucidated in detail.

Although this receptor is not essential for survival, AHR-deficient mice age prematurely and, furthermore, exhibit abnormal vascular development, with spontaneous development of lesions in several organs (Nebert et al. 1984; Schmidt et al. 1996; Fernandez-Salguero et al. 1997; Lahvis et al. 2000; Sauzeau et al. 2011; Hu et al. 2013). Moreover, deletion of the AHR affects endocrine, and reproductive functions (Abbott et al. 1999; Benedict et al. 2000; Baba et al. 2005, 2008; Moran et al. 2012); results in numerous pathologies related to defects in hematopoietic progenitor cells (Fernandez-Salguero et al. 1997; Bennett et al. 2015); disturbs hepatic energy homeostasis (Girer et al. 2016); and disrupts immune responses in a manner that enhances sensitivity to inflammatory stimuli (Thatcher et al. 2007; Sekine et al. 2009; Kiss et al. 2011). Thus, the emerging view is that this receptor plays an essential role in maintaining cellular homeostasis, control, and balancing physiological functions in response to endogenous ligands. One of its highly important functions appears to involve maintenance of progenitor cell populations, e.g. coordinating proliferation and differentiation and protecting against cellular senescence (Singh et al. 2009; Gasiewicz et al. 2014; Bennett et al. 2015). Indeed, persistent repression of the AHR appears to be required in order to safeguard the pluripotency of stem cells (Ko et al. 2016).

Most of our current knowledge in this context originates from decades of studies employing the metabolically inert, high-affinity ligand TCDD, which have clearly revealed the negative consequences of over-activating AHR. Indeed, TCDD is a highly toxic man-made chemical. Additional knowledge has been gained from the phenotypes of AHR knockout (AHR KO) and AHR-overexpressing strains of mice, extreme situations in which primarily physiological processes are altered. Interestingly, the first strains of mice carrying AHR null-alleles appeared rather normal but important functions of the AHR for development and for physiologic homeostasis are now well established (see Section 4). The present review focuses on novel information obtained using the high-affinity endogenous ligand 6-formylindolo[3,2-b]carbazole, FICZ, which indicates that the AHR is a highly dynamic regulator of numerous homeostatic processes.

2. Exogenous and endogenous activation of AHR signaling

The AHR is activated by binding of many low-molecular-weight substances, having dissociation constants (K_d) ranging from 10⁻¹³ to 10⁻³M. CYP1A1, which encodes a member of the cytochrome P450 superfamily of enzymes, is typically the gene most highly up-regulated in response to activation of this receptor. CYP1A1 mediates the oxidative metabolism of a wide variety of exogenous, as well as endogenous substrates (Nebert and Dalton 2006).

In the absence of ligands, AHR is part of a cytoplasmic chaperone complex that also includes a dimer of Hsp90, and the co-chaperones p23 and AIP. In response to a ligand, the conformation of the receptor changes to expose a nuclear localization sequence that allows dissociation from the chaperone complex and translocation into the nucleus (Petrulis and Perdew 2002). Once inside the nucleus, the AHR heterodimerizes with the aryl hydrocarbon receptor nuclear translocator (ARNT) and the resulting complex binds to response elements (AHREs) in the promotor regions of target genes, where it recruits transcription cofactors and proteins that remodel chromatin, promoting gene transcription ( reviewed in Beischlag et al. 2008; Jackson et al. 2015). Interestingly, genes involved in developmental and tumorigenic processes, including those encoding components of the Wnt/β-catenin and TGFβ signaling pathways, along with genes encoding xenobiotic-metabolizing enzymes are all targets (Chang et al. 2007; Sartor et al. 2009; Wang et al. 2010; Faust et al. 2013; Salisbury et al. 2014).

Identification of endogenous agonists of AHR remains the focus of intensive research. Among these, FICZ, derived from tryptophan (Trp), binds to this receptor with higher affinity than any other compound yet tested, including TCDD giving an estimated dissociation constant K_d of 0.07 nM in competitive binding assays with 1 nM TCDD using the hydroxylapatite assay (Rannug et al. 1987).

FICZ was initially detected as a photooxidation product of Trp (Rannug et al. 1987, 1995). The chemical structures of two Trp-derived AHR ligands were determined by means of mass spectrometry, ¹H-NMR, and ¹³C-NMR, also fluorescence, UV, and infrared spectra were recorded. The two Trp-derived ligands were indolo[3,2-b]carbazoles identified as the monosubstituted 6-formylindolo[3,2-b]carbazole, i.e. FICZ (MW 284) and the symmetrical 6,12-diformylindolo[3,2-b]carbazole (dFICZ, MW 312), respectively (Rannug et al. 1995). Ultraviolet A and B radiation (UVA and UVB) and visible light can all generate FICZ upon irradiation of Trp solutions, although UVB is most efficient (Wincent et al. 2009). Recently it was shown that the generation of hydrogen peroxide (H₂O₂) during irradiation is involved in the generation of FICZ, but also that H₂O₂ alone in the absence of light can generate FICZ from Trp (Smirnova et al. 2016). Other light-independent pathways for the formation of FICZ were also described in detail by Smirnova et al. (2016). Briefly, the Trp metabolites tryptamine and indole-3-pyruvic acid give rise to FICZ and its oxidation product indolo[3,2-b]carbazole-6-carboxylic acid (CICZ) via the common denominator indole-3-acetaldehyde. The human mitochondrial enzymes monoamine oxidase A and B are able to catalyze the deamination of tryptamine to give indole-3-acetaldehyde, and indole-3-pyruvate can be de-carboxylated to form indole-3-acetaldehyde that in multistep reactions yields FICZ (Figure 1).

Figure 1. Formation pathways of 6-formylindolo[3,2-b]carbazole (FICZ) and indolo[3,2-b]carbazole-6-carboxylic acid (CICZ) from tryptophan directly or from the tryptophan metabolites tryptamine and indolo-3-pyruvic acid, respectively, via the common precursor indolo-3-acetaldehyde. All three pathways produce the same precursor of FICZ.

Microbiota, both on the human skin and in the gut, can convert Trp to several metabolites with little or no affinity for the AHR, as well as pro-ligands and high affinity ligands (reviewed in Murray and Perdew 2017). The lipophilic yeast Malassezia furfur, constituting a part of the normal skin microbiota can produce several Trp derived indoles and indolocarbazoles (Magiatis et al. 2013) (summarized in Janosik et al. 2018). The production of FICZ by M. furfur cultures as reported by Magiatis et al. (2013) was later confirmed by Smirnova et al. (2016), who also detected the oxidation product CICZ in the same cultures (Smirnova et al. 2016). Thus, the processes that lead to the formation of FICZ from Trp involve in addition to microbial enzymes also common mammalian/human enzymes as well as non-enzymatic reactions. Therefore, it is likely that these processes would take place in mammalian and human cells in vivo (Figure 1).

The CYP1A1 enzyme is the most important enzyme for the monohydroxylation of FICZ, the primary step in the mammalian and human metabolism of FICZ. The NADPH-dependent FICZ metabolism is catalyzed with extreme efficiency by human recombinant CYP1A1, with k_cat/K_m of 8.1 × 10⁷ M⁻¹ s⁻¹ approaching the limit of diffusion, and being 50-fold more efficient than the metabolism of the standard substrate 7-ethoxyresorufin (Wincent et al. 2009). CYP1A2 has an overlapping specificity with CYP1A1 and forms the same major metabolites, although with different kinetics. CYP1B1 seems to be preferentially involved in the further metabolism. By applying liquid chromatograhy-mass spectrometry (LC-MS) and NMR spectroscopy two primary metabolites, 2- and 8-hydroxyindolo[3,2-b]carbazole-6-carboxaldehyde (2-OH- and 8-OH-FICZ), respectively were identified. The further CYP-catalyzed metabolism of these two metabolites resulted in three identified dihydroxylated FICZ metabolites (Bergander et al. 2003, 2004). These metabolites are subject to further metabolism by e.g. sulfotransferases and UDP-glucuronosyltransferase, especially sulfoconjugation seems to be an important pathway (Bergander et al. 2004). Detailed studies with six different recombinant human sulfotransferases showed that four of them (SULT1A1, -1A2, -1B1, and -1E1) exhibited high catalytic efficiencies, especially for monohydroxylated FICZ metabolites (Wincent et al. 2009). Since sulfate conjugates are excreted in the urine, the authors analyzed seven human urine samples for sulfate conjugates of the mono hydroxylated metabolites 2- and 8-SO₄-FICZ (the only standards available) of FICZ by LC/MS/MS. The samples contained several FICZ-derived metabolites and 8-SO₄-FICZ was identified in two of the samples, demonstrating the presence of metabolites of the high affinity agonist of AHR in humans (Wincent et al. 2009).

Certain other indole derivatives – including indolo[3,2-b]carbazole (ICZ) (Bjeldanes et al. 1991), 2-(1′H-indole-3′-carbonyl)-thiazole-4-carboxylic acid methyl ester (ITE) (Song et al. 2002), the indigoids indigo and indirubin (Rannug, Bramstedt, et al. 1992; Adachi et al. 2001), as well as dFICZ, and two monohydroxylated FICZ metabolites (2-OH- and 8-OH-FICZ) (Wincent et al. 2009) – as well as lipoxin A4 (Schaldach et al. 1999) and the bacterial virulence factor 1-hydroxyphenazine (Moura-Alves et al. 2014) also bind to the AHR with high affinities.

Fecal bacteria may contribute significantly to the formation of, e.g. indigo, indirubin, indole, 3-indoxyl sulfate, indole-3-acetic acid, indole-3-acetaldehyde, indole-3-aldehyde, indole-3-lactic acid, indole-3-proprionate, indole-3-pyruvate, 3-methyl indole, and tryptamine (Kawajiri et al. 2009; Chung and Gadupudi 2011; Zelante et al. 2013; Jin et al. 2014; Murray and Perdew 2017). Of these, indole-3-acetaldehyde, indole-3-pyruvate, and tryptamine can be converted to FICZ (see above) and possibly other potent AHR ligands (Rannug et al. 1995; Bittinger et al. 2003; Smirnova et al. 2016) (Figure 1). Such formation of FICZ by intestinal bacteria might explain the report by Perdew and Babbs in 1991 that diluted suspensions of rat fecal material can transform Trp into effective activators of the AHR (Perdew and Babbs 1991).

AHR signaling can also be activated by a multitude of exogenous and endogenous factors that do not bind directly to this receptor but that transiently suppress CYP1A1. Such activation by UVB (Goerz et al. 1983, 1996; Wei et al. 1999; Katiyar et al. 2000; Fritsche et al. 2007; Luecke, Wincent, et al. 2010), hyperbaric oxygen and ozone (Okamoto et al. 1993; Afaq et al. 2009), and H₂O₂ (Luecke, Wincent, et al. 2010; Wincent et al. 2012) have been shown in the case of UVB and H₂O₂ to involve enhanced formation of FICZ from Trp (Wei et al. 1999; Fritsche et al. 2007; Wincent et al. 2012; Smirnova et al. 2016), as well as attenuated turnover of FICZ by CYP1A1 (Luecke, Wincent, et al. 2010; Wincent et al. 2012). In addition, activation by, e.g. phorbol-12-myristate-13-acetate (PMA) (Crawford et al. 1997), bacterial endotoxin lipopolysaccharide (LPS) (Baron et al. 1998; Marcus et al. 1998; Lee et al. 2015), and tumor necrosis factor-α (TNFα) (Drozdzik et al. 2014) may also reflect slower metabolic degradation of FICZ, since mediators of inflammation or infection such as LPS, TNFα, and the interleukins IL-6 and IL-1β repress expression and/or activity of drug metabolizing enzymes, including CYP1A1, both in vitro and in vivo (reviewed in Gerbal-Chaloin et al. 2013). Accordingly, induction of CYP1A1 expression in primary human monocytes by FICZ was recently shown to be inhibited by pathogen-associated molecular patterns, PAMPs, that include bacterial lipoproteins, lipopolysaccharides, peptidoglycans, and unmethylated CpG dinucleotides, important regulatory components of innate immune responses (Peres et al. 2017).

As illustrated in Figure 2, various types of agents may attenuate the CYP1A1 enzyme and/or suppress the expression of the CYP1A1 gene and thereby slow down the CYP1A1-mediated degradation of FICZ. Many clinical drugs, metals, industrial chemicals, and phytochemicals that activate the AHR have been shown to inhibit CYP1A1 activity or gene expression (Wincent et al. 2012). Ke et al. (2001) described suppression of cyp1a1 gene expression in mouse Hepa1c1c7 cells by TNFα and LPS in a manner dependent on activation of the nuclear factor κB (NF-κB) and chromatin remodeling in the promoter region of this gene. In accordance with such a mechanism, we have shown that reactive oxygen species (ROS), UVB, H₂O₂, curcumin, and sodium arsenite transiently suppress and subsequently induce CYP1A1 expression in FICZ-treated human cells and that the pleiotropic transcription factors NF-κB and Nrf2, reduced glutathione (GSH), and epigenetic modifications are involved in this process (Luecke, Wincent, et al. 2010; Wincent et al. 2012; Mohammadi-Bardbori et al. 2015, 2016). Thus, activation of the AHR by agents which bind poorly or not at all might be explained by formation of FICZ and/or attenuated metabolic degradation of FICZ. There is also an influence that comes from enhanced efficiency of CYP1A1 transctiption which occurs subsequent to elevated intracellular levels of antioxidants and chromatin remodeling (Wincent et al. 2012; Mohammadi-Bardbori et al. 2015, 2016).

Figure 2. The FICZ/AHR/CYP1A1 feedback loop can be blocked by various types of inhibitors. Abbreviations: AHR, aryl hydrocarbon receptor; AHRE, AHR response element; ARNT, AHR nuclear translocator; CYP1A1, cytochrome P4501A1; FICZ, 6-formylindolo[3,2-b]carbazole; kcat/Km, catalytic efficiency; Kd, dissociation constant; PAMPs, pathogen associated molecular patterns; ROS, reactive oxygen species.

In contrast to the binding of TCDD, which displays species differences, the binding of FICZ to different isoforms of the AHR is evolutionarily well conserved (Jonsson et al. 2009; Laub et al. 2010; Farmahin et al. 2014; Kim et al. 2016). Murine AHR containing Val375 (Val381 in the human AHR) is less responsive to TCDD than the variant with Ala at this position but both exhibit similar sensitivity to FICZ (Cho et al. 2015; Seok et al. 2018; Smith et al. 2018) and FICZ accelerates wound healing similarly in C57BL/6 (Ala375) and DBA/2 (Val375) mice (Morino-Koga et al. 2017). Work with recently developed crystal AHR structures (Schulte et al. 2017; Seok et al. 2017) have not yet revealed important FICZ-interacting residues but in studies using homology models of the AHR PAS domain several residues in the binding pocket that are critical for strong binding of FICZ have been revealed. The somewhat similar binding of TCDD and FICZ involves hydrogen bonding with residues Ser359 and Gln377 in the PASB region of murine AHR (equivalent to Ser365 and Gln383 of the human AHR) (Bisson et al. 2009; Nuti et al. 2014; Hirano et al. 2015; Miyagi et al. 2015).

Reported changes in gene expression in response to FICZ, ITE, and TCDD are very similar (Henry et al. 2010; Farmahin et al. 2016; Prochazkova et al. 2018). Therefore, since they elicit analogous changes in the conformation, nuclear translocation, and binding of the AHR to AHREs, ligands derived from Trp such as FICZ, ICZ, and ITE and possibly many others, most probably activate the same target genes as TCDD.

3. The FICZ/AHR/CYP1A1 feedback loop

As exemplified by the extreme toxicity of TCDD, turning off AHR-mediated signaling when it is no longer required is critical. Already in the 1980s, Nebert and coworkers (Hankinson et al. 1985) proposed that a feedback loop involving an endogenous AHR ligand that is also a substrate for CYP1A1 regulates this signaling. Later, some hydrophobic endogenous AHR-ligand which is metabolized by CYP1A1, 1A2, and 1B1 was proposed to be present in AHR-deficient cells (Weiss et al. 1996; Chang and Puga 1998; Roblin et al. 2004; Chiaro et al. 2007). The FICZ/AHR/CYP1A1 transcriptional-translational feedback loop (Figure 2) described here provides a mechanism that can explain accumulation of FICZ in the cell when AHR-mediated induction of CYP1 enzymes is blocked and how such accumulation can effectively induce CYP1 enzymes and thereby facilitate transient AHR-mediated responses (Wei et al. 1998, 2000; Bergander et al. 2004; Wincent et al. 2009, 2012).

Despite the efficient CYP1 enzyme-mediated metabolic clearance of FICZ, described earlier, when 10 ng FICZ (about 0.5 μg kg⁻¹) was applied percutaneously to the back of one ear of female C57BL/6J mice, the levels of Cyp1a1 mRNA in the liver and adipose tissue rose several hundred-fold, peaking 9 h after application (Wincent et al. 2012). The different transcriptional responses, transient versus prolonged, observed in the liver, and adipose tissues compared to the ear illustrate the different kinetics of responses to FICZ when the compound is present at low or high concentrations. At higher concentrations, FICZ can, similar to ICZ and CICZ, efficiently inhibit the CYP1A1 enzyme and cause prolonged CYP1A1 transcription (Wei et al. 1998; Wincent et al. 2009; Smirnova et al. 2016). A more sustained expression was seen at the site of application in the ear as a result of the higher tissue concentration of FICZ that could inhibit its clearance.

To date, only FICZ has been demonstrated to be an efficient endogenous substrate for CYP1A1 thereby participating in the proposed CYP1A1-dependent feedback control mechanism (Wincent et al. 2009). To the best of our knowledge, no other high-affinity AHR ligands have yet proven to be as excellent CYP1A1 substrate as FICZ. A comparison between ICZ and FICZ showed that in the presence of human CYP1A1 the effects of FICZ disappeared faster than the effects of ICZ indicating a more efficient CYP1A1-dependent clearance of FICZ (Wei et al. 1998).

4. Physiological functions mediated by FICZ

4.1. Self-renewal and differentiation of stem/progenitor cells

Presently, there is considerable interest in the role of the AHR in the maintenance and development of mammalian tissues by regulating the expansion and differentiation of stem cells. A large body of evidence indicates that AHR is involved in the growth of stem cells and, in particular, that its activation may eliminate the pluripotency of and attenuate the potential for self-renewal by various types of stem cells. Thus, the AHR regulates the balance between the quiescence and proliferation of intra-thymic progenitor cells (Laiosa et al. 2003), hematopoietic stem cells (HSCs) (Singh et al. 2009; Boitano et al. 2010; Casado et al. 2011; Gasiewicz et al. 2014; Rentas et al. 2016; Unnisa et al. 2016), pulmonary stem cells (Morales-Hernandez et al. 2017), and neuro-epithelial stem cells (Latchney et al. 2011).

The AHR also appears to play multiple roles in connection with normal embryonic development, with reversible repression of this receptor seeming to be essential for the maintenance of embryonic stem cell (ESC) pluripotency as described by Puga and coworkers. In 2010 these investigators showed that sustained activation with TCDD disrupts the temporal expression of genes whose products are involved in diverse pathways underlying differentiation (Wang et al. 2010) and they later emphasized the importance of the AHR in coordinating embryonic development (Ko et al. 2014, 2016; Ko and Puga 2017). In agreement with such a role, the AHR is not expressed by pluripotent murine ESCs (Ko et al. 2014) and the Cyp1a1 promoter is intensely and constitutively active in a temporal and spatially restricted manner throughout the E7–E14 stages of embryonic development (Campbell et al. 2005). Moreover, in murine ESCs treated with 250 nM FICZ, the AHR and ARNT immunoprecipitate together with CHD4, the major component of the pluripotency regulator NuRD, and with the transcription factor SALL4, suggesting that these are components of the same complex (Gialitakis et al. 2017).

While the underlying mechanism(s) remains unclear, activation of the AHR by FICZ can apparently either attenuate or intensify the expansion of stem cells. As shown by Rentas et al. (2016), expansion of early progenitor murine HSCs was promoted through down-regulation of AHR signaling by the self-renewal-stimulating RNA binding protein Musashi-2 (MSI2) and 200 nM FICZ reversed this effect. Likewise, expansion of human induced pluripotent stem cells (iPSCs) was enhanced by the AHR inhibitor CH223191 and blocked by FICZ (Leung et al. 2018). Administration of 30 µg FICZ kg⁻¹ to 6-week-old female mice, by gavage every second day for 10 days significantly reduced the number of Paneth cells in the small intestine, as well as the depths of the crypts of both the small and large intestines, as well as inhibiting Wnt signaling and lowering the level of active β-catenin (Park et al. 2016). In contrast, AHR activation by FICZ drove an unprecedented expansion of hematopoietic cells into cells of the megakaryocyte, and erythroid linages (Smith et al. 2013). In addition, ex vivo treatment of rat hepatic progenitor cells with 1–100 nM FICZ led to both sustained AHR activation and triggering of cell proliferation (Harrill et al. 2015).

Thus, the FICZ/AHR/CYP1A1 feedback loop, with its extraordinary efficiency in clearing FICZ via enzymes whose expression is regulated by the AHR, may be involved in the regulation of stem cell niches. Interaction of FICZ with SALL proteins and the NuRD complex in ESCs indicates that transient AHR-mediated regulation of stem/progenitor cell homeostasis can be maintained by this loop. Temporary AHR activation by FICZ might potently stimulate the expansion of differentiated cells, with the loop guaranteeing the subsequent silencing of the AHR required for the self-renewal of stem cells.

Various effects of the AHR and FICZ in cancer cells and cancer stem cells (CSCs) have also been described. This receptor is expressed at high levels in a variety of human and rodent tumors and in many human tumor cell lines (reviewed in Wang, Monti, et al. 2017). Furthermore, it is generally believed that sustained AHR activation can be pro-oncogenic, dysregulating key physiological processes and thereby influencing tumor initiation, promotion, and progression, as well as metastasis in a cell-type specific manner (Mitchell and Elferink 2009; Safe et al. 2013; Murray et al. 2014; Wang, Monti, et al. 2017).

Treatment with 500 nM FICZ elevated the activity of the stem cell-associated factor ALDH1 in both breast cancer cells and oral squamous carcinoma cells, while also enhancing migration of the latter in scratch-wound assays, a phenomenon possibly associated with tumor metastasis (Novikov et al. 2016; Stanford, Ramirez-Cardenas, et al. 2016; Stanford, Wang, Zhou, et al. 2017). Such treatment also up-regulated the expression of the genes encoding the stem cell-associated factors ALDH1 and Oct4, as well as of the Snai1, Twist1, Twist2, and Vim genes, which are associated with increased migration and/or invasion by breast cancer cells (Stanford, Wang, Zhou, et al. 2017). In addition, exposure of head and neck squamous carcinoma cell lines to 100 nM FICZ up-regulated expression of the growth factors AREG, EREG, and PDGFA (John et al. 2014). Furthermore, in a validated in vitro model of the movement of breast cancer cells into and out of the lymphatic vasculature, 5 µM FICZ stimulated the synthesis of 12(S)-hydroxyeicosatetraenoic acid (12(S)-HETE) and enhanced tumor breaching of the lymphatic endothelial barrier (Nguyen et al. 2016).

However, recent evidence suggests that the AHR can attenuate the proliferation of CSCs by down-regulating the activity of the stem cell-associated factors Oct4 and ALDH1 (Prud'homme et al. 2010; Bunaciu and Yen 2011). FICZ (100 nM) augmented retinoic acid (RA)-induced cytostatic mechanisms that drive the differentiation of human myeloblastic leukemia HL-60 cells (Bunaciu and Yen 2013; Bunaciu et al. 2015). Furthermore, in B16F10 murine melanoma cells exposed to 5 nM FICZ the level of Aldh1a1 mRNA was reduced and an inverse correlation between the levels of AHR and ALDH1 was observed (Contador-Troca et al. 2015). In addition, Morales-Hernandez et al. (2016) found that 10 nM FICZ stimulated the differentiation of human embryonic teratocarcinoma (NTERA) cells, possibly by inhibiting the expression of the stem cell-associated POU5F1 (Oct4) and NANOG genes.

In yet another study on human cancer cells, AHR ligands were shown to inhibit cell expansion. When Cheng et al. (2015) added 50 nM FICZ or 10 µM of the AHR ligand ITE to human U87 glioblastoma cells, both repressed expression of the POU5F1 gene. In addition, ITE reduced the size of tumors obtained by injecting U87 cells into nude mice but, unfortunately, similar experiments were not performed with FICZ (Cheng et al. 2015). FICZ (100 nM) down-regulated cyclin D1, and up-regulated p27 and caused cell cycle and proliferation arrest in colon cancer cells (Yin et al. 2016). Moreover, in 2013, Shin and coworkers reported that in mice implanted subcutaneously with murine T-cell lymphoma or B16 melanoma cells, three intraperitoneal (i.p.) injections of 3 µg FICZ (approximately 150 μg kg⁻¹) at two day intervals starting on the day of implantation inhibited tumor growth in a manner dependent on the AHR and natural killer (NK) cells (Shin et al. 2013). In another study, 5 nM FICZ raised the level of E-cadherin in murine NMuMG epithelial cells and repressed their migration in a wound healing assay suggesting that FICZ can modulate the epithelial to mesenchymal transition (EMT) (Rico-Leo et al. 2013).

In conclusion, investigations in which FICZ has been employed to elucidate mechanisms of cancer cell growth have reported different outcomes. Therefore, endogenously activated AHR can be postulated to play a dual role in tumorigenesis: (1) stimulation of tumor cell migration and invasiveness by up-regulating the expression of genes associated with growth and migration and (2) inhibition of cancer cell proliferation through inhibition of the expression of factors associated with CSC, repression of EMT and epithelial cell migration, and promotion of the control of tumor development by immune cells.

4.2. Differentiation of immune cells and immune responses

The levels of the AHR expressed by cells of the immune system are particularly high and two high-profile publications in 2008 describing an important role for AHR in a murine model of multiple sclerosis have intensified focus on the involvement of this receptor in infection and autoimmunity (Quintana et al. 2008; Veldhoen et al. 2008). With respect to the adaptive immune system, AHR activation by FICZ was reported in these two studies to enhance pro-inflammatory responses only by stimulating the differentiation of CD4⁺ T helper 17 cells (Th17) that express the cytokines IL-17, IL-17F, and IL-22 and blocking the formation of regulatory T cells (Tregs) that express the transcriptional regulator forkhead box P3 (Foxp3) (Table 1). In addition, oral administration of 1 µg TCDD to each mouse induced Tregs and markedly reduced disease severity (Quintana et al. 2008). Quintana et al. (2008) proposed that TCDD diverts T cells toward regulator T cells and FICZ generates effector T cells. Support for differential effects of FICZ and TCDD was provided in several subsequent studies (Mezrich et al. 2010; Singh et al. 2011; Schulz et al. 2012; Singh et al. 2016). TCDD also suppressed sensitization of mice to peanuts, at least in part by elevating the percentage of Treg cells, whereas FICZ did not increase this percentage or influence the cytokine production induced by peanut extract (Schulz et al. 2012) (Table 1). However, both FICZ and TCDD enhanced the severity of collagen-induced arthritis in transgenic, shared epitope (SE)-positive mice and elevated the numbers of cells expressing IL-17 both in popliteal lymph nodes and the spleen, without influencing the abundance of FoxP3 expressing cells (Fu et al. 2018) (Table 1).

Table 1. Effects by FICZ on disease severity in murine models of human diseases.

Disease	Murine model	FICZ treatment	Biological responses	Disease outcome	Reference
Allergy	Passive systemic anaphylaxis induced with 2,4-dinitrophenyl (DNP)-specific IgE in mice	One i.v. administration of 100 µg FICZ kg^−1 during the challenge phase	↑ Plasma histamine	Increased severity	(Sibilano et al. 2012)
		Two i.v. administrations of 100 µg FICZ kg^−1 3 and 16 h before antigen challenge	↓ Plasma histamine	Less severity
	Passive cutaneous anaphylaxis induced with ovalbumin (OVA)-specific IgE in mice	One intra-dermal administration of 0.03 or 0.3 µg FICZ kg^−1 during the sensitization phase	↑ Vascular leakage	Increased severity	(Zhou et al. 2013)
	Peanut allergy in mice	Multiple i.p. injections of 50 or 500 µg FICZ kg^−1 before and during sensitization to antigen	No effect	No difference	(Schulz et al. 2012)
	OVA-induced allergic asthma in mice	One i.p. injection of 3 or 30 µg FICZ kg^−1 during the sensitization phase	↓ IL-4, IL-5, IgE	Less severity	(Jeong et al. 2012)
Atopic dermatitis	Mite-induced topic dermatitis in mice	Topical treatment with 100 mg of an ointment containing 1 µM FICZ three times a week for 2 weeks	↓ IL-22, IFNγ	Less severity	(Kiyomatsu-Oda et al. 2018)
Cancer	Tumor growth in wild type mice injected with murine lymphoma RMA-S cells	Three i.p. injections of 150 μg FICZ kg^−1	↑ NK cells, reduced tumor growth	Less severity	(Shin et al. 2013)
	Growth of tumors in immune deficient recombinase-activating gene 1 (RAG1) knock out mice injected with B16 melanoma cells	Three i.p. injections of 150 μg FICZ kg^−1	↓ Tumor cells and lung metastases	Less severity
Diabetes	Mouse streptozotocin (STZ) model	Five i.p. injections of 5.7 µg FICZ kg^−1	↑ Foxp3	Less severity	(Abu-Rezq and Millar 2013)
Graft-versus-host disease (GVHD)	A parent-into-F1 model of GVHD	Four i.p. injections of 10 mg FICZ kg^−1	↑ Foxp3	Less severity	(Ehrlich et al. 2018)
Inflammatory bowel disease (IBD)	Dextran sulfate sodium (DSS)- or trinitrobenzene sulfonic acid (TNBS)-induced colitis in mice	One i.p. injection of 50 µg FICZ kg^−1	↑ IL-22 ↓ IL-17, IFNγ, TNFα	Less severity	(Monteleone et al. 2011)
	T cell transfer to immune deficient RAG1 knock out mice	Five i.p. injections of 50 µg FICZ kg^−1	↑ IL-22 ↓ IFNγ	Less severity	(Monteleone, Marafini, et al. 2016)
	TNBS-induced intestinal fibrosis in mice	Seven i.p. injections of 50 µg FICZ kg^−1	↓ IL-17, TNFα	Less severity	(Monteleone, Zorzi, et al. 2016)
	DSS-induced colitis in mice	Five i.p. injections of 50 µg FICZ kg^−1	↓ IL-7	Less severity	(Ji et al. 2015)
	DSS-induced colitis in mice	Five i.p. injections of 50 µg FICZ kg^−1	↑ IL-10 ↓ IFNγ	Less severity	(Chen et al. 2017)
	DSS-induced colitis in mice deficient in the caspase recruitment domain family member 9 (CARD9) protein	One i.p. injection of 50 µg FICZ kg^−1	↑ IL-22	Less severity	(Lamas et al. 2016)
Multiple sclerosis (MS)	Experimental autoimmune encephalomyelitis (EAE) in mice	One subcutaneous administration of 30 µg FICZ kg^−1	↑ IL-17	Increased severity	(Veldhoen et al. 2008)
		One i.p. injection of 50 µg FICZ kg^−1	↑ IL-17	Increased severity	(Quintana et al. 2008)
		One i.p. injection of 10 mg FICZ kg^−1	↓ IL-17 ↑ IL-10	Less severity	(Duarte et al. 2013)
Peritonitis	Alum-induced peritonitis in mice	One i.p. injection of 200 µg FICZ kg^−1	↓ IL-1β, TNFα, IL-6	Less severity	(Huai et al. 2014)
Psoriasis	Mouse imiquimod (IMQ) model	Daily i.p. injections of 100 µg FICZ kg^−1 over the course of IMQ treatments	↓ IL-22, IL-17	Less severity	(Di Meglio et al. 2014)
		Daily topical applications of 10 ng of FICZ over the course of the IMQ treatment	↓ IL-17	Less severity	(Smith, Jayawickreme, et al. 2017)
Rheumatoid arthritis	Collagen-induced arthritis in transgenic mice carrying human shared epitope-coding alleles	One weekly i.p. injection of 28 µg FICZ kg^−1 for the entire disease course	↑ IL-17, osteoclast infiltration	Increased severity	(Fu et al. 2018)

With immune cells in vitro, FICZ stimulated the production of IL-22 in human cells (Trifari et al. 2009; Ramirez et al. 2010; Dorgham et al. 2015; Effner et al. 2017) and both IL-17 and IL-22 in mouse cells (Rutz et al. 2011; Pauly et al. 2012; Hayes et al. 2014; Schiering et al. 2017) and the inability of FICZ to promote the formation of Foxp3-expressing Tregs under the conditions employed was noted in some studies (Vogel et al. 2008; Mezrich et al. 2010; Singh et al. 2011; Pauly et al. 2012; Duarte et al. 2013). Under conditions that induce Th17 cells, both FICZ and TCDD up-regulated the expression of a microRNA-132/212 cluster that enhances the Th17 differentiation of naïve murine T cells (Nakahama et al. 2013). Additionally, in their assessment of the responses of Th17 cells to treatment with FICZ, TCDD, ITE, β-naphthoflavone (βNF) or 3-methylcholanthrene (3MC), Duarte et al. (2013) found no indications of any intrinsic difference in the mode of action of these ligands and, moreover, a high dose of FICZ administered systemically proved to suppress experimental autoimmune encephalomyelitis (EAE) substantially. Thus, chronic activation of the AHR can prevent Th17-mediated inflammation in the central nervous system and the different regulation of Treg and Th17 cell differentiation by FICZ and TCDD proposed by Quintana and coworkers in 2008 was refuted (Table 1).

Confirmation that the dose and duration of AHR activation by high-affinity ligands are the primary drivers of T cell differentiation was provided by a recent examination of allogenic responses in mice (Ehrlich et al. 2018) (Table 1).

Activation of anti-inflammatory responses and tolerogenic cell phenotypes by FICZ, as well as TCDD, is now firmly established and several laboratories have detected induction of Foxp3-expressing Tregs by FICZ (Kimura et al. 2008; Abu-Rezq and Millar 2013; Mohinta et al. 2015; Jurado-Manzano et al. 2017; Ye et al. 2017; Ehrlich et al. 2018). Such induction attenuated disease severity in murine models of diabetes and graft-versus host disease (Abu-Rezq and Millar 2013; Ehrlich et al. 2018) (Table 1). Treatment of murine CD4⁺ T cells originally expressing the proinflammatory IL-17 with 100 nM FICZ in the presence of TGFβ caused these cells to acquire an anti-inflammatory phenotype (Gagliani et al. 2015). Furthermore, when 100 nM FICZ or 100 nM TCDD was added in the presence of IL-27 and TGFβ to murine CD4⁺ T cells, these cells acquired an anti-inflammatory Tr1 phenotype, with enhanced secretion of the anti-inflammatory IL-10 (Apetoh et al. 2010). In addition, expression of IL-10 by naïve human T cells was enhanced by differentiation in the presence of FICZ (Gandhi et al. 2010).

Wheeler et al. (2014) observed no influence of FICZ on the immune response of wild-type mice with functional CYP1A1 to influenza virus A. However, when FICZ was administered to CYP1A1-defective mice via micro-osmotic pumps (10 µg FICZ kg⁻¹ h⁻¹), like TCDD, this compound diminished the virus-specific CD8⁺ T cell response (Wheeler et al. 2014). In other mice infected with the influenza A virus, ten daily i.p. administrations of 100 μg FICZ kg⁻¹ attenuated virus-specific immunoglobulin M (IgM) levels and elevated the proportion of Treg cells, without reducing CD8⁺ T cell responses (Boule et al. 2018). In other studies, treatment of mice with 2 µg FICZ (about 100 µg kg⁻¹) suppressed immune responses to the influenza A, vesicular stomatitis, and Sendai viruses (Yamada et al. 2016), whereas treatment of mice infected with ocular herpes simplex virus intraperitoneally with 1 µg FICZ each (about 50 µg kg⁻¹) once every day for eleven consecutive days had no effect on the occurrence of ocular lesions (Veiga-Parga et al. 2011).

There are recent indications that AHR activation by FICZ may help reduce inflammation via activation of the Trp-degrading enzyme indoleamine 2,3-dioxidase (IDO) in the dendritic cells (DCs) present in various epithelial tissues. Vogel et al. (2008) found that both TCDD and FICZ activated IDO in human DCs derived from the monocytic U937 cell line, whereas only TCDD (10 nM), but not FICZ (100 nM) stimulated the formation of Foxp3-expressing Tregs. In several subsequent investigations, FICZ was also shown to elicit phenotypical alterations in DCs similar to those caused by TCDD, including attenuation of the responses of Th17 cells (Bankoti et al. 2010; Simones and Shepherd 2011; Kado et al. 2017; Thatcher et al. 2016). Expression of IDO1 was elevated in murine bone marrow DCs (BMDCs) by exposure to 100 nM FICZ (Pauly et al. 2012) and in Langerhans cells obtained from hematopoietic stem cells derived from the umbilical cord by treatment with 3.5 µM (1 µg/ml) FICZ (Koch et al. 2017).

Exposure of DCs derived from human monocytes to 50 nM or 300 nM FICZ elevated the expression of the IDO enzyme, induced a tolerogenic phenotype by promoting Treg-like cell differentiation and suppressed the activation of naïve T lymphocytes (Jurado-Manzano et al. 2017). Such activation of IDO does not appear to be AHR ligand-specific, since TCDD, kynurenine, ITE, and the synthetic compound SU5416 all had this same effect (Simones and Shepherd 2011; Mezrich et al. 2012; Bessede et al. 2014). The absence of IDO-activation by FICZ in certain cases (Ilchmann et al. 2012; Mezrich et al. 2012) may reflect its rapid CYP1A1-mediated degradation.

Therefore, taking rapid metabolic degradation of FICZ into account, there appears to be no intrinsic difference in the effects of FICZ and TCDD on T cell differentiation and T cell-mediated adaptive immune responses.

A potential role for the AHR and its ligands in maintaining activation of B cells by antigen was indicated by the findings by Villa et al. (2017) that treating anti-IgM- and IL-4-stimulated B cells with 250 nM FICZ potently elevated the expression of the Ccno gene the product of which is involved in controlling the cell cycle and, thereby, probably B cell proliferation.

Mast cells (MCs), which regulate murine and human mucosal innate immunity and allergic and anaphylactic responses by releasing pro-inflammatory mediators (including histamine, mast cell proteases, and pro-inflammatory cytokines) were recently found to express the AHR constitutively (Maaetoft-Udsen et al. 2012; Sibilano et al. 2012) as well as to be highly responsive to FICZ (Sibilano et al. 2012; Zhou et al. 2013). Another study did, however, not confirm any role of the AHR in mast cell homeostasis (Pilz et al. 2016). In their studies on the degranulation (i.e. release of the granule-associated enzyme β-hexosaminase (Hex)) of mast cells derived from murine bone marrow (BMMCs) in response to a challenge with dinitrophenyl-specific IgE, Sibilano et al. (2012) found that 300 nM FICZ enhanced the levels of cyclic adenosine monophosphate (cAMP) and ROS in the absence of this antigen and promoted degranulation when added together with antigen. However, treatment with 300 nM FICZ both 3 and 16 h prior to the antigen challenge lowered the cAMP/ROS response and impaired degranulation. Thus, depending on the timing, administration of FICZ to mice could either attenuate or enhance the severity of passive systemic anaphylactic responses (Sibilano et al. 2012) (Table 1).

Importantly, Zhou et al. (2013) described the critical role of AHR signaling in maintaining MC homeostasis. When FICZ was added to BMMCs during sensitization with anti-ovalbumin (OVA) IgE, inverted U-shaped dose–response curves were obtained. FICZ alone was unable to promote release of Hex, but when added together with IgE low (0.1–1 nM), but not higher concentrations (10–100 nM) of this compound increased both degranulation and cytokine responses when challenged with antigen. When sensitized in the presence of 1 nM FICZ, the MCs responded with increased intracellular levels of Ca²⁺ and ROS and the higher levels of ROS oxidized the tyrosine phosphatase SHP-2 protein, thereby enhancing MC signaling. Likewise, these same investigators showed that in a murine model of IgE-mediated passive cutaneous anaphylaxis, a single intra-dermal administration of FICZ (0.03 or 0.3 µg FICZ kg⁻¹) together with IgE during the sensitization phase aggravated responses to a subsequent OVA challenge (Zhou et al. 2013) (Table 1). In other cases, IgE-mediated degranulation, and changes in the levels of ROS/Ca²⁺, leukotriene C4 (LTC4), IL-13, and PKR-like ER kinase (PERK) signaling in BMMCs were all augmented by 1 nM FICZ (Kawasaki et al. 2014; Wang, Zhou, et al. 2017). Alltogether, these observations indicate that endogenous activation of the AHR plays a key role in the homeostasis of MCs, a role that depends on the extent and timing of activation.

In the case of primary murine, microglial cells co-treatment with 20 nM FICZ influenced pro-inflammatory responses to LPS, suggesting that the AHR could be involved in regulating anti-inflammatory processes in the CNS (Lee et al. 2015). Multiple investigations have observed a potential association between environmental AHR ligands and an increased risk for developing ALS. In this context Ash et al. (2017) reported that AHR ligands elevated the amount of aggregated TAR DNA-Binding Protein-43 (TDP-43) in human cell lines, in iPSCs, and even in the murine brain. FICZ (100 nM) enhanced the level of TDP-43 protein in iPSCs derived from a patient with ALS that had differentiated into neurons. Moreover, in human neuroblastoma M17 cells 500 nM FICZ raised TDP-43 levels and stabilized this protein; while in H4 neuroglioma cells this same concentration increased the expression of the TARDBP, SOD1, PON2, and C9ORF72 genes associated with ALS (Ash et al. 2017). Accordingly, environmental chemicals that lead to sustained activation of AHR signaling or block metabolic degradation of FICZ might be significant risk factors for ALS.

In conclusion, AHR activation by FICZ can promote the development of Th17 cells, which may lead to tissue inflammation and autoimmunity; but FICZ can also stimulate immunosuppressive activity by expanding the population of regulatory T cells. Similarly, FICZ can stimulate or inhibit cytokine production and the maturation and homeostasis of MCs in vitro, as well as anaphylactic responses in vivo, depending on the dose and timing of exposure.

5. Effects of FICZ on immune barriers

Dysregulation of immune barriers is associated with many common diseases and the AHR is expressed at high levels in cells of these barriers, which are routinely exposed to environmental ligands (Esser et al. 2009; Esser and Rannug 2015). In response to secretion of cytokine IL-22 by hematopoietic cells, probably in combination with IL-17, the epithelial cells of the gut, lung, and skin produce inflammatory cytokines, chemokines that recruit neutrophils (e.g. CXCL1 and CXCL5), anti-microbial proteins and mucins to maintain homeostasis, fight invading pathogens, and accelerate regeneration of the barrier whenever required (Sonnenberg et al. 2011; Stockinger and Omenetti 2017). Thus, IL-22 and IL-17 appear to have both pro-inflammatory and protective functions.

In Figure 3, a model is proposed that explains how FICZ through activation of the IL-22–IL-22 receptor (IL-22R) pathway in immune barriers can reduce the severity in animal models of several long-term diseases when administered systemically. Under such conditions, FICZ escapes the CYP1A1-dependent clearance in the non-hematopoietic cells in the epithelial cell lining. This model can also explain how FICZ in the presence of CYP1A1 inhibitors can act pro-inflammatory and provide resistance to pathogenic bacteria.

Figure 3. Activation of the interleukin 22 (IL-22)–IL-22 receptor (IL-22R) pathway by FICZ can regulate immunity, inflammation and tissue homeostasis at immune barriers of, e.g. the skin, lung, and intestine. (1). FICZ formed externally from tryptophan, e.g. by enzyme-catalyzed conversions in microorganisms and by UV-light exposure or taken up from the diet, is efficiently metabolized and cleared in epithelial cells at barrier surfaces after AHR-mediated upregulation of the CYP1A1 enzyme. (2). FICZ formed from tryptophan through host metabolism or added systemically, e.g. via intraperitoneal injection, activates AHR-mediated formation of IL-22 in various immune cells. IL-22 regulates immune responses via ligation of the IL-22R on epithelial cells leading to release of tissue specific cytokines, chemokines and antimicrobial agents. (3). FICZ formed externally may also activate the IL-22–IL-22R pathway if CYP1A1 inhibitors are present that inhibit the clearance of FICZ. For abbreviations, see Figure 2.

5.1. Gut immunity

The intestinal epithelium constitutes a relatively impermeable physical and immunological barrier against the contents in the lumen, while the intestinal epithelial cells (IECs) take up nutrients and participate in the inflammatory and immune responses when activated by IL-22 and IL-17 (Sonnenberg et al. 2011; Conti and Gaffen 2015). The general sources of IL-22 and IL-17 in the gut include innate lymphoid cells (ILCs), Th17 and Th22 cells, and various types of intraepithelial lymphocytes (IELs) (e.g. T cell receptor (TCR) γδ cells). One dominant endogenous source of IL-22 appears to be the heterogeneous population of ILCs that includes NK cells and lymphoid tissue-inducer (LTi) cells (Sonnenberg et al. 2011).

AHR-deficient mice have strongly reduced numbers of intestinal IELs and ILCs and are therefore unable to mount an IL-22 response, succumbing to the intestinal pathogen Citrobacter rodentium (Kiss et al. 2011; Lee et al. 2011; Li et al. 2011; Qiu et al. 2012). Qiu et al. (2012) found that in wild-type mice, FICZ increased the production of IL-22 by the ILCs. Moreover, i.p. injection of 12-day-old mice with 0.5 µg FICZ (about 50 µg kg⁻¹) for 6 days promoted accumulation of ILCs expressing the transcription factor RORγt required for production of IL-22 (Qiu et al. 2012). Daily i.p. injections of 100 µg kg⁻¹ FICZ to adult mice dramatically reduced their mortality following infection with the intestinal pathogen Listeria monocytogenes (Kimura et al. 2014). Furthermore, in an earlier study on adult mice, 1 µg FICZ (50 µg kg⁻¹) injected i.p. together with heat-killed Mycobacterium tuberculosis markedly increased the subset of IL-17-expressing γδ T cells in the peritoneum, as well as their production of IL-22 (Martin et al. 2009). Thus, FICZ appears to enhance gut protection against intestinal pathogens.

Intestinal epithelial cells can be viewed as gatekeepers that regulate the passage of CYP1A1-degradable AHR ligands from the lumen into the blood (Figure 3). The role of AHR/ARNT in metabolic degradation of AHR agonists by the gut epithelium was first demonstrated in 2007 by Gonzalez and coworkers, who disrupted the Arnt gene in mice predominantly in this epithelium, and found loss of intestinal Cyp1a1 expression, in addition to an endogenous factor that caused systemic expression of Cyp1a1 (Ito et al. 2007). This factor only activated the AHR when the mice were fed natural diets containing phytochemicals or synthetic diets supplemented with indole-3-carbinol (I3C), derived from the glucosinolate glucobrassicin.

Recently, Stockinger and coworkers demonstrated that CYP1A1 in the IECs plays a vital role in intestinal immunity by controlling endogenous availability of FICZ (Schiering et al. 2017). Employing their mouse strain that expressed Cyp1a1 constitutively, these investigators found that Th17 cells differentiated in vitro metabolized FICZ rapidly, and compared to wild-type Th17 cells, showed a lower production of IL-22 in response to 0.01 nM FICZ. Furthermore, overexpression of Cyp1a1 specifically by IECs showed that these cells performed an important CYP1A1-mediated feed-back function involved in regulating intestinal immunity. Actually, the metabolic clearance of natural AHR ligands in mice that overexpress Cyp1a1 in the gut epithelium led to a pseudo AHR-deficient state. Upon infection with Citrobacter rodentium, these animals exhibited markedly reduced numbers of group 3 ILCs and Th17 cells and succumbed rapidly, with augmented intestinal pathology. Conversely, mice deficient in CYP1A1, CYP1A2, and CYP1B1 (CYP1 triple-KO mice), as well as wild-type mice fed diets free of phytochemicals, but supplemented with I3C demonstrated enhanced IL-22 production, as well as less histopathology upon bacterial infection. Likewise, a follow-up in vitro study with Th17 cells derived from CYP1 triple-KO mice or wild-type murine Th17 cells exposed to the polycyclic aromatic hydrocarbons (PAHs) fluoranthene, pyrene, or phenanthrene that competitively inhibit CYP1A1 (Shimada and Guengerich 2006) led to the conclusion that IL-22 production by Th17 cells can be enhanced profoundly by inhibiting CYP1 activity (Schiering et al. 2018). Both of these investigations indicate strongly that inhibitors of CYP1A1 activity can promote AHR-mediated protection against pathogens by elevating the endogenous level of FICZ (Schiering et al. 2017, 2018).

In addition, AHR acted as a negative regulator of the inflammasome, another component of host defenses against microbial pathogens, by inhibiting induction of the pattern recognition receptor NLRP3 by LPS (Huai et al. 2014). FICZ (100 nM) also suppressed recruitment of inflammatory cells and consequently peritonitis upon exposure to alum (Table 1).

The composition of the murine commensal microbiota influences susceptibility to gastrointestinal infections and induced colitis (Ivanov et al. 2008; Gao et al. 2018). Recently, specific components of this microbiota were found to regulate gut immunity by promoting production of AHR ligands from Trp and thereby transcription of genes encoding anti-inflammatory cytokines, resulting in protection against intestinal damage induced by dextran sulfate sodium (DSS) (Zelante et al. 2013; Lamas et al. 2016; Hou et al. 2018). Lamas et al. (2016) showed that mice with dysbiotic microbiota due to their lack of the caspase recruitment domain 9 (CARD9) produced lower levels of endogenous AHR agonists and recovered more poorly from DSS-induced colitis (Table 1). When 1 μg FICZ (50 µg kg⁻¹) was injected i.p. one day after DSS administration, the severity of colitis in these animals was reduced significantly. Moreover, the effects of defects in the colonic expression of IL-22 and in genes coding for antimicrobial proteins in Card9 KO mice could be reversed by FICZ (Lamas et al. 2016).

Following the first report by Monteleone et al. (2011) that FICZ can alleviate DSS-induced colitis in a murine model of human inflammatory bowel disease (IBD), several studies involving single or repeated i.p. injections of 1 µg FICZ per mouse (50 µg kg⁻¹) have confirmed that this ligand can trigger regulatory signals in the gut that promote production of IL-22 and suppress colitis (Ji et al. 2015; Monteleone, Marafini, et al. 2016; Monteleone, Zorzi, et al. 2016; Chen et al. 2017; Wang et al. 2018; Yu et al. 2018). Moreover, upon exposure to 2–200 nM FICZ, a dose-dependent increase in IL-22 with a reduction in the level of interferon-γ (IFNγ) occurred in lamina propria mononuclear cells from both patients with Crohn’s disease and healthy individuals (Monteleone et al. 2011). As mentioned earlier, IL-22 is important for the regeneration of intestinal stem cells and, consequently, in several cases FICZ restored the crypt architecture and lengthened the colon of mice treated with DSS.

In addition to IL-22, the anti-inflammatory cytokine IL-10 plays a key role in regulating autoimmune diseases and FICZ doubled the number of Tr1 cells producing this cytokine, and augmented its secretion (Apetoh et al. 2010). In addition, FICZ stimulated the homing of FoxP3-expressing Tregs in the large intestine of mice (Ye et al. 2017), and up-regulated epithelial expression of the IL-10 receptor (IL-10R) (Lanis et al. 2017), which helps protect the epithelial integrity. Thus, in addition to stimulating formation of IL-22, FICZ might also suppress colitis by regulating the expression of the IL-10R.

Altogether, these data reveal that control of FICZ metabolism in epithelial cells by the FICZ/AHR/CYP1A1 feedback loop can explain the effects of this compound on gut immunity.

5.2. Pulmonary and eye immunity

Experiments by Thatcher et al. (2016) indicated that endogenous AHR ligands might be involved in the physiological regulation of Th2-mediated immunity in the lung via a dendritic cell-dependent mechanism. They found that after murine pulmonary dendritic cells were incubated overnight with LPS and OVA, addition of 100 or 200 nM FICZ markedly inhibited the proliferation of T cells (Thatcher et al. 2016). FICZ was also suggested to exert anti-asthmatic effects on the basis of the observation that when injected i.p. to a murine model of allergic asthma during sensitization to OVA, 3 or 30 μg of FICZ kg⁻¹ reduced pulmonary eosinophilia and expression of Th2 cytokines (Jeong et al. 2012) (Table 1). In another examination of allergic asthma in mice, 30 nM FICZ activated Jagged 1 (Jag1), a component of the pro-inflammatory Notch pathway, in BMDCs (Xia et al. 2015).

In thyroid eye disease (TED) mast cells, together with infiltrating T cells, cause chronic orbital inflammation and promote production of myofibroblasts. In orbital fibroblasts from patients with TED, several of the detrimental effects, such as remodelning of orbital fibroblasts (including TGFβ-induced proliferation), production of collagen, contraction of myofibroblasts, and formation of α-smooth muscle actin (α-SMA) filament were blocked by 100 nM–10 µM FICZ. FICZ (100 nM) prevented TGFβ-induced profibrotic Wnt signaling by inhibiting phosphorylation of glycogen synthase kinase 3β (GSK-3β), destabilizing β-catenin and blocking transcription of genes involved in myofibroblast production (Woeller et al. 2016).

5.3. Skin immunity

In addition to being formed from Trp upon exposure of human keratinocytes to light and oxidation by H₂O₂ in the skin of patients with vitiligo Schallreuter et al. 2012, FICZ can be synthesized by the microbial flora on the skin (Magiatis et al. 2013). Malassezia yeasts, a normal component of this flora, interact with melanocytes, fibroblasts, keratinocytes, and dendritic cells to modulate the immune responses against them (Ashbee 2006). In extracts of skin scale from patients with Malassezia-associated seborrheic dermatitis, the capacity to activate AHR was 10- to 1000-fold higher than in healthy controls, a difference ascribed to FICZ and other highly potent AHR agonists present in these extracts (Magiatis et al. 2013).

The key roles played by dynamic AHR signaling in skin immunity have been emphasized by Stockinger and coworkers (Di Meglio et al. 2014). They found that the AHR, primarily in keratinocytes, prevented excessive skin inflammation by controlling the expression of certain chemokine ligands involved in neutrophil attraction (Csf2, Csf3, Cxcl1, Cxcl5) and of antimicrobial peptides (S100a7a, S100a8) in response to inflammatory stimuli. When AHR signaling in full-thickness biopsies from the lesional skin of patients with psoriasis was activated by 250 nM FICZ, 29 genes belonging to the so-called psoriasis transcriptome were down-regulated. Furthermore, activation of the AHR pathway in a murine model of psoriasis by i.p. injection of 100 µg FICZ kg⁻¹ for 6 days over the course of treatment with imiquimod (IMQ) ameliorated skin pathology (Table 1) and lowered the expression of several mediators of inflammation, including Csf2, Csf3, Cxcl5, S100a7a, and S100a8, as well as IL-17 and IL-22 (Di Meglio et al. 2014). A later murine study confirmed that FICZ decreased IL-17 expression and lessened the severity of psoriasis (Smith, Jayawickreme, et al. 2017) (Table 1). Moreover, low concentrations of FICZ attenuated the expression of the chemokine ligand 5 (CCL5) by HaCaT cells (Morino-Koga et al. 2013) and augmented the expression of filaggrin, an epithelial barrier protein, by normal human keratinocytes in vitro (Tsuji et al. 2017; Kiyomatsu-Oda et al. 2018). However, in primary mouse keratinocytes treated simultaneously with IL-1β and 100 nM FICZ, expression of the Cxcl5 chemokine was enhanced (Smith, Boyer, et al. 2017).

Similarly, FICZ and AHR may be involved in the etiology of cutaneous systemic lupus erythematosus (SLE) (Dorgham et al. 2015) and atopic dermatitis (AD) (Kiyomatsu-Oda et al. 2018). In a study on chronic eczema induced by mites, application of 2.5 μg FICZ kg⁻¹ reduced the AD-like inflammation of the skin and transepidermal water loss, and as in the model of psoriasis involving induction by IMQ, the expression of IL-22 was elevated (Kiyomatsu-Oda et al. 2018) (Table 1).

More evidence for an important physiological role for FICZ and AHR-dependent expression of IL-22 in the skin has been obtained by Cibrian et al. (2016), who noted that uptake of Trp and intracellular accumulation of FICZ in skin γδ T cells is regulated by the activation marker CD69 in combination with the aromatic-amino-acid-transporter complex LAT1-CD98. These results revealed the importance of Trp uptake for AHR-dependent secretion of IL-22 by γδ T cells during the development of psoriasis (Cibrian et al. 2016).

6. The toxicity of FICZ

The median lethal concentrations (LC₅₀) of FICZ, in the case of fibroblast-like monkey kidney COS-7 cells and primary chicken hepatocytes, were estimated to be 9.4 and 14 µM, respectively (Farmahin et al. 2014, 2016).

Treatment with 0.1–1 µM FICZ reduced the proliferation of several human tumor cell lines (Yin et al. 2016; Mohammadi-Bardbori et al. 2017). Release of Ca²⁺, depletion of GSH, an increase in the level of intracellular ROS, impairment of mitochondrial membrane integrity, stimulation of caspase 3 activity, and release of cytochrome c all occurred in HepG2 cells and mouse hepatoma (Hepa-1) cells exposed to 1 µM FICZ for 24 h (Mohammadi-Bardbori et al. 2017). Increase in ROS and elevated apoptosis signaling were also obtained when 1 nM FICZ was added together with 50 nM of 3′-methoxy-4′-nitroflavone (MNF), which inhibits CYP1A1. Interestingly, Mohammadi-Bardbori and coworkers also showed that the growth of HepG2 and Hepa-1 cells was stimulated by 0.01 nM FICZ, the level ordinarily present in cell culture medium (Oberg et al. 2005; Wincent et al. 2012). Exposure of BMMCs to as little as 1 nM FICZ led to accumulation of Ca²⁺ in the cytosol and higher levels of intracellular ROS that peaked 2 h after treatment, along with a decline in intracellular GSH (Wang, Zhou, et al. 2017). In addition, there was a rapid but transient decrease in mitochondrial ATP production and reduction of the mitochondrial membrane potential, as well as up-regulation of Nrf2, a strong inducer of antioxidant genes. Altogether, these thought-provoking findings indicate that high levels of FICZ can exert ROS-dependent toxicity, whereas under certain conditions low doses of FICZ can transiently elevate local levels of second messengers such as ROS/Ca²⁺, thereby promoting cellular adaptation, survival, and proliferation.

Recently, FICZ has proven to be potently embryotoxic toward fish and birds. FICZ is a strong agonist of fish AHR2, inducing CYP1A and CYP1B1 in zebrafish embryos approximately fivefold more potently than the polybrominated biphenyl PCB126 (Jonsson et al. 2009). Wincent et al. (2016) demonstrated that FICZ dramatically increased mortality and the incidence and severity of pericardial edema and circulatory failure, while reducing hatching and blocking inflation of the swim bladder in zebrafish embryos. Importantly, these toxic manifestations were only seen when CYP1A was inhibited by morpholino antisense oligonucleotides or α-naphthoflavone (αNF), this latter compound augmenting the mortality caused by FICZ even at concentrations as low as 5 pM (Wincent et al. 2016). Proper development of zebrafish embryos appears to require crosstalk between the AHR and the Wnt/β-catenin pathway (Yin et al. 2011; Yoshioka et al. 2011) and this latter pathway was a key target for FICZ during early zebrafish development (Wincent et al. 2015).

FICZ was recently also tested for embryo toxicity in chicken and Japanese quail embryos (Jonsson et al. 2016). Injection of FICZ into the air sac of chicken embryos was found to cause dose-dependent mortality with LD₅₀ values below 20 μg kg⁻¹. Notably, when 4 μg FICZ kg⁻¹ was injected into the yolk of a chicken egg together with the antifungal drug ketoconazole, which inhibits metabolism of FICZ by chicken CYP1A4- and CYP1A5, its toxic potency was similar to that of PCB126 (Jonsson et al. 2016).

Routine rodent toxicity studies with high doses of FICZ have yet to be performed, but certain reports involving exposure of adult mice to relatively high concentrations of FICZ have appeared with no mention of signs of pathology (Schulz et al. 2012; Wheeler et al. 2014).

One conclusion based on the relevant in vivo studies to date is that FICZ (unless its metabolic clearance is compromised) is not toxic toward adult vertebrates, since the rapid turnover of FICZ, regulated by the FICZ/AHR/CYP1A1 feedback loop, efficiently controls endogenous AHR functions. It is, however, apparent that FICZ can disrupt the embryonic development of fish and birds.

7. Other effects of FICZ

7.1. FICZ acts as a powerful sensitizer to blue light, which may result in photo-aging of the skin and the retinal epithelium

FICZ has been found to be an endogenous chromophore that acts as a photosensitizer, by potentiating UVA-induced oxidative stress even at nanomolar concentrations. Park et al. (2015) concluded that the photodynamic potency of FICZ exceeds that of the well-known endogenous photosensitizers riboflavin (vitamin B2) and protoporphyrin IX and described oxidative damage (including 8-oxo-dG the major DNA lesion induced by ROS) caused by combinations of FICZ and UVA in HaCaT human keratinocytes. Phototoxicity was also seen in primary human epidermal keratinocytes, organotypic epidermal skin reconstructs, and mouse skin. Repair of 8-oxo-dG by base excision repair (BER) was not affected, while the more versatile repair system, nucleotide excision repair (NER) was impaired (Brem et al. 2017). Thus, if not rapidly degraded, FICZ could cause dermal photocarcinogenesis, as well as photo-aging. Indeed, 100 nM FICZ inhibited TGFβ-induced up-regulation of collagen I synthesis (Murai et al. 2018). In addition, the AHR protects against blue light-induced degeneration of retinal pigmented epithelial cells (Kim et al. 2014).

The AHR regulates the stress response to UVB-light in human keratinocytes and in wild-type mice (Fritsche et al. 2007), as well as melanogenesis in human melanocytes and mouse skin (Luecke, Backlund, et al. 2010; Jux et al. 2011). Treatment of primary human melanocytes with 100 nM FICZ doubled the activity of the melanin-producing enzyme tyrosinase, an effect weaker and more transient than that of 10 nM TCDD (Luecke, Backlund, et al. 2010).

7.2. FICZ stimulates intrachromosomal exchange and retrotransposition

Concentration-dependent increases in sister chromatid conversion and gene conversion in the RS112 strain of Saccharomyces cerevisiae occurred upon exposure to 7.5–15 μM FICZ, with analogous increases in sister chromatid exchanges (SCE) in Chinese hamster ovary cells treated with 0.12–1.2 μM FICZ (Rannug, Agurell, et al. 1992). Since neither type of cells express AHR or CYP1A1 the effects are not AHR-dependent or generated by CYP-metabolites of FICZ. However, to suggest a plausible mechanism more data are needed.

Transposable elements (TEs), including LINEs and SINEs are estimated to comprise 45% of the human genome (Hancks and Kazazian 2012). Picomolar levels of FICZ induced retrotransposition (RTP) of long interspersed nucleotide element-1 (LINE-1 or L1) in human hepatocellular carcinoma HuH-7 cells, in a manner dependent on the activation of mitogen-activated protein kinase (MAPK) and ARNT, but not the AHR (Okudaira et al. 2010). In addition, FICZ (10 nM) regulated the transcription of Alu retrotransposons located in the upstream promoter regions of the pluripotency genes NANOG and OCT4 (Morales-Hernández et al. 2016). When administered by injection into mouse testes, 4 μg FICZ kg⁻¹ also repressed the germ-line specific genes MVH, Mili, and Miwi, effects mediated by the AHR and proposed to control the expression of small non-coding TEs involved in the maturation of germ cells (Rico-Leo et al. 2016).

7.3. FICZ activates nuclear hormone receptors

In addition to being a high affinity ligand for the AHR, FICZ activates receptors belonging to the superfamily of nuclear hormone receptors (NRs). In human HepG2 cells carrying expression vectors for full-length receptors, low micromolar amounts of FICZ activated most of the vertebrate vitamin D receptors (VDRs) studied, the invertebrate pregnane X receptor (Ciona VDR/PXR), and invertebrate and vertebrate liver X receptors (LXRs) (including the human LXRα and LXRβ) (Reschly et al. 2007, 2008; Ekins et al. 2008). Utilizing a reporter system that allows quantitative assessment of the activities of multiple transcription factors in HepG2 cells, 1 µM FICZ was also shown to activate vertebrate peroxisome proliferator-activated receptors (PPARs) (Romanov et al. 2008).

7.4. FICZ influences behavior

The expression of cyp1a1 exhibits circadian rhythmicity, with an acrophase that peaks several hours after the peak AHR expression under light–dark (LD) conditions (Huang et al. 2002), as well as in continuous darkness (DD) (Carmona-Antonanzas et al. 2017). TCDD and FICZ altered the expression of the clock genes Per1, Cry1, Cry2, and Bmal1 and attenuated phase shifts in response to light (Garrett and Gasiewicz 2006; Mukai and Tischkau 2007; Mukai et al. 2008; Xu et al. 2010).

Similar to the avoidance of novel items of food by rats exposed to TCDD (Lensu et al. 2011), Sprague-Dawley rats exposed to 100 μg FICZ kg⁻¹ by gavage avoided chocolate almost totally for only a few hours, although this avoidance was still clearly present two weeks later (Mahiout and Pohjanvirta 2016).

8. Conclusions and future perspective

Collectively, the findings reviewed here reveal that at high concentrations, various AHR ligands behave similarly, i.e. sustained AHR activation by high doses of the endogenous ligand FICZ resembles that caused by toxic ligands such as TCDD. However, at pico- or nanomolar concentrations, FICZ is different, because of its high affinity for the AHR in combination with its being an almost perfect substrate for CYP1A1, thus creating a FICZ/AHR/CYP1A1 negative feedback loop that makes FICZ an ideal regulator of transient and dynamic cellular processes. It is mandatory that this loop be regulated, because both endogenous and exogenous compounds that inhibit CYP1A1 might otherwise lead to accumulation of toxic levels of FICZ. Combustion products in the air we breathe, a major threat to human health, include the PAHs fluoranthene, pyrene, and phenanthrene that can cause such inhibition and thereby most likely dysregulate the growth and differentiation of embryonic, immune, and cancer cells.

The investigations described here also highlight the impact of FICZ on pathological processes, which motivates its further assessment as a potential therapeutic agent. In particular, the realization that FICZ can regulate immune functions suggests that this compound might be used in treating, e.g. IBD, MS, psoriasis, diabetes, allergies, and cancer. FICZ might also be valuable in connection with photodynamic therapies, wound healing, and regenerative therapies, due to its ability to expand populations of stem cells. At the same time, the observations that FICZ can also aggravate diseases such as MS, ALS, and rheumatoid arthritis in the animal models of these diseases underscores the importance of appropriate dose regimens.

Acknowledgements

The authors gratefully acknowledge the review comments provided by three external reviewers selected by the Editor and anonymous to the authors. These comments were very helpful in strengthening the manuscript and clarifying key points.

Declaration of interest

The employment affiliations of the authors are shown on the cover page. The authors report no conflicts of interest. The preparation of the manuscript was financially supported by departmental grants from The Department of Molecular Biosciences, The Wenner-Gren Institute, Stockholm University and Institute of Environmental Medicine, Karolinska Institutet. The manuscript has not been sent to any other journal. It has not reviewed by others. The authors have not appeared in any regulatory or legal proceedings in the last 5 years that draw on the contents of the paper. The paper and its conclusions are the exclusive professional work product of the authors.