Circadian tryptophan metabolism contributes to systemic aryl hydrocarbon receptor activity

Summary

Diet-phytochemical derived activation, host and microbial tryptophan metabolism represent the dominant route to endogenously mediated stimulation of physiological Ah receptor (AHR) activity. Whether host tryptophan metabolism provides a phytochemical independent circadian AHR tone has not been established. Using mice maintained on a nocturnally restricted feeding schedule with a nutritionally defined diet, we utilized quantitative gene/protein expression analyses in conjunction with targeted metabolomics to examine the temporal relationship between host tryptophan metabolism and circadian AHR activity. Time-resolved, targeted LCMS metabolomic, gene, and protein expression analyses reveal circadian cycling of hepatic tryptophan metabolizing enzymes (TDO2, TAT, GOT1, GOT2, KAT1, KAT2, and IL4I1) and serum tryptophan metabolites (indole-3-acetate, indole-3-lactate, indole-3-propionate, indole aldehyde, kynurenine, kynurenic acid) previously established as AHR ligands. We observed cyclical hepatic AHR activity directed by circadian feeding. These data suggest that a circadian rhythm of tryptophan metabolism orchestrates a daily tone in AHR activity that likely modulates AHR dependent physiology.

Graphical abstract

Highlights

• •Key tryptophan metabolites were assessed in serum and/or liver
• •Temporal ordered sequence of tryptophan metabolism and CYP1A1 expression was observed
• •A robust circadian rhythm of AHR activity and transaminase expression was observed
• •Dexamethasone intervention induces nocturnal hepatic tryptophan metabolism and AHR activity

Animal biotechnology; Metabolomics

Introduction

At the lowest resolution, organismal gene expression and physiology are orchestrated by an autonomous circadian rhythm, typically oriented around a ∼24 h cycle with opposing periods of physiological activity e.g., sleep/wake. Such rhythms are required to maintain system wide cooperativity and coordination. Circadian oscillations in gene expression are governed by a core network of synchronous, self-regulating ‘clock’ proteins comprised of CLOCK, BMAL, NR1D1, PERIODs, CRYPTOCHROMEs, and other ancillary factors that ultimately influence generalized gene expression patterns.^1,2 The core clock is autonomous, but it is acutely sensitive to both internal and external stimuli (Zeitgebers), including light, sound, temperature, humidity, nutritional factors, xenobiotics, psycho-social stress, and disease. Such stimuli can alter the synchronicity, phase, and amplitude of the circadian rhythm and, as such, have profound physiological consequences. Exposure to natural light, dictated by the Earth’s rotation, axis, and latitude, manifests as a dominant Zeitgeber, with signals translated from the retina to the superchiasmatic nucleus (SCN) within the hypothalamus, which houses the central clock signaling machinery. Neuronal and hormonal signals are disseminated from the SCN to peripheral organs to establish a central circadian rhythm that ultimately coordinates interorgan physiology. Such coordination introduces additional Zeitgebers, including feeding, hormone production, increased locomotor activity, changes in core body temperature, and so forth, that in turn signal back to the central clock in the SCN to reinforce its autonomous oscillation. Such feedback to the central clock is driven by the peripheral organs, which exhibit intrinsic circadian rhythms significantly influenced by, but independent from, the central clock. These peripheral clocks likely provide a decentralized cycle that ensures any given tissue functions in a synchronized, holistic manner and fully coordinates with the central clock to reinforce physiology.

Consistent with its significance as an environmental sensor and modulator of physiological function, the aryl hydrocarbon receptor (AHR) and its expression have been revealed to be under circadian control.^3,4 The AHR is reported to contribute to metabolic detoxification, lymphocyte differentiation, lipid metabolism, mucosal and dermal barrier function, embryonic development, cell cycle progression, tissue repair, and so forth.^5,6 However, it has not been established if the circadian nature of these physiological functions is, in part, a consequence of oscillations in AHR activity. While AHR protein expression has been linked to its overall transcriptional activity, its defining feature as the sole ligand binding member of the basic-Helix-Loop-Helix/Per/Sim/ARNT family of transcription factors suggests that direct transcriptional AHR activity is predominantly a function of ligand binding, and thus, ligand bioavailability. The circadian nature of AHR activity is likely linked to oscillations in the types, abundance, and disposition of its ligands in each tissue.

AHR was initially characterized as the high affinity receptor and toxicity mediator for environmental polycyclic and halogenated aromatic hydrocarbon (PAHs and HAHs, respectively) pollutants, including 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD), benzo(a)pyrene, and polychlorobiphenyls. However, these anthropogenic, exogenous AHR ligands clearly cannot form the basis for the circadian oscillation observed in physiological AHR activity. On the other hand, it has become apparent that non-toxicological AHR ligand activation is driven by both diet-derived phytochemical products, e.g., flavonoids, indolo[3,2b]carbazole (ICZ) and so forth, and a repertoire of endogenously (host and/or microbial) generated tryptophan (Trp) metabolites including kynurenine (KYN), kynurenic acid (KA), indole-3-pyruvate (I3Pyr), indole-3-acetate (IAA), indole-3-lactate (ILA), indole-3-aldehyde (IAld), and indole-3-propionate (IPA).⁷ Importantly, in mammals, the hepatic TDO2-mediated kynurenine pathway, which accounts for the majority (∼95%) of host tryptophan metabolism producing both KYN and KA, follows a circadian rhythm. Additional tryptophan metabolic pathways involving tryptophan trans/deamination by aminotransferase and deaminase enzymes are proposed to generate I3Pyr and downstream products with the capacity to activate AHR.⁸ Although the circadian nature of tryptophan trans/deamination and its relevance to physiological AHR activity have not been fully investigated. Intestinal microbial tryptophan metabolism, similar to host TDO2 activity, exhibits rhythmicity due to circadian host feeding cycles and associated dietary tryptophan availability and thus likely contributes to the circadian abundance of tryptophan metabolites.⁹

Data from independent studies in mice suggest that circadian host/microbial tryptophan metabolism and AHR transcriptional activity are in phase with each other, and therefore, may explain circadian AHR function, which typically achieves maximal activity (acrophase) during the nocturnal, active period.^3,10 Since these studies have been performed in the context of standard rodent chow, a rich source of AHR-activating phytochemicals, the potential synchronization of circadian AHR activity and tryptophan metabolism cannot be segregated from nocturnal dietary phytochemical exposure. Furthermore, whether the expression of amino acid transaminases is under circadian regulation that temporally correlates with increased tryptophan metabolite production has not been examined. Mice maintained on a semi-purified diet, deficient in phytochemicals, were utilized to evaluate whether AHR activity and specific tryptophan metabolite production are circadian, and if the potential relationship between the two could account for oscillations in AHR-dependent physiological processes.

Results

Ah receptor transcriptional activity and serum tryptophan metabolites conform to diurnal rhythms

To address the circadian nature of AHR activity and tryptophan metabolism and their potential synchronization, we employed a controlled 12 h light/dark regimen in which mice were trained for one week on a nocturnally restricted but otherwise ad libitum feeding schedule. Mice were fed a diet (AIN93G) deficient in phytochemicals, yet nutritionally complete, and housed in wire-bottom cages without edible bedding material. Samples were collected at 4 h intervals over a 24 h period (Figure 1A). To determine the circadian nature in these subjects, the expression of the opposing circadian indicators, Bmal and Nr1d1, was assessed. Quantitative PCR analysis of hepatic Bmal and Nr1d1 expression revealed significant (p < 0.01) but reciprocal periodicity, as determined by COSINOR analysis, associated with the expression of both circadian indicators. Hepatic Bmal demonstrated clear nocturnal (ZT12-24) expression, achieving acrophase at ZT20 (Figure 1B). Conversely, hepatic Nr1d1 expression was constrained to the diurnal phase, achieving acrophase at ZT8 (Figure 1B). Extra-hepatic examination of Bmal and Nr1d1 expression in cardiac, pulmonary, and renal tissues also identified significant (p < 0.01) periodicity, paralleling that observed in the liver (Figure S1). To further assess the circadian rhythm under this experimental regimen, serum corticosterone was quantified by LC-MS analysis (Figure 1B). Consistent with the reciprocal expression of Bmal and Nr1d1, serum corticosterone demonstrated significant (p < 0.01) periodicity, with increasing diurnal concentration, achieving acrophase at ZT12. These data establish a robust circadian rhythm entrained to the light/dark cycle, and that the absence of dietary phytochemicals does not overtly alter the expected expression of circadian indicators.

Figure 1.

AHR transcriptional activity conforms to a diurnal rhythm

(A) Scheme depicts the experimental regimen used for assessing the diurnal rhythm.

(B) Diurnal hepatic mRNA expression profile of the circadian markers Bmal and Nr1d1 together with serum corticosterone, as determined by QPCR and LC-MS, respectively.

(C) Diurnal hepatic mRNA expression profile for Ahr.

(D) Diurnal hepatic mRNA expression profile for AHR target genes Cyp1a1, Cyp2e1, and Ahrr, as determined by QPCR.

(E) Diurnal hepatic protein expression profile for CYP1A1, as determined by capillary electrophoresis.

(F) Visual representation illustrates the cumulative temporal acrophase patterning of indicated mRNA and protein expression combined with indicated metabolites. Where appropriate, data represent mean mRNA expression ±SEM (n = 5/ZT) normalized to 18S, mean serum concentration ±SEM (n = 5/ZT), and mean protein expression ±SEM (n = 3/ZT) normalized to total protein. Significant periodicity, as determined by COSINOR analysis, is indicated by p < 0.05.

Verification of a robust circadian rhythm allowed a valid assessment of AHR activity in the absence of dietary phytochemicals over the course of 24 h. The liver represents a major site of AHR activity; therefore, we first investigated the circadian nature of hepatic AHR expression and activity. Quantitative PCR analysis of hepatic Ahr expression demonstrated significant (p < 0.01) periodicity, achieving acrophase during the diurnal phase at ZT4 (Figure 1C). Further analysis revealed significant (p < 0.01) periodicity associated with the hepatic expression of the AHR regulated genes Cyp1a1 (p < 0.01), Cyp2e1 (p < 0.01), and Ahrr (p < 0.02) (Figure 1D). Both Cyp1a1 and Cyp2e1 reached acrophase toward the end of the diurnal phase between ZT8-12, shifted approximately +4 h relative to the peak expression of Ahr. Cyp2e1 mRNA and protein levels has been previously shown to be regulated by the AHR in mouse liver.¹¹ Hepatic Ahrr, although rhythmic, exhibited a phase shift in expression by +10 h, relative to Cyp1a1 and Cyp2e1, achieving acrophase during the nocturnal phase at ZT20 (Figure 1D). To determine if the observed oscillation in hepatic AHR transcriptional activity, as assessed by Cyp1a1, is reflected at the level of protein expression, time-resolved CYP1A1 protein levels were analyzed by capillary electrophoresis. Data identify a significant (p < 0.01) periodicity associated with CYP1A1, achieving a nocturnal ZT20 acrophase (Figure 1E). Comparison between Cyp1a1 and its protein level reveal a temporal delay of 8 h between maximal AHR transcriptional activity and maximal CYP1A1 protein expression (Figures 1D and 1E). These data, in the context of a phytochemical-free diet, reveal hepatic Ahr expression and its activity to be circadian in nature. Although the liver represents a major site of AHR activity, AHR exhibits almost ubiquitous systemic expression. We therefore examined if circadian transcriptional AHR activity is apparent in extra-hepatic tissues. Analysis of renal, pulmonary, and cardiac Cyp1a1 expression revealed significant (p < 0.01) periodicity, with diurnal acrophase achieved in each tissue at ZT8-12, similar to that observed in the liver (Figure S1). Interestingly, the amplitude of the diurnal acrophase is much greater in the lung compared to other tissues. These data indicate a coordinated, interorgan synchronization of AHR activity (Summarized in Figure 1F).

The circadian nature of AHR-dependent gene expression, across different organs, in the absence of dietary phytochemical exposure, indicates an oscillating endogenous and, perhaps, systemically circulating mode of AHR ligand activation. The reported role of numerous tryptophan metabolites as AHR modulators (Figure 2A) prompted a time-resolved analysis of such metabolites to identify if metabolite periodicity is evident and whether such fluctuations may account for the observed oscillation in AHR activity. Utilizing LC-MS analysis, the time-resolved serum and hepatic concentrations of tryptophan, IAA, IAld, ILA, IPA, KYN, and KA were quantified. These metabolites were examined because we previously have shown that they are the predominant metabolites that can activate the AHR found in mouse and human serum.¹² Tryptophan, as an essential, diet derived amino acid, exhibited significant (p < 0.01) periodicity. Serum tryptophan concentrations oscillated over a range of 40–60 μM throughout the 24 h window and were oriented around the nocturnal feeding schedule, achieving acrophase at ZT20 (Figure 2B). All Trp metabolites examined exhibited significant (p < 0.05) periodicity in serum concentrations (Figure 2B). Except for IPA, each host/microbial derived metabolite displayed a largely similar phase, achieving nocturnal acrophase at ZT20-ZT24. Conversely, the obligate microbial tryptophan metabolite IPA appeared out of phase by ∼15 h, achieving a diurnal acrophase at ZT4, thus indicating that microbial tryptophan metabolism is asynchronous with respect to host tryptophan metabolism (Figures 2B and 3B). Notwithstanding the observed synchronicity between the phases associated with serum Trp metabolites (Trp, IAA, ILA, IAld, KYN, and KA) and the asynchronous nature of IPA, the absolute serum concentrations of each metabolite at acrophase varied (Summarized in Figure 2C). Normalization of each metabolite to its rhythm-adjusted mean concentration (MESOR) identified KA as the Trp metabolite with the greatest relative amplitude, ∼3-fold (Figure 2D).

Figure 2.

Tryptophan metabolism conforms to a diurnal rhythm

(A) Scheme depicting the metabolism of tryptophan to host and microbial derived metabolites.

(B) Diurnal serum concentrations of tryptophan, KYN, KA, IPA, ILA, IAA, and IAld, as determined by LC-MS. Data represent mean serum concentration ±SEM (n = 5/ZT). Significant periodicity, as determined by COSINOR analysis, is indicated by p < 0.05.

(C) Visual representation illustrates the cumulative temporal acrophase patterning of indicated mRNA and protein expression combined with indicated metabolites.

(D) Heatmap depicts min-max concentrations of metabolites indicated in (B). Table listing individual metabolite MESOR values and % change during the diurnal cycle.

Figure 3.

Hepatic tryptophan metabolism conforms to a diurnal rhythm

(A) Scheme depicting the metabolism of tryptophan to host and microbial derived metabolites.

(B) Diurnal serum concentrations of tryptophan, Kyn, KA, IPA, ILA, IAA, and IAld, as determined by LC-MS. Data represent mean serum concentration ±SEM (n = 5/ZT). Significant periodicity, as determined by COSINOR analysis, is indicated by p < 0.05.

(C) Visual representation illustrates the cumulative temporal acrophase patterning of indicated mRNA and protein expression combined with indicated metabolites.

(D) Heatmap depicts min-max concentrations of metabolites indicated in (B). Table listing individual hepatic metabolite MESOR values and % change during the diurnal cycle.

Circadian comparisons between hepatic AHR activity and serum Trp metabolites demonstrate an apparent acrophase phase shift with hepatic Cyp1a1 expression preceding the nocturnal increase in tryptophan and its metabolites by ∼ 6 h. To determine if hepatic Trp metabolite abundance rather than serum concentration could account for the observed shift, time-resolved LC-MS analysis was performed on hepatic tissue samples (Figure 3A). Data reveal Trp metabolite-dependent discrepancies between hepatic and serum periodicity (Figures 2B and 3B). Notably, serum fluctuations in tryptophan are not evident within hepatic tissue. Notwithstanding, hepatic Trp metabolites exhibit significant (p < 0.05) periodicity except for the microbial product IPA. Those hepatic Trp metabolites that retain periodicity, similar to serum metabolites, display nocturnal acrophase. However, hepatic KA and IAld exhibit a −8 h shift relative to serum and thus hepatic AHR activity appears out of phase with serum and hepatic Trp metabolites (Figures 2B and 3B). The peak of Trp metabolite concentrations in serum exhibited a 4 h delay relative to peak liver concentrations, consistent with the concept that these metabolites are produced in the liver (Summarized in Figure 3C). Normalization of each metabolite to its rhythm-adjusted mean concentration (MESOR) identified KYN as the Trp metabolite with the greatest relative amplitude (Figure 3D).

Circadian serum tryptophan metabolite concentrations confer differential Ah receptor activation potential

Despite the observation that AHR activity precedes Trp metabolite levels, previous studies have indicated that the human or mouse serum concentrations of Trp metabolites, either individually or in a combined pool, are sufficient to promote AHR activity in vitro. To determine whether the observed oscillation in hepatic (and extra-hepatic) AHR activity may be associated with serum Trp metabolite concentrations, quantitative PCR analysis on mouse AHR-dependent Cyp1a1 expression was performed in Hepa1.1 cells treated for 4 h with differential, reconstituted ZT Trp pools. Trp pools were derived from mouse serum concentrations quantified at opposing circadian phases i.e., diurnal ZT4 and nocturnal ZT16 (Figure 4A: ZT4/16 metabolite concentration and % change). Exposure to either ZT4 or ZT16 Trp pools facilitated significant (p < 0.001) 11.78 ± 0.79 and 13.66 ± 1.64-fold increases, respectively, in Cyp1a1 expression relative to vehicle treatment (Figure 4B:ZT4/16 Hepa1 reporter assay). Analysis of ZT4 and ZT16 Trp pool mediated induction further revealed the AHR activation potential of the ZT16 pool to be significantly (p < 0.05) greater than the corresponding ZT4 pool. Similar studies, analyzing human AHR-dependent CYP1A1 expression, performed in human HepG2 cells also revealed significant (p < 0.001) 21.50 ± 4.39 and 32.66 ± 1.49-fold increases, respectively in CYP1A1 expression relative to vehicle treatment. Thus, analysis of ZT4 and ZT16 Trp pool mediated induction in human HepG2 cells revealed the AHR activation potential of the ZT16 pool to be significantly (p < 0.01) greater than the corresponding ZT4 pool (Figure 4C).

Figure 4.

Diurnally oscillating tryptophan metabolites confer differential AHR activation potential

(A) Table listing tryptophan metabolite concentrations quantified by LC-MS at ZT4 and ZT16 together with % change between these time points. These values were used to construct a synthetic Trp metabolite pool to assess relative AHR activation potential.

(B) Mouse Hepa1 cells were treated in triplicate for 4 h with synthetic Trp-metabolite pools constructed from either ZT4 or ZT16 serum quantification, as indicated in (A). Total RNA was isolated and Cyp1a1 expression assessed by QPCR.

(C) Human HepG2 cells were treated in triplicate for 4 h with synthetic Trp-metabolite pools constructed from either ZT4 or ZT16 serum quantification, as indicated in (A). Total RNA was isolated and CYP1A1 expression assessed by QPCR. Data represent relative gene expression (mean ± SEM) normalized to β-actin (n > 3/group). The value of p < 0.05 was considered statistically significant (∗p < 0.05).

These data indicate that differential, circadian Trp metabolite concentrations, independent of additional serum components, may be associated with the oscillation of hepatic AHR activity and that, within the context of mouse serum concentrations, are not constrained by AHR species.

Hepatic tryptophan metabolic enzymes conform to a diurnal rhythm

The essential nature of tryptophan, combined with the observed fluctuations in hepatic and serum Trp metabolites, likely affects tryptophan availability during the nocturnal feeding phase. However, it is less clear whether, in the context of variable substrate availability, whether the hepatic expression of tryptophan metabolic enzymes that can generate Trp-derived AHR ligands are themselves under circadian regulation (Figure 5A). Hepatic tryptophan dioxygenase 2 (TDO2) represents the dominant route for tryptophan metabolism, generating the AHR ligand KYN. Quantitative PCR analysis of time-resolved hepatic samples for Tdo2 expression demonstrates significant (p < 0.05) periodicity, achieving a diurnal acrophase at ZT8-12 (Figure 5B). Similarly, TDO2 protein expression analysis by capillary electrophoresis demonstrates significant (p < 0.01) periodicity with a nocturnal ZT16-20 acrophase, indicating a 4–6 h temporal shift between mRNA and protein maxima (Figure 5B). These observations indicate that hepatic Tdo2/TDO2 is diurnally regulated and hepatic/serum KYN concentrations are appropriately in-phase (Figure 5C).

Figure 5.

Hepatic expression of enzymes associated with the production of tryptophan metabolites exhibits diurnal oscillation

(A) Scheme depicts enzymes associated with the metabolism of tryptophan to host and microbial derived metabolites.

(B) Diurnal hepatic mRNA and protein expression profiles for Tdo2 and TDO2, as determined by QPCR and capillary electrophoresis, respectively. Data represent mean ± SEM mRNA (n = 5/ZT), normalized to 18S, and mean ± SEM for protein (n = 3/ZT), normalized to total protein. Significant periodicity, as determined by COSINOR analysis, is indicated by p < 0.05.

(C) Visual representation illustrates the cumulative temporal acrophase patterning of indicated mRNA and protein expression combined with indicated metabolites.

(D) Diurnal hepatic mRNA expression profiles for Kyat1, Aadat, and Got2, as determined by QPCR. Data represent mean ± SEM mRNA (n = 5/ZT), normalized to 18S. Significant periodicity, as determined by COSINOR analysis, is indicated by p < 0.05.

(E) Diurnal hepatic protein expression profiles for KYAT1, KYAT2, and GOT2, as determined by capillary electrophoresis. Data represent mean ± SEM protein (n = 3/ZT), normalized to total protein. Significant periodicity, as determined by COSINOR analysis, is indicated by p < 0.05.

(F) Visual representation illustrating the cumulative temporal acrophase patterning of indicated mRNA and protein expression combined with indicated metabolites.

(G) Diurnal hepatic mRNA expression profiles for Got1, Tat, Il4i1, Ldha, and Ldhb, as determined by QPCR. Data represent mean ± SEM mRNA (n = 5/ZT), normalized to 18S. Significant periodicity, as determined by COSINOR analysis, is indicated by p < 0.05.

(H) Diurnal hepatic protein expression profiles for GOT1, TAT, IL4I1, and LDHA, as determined by capillary electrophoresis. Data represent mean ± SEM protein (n = 3/ZT), normalized to total protein. Significant periodicity, as determined by COSINOR analysis, is indicated by p < 0.05.

(I) Visual representation illustrates the cumulative temporal acrophase patterning of indicated mRNA and protein expression combined with indicated metabolites.

Hepatic mRNA and protein analysis of enzymes [Kynurenine aminotransferases 1, Kyat1; Aminoadipate aminotransferase, Aadat (also known as Kyat2); and Glutamic-oxaloacetic acid transaminase 2, Got2 (also known as Alt)] associated with the conversion of the TDO2-derived product KYN into the AHR ligand KA were also assessed. Although demonstrating variable expression, Kyat1 failed to exhibit significant periodicity (Figure 5D). Time-resolved KYAT1 protein levels, however, demonstrated significant (p < 0.05) periodicity with a nocturnal ZT20 acrophase (Figure 5E). Unlike Kyat1, Aadat exhibited significant (p < 0.01) periodicity with a diurnal ZT8 acrophase. AADAT demonstrated significant (p < 0.04) periodicity but with a +6–8 h shift in acrophase relative to its mRNA (Figure 5E). Although demonstrating variable expression, Got2 failed to exhibit significant periodicity (Figure 5D). However, time-resolved GOT2 protein abundance demonstrated significant (p < 0.01) periodicity with a nocturnal ZT20 acrophase (Figure 5E). These data suggest that hepatic enzymes involved in KYN transamination to KA are diurnally regulated and that hepatic/serum KA concentrations are appropriately in-phase (Figure 5F).

In addition to KYN and KA, ancillary Trp metabolites derived from I3Pyr, through the product of tryptophan deamination, are known to promote AHR activity. I3Pyr, unless otherwise stabilized, is extremely labile in aqueous environments such as serum and thus difficult to quantify. Nonetheless, we demonstrate that the I3Pyr-derived products ILA, IAA and IAld exhibit serum/hepatic periodicity with nocturnal maxima (Figures 2B and 3A). Hepatic mRNA and protein analysis of enzymes (glutamic-oxaloacetic acid transaminase 1, Got1; Tyrosine aminotransferase, Tat, and Interleukin 4 induced 1, Il4i1) with a reported capacity to deaminate tryptophan to I3Pyr were assessed by quantitative PCR and capillary electrophoresis. Got1 exhibited significant (p < 0.01) periodicity with a diurnal ZT8 acrophase (Figure 5G). GOT1 similarly demonstrated significant (p < 0.01) periodicity but with a +12 h nocturnal shift in acrophase relative to its mRNA (Figure 5G). Tat exhibited significant (p < 0.05) periodicity, achieving acrophase at the ZT12 diurnal-nocturnal transition (Figure 5G). TAT protein also demonstrated significant (p < 0.05) periodicity associated with a +4 h shift in acrophase relative to its mRNA (Figure 5H). Similar to GOT1 and TAT, IL4I1 exhibited significant (p < 0.01) periodicity, achieving a diurnal acrophase at ZT6-8 (Figure 5H). IL4I1 protein, however, failed to demonstrate significant periodicity (Figure 5H). I3Pyr is reported to be a retrograde substrate for lactate dehydrogenase (LDH), leading to the production of ILA. As such, we examined if hepatic Ldha expression exhibits rhythmicity consistent with the observed diurnal oscillation in hepatic/serum ILA. Quantitative PCR analysis reveals the absence of significant (p=0.67) periodicity associated with hepatic Ldha expression (Figure 5G). In contrast, LDHA protein levels demonstrate significant (p < 0.01) periodicity with nocturnal ZT20 acrophase (Figure 5H). With the exception of IL4I1, these data suggest that hepatic enzymes involved in tryptophan transamination to I3Pyr and subsequent conversion to IAA, IAld, and ILA are diurnally regulated (Summarized in Figure 5I).

Diurnal glucocorticoid supplementation increases Ah receptor activity in parallel with hepatic tryptophan metabolism

The combined observations that tryptophan, hepatic enzymes involved in tryptophan metabolism, and associated Trp metabolites reported to be AHR ligands, all demonstrate rhythmicity, suggest such oscillations may be integral to circadian AHR activity. Both Tdo2 and Tat circadian targets, suspected mediators of AHR activity through the production of KYN and I3Pyr, are sensitive to glucocorticoid-mediated induction. As such, supplementary exposure to dexamethasone (Dex) is expected to increase TDO2 and TAT expression, tryptophan metabolism, and consequently AHR activity. To test this hypothesis, we adopted an intervention approach using in vivo diurnal glucocorticoid agonist administration (Figure 6A). Mice were intraperitoneally administered 5 mg/kg Dex at ZT4, coinciding with the onset of endogenous adrenal glucocorticoid release. 12 h post-Dex administration i.e., nocturnal ZT16, hepatic tissue and serum were isolated for gene expression, protein, and LC-MS analyses. Quantitative PCR analysis of hepatic Tdo2 demonstrates a significant (p < 0.01) 6.0 -fold induction relative to vehicle (10.6 ± 0.2), in response to diurnal ZT4-Dex administration (1.8 ± 0.4) (Figure 6B). Hepatic Tat exhibited a modest (albeit non-significant p = 0.056) change, relative to vehicle (11.6 ± 3.5), following ZT4-Dex administration (23.6 ± 12.2) (Figure 6B). Protein expression analyses of hepatic TDO2 demonstrate a significant (p < 0.01) 6.1-fold induction, relative to vehicle (0.3 ± 0.04), in response to ZT4-Dex administration (1.6 ± 0.18) (Figure 6B). In contrast to its cognate mRNA, hepatic TAT exhibited a modest yet significant (p < 0.01) 2-fold induction, relative to vehicle (4.2 ± 0.96), following ZT4-Dex exposure (8.5 ± 2.2) (Figure 6B).

Figure 6.

Diurnal dexamethasone intervention induces nocturnal hepatic tryptophan metabolism and AHR activity

(A) Scheme depicts experimental Dex intervention regimen.

(B) Influence of ZT4 Dex (5 mg/kg, i.p.) on ZT16 hepatic Tat and Tdo2 mRNA and protein expression, as determined by QPCR and capillary electrophoresis, respectively. Data are presented as box blots illustrating min-max quantification with a bar indicating mean value (n = 5/group). Protein expression data represent TAT or TDO normalized to total protein. Gene expression data represent Tat or Tdo2 mRNA normalized to 18S.

(C) Influence of ZT4 Dex on ZT16 serum tryptophan and kynurenine levels, as determined by LC-MS. Data are presented as boxplots illustrating min-max quantification with bar indicating mean value (n = 4–5/group, one sample from ZT4-Dex was eliminated due to serum hemolysis). Serum Trp and KYN were normalized to the chlorpropamide internal control.

(D) Influence of ZT4 Dex on ZT16 hepatic tryptophan metabolic component mRNA expression. Data are presented as boxplots illustrating min-max mRNA quantification with a bar indicating the mean value (n = 5/group), normalized to 18S.

(E) Influence of ZT4 Dex on ZT16 serum KA, IAA, IAld, and ILA levels, as determined by LC-MS. Data are presented as min-max metabolite quantification with bar indicating mean value (n = 4–5/group, one sample from ZT4-Dex was eliminated due to serum hemolysis). Serum metabolites were normalized to the chlorpropamide internal control.

(F) Table listing concentrations and % change in ZT16 serum Trp metabolite concentrations following ZT4 exposure to Dex relative to ZT4 vehicle.

(G) Influence of ZT4 Dex on ZT16 hepatic Cyp1a1 and Ahr mRNA expression. Data are presented as boxplots illustrating min-max mRNA quantification with a bar indicating mean value (n = 5/group), normalized to 18S. Statistical analyses were performed using the Mann-Whitney test. The value of p < 0.05 was considered statistically significant (∗p < 0.05; ∗∗p < 0.01).

To examine if the observed ZT4-Dex dependent amplification of hepatic TDO2 and TAT is functionally relevant regarding Trp metabolite production, LC-MS analyses were performed. Targeted analysis revealed nocturnal ZT16 serum tryptophan to be not significantly different between vehicle (92 ± 11 μM) and ZT4-Dex (110 ± 14 μM) (Figure 6C). Consistent with the observed ZT4-Dex dependent induction of hepatic TDO2, analysis of serum KYN identified a significant (p < 0.05) 2-fold increase in KYN (2.1 ± 0.2 μM) following ZT4-Dex, relative to vehicle (1.0 ± 0.1 μM) (Figure 6C). Functional analysis of increased TAT activity was also assessed through the serum quantification of the tyrosine and phenylalanine metabolites, 4-hydroxyphenylacetate and phenyllactate, respectively. Consistent with the observed ZT4-Dex dependent increase in TAT, serum 4-hydroxyphenylacetate exhibited a significant (p < 0.001) nocturnal 2-fold relative increase between ZT4-Dex (0.9 ± 0.1) and vehicle (0.5 ± 0.1) (Figure S2). Similarly, phenyllactate exhibited a significant (p < 0.05) 1.7-fold relative increase with ZT4-Dex (2.2 ± 0.8), compared to vehicle (1.3 ± 0.2) (Figure S2). These data thus support hepatic TDO2 and TAT protein expression and activity to be responsive to supplementary glucocorticoids.

We also examined the effect of ZT4-Dex upon the nocturnal hepatic expression of additional tryptophan metabolic pathway components that may be involved in the downstream production of KA and I3Pyr-derived AHR ligands. Quantitative PCR analysis of nocturnal ZT16 Kyat1 demonstrated a significant (p < 0.01) 2.7-fold induction in response to ZT4-Dex (37.8 ± 13.0), relative to vehicle (14.1 ± 2.1) (Figure 6D). In contrast, hepatic Aadat expression was not significantly influenced by ZT4-Dex (1.65 ± 0.2), relative to vehicle (1.85 ± 0.1) (Figure 6D). Analysis of nocturnal ZT16 Got1 mRNA expression identified a significant (p < 0.05) 2.5-fold induction in response to ZT4-Dex administration (0.98 ± 0.8), relative to vehicle (0.38 ± 0.1) (Figure 6D). In contrast, Got2 mRNA expression was not significantly influenced by ZT4-Dex administration (0.99 ± 0.2), relative to vehicle (1.1 ± 0.2) (Figure 6D). Analysis of Ldha and Ldhb, which may contribute to the interconversion between I3Pyr and I3L, revealed a significant (p < 0.01) 50% suppression associated with Ldha in response to ZT4-Dex (1.48 ± 0.3), relative to vehicle (3.3 ± 0.7). However, Ldhb expression was not significantly influenced by ZT4-Dex administration (1.4 ± 0.4), relative to vehicle (1.5 ± 0.3) (Figure 6D). These data suggest diurnal glucocorticoid supplementation may influence the subsequent nocturnal abundance of KYN, KA, and I3Pyr-derived Trp metabolites.

To examine if the observed ZT4-Dex dependent amplification of additional hepatic tryptophan metabolizing components influences serum Trp metabolite abundance, LC-MS analyses were performed. In contrast to the observed increase in KYN, we observed a significant (p < 0.05) 1.4-fold (121 ± 11 nM) nocturnal decrease in serum KA, compared to vehicle (177 ± 30 nM) (Figure 6E). Serum analysis of the additional circadian I3Pyr-derived Trp metabolites IAA, IAld, and ILA was also assessed. No significant change in nocturnal IAA was identified following ZT4-Dex exposure (125 ± 19 nM), relative to vehicle (97 ± 41 nM) (Figure 6E). IAld, the downstream product of IAA however, demonstrated a significant (p < 0.01) 1.6-fold nocturnal increase associated with ZT4-Dex administration (872 ± 140 nM), relative to vehicle (557 ± 79 nM) (Figure 6E). ILA, a product/substrate of LDH activity, exhibited a significant (p < 0.05) 2.1-fold nocturnal increase following ZT4-Dex (2.5 ± 0.9 μM), compared to vehicle (1.2 ± 0.3 μM) (Figure 6E). The obligate microbial Trp metabolite, IPA, exhibited no significant change in response to ZT4-Dex, relative to vehicle (Figure S2). These data indicate that diurnal glucocorticoid supplementation influences Trp metabolite abundance (Summarized in Figure 6F).

To investigate whether the observed coordinated increases in hepatic tryptophan metabolic enzymes and associated Trp metabolites, in response to diurnal ZT4-Dex, exert a subsequent functional impact on hepatic AHR activity, nocturnal Ahr and Cyp1a1 expression were assessed. Quantitative PCR analysis identified no significant change in nocturnal hepatic Ahr expression in response to ZT4-Dex (0.8 ± 0.2), relative to vehicle (0.9 ± 0.3) (Figure 6G). In contrast, nocturnal Cyp1a1 exhibited a significant (p < 0.01) and robust 2.2-fold increase in expression following ZT4-Dex administration (0.9 ± 0.3), relative to vehicle (0.4 ± 0.2) (Figure 6G). Altogether, these data suggest that circadian-regulated glucocorticoid exposure influences tryptophan metabolism and the subsequent magnitude of AHR transcriptional activity.

Discussion

Conservation across vertebrate evolution, combined with near ubiquitous tissue expression, indicates a physiological imperative for AHR transcriptional activity. Indeed, toxicological, inhibitor, and genetic knockout/overexpression studies demonstrate that the AHR is a participant in numerous physiological processes, including xenobiotic detoxification, adaptive and humoral immunity, embryonic vascular development, wound healing, maintenance of epithelial integrity, microbial colonization, and defense, and so forth. Consistent with its significance as a physiological modulator, AHR activity, in the apparent absence of known xenobiotic exposure, follows a circadian rhythm, oscillating around the 24 h light/dark cycle.^13,14 Many of the processes in which AHR participates are themselves circadian, suggesting that the temporal regulation of AHR activity is likely of physiological consequence. For example, the presence of endogenous AHR ligands—such as those derived from tryptophan metabolism— in significant concentrations throughout the day should maintain basal AHR activity. Even after fasting overnight, sufficient CYP1A1 expression would essentially be able to detoxify subsequent exposure to potentially deleterious substrates in the next meal. In addition, the presence of tryptophan in a meal would further transiently increase tryptophan metabolite levels, leading to increased CYP1A1 expression. Indeed, the fluctuating yet consistent presence of these metabolites may, in effect, prime the promoters of AHR target genes to allow a more rapid response to dietary ligands and/or toxic xenobiotics. Evidence supporting the protective effect of enhanced CYP1A1 expression has been demonstrated using transgenic mice with constitutive intestinal CYP1A1 expression, which mitigates the effect of exogenous FICZ on systemic AHR activation.¹⁵ Our results provide evidence for a role of endogenous Trp metabolites in mediating temporal AHR activity, which likely modulates the systemic dissemination of exogenously derived CYP1A1 substrates and AHR ligands. Consistent with protective CYP1A1 activity, we observe a logical temporal order comprising nocturnal induction of tryptophan metabolism→increased Trp metabolites→diurnal induction of Cyp1a1 mRNA (AHR activity)→nocturnal expression of CYP1A1 protein during the active feeding phase and time of likely exposure to diet-derived xenobiotics. Additional factors, including hepatic zonation and substrate availability, together with mRNA and protein turnover rates, likely contribute to and reinforce such temporal regulation of hepatic CYP1A1 (and other AHR targets).¹⁶

It is clear from a circadian perspective that AHR-regulated physiological processes cannot be solely dependent upon sporadic exposure to xenobiotics or spontaneously, non-enzymatically generated AHR agonists. It has become increasingly evident that AHR activity can be stimulated by a range of endogenous mediators, principally derived from tryptophan (e.g., KYN, KA, and IAA), thus positioning the AHR as an endobiotic Trp metabolite sensor.¹² Importantly, under non-pathological conditions, tryptophan and its metabolism have been shown to be integrated signaling components governing and responding to circadian rhythms.^17,18 Consistent with these observations, we demonstrate that maximal AHR activity, regardless of tissue origin and serum tryptophan concentration are coincident, both achieving a nocturnal acrophase. However, such parallel increases may be unconnected and could be attributed to nocturnal feeding behavior since tryptophan is an essential diet-derived amino acid and standard rodent chow provides a rich source of uncharacterized phytochemical AHR ligands. Indeed, it has been demonstrated that asynchronous diurnal feeding of standard chow elicits a robust out-of-phase induction of AHR activity that is distinct from the typical nocturnal induction.¹⁴ For this reason, we adopted a semi-purified rodent chow nutritionally largely devoid of phytochemicals and can thus eliminate a direct dietary component as the mediator of nocturnal AHR activity. To the best of our knowledge, this represents the first report describing an intact AHR rhythm in the absence of dietary phytochemical activators. This observation supports the hypothesis that oscillations in AHR activity are intrinsic and a function of endogenous signaling.

Evidence exists for the occurrence of AHR binding, nuclear translocation, and transcriptional activity in response to either L- or D-tryptophan, which suggests agonist activity.¹⁹ However, these studies used supraphysiological (mM) concentrations and long incubations, did not account for in situ tryptophan metabolism—particularly the action of D-amino oxidase in generating I3Pyr-derived agonists—nor for the competitive displacement from albumin of preexisting AHR agonists (e.g., KYN) by tryptophan.^20,21 In the absence of cell-free evidence to the contrary, tryptophan is not considered to be a direct AHR ligand, and thus, circadian tryptophan is unlikely to be directly responsible for increased nocturnal AHR activity. In contrast, we and others have previously established a link between endogenous tryptophan metabolites and AHR activity.¹² An obvious candidate for a tryptophan metabolite facilitating circadian AHR activity is the pineal hormone melatonin, responsible for orchestrating sleep/wake cycles. Melatonin does exhibit weak AHR agonist activity, but only at extreme supraphysiological concentrations.²² Furthermore, the C57BL6/J strain used here has been shown to be deficient in melatonin synthesis. Therefore, melatonin is not considered as a potential physiological modulator of AHR activity in the context of this study.²³ Consistent with feeding behavior, an increase in nocturnal tryptophan is associated with elevated Trp metabolite concentrations. The observed oscillation in hepatic enzyme expression associated with tryptophan metabolism suggests that Trp metabolite levels are not solely a function of substrate, i.e., tryptophan availability, but rather a consequence of regulated expression and catalytic activity. TDO2 expression and activity are established as being under circadian and substrate regulation due to its sensitivity to glucocorticoid mediated induction and allosteric control by tryptophan.^24,25,26 Such regulation appears to be critical in the avoidance of systemic tryptophan toxicity.^27,28 Importantly, KYN, the product of TDO2 activity, exhibits AHR activity at nocturnal concentrations detected here.¹² A similar contributing factor to nocturnal AHR activity can be ascribed to nocturnal KA production by KYAT activity on KYN. KA levels appear to be the most dynamic of the Trp metabolites examined, suggesting KYAT expression/activity to be a rate-limiting factor. In fact, we observed synchronous expression of KYAT1, AADAT, and GOT2, each with the capacity to generate KA. While detected in the nM range, KA manifests as a potent AHR agonist.^12,29 Although circadian patterning of KYN and KA is previously described, such patterns have not been correlated with oscillations in AHR activity.

Utilizing both mouse and human hepatoma cell lines, Trp metabolites at physiological concentrations in mouse serum were examined at ZT4 and ZT16. Both mouse and human cell lines exhibited significant AHR activation of the target gene Cyp1a1, thus indicating that serum levels of Trp metabolites can drive basal AHR activity. Importantly, the human hepatoma line revealed a much greater difference between the two time points, compared to the mouse hepatoma line. While there are a number of reasons that might explain this difference, the most likely reason is that several Trp metabolites have previously been shown to be more potent activators of the human AHR compared with the mouse AHR.^30,31 In particular, kynurenic acid exhibits a 100-fold greater potency to activate the human AHR relative to the mouse AHR in reporter cell lines.²⁹ These results suggest that fluctuating Trp metabolite levels in humans may have a more profound effect on AHR activation than that observed in mice.

The oscillation of KYN and KA concentrations clearly contributes toward AHR activity. We also report for the first time, in-phase nocturnal increases in IAA, ILA, and IAld, each of which similarly presents as AHR ligands and likely contributes to AHR activation in combination with KYN and KA.¹² The production of IAA, ILA, IAld, and other AHR-sensitive Trp metabolites is often attributed to microbial activity, suggesting that circadian microbial metabolism may facilitate oscillations in AHR activity. However, we have previously demonstrated the presence of IAA in germ-free mice, indicating that nocturnal IAA concentrations coincide with AHR activity, and thus do not necessarily require IAA of microbial origin.^12,32 Furthermore, circadian levels of IPA, an indicator of microbial tryptophan metabolism and a relatively weak AHR ligand, are diametrically out-of-phase with the other Trp metabolites examined and AHR activity (diurnal v. nocturnal), thus arguing against a direct microbial determinant for circadian systemic AHR activity.

The most likely host-derived source of these non-TDO2-derived Trp metabolites is through the deamination of tryptophan to I3Pyr and subsequent metabolism. A number of aminotransferases have been shown to accept tryptophan as a substrate, albeit with diminished affinity relative to prototypical substrates.^33,34,35,36 We establish rhythmic expression associated with the transaminases TAT, GOT1, GOT2, and associated LDHA that are synchronous and consistent with Trp metabolite abundance and AHR activity. Importantly, studies using the controlled administration of dexamethasone in humans identified serum IAA, ILA, and KA as being significant glucocorticoid induced metabolites.¹⁰ To our knowledge, I3Pyr (and hence transamination) represents the only precursor to ILA, suggesting circadian glucocorticoids influence tryptophan transamination. The in vivo contribution of these transaminases toward AHR ligand generation is impossible to determine within the scope of this study. Biochemical studies, however, have demonstrated that TAT has the highest specificity and catalytic rate regarding the deamination of tryptophan, indicating that it may represent the dominant circadian factor in the in vivo production of I3Pyr.³⁴ Supporting evidence is also provided through our glucocorticoid intervention, which revealed significant induction of TAT (and TDO2) protein despite a modest (albeit non-significant) increase in Tat mRNA, likely due to the recorded short half-life of Tat mRNA.³⁷ Significantly, functionally elevated TAT, as evidenced by the increased production of TAT-dependent tyrosine/phenylalanine catabolites, correlates with increased I3Pyr-derived Trp metabolites and subsequent AHR activity. Importantly for this assessment, glucocorticoids have not been shown to directly activate AHR. The amino acid oxidase IL4I1, while not exhibiting periodicity within the context of this study, may impact circadian AHR activity under pathological conditions.^8,33 Its reported capacity to deaminate tryptophan to I3Pyr in response to induction by IL4 would likely promote AHR activation independent of circadian Trp metabolites. It remains to be established if such asynchronous AHR activity would contribute to or mitigate IL4-mediated pathology or physiological AHR activity.

We and others demonstrate that AHR mRNA and protein expression exhibit a diurnal oscillation that may account for fluctuations in AHR activity, independent of Trp metabolites.^3,4 Interestingly, the expression of the SP1 family of transcription factors displays a circadian pattern and is also involved in the basal expression of both AHR and CYP1A1.^38,39,40 However, our observation that glucocorticoid intervention did not induce Ahr expression combined with our in vitro assessment using quantified circadian Trp metabolite pools, indicate ligand availability, not AHR expression levels, are necessary and sufficient to facilitate differential AHR activity. Recent analysis of Trp metabolites from overnight-fasted human sera from a controlled diet study indicates inter-metabolite and total Trp metabolite (and corresponding AHR activation) variability across individuals.¹² The source of variability, despite a fasted state, was not established. Future studies examining urinary excretion rates of Trp metabolites, the polymorphic nature of tryptophan metabolic components, and internal circadian time (e.g., glucocorticoid and melatonin concentrations), and recorded sleep duration of individuals correlated with Trp metabolite levels may provide some rationale for the observed variability.

In summary, we provide a circadian driven, Trp metabolite-AHR ligand production/bioavailability rationale to account for observable rhythmic oscillations of AHR activity in mice. The diversity of Trp metabolites may also facilitate tissue specific AHR activity through the differential expression of individual Trp metabolite transporters and downstream metabolic pathways. While it remains to be established if such a rationale is applicable to humans, these data suggest that physiological processes governed by AHR activity are likely sensitive to circadian disruption, and thus likely a contributing factor to circadian-related pathologies—e.g., insomnia, depression, stress, adrenal insufficiency, inflammation, cancer, and so forth.^41,42,43 Moreover, the recent clinical advancement of AHR-directed therapies using agonists or antagonists suggests time of day administration (chronotherapy) and ligand half-life may have circadian-related consequences for both efficacy and potential side effects.

Limitations of the study

A major focus of this study was the production of indole-3-pyruvate by tryptophan-transaminating or -deaminating enzymes, which raises two key limitations. First, indole-3-pyruvate is known to rapidly decompose under aqueous conditions into numerous secondary products, many of which remain unidentified. It is therefore possible that additional, uncharacterized tryptophan-derived metabolites with AHR-activating properties were present but remain undetected. Second, numerous host enzymes are capable of transaminating or deaminating tryptophan, making it difficult to deconvolute the relative contributions of individual enzymes to metabolite production in vivo. Furthermore, it is well established that gut-colonizing bacteria can produce the indole metabolites examined in this study. However, the extent to which the microbiota contributes to circulating concentrations of these metabolites, and whether this influences their circadian rhythms, remains unclear. Germ-free or antibiotic-treated mice may help address this question, but both models disrupt normal host physiology and therefore represent imperfect controls. As such, our study does not fully resolve the role of the microbiome. Finally, our work was conducted in mice, whose microbiomes and temporal gene expression patterns differ markedly from those of humans, raising questions about translatability. This concern is highlighted by the lower sensitivity of the mouse AHR to tryptophan-derived metabolites than the human AHR. Differences in lifestyle further complicate this comparison: for example, human populations such as shift workers often experience circadian disruption alongside dietary patterns enriched in ultra-processed foods and caffeinated beverages, factors not examined in our model. Finally, only female mice were used in this study, and male mice could vary in the results obtained.

Resource availability

Lead contact

Requests for further information and resources should be directed to and will be filled by the lead contact. Gary Perdew (ghp2@psu.edu).

Materials availability

This study did not generate new unique reagents. Biological material is not available because the small amount available for the experiments has been used up.

Data and code availability

• •The metabolomics data are available at MassIVE at https://doi.org/10.25345/C5901ZV1Z.
• •This article does not report original code.
• •Any additional information required to reanalyze the data reported in this article is available from the lead contact upon request.

Acknowledgments

The coauthors acknowledge the Huck Institutes of the Life Sciences Metabolomics Core Facility (RRID:SCR_023864). Graphical representations were created using BioRender. We thank Marcia H. Perdew for critically reviewing this article. This research was supported by the National Institutes of Health under grants National Institutes of Environmental Health Grants ES028244 (GHP), ES035027 (ADP), and T32DK120509 (EWM). This work was also supported by the USDA National Institute of Food and Federal Appropriations under Project PEN04916 and Accession number 1009993.

Author contributions

E.W.M., I.A.M., A.D.P., and G.H.P. designed and conceived the work performed. G.H.P. supervised the study. E.W.M. performed the metabolomic analysis. E.W.M., I.A.M., and D.M.C. performed all other experiments. E.W.M., I.A.M., D.M.C., A.D.P., and G.H.P. contributed to data analysis. E.W.M. and I.A.M. wrote the initial draft of the article. A.D.P. and G.H.P. reviewed and edited the article.

Declaration of interests

The authors declare no competing interests.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Rabbit polyclonal anti-CYP1A1	Proteintech	13241-1-AP
Rabbit polyclonal anti-TDO2	Proteintech	15880-1-AP
Rabbit polyclonal anti-IL4I1	Invitrogen	PA5-113266
Rabbit polyclonal anti-LDHA	Proteintech	19987-1-AP
Rabbit polyclonal anti-TAT	Invitrogen	PA5-80097
Rabbit polyclonal anti-KAT1	Invitrogen	PA-5100131
Rabbit polyclonal anti-AADAT	Proteintech	13031-1-AP
Rabbit polyclonal anti-GOT1	Proteintech	14886-1-AP
Rabbit polyclonal anti-GOT2	Proteintech	14800-1-AP
Rabbit monoclonal anti-BACT	Cell Signaling	4970
Chemicals, peptides, and recombinant proteins
Indole-3-propionic acid	Alfa Aesar	L04877; CAS: 830-96-6
Indole-3-acetic acid	Alfa Aesar	A10556; CAS: 87-51-4
Indole-3-lactic acid	Sigma-Aldrich	I5508; CAS: 832-97-3
L-kynurenine	Cayman Chemical	11305; CAS: 2922-83-0
Indole-3-aldehyde	Sigma-Aldrich	129445; CAS: 487-89-8
corticosterone	Cayman Chemical	16063; CAS: 50-22-6
tryptophan	Sigma-Aldrich	T0254; CAS: 73-22-3
kynurenic acid	Sigma-Aldrich	K3375; CAS: 492-27-3
Deposited data
Raw LC-MS data files	MassIVE	https://doi.org/10.25345/C5901ZV1Z
Experimental models: Cell lines
Hepa1c1c7	ATCC	CRL-2026
HepG2	ATCC	HB-8065
Experimental models: Organisms/strains
Mouse: C57BL6/J	Jackson Laboratory	000664
Oligonucleotides
Oligonucleotide primers for RT-PCR, see Table S1	Integrated DNA Technologies	N/A
Software and algorithms
GraphPad Prism	GraphPad	v10.4.1
MS-DIAL	N/A	v5.5.250221
SciOS	Sciex	V3.3.0.12027
COSINOR.online	Molcan, L. (2023)	https://www.biorxiv.org/content/10.1101/805960v2

Experimental model and subject details

Mouse model

Female C57BL6/J mice were purchased from Jackson Laboratories (Bar Harbor, ME, USA) and housed at the Pennsylvania State University vivarium. Mice between 8-10 weeks of age were used for experiments. Purchased mice were acclimated to the vivarium environment for one week and housed as five mice per cage. Mice were housed on corncob bedding in a temperature- and light-controlled facility and given access to food and water ad libitum. Mice were maintained in a pathogen-free facility and treated humanely with approval from the Animal Care and Use Committee of the Pennsylvania State University and methods were carried out in accordance with approved guidelines.

Cell lines

Hepa1 and HepG2 (mouse and human, respectively) hepatoma cell lines were maintained in αMEM (Sigma) supplemented with 8% FBS (Hyclone) and penicillin/streptomycin (Sigma) and were authenticated by obtaining them from ATCC.

Method details

Chemicals and reagents

AIN93G was purchased from Dyets, Inc., (Bethlehem, PA, USA) and the composition has been previously published.⁴⁴ LC-MS grade solvents including, methanol and water, were purchased from Fisher Scientific (Hampton, NH, USA) and acetonitrile was purchased from Honeywell (Charlotte, NC, USA). Stock solutions of all reference standards for liquid chromatography were prepared in 10% acetonitrile (v/v) containing 1 μM chlorpropamide (internal standard). The mixed standard solutions were serially diluted with 10% acetonitrile (v/v) containing 1 μM chlorpropamide for generating the calibration curves.

Mouse experiments

Mice were fed standard laboratory chow prior to a 3-day washout period on the phytochemical free AIN-93G diet before commencement of the temporally controlled feeding schedule. The experimental regimen consisted of mice being transferred to wire-bottomed cages containing non-edible enrichment to restrict coprophagia and consumption of bedding. A strict 12 h light/dark cycle was established with lights on at 06:00 h and off at 18:00 h (defined as Zeitgebers TIME (ZT)0/24 and ZT12, respectively). Mice were allowed ad libitum access to water throughout the experimental regimen. The temporally restricted feeding schedule was established by allowing ad libitum access to AIN-93G only during the dark cycle (18:00-06:00 h; ZT12-ZT24). The regimen was maintained for 7 days prior to euthanasia and collection of serum and tissues at 4 h intervals starting at ZT4. Dark cycle subjects were maintained in the dark during euthanasia.

An identical regimen was followed for the dexamethasone administration study. Dexamethasone (Sigma, NJ) was administered by intraperitoneal injection (5 mg/kg) in corn oil at ZT4, control mice received corn oil alone. Mice were returned to cages for completion of the study under the established temporal regimen. Mice were euthanized at ZT16 and indicated tissues collected.

Whole blood was collected through cardiac puncture with a 25G needle and transferred to BD Microtainer SST™ tubes (Becton Dickinson, NJ). Samples were incubated at room temperature for 30 min prior to centrifugation (1,200 x g, 15 min, 4°C). Sera were transferred to tubes, flash frozen in liquid nitrogen and stored at -80°C for further processing. Hepatic, cardiac, pulmonary and renal tissues were excised, rinsed in ice-cold PBS prior to being flash frozen in liquid nitrogen and stored at -80°C for further processing. Individual processing of each mouse was completed within 5 mins of euthanasia. All animal protocols were reviewed and approved by the Animal Care and Use Committee of The Pennsylvania State University (protocol # 201901049).

RNA isolation and quantitative PCR

Total RNA was isolated from indicated tissues with TRI Reagent (Sigma, NJ). Briefly, ∼100 mg tissue was homogenized in 1 ml TRI Reagent with 1-2 mm zircon beads using a BeadBlaster 24 (Benchmark) homogenizer (3 cycles of 6.5 m/s for 30 s). Homogenates were centrifuged (1,000 x g, 5 min, 4°C) to pellet insoluble material. Supernatants were transferred to fresh tubes and 0.2 ml chloroform added. Samples were shaken and then centrifuged (12,000 x g, 15 min, 4°C). Aqueous phase samples were transferred to fresh tubes and RNA precipitated with an equal volume of 100% isopropanol. Samples were centrifuged (16,000 x g, 30 min, 4°C) and the RNA pellet washed with 0.5 ml 70% ethanol prior to air drying. Samples were solubilized in DEPC-treated nuclease-free water at a concentration of 400 ng/μl.

2 μg of RNA was converted to cDNA using a High-Capacity cDNA synthesis kit (Applied Biosystems) following manufacturer’s instructions. Quantitative PCR was performed as previously described using validated oligonucleotide primers.⁴⁵ Oligonucleotide sequences are provided in Table S1.

Metabolite profiling and quantification by liquid chromatography-mass spectrometry

Tryptophan metabolites were quantified using LC-MS, according to previously established methods.¹² Briefly, 25 μl sera were mixed with 4 volumes ice-cold methanol supplemented with 1 μM deuterated internal standards (indole-3-acetic acid-d4 and kynurenic acid-d5) (Cayman Chemical). Samples were vortexed, incubated at -20°C for 30 min and then centrifuged (12,000 x g, 15 min, 4°C). 90 μl of extract from each sample was collected and dried under vacuum using a SpeedVac (Thermo Scientific, Waltham, MA). Dry samples were solubilized in 45 μl 1:9 acetonitrile:water supplemented with 1 μM chlorpropamide (internal validation standard). Samples were centrifuged (12,000 x g, 15 min, 4°C) and soluble fractions transferred to autosampler vials for LC-MS analysis.

Metabolite quantification was performed using reverse phase UHPLC on a Nexera system (Shimadzu, Columbia, MD) equipped with a Waters Acquity BEH C18 column (2.1 x 100 mm, 1.7 μm particle size) maintained at 55°C. An aqueous acetonitrile gradient was applied at a flow rate of 250 μL/min for 20 min using Solvent A (water/0.1% formic acid) and solvent B (acetonitrile/0.1% formic acid). Initial gradient was established at 97% A/3% B increasing to 55% A/45% B after 10 min, 25% A/75% B after 12 min. The 25% A/75% B gradient was held until 17.5 min before returning to initial conditions. Eluates were introduced into a ZenoTOF 7600 using a Turbo V ion source (SCIEX, Framingham, MA). Data were acquired in high resolution multiple reaction monitoring mode, with mass accuracy calibrated against a SCIEX standard mix. Metabolite validation was assessed by using authentic standards with matching retention time, parent mass and MS/MS fragmentation. Analyte recovery and matrix effects for each metabolite were quantified using the following formulae. PA: Peak area of analyte; PApre: Peak area of analyte spiked in pre-extraction; PApost: Peak area analyte spiked in post-extraction; PAend: Peak area of endogenous analyte, PAwater: Peak area of spiked analyte in water.

$$\%recovery=\left(\frac{\left(PApre-PAend\right)}{PAend}\right)X100$$

$$\%matrix=\left(\frac{PAspike-PAend}{PAwater}\right)X100$$

Calculated values were used to compensate quantification. Limits of detection and quantification were defined as signal/noise >3 and > 10, respectively, as determined using MSDIAL v.5.4.241004. Metabolite quantification was interpolated using non-weighted linear regression based on a six-point standard curve containing chlorpropamide, indole-3-acetic acid-d4 and kynurenic acid-d5 as internal standards.

Protein analysis-capillary gel electrophoresis

Approximately 100 mg of each tissue samples was homogenized in 1 ml 2X Cell Lysis Buffer (Cell Signaling Technology) supplemented with protease inhibitor cocktail (Roche) with 1-2 mm zircon beads using a BeadBlaster 24 (Benchmark) homogenizer (3 cycles of 6.5 m/s for 30 s). Lysates were centrifuged (1,000 x g, 5 min, 4°C) to pellet insoluble material. Supernatants were transferred to fresh tubes and centrifuged (10,000 x g, 10 min, 4°C) again. Protein concentration was determined using BCA kit (Pierce), following manufacturer’s instructions, and samples normalized to 1 mg/ml with lysis buffer.

Immuno-detection/quantification was performed by capillary gel electrophoresis using the Jess™ system (ProteinSimple) following recommended protocol for single immunoassay and total protein asay. Protein samples were diluted to 0.2 mg/ml with 0.1X ProteinSimple sample buffer prior to preparation of plate conditions. Specific detection was achieved using anti-TAT (1:200), anti-KAT1 (1:400), anti-KAT2 (1:50), anti-IL4I1 (1:50), anti-LDHA (1:5000), anti-TDO2 (1:200), anti-AHR (1:200), anti-CYP1A1 (1:100), anti-GOT1 (1:400), and anti-GOT2 (1:400). Secondary detection was achieved using HRP-conjugated anti-rabbit (Protein Simple). Total protein was determined using supplied Total Protein Module. Quantification was achieved by normalizing baseline subtracted target peak area to in-well total protein peak area.

Cell culture

Prior to treatment with Trp metabolite pools, cells were washed with PBS and incubated for 24 h in serum-free αMEM supplemented with 1 mg/ml BSA and penicillin/streptomycin. Cells were subsequently treated for 4 h with vehicle (0.1 % v/v DMSO) or Trp metabolite pools derived from serum concentrations quantified at ZT4 and ZT16. Relative AHR activity was assessed by quantitative PCR analysis as indicated. Cell lines were routinely tested for mycoplasma contamination and were negative.

Quantification and statistical analysis

Gene expression, protein expression and metabolite periodicity were assessed using the COSINOR online (Molcan, L. [2023]). Time-distributed data analysis by Cosinor. Statistically significant periodicity was assigned with a p < 0.05. The specific statistical tests and the value of n either for animals or cells is stated in each figure legend. All data in graphs is expressed as the mean ± SEM. GraphPad Prism v10.4.1. (GraphPad Software, San Diego, CA) in-built unpaired, non-parametric t-tests were applied where appropriate with α significance set at 0.05. Statistical comparisons between groups were analyzed using two-tailed non-parametric Mann-Whitney with α significance set at 0.05 (∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001 and ∗∗∗∗p < 0.0001).

Published: January 12, 2026