2,3,7,8-Tetrachlorodibenzo-p-dioxin (TCDD) dysregulates hepatic one carbon metabolism during the progression of steatosis to steatohepatitis with fibrosis in mice

2,3,7,8-Tetrachlorodibenzo-p-dioxin (TCDD), a persistent environmental contaminant, induces steatosis that can progress to steatohepatitis with fibrosis, pathologies that parallel stages in the development of non-alcoholic fatty liver disease (NAFLD). Coincidently, one carbon metabolism (OCM) gene expression and metabolites are often altered during NAFLD progression. In this study, the time- and dose-dependent effects of TCDD were examined on hepatic OCM in mice. Despite AhR ChIP-seq enrichment at 2 h, OCM gene expression was not changed within 72 h following a bolus dose of TCDD. Dose-dependent repression of methionine adenosyltransferase 1A (Mat1a), adenosylhomocysteinase (Achy) and betaine-homocysteine S-methyltransferase (Bhmt) mRNA and protein levels following repeated treatments were greater at 28 days compared to 8 days. Accordingly, levels of methionine, betaine, and homocysteic acid were dose-dependently increased, while S-adenosylmethionine, S-adenosylhomocysteine, and cystathionine exhibited non-monotonic dose-dependent responses consistent with regulation by OCM intermediates and repression of glycine N-methyltransferase (Gnmt). However, the dose-dependent effects on SAM-dependent metabolism of polyamines and creatine could not be directly attributed to alterations in SAM levels. Collectively, these results demonstrate persistent AhR activation disrupts hepatic OCM metabolism at the transcript, protein and metabolite levels within context of TCDD-elicited progression of steatosis to steatohepatitis with fibrosis.

One carbon metabolism (OCM) comprises the interlinking methionine and folate cycles to provide one carbon units for biosynthetic reactions 1 . This includes the biosynthesis of S-adenosylmethionine (SAM), the primary cellular methyl donor for methyltransferase reactions, and the second most utilized enzymatic cofactor after ATP 2 . SAM is essential for the biosynthesis of several products required for maintaining and regulating cell structure and function including creatine for ATP regeneration, phospholipids such as phosphatidylcholine for membrane integrity and lipid transport, and epigenetic gene regulation via the methylation of histones, DNA, and RNA 3,4 . In addition, decarboxylated SAM serves as the source of aminopropyl groups for polyamine biosynthesis which is important for cell growth, survival and proliferation 5 .
Alterations in the levels of SAM, as well as related OCM metabolites, can have profound effects on cell growth, differentiation, response to injury, and tissue regeneration 1 . For example, SAM reduces homocysteine remethylation by allosterically inhibiting betaine homocysteine methyltransferase (BHMT) in the methionine cycle Scientific RepoRtS | (2020) 10:14831 | https://doi.org/10.1038/s41598-020-71795-0 www.nature.com/scientificreports/ liver using perchloric acid (PCA). Briefly, ~ 25 mg liver tissue was added to ice cold 0.4 M PCA containing internal standards (SAM-13 C 5 , SAH-d4, and cell free amino acid mixture- 13 C, 15 ) and AcN (solution B). The QTof was run in positive ionization mode with continuum data acquisition and leucine enkephalin used as the lockspray reference compound. The MS total useful signal (MSTUS) method was used to normalize urine samples 35 . Progenesis software (Waters, Milford, MA) was used to determine total useful signal for each sample by summing metabolite peak areas common to all samples. Derivatized polyamine extracts were separated with a Waters Acquity UPLC BEH C18 column (1.7 µM particle size, 2.1 × 100 mm) held at 30 °C. The mobile phases were water containing 0.1% formic acid (solution A) and AcN (solution B). The QTof was run in positive ionization mode with continuum data acquisition and leucine enkephalin used as the lockspray reference compound. Liver extracts were normalized to total protein of each sample.
Geneexpressionanalysis. Hepatic RNA-seq data sets were previously published 22 . Genes were considered differentially expressed when |fold-change|≥ 1.5 and posterior probability values (P1(t)) ≥ 0.8 as determined by an empirical Bayes approach 36,37 . For figures, relative transcript counts represents the maximum raw number of aligned reads to each transcript across all treatments indicating the potential level of hepatic expression, where low level of expression ≤ 500 reads, and higher level of expression ≥ 10,000 reads. Sequencing data for the 72 h time course and 28 day dose response study are available at the Gene Expression Omnibus (GEO; GSE109863 and GSE87519, respectively). Hepatic gene expression in the 8-day study and renal gene expression in the 28 day study were assessed using quantitative real-time polymerase chain reaction (qRT-PCR). Total RNA was reverse transcribed by SuperScript II (Invitrogen) using oligo dT primer according to the manufacturer's protocol. PCR amplification was conducted on a Bio-Rad CFX Connect Real-Time PCR Detection System. Gene expression relative to vehicle control was calculated using the 2 −ΔΔCT method. Liver samples were normalized to the housekeeping genes ActB, Hprt, and

Results
TCDDeliciteddose-dependenteffectsonOCM. AhR activation following acute or repeated treatment with TCDD elicits NAFLD pathologies in mice that include dose-dependent hepatic lipid accumulation, immune cell infiltration, and periportal fibrosis with bile duct proliferation occurring only in males 22,[40][41][42][43][44][45] . Dysregulation of OCM, most notably SAM and SAH levels, is reported in human NAFLD and rodent models 2,9 . Gene expression, protein levels, and metabolite levels were integrated to further investigate the time and dose-dependent effects of TCDD on OCM including the SAM-dependent creatine and polyamine biosynthesis pathways.
To assess the effects of TCDD on SAM biosynthesis and metabolism (Fig. 1a), gene expression, protein levels, and metabolite levels were analyzed in liver samples after mice were orally gavaged every 4 days for 8 or 28 days. At 8 days, TCDD elicited a dose-dependent decrease in the SAM/SAH ratio (Fig. 1b). By 28 days, the SAM/ SAH ratio exhibited a non-monotonic dose-response, with a decreasing trend between 0.3 and 10 µg/kg TCDD (Fig. 1c). The effects on SAM and the SAM/SAH ratio are consistent with changes in OCM gene expression and protein levels. At 8 days, TCDD dose-dependently repressed Mat1a (BMDL 0.5 µg/kg; Fig. 1c). At 28 days, TCDD repressed Mat1a 4.5-fold (BMDL 0.1 µg/kg). However, at 30 µg/kg TCDD, Mat2a was induced 2.1-fold while repressing highly expressed Gnmt (12.1-fold), a known regulator of SAM levels, and Sardh (19.6-fold), which catalyzes the oxidative demethylation of sarcosine back to glycine (Fig. 1c). The dose-dependent decreases in MAT1A and GNMT protein levels were in agreement with respective gene repression (Fig. 1d). In addition, other highly expressed SAM-dependent methyltransferases including guanidinoacetate N-methyltransferase (Gamt), indolethylamine N-methyltransferase (Inmt), nicotinamide N-methyltransferase (Nnmt), and phosphatidylethanolamine N-methyltransferase (Pemt) were repressed 2.0-, 636.4-, 2.9-and 3.7-fold, respectively at 28 days (Fig. 1c). The non-monotonic dose-response for the SAM/SAH ratio at 28 days likely involves dysregulation of Mat1a and Mat2a expression, as well as the repression of Gnmt, in addition to the repression of SAM-dependent Gamt, Inmt, Nnmt, and Pemt methylation reactions (Fig. 1b,c). Despite pDRE-independent AhR enrichment at 2 h for most of these genes, there was negligible gene repression within the first 72 h following treatment with 30 µg/kg TCDD (Fig. 1e). Moreover, repression of the above genes was greater at 28 days compared to 8 days.
Other metabolites of the OCM and transsulfuration pathways (Fig. 2a) are important for methylation and were also affected by TCDD. Homocysteine is the product of SAH hydrolysis catalyzed by adenosylhomocysteine (AHCY) which was dose-dependently repressed by TCDD at the mRNA and protein levels in the absence of AhR enrichment (Fig. 2b). Under normal conditions, BHMT uses betaine as a donor to methylate homocysteine in the re-synthesis of methionine, producing N,N-dimethylglycine as a byproduct. At 28 days, Bhmt mRNA and protein levels were dose-dependently repressed by TCDD (Fig. 2b,c). Accordingly, there was an 2.6-fold increase in betaine and a non-significant 1.4-fold decrease in N,N-dimethylglycine (Fig. 2d). Alternatively, homocysteine can enter the transsulfuration pathway. However, cystathionine β-synthase (Cbs) mRNA and protein levels were also dose-dependently repressed with a corresponding decrease in cystathionine levels (Fig. 2b,c,e). At 28 days, cystathionine levels recovered following treatment with 10 and 30 µg/kg TCDD which may be due the allosteric activation of CBS by increasing SAM levels. Repression of BHMT in the methionine cycle and CBS in the transsulfuration pathway is consistent with the dose-dependent increase in homocysteic acid (Fig. 2e), Hepatic levels of SAM and SAH were determined by LC-MS/ MS (mean ± s.e.m., n = 5-6) at 8 and 28 days of repeated TCDD exposure and (c) hepatic gene expression of genes involved in the biosynthesis, regulation, and utilization of SAM and SAH were assessed at 8 and 28 days by RT-qPCR or RNA-seq, respectively (n = 8). (d) Fold change for hepatic MAT1A and GNMT protein levels after 28 days measured by the WES capillary electrophoresis system (mean ± s.e.m., n = 4). (e) Hepatic gene expression associated with SAM metabolism was determined by RNA-seq for a time-course after a bolus dose of 30 µg/kg TCDD (n = 5). For the heatmaps, the median effective dose (ED 50 ) and benchmark dose lower limit (BMDL) and relative transcript count (rel. count, ) are denoted. The red/blue color scale represents the log 2 (fold change) for differential gene expression. Orange represents the presence of putative dioxin response elements (pDREs). AhR enrichment peaks (FDR ≤ 0.05) are denoted by light green. pDREs found within AHR ChIPseq enrichment peaks are denoted by garnet. Asterisks (*) denote p < 0.05 determined by one-way ANOVA with a Dunnett's post-hoc test. Pound signs (#) denote posterior probabilities P1(t) ≥ 0.80 compared to vehicle. Official gene name and symbol, and metabolite abbreviations: www.nature.com/scientificreports/ produced as a result of the spontaneous oxidation of accumulating homocysteine 46 . Collectively, these changes would be expected to reduce hepatic methionine levels, but they were increased 5.8-fold (Fig. 2e). This may be due to increased methionine import with the induction of Slcs 1a5, 7a5, 7a7, 7a8, 38a1, 38a2 and 43a2. Slc3a2, the heavy chain heterodimeric partner for many amino acid transporters, was also dose-dependently induced by TCDD and contained a ChIP-seq peak with a pDRE (Fig. 2f).
Effectsonpolyaminemetabolism. To further investigate the effects of TCDD-elicited alterations in the SAM/SAH ratio, subsequent effects on SAM-dependent pathways were examined. Polyamines are ubiquitous polycationic alkylamines that are crucial for a broad range of cellular functions including cell cycle modulation, scavenging reactive oxygen species, and the control of gene expression. S-adenosylmethionine decarboxylase (Amd1), which was modestly induced (1.6-fold at 3 µg/kg TCDD, Fig. 3a), catalyzes the decarboxylation of SAM to produce decarboxylated SAM. The donation of the propylamine group from decarboxylated SAM to putrescine is catalyzed by spermine synthase (Srm) which was induced 2.5-fold at 30 µg/kg TCDD (Fig. 3a). Surprisingly, putrescine levels, which are typically kept low within cells 47 , were increased 8.0-fold at 30 µg/kg TCDD (Fig. 3b), possibly due to the 1.9-fold induction of ornithine decarboxylase (Odc1), and the 2.3-fold repression of ornithine decarboxylase antizyme (Oaz1), which promotes ODC1 degradation (Fig. 3a) 5 . There were also 1.6-and 2.2-fold increases in spermidine and N1-acetylspermidine at 30 µg/kg TCDD, respectively (Fig. 3b). The 4.9-fold induction of the polyamine transporter, Slc22a3, and the 3.9-fold induction of spermine oxidase (Smox) at 30 µg/kg TCDD, which oxidizes spermine to spermidine, may also contribute to putrescine accumulation (Fig. 3a). Like methionine and homocysteine metabolism, polyamine associated gene expression exhibited only moderate changes in the first 72 h following a bolus dose of 30 µg/kg TCDD (Fig. 3c). Collectively, TCDD dysregulated polyamine biosynthesis and transport, consistent with increased putrescine, spermidine and N1-acetylspermidine levels (Fig. 3d).

Discussion
Alterations in OCM, including human polymorphisms, correlate with NAFLD progression and severity [50][51][52] . Previous work has shown that persistent AhR activation by TCDD induces a NAFLD-like phenotype that includes hepatic fat accumulation, inflammation, and mild fibrosis in mice 22,41,45,53 . However, the effects of TCDD on OCM within the context of this NAFLD model have not been comprehensively investigated. Herein, we show TCDD dose-dependently alters OCM at the transcript, protein, and metabolite levels. The SAM/SAH ratio, an indicator of methylation potential, and creatine and polyamine biosynthesis pathways were altered.
In this study, TCDD is used as a surrogate for the cumulative burden of all AhR ligands. Mice were orally gavaged with 0.01-30 µg/kg TCDD starting at post-natal day 28 with TCDD every 4 days for 8 and 28 days to Hepatic gene expression associated with homocysteine metabolism was measured at 8 or 28 days by qRT-PCR and RNA-seq, respectively (n = 8). (c) Hepatic protein levels (mean ± s.e.m.) were determined by capillary electrophoresis for AHCY, BHMT, and CBS in male mice at 28 days (n = 4). (d) Metabolite fold change at 8 days (mean ± s.e.m., n = 3-6) or 28 days (mean ± s.e.m., n = 4-5) were determined by LC-MS/MS for betaine, N,N-dimethylglycine and (e) cystathionine (8 and 28 days), or methionine and homocysteic acid (28 days only). (f) Hepatic gene expression of methionine transporters at 28 days (n = 8). (g) Hepatic gene expression associated with homocysteine metabolism was determined by RNA-seq for a time-course after a bolus dose of 30 µg/kg TCDD (n = 5). For the heatmaps, the effective dose (ED 50 ), benchmark dose lower limit (BMDL), and relative transcript counts (rel. count) are denoted. The red/ blue color scale represents the log 2 (fold change) for differential gene expression. Orange represents the presence of putative dioxin response elements (pDREs). AhR enrichment peaks (FDR ≤ 0.05) are denoted by light green. pDREs found within AHR ChIP-seq enrichment peaks are denoted by garnet. Asterisks (*) denote p < 0.05 determined by one-way ANOVA with a Dunnett's post-hoc test. Pound signs (#) denote posterior probabilities P1(t) ≥ 0.80 compared to vehicle. Official gene name and symbol: Ahcy adenosylhomocysteinase, Bhmt betaine homocysteine S-methyltransferase, Cbs cystathionine beta-synthetase.   33,44,[61][62][63] . Although there was an increase in serum ALT levels following oral gavage with 30 µg/kg TCDD every 4 days for 28 days for a total of 7 treatments, there was no evidence of overt toxicity, no body weight loss > 15%, no significant change in food consumption, and no histopathological evidence of necrosis or apoptosis 23,33,64 . Consequently, the dose-dependent effects of TCDD on OCM, the SAM/SAH ratio and the biosynthesis of polyamines and creatine cannot be attributed to overt toxicity.
Mat1a is highly expressed in the adult liver making it the major site of SAM biosynthesis. As the primary methyl group donor for methyltransferase reactions, hepatic SAM and SAH levels are maintained in a narrow range with increases or decreases to the SAM/SAH ratio outside this window potentially affecting numerous cell functions 2,65 . Our studies show TCDD dose-dependently repressed Mat1a mRNA and protein levels (Fig. 1c,d), as previously reported with other AhR ligands 66,67 . TCDD also altered the SAM/SAH ratio with trends suggesting SAM levels decreased while SAH levels increased at higher doses at 8 days (Fig. 1b). A similar trend was observed at 28 days, except for a reproducible increase in the SAM/SAH ratio at 30 µg/kg TCDD (Fig. 1b) that coincided with the induction of Mat2a and the repression of several highly expressed methyl transferases including GNMT (Fig. 5a). GNMT is the most abundant hepatic methyltransferase that acts as a sink by transferring methyl groups from SAM to glycine to reduce SAM levels in order to regulate methionine consumption and SAM levels 7 . Interestingly, repression of Mat1a with the induction of Mat2a, which is usually only expressed during liver development, is characteristic of NAFLD and aggressive hepatocellular carcinoma progression [68][69][70][71][72] .
SAH is a potent competitive methyltransferase inhibitor 73 . It is readily hydrolyzed by AHCY to adenosine and homocysteine (not measured in our analysis) to allow OCM to proceed. However, AHCY mRNA and protein levels were dose-dependently repressed by TCDD (Fig. 2b,c). Homocysteine can be re-methylated back to methionine by BHMT or MTR which catalyzes the transfer of a methyl group from betaine or 5-methyltetrahydrofolate, respectively. In mice, BHMT is highly expressed in the liver and is the primary methionine biosynthesis pathway 74 . TCDD dose-dependently repressed BHMT coincident with an increase in betaine and decrease in N,N-dimethylglycine (Fig. 2c,d). An alternative pathway to re-methylation, homocysteine can be converted to cystathionine by CBS, which would undergo further metabolism by cystathionine gamma-lyase to produce cysteine and support glutathione biosynthesis. However, TCDD dose-dependently repressed Cbs mRNA and protein levels, and reduced cystathionine levels (Fig. 2b,c,e). CBS activity may also be further allosterically repressed by TCDD induced oxidative stress 75,76 . The partial recovery of cystathionine levels at higher TCDD doses is consistent with Mat2a induction with increased SAM levels allosterically activating CBS 6 . Consequently, homocysteine is not consumed in the transsulfuration pathway or by re-methylation to methionine, and undergoes oxidation as indicated by the increase in homocysteic acid levels, the spontaneous oxidation product of homocysteine (Fig. 2e). Despite BHMT transcriptional repression and inhibition by SAM, hepatic methionine levels increased due to the induction of transporters, and the repression of Mat1a and major methyltransferases that consume SAM (Fig. 5).
Given TCDD affected OCM, we next examined the potential consequences of an altered SAM/SAH ratio on polyamine and creatine biosynthesis. Polyamines are low molecular weight aliphatic polycations present in all living cells. De novo synthesis, interconversion, degradation and transport ensure levels of putrescine, spermidine and spermine are maintained in a narrow range since low levels inhibit cell proliferation and high levels or catabolic byproducts are toxic 5,47 . Approximately 5% of hepatic SAM is used to produce polyamines 48,77 . Consistent with previous reports that TCDD increases ODC activity, our results showed TCDD increased Odc expression and disrupted polyamine biosynthesis [78][79][80] , in contrast to short term studies that reported TCDD  www.nature.com/scientificreports/ decreased polyamine levels [81][82][83] . Adaptive compensatory responses such as increased Smox expression, catalyzing the interconversion of spermine to spermidine, and the differential expression of transporters, may partially explain the different effects of TCDD on polyamine levels we observed after 28 days of treatment (Fig. 3). The increase in putrescine by TCDD is comparable to levels induced by 12-O-tetradecanoylpphorbol-13-acetate (TPA), an inducer of ODC activity [84][85][86] . Putrescine levels are normally low when demand for polyamines are low due to multiple levels of ODC activity regulation and the allosteric activation of SAM decarboxylase (AMD1) 5 . Yet, putrescine levels increased despite the TCDD-elicited dose-dependent induction of Odc and spermidine synthase (Srm) suggesting decarboxylated SAM may be limiting (Fig. 3a,b). Paradoxically, there was an increase in spermidine and N1-acetylspermidine levels with negligible effects on spermine levels possibly due to interconversion and/or transport to maintain cellular polyamine homeostasis. Phosphocreatine is an important phosphate donor that can quickly regenerate ATP via substrate level phosphorylation reactions. As much as 70% of hepatic SAM is consumed in creatine biosynthesis, although hepatic levels are low with > 90% of creatine stored in muscle 87 . TCDD increased renal Gatm expression while repressing hepatic Gamt with modest effects on the levels of creatine and creatinine despite an increase in serum GAA (Fig. 4) comparable to levels reported in GAMT −/− mice and humans deficient in GAMT activity 88,89 . TCDD also decreased creatinine levels in short-term in vitro studies 82,90 . Similar to polyamines, adaptive responses following prolonged TCDD exposure may account for the modest changes in hepatic creatine and creatinine levels despite the repression of Gamt. For instance, induction of the creatine importer, Slc6a8, in the liver was only observed after 28 days of treatment (Fig. 4b). Collectively, the SAM-dependent biosynthesis of both creatine and polyamine demonstrated differential gene expression and metabolite levels. Despite disruption of OCM by TCDD, the effects on polyamine and creatine biosynthesis cannot be adequately explained due to alterations on the SAM/SAH ratio alone.  , 7a5, 7a7, 7a8, 16a9, 38a1, 38a2, and 43a2 solute carrier family.
Scientific RepoRtS | (2020) 10:14831 | https://doi.org/10.1038/s41598-020-71795-0 www.nature.com/scientificreports/ Many genes associated with OCM and the transsulfuration pathway exhibited BMDLs in the sub to low µg/ kg range only after 8 and 28 days of treatment. In addition, OCM and transsulfuration pathway disruption was time-dependent with the greatest effects after 28 days, more modest changes at 8 days and modest effects following a single bolus dose. The canonical mechanism of action of TCDD and related compounds involves binding to the cytoplasmic AhR, translocation to the nucleus, and heterodimerization with ARNT. The ligand-bound AhR/ARNT complex then binds to DREs within the promoter region of target genes, leading to recruitment of transcriptional co-regulators and differential gene expression 91 . Numerous studies have also reported differential gene expression following AhR binding within DNA regions lacking a DRE [92][93][94] . Despite ChIP-seq evidence of AhR enrichment at 2 h, only modest changes in OCM gene expression were observed in the first 72 h after treatment. In contrast, AhR targets such as Cyp1a1, Cyp1a2 and Tiparp were induced within 2 h 95 . Moreover, many differentially expressed OCM genes exhibited AhR genomic enrichment in the absence of a pDRE. Collectively, these results suggest that (i) although AhR activation is required, it in itself is not sufficient and likely requires unknown additional responses, (ii) TCDD-elicited OCM disruption involves DRE-dependent and -independent changes in gene expression, and (iii) the effects of TCCD are not immediate and require persistent AhR activation.
Given that OCM and transsulfuration pathway enzyme activity is subject to allosteric activation and competitive inhibition by intermediate metabolites (Fig. 5b), and are regulated by post-translational modification, more integrative approaches such as tracer studies are required to identify the key steps affected by TCDD that alter the flux of 13 C-labelled intermediates through OCM and its associated pathways. Moreover, an examination of other SAM-dependent reactions would expand our understanding of additional OCM mechanisms disrupted by TCDD such as the methylation of histones, DNA, and RNA associated with epigenetic regulation, and the biosynthesis of phosphatidylcholine via the PEMT and Kennedy pathways. Phosphatidylcholine is not only critical for membrane integrity, but also the secretion of very low-density lipoprotein (VLDL) 2,65 . Interestingly, the inhibition of VLDL secretion by TCDD contributes to steatosis in mice 41,53 . Additional studies are required to determine the relevance of these effects in humans due to the species-specific effects of TCDD and related compounds.