EWS‐FLI1 impairs aryl hydrocarbon receptor activation by blocking tryptophan breakdown via the kynurenine pathway

Ewing sarcoma (ES) is an aggressive pediatric tumor driven by the fusion protein EWS‐FLI1. We report that EWS‐FLI1 suppresses TDO2‐mediated tryptophan (TRP) breakdown in ES cells. Gene expression and metabolite analyses reveal an EWS‐FLI1‐dependent regulation of TRP metabolism. TRP consumption increased in the absence of EWS‐FLI1, resulting in kynurenine and kynurenic acid accumulation, both aryl hydrocarbon receptor (AHR) ligands. Activated AHR binds to the promoter region of target genes. We demonstrate that EWS‐FLI1 knockdown results in AHR nuclear translocation and activation. Our data suggest that EWS‐FLI1 suppresses autocrine AHR signaling by inhibiting TDO2‐catalyzed TRP breakdown.

Ewing sarcoma (ES) is the second most common primary malignant bone tumor in children and young adults. ES is characterized by the presence of a gene rearrangement between EWSR1 and one of five different ETS transcription factor (ETS) genes, with FLI1 being most commonly affected [1]. The expression of EWS-FLI1 (EF) results in the modulation of hundreds of different target genes [2]. RNA-sequencing data analysis revealed tryptophan 2,3-dioxygenase-2 (TDO2) as one of the genes being significantly up-regulated after silencing EF in A673sh cells [3]. We therefore investigated the involvement of EF in tryptophan (TRP) metabolism since little is known about the metabolic alterations caused by EF in cells, and also the repressive EF regulatory network is still poorly understood. Most of the dietary TRP, an essential amino acid, is metabolized along the kynurenine (KYN) pathway leading to the synthesis of NAD + together with intermediate products, including KYN, 3-hydroxykynurenine (3-HK), and quinolinic acid [4]. A secondary path from KYN leads to the generation of kynurenic acid (KYNA) via kynurenine aminotransferase-1 (KAT1) [5]. TDO2 and indoleamine 2,3-dioxygenase-1 (IDO1) are the first enzymes of the pathway (see Fig. 1A) with TDO2 being almost exclusively expressed in the liver and the brain [6], whereas IDO1 is found in tissues throughout the body [7]. The KYN pathway has been implicated in a variety of diseases and disorders such as AIDS, Alzheimer's disease, depression, schizophrenia, Huntington's disease, amyotrophic lateral sclerosis, and neoplasia [8]. Different metabolites of the KYN pathway have been associated with immune active properties [9], and because of their roles in immunity and the central nervous system, the KYN pathway has emerged as an attractive target for drug development [7]. Several downstream metabolites of the KYN pathway are biologically active in various physiological and pathological processes, including KYN, KYNA, 3-HK, anthranilic acid, 3-hydroxyanthranilic acid, and quinolinic acid [10]. KYNA has agonistic activity on the G protein-coupled receptor GPR35 [11] and antagonistic effects on glutamate receptors, in particular, the glycine co-agonist site of the N-methyl D-aspartate (NMDA)-receptor, and the cholinergic a7 nicotinic receptor, but it is also a ligand for the aryl hydrocarbon receptor (AHR) [12,13]. Similarly, KYN was identified as an endogenous AHR ligand in immune and tumor cells, acting both in an autocrine and paracrine manner, and promoting tumor cell survival [14]. AHR belongs to the subgroup of basic-helix-loop-helix Per-ARNT-Sim (bHLH-PAS) transcription factors [15], best known as a receptor for xenobiotics such as polycyclic aromatic hydrocarbons [16]. Ligand activation of AHR leads to its  breakdown is initiated by TDO2 or IDO1 and the product is hydrolyzed to KYN. KYN itself can act as signaling molecule or can be the substrate for KMO, KAT, or kynureninase in order to fuel different pathways. Multiple arrows depict several enzymatic steps; gray triangle highlights KYN pathway. (B) A673sh cells harbor a dox-inducible shRNA against EWS-FLI1 allowing for switching from high (EWS-FLI1-high, EFH) to low (EWS-FLI1-low, EFL) EWS-FLI1 expression levels. A673sh protein lysates were immunoblotted after 72 h induction of the shRNA. Western blot shows a representative experiment and quantification of protein expression AE SD from three replicates using LICOR Odyssey Infrared Imaging System is displayed in lower panel. ***P < 0.001. (C) Differential RNA expression of enzymes of the KYN pathway after EWS-FLI1 depletion. TDO2 expression increased and expression of KMO and KAT1 decreased strongly. Values are shown as mean log2 fold change relative to EWS-FLI1 expressing cells (data taken from [3]). ***P-value <0.001; Voom test statistic [71] adjusted for multiple testing by the Benjami Hochberg method.
translocation to the nucleus, dimerization with AHR nuclear translocator (ARNT) [17], and binding to dioxin response elements (DRE) in the promoter region of target genes. Among these targets are genes encoding enzymes for xenobiotic metabolism, such as the cytochromes P450 CYP1A1/2 and CYP1B1, several phase II conjugating enzymes [15,18], and pro-inflammatory interleukins (IL)-8 [14], IL-1b, and IL-6 [19]. Analysis of large-scale gene expression, chromosomal copy number, and massively parallel sequencing data of 947 human cancer cell lines from the Cancer Cell Line Encyclopedia identified elevated AHR as mechanistic biomarker for enhanced MEK inhibitor sensitivity in NRAS-mutant cell lines [20]. In addition, exposure to toxic polycyclic aromatic hydrocarbons stimulating the AHR has been implicated in a variety of cancers in experimental animals and humans [21].
The majority of genes activated by EF annotates to cell-cycle regulation and proliferation, in contrast to genes down-regulated by EF that mainly associate with cell differentiation and cell communication [22]. Our study was performed in A673sh cells [23], where EF can be silenced via doxycycline (dox)-inducible RNA interference (RNAi). Here, we identify a new signaling pathway that is activated when EF expression is low. It is induced via TRP breakdown and accumulation of intermediate metabolites in A673sh cells, which might play an important role in the pathogenesis of ES. This is the first report, to our knowledge, investigating the regulatory function of EF on the KYN pathway.

Cells
A stably transfected subclone of A673 cells (A673sh) with a dox-inducible shRNA against the EF fusion protein was used [23]. Cells were kept in DMEM + GlutaMax supplemented with 10% fetal calf serum (FCS), 100 UÁmL À1 penicillin, and 100 lgÁmL À1 streptomycin (Gibco by Life Technologies, Carlsbad, CA, USA), 50 lgÁmL À1 zeocin, and 2 lgÁmL À1 blasticidin (InvivoGen, San Diego, CA, USA). In order to induce the EF shRNA, 1 lgÁmL À1 dox (Sigma-Aldrich) was added to the medium. Cells were transfected using the Lipofectamine Plus reagent (Invitrogen, Groningen, NL, USA) and on the following day, selection for efficiently transfected cells was performed with puromycin (InvivoGen). Preparation of fractionated cell extracts was accomplished using the NE-PER Nuclear and Cytoplasmic Extraction Reagents Kit according to the manufacturer's instructions (Thermo Fisher Scientific, Waltham, MA, USA). TRP and KYN were measured as described previously [24]. In brief, their concentrations were determined by HPLC on reversed-phase C18 columns and subsequent monitoring of their UV absorption at 360 nm (KYN) and fluorescence (TRP) at 286 nm excitation and 366 nm emission wavelengths. For the analysis of KYNA 20 lL of each cell culture supernatant was injected on an HPLC column (Reprosil 100 C18, 3 lm particles, 100 9 4 mm, Dr. Maisch HPLC GmbH, Ammerbuch, Germany). The analytes were eluted with an acetonitrile/sodium acetate (6.8%/ 30 nM) solution at a flow rate of 0.5 mLÁmin À1 . The eluate was then mixed on line with zinc acetate (final concentration 0.125 M) and KYNA was determined with a fluorescence detector set at excitation and emission wavelengths of 344 and 398 nm, respectively. The signals from the fluorescence and UV-VIS detectors were transferred to a computer and analyzed by DATALYS AZUR software (Grenoble, France). The concentration of KYNA was extrapolated from freshly prepared standard curves. To determine 3-HK concentration, 20 lL was subjected to analysis utilizing an isocratic reversed-phase HPLC system coupled to an electrochemical detector (Coulochem III; ESA Inc.) with an applied potential of 100-150 mV. A mobile phase consisting of 20 mM sodium phosphate, 0.7 mM octane sulfonic acid, and 8% acetonitrile (pH 3.2) was pumped through a Reprosil-Pur C18 column (4 9 150 mm; Dr. Maisch HPLC GmbH), at a flow rate of 0.6 mLÁmin À1 . Signals from the detector were analyzed using the CLARITY software (Data Apex Ltd, Prague, Czech Republic). The concentrations were calculated based on standard solutions.

Luciferase reporter assay
The wild-type pT81/3xDRE and mutant pT81/CDEF plasmids were a kind gift from Peter A. M€ unzel (Department of Toxicology, University of T€ ubingen, Germany). The pT81/3xDRE reporter construct was generated using the motif of DRE3 from the CYP enhancer region fused in triplicate tandem array into pT81Luc as previously described [25]. Cells were cotransfected with the pT81based reporter constructs and pRL-TK (Promega, Madison, WI, USA) using Lipofectamine Plus reagent (Invitrogen) at 20% density. The cells were treated with dox, KYN (50 lM), KYNA (150 nM), or AHR inhibitor (6 lM; CH-223191) 24 h after transfection, and gene reporter assays were carried out with the Dual Glo Luciferase assay kit (Promega) 96 h after transfection (48 h compound/dox treatment). Renilla Luciferase activity served as a measure of transfection efficiency.

Immunofluorescence microscopy
Cells were fixed with 4% paraformaldehyde in PBS for 10 min at room temperature. Cells were permeabilized with 0.3% Triton TM X-100 in PBS with 5% goat serum (Dako, Agilent Technologies, Santa Clara, CA, USA) for 30 min. Subsequently, the primary rabbit anti-AHR antibody (Abcam) was added in 0.1% Triton TM X-100/1% bovine serum albumin (BSA)-PBS with 1% goat serum overnight at 4°C. The secondary goat anti-rabbit Alexa Fluor 488 antibody (Life Technologies) was diluted in 0.1% Triton TM X-100 in 2% BSA-PBS with 1% goat serum and added for 30 min at room temperature. Cells were mounted with Vectashield mounting medium containing 4 0 ,6-diamidino-2-

Statistical analysis
Results are shown as representative images or as means AE SD of at least three independent experiments. If not stated otherwise, data were analyzed using the unpaired t-test with Welch's correction or with the one-sample t-test using the PRISM 5 for Windows (version 5.02) statistical software (GraphPad Prism Software Inc., La Jolla, CA, USA). Data shown in graphical format represent the means (AESD) and a P-value of ≤ 0.05 is considered statistically significant.

EWS-FLI1 knockdown activates tryptophan metabolism via the KYN pathway
The addition of dox to A673sh cells induces a knockdown of EF (Fig. 1B) (Fig. 1C [3]). Therefore, we interrogated the pathway by measuring upand downstream metabolites of KYN (Fig. 2), a readout metabolite of TDO2 activity. Cells were incubated with fresh growth medium (including antibiotics and 10% FCS as described in 'Materials and Methods') in absence or presence of dox for 48 and 72 h and media from EWS-FLI1-high (EFH) and EWS-FLI1-low (EFL) cells were collected and subjected to HPLC and mass spectrometry (MS) analyses for the metabolites TRP, KYN, KYNA, and 3-HK. In EFL cell supernatants, TRP levels were significantly reduced after 72 h ( Fig. 2A). KYN was hardly detectable in the media of EFH cells, but was highly enriched (up to 40-fold) upon EF depletion (Fig. 2B). Although EFH cells produced small amounts of the downstream metabolite KYNA, the levels were highly elevated after EF knockdown, most prominently at 72 h of dox treatment (Fig. 2C). KYNA is known to inhibit the proliferation of various cancer cell lines [29,30] and also acts as a neuroprotective agent in the central nervous system [31]. Up to 72 h of incubation, 3-HK increased in the supernatants of A673sh cells under both conditions, but slightly faster upon EF silencing (Fig. 2D). The metabolite measurements suggest that in the absence of EF, TRP breakdown was initiated and led to the accumulation of KYN and KYNA, which are both known AHR ligands [12,32].

TDO2 is responsible for tryptophan uptake in EWS-FLI1-low cells
In order to verify whether TDO2 is required for TRP uptake in EF silenced A673sh cells, we investigated TRP consumption after silencing TDO2 with two distinct shRNA constructs (sh1-TDO2, sh2-TDO2). One day post transfection, fresh full medium with puromycin (for selection) and with/without dox (for EF silencing) was added to the cells and kept throughout 72 h. Subsequently, medium was tested for TRP and KYN abundance (Fig. 3A,B). Both constructs led to a similar reduction in TDO2 mRNA and protein levels of 70-90% (Fig. 3C,D). Cells harboring the sh-scrambled control shRNA showed a strong decrease in TRP levels upon EF knockdown. However, in sh-TDO2 transfected cells, TRP levels in the supernatant of EFL cells remained comparable to those of EFH cells (Fig. 3A), indicating that TDO2 activity is the main cause for TRP consumption after EF knockdown. Similarly, knockdown of TDO2 prohibited KYN accumulation associated with EF silencing (Fig. 3B). As IDO1 is not expressed in EFH or EFL A673sh cells (Fig. S1), these data confirm that KYN levels are regulated by TDO2 in A673sh cells in the absence of EF.

Functional activation of AHR is regulated by EWS-FLI1
Since we have demonstrated that KYN and KYNA are significantly up-regulated in the culture medium of EF silenced A673sh cells, we wanted to elucidate functional consequences of these enriched metabolites. As KYN and KYNA are both ligands of the AHR [12,32], we speculated that ligand activation of AHR signaling might be efficiently suppressed in the presence of EF and only initiated in its absence. In an inactive state the AHR rests in a cytosolic multiprotein complex including the heat shock protein 90 [33]. . Thus, the subcellular localization of the AHR transcription factor depends on ligand binding. To investigate whether AHR is present in the nucleus of EFL cells, we separated the cytoplasmic from the nuclear cellular fractions and immunoblotted them with anti-AHR antibody to visualize endogenous AHR expression (Fig. 4A). Nuclear localization of EF and its depletion upon dox treatment served as control. AHR protein was found in the cytoplasm and, to a much lesser extent, in the nucleus of EFH cells (Fig. 4A). However, upon EF silencing, AHR strongly increased in the nuclear fraction with the highest level of induction at 72 h of EF depletion. This finding is strong evidence for successful AHR activation and subsequent translocation to the nucleus in EFL cells. In concordance with western Blot analysis, immunofluorescence microscopy including quantification of nuclear AHR signal intensity supported these results (Fig. 4B).
Although already present in the nucleus in EFH cells, a strong increase in nuclear AHR staining was observed after dox treatment. To further confirm the influence of EF on the activation of AHR, we performed luciferase reporter assays. After functional stimulation of DREs on the reporter vector, luciferase activity can serve as readout (Fig. 4C). EFL cells displayed significantly elevated luciferase activity which was lost in the presence of AHR inhibitor CH-223191. Consistent with this finding, mutation of the DRE in pCDEF abolished reporter activity under EFL conditions (Fig. 4C). To follow up on the hypothesis that AHR gets activated after EF knockdown due to higher KYN and KYNA levels, the metabolites were added under EFH and EFL conditions. The concentrations of KYN and KYNA in the experiments were chosen according to their abundance in EFL conditions from MS analysis. Addition of KYN and KYNA led to increased DRE reporter activity. Consistent with the reporter assay results, EF silencing drastically up-regulated endogenous AHR-target genes IL8, IL6, IL1B, FAM65B, TUFT1, CYP1B1, and CYP1A1. Knocking down TDO2 in EFL conditions significantly decreased this up-regulation in AHR-target mRNA expression for most of the genes (IL8, P = 0.0027; IL6, P = 0.0001; IL1B, P = 0.012; FAM65B, P = 0.014; and TUFT1, P = 0.039) compared to sh-scrambled EFL (Fig. 4D), but not for CYP1B1 or CYP1A1 (not shown). Collectively, these data suggest that EF represses AHR activity in A673sh cells. The receptor gets activated once EF is down-regulated, which is most probably accomplished via initiation of TRP breakdown followed by an enrichment of KYN and KYNA.

Discussion
With the advent of sensitive single-cell gene expression analysis methods, it is becoming increasingly clear that stochastic variation in oncogene expression levels exist between individual cells of a tumor. This was recently demonstrated for FUS-DDIT3 positive myxoid liposarcoma [36] and discussed for EF in ES at the 2nd European Ewing Sarcoma Research Summit [37]. Here, such stochastic variations in EF expression may translate into differential metastatic behavior, since EF-low cells were demonstrated to have drastically increased migratory and metastatic potential [38]. In this study, we have used RNAi induced EF silencing as an experimental approach to study the phenotype of A673sh cells under EF-low conditions. We report activation of TDO2 resulting in TRP degradation and KYN/KYNA dependent activation of AHR signaling and speculate that TRP degradation, KYN/KYNA accumulation and AHR activation contribute to the survival of A673sh cells in an autocrine manner under conditions when EF expression is low.
TRP degradation in cancers has mostly been attributed to the activity of IDO1 in cancer cells and tumordraining lymph nodes [39]. Alternatively, in the absence of IDO1, TDO2 can overtake the constitutive TRP breakdown and KYN production in some human cancers and specialized myeloid cells [14]. Several reports point out that IDO1 activity and IDO1 pathway play key roles in regulating immune evasion by tumors [40]. IDO1 is expressed in several human cancers such as malignant melanoma, ovarian cancer, and colorectal cancer [41,42]. In the tumor microenvironment, IDO1-mediated TRP deficiency leads to the induction of a stress response which finally results in cell-cycle arrest of T-cells [43], differentiation of T regulatory cells and an immunosuppressive environment [44]. Additionally, the TRP degradation product KYN stimulates tumor promoting immune tolerance via activation of the AHR pathway [40,45]. Thus, IDO1 inhibition seems a promising strategy for cancer treatment and some antitumor effects have already been reported for human endometrial carcinoma and murine glioma [46][47][48].
However, IDO1 is not expressed in various tested ES cell lines (including A673) independently of EF levels, unless interferon (IFN) c treated (Fig. S1) [49]. Instead, we find TDO2 as the exclusive TRP degrading Both, AHR and ARNT are expressed in EFH and EFL cells [3], but nuclear localization of AHR was predominantly observed under EFL conditions (Fig. 4B,C). Expression of TDO2 as well as genes with AHR binding sites such as IL8, IL1B, IL6, FAM65B, TUFT1, CYP1B1, and CYP1A1 [3,50] were low in the presence of EF in vitro (A673sh), but up-regulated after EF silencing [3]. Interestingly, CYP1A1 is predominantly involved in detoxification, whereas CYP1B1 is required for metabolic activation in favor of tumor initiation [51]. In the AHR expressing state of EFL cells, CYP1B1 is expressed to a very high extent compared to CYP1A1 [3,50], probably a result of metabolic activation and proliferation. This speculation might be counterintuitive considering the fact that depleting ES cells of EF has been associated with growth inhibition and G1 cell-cycle arrest [52,53]. However, although a large proportion of EFL cells enters cell-cycle arrest [53], cells do not stop growing in vitro, which might again involve an alternative survival strategy that can partially be explained by AHR activation. Microarray data for several human tumors, among them ES, revealed a correlation of TDO2 expression with the expression of CYP1B1 [14], suggesting that some ES tumors express TDO2, which might have therapeutic potential. IL-8, a pro-inflammatory cytokine, is another AHRactivated target gene [54] and mRNA strikingly increases in EFL cells, but can be down-regulated by silencing TDO2 (Fig. 4D). IL-8 is involved in processes such as chemotaxis of target cells to the site of inflammation, stimulation of phagocytosis, and release of TNF-related apoptosis-inducing ligand (TRAIL) [55]. Depending on the combination of stimuli, IL-8 might also exert anti-inflammatory signals [54]. Intriguingly, in a concordant cytokine array, IL-8 and IL-6 were identified as up-regulated upon low expression of EF, and functional analysis showed IL-6-mediated phosphorylation of signal transducer and activator of transcription 3 (STAT3) [56]. Phosphorylated activated STAT3 is associated with tumor progression via favoring cell survival and proliferation [57]. IL-6 is one of AHR's target genes [14] and we can now speculate that ligand-activated AHR functions as a mediator in this pathway.
To a certain extent, protumorigenic properties of TDO2 are mediated via the microenvironmental accumulation of breakdown products like KYN, KYNA, and their binding to AHR [58]. The endogenous levels of KYN and KYNA in A673sh cells are in a high nanomolar to micromolar range which is in concordance with what has been reported to be sufficient for AHR activation [14,59], and results in AHR translocation to the nucleus and activation. In vivo oncogenic potential of constitutively active AHR was described for hepatocarcinoma [60] and stomach tumor [61] progression, and in vitro the AHR has been reported to be overexpressed in several cancers, including lung carcinoma, gastric carcinoma, and medulloblastoma [62]. In general, there is strong evidence of AHR activation being involved in tumor initiation, promotion and progression. Although AHR was discovered for its implication in detoxification, it is also involved in the activation of pro-carcinogens causing DNA adduct formation [63] and favoring ligand-mediated cell-cycle progression [64]. Intriguingly, AHR even mediates anti-apoptotic effects, but only in a ligand-dependent manner, as shown in Myc transgenic mice [65], in the promotion of ovarian tumors in rats [66], and cultured cell lines [67]. However, AHR activation can stimulate anti-or protumorigenic pathways dependent on the cellular background or tumor context [68,69]. In general, ligand-activated AHR is supposedly involved in carcinogenesis and tumor development, whereas the sole expression of AHR is mostly interconnected with low tumor burden or tumor suppression [21].
In summary, this is the first report to show ligandstimulated AHR activation in A673sh cells, caused by TDO2 activity and TRP breakdown products KYN and KYNA. In cells with low EF expression, this metabolic pathway can be switched on and might represent a novel alternative route of tumor cell survival, at least in the specific case of A673sh cells. It should be noted, however, that out of five ES cell lines studied by us, only A673sh cells induced TDO2 after EF knockdown, while IDO1 was consistently induced by IFNc. However, analysis of primary human ES mRNA expression data sets reveal considerable variation in TDO2 levels, correlated with CYP1B1 as a sign of putative AHR activation. Thus, our finding, so far restricted to A673sh cells, may be of relevance to a subset of primary ES and/or subsets of primary tumor cells within a given tumor. Since the development of small molecule inhibitors against TDO2 [70], targeting strategies have been discussed for many types of cancers where active TRP catabolism maintains the immunosuppressive environment [58]. In the future, closer investigation of the tumor microenvironment from TDO2-positive ES tumors could shed light onto metabolic alterations, potential AHR activation and, thus, help to identify those patients who might potentially profit from TDO2 directed therapy.