Female wild-type C57BL/6 mice of 6-wks of age were purchased from Charles River Breeding Laboratories Italia (Milan, Italy). All animals were housed and fed under specific pathogen free conditions in the animal facility of the University of Perugia. All in vivo studies were compliant with National (Italian Animal Welfare Assurance A-3143-01) and Perugia University Animal Care and Use Committee guidelines. DMSO was purchased from Sigma-Aldrich (MO, USA). The H-2Db-restricted HY peptide (WMHHNMDLI) — containing the immunodominant epitope of male mouse specific minor transplantation antigen — was synthesized by BioFab Research (Rome, Italy). Epacadostat was purchased from Selleckchem (Huston, TX, USA).

Cells of P1.HTR, a highly transfectable clonal variant of mouse mastocytoma P815 (14), were transfected by electroporation with a plasmid construct coding for mouse wild type (wt) (P1.IDO1), or mutant IDO1 lacking the catalytic activity, namely H350A (P1.H350A), as described in (11, 15). The stable transfectant cell line was obtained by puromycin selection and cultured in Iscove’s Modified Dulbecco’s Medium (Gibco, Invitrogen CA, USA) supplemented with 10% FCS (Gibco, Invitrogen CA, USA), 1mM glutamine (Gibco, Invitrogen CA, USA) and penicillin/streptomycin (Gibco, Invitrogen CA, USA) at 37°C in a humidified 7% CO₂ incubator. In every cellular assay, the cell line was used at a passage number not exceeding the 10^th. Dose-response curve for extrapolating the IC₅₀ (half-maximal inhibitory concentration) of epacadostat in P1.IDO1 cells was built in a cellular assay, as described in (16) by incubating cells with 2-fold serial dilutions of the compound. After the incubation, supernatants of cell cultures were recovered and the kynurenine concentration was detected by HPLC. The control was represented by cells incubated with an equivalent volume of DMSO, the vehicle in which epacadostat has been solubilized. Every cell assay was conducted in triplicate and repeated three times.

Detection of kynurenine concentrations was performed by using a Perkin Elmer, series 200 HPLC instrument (MA, USA). A Kinetex^® C18 column (250×4.6 mm, 5 μm, 100 Å; Phenomenex, USA), maintained at the temperature of 25°C and pressure of 1800 PSI, was used. A sample volume of 300 μl was injected and eluted by a mobile phase containing 10mM NaH₂PO₄ pH 3.0 (99%) and methanol (1%) (Sigma-Aldrich, MO, USA), with a flow rate of 1 ml/min. Kynurenine was detected at 360 nm by an UV detector. The software TURBOCHROM 4 was used for evaluating the concentration of kynurenine in samples by mean of a calibration curve. The detection limit of the analysis was 0.05 μM.

P1.HTR cells (25 × 10³ cells/well) were seeded into a flat-bottom 96-well plate and incubated overnight at 37°C with scalar concentrations of epacadostat or DMSO as a control. Then, medium containing MTT (Sigma Aldrich, St. Louis, MO, USA) was added and cells were incubated for 4 h at 37°C. Subsequently, solubilization buffer (SDS 10% in HCl 0.01 M) was added, the plate was incubated overnight at 37°C, and absorbance at 570 nm was measured using a UV/visible spectrophotometer (TECAN, Thermo Fisher Scientific, Waltham, MA, USA) to construct a graph of cell viability. The assay was performed in triplicate for each concentration.

Constructs used for FRET analysis were SHP-1-GFP (Ptpn6 Mouse Tagged ORF Clone, Origene) and IDO1-RFP (mouse Ido1 were cloned in pTagRFP-N vector, Evrogen). Both constructs were confirmed by DNA sequencing. HeLa cells (4 x 10⁵ cells/96-mm dish) were plated on coverslips for 24 h and then co-transfected with SHP-1-GFP and IDO1-RFP plasmids DNA using Lipofectamine 3000 (Thermo Fisher Scientific), according to the manufacturer’s protocol. After 16 h, transfected cells were incubated for 4 h with epacadostat 1 μM or DMSO as control. Prior to the FRET imaging, cells were washed twice with PBS 1X, treated with PFA 4%, and then mounted with Prolong-DAPI (ThermoFisher). Samples were imaged using a Nikon inverted microscope (60X magnification) using a GFP/RFP FRET pair and processed using ImageJ software (National Institute of Health). Acquired fluorescent images were background-subtracted, and multiple regions of interest (ROIs) were selected for quantitative data analysis (~10–20 ROIs per cell, and 6–8 cells per condition were averaged). FRET efficiency values presented as a pseudo-color scale by ImageJ. Cells transfected with either SHP-1-GFP or IDO1-RFP were used to calculate the coefficient A (CoA) and B (CoB), that correct for the amount of acceptor bleedthrough in the absence of a donor and that of donor bleedthrough in the absence of an acceptor, respectively.FRET efficiency was calculated as follows:

For immunoprecipitation of IDO1, a rabbit monoclonal anti-mouse IDO1 antibody (CV152) obtained in our laboratory (11) was used. Antibodies used to analyze immunoprecipitated samples were SH-PTP1 (clone C-19 Cat# sc-287) and SH-PTP2 (clone B-1 Cat #sc-7384), both from Santa Cruz Biotechnology (TX, USA), and PI3K p85 (rabbit mAb clone 19H8 Cat #42575) from Cell Signaling Technology (MA, USA). Anti-Rab5 (rabbit mAb clone C8B1 Cat# 3547) were from Cell Signaling Technology (MA, USA).

P1.IDO1 and P1.H350A cells were pre-incubated for 4 h with epacadostat 1 μM or DMSO as control, before assessing the phosphatase activity associated with the IDO1 protein. For each condition, the IDO1 protein was immunoprecipitated from 6 x 10⁶ cells lysed in 400 μl of Lysis Buffer (50 mM Tris-HCl pH 7.5, 150 mM NaCl, 1% Nonidet P-40 and a cocktail of protease inhibitors). The IDO1 protein was immunoprecipitated overnight at 4°C by a specific monoclonal rabbit antibody raised in our lab, followed by addition of Protein G-Sepharose (Sigma-Aldrich, MO, USA) for 2 h. The phosphatase cactivity associated with immunoprecipitated IDO1 was assessed by a specific kit purchased from Promega (WI, USA), according to the manufacturer’s instructions.

Endosomes were isolated using a Minute Endosome Isolation and Cell Fractionation Kit (Invent Biotechnologies, Inc., MN, USA) according to the manufacturer’s instructions. Briefly, 3×10⁷ cells were collected, washed with cold PBS and resuspended in buffer A. After incubation, the cell suspension was transferred to a filter cartridge and centrifuged twice. Next, the pellet was resuspended and centrifuged twice. After centrifugation, supernatant was transferred to a fresh tube, mixed with buffer B, incubated overnight at 4°C, and centrifuged. The pellet resulting from the centrifugation contains isolated endosomes, whereas the supernatant represents the cytosolic fraction. Both fractions were analyzed by western blotting.

All purification procedures and purities of splenic pDCs and CD8⁻ DCs (hereafter referred to as cDCs, for conventional DCs) were conducted as previously described in (17, 18). Briefly, splenic DCs were purified using CD11c MicroBeads (Miltenyi Biotec) in the presence of EDTA to disrupt DC-T cell complexes. Cells were >99% CD11c⁺, >99% MHC I-A⁺, >98% B7–2⁺,<0.1% CD3⁺, and appeared to consist of 90-95% CD8^-, 5–10% CD8⁺, and 1–5% B220⁺ cells. DC populations were further fractionated according to CD8 expression to obtain purified CD8⁺ DCs and CD8^- DCs by means of CD8α MicroBeads (Miltenyi Biotec). After cell fractionation, the recovered CD8^- cells typically contained<0.5% contaminating CD8⁺ DCs, whereas the CD8⁺ fraction was made up of >95% CD8⁺ DCs. Less than 3% CD11c⁺ cells expressed the mPDCA-1 marker. A skin test assay was used for measurements of major histocompatibility complex class I-restricted delayed-type hypersensitivity (DTH) responses to the HY peptide in C57BL/6 female recipient mice (19).

For the antigen-specific immunization of female recipient mice was used the HY peptide because is a male histocompatibility antigen that causes rejection of male skin grafts by female recipients (20).

For in vivo immunization, 3×10⁵ peptide-loaded cDCs, combined with a minority fraction (5%) of peptide-loaded pDCs, were injected subcutaneously into recipient mice. After two weeks, mice were challenged intrafootpad with the HY peptide in the absence of DCs (experimental footpad) or with vehicle alone (control counterpart). Twenty-four hours later, mice were sacrificed, footpads were removed and the change in the weight of the peptide-injected footpad over that of the vehicle-injected counterpart (i.e., internal control) was measured. A significant increase in the mean weight of the experimental over the control footpads will then demonstrate and quantify any delayed type, antigen-specific hypersensitivity response. Representative pictures of the footpads from a positive and negative DTH response have been already published in (11). Results are represented as the mean weight ± SD of the experimental footpads over that of control, vehicle-injected counterparts. Measurements were performed in a blinded fashion. The minority cell fraction, constituted by pDCs, was left untreated or treated overnight with DMSO or epacadostat 1 μM.

Expression of Ido1 and Tgfb1 genes was analyzed by Real-Time PCR, as described [21] using the specific primers reported in (16). Real-time PCR assays were performed using Stratagene Mx3005P (Agilent Technologies, CA, USA). mRNA levels were normalized to the expression of the Gapdh housekeeping gene. The Ct number was measured using the MxPro-Mx3005P software (Agilent Technologies, CA, USA). Values were expressed as the ratio of Gapdh normalized transcript expression of compound treated-pDCs to Gapdh normalized transcript expression.

pDCs were harvested by centrifugation at 1400 r.p.m for 5 minutes, followed by surface staining for 30 minutes at 4°C in 1x PBS + 0.5% BSA and 2mM EDTA (MACS buffer). Subsequently, surface antigens were stained in MACS buffer for 30 minutes at 4°C. Samples were fixed by incubation in 1x PBS + 1% (v/v) paraformaldehyde. Cells were analyzed with FACS Fortessa and data was analyzed with FlowJo software (Tree Star).

All analyses were performed using Prism version 8.0.1 (GraphPad Software). Data usually met normality and were analyzed by two-tailed unpaired Student’s t-test when two samples were under comparison or ANOVA followed by post-hoc Bonferroni’s test, when three or more samples were under comparison. A P value less than 0.05 was considered significant. Overall results, obtained by at least three replicates per experimental parameter, were shown as mean ± SD.

Before investigating the effect of epacadostat on IDO1-mediated signaling, we tested the ability of this compound to inhibit the enzymatic activity of IDO1 in the mastocytoma cell line P1.HTR stably expressing the mouse IDO1 protein (P1.IDO1). By treating the cells with different concentration of the compound, we could confirm that epacadostat can inhibit the conversion of tryptophan into kynurenine also in P1.IDO1. The obtained concentration curve provided a cellular IC50 of 54.46 nM ± 11.18 nM, a value very similar to that reported in the literature (i.e. IC50 88 nM (21)) ( Figure 1A ). In order to exclude a cytotoxic effect of the compound, we first performed a MTT assay that measures cellular metabolic activity as an indicator of cell viability. By using seven 2-fold dilutions in the 1–30 μM concentration range, we did not observe any cytotoxic effect of epacadostat for all the concentrations tested ( Figure 1B and Supplementary Figure 1 ).

Whether epacadostat could affect also the non-enzymatic function of IDO1 was never investigated. Because the first event which activates the IDO1-mediated signaling pathway is the association with tyrosine phosphatases SHPs (11, 12, 18, 22), we examined whether this compound could have a role in the IDO1-SHPs binding. HeLa cells transiently co-expressing IDO1–RFP and SHP-1–GFP were treated with epacadostat or the vehicle alone to determine if the compound could affect the association between these two proteins by fluorescence resonance energy transfer (FRET) microscopy. We found that IDO1 interacts with SHP-1 ( Figure 2A ) and SHP-2 ( Supplementary Figure 2 ) in HeLa cells in absence of any stimulus, consistently with previous reports in P1.IDO1 cells (11). Surprisingly, epacadostat enhanced the association between these proteins ( Figures 2A, B and Supplementary Figures 2A, B ). This finding was confirmed by the evaluation of the ability of IDO1 immunoprecipitates to remove a phosphate group from a tyrosine-phosphorylated substrate, as described (11, 13, 15). Specifically, we quantified the amount of free phosphate produced by IDO1 immunoprecipitated from P1.IDO1 cells untreated, treated with epacadostat or the vehicle as a control. As previously demonstrated (11), we observed a basal phosphatase activity associated with IDO1 immunoprecipitates obtained from P1.IDO1 cells. However, preconditioning of the cells with epacadostat for 4 h generated a significantly increased level of free phosphate, as compared to the control ( Figure 2C ). Western blot analysis of the coimmunoprecipitates and quantitative analysis of molecular partners confirmed that both the SHPs enhance their binding to IDO1 in the presence of epacadostat ( Figures 2D, E and Supplementary Figure 3 ). Interestingly, by using our previously characterized IDO1 mutant lacking the histidine residue required for catalytic activity (i.e. H350A (11)), we found that IDO1 can bind the same epacadostat, which still improves the first event in the IDO1-signaling functions ( Figures 2F–H and Supplementary Figure 4 ).

Collectively, these data indicate that in the mastocytoma tumor cell stably expressing IDO1, epacadostat can effectively inhibit the catalytic activity of this enzyme. However, it can also improve the ability of IDO1 to interact with SHP phosphatases, previously described as the first molecular event triggering the signaling function of IDO1 (11).

In the literature, IDO1 is described as a cytoplasmic enzyme. However, we recently demonstrated that under conditions that favor the signaling rather than the catalytic activity of IDO1, the protein can shift from the cytosol to early endosomes (EEs) where it plays a signalling activity (12, 13). We thus investigated whether epacadostat may regulate IDO1 intracellular trafficking. A mandatory event for the anchoring of IDO1 to EEs is its direct interaction with class IA PI3Ks p85 via a YxxM motif (12, 13). Therefore, we co-immunoprecipitated IDO1 from P1.IDO1 cell lysates pretreated with epacadostat or the vehicle alone, as control. We found that the p85 subunit of PI3K efficiently co-precipitated with IDO1 and the association was enhanced in epacadostat-treated cells ( Figures 3A, B and Supplementary Figure 5 ). To further investigate the effect of epacadostat on the cellular topology of IDO1, we isolated EEs from P1.IDO1 cells using a subcellular fractionation approach. In accordance with our previous data (13), the immunoblotting assay verified that in a tumor transfectant cell the IDO1 protein is not just confined to cytosol but can also associate with EEs. However, this intracellular localization of IDO1 was significantly favored by the presence of epacadostat ( Figures 3C, D and Supplementary Figure 6 ).

As a whole, these data indicate that epacadostat promotes the molecular events and the intracellular trafficking required for IDO1-mediated signaling pathway, by increasing the interaction of IDO1 with PI3K and making it move toward the EEs.

The IDO1 signaling functions were first described in pDCs and cDCs, immune cells characterized by a plastic phenotype prompt to respond to any environmental signal (11, 18). In particular, pDCs ( Supplementary Figure 7 ) represent the most flexible DC subset and can shift from an immunostimulatory toward an immunosuppressive program by sensing the extracellular milieu (23). In order to analyze the effect of epacadostat on the immunoregulary function of pDCs, we investigated the in vivo priming ability of these antigen presenting cells after in vitro pre-conditioning with epacadostat. Specifically, we sensitized female mice with the H-2Db–restricted HY peptide presented by a combination of immunogenic cDCs and a minority fraction of pDCs preincubated in vitro overnight with epacadostat or DMSO, as control. Two weeks later, we evaluated the antigen-specific reactivity through an intrafootpad challenge of the HY antigen, according to an established protocol for measuring the induction of immune reactivity versus tolerance (19). As expected, sensitization of mice with a combination of immunogenic cDCs and vehicle-treated pDCs elicited a positive response against the HY antigen in mice. In contrast, epacadostat-pretreated pDCs abrogated the immunoadjuvant response elicited by cDCs, suggesting that this compound could induce in pDCs a tolerogenic phenotype ( Figure 4A and Supplementary Figures 8A, B ).

We previously demonstrated that in a microenvironment dominated by the immunoregulatory cytokine TGF-β, the activated PI3K-IDO1-SHPs axis promotes the nuclear translocation of the noncanonical pathway of NF-κB, leading to the de novo expression of Ido1 and Tgfb1 genes. This mechanism establishes a positive immunoregulatory circuitry responsible for a long-term immunoregulatory phenotype of the DCs and pDCs (11–13, 18). Thus, we investigated whether the epacadostat-dependent induction of the tolerogenic phenotype in pDCs could be due to the signaling pathway activated by IDO1. To this purpose, we conditioned pDCs with the compound or the vehicle alone and we then analyzed the gene transcription of key molecules involved in the IDO1-mediated self-maintaining regulatory circuitry, namely Ido1 and Tgfb1 ( Figure 4B ). We did detect induction of the above genes after stimulation with epacadostat, confirming that the compound favors the activation of a positive immunoregulatory circuitry in pDCs ( Figure 5 ).