The human ovarian adenocarcinoma cell line SKOV-3, purchased from CLS Cell Lines Service GmbH, was grown in RPMI-1640 medium supplemented with 10% FCS and maintained in a humidified atmosphere at 37°C in 5% CO₂. Cells were routinely screened to confirm the absence of mycoplasma contamination. For cellular assays, SKOV-3 cells were seeded at the final concentration of 0.15 x 10⁶ cells/ml in 12-well plate, unless otherwise specified, and then exposed to epacadostat for the indicated times. Epacadostat was purchased from SelleckChem and was used at the final concentration of 1 µM unless otherwise indicated. Cycloheximide was obtained from Sigma-Aldrich and used at the final concentration of 50 µg/ml. For cycloheximide-chase assay, cells were pre-treated with the protein synthesis inhibitor for 1 hour prior the addition of epacadostat or vehicle. Cells were then incubated up to 8 and 16 hours for the short and long kinetics analysis, respectively. For washout experiments, SKOV-3 cells were pre-conditioned with epacadostat or vehicle for 16 hours at 37°C, and then washed twice with medium to completely remove the stimulus. Then, cells were exposed to cycloheximide and incubated for 4 or 8 hours (post-washout) to measure IDO1 protein expression and kynurenine concentration. DMSO, used as control, was purchased from Sigma-Aldrich. IDO1 knockdown experiments were performed by using Silencer^® Select human IDO1 siRNA (ID s7425) and Silencer^® Select Negative Control No.1 siRNA from Thermo Fischer Scientific, according to the manufacturer’s instructions. Briefly, SKOV-3 cells were incubated overnight at 37°C at the final concentration of 0.1 x 10⁶ cells/ml in 6-well plate. The day after, cells were transfected with 75 pmol of IDO1-specific or negative control siRNAs, by using Lipofectamine™ 3000 Transfection Reagent (Thermo Fischer Scientific). After 48 hours, IDO1-silenced SKOV-3 cells were used for the scratch wound healing assay and assessed for IDO1 silencing, by analyzing IDO1 protein expression and catalytic activity, as represented in the experimental workflow in Supplementary Figure 2A .

IDO1 tyrosine-phosphorylation and its interaction with SHP-2 were assessed by immunoprecipitation experiments with 2.5 x 10⁶ cells/sample. Cells were lysed in M-PER buffer (Thermo Fisher Scientific), containing protease and phosphatase inhibitor cocktails (Thermo Fischer Scientific), and a representative aliquot per sample was collected as a whole cell lysate (WCL) control. The remaining lysates were used for immunoprecipitation studies, performed following the manufacturer’s protocol (Thermo Fisher Scientific) and as previously described (14). For each sample, 12.5 μl of magnetic beads (Dynabeads Protein G, Thermo Fischer Scientific) were blocked with PBS containing 0.5% BSA (w/v) and conjugated overnight either with an anti-phospho-tyrosine (p-Tyr-1000) MultiMab™ Rabbit monoclonal antibodies mix (Cell Signaling Technology), or a mouse monoclonal anti-SH-PTP2 antibody (clone B-1, Santa Cruz Biotechnology), for IDO1 tyrosine-phosphorylation and SHP-2/IDO1 interaction analyses, respectively. Beads without the specific antibody were used for negative control samples. Antibody-conjugated beads were then incubated with cell lysates, overnight at 4°C, and then washed with washing buffer (25 mM citric acid, 50 mM dibasic sodium phosphate dodecahydrate, pH 5). Subsequently, the immuno-complexes were eluted with elution buffer (0.1 M sodium citrate dihydrate, pH 2-3) plus Laemmli buffer containing 2-mercaptoethanol and used for western blotting analysis.

IDO1 protein expression was analyzed by means of a rabbit monoclonal anti-human IDO1 antibody (D5J4E™, Cell Signaling Technology), whereas SHP-2 protein level was detected by the use of a mouse monoclonal anti-SH-PTP2 antibody (clone B-1, Santa Cruz Biotechnology). Moreover, a mouse monoclonal anti-β-tubulin antibody (clone AA2, Sigma-Aldrich) was used as normalizer. SKOV-3 cells were lysed in Laemmli buffer containing 2-mercaptoethanol and used for sodium dodecyl sulfate/polyacrylamide gel electrophoresis (SDS/PAGE). After incubation with the specific horseradish peroxidase-conjugated secondary antibodies (Sigma-Aldrich) and the addition of Clarity™ Western ECL (Enhanced ChemiLuminescence, Bio-Rad Laboratories) substrate, signals were detected by the use of ChemiDoc™ MP Imaging System (Bio-Rad Laboratories). Protein expression was quantified by densitometric analysis, using ImageLab Software (Bio-Rad Laboratories), as previously described (14). The enzymatic activity of IDO1 was evaluated in cell culture supernatants in terms of the ability of the enzyme to metabolize tryptophan into kynurenine, by HPLC analysis as previously described (20).

For gene expression analysis, SKOV-3 cells were treated with epacadostat for 24 hours and lysed in TRIzol reagent (Thermo Fischer Scientific) to isolate RNA, according to the manufacturer’s instructions. One µg of RNA was reverse-transcribed using QuantiTec Reverse Transcription Kit (Qiagen) and analyzed by Real-Time PCR, using Stratagene Mx3005P (Agilent Technologies). Specific primers were used to determine the expression of the human IDO1 gene (sense, 5’-TCACAGACCACAAGTCACAG-3’; antisense, 5’-GCAAGACCTTACGGACATCT-3’). Data were calculated as the ratio of the gene of interest to that of human β-actin expression (ACTB) (sense, 5’-CTCGTCGTCGACAACGGCT-3’; antisense, 5’-TCAGGGTGAGGATGCCTCTC -3’) by the relative quantification method (ΔΔCt). Values are presented as normalized transcript expression in samples relative to normalized transcript expression in control cells (1-fold).

The cell viability of SKOV-3 cells exposed to epacadostat was analyzed by MTT assay. Cells were seeded at the final concentration of 5 x 10³ cells/100 µl/well in a 96-well plate, and treated with epacadostat for 24, 48, and 72 hours. Vehicle-treated cells were used as control. At each time point, the stimulus was removed and cells were provided with 110 µl/well of complete medium containing MTT (Sigma-Aldrich) at the final concentration of 0.45 mg/ml. After 4 hours at 37°C, 100 µl/well of the solubilization buffer (SDS 10% in HCl 0.01 M) were added and cells were incubated overnight at 37°C. The day after, cell viability was determined by measuring the absorbance at 570 nm, by using a UV/visible spectrophotometer (TECAN, Thermo Fisher Scientific).

For the scratch wound healing assay, SKOV-3 cells were seeded at 0.2 x 10⁶ cells/ml in 6-well plate and cultured overnight to reach approximately over 80% of confluence. The day after, a 10-µl sterile pipette tip was used to make a scratch line on the monolayer of confluent cells, at the bottom of the well. Cells were washed twice to remove cellular debris and then treated with epacadostat or DMSO (as control). Cells were incubated at 37°C, in a humidified 5% CO₂ incubator, for 48 hours. Over this incubation time, the wound healings were continuously observed by using a Nikon Eclipse Ti inverted microscope (Nikon) with 10X magnification, and pictures were acquired every 4 hours. The area of the wound healing was determined by using the MRI Wound Healing tool (ImageJ Software, NIH) and the data were reported as percentage of the wound closure respect to time 0.

Anchorage-independent growth of SKOV-3 cells was studied by performing a soft agar colony formation assay, after conditioning cells with epacadostat for 16 hours. Firstly, the bottom of a 6-well plate was precoated with a mix (1:1 ratio) of 1% agarose solution and 2X complete RPMI-1640 medium (1.5 ml/well). Then, pre-conditioned cells were resuspended in 2X complete RPMI-1640 medium (containing epacadostat or vehicle) and 0.6% agarose solution (1:1 ratio) and plated (5 x 10³ cells/well) on the precoated 6-well plate (1.5 ml/well). The stimulus renewal was performed every other day, by the addition of 200 µl of complete medium (containing epacadostat or DMSO) over the agarose’s upper layer. After 21 days, colonies were imaged by using a Nikon Eclipse Ti inverted microscope (Nikon) with 40X magnification. The size of colonies was analyzed in 30 randomly selected visual fields in each well and measured using the NIS-Elements AR Software (Nikon). Subsequently, the colonies were stained by adding MTT, at the final concentration of 1 mg/ml, and incubated overnight at 37°C.

All analyses were performed using Prism version 10.0.3 (GraphPad Software). Data were shown as mean ± SD of at least three independent experiments. A P value less than 0.05 was considered significant. Data were analyzed by 2-tailed unpaired Student’s t-test or two-way ANOVA followed by post-hoc Bonferroni’s test when three or more samples were under comparison. The IC₅₀ and EC₅₀ were calculated by nonlinear regression analysis of raw data from the concentration-response curve, by applying the “log[inhibitor] vs response” and “log[agonist] vs response” equations, respectively. The negative logarithm of the IC₅₀ and EC₅₀ in unit of Molar was used to express the pIC₅₀ and pEC₅₀ values. The half-life (t_1/2) and the degradation speed (K) of IDO1 protein were calculated by nonlinear regression analysis of raw data, by using the “one-phase decay” equation.

Besides being unable to degrade tryptophan, the apo-IDO1 assumes a conformation prone for transducing an intracellular signal, as previously demonstrated in tumor cells ectopically expressing a loss-of-function mutant of IDO1 (namely, the apo-form) as well as in response to the IDO1 catalytic inhibitor epacadostat (13, 14, 18). To investigate the effect of epacadostat on endogenous IDO1, we resorted to the human ovarian cancer cell line SKOV-3 that constitutively expresses the highest level of IDO1 transcript among different human ovarian cancers ( Supplementary Figure 1A ) (21). Moreover, in SKOV-3 cells, IDO1 protein is present in a dynamic balance between the apo- and the holo-form (19). By treating cells with increasing concentration of the molecule, we confirmed that epacadostat inhibited the catalytic activity of IDO1 with a pIC₅₀ of 7.76 ± 0.06 ( Figure 1A ). Of note, we found that the protein level of IDO1 increased in the cells exposed to epacadostat with a pEC₅₀ of 7.66 ± 0.26 ( Figure 1B ), in the same concentrations range as pIC₅₀ for the catalytic inhibition, suggesting a dual effect of epacadostat on IDO1. Interestingly, the catalytic inhibitor linrodostat also exerts a similar dual effect on IDO1 protein (data not shown). Thus, to investigate whether epacadostat affected the transcript level of IDO1, SKOV-3 cells were treated with a fixed concentration of the molecule and the IDO1 mRNA level was evaluated. Based on the previous literature (18, 22), we selected a final concentration of 1 µM, that has proved to be not cytotoxic ( Supplementary Figure 1B ) and able to inhibit IDO1 catalytic activity ( Supplementary Figure 1C ) as well as increase IDO1 protein expression in SKOV-3 ( Supplementary Figure 1D ). Results from Real-Time PCR analysis demonstrated that the level of IDO1 was not significantly different in cells exposed to epacadostat and the control counterpart ( Figure 1C ), suggesting that post-translational regulatory mechanisms underlie the sustained IDO1 protein levels observed in SKOV-3 in response to epacadostat.

In B16 melanoma cells, the ectopic apo-IDO1 protein has a greater stability compared to the holo-IDO1 protein (14). Thus, we investigated whether epacadostat – by binding IDO1 – favored a protein conformation less prone to post-translational degradation. To this purpose, the kinetic of the intracellular protein turnover was analyzed by a cycloheximide-chase assay followed by immunoblotting. Results indicated that the protein expression of IDO1 remained more stable in SKOV-3 cells treated with epacadostat as compared to control cells ( Figure 2A ) and this effect lasted for at least 16 hours ( Figure 2B ). In particular, the IDO1 half-life is significantly longer in the presence of epacadostat (t_1/2 > 29.420 hours ± 10.64; degradation speed K, 0.025 hour⁻¹ ± 0.01) than that of the vehicle-treated cells (t_1/2 > 4.787 hours ± 1.63; degradation speed K, 0.154 hour⁻¹ ± 0.05) that resulted similar to the IDO1 turnover previously observed in other cell types (23, 24). Moreover, the stabilized IDO1 protein is still catalytically inactive ( Figure 2C ). Interestingly, the complete washout of epacadostat from SKOV-3 culture media – after 16 hours of cell pre-treatment – did not restore the conventional IDO1 protein turnover ( Figure 2D ), nor the catalytic activity of the enzyme ( Figure 2E ). Overall, these findings indicated that the pharmacological effect of epacadostat persists long after the initial exposure and prolongs the half-life of IDO1 protein in the SKOV-3 cells.

IDO1 is the prototype of a dynamic protein, as it acquires mutually exclusive conformations endowed with distinct functions, namely the enzymatic and the non-enzymatic forms (12). The phosphorylation of specific tyrosine residues in the ITIMs is the critical event to trigger the IDO1 non-enzymatic function (6, 10). Once phosphorylated, IDO1 interacts with different molecular partners, including the phosphatase SHP-2 (14, 18). Thus, we investigated whether epacadostat stabilized the conformation of IDO1 amenable of phosphorylation and interaction with SHP-2 phosphatase. Results from immunoprecipitation followed by immunoblotting analysis demonstrated that the tyrosine-phosphorylation and the association of IDO1 with SHP-2 significantly increased in SKOV-3 cells in response to epacadostat (1.5 and 9.8 folds, respectively) ( Figure 3A, B ).

SHP-2 is a well-recognized oncogene enabling a broad spectrum of human malignancies, including ovarian cancers (25–27). Thus, we analyzed the effect of epacadostat on SKOV-3 cells phenotype by measuring cellular motility and ability to growth in an anchorage-independent manner. In the wound healing assay, epacadostat accelerated the migratory capacity of SKOV-3 cells by promoting the wound closure faster than the vehicle-treated cells that did not close the wound in 48 hours ( Figure 3C ). Differently, IDO1 knockdown of SKOV-3 cells slowed their migratory capacity compared to IDO1-competent SKOV-3 cells. In fact, IDO1-silenced SKOV-3 cells achieved 50% wound closure after 24 hours compared to negative control siRNA-treated SKOV-3 cells, which only achieved wound closure after 12 hours ( Supplementary Figure 2 ), confirming a role of IDO1 protein in the migratory ability of SKOV-3 cells. In addition, epacadostat treatment increased the ability of SKOV-3 cells to form colonies in semisolid media ( Figure 3D ). Overall, these data suggested that epacadostat – although inhibits the catalytic activity of IDO1 – promotes the signal conformation of IDO1 that potentiates in vitro the tumorigenic phenotype of SKOV-3 cells.