The present study was conducted with local regional ethics committee approval, and written informed consent for publication of this study was obtained from all patients. Samples included serum collected for the purposes of research and excess liver biopsy tissue from diagnostic procedures. The mean age of patients was 58 years, with an age range between 35 and 66 years. All samples were anonymized. To discount the variable effects due to gender difference, all patients with PBC and all healthy controls included in the study were female.

The H69 immortalized biliary epithelial cell (iBEC) line was created and obtained from Grubman et al (29) (Tufts University School of Medicine, Boston, MA, USA) from human intrahepatic biliary epithelial cells and was used for the current study. These cells exhibit characteristics of normal human biliary epithelium, with expression of cytokeratin (CK)7 and 19. The iBECs were cultured in 75-cm² flasks, at 37°C in an atmosphere containing 5% CO₂, in a 3:1 mix of Dulbecco's modified Eagle's medium and Nutrient Mixture F12 Ham (Sigma-Aldrich; Merck KGaA, Darmstadt, Germany) supplemented with 1.8×10⁻⁴ M adenine (24.3 mg/l), 2×10⁻⁹ M triiodothyronine (1.345 µg/l), 5.5×10⁻⁶ M epinephrine (1.0 mg/l) (all from Sigma-Aldrich; Merck KGaA), ITS-X supplement (10 mg/l insulin, 5.5 g/l transferrin, 2.0 g/l ethanolamine and 6.7 µg/l sodium selenite; Gibco; Thermo Fisher Scientific, Inc., Waltham, MA, USA), 1 µM hydrocortisone solution (362.46 µg/l), 10% heat-inactivated fetal bovine serum and 100 U/ml each of penicillin and streptomycin (all from Sigma-Aldrich; Merck KGaA).

H69 cells were treated with 20 ng/ml IFN-γ. Once confluent (>80%), the cells were then left post- IFN-γ treatment for 0, 6, 12, 24, 48, 72, 96 and 120 h prior to RNA isolation or harvesting of supernatants.

H69 cells were incubated with an ALK5 inhibitor (SB-505124; catalog no. 3263; TOCRIS Bioscience, Bristol, UK) at an optimal concentration of 1 µΜ (30) for 1 h prior to stimulation with IFN-γ for 6 and 24 h.

RNA isolation was performed according to the method developed by Chomczynski and Sacchi (31). TRIzol reagent (Sigma-Aldrich; Merck KGaA) was used according to the manufacturer's instructions. RNA was quantified and quality-assessed using a NanoDrop spectrophotometer (NanoDrop ND-1000; Thermo Fisher Scientific, Logan, UT, USA). Isolated RNA was reverse transcribed to cDNA using the AffinityScript Multi Temperature cDNA Synthesis kit (Agilent Technologies, Inc., Santa Clara, CA, USA) according to the manufacturer's protocol. The qPCR experiments in this project utilized TaqMan chemistry and were performed in a MicroAmp Optical 96-well plate (both from Applied Biosystems; Thermo Fisher Scientific, Inc.) on a StepOnePlus Real-Time PCR machine (Applied Biosystems; Thermo Fisher Scientific, Inc.), with each well containing 1 µl TaqMan primer-probe (all primers were exon-spanning), 1 µl cDNA, 10 µl Master Mix and 8 µl RNase-free water. Primers used were for GAPDH (catalog no. Hs02758991_g1), IDO-1 (catalog no. Hs00984148_m1), IDO-2 (catalog no. Hs01589373_m1), TGF-β1 (catalog no. Hs00998133_m1), TGF-β2 (catalog no. Hs00234244_m1) and were all purchased from Applied Biosystems; Thermo Fisher Scientific, Inc. The thermocycling conditions were as follows: 50°C for 1 min, 95°C for 10 min, followed by 40 cycles of 95°C for 15 sec and 60°C for 1 min. Data were normalized to the expression of GAPDH mRNA using the 2^−ΔΔCq method (32) and analyzed using REST 2009 software (Qiagen GmbH, Hilden, Germany).

Formalin-fixed, paraffin-embedded liver biopsy sections from patients were deparaffinized in xylene and rehydrated. Antigen retrieval was conducted by pressure-cooking (pressure, 15PSI; temperature, 121°C) for 1 min in Tris/EDTA (pH 8.0). To block endogenous peroxidase activity, 3% H₂O₂ in methanol was utilized and a biotin block was performed according to the manufacturer's protocol (Avidin/Biotin Blocking kit; catalog no. SP-2001; Vector Laboratories, Ltd., Peterborough, UK). Sections were incubated either with anti-IDO antibody (catalog no. ab55305; dilution, 1:100) or anti-Foxp3 antibody (catalog no. ab20034; dilution, 1:100) (both from Abcam, Cambridge, UK) in normal swine serum at 4°C, in a humidified chamber overnight. Following washing with 0.1% Tween-20 in PBS, sections were incubated with biotinylated anti-mouse secondary antibody (catalog no. BA-9200; dilution, 1:200; Vector Laboratories, Ltd.) for 1 h at room temperature in a humidified chamber. Sections were then stained with Vectastain peroxidase ABC kit (Vector Laboratories, Ltd.), according to the manufacturer's protocol. Staining was developed with 3,3′-diaminobenzidine (Sigma-Aldrich; Merck KGaA) and counterstained with Mayer's hematoxylin (Dako; Aglient Technologies, Inc.), according to the manufacturers' protocols. Sections stained with no primary antibody was used as a negative control. Visualization was performed using an optical microscope (Provis AX-70; Olympus Corporation, Tokyo, Japan).

HPLC was performed on a Shimadzu LC-10ADVP system (Shimadzu Corporation, Kyoto, Japan). The separation was performed isocratically using a Cronusil-S ODS1 (250×4.6-mm) column (SMI-LabHut, Ltd., Gloucester, UK) at 25°C at a flow rate of 1 ml/min. The mobile phase contained 15 mM sodium acetate (pH 5.0) and acetonitrile [94:6% (v/v)]. Following injection (25 µl), the eluted Kyn was monitored at 360 nm between 0 and 8.5 min, and after 8.5 min, tryptophan was monitored at 278 nm. A working-top standard solution containing 10 µM Kyn and 100 µM tryptophan was made in mobile phase and further diluted to create a standard curve. Standard or sample (200 µl) was mixed with 50 µl sulfosalicylic acid, incubated on ice for 15 min and clarified by centrifugation for 15 min at 18,000 × g at 4°C. Supernatant (25 µl) was injected into the HPLC system. Kyn typically eluted at ~6.5 min and tryptophan at 9.4 min. Standard curve and sample quantification was based on peak area using LCSolution software (version 1.11SP1; Shimadzu Corporation).

Kyn was measured spectrophotometrically, as described previously (33). A total of 75 µl 30% trichloroacetic acid was added to 100 µl of the culture supernatant, vortexed and centrifuged at 10,000 × g for 5 min at 4°C. Following this, a 75-µl volume of the supernatant was then added to an equal volume of Ehrlich's reagent (100 mg dimethylbenzaldehyde and 5 ml glacial acetic acid) in a microtiter plate well. Optical density was measured at 492-nm using a Multiskan MS microplate reader (Thermo Labsystems, Santa Rosa, CA, USA). A standard curve of defined Kyn concentration (0–100 µM) permitted analysis of unknowns.

3-Hydroxykynurenine (3-HK), Kyn and 3-hydroxyanthranilic acid (3-HAA) (all from Sigma-Aldrich; Merck KGaA) were dissolved in RPMI-1640 medium and added to CD4⁺ T cells in culture at concentrations between 0 and 100 µM (pH 7.5–8.5).

Isolation of human CD4⁺ T cells was performed using Rosette Sep (Stemcell Technologies, Vancouver, BC, Canada) human CD4⁺ T-cell enrichment cocktail at 50 µl/ml of whole blood (500 µl/10 ml), according to the manufacturer's protocol. Briefly, the whole blood was taken from a healthy donor and incubated with the cocktail for 20 min at room temperature. The sample was diluted with an equal volume of PBS (plus 2% FBS) and mixed gently. This was layered on top of the density medium and centrifuged at 1,200 × g for 20 min at room temperature. The enriched cells were isolated from the density medium-plasma interface and washed with PBS with 2% FBS. Isolated CD4⁺ T cells (1×10⁵ cells/well) were stimulated with anti-human CD3 and CD28 beads at the ratio of one bead per cell (Dynabeads Human T-activator; Invitrogen; Thermo Fisher Scientific, Inc.) in the presence of tryptophan metabolites, 3-HK, 3-HAA and L-Kyn (Sigma-Aldrich; Merck KGaA) at 50 µΜ concentration, with 10 ng/ml TGF-β1 (R&D Systems Europe, Ltd., Abingdon, UK) used as a control; cells were incubated for 5 days.

Intracellular staining for Foxp3 was performed using a Foxp3 intracellular staining kit (eBioscience; Thermo Fisher Scientific, Inc.), according to the manufacturer's protocol. This kit has been formulated and optimized to stain intracellular antigen Foxp3 (clone: PCH101; allophycocyanin, catalog no. 17-4776-42; eBioscience™, Invitrogen; Thermo Fisher Scientific, Inc.). Cells were stained for intracellular staining. Briefly, cells suspended in FACS tubes were pelleted by centrifugation at 500 × g for 5 min at 4°C. Following removal of the supernatant, cells were fixed in a fixation/permeabilization buffer (1-part fixation/permeabilization concentrate to 3 parts fixation/permeabilization diluent) for 30 min at 4°C. Cells were then washed and centrifuged in permeabilization buffer at 500 × g for 5 min (at 4°C) followed by 1-h incubation with the primary antibody. Finally, cells were washed twice in permeabilization buffer prior to re-suspension in FACS buffer (PBS with 2% FBS) and analyzed using a FACSCanto II flow cytometer and BD FACSDiva™ software, version 7.0 (both from BD Biosciences, Franklin Lakes, NJ, USA).

Statistical analysis was performed using GraphPad Prism 5.0 (GraphPad Software, Inc., La Jolla, CA, USA). Unless otherwise noted, data represent the mean value of an experiment conducted in triplicate ± standard error of the mean. For comparison between 2 groups, unpaired two-tailed Students t-tests or Mann Whitney unpaired t-test were performed. For ≥3 groups, a one-way analysis of variance test was performed and, if significant, a post-hoc Tukey's analysis was performed. P<0.05 was considered to indicate a statistically significant difference.

A prior study established an in vitro model using iBECs to investigate the link between senescence and epithelial cell de-differentiation (30). The present study assessed the expression of mRNA encoding IDO-1, IDO-2, TGF-β1 and TGF-β2 following stimulation of iBEC with IFN-γ, for 6 and 24 h. Fig. 1 demonstrates the significant increase (P<0.05) in the expression of IDO-1 (6,644-fold) with IFN-γ at 6 h. The results shown were normalized against the housekeeping gene GAPDH. High levels of IDO-1 expression were sustained at 24 h in IFN-γ treated cells; whereas no detectable expression of IDO-1 was found in untreated iBECs (data not shown).

Increase in IDO-2 expression was observed in cells treated with IFN-γ (P<0.01). Expression of TGF-β2 was also upregulated (12-fold) upon treatment with IFN-γ at 6 h, but expression was reduced at 24 h. There was no significant increase in expression of TGF-β1 following stimulation with IFN-γ; however, TGF-β2 expression was significantly increased.

To determine whether IDO-1 expression is dependent on the TGF-β pathway, cells were incubated with an ALK5 inhibitor (SB-505124) at an optimal concentration of 1 µΜ (30) for 1 h prior to stimulation with IFN-γ for 6 and 24 h. The expression of IDO-1 was not significantly different in the treated or the untreated group (P>0.05, data not shown). These findings suggest that IDO-1 expression in iBEC is inducible, with a marked increase following IFN-γ stimulation, but independent of the TGF-β pathway.

Certain studies have revealed an inconsistency between the expression of IDO and its enzymatic activity (34,35); thus, the present study examined the enzymatic activity of IDO in iBECs. A colorimetric assay was used to measure the levels of Kyn in the culture supernatant (Fig. 2). Cells were stimulated with IFN-γ for up to 120 h. The supernatant of unstimulated iBEC produced no Kyn. The difference in Kyn levels between the IFN-γ-stimulated and the unstimulated iBECs was significant (P<0.05).

IDO activity was measured by assessing the ratio of the enzyme substrate tryptophan and its product Kyn (Fig. 3). Higher IDO activity was observed in the sera of PBC patients (n=47) compared with the healthy controls (n=24). Statistically significant differences were found between the two groups when measuring the area and height of peaks as determined by HPLC (Fig. 3).

To assess IDO expression in PBC, an immunohistochemical assay was performed. Expression of IDO in the liver tissue was examined in samples from healthy controls (n=5) and patients with PBC (n=5). No expression was observed in the cholangiocytes of healthy controls (Fig. 4A). Fig. 4B-D reveals IDO expression in late-stage PBC. Discrete staining is revealed in the BECs (Fig. 4B). IDO expression was increased in periportal hepatocytes, in conjunction with notable interface hepatitis (Fig. 4C and D). Sections from matched PBC patients stained for Foxp3 are presented in Fig. 4E and F; Foxp3⁺ cells were morphologically small lymphocytes. There was an evident increase in the number of inflammatory cells, particularly those with interface hepatitis and IDO expression.

IDO may induce T-regs, at least in part, through certain tryptophan metabolites. To test this, human CD4⁺ T cells were stimulated with anti-CD3/CD28 beads in the presence of 50 µM of the Kyns, L-Kyn, 3-HK and 3-HAA for 5 days. T-reg development was assessed by analyzing Foxp3 expression with flow cytometry.

There was no significant difference in the proportion of CD4⁺ cells expressing Foxp3 in the presence of L-Kyn and 3-HAA compared with the control after 5 days (Fig. 5); however, 3-HK was significantly upregulated in the fraction of CD4⁺ cells expressing Foxp3.