Eight- to 10-week-old male C57BL/6 mice were obtained from Charles River Breeding Laboratories. IDO1-deficient mice (Ido1^–/–) were purchased from The Jackson Laboratory. B6.129-Ahrtm1Bra/J Ahr-deficient (Ahr^–/–) mice were kindly supplied by B. Stockinger (MRC National Institute for Medical Research, London, United Kingdom). All in vivo studies were in compliance with national (Italian Parliament DL 116/92) and Perugia University Animal Care and Use Committee guidelines, and the overall study was approved by the Bioethics Committee of the University of Perugia.

The mouse fibroblast line, SYF, was obtained from ATCC, and cells were cultured in DMEM (Gibco) supplemented with 10% FBS at 37°C. This cell line was transfected to over-express c-Src and AhR and used for immunoprecipitation assays and as a source of nuclear extracts, as described below. Mouse embryonic fibroblasts (MEFs) were obtained according to the guidelines of the University of Perugia Ethical Committee and the European Communities Council Directive 2010/63/EU. MEFs and Ahr^–/– MEFs were prepared as described (14). Briefly, pregnant mice were sacrificed at 13.5 days post-coitum (dpc) by cervical dislocation. The embryos were separated from placenta and membranes and were placed in 10 cm culture dishes in sterile phosphate-buffered saline. Then, the liver, heart, and brain were removed and discarded. The remaining part of each embryo was washed and minced with cool razor blades and incubated 20 min at 37°C with trypsin-EDTA (500 mg L^–1). The minced tissues were chopped by repeated pipetting, and then the cell suspension was plated on 10 cm tissue culture dishes, and DMEM medium (10 ml) containing 10% FBS (Euroclone) was added. Electroporation was used to transfect MEFs.

Dendritic cells were isolated from bone marrows of C57BL/6 mice by crushing tibias, femurs, and hips in MACS buffer as previously described (15). Extracted cells were filtered through a 40 μm cell strainer and centrifuged at 300 × g for 5 min. Cells were re-suspended in red blood cell lysis buffer to remove erythrocytes. For DC differentiation, BM cells were cultured at a density of 2 × 10⁶/ml in culture media (IMDM from Gibco) conditioned with 5% Flt-3l for 9 days at 37°C with 7% CO₂. At the end of the culture, 25–30 million/mouse total DCs were obtained. Purification of cDCs was performed by MACS column (Miltenyi) and biotin antibodies (Biolegend). Total DCs were incubated with biotin mouse monoclonal antibodies against B220 (a marker of pDCs and B cells). After this, cells were incubated with MagniSort Streptavidin Negative Selection Beads (Thermo Fisher) followed by depletion of B220⁺ cells. We collected the B220^– cell fraction and used this to verify cell purity. cDCs were stained with fluorescent antibodies, identified as B220^– CD3^– and MHC-II⁺ CD11c⁺ CD24⁺ CD172⁺ by cytofluorimetric analysis by LSRFortessa (BD BioSciences), and analyzed by flowJo data analysis software. Cell purity was more than 90%. Cells (1 × 10⁶/ml) were primed with 250 ng/ml LPS (055:B5 Sigma-Aldrich) overnight before treating with II° LPS (1 μg/ml) or 50 μM L-kynurenine (Sigma-Aldrich) for 24 h.

Purified DCs were re-suspended at 1 × 10⁶/ml in fresh media in the presence or absence of LPS and L-kynurenine for a total of 36 h; supernatants were collected and analyzed for TGF-β by ELISA according to the manufacturer’s instructions (R&D system). DCs treated with LPS with either LPS or L-kynurenine were analyzed by ELISA for IL-10 and TGF-β contents according to the manufacturer’s instructions (R&D system) (16).

Real-time PCR (for mouse Ido1, Ido2, and Tdo2 expressions) analyses were carried out as described (17) using the specific primers listed in Supplementary Table S1. In all figures depicting RT-PCR data, bars represent the ratio of the relevant gene to β-actin-encoding gene expression, as determined by the relative quantification method (ΔΔCT; means ± SD of triplicate determinations).

SYF cells expressing c-Src kinase and LPS-primed cDCs were stimulated with L-kynurenine at 50 μM for different lengths of time (0.5, 1, and 2 h). Nuclear and cytoplasmic fractions were prepared from cells lysed on ice with Buffer N (15 mM Tris–HCl, pH 7.5, 15 mM NaCl, 60 mM KCl, 5 mM MgCl₂, 25 mM sucrose, 0.6% Non-idet P-40, 1 mM DTT, 2 mM Na₃VO₄). Lysates were then immunoprecipitated by means of sheep polyclonal antibody recognizing AhR, previously complexed with G Dynabeads (Invitrogen). Alternatively, cytosolic and nuclear lysates were run directly on SDS/PAGE.

IDO1 and pIDO1 expressions were investigated in cDCs by immunoblot with a rabbit monoclonal anti-mouse IDO1 antibody (cv152) or a rabbit polyclonal antibody to the phosphorylated immunoreceptor tyrosine-based inhibitory motifs (ITIM) motif of IDO1, respectively, both raised in our laboratory (18). c-Src and its phosphorylated form were revealed by specific anti-Src and -pSrc antibodies (Tyr416; Cell Signaling Technology). Anti-AhR antibody was from R&D system. Anti-β-actin and β-tubulin antibodies (Sigma-Aldrich) were used as a normalizer. Anti-lamininB antibody (Thermo Fisher) was used as a nuclear extract control. Whole-cell extracts, immunoprecipitates, and nuclear and cytosolic extracts were denatured in Laemmli sample buffer at 95°C for 5 min. Samples were run on a gel and transferred onto a nitrocellulose membrane (Bio-Rad). Immunoblots were blocked in TBS containing 3% BSA or 5% non-fat milk and 0.1% Tween 20 at room temperature for 1 h and then incubated with primary antibody overnight at 4°C. After washing, membranes were incubated with goat anti-rabbit, goat anti-mouse IgG or anti-sheep IgG conjugated to horseradish peroxidase (Jackson ImmunoResearch Laboratories, Inc) for 1 h at room temperature in 5% non-fat milk and 0.1% Tween 20 TBS. Then, membranes were washed and developed with ECL BioRad.

LPS-unprimed and LPS-primed cDCs treated or not with L-kynurenine, as described above, were cultured with CD4⁺ T cells isolated from the spleens of Foxp3^YFP mice. Briefly, CD4⁺ T cells were isolated via cell separation by two steps. A first step was meant to enrich the CD3⁺ cell fraction, using biotin mouse monoclonal antibodies against B220 (a marker of pDCs and B cells) and CD11c (a marker of DCs). After this step, cells were incubated with MagniSort Streptavidin Negative Selection Beads (Thermo Fisher) followed by depletion of B220⁺CD11c⁺. B220^– CD11c^– collected cell fractions were incubated with magnetic anti-mouse CD4 beads (Miltenyi Biotec) to select CD4⁺ T cells. The purity of CD4⁺ T cells was verified by FACS analysis by cell staining with anti-B220, anti-CD3, anti-CD11c, anti-CD4, and anti-CD8 specific antibodies. CD4⁺ T cells (2 × 10⁵) were activated with 5 μg/ml anti-CD3 mAb (clone OKT3) and co-cultured for 4 days with unprimed and primed cDCs (5 × 10⁴) treated or not with L-kynurenine for 36 h (19). Treg cells were evaluated by FACS, analyzing the induction of Foxp3 expressing YFP.

In the induction of primary endotoxemia, mice of the different genotypes (WT, Ahr^–/–, and Ido1^–/–) were randomly grouped (10 per group) and injected intravenously with 1 × 10⁶ cDCs, primed or not with LPS and treated or not with L-kynurenine, as described above. After 48 h, mice were injected intraperitoneally with a lethal dose of LPS (40 mg/kg). Vehicle-treated groups (no cDCs) were used as controls. Animals were monitored daily for 1 week for mortality or the presence of signs of moribundity, including lack of responsiveness to manual stimulation, immobility, and inability to eat or drink.

For histology, morphological analysis of paraffin-embedded lung sections (4 μm) from mice treated as above included stained with H&E at 48 h after challenge and examination by light microscopy, using representative specimens from one of three experiments.

For animal studies, randomization consisted of generating a random permutation of a sequence and selecting random numbers. On assessing outcomes involving scores, the investigator was totally blinded to group allocation. The log-rank test was used for paired data analyses of Kaplan–Meier survival curves. All in vitro determinations are means ± SD from three independent experiments and were evaluated by two-way and one-way ANOVA. GraphPad Prism version 6.0 (San Diego, CA, United States) was used for all analyses and graph preparation.

Unlike a single LPS stimulation, repeated exposure in vitro of splenic cDCs to endotoxin significantly induces IDO1 protein expression and L-kynurenine production, the main IDO1 enzymatic product. Such conditioned cDCs protect mice from lethal endotoxemia, and this effect requires that the cDCs be competent for IDO1 and TGF-β production (17). We here examined whether L-kynurenine could substitute for the second LPS challenge in mimicking an endotoxin-tolerant state in cDCs. We preliminarily set up cultures of bone marrow-derived cDCs (Supplementary Figure S1A). Then, IDO1 induction was assessed in these cells after priming with LPS overnight, followed by treatment with LPS or L-kynurenine for an additional 24 h, and unprimed cDCs were used as control. Priming with LPS alone modestly induced IDO1 at both transcript and protein levels; however, sequential treatment with LPS and L-kynurenine significantly enhanced this effect. In LPS-primed cDCs derived from AhR-deficient mice, IDO1 induction by L-kynurenine as a second stimulus was instead lost (Figures 1A,B). Under the same conditions, the expressions of the IDO1 paralog, Ido2, and of Tdo2 were not significantly affected by the treatment with LPS and L-kynurenine. Of note, Ido2 expression was induced by LPS alone on first exposure but not by a subsequent exposure to L-kynurenine or LPS as a repeated stimulus (Supplementary Figure S1B). Treatment of cDCs with L-kynurenine in the absence of prior LPS did not affect Ido1 expression (Figure 1A), implying that early TLR4-dependent events are necessary for initiating the loop whereby L-kynurenine will later reinforce AhR-dependent Ido1 transcription. The transition from endotoxin susceptibility to tolerance in cDCs is marked by an increased TGF-b secretion (3). We found that in unprimed cDCs, L-kynurenine alone treatment would increase TGF-β production according to a slower kinetic pattern than when used after LPS pre-conditioning. The ideal conditions for TGF-β production for the former involved removing L-kynurenine by extensive washing, followed by additional culturing for 48 h in medium alone (Figure 1C). Yet, production of the cytokine was strongly enhanced by previous LPS conditioning (Figure 1C). As expected, TGF-β production was unaffected by LPS + L-kynurenine treatment in cDCs derived from Ahr^–/– and Ido1^–/– mice.

Based on the finding that L-kynurenine binds AhR on the Gln377 residue, and in order to validate the ability of L-kynurenine to induce IDO1 via AhR in cDCs, we reconstituted AhR-deficient cDCs with a vector encoding the mutated form of AhR where Gln at position 377 is replaced by an Ala residue (Q377A) as described (3). The engineered cDCs were primed with LPS and then treated with L-kynurenine. Cells, either reconstituted with empty vector or the Q377A mutant, failed to express IDO1 (Figures 1D,E) and to produce TGF-β (Figure 1F). While confirming the binding mode of L-kynurenine to the Gln377 domain of the PAS-B domain of AhR, these results showed a crucial requirement for the Gln377 residue of AhR in L-kynurenine-mediated induction of IDO1 in LPS-primed cDCs, and for reprogramming those cells toward a tolerogenic phenotype. Overall, these data suggest that, in vitro, externally added L-kynurenine contributes to reprogramming gene expression—including Ido1’s—in endotoxin tolerance via effects involving AhR.

Phosphorylation of a specific domain in IDO1 is involved in its non-enzymatic function leading to reprogramming gene expression and the induction of a stable regulatory phenotype in splenic plasmacytoid DCs that will produce less IL-6 and more TGF-β in response to TLR signaling (20). We investigated whether IDO1 phosphorylation occurs in LPS-primed, IDO1-competent cDCs upon L-kynurenine exposure. We found that phosphorylation of IDO1 occurs after 30 min of L-kynurenine treatment in wild type cDCs but not in AhR-deficient cells (Figure 2A). The inhibition of c-Src—the most widely represented kinase in mouse cDCs (13)—with PP2 (but not its inactive analog PP3) negated IDO1 phosphorylation in LPS-primed cDCs exposed to L-kynurenine (Figure 2A).

It has been previously demonstrated that several AhR ligands can promote the activation of c-Src (21). We thus assessed c-Src activity by measuring the phosphorylation status of the kinase in LPS-primed cDCs upon challenge with L-kynurenine. Western blot analysis showed that c-Src activity increased at 5 and 15 min of L-kynurenine exposure (Figure 2B), an effect that was contingent on AhR activation in that it was lost in AhR-deficient cDCs. Moreover, we found that induction of c-Src activity by L-kynurenine is necessary for TGF-β production (Figure 2C). Overall, these data indicated that L-kynurenine can mimic the effects of a second LPS challenge in inducing an endotoxin-tolerant state. In particular, L-kynurenine promotes the activation of c-Src and the subsequent phosphorylation of IDO1 in AhR-competent cDCs.

The c-Src protein may exist in at least two different conformations (namely, inactive and active), whose relative stability determines the overall activity of the enzyme (22). Post-translational modifications, as well as binding of regulatory substrates and ligands, may alter this equilibrium by favoring one conformational population over the other and thus change the overall kinase activity. Several lines of evidence have shown that, upon tetrachlorodibenzo-p-dioxin (TCDD) binding, the kinase is released from the AhR complex and becomes activated (21). We thus investigated whether c-Src kinase activation mediated by L-kynurenine would require the dissociation from the AhR complex. SYF cells (i.e., a fibroblast cell line that does not express the ubiquitous c-Src, c-Fyn, and c-Yes kinases) were reconstituted with a vector coding for c-Src and exposed to L-kynurenine for different lengths of time. By co-immunoprecipitation assays, we found that AhR interacts with c-Src kinase and that L-kynurenine significantly reduces this association (Figure 3A). Moreover, by analyzing the cellular fractions of cells exposed to L-kynurenine, we observed that AhR localizes to the nucleus, while Src is mostly present in the cytosol in its active state (Figure 3B).

Co-immunoprecipitation and cellular fractionation studies confirmed that these molecular events (i.e., Src disjunction from the complex and AhR shuttling to the nucleus) occur in LPS-primed cDCs exposed to L-kynurenine as well (Figures 3C,D). Overall, these data demonstrated that L-kynurenine can indirectly regulate c-Src kinase activity. By binding AhR, L-kynurenine promotes the receptor’s conformational changes required for the release of c-Src, which then becomes activated and capable of phosphorylating downstream target proteins.

IDO1⁺ cDCs are known to be involved in the regulation of hyperinflammatory conditions via the generation of regulatory T cells (23). In particular, cDC co-treatment with LPS and L-kynurenine promotes TGF-β release in an AhR- and IDO1-dependent manner (Figure 2C) that is essential for Treg cell differentiation in vitro and in vivo.

To evaluate if the combined effects of L-kynurenine and LPS promoted functional regulatory cDCs, those cells—either unprimed or primed with LP—were incubated with L-kynurenine and cultured with naïve CD4⁺CD25^– T cells, after removing the stimuli by extensive washing. L-Kynurenine, either alone or in combination with LPS, significantly increased LAP expression on the T-cell surface relative to control samples (Figures 4A,B). By contrast, when the cDCs had been derived from AhR- and IDO1-deficient mice, this effect was abolished (Figures 4A,B). Foxp3 expression was evaluated in the same co-cultures. We failed to observe any differences between samples (data not shown). Moreover, the primed cDCs further treated with L-kynurenine decreased TNFα production as well as increased IL-10 secretion by the T cells in AhR- and IDO1-dependent fashion (Figure 4C). Overall, these findings validate the hypothesis that L-kynurenine progressively suppresses T CD4⁺ cell effector function, favoring instead the emergence of a regulatory T-cell phenotype.

Because cDCs stimulated twice in vitro with LPS protect mice from a primary, otherwise lethal LPS challenge (3), we investigated whether the combined effects of LPS and L-kynurenine treatment in vitro would confer a protective potential on those cells in vivo. We established an adoptive transfer model using Flt-3L-induced, BM-derived cDCs (Supplementary Figure S1A). LPS-primed or unprimed cDCs, treated or not with L-kynurenine, were transferred i.v., into naïve mice to be challenged at 48 h with a lethal i.p., dose of LPS (Figure 5A). We found that all vehicle-treated control animals died within 48–72 h after LPS challenge. Among mice treated with differentially treated DCs, we found that as many as 80% mice receiving LPS + L-kynurenine-treated DCs survived challenge over 200 h (Figure 5B). To further assess the effects of cDCs in vivo, groups of mice were sacrificed at 48 h of LPS challenge, and lung injury—a typical feature of septic shock and endotoxemia—was sought for on gross pathology. Interestingly, lungs of mice receiving primed cDCs further conditioned by L-kynurenine were characterized by reduced inflammatory cell infiltration and vascular congestion relative to control or to the other DC treated groups (Figure 5C). Based on in vitro evidence of AhR and IDO1 dependence for pharmacologic induction of endotoxin tolerance by LPS and L-kynurenine in cDCs, the same experiment was performed using LPS-primed or unprimed cDCs, treated or not with L-kynurenine from AhR- and IDO1-deficient mice. The protective effect of the double (LPS + L-kynurenine) pulsed cDCs on recipient mice was lost when the transferred cells were not competent for AhR and IDO1 (Figure 5D). Thus, successful transfer of endotoxin tolerance by fully tolerogenic DCs requires that AhR and IDO1 be fully functional.