Blockade of the AHR restricts a Treg-macrophage suppressive axis induced by L-Kynurenine

Abstract

Tryptophan catabolism by the enzymes indoleamine 2,3-dioxygenase 1 and tryptophan 2,3-dioxygenase 2 (IDO/TDO) promotes immunosuppression across different cancer types. The tryptophan metabolite L-Kynurenine (Kyn) interacts with the ligand-activated transcription factor aryl hydrocarbon receptor (AHR) to drive the generation of Tregs and tolerogenic myeloid cells and PD-1 up-regulation in CD8⁺ T cells. Here, we show that the AHR pathway is selectively active in IDO/TDO-overexpressing tumors and is associated with resistance to immune checkpoint inhibitors. We demonstrate that IDO-Kyn-AHR-mediated immunosuppression depends on an interplay between Tregs and tumor-associated macrophages, which can be reversed by AHR inhibition. Selective AHR blockade delays progression in IDO/TDO-overexpressing tumors, and its efficacy is improved in combination with PD-1 blockade. Our findings suggest that blocking the AHR pathway in IDO/TDO expressing tumors would overcome the limitation of single IDO or TDO targeting agents and constitutes a personalized approach to immunotherapy, particularly in combination with immune checkpoint inhibitors.

Fig. 1. An active IDO/TDO-Kyn-AHR pathway associates with immune suppressive features in human cancers.

a Top: heat-map representing the gene-expression analysis of Kyn-degrading enzymes (KMO, KYNY, HAAO) and KD-Score (grayscale) in responsive (complete response/CR or partial response/PR) highlighted in black (n = 8), or non responsive (stable disease/SD or progressive disease/PD) highlighted in gray (n = 31), melanoma patients after PD-1 blockade. Bottom: bar graph with quantification of KD-Score. b ELISA quantification of L-Kynurenine in blood serum of melanoma patients (left) categorized as IDO^high or IDO^low (n = 4 IDO^high and n = 4 IDO^low) based on intracellular IDO staining by FACS of melanoma cell suspensions (right) (n = 8 IDO^high and n = 10 IDO^low). c mRNA of immunoregulatory markers by qRT-PCR analysis in IDO^high and IDO^low melanoma cell suspensions. d mRNA of AHR-target genes CYP1A1 and CYP1B1 by qRT-PCR analysis in IDO^high and IDO^low melanoma cell suspensions after treatment with selective AHR inhibitor KYN-101 for 24 h (n = 6 IDO^high and n = 6–7 IDO^low). e Distribution of log-transformed expression levels of AHR-related genes (TDO2, AHR, and CYP1B1) across six immune subtypes of cancer (C1: wound-healing, C2: IFN-γ dominant, C3: inflammatory, C4: lymphocyte-depleted, C5: immunologically quiet, C6: TGF-β dominant). Data plotted as box and whiskers with the median and limits within the 10–90% percentile. f Correlation analysis between Treg marker (FOXP3), myeloid-cell marker (MRC1/CD206), inhibitory checkpoint (PDCD1/PD-1), and AHR-related genes IDO1, TDO2, and CYP1B1 in TCGA RNAseq data of skin melanoma (SKCM), squamous lung (LUSC) and pancreatic adenocarcinoma (PDAC) analyzed by Spearman rank correlation. Data shown are represented as mean values ± SEM with two-tailed unpaired Student’s t test in (a–d), one-way ANOVA test with Tukey correction in (a) and Kruskal–Wallis with Dunn correction in (e). P value: *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Fig. 2. IDO-overexpressing melanomas present a dysfunctional myeloid-enriched immune landscape.

a Top: schematic representation of the experimental setup for the 3D collagen-fibrin gel killing assay and bottom, quantification of target cell killing by tumor-isolated CD8⁺ T cells or tumor-antigen (gp100)-specific CD8⁺ T cells (pmels) (Data is representative of two independent experiments). b Multiplex analysis of cytokines from ex vivo tumor-isolated CD8⁺ T cells stimulated with aCD3/CD28 beads for 48 h. c, flow cytometric analysis and quantification of M2-like TAMs (CD206⁺MHCII⁺ in CD11b⁺F4/80⁺) cell populations in B16^IDO or B16^WT tumors at day 14 post implantation (n = 10 (B16^WT) and n = 9 (B16^IDO)). Results are representative of three independent experiments. d Heat-map of mRNA expression of M1 and M2 markers and e, of AHR-target genes in TAMs as determined by RT-PCR, data were relative to GAPDH expression and z-score normalized (n = 3). f in vitro suppressive activity of TAMs purified from B16^IDO or B16^WT tumor-bearing mice at day 14 post implantation in CD8⁺ T cell activation state at a ratio of 1:2 CD8⁺ to TAMs (n = 3). g Tumor progression of B16^IDO or B16^WT implanted in mice treated with clodronate liposome (1 mg/mouse) (n = 8) or control (n = 8). Results are representative of two independent experiments. Data is represented as mean values ± SEM. Two-tailed unpaired Student’s t test in (c–e, g) and one-way ANOVA test with Tukey correction in (f). P value: *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Fig. 3. FoxP3⁺T cells co-operate with TAMs in promoting immune suppression of IDO + tumors.

a Flow cytometric analysis of FoxP3 expression in tumor-infiltrating CD4⁺ populations (left), and of Tregs (CD4⁺FoxP3⁺) (right) (n = 7 (B16^WT) and n = 8 (B16^IDO)). Results representative of three independent experiments. b Tumor progression of B16^WT (left) and B16^IDO (right) tumors implanted in FoxP3^DTR mice treated with PBS or DT (1 μg/mouse) (n = 5 per group). c FACS analysis of AHR expression (left) (n = 7 (B16^WT) and n = 8 (B16^IDO)) and mRNA analysis by qRT-PCR analysis of AHRR expression in B16^IDO and B16^WT tumor-isolated T-regs (right). d In vitro suppressive activity of tumor-isolated Tregs isolated from B16^IDO or B16^WT-bearing FoxP3^GFP mice. Left: Representative histograms of CD8⁺ T cell proliferation at a ratio of 1:1 (Effectors:Targets). Right: FACS quantification of T cell proliferation (CTV^low) (n = 3). e mRNA of Treg markers by qRT-PCR analysis in tumor-isolated Tregs from B16^WT and B16^IDO tumors (n = 3). f FACS analysis of TAMs (top) and M2-like myeloid populations (CD206^highCD11b⁺) (bottom) in B16^IDO tumors after Treg depletion in FoxP3^DTR mice injected with DT (1 μg/mouse) or PBS (B16^WTn = 3; B16^IDO PBS n = 4, DT n = 5 mice/group). g Correlation analysis between myeloid marker (MRC1/CD206) and Treg marker (FoxP3) in TCGA RNAseq data of of skin melanoma (SKCM), squamous lung (LUSC) and pancreatic adenocarcinoma (PDAC). h Correlation analysis in TCGA RNAseq data of IDO^high and IDO^low skin melanomas (SKCM) analyzed by Spearman rank correlation. Bottom: correlation analysis between frequency of Tregs (CD4⁺FoxP3⁺) and myeloid cells (CD11b⁺) in FACS analyzed IDO^high and IDO^low melanoma patients’ biopsies by Spearman rank correlation. Results in (b–f) are representative of two independent experiments. Data is represented as mean values ± SEM. Two-tailed unpaired Student’s t test was used when only two groups were compared, and one-way ANOVA was applied with comparison of more than two groups. P value: *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Fig. 4. Selective AHR inhibition reverses IDO/TDO-mediated tumor progression and improves the efficacy of PD-1 blockade.

a Top left: scheme figure with experimental approach of targeting multiple steps of Trp-catabolic pathway. Top right: therapy regimen (PO, QD). Bottom: tumor progression of orthotopically injected B16^WT, B16^IDO and B16^TDO tumors in mice treated with vehicle or inhibitors (Epacadostat/IDO inhibitor 300 mg/kg/d, 680C91/TDO inhibitor 20 mg/kg/d, CH-223191/AHR inhibitor 50 mg/kg/d). (n = 5 mice per group for two independent experiments). b FACS analysis of MHCII expression (n = 4 per group, n = 5 B16^IDO vehicle). c Quantification of frequency and representative plots of intratumor Tregs (CD4⁺FoxP3⁺CD25⁺/% of CD45⁺) and d, quantification of frequency and representative plots of Ki67⁺PD-1⁺ CD8⁺ T cells from B16^WT and B16^IDO-bearing mice treated with vehicle or AHR inhibitor (CH-223191) (n = 4 per group). e Top: therapy regimen. bottom: mean tumor size and overall survival of B16^IDO and B16^TDO tumor-bearing mice treated with AHR inhibitor (CH-223191), anti-PD-1 alone or in combination with AHR inhibitor (combo) (n = 5 B16^TDO and n = 10 B16^IDO). f Mean tumor size of B16^IDO and CT26 and g, overall survival of B16^IDO tumor-bearing mice treated with the optimized AHR inhibitor (KYN-101), anti-PD-1 alone or in combination with AHR inhibitor (combo) (n = 10 mice per group). Results are representative of two independent experiments. Data represented as mean values ± SEM. Two-tailed Student’s t test was used when only two groups were compared and log-rank (Mantel–Cox) test was used for survival comparison. P value: *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.