Counteracting tryptophan metabolism alterations as a new therapeutic strategy for rheumatoid arthritis

Abstract

Objectives: Alterations in tryptophan (Trp) metabolism have been reported in inflammatory diseases, including rheumatoid arthritis (RA). However, understanding whether these alterations participate in RA development and can be considered putative therapeutic targets remains undetermined.In this study, we combined quantitative Trp metabolomics in the serum from patients with RA and corrective administration of a recombinant enzyme in experimental arthritis to address this question.

Methods: Targeted quantitative Trp metabolomics was performed on the serum from 574 previously untreated patients with RA from the ESPOIR (Etude et Suivi des POlyarthrites Indifférenciées Récentes) cohort and 98 healthy subjects. A validation cohort involved 69 established patients with RA. Dosages were also done on the serum of collagen-induced arthritis (CIA) and collagen antibody-induced arthritis (CAIA) mice and controls. A proof-of-concept study evaluating the therapeutic potency of targeting the kynurenine pathway was performed in the CAIA model.

Results: Differential analysis revealed dramatic changes in Trp metabolite levels in patients with RA compared with healthy controls. Decreased levels of kynurenic (KYNA) and xanthurenic (XANA) acids and indole derivatives, as well as an increased level of quinolinic acid (QUIN), were found in the serum of patients with RA. They correlated positively with disease severity (assessed by both circulating biomarkers and disease activity scores) and negatively with quality-of-life scores. Similar profiles of kynurenine pathway metabolites were observed in the CAIA and CIA models. From a mechanistic perspective, we demonstrated that QUIN favours human fibroblast-like synoviocyte proliferation and affected their cellular metabolism, through inducing both mitochondrial respiration and glycolysis. Finally, systemic administration of the recombinant enzyme aminoadipate aminotransferase, responsible for the generation of XANA and KYNA, was protective in the CAIA model.

Conclusions: Altogether, our preclinical and clinical data indicate that alterations in the Trp metabolism play an active role in the pathogenesis of RA and could be considered as a new therapeutic avenue.

Keywords: antirheumatic agents; inflammation; rheumatoid arthritis.

Figure 1

Targeted tryptophan metabolomics analysis: Differential abundance between patients with RA (ESPOIR cohort) and healthy controls. Targeted tryptophan metabolomics analysis was conducted on the serum of 574 patients with early RA from the ESPOIR cohort and on 98 healthy subjects. (A) Differential analysis between RA and HS. (B) Ratio of QUIN over 3HAA in RA and HS serum. (C) Level of KYNA (pmol/ml) in RA and HS serum. (D) Ratio of XANA over 3H-Kyn. 3HAA, 3-hydroxyanthranilic acid; 3H-Kyn, 3-hydroxykynurenine; HS, healthy subject; KYNA, kynurenic acid; QUIN, quinolinic acid; RA, rheumatoid arthritis; XANA, xanthurenic acid.

Figure 2

Correlation between RA activity and Trp metabolites in serum. Pairwise correlation using serum samples from the ESPOIR cohort. Statistical significance was determined for all pairwise comparisons using Spearman test; only significant correlations (p value <0.05 after false discovery rate correction, q<0.2) are displayed. Cytokine were measured in serum. DAS28, Disease Activity Score-28 for Rheumatoid Arthritis; RA, rheumatoid arthritis; Trp, tryptophan.

Figure 3

Correlation between quality of life and Trp metabolites in serum. Pairwise correlation using serum samples from the ESPOIR cohort. Statistical significance was determined for all pairwise comparisons using Spearman test; only significant correlations (p value <0.05 after false discovery rate correction, q<0.2) EQ5D, standardised measure of health-related quality of life developed by the EuroQol Group; MHI5, Mental Health Inventory 5; SF36_phys, physical component of the SF36 score (The Short Form (36) Health Survey); SF36_psy, psychological component of the SF36 score.

Figure 4

Targeted tryptophan metabolomics analysis in collagen-induced arthritis (CIA). Targeted tryptophan metabolomics analysis was conducted on serum of CIA mice at day 35 and on control mice. (A) Study design of CIA. (B) Level of indole derivatives in the serum of CIA mice and controls. (C–E) Ratio of KYNA, XANA, QUIN over IDO metabolites. (F) Ratio of KYNA+XANA over QUIN+KYN in the serum of CIA mice and controls. (G) Correlation of Trp metabolites and clinical activity and circulating IL-6. Statistical significance was determined for all pairwise comparisons using Spearman test; only significant correlations (p value <0.05 after false discovery rate correction, q<0.2 are displayed). KYNA, kynurenic acid; QUIN, quinolinic acid; Trp, tryptophan; XANA, xanthurenic acid.

Figure 5

Kynurenine pathway (A) and AADAT level (B). Serum samples from the ESPOIR cohort compared with HS. (A) Kynurenine pathway showing AADAT involvement in KYNA and XANA production. (B) Level of AADAT measured by ELISA in human serum samples from the ESPOIR cohort. P value results from Mann-Whitney (****p<0.0001). AADAT, aminoadipate aminotransferase; HS, healthy subject; KYNA, kynurenic acid; XANA, xanthurenic acid.

Figure 6

AADAT protects from collagen-antibody induced arthritis (CAIA) in mice. (A) Clinical effect of AADAT administration in collagen antibody-induced arthritis (CAIA) mice. (B) Histological analysis confirms AADAT protective effects in CAIA on synovial inflammation (H&E) and cartilage lesions (safranin O staining). Data are the pool of two independent experiments, n=16 mice per group. Clinical score and oedema volume: p value results from analysis of variance followed by Sidak correction for multiple comparison (for each time point); Oedema volume: p value results from unpaired t-test; Histological score: p value results from Mann-Whitney. (C) Ratio of KYNA/Kyn, KYNA+XANA over Kyn+3H-Kyn and XANA/3H-Kyn in the serum of CAIA mice treated or not with AADAT, p value results from unpaired t-test. 3H-Kyn, 3-hydroxykynurenine; AADAT, aminoadipate aminotransferase; Kyn, kynurenine; KYNA, kynurenic acid; XANA, xanthurenic acid.

Figure 7

Effects of KYNA, XANA and QUIN on human synoviocyte metabolism. (A) Proliferation of hFLS after incubation with Trp metabolites at 1 µM in normal and inflammatory conditions. Control corresponds to hFLS proliferation without stimuli. (B) Cell cycle analysis by using DRAQ5 DNA intercalating agent in flow cytometry on hFLS. (C) Relative expression of IL-1β in hFLS. (D) Real-time oxygen consumption rates (OCR) and real-time extracellular acidification rates (ECAR) of incubated hFLS. Data are representative of three individuals. P value results from analysis of variance followed by Sidak correction for multiple comparisons (for each point). KYNA, kynurenic acid; hFLS, human fibroblast-like synoviocyte; QUIN, quinolinic acid; XANA, xanthurenic acid.