Background/Purpose: A subset of human FoxP3+ regulatory T (Treg) cells expresses CD39 (Treg39+) and hydrolyses proinflammatory adenine nucleotides released at inflammatory foci, rendering the antiinflamatory adenosine. Methotrexate (MTX), inhibiting AICAR transformylase, enhances the extracellular release of adenine nucleotides and may cooperate with Treg39+ cells to control inflammation. Therefore we examined whether Treg cells may help predict the response to MTX independently of clinical parameters.

Methods: PBMCs from 82 ERA patients (duration < 24 weeks) and 82 healthy controls (HCs) were examined by cytometry to assess the frequency of CD4+CD25+CD127-FopxP3+ [total circulating Treg (cTreg)] and CD4+CD27+CD127-FoxP3+CD39+ Treg [circulating Treg39+ (cTreg39+)]. PBMCs from 40 patients were reexamined 12 months (12m) after initiating MTX. Disease activity was assessed by DAS28-ESR. High, moderate or low disease activity (HDA, MDA, LDA) were defined as a DAS28>5.1, >3.2 and ≤ 5.1, or DAS28 ≤ 3.2, respectively.

Results: As compared with HC, ERA patients showed an increased baseline (0 months, 0m) frequency (fr.) of cTreg with increased proportions of cTreg39+ cells. Patients who attained LDA 12m after initiating MTX had significantly higher 0m fr. of cTreg [logistic regression OR 7.18 (95% CI, 3.26-21.75); ROC AUC 0.93 (0.87-0.99), p< 0.0001] and cTreg39+ cells [OR 1.55 (1.29-2.10); AUC 0.97 (0.95-1.0), p< 0.0001], as compared with those who did not; multiple regression indicated that this was independent of age, 0m disease activity, RF or ACPA titres. The baseline cutoff fr. for 12m LDA was cTreg >7.8% [sensitivity (S) 0.82, specificity (Sp) 0.91%] and cTreg39+ > 42.0% (S 90.4, Sp 96.8%). We next analyzed separately patients who had presented with 0m-HDA (n=44), 0m-MDA (n=34) or 0m-LDA (n=4). The number of patients who attained 12m LDA was 23 (52%) in the 0m-HDA, 24 (71%) in the 0m-MDA and 4 (100%) in the 0m-LDA group. Interestingly, in either the 0m-HDA or the 0m-MDA group, the basal Treg and Treg39+ cell fr. significantly correlated with the ΔDAS28, and were good predictors of 12m-LDA [Treg OR for 0m-HDA = 2.60 (1.50-5.35), AUC 0.81 (0.68-0.94), p< 0.0005; Treg OR for 0m-MDA = 3.03 (1.55-8.50), AUC 0.88 (0.77-0.99), p< 0.0005; Treg39+ OR for 0m-HDA = 1.79 (1.27-3.90), AUC 0.97 (0.91-1.0), p< 0.0001; Treg39+ OR for 0m-MDA = 1.49 (1.18-2.31), AUC 0.95 (0.87-1.0), p< 0.0001]. After removing the seronegative patients from the 0m-HDA (n=9, 20%) or the 0m-MDA (n=6, 18%) groups, the relation of the basal Treg or Treg39+ cell fr. with the response to MTX did not vary. The small number of patients who presented with 0m-LDA (n=4) did not allow further analysis of this group. At 12m, patients demonstrated a significant reduction of cTreg and cTreg39+ with no variation observed in HC, and differences between HC and ERA were no longer apparent, likely reflecting the 0m upregulation of the genetically determined CD39 expression levels; of note, the 12m cTreg [OR 3.19 (1.68-8.0)] and cTreg39+ [OR 1.31 (1.13-1.80)] cell fr. remained associated with LDA.

Conclusion: The baseline cTreg39+ fr. in untreated ERA predicts the clinical response to MTX across baseline disease activity strata, and facilitates the development of precision medicine strategies.

« Back to ACR Convergence 2023

ACR Meeting Abstracts - https://acrabstracts.org/abstract/circulating-cd4cd25cd127-foxp3cd39-t-cells-predict-the-response-to-methotrexate-across-basal-disease-activity-strata-in-early-rheumatoid-arthritis/