Background/Purpose: Chemokine (C-X-C motif) ligand 13 (CXCL13), produced by follicular helper T (Tfh) cells, plays a pivotal role in B-cell homing and activation in germinal centers in salivary glands.^1,2 CXCL13 is associated with systemic disease activity and clinical activity in primary Sjögren’s syndrome (pSS) and increased risk of lymphoma.^1,2 In patients with pSS, in an open-label trial (ASAP), abatacept (ABA) treatment decreased Tfh and CD4 memory T (TEM) cells.¹ In the Phase III, double-blind, placebo-controlled study (NCT02915159), ABA significantly impacted biomarkers of disease activity, but did not improve clinical disease measures, vs placebo.³ This analysis aimed to assess serum CXCL13 levels and immune cell phenotypes of study patients.

Methods: Adult patients meeting the 2016 ACR/EULAR criteria for pSS with EULAR Sjögren’s Syndrome Disease Activity Index ≥5 were randomized to receive ABA 125 mg SC weekly or placebo for 168 days. Serum and whole blood samples were collected at follow-up. Serum CXCL13 was measured using the SIMOA™ assay from Myriad RBM (all patients). Immune cell phenotyping of whole blood samples in a subgroup of patients from participating countries with available samples (cellular phenotyping population [CPP]) was assessed by flow cytometry, with data analyzed using BD FACSDiva™ software. Estimates of adjusted mean change from baseline (BL) in immune cells were from a repeated measures mixed model with statistical significance determined by the Benjamini–Hochberg procedure.

Results: In total, 187 (ABA, 92; placebo, 95)³ patients were randomized and received ≥1 dose of study drug. At BL, CXCL13 levels were comparable between treatment arms. At Days 85 and 169, ABA reduced levels of serum CXCL3 from BL, and ABA-treated patients had significantly lower levels of serum CXCL13, compared with placebo (p< 0.0001; Figure). Seventy-eight (ABA, 32; placebo, 46) patients were included in the CPP, with no observed differences in BL characteristics between treatment groups or populations (CPP vs non-CPP). ABA-treated patients had significantly reduced proportions of disease-relevant cell subsets (Tfh, activated Tfh, CD4 TEM, helper T cell 1 [Th1] and regulatory T cells [Treg]) at Day 169 compared with BL (data not shown). At Day 169, there was a significant adjusted mean difference in adjusted mean change from BL for ABA vs placebo for Tfh, ICOS-expressing activated Tfh and Treg cells (Table), and CD4 TEM, Th1 cells and naïve CD4 T cells (data not shown), and a numerical difference for plasmablasts (Table).

Conclusion: A significant pharmacodynamic effect of abatacept was observed on disease-relevant serum (CXCL13) and pathogenic cell subpopulations. The disconnect between the effectiveness of abatacept in lowering serum CXCL13 (a relevant biomarker in pSS) and pathogenic cell subsets, and the previously-reported lack of demonstrable benefit (vs placebo) in clinical outcomes, warrants further investigation.

¹Verstappen GM, et al. Arthritis Rheumatol 2017;69:1850–61.

²Nocturne G, et al. Arthritis Rheumatol 2015;67:3226–33.

³Baer AN, et al. EULAR 2019:abstract OP0039.

Professional medical writing: Rachel Rankin, Caudex, funded by Bristol-Myers Squibb.

« Back to 2019 ACR/ARP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/abatacept-reduces-serum-cxcl13-and-disease-relevant-immune-cell-phenotypes-in-a-double-blind-placebo-controlled-primary-sjogrens-syndrome-clinical-trial/