Background and objectives Rheumatoid arthritis (RA) is characterised by increased numbers of circulating CD28neg T-cells, a population displaying functional characteristic of cytotoxic memory cells, including expression of granzymes, a family of serine proteases playing key roles in the induction of target-cell death. Soluble granzymes (as granzyme B (GrB)) are elevated in sera from patients with RA, in which are associated with development of radiographic erosions. Abatacept (CTLA4-Ig; ABA) competing with the engagement of CD28 on T-cells, influences the subsequent T-cell activation. In patients treated with ABA, the number of circulating CD28neg T-cells decreases, suggesting that the drug can prevent the generation of CD28neg T-cell populations. Since CD28neg T-cells may be an important source of GrB, we have evaluated the effect of ABA therapy on GrB serum levels in patients with RA.
Materials and methods Fifty-three RA patients, treated for at least 3 months with ABA were evaluated. Disease activity and response to the treatment were measured with DAS28 (based on CRP) and EULAR Criteria. T-cell counts were determined by flow cytometry. Serum GrB samples were collected before the first administration of ABA (T0) and then after 6 months (T6), and measured by an indirect solid-phase enzyme immunoassay with a sensitivity limit of 20 pg/ml.
Results The percentage and the absolute number of circulating CD4+CD28-neg T-cells decreased after ABA therapy (T0 vs. T6: p = 0.018; p = 0.018, respectively), as well as those of CD8+CD28-neg T cells (T0 vs. T6: p = 0.005; p = 0.008, respectively). At T0, GrB serum levels were detectable in all RA patients, and were correlated with disease activity (p = 0.0022), and percentages of circulating CD4+CD28- (p = 0.007) and CD8+CD28- T-cells (p = 0.031). In 25 patients serum level of GrB were evaluated at T6: in 18 patients with a moderate or good clinical response to ABA the levels of GrB significantly decreased from T0 (median: 62.8 pg/mL[10th-90th percentile:45.8–116]) to T6 (53.8 [46.4–96.6]; p = 0.023), whereas no variation was observed in 7 non responders. The variation of GrB levels was directly correlated with that of DAS28-PCR (p = 0.040), but not with those of circulating CD28-neg T-cell subsets.
Conclusions Costimulation blockade by ABA can decrease the serum levels of GrB in RA patients responding to the treatment, suggesting that this might be one of the mechanism by which ABA can prevent radiographic erosions. However, the lack of correlation of such decrease with the numbers of circulating CD28-neg T-cells suggests that these cells are not the main source of serum GrB.