Background: Tumour necrosis factor-αlpha inhibitors (TNFαi) are the main biologics (b-MARDs) used to treat active rheumatoid arthritis (RA) in patients that have failed disease modifying treatment (DMARD). However, 10% of patients with rheumatoid arthritis, TNFαi don’t t work at all and 40% of patients only respond partially and require ongoing treatment with DMARDs/prednisone. Biomarkers may identify before treatment which patients will be responders and whether prednisone and DMARDS can be withdrawn. In a longitudinal study our aim was to evaluate whether cytokine profiles predicts: (a) sustained DMARD treatment responders (b) sustained TNFαi treatment responders, (c) TNFαi treatment failures.
Methods: We used the Millipore’s MILLIPLEX MAP Human Th17 Magnetic Bead kit for the simultaneous quantification of the following cytokines: IL-1β, IL-2, IL-4, IL-5, IL-6, IL-9, IL-10, IL- 12p70, IL-13, IL-15, IL-17A, IL-17E/IL-25, IL-17F, IL-21, IL-22, IL-23, IL-27, IL-28A, IL-31, IL-33, GMCSF, IFNγ, MIP-3α, TNFα and TNFβ. γ, MIP-3α, TNFα and TNFβ. RA disease activity was measured using the RA Disease Activity Score (DASCRP28) which includes joint counts, CRP and a patient-reported outcome of health status. All samples were blocked with Heteroblock to reduce rheumatoid factor and other heterophilic antibodies. Rheumatoid factor was measured before and after blocking. The same negative and positive controls were included across all plate runs. All assays were done in singlet to accommodate longitudinal samples. Mixing studies were undertaken to evaluate whether cytokine results could be analysed using quantitative statistics. Only p values <0.005 are reported given that 25 cytokines were included in the model.
Results: We included 14 patients with RA starting on a DMARD and 26 patients with RA starting on a TNFαi after failing DMARDs. Cytokines were assayed monthly 2-3 months prior treatment and to evaluate month-to-month cytokine variability, and monthly up to 5 months after treatment initiation. There were 67 serum samples in the DMARD treated group and 202 serum samples in the TNFαi treated group. Using mixed effects linear regression to account for longitudinal data in a model that included all 25 cytokines, treatment-time and treatment type (DMARD or TNFαi+/-DMARD), we found that in patients on DMARDs, IL-6, IL-1β, IL-28A, and TNFβ were associated with treatment response. However, in TNFαi treated patients, TNF-a, GM-CSF and IL-6 were associated with treatment response.
Conclusions: In a small longitudinal cohort of 26 patients with a total of 202 samples, measuring TNFα and GM-CSF predict early TNFαi responders. Mixing studies and longitudinal variability evaluation indicated that cytokine assays were robust even though it was not possible to remove all RF from samples.