Background Diagnosis of the preclinical phase of rheumatoid arthritis (RA) allows timely start of treatment with the potential to prevent disease progression. It is known that antibodies against citrullinated proteins (ACPA) and rheumatoid factor (RF) have diagnostic value to identify pre-clinical RA. However, since only 20-40% of ACPA+/RF+ arthralgia patients develop arthritis within 5 years, better prognostic markers are needed. Recently, we demonstrated involvement of Interferon (IFN) response and B-cell gene signatures in pre-clinical RA.
Objectives The objective of this study is to demonstrate the diagnostic value of the IFN and B-cell gene signatures in the diagnosis of pre-clinical RA.
Methods Peripheral blood (Paxgene) was collected from 115 ACPA+/RF+ arthralgia patients from the Jan van Breemen Research Institute | Reade Amsterdam. Patients where clinically followed for arthritis development, one or more swollen joints, with a mean follow-up time of 23 months (IQR 12-30). IFN response and B-cell related gene expression was measured by multiplex qPCRs. An IFN score was calculated based on 7 highly correlating Type I IFN response genes. A B-cell score was calculated based on three highly correlating B-cell related genes. Cut-off levels for the IFN and B-cell high or low definition were determined by the 85% specificity of the individual IFN and B-cell Receiver Operating Characteristics (ROC)-curves. Cox regression analysis and ROC-curve analysis were used to demonstrate prognostic and diagnostic significance.
Results Out of 115 arthralgia patients 44 developed arthritis after a median time of 8 months (IQR 5-13). Stratification of these individuals based on the IFN score revealed that 60% of the IFNhigh patients converted to arthritis compared to 32% in IFNlow patients (P=0.011). For the B-cell signature, 58% in B-celllow patients developed arthritis, compared to 33% of B-cellhigh patients (P=0.020). Combined analysis revealed a significant high risk for arthritis development in IFNhigh/B-celllow patients (80%, hazard ratio (HR)(6.22, P=0.003) and a low risk for IFNlow/B-cellhigh patients (26%, HR 0.16, P=0.003). To demonstrate clinical utility a ROC-curve was constructed of ACPA+/RF+ alone and in combination with both signatures. The area under the curve reached 0.619 (P=0.032, C.I. 0.514-0.724) for ACPA+/RF+ and increased to 0.803 (P=0.0001, C.I. 0.718-0.888) with IFN and B-cell signatures included. The sensitivity to diagnose pre-clinical RA increased from 16% to 52% when both signatures are included, with a cut-off of 94% specificity.
Conclusions These findings demonstrate the clinical value of IFN and B-cell gene signatures as biomarkers for the diagnosis of pre-clinical RA.
Acknowledgements This research was supported by the Center for Translational Molecular Medicine (CTMM) consortium “TRACER”.
Disclosure of Interest J. Lubbers Grant/research support from: CTMM TRACER, L. van de Stadt Grant/research support from: Dutch Arthritis Foundation, S. Vosslamber: None Declared, J. G. Wesseling Grant/research support from: CTMM TRACER, D. van Schaardenburg Grant/research support from: Dutch Arthritis Foundation, C. Verweij Grant/research support from: CTMM TRACER