Background Presence of antibodies against cyclic citrullinated peptides (Anti-CCP2) of IgG, IgA and IgM isotypes has been demonstrated to precede the development of rheumatoid arthritis (RA) by several years. The underlying process for the development of antibodies against certain citrullinated peptides (ACPA) is largely unknown.

Objectives We have investigated antibodies against thirteen citrullinated proteins, besides anti-CCP2, to analyse the development and autoantigen reactivity in the pre-patient phase of RA.

Methods This study comprised 406 individuals, with 717 samples, who were identified before onset of symptoms of RA (median 7.4 years IQR 3.3 – 12.6 years), as donors to the Medical Biobank of Northern Sweden. 204 of them were also sampled at the time of diagnose, and 1305 population controls were identified from the Medical Biobank for analysis of antibodies towards thirteen different citrullinated peptides in plasma; Fibrinogen (Fib) β36-52, Fib 72, Fib 74, Fib α 36-50, Fib α621-635, Fib β60-74, Fib 573,Fib 591, α-enolase, collagen citC1, CCP-1, Vimentin (Vim) 2-17, Vim 60-75 using the microarray based ImmunoCAP ISAC® system (Phadia Diagnostics, Uppsala). All samples were also analysed for anti-CCP2 antibodies with ELISA (Euro-Diagnostics). Cut-off levels were identified by receiver operating characteristic (ROC) curves with 95% specificity as lower limit.

Results The concentrations of all of the analysed antibodies against citrullinated peptides except for antibodies against Vim 60-75, Fib β72, Fib α 591, Fib α 573 and Vim 2-17 using the multiplex assay were significantly increased in pre-diseased individuals compared with controls (p<0.001). Median (IQR) time for the first antibody (Fib α36-50) to appear was 9.8 (11.4) years. The first appearing antibodies (Fib α36-50, Fib α 591, Fib β72 and Vim 60-75) had showed fluctuations in levels over time and only a slight increase after disease onset. Antibodies against Fib β36-52, α-enolase and CCP1 gradually increased in parallel with the highest concentrations before symptom onset. Antibodies against citC1, Fib α621-635 and Fib β74 constituted a third cluster with only a slight increase pre-dating disease onset but a prominent increase after onset. All antibodies had increased concentrations in the RA patients compared with controls (p<0.05- 0.001). The relative risk for future development of RA in individuals with the combination of α-enolase and Fib β36-52 antibodies compared with those having none of the two antibodies all sampled less than 3.35 years before onset of symptoms was 40.4 (CI 95% 19.8-82.3).

Conclusions These results indicate that the development of an ACPA immune response is initially restricted and largely unspecific and expands over time to involve a more specific response with increasing concentrations towards onset of symptoms, most evident for antibodies against α-enolase, Fib β36-52 and CCP1.

Disclosure of Interest None Declared