Circulating DNA in rheumatoid arthritis: pathological changes and association with clinically used serological markers

Abstract

Background: Early diagnosis of rheumatoid arthritis (RA) is crucial to providing effective therapy and often hampered by unspecific clinical manifestations. Elevated levels of extracellular circulating DNA (cirDNA) in patients with autoimmune disease were found to be associated with etiopathogenesis. To our knowledge, this is the first study to investigate the putative diagnostic use of cirDNA in RA and its association with disease activity.

Methods: Blood samples were taken from 63 healthy subjects (HS) and 74 patients with RA. cirDNA was extracted from plasma and cell surface-bound cirDNA fractions (csbDNA). cirDNA concentration was measured by quantitative real-time polymerase chain reaction. Rheumatoid factor was analyzed by immunonephelometry, whereas C-reactive protein and anticitrullinated protein/peptide antibodies (ACPA) were detected by enzyme-linked immunosorbent assay.

Results: Plasma cirDNA was significantly elevated in patients with RA compared with HS (12.0 versus 8.4 ng/ml, p < 0.01). In contrast, nuclear csbDNA (n-csbDNA) was significantly decreased (24.0 versus 50.8 ng/ml, p < 0.01), whereas mitochondrial csbDNA (m-csbDNA) was elevated (1.44 × 10⁶ copies/ml versus 0.58 × 10⁶ copies/ml, p < 0.05) in RA. The combination of csbDNA (mitochondrial + nuclear) with ACPA reveals the best positive/negative likelihood ratios (LRs) for the discrimination RA from HS (LR+ 61.00, LR- 0.03) in contrast to ACPA (LR+ 9.00, LR- 0.19) or csbDNA (LR+ 8.00, LR- 0.18) alone.

Conclusions: Nuclear and mitochondrial cirDNA levels in plasma and on the surface of blood cells are modulated in RA. Combination of cirDNA values with ACPA can improve the serological diagnosis of RA.

Keywords: Antibodies to citrullinated protein/peptide; Circulating nuclear DNA; Mitochondrial DNA; Real-time PCR; Rheumatoid arthritis; Rheumatoid factor.

Fig. 1

Challenge data inspection by principal component analysis. Scatterplots with representation of the various classes were produced with the s.class command of the ade4 R package. The various classes are control healthy donors, rheumatoid arthritis (RA) activity 1 (with classes I–II RA activity), RA activity 2 (with class III RA activity). The variables analyzed are: cell surface-bound nuclear DNA, cell surface-bound mitochondrial DNA, anticitrullinated protein/peptide antibodies, and rheumatoid factor

Fig. 2

Random forest classification tree plot. Diagram with visualization of patients with rheumatoid arthritis and discrimination from healthy donors by random forest classification tree algorithm, based on cell surface-bound mitochondrial DNA and anticitrullinated protein/peptide antibodies estimates. Red circles represent individual healthy donors; green circles represent individual patients with rheumatoid arthritis