Anti-cyclic citrullinated peptide (anti-CCP) antibody is a highly specific biomarker for rheumatoid arthritis (RA), and is recognised as a predictor of the development of RA.1 2 It has been demonstrated that some HLA-DRB1 alleles, such as shared epitope (SE) alleles, significantly contribute to the positivity of anti-CCP antibody and the susceptibility of anti-CCP antibody-positive RA.3 4 However, the quantitative effect of HLA-DRB1 alleles on anti-CCP antibody levels in RA patients is controversial.3 5 6 Therefore, we carried out a large-scale study to study the quantitative effects of HLA-DRB1 alleles on anti-CCP antibody levels in Japanese patients with RA.

A total of 2384 Japanese patients with RA was recruited through several medical institutes of Japan under the support of the BioBank Japan Project7 (age 61.5±11.8 years (mean±SD), 81.2% were women). All patients fulfilled the American College of Rheumatology revised criteria for RA, and provided written informed consent. Serum anti-CCP antibody levels were assayed using the second-generation anti-CCP ELISA kit (DIASTAT Anti-CCP; MBL, Nagoya, Japan) with a cut-off value of 4.5 U/ml, and high-resolution (4-digit) genotyping of HLA-DRB1 allele was performed by the WAKFlow HLA typing kit (Wakunaga, Hiroshima, Japan) with the Luminex multi-analyte profiling system (xMAP; Luminex Corporation, Texas, USA). HLA-DRB1 alleles that had the amino acid sequence [QK]/[QR]/[RR]RAA at positions 70–74 of the third hypervariable region were classified as SE positive and the others as SE negative. We selected only anti-CCP antibody-positive RA patients for the subjects of analysis. For each HLA-DRB1 allele, the association between the genotypes of the HLA-DRB1 allele and anti-CCP antibody levels was tested using the two-sided Jonckheere–Terpstra test, a non-parametric test that evaluates the trend of ranks among ordered categories. Bonferroni's correction was applied based on the number of HLA-DRB1 alleles identified.

Anti-CCP antibody was positive in 1883 RA patients (79.0%) with a median level of 100.2 U/ml (interquartile range (IQR) 29.3–240.4 U/ml). Thirty-five HLA-DRB1 alleles (10 SE positive and 25 SE negative) were identified (table 1), with frequencies similar to the previous report in Japanese patients with RA.8 The frequencies of HLA-DRB1 alleles were different among anti-CCP antibody-positive and antibody-negative RA. We had evaluated their associations with anti-CCP antibody positivity, using some of the patients in this study.4 Four HLA-DRB1 alleles demonstrated significant quantitative associations with anti-CCP antibody levels (p<0.05 after Bonferroni's correction). DRB1*0901, accounting for 17.2% of HLA-DRB1 alleles, significantly lowered anti-CCP antibody levels (p=4.5×10⁻²⁰). The median anti-CCP antibody level was 118.7 U/ml (IQR 37.9–293.6 U/ml), 51.6 U/ml (IQR 22.4–143.5 U/ml), or 56.5 U/ml (IQR 17.9–113.7 U/ml) for patients with zero, one, or two DRB1*0901 alleles (figure 1). Two SE-positive alleles, DRB1*0405 and DRB1*1406, and one SE-negative allele, DRB1*1501, showed significant associations with higher anti-CCP antibody levels (p<0.0005). When SE-positive alleles were grouped into SE, SE significantly raised anti-CCP antibody levels (p=4.9×10⁻¹²). We evaluated the association of DRB1*0901 independent of SE, by stratified analysis using the samples not carrying SE, and found a significant association (p=6.1×10⁻⁸).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Distribution of anti-cyclic citrullinated peptide antibody (anti-CCP) levels in 1883 Japanese patients with anti-CCP antibody-positive rheumatoid arthritis, stratifi ed by the number of HLA-DRB1*0901 alleles. Boxes depict medians and interquartile ranges. Whiskers show 5th and 95th percentiles. Dots represent levels less than 5th percentiles or more than 95th percentiles. The widths of the boxes are proportional to the square roots of the numbers of the patients in the stratifi ed groups. The two-sided p value by Jonckheere–Terpstra test is indicated in the legend.

In conclusion, this study demonstrated that one of the non-SE HLA-DRB1 alleles, DRB1*0901, significantly lowers anti-CCP antibody levels in Japanese patients with anti-CCP antibody-positive RA. The positive correlation between SE and anti-CCP antibody levels was replicated.3 5 Although we demonstrated that the possession of DRB1*0901 was associated with a lower level of anti-CCP antibody, the relation between DRB1*0901 and RA and its subtypes of anti-CCP antibody is complicated.8,–,10 DRB1*0901 is believed to be the risk allele to develop RA in Asian populations.8 9 However, its effect to develop an anti-CCP antibody-positive or negative RA subgroup is not conclusive.10 These results might suggest that the disease onset of RA, the production of anti-CCP antibody and the increase in the anti-CCP antibody level would not simply be related to each other when considering the role of HLA-DRB1 alleles in RA pathology, and it is a topic to be investigated further.