Geo-epidemiology of autoantibodies in rheumatoid arthritis: comparison between four ethnically diverse populations

Abstract

Background: Rheumatoid arthritis (RA) occurs across the globe in different ethnic populations. Most RA patients harbor anti-modified protein antibodies (AMPA); however, it is unclear whether differences exist in autoantibody responses at different geographic locations and between different ethnic groups, which could provide new clues regarding factors underlying autoantibody development. We therefore investigated AMPA prevalence and association with HLA DRB1 alleles and smoking in four ethnically diverse populations on four different continents.

Methods: Anti-carbamylated (anti-CarP), anti-malondialdehyde acetaldehyde (anti-MAA), and anti-acetylated protein antibodies (anti-AcVim) IgG were determined in anti-citrullinated protein antibody-positive Dutch (NL, n = 103), Japanese (JP, n = 174), First Nations Peoples in Canada (FN, n = 100), and black South African (SA, n = 67) RA patients. Ethnicity-matched local healthy controls were used to calculate cut-offs. Risk factors associated with AMPA seropositivity in each cohort were identified using logistic regression.

Results: Median AMPA levels were higher in First Nations Peoples in Canada and especially South African patients, as reflected by percentage seropositivity: NL, JP, FN, and SA: anti-CarP: 47%, 43%, 58%, and 76% (p < 0.001); anti-MAA: 29%, 22%, 29%, and 53% (p < 0.001); and anti-AcVim: 20%, 17%, 38%, and 28% (p < 0.001). Total IgG levels also differed markedly, and when autoantibody levels were normalized to total IgG, differences between cohorts became less pronounced. Although there were some associations with AMPA and HLA risk alleles and smoking, none was consistent across all four cohorts.

Conclusions: AMPA against various post-translational modifications could consistently be detected on different continents across ethnically diverse RA populations. Differences in AMPA levels corresponded to differences in total serum IgG levels. This suggests that, despite differences in risk factors, a common pathway may be involved in AMPA development across geographic locations and ethnicities.

Keywords: Autoantibodies; Epidemiology; Ethnicity; Genetics; Rheumatoid arthritis.

Fig. 1

AMPA levels in geographically and ethnically diverse RA populations. Levels in arbitrary units (aU/mL) for four AMPAs and grams per liter (g/L) for total IgG in the serum of four geographically and ethnically diverse RA populations. Patients clustered at the maximum were above the highest standard of the ELISA. Blue shading indicates the patients falling below the cohort-specific positivity cut-off (see Supplementary Fig. 1 for data regarding cut-off determination in healthy controls and reactivity to the control peptide); green shading is the normal range for total IgG (established in European cohorts). Lines indicate the median and interquartile range. p-values correspond to chi-square tests on the proportion of patients considered positive, indicated in percentages above graphs

Fig. 2

AMPA to IgG ratios in geographically and ethnically diverse RA populations. Ratios of levels in arbitrary units (aU/mL) for four AMPAs, per grams per liter (g/L) total IgG in the serum of four RA populations. Ratios are only shown in patients that were positive for the AMPA. Lines indicate median and interquartile range; p-values correspond to Mann–Whitney U tests. The range of aU/mL per g/L is not directly comparable between AMPAs

Fig. 3

Association of the different AMPA with smoking in the four populations. The association with smoking was corrected for gender. Please note that in South Africa, correction for gender had a large effect on odds ratios because very few women smoke

Fig. 4

Associations of the different AMPA with the presence of HLA DRB1*03 alleles. Genetic associations were not corrected for gender

Fig. 5

Associations of the different AMPA with the presence of shared epitope alleles. Genetic associations were not corrected for gender