Ankylosing spondylitis in three Sub‐Saharan populations: HLA‐B*27 and HLA‐B*14 contribution

Abstract

Ankylosing spondylitis (AS) is a chronic inflammatory rheumatic disease affecting predominantly the axial skeleton and sacroiliac joints, of unknown etiology and strongly associated with human leukocyte antigen (HLA)-B27 (1). The clinical presentation of AS in Sub-Saharan populations is rare (2) due to the very low frequency in Africans of the most important genetic predisposing condition for the development of AS, HLA-B27 antigen, comparing to Europeans or North Americans. However, the influence of HLA markers other than B27, like other HLA-B alleles, has not been investigated thoroughly in these regions. We previously conducted two studies in the Togolese and Zambian population that showed genetic evidence for implication of HLA-B*1403 in AS (3, 4). This allele was found in 50% of the AS patients from Togo and absent in healthy controls, whereas two of the three AS patients from Zambia were B*14:03 positive and none of the healthy controls had this allele. Here, 14 cases of AS were diagnosed over a 2-year period in the University Hospital Yalgado Ouedraogo of Ouagadougou (Burkina Faso, Africa) according to the New York criteria (5). We also selected 27 controls matched for sex, age and ethnic origin. None of the patients or the controls showed other types of spondyloarthropathies (SpA) (reactive arthritis, arthritis associated to human immunodeficiency virus, psoriasic arthritis or undetermined SpA). A sample of peripheral blood per subject was taken in the University Hospital Yalgado Ouedraogo using PAXgene Blood DNA Tubes (Qiagen, Valencia, CA) for blood collection and stabilization. DNA purification and HLA-B genotyping with the HLA-SSO Typing Kit (Tepnel Lifecodes, Stanford, CT) was performed at the Histocompatibility Unit of the Hospital Universitario Central de Asturias, Oviedo, Spain. We analyzed the HLA-B alleles’ distribution between AS patients and controls from Burkina Faso. One finding was the presence of B*14:03 in two of the 14 AS patients (14.3%) and its absence in the 27 controls. We also found that the presence of B*27:05 was increased in AS patients compared with controls (50% vs 3.7%, P = 0.001, OR = 26). To date, we have described 25 AS patients from Sub-Saharan populations (n = 3, 8 and 14 from Zambia, Togo and Burkina Faso, respectively), and we have compared the HLA-B alleles distribution to healthy subjects (n = 204; 92 Zambians, 85 Togoleses and 27 from Burkina Faso), all of them without other types of SpA. Although the distribution of each HLA-B allele (B*27:05 and B*14:03 ) was distinctive, we concluded that the influence of HLA-B alleles in AS is homogenous in all populations studied (B*14:03+B*27:05 ) (Table 1). In this regard, we take into account the three AS Sub-Saharan populations and showed that the frequency of HLA-B*14:03 or HLA-B*27:05 alleles was significantly increased in AS patients compared to healthy controls (32% vs 0.2%, P = 5.65 × 10−16, OR = 191.5) (Table 1). It has been argued that the AS in Sub-Saharan populations seems to represent a subgroup of the disease (2). The disease has a similar clinical presentation compared to Western Europe and North America, but according to several studies the AS patients are older at onset of the disease, the most of them lack extra-articular manifestations and many affected subjects do not have a family history of AS (2). To date, we have analyzed 25 AS cases in Sub-Saharan populations and the average age at the beginning of the disease was 26 years, similar to the other parts of the world. Moreover, the most important issue for assuming that AS in African is a subgroup of the disease was that the most of Sub-Saharan African patients were HLA-B27 negative, and this subtype was rare in black Africans (2). However, we showed that the influence of HLA class I allotypes is homogeneous in all populations studied, suggesting a common mechanism of predisposition to AS. The pathogenic behavior of B*14:03 and B*27:05 may be related to a common feature, taking on that in both cases they are involved in the disease mechanism. The objective of future studies should be evaluate features compatibles with the association of B*14:03 and B*27:05 with AS. Accordingly, results inconsistent with a pathogenic role of misfolding or free surface-expressed heavy chain have been described (6), showing relevance of peptide binding as a important purpose of study for searching common features between B*14:03 and B*27:05. Structural studies have also been reported in order to shed light on the causes of association of these alleles with AS (7, 8). It has been observed that alloreactive cytotoxic T-lymphocyte may be generated against cells expressing B*14:02 (not associated with AS) that cross-react with HLAB27 molecules, suggesting that many shared ligands adopt antigenically similar conformations when bound to different