Exome sequencing identifies novel rheumatoid arthritis-susceptible variants in the BTNL2.

The butyrophilin-like protein 2 gene (BTNL2) within the class III region of the major histocompatibility complex genomic region was identified as a rheumatoid arthritis (RA) susceptibility gene by exome sequencing (19 RA cases) with stepwise filtering analysis, and then validated by Sanger sequencing and association analysis using 432 cases and 432 controls. Logistic regression of the Sanger-sequenced single-nucleotide variants in an association study of 432 cases and 432 controls showed that 12 non-synonymous single-nucleotide polymorphisms (SNPs) in BTNL2 were significantly associated with RA. The lowest P-values were obtained from three SNPs, rs41521946, rs28362677 and rs28362678, which were in absolute linkage disequilibrium: P=4.55E-09, odds ratio=1.88, 95% confidence interval=1.52-2.33. The BTNL2 locates on chromosome 6 between HLA-DRB1 and NOTCH4, and is 170 kb apart from these two genes. Although DRB1 and NOTCH4 were reported to be RA-susceptible, the three BTNL2 SNPs retained significant association with RA when evaluated by the logistic regression with the adjustment for RA-susceptible HLA-DRB1 alleles in Japanese or rs2071282-T in NOTCH4: P=0.0156 and P=0.00368, respectively. These results suggest that the three non-synonymous SNPs in BTNL2 confer RA risk independently from HLA-DRB1 and NOTCH4.

Exact break point of a 50 kb deletion 8 kb centromeric of the HLA-A locus with HLA-A*24:02: the same deletion observed in other A*24 alleles and A*23:01 allele.

In a structural aberration analysis of patients with arthritis mutilans, a 50 kb deletion near the HLA-A locus with HLA-A*24:02 allele was detected. It was previously reported that HLA-A*24:02 haplotype harbored a large-scale deletion telomeric of the HLA-A gene in healthy individuals. In order to confirm that the deletion are the same in patients with arthritis mutilans and in healthy individuals, and to identify the break point of this deletion, the boundary sequences across the deletion in A*24:02 was amplified by polymerase chain reaction (PCR) as a 3.7 kb genomic fragment and subjected to nucleotide sequence determination. A comparison of these genomic sequences with those of the non-A*24:02 haplotype revealed that the deleted genomic region spanning 50 kb was flanked by 3.7 kb repetitive element-rich segments homologous to each other on both sides in non-A*24. The nucleotide sequences of the PCR products were identical in patients with arthritis mutilans and in healthy individuals, revealing that the deletion linked to A*24:02 is irrelevant to the onset of arthritis mutilans. The deletion was detected in all other A*24 alleles so far examined but not in other HLA-A alleles, except A*23:01. This finding, along with the phylogenic tree of HLA-A alleles and the presence of the 3.7 kb highly homologous segments at the boundary of the deleted genomic region in A*03 and A*32, may suggest that this HLA-A*24:02-linked deletion was generated by homologous recombination within two 3.7 kb homologous segments situated 50 kb apart in the ancestral A*24 haplotype after divergence from the A*03 and A*32 haplotypes.