Candidate genes in rheumatoid arthritis

Abstract

Rheumatoid arthritis (RA) is a systemic inflammatory disease with autoimmune features and primarily manifests in the synovial joints. There is a prominent genetic component. We investigated candidate polymorphisms in the genes encoding several molecules implicated in RA: HLA-DRB1, CCR5, CTLA-4, IFN-γ, TNFα, MBL, TGFβ1, and TcRBV14SI. The Caucasian and Maori and Polynesian study populations were derived from 563 unrelated RA patients and 1022 healthy controls from New Zealand (NZ), providing good statistical power for association genetic analysis. 126 of the patients were from a prospective early RA (ERA) cohort with detailed assessment, and including 42 examined with magnetic resonance imaging of the dominant wrist. HLA-DRB1 typing by exon 2 DNA sequencing confirmed the relationship between the ‘shared epitope’ (codons 70 - 74 QKRAA, QRRAA, or RRRAA) and disease severity in Caucasians in the early RA cohort (OR for erosions at 2 years 6.41, 95% CI 2.96 - 13.84, p < 0.0001) but this was not seen in Maori and Polynesian ERA patients. A low-resolution study of the larger cohort (563 RA, 1022 healthy controls) confirmed the HLA-DR4 association (OR 4.3, 95% CI 3.34 - 5.40, p < 0.0001), validating the collections. We found a significant protective effect of the common non-functional CCR5 ∆32 variant on RA prevalence (RA heterozygotes 17%, healthy controls 11%, p = 0.007; no ∆32 homozygote patients detected), but with CCR5 (and the other non-HLA genes examined) there was no impact on RA severity. The HUMCTLA-4A position +49 G allele (associated with reduced CTLA-4 down-regulatory function on T-cells) was assessed by PCR-RFLP and was significantly higher in RA patients than in controls (p = 0.0003), consistent with a genetic contribution from impaired T-cell immunoregulation. An association between the VNDR microsatellite in intron A of the gene for pro-inflammatory IFN-γ (HUMIFNG) and RA was noted (124 bp (CA)12 allele frequencies RA 0.49 and healthy controls 0.40, p < 0.0001, p (corr.) < 0.0005; 126 bp 0.41 and 0.46, p = 0.009, p (corr.) = 0.04; 128bp 0.05 and 0.09, p = 0.0005, p (corr) = 0.002, although a reported strong effect of the 126 bp (CA)13 on RA severity could not be confirmed. Comparison of the TNFa variable number dinucleotide repeat (VNDR) microsatellite polymorphism allele length frequencies showed a strong negative association with TNFa7 (p < 0.0001, p corrected (corr.) < 0.001 and TNFa12 (p = 0.0007, p (corr.) = 0.009). An interaction between TNFa6 and DRB1*04 was noted, but this relationship is complicated by the strong linkage between the TNFA and DRB1 loci. A reported association between a non-conservative single nucleotide polymorphism (SNP) of the mannose binding lectin gene HSMBPCA1 (MBL 54B) and RA was examined by SSP-PCR, but this was not confirmed (RA allele frequency 11%, healthy controls 10%; p = 0.70). Similarly, an association between the TGFβ1 (HSTGFβ1) cytokine gene (T/C at +869; Leu→Pro substitution in codon 10) was not confirmed when the NZ samples were examined using PCR-RFLP. Based on previous reports of elevated T cell receptor BV14SI (Vβ14) transcript levels in RA, we undertook sequence-based mutation screening of the TcRBV14SI gene and found two novel SNP's in non-coding regions. Although one (G3’137A) was present in 6 of 17 RA but none of 12 healthy controls initially sequenced, by using PCR-RFLP on larger samples no difference was shown. In summary, we found genetic association evidence supporting a role in RA for HLA-DRB1, CCR5, CTLA-4, IFN-γ and TNFα. It is likely that other genes contribute, with individually weak effects that require study with large samples. The lack of association for the SNP’s in MBL, TGFβ1 and TcRBV14SI does not rule out their role in RA. The genetics of RA are as complex as the disease itself.