Association of estrogen receptor alpha gene polymorphisms and its gene-gene interactions with knee osteoarthritis susceptibility in Chinese Han population

Aims: To investigate the association of estrogen receptor alpha (ESR1) gene polymorphisms and its gene-gene interactions with knee osteoarthritis (OA) susceptibility. Methods: Generalized multifactor dimensionality reduction (GMDR) was used to screen the best interaction combination among SNPs. Logistic regression was performed to investigate association between 3 SNPs within ESR1 gene and knee OA risk and additional genegene interaction between rs2228480 and rs9340799. Results: Logistic analysis showed that rs2234693 and rs2228480 within ESR1 gene were both associated with increased knee OA risk in both additive and dominant models. The carriers of homozygous mutant of rs2234693 and rs2228480 polymorphism have higher OA risk than those with wild-type homozygotes, OR (95% CI) were 1.61 (1.22-2.12) and 1.75 (1.40-2.13), respectively. GMDR model shows a significant two-locus model (P=0.0010) involving rs2228480 and rs9340799, the crossvalidation consistency of this two-locus model was 9/10, and the testing accuracy was of 60.72%. Participants with rs2228480-TC or CC and rs9340799-AG or GG genotype have the highest knee OA risk, compared to participants with rs2228480-TT and rs9340799-AA genotype, OR (95% CI) was 2.32 (1.52-3.23). Haplotype containing the s2234693-C and rs2228480-A allele were associated with a statistically increased knee OA risk, OR (95% CI)=1.76 (1.38-2.18), P<0.001. Conclusions: We found that rs2234693 and rs2228480 within ESR 1 gene, interaction between rs2228480 and rs9340799 and haplotype containing the s2234693-C and rs2228480-A allele were associated with a statistically increased knee OA risk.