Epistasis between IGF2R and ADAMTS19 polymorphisms associates with premature ovarian failure

Abstract

Do single nucleotide polymorphisms (SNPs) or synergistic interactions between SNPs and diplotypes within the insulin-like growth factor 2 receptor (IGF2R) and ADAM metallopeptidase with thrombospondin type 1 motif, 19 (ADAMTS19), contribute to premature ovarian failure (POF)? Synergistic interactions were detected between SNPs, including a non-synonymous SNP, and diplotypes within IGF2R and ADAMTS19 which may contribute to POF; however, there was no correlation with POF in a single SNP model after Bonferroni correction. IGF2R regulates free IGF2 level, which is involved in steroidogenesis in bovine granulosa cells. ADAMTS19 expression is higher in the murine embryonic ovary than in the embryonic testis during sexual differentiation, and an ADAMTS19 SNP (rs246246) showed a possible association with POF in a genome-wide association study in Caucasian women. This study analyzed interactions between SNPs and diplotypes within IGF2R and ADAMTS19 as well as SNPs within the two genes. In Stage I, a total of 120 patients with POF and 152 female controls were recruited. All patients were diagnosed with POF at the CHA hospital in Seoul, Korea, and were recruited between 1994 and 2004. The 152 controls were recruited from Chungju, Korea, as part of another study that was conducted from April 2002 to March 2004. For Stage II, we obtained genotype data for an additional 1641 female controls, recruited in Ansung and Ansan from 2001 to 2008, from the Korean Genome Epidemiology Study (KoGES). In Stage I, the GoldenGate assay with VeraCode technology was used to genotype SNPs in IGF2R and ADAMTS19. In Stage II, we obtained genotype data for IGF2R and ADAMTS19 using Affymetrix Genome-Wide Human SNP array 5.0 and imputed data by the IMPUTE program from the KoGES. To identify POF-associated SNPs, logistic regression analysis in an additive model was performed using the PLINK tool. Synergistic interactions between SNPs and diplotypes within IGF2R and ADAMTS19 were analyzed by logistic regression analysis in three alternative models. In Stage I, 13 combinations of SNPs showed significant synergistic interactions after Bonferroni correction [the strongest association had odds ratio (OR) = 5.77, 95% confidence interval (CI): 2.26-14.75, P = 0.00025]. In Stage II and combined analyses, two and four combinations, respectively, of the significant results in Stage I showed significant synergistic interactions after Bonferroni correction. For interactions between diplotypes in block 2 of IGF2R and block 3 of ADAMTS19 in Stage I, we found 17 synergistic interactions with P < 0.0001, but there was no significant interaction after Bonferroni correction. In Stage II and combined analyses, we found that three and seven combinations in the same blocks, respectively, showed significant synergistic interactions after Bonferroni correction (strongest association: OR = 4.12, 95% CI: 2.22-7.62, P = 6.74E-06). The sample size for patients with POF in this study was small but, compared with recent reports describing associations between SNPs and POF and considering the low prevalence of POF (1%), the sample size is considered to be reasonable. These results should be confirmed in large-scale studies involving different ethnic groups. Our results may ultimately provide predictive markers for women at a high risk of POF. This study was supported by grants from Basic Science Research Program through the National Research Foundation of Korea (NRF), which is funded by the Ministry of Education (2009-0093821, 2011-0010637). There are no competing interests.