Rs870142

rs870142 is located on chromosome 4p16 in a 300 kb-interval between the STX18 and MSX1 genes. Table 1 of the GWAS refers to a single nucleotide variation between A/G at rs870142, whereas dbSNP refers to the other T/C strand. According to dbSNP, C is the ancestral allele. The overall minor allele frequency for T across all populations is 23% (and 21% according to 1000 Genomes).

STX18 encodes a protein involved in transportation between the endoplasmic reticulum (ER) and the Golgi apparatus. MSX1 encodes a homeobox transcription factor that is expressed in the atrial septum throughout heart development in mouse and chick models. MSX1 interacts with another transcription factor, TBX5, which is also critical for proper atrial septal development. Msx1 and Msx2 double knockout mice have congenital heart defects. In humans, loss-of-function mutations in MSX1 can cause tooth agenesis, cleft lip, and tooth-and-nail syndrome but not typically congenital heart defects. Therefore, it is unlikely that rs870142 acts solely through the regulation of MSX1 expression.

A GWAS of congenital heart disease found a significant association between rs870142 and atrial septal defect (ASD) in a European Caucasian population. Genotype data in this study were generated at 557,124 SNPs across the genome for 1,995 individuals with congenital heart defects (CHD) and 5,159 controls. Originally, three major CHD categories (septal, obstructive, and cyanotic defects) were considered together, but this resulted in no region achieving genome-wide significant association. However, when specific CHD categories were considered separately, rs870142 was associated with ASD (p-value = 9.5 x 10-7) in the discovery cohort, and this association was further confirmed in a replication study of 417 additional ASD cases and 2,520 controls (p-value = 5.0 x 10-5). The odds ratio in the discovery cohort was found to be 1.52 per copy of the minor allele and 1.40 in the replication cohort. The overall odds ratio was 1.46 upon combination, and the 95% confidence interval ranged from 1.19 to 1.65. The combined p-value from discovery and replication cohorts was 2.6 x 10-10. Using Levin&#8217;s formula, it was calculated that the genotype at rs870142 accounts for 9% of the population-attributable risk of ASD.

Atrial septal defects account for 7-10% of congenital heart defects in children and 25-30% of such defects in adults. People with ASD have higher morbidity and mortality rates, but this difference becomes more pronounced at older ages. The association between the SNPs at 4p16 and ASD was not seen for any other CHD condition. These results suggest that genetic associations between CHDs and SNPs have a high degree of phenotypic specificity. As this is the only study so far that links rs870142 with ASD, further work in the developmental context is needed to identify the mechanism responsible for this association.