Genetic Associations with Gestational Duration and Spontaneous Preterm Birth

Abstract

Background: Despite evidence that genetic factors contribute to the duration of gestation and the risk of preterm birth, robust associations with genetic variants have not been identified. We used large data sets that included the gestational duration to determine possible genetic associations.

Methods: We performed a genomewide association study in a discovery set of samples obtained from 43,568 women of European ancestry using gestational duration as a continuous trait and term or preterm (<37 weeks) birth as a dichotomous outcome. We used samples from three Nordic data sets (involving a total of 8643 women) to test for replication of genomic loci that had significant genomewide association (P<5.0×10^-8) or an association with suggestive significance (P<1.0×10^-6) in the discovery set.

Results: In the discovery and replication data sets, four loci (EBF1, EEFSEC, AGTR2, and WNT4) were significantly associated with gestational duration. Functional analysis showed that an implicated variant in WNT4 alters the binding of the estrogen receptor. The association between variants in ADCY5 and RAP2C and gestational duration had suggestive significance in the discovery set and significant evidence of association in the replication sets; these variants also showed genomewide significance in a joint analysis. Common variants in EBF1, EEFSEC, and AGTR2 showed association with preterm birth with genomewide significance. An analysis of mother-infant dyads suggested that these variants act at the level of the maternal genome.

Conclusions: In this genomewide association study, we found that variants at the EBF1, EEFSEC, AGTR2, WNT4, ADCY5, and RAP2C loci were associated with gestational duration and variants at the EBF1, EEFSEC, and AGTR2 loci with preterm birth. Previously established roles of these genes in uterine development, maternal nutrition, and vascular control support their mechanistic involvement. (Funded by the March of Dimes and others.).

Figure 1. Manhattan plots of discovery stage genomewide-associated results.

Top: gestational length as quantitative trait; bottom: preterm birth as dichotomous trait. Regions reached genome wide significance (P < 5×10^-8) and suggestive significance (P < 1×10^-6) were highlighted in red and orange respectively. The six replicated loci were highlighted in bold.

Figure 1. Manhattan plots of discovery stage genomewide-associated results.

Top: gestational length as quantitative trait; bottom: preterm birth as dichotomous trait. Regions reached genome wide significance (P < 5×10^-8) and suggestive significance (P < 1×10^-6) were highlighted in red and orange respectively. The six replicated loci were highlighted in bold.

Figure 2. ESR1 binding at the WNT4 locus.

a. The rs3820282 T allele creates a stronger ESR1 binding site. The ESR1 binding motif ‘sequence logo’ (taken from the CisBP web server) illustrates the DNA binding preferences of ESR1. Tall nucleotides above the X-axis indicate DNA bases preferred by ESR1. Bases below the X-axis are disfavored. The sequence located in the WNT4 promoter is shown below, with the T allele for rs3820282 shown at the bottom. Note that the T allele changes the sequence from C (most disfavored) to T (most preferred). b. rs3820282 overlaps ATAC-seq and H3K4me3 signals in decidual stromal cells at the WNT4 locus. The red vertical line indicates the position of rs3820282. The location of the WNT4 gene is depicted at the bottom. Tall blocks indicate exons, medium height blocks indicate UTRs, and thin lines indicate introns. Arrows within introns indicate the direction of transcription. c. Experimental validation of allele-dependent binding of ESR1 to rs3820282 by electrophoretic mobility shift assay (EMSA). Fluorescently-labeled rs3820282 probe with either the C or T allele was incubated with nuclear extracts of decidual stromal endometrial cells in the presence or absence of purified ESR1 and/or antibody against ESR1. Lane pairs indicate C and T alleles. Preferential binding of ESR1 to the T allele is observed through an increased signal intensity of the shifted band. Upper and lower arrows indicate the locations of supershifted and shifted bands, respectively. Left to right - Lanes 1+2: negative control lanes containing only oligos; Lanes 3+4: increased binding of purified ESR1 to the T allele. Lanes 5+6: limited binding in the presence of nuclear extract only, due to low expression of ESR1 in these cells; Lanes 7+8: substantial allelic binding is detected with the addition of purified ESR1 to the nuclear extract; Lanes 9+10: Supershift using an ESR1 antibody