Nagendra K. Monangi, Ge Zhang, Mikko Hallman, Kari Teramo, Bo Jacobsson, and Louis J. Muglia
A wide variety of maternal physical and physiological traits and environmental exposures have been associated with various pregnancy outcomes (1). With the growing number of genotyped maternal-fetal dyads now becoming available, human genomics encompassing both classical human genetics and recent advances in understanding genome organization and function as a whole provides unique opportunities to dissect the relative contributions of pregnancy programming by maternal environment or fetal genotype on determining the pregnancy outcomes (2).
The specific genetic variants in humans that modify the risk for adverse pregnancy outcomes have been difficult to determine due to multifactorial environmental influences, heterogenous phenotypes, and involvement of two genomes, those of the mother and the fetus. The “Genetic Conflict” hypothesis describes the potentially competing goals for the mother and infant in pregnancy outcomes and traits, with selection during evolution balancing these differing priorities (3). Many pregnancy characteristics, and the evolution of genomic imprinting itself, are consistent with this theory.
There have been several recent reviews summarizing the current state of knowledge for candidate gene association studies with various pregnancy outcomes. In general, compelling evidence for specific genes as substantially contributing to or protecting from various adverse outcomes has not emerged with this approach. Because of the limited progress with these targeted candidate methods, several groups embarked on high-throughput genome-wide studies. In this review, we describe the available evidence for maternal genetic contributors to various pregnancy outcomes and results from the high dimensional human genetic and genomic investigations.
Integrating genomics and epidemiology: Mendelian randomization
Pregnancy is a complex process that is influenced by multiple factors. Pregnancy outcomes are based on numerous risk factors and represent the aggregation of heterogenous phenotypes. For example, maternal height (4), weight (5), BMI (6), blood pressure (7), and environmental exposures (8) are reported to be associated with birth weight and preterm birth (PTB). Many potential mechanisms have been proposed to explain these observed associations (Figure 14.1).
Figure 14.1 Potential causal mechanisms that can lead to the associations between maternal phenotype and pregnancy outcomes. (a) Direct causal effect of maternal phenotype on pregnancy outcomes. (b) Associations of social and nutritional confounders that have impacts on both maternal phenotype and pregnancy outcomes. (c) Fetal genetics that directly influence pregnancy outcomes. (Source: Ref. 4, with permission.)
To distinguish these mechanisms, Mendelian randomization (MR) has been utilized to define the causal relationships between various exposures and different health outcomes beyond simply association (9). However, as illustrated in Figure 14.1, the approach is complicated by the transmission of maternal alleles. To curtail this problem, Zhang et al. (4) proposed a novel extension of the method to utilize the nontransmitted alleles as a valid instrument for maternal environment, thus avoiding the interference of genetic transmission. This modification provides a new method, using maternal nontransmitted haplotype, to distinguish maternal environment from fetal genetic effects. In this way, using an aggregate genetic score for approximately 700 single nucleotide polymorphisms (SNPs) influencing adult stature, the authors demonstrated that maternal height, and the in utero growth environment it shapes, define the duration of gestation. In contrast, these same height-associated polymorphisms defined birth weight and birth length of the fetus by direct transmission to and action in the fetus.
Genetics in human preterm birth
Preterm birth, defined as birth at <37 weeks of completed gestation, is the leading cause of infant mortality that can arise from many physiological and pathological processes with involvement of a wide variety of molecular pathways (10,11). Multiple lines of evidence suggested a substantial role of genetics on the risk of preterm birth. A woman’s risk of delivering preterm increases up to fourfold if one of her previous infants was delivered preterm (12). Epidemiological studies also showed that women who are born preterm or have sisters who were also preterm have an increased risk of preterm delivery (13). Twin studies and segregation analysis of traits of families demonstrate that approximately 30% of variation in gestational duration is due to hereditary factors (14). Large birth registry data-based quantitative studies indicated that the majority of the genetic influence is due to maternal genes (15). It is estimated that 21% of variation in gestational duration was explained by maternal genetic factors, and fetal genetics account for 13% (16).
Advances in genomic techniques allow dissection of genetic mechanisms of PTB as well as gestational duration in humans. However, similar to other complex human traits, birth timing is likely influenced by multiple common genetic variations of small effect size. Detecting these small effects requires large sample sizes and careful assessment of population substructure to avoid confounding effects due to ancestry.
Many candidate gene studies have been performed focusing on genes presumed to influence contributory pathways of preterm birth such as immunity and inflammation, tissue modeling, progesterone signaling, vascular regenesis, metabolism, and placental function (17). Currently, several hundred gene variants have been interrogated with some suggestive findings (18). However, these variants have either not been replicated or have had a nominal association in genomic studies.
Genome-wide association studies (GWAS) are a robust and unbiased method typically with a case control design for the discovery of disease-associated SNPs utilized to identify shared loci of interest in complex diseases. The whole genome consists of ∼3.1 billion base-pairs, while the whole exome comprises ∼180,000 exons, 35 million base-pairs and 22,000 genes (about 1.2% of the whole-genome sequence). This hypothesis free approach systematically screens the whole genome without prior preference for specific regions or genes (19).
The first published GWAS study of spontaneous PTB using the Danish National Birth Cohort (20) investigated maternal as well as fetal genomes. None of the suggestive loci in the discovery phase were replicated in the validation cohort (1). The relatively small sample size of approximately 1,000 cases and 1,000 controls in this study likely was underpowered to detect true associations.
To overcome this limitation, a large two-stage GWA study was performed using a discovery set of samples obtained from 43,568 women of European ancestry who were research participants of 23andMe using gestational duration as a continuous trait and term and PTB as a dichotomous outcome (21). Samples from three Nordic data sets (∼8,000 samples collected from Denmark, Norway, and Finland) were used to test for replication of genomic loci that had significant GWA or an association with suggestive significance in the discovery set. This GWAS found that variants at the EB1, EEFSEC, AGTR2, WNT4, ADCY5, and RAP2C loci in the maternal genome were associated with gestational duration and/or PTB (Figure 14.2) (21). All associations were of maternal origin, as suggested by the analysis of mother-infant dyads.
Figure 14.2 Results of the Discovery-Stage Genome-wide Association Study. (a) The 12 loci that were associated with gestational duration with suggestive significance (p < 1.0 × 10−6), four of which were associated with genome-wide significance (p < 5.0 × 10−8). (b) The five loci that had an association with preterm birth with suggestive significance, two of which had an association of genome-wide significance. The top three loci that were associated with gestational duration (EBF1, EEFSEC, and AGTR2) were also associated with preterm birth. The names of the six replicated loci are highlighted in bold. (Source: Ref. 21, with permission.)
Functional significance of these loci has not been previously studied in relation to preterm birth; however, previously established roles of these loci in maternal nutrition, vascular control, and uterine development support their mechanistic role in control of parturition (Table 14.1). Although the effect sizes of these loci are small, these findings reveal new biological insights into previously unsuspected and potential intervenable pathways.
Table 14.1 Genetic loci associated with gestational duration and preterm birth
Known genome-wide association
Regulates adipocyte differentiation and development (9)
Involved in immune system development
Reduced selenium concentration is associated with preterm birth risk (43)
Bone mineral density
Development of female reproductive system
Associated with endometriosis (44)
Angiotensin II receptor
Modulating uteroplacental circulation (45)
Diabetes, obesity, birth weight
Membrane-bound adenylyl cyclase
Cell energy, metabolism (46)