Recurrent pregnancy loss and preeclampsia most often manifest at different stages of a pregnancy, yet both originate from deficient trophoblast invasion during early gestation [
20]. Between these two extreme ends, variable degrees of trophoblast invasion are compatible with on-going pregnancy, but result in inappropriate conversion of the spiral arteries and thus inadequate placentation might develop [
20]. Moreover, although a predisposition to vascular dysfunction may exist pre-conceptionally, it is ultimately the disruption of the finely tuned balance between proangiogenic and antiangiogenic factors secreted by the dysfunctional placenta that converge on the vasculature that lead to this syndrome [
21]. However, because hypertensive disorders in pregnancy, including preeclampsia, are a heterogeneous, polygenic and multifactorial group of disorders, no single factor, has so far been described as being responsible for the condition [
22,
23]. On the contrary, multiple genes involved in various biological pathways must be deranged for the syndrome to develop [
23]. It seems therefore more appropriate to claim that there are multiple candidate genes, that act as susceptibility loci, which, after complex interaction with environmental modifiers (such as obesity or smoking), enhance or lower a woman’s threshold for developing gestational hypertensive disorders.
To date, it has been shown that low levels of serum HRG during pregnancy have been associated with early-onset preeclampsia [
24]. Furthermore, in preeclampsia models, VEGF (Vascular Endothelial Growth Factor) and PlGF (Placental Growth Factor) exhibit a proangiogenic role, whereas sFlt-1 (soluble form of the receptor for VEGF) acts as an antagonist of VEGF and PlGF, resulting in endothelial cell dysfunction and vasoconstrictor effects in maternal circulation [
23]. Moreover, HRG peptides reflecting the C633T polymorphism added in endometrial culture models have been associated with low VEGF-induced proliferation, but increased migration of human endometrial endothelium and enhanced tube formation, leading to affected angiogenesis [
16].
Proline to serine missense mutation on position 204 (corresponding to the HRG C633T SNP) allows glycosylation at position 202, which is located near a disulphide bond(Cys 185-Cys 407), entailing the proteolytic HRR fragment of the HRG protein. It has been postulated that differences in glycosylation could possibly affect protein stability and/or degradation rates and thereby affect HRG concentration and its antiangiogenic properties [
25]. The exact pathophysiological mechanism of how HRG contributes to preeclampsia still remains to be discovered.
The allele frequency of the HRG C633T SNP in a European population (Hap-Map Eu) is 0.69 for the proline (C-allele) and 0.31 for the serine (T-allele) alleles, respectively [
26]. In accordance, in the entire study population it was estimated that MAF = 0.32. We regard that it is a case of co-dominance inheritance pattern [
27] as the disease risk in heterozygous lies between that of homozygous C/C or T/T but not in a specific way, such as the one in a multiplicative or additive model. Thus, a logistic regression model was regarded as being appropriate to examine the association [
27].
Strengths and limitations
This is, to the best of our knowledge, the first study examining the association between gestational hypertensive disorders and HRG C633T SNP carrier status. No power calculation could be made, as no similar study has been performed before. The study can therefore be regarded as a pilot study. Additionally, when the study was initiated (year 2011) genetic studies were not that common and the laboratory analysis had a higher cost which prohibited including a higher number of participants. Larger studies are thus needed to verify the results.
One of the major strengths of the study is the well-defined study population with the inclusion of accurate information, both in relation to fertility/infertility status, as well as in relation to the diagnosis of preeclampsia and other hypertensive disorders of pregnancy. The nested case-control study design, with participants primarily selected on the basis of their outcome within the same cohort (BASIC), followed prospectively starting at pregnancy week 17, is considered appropriate. The risk of recall bias and uncertainty was reduced, as all clinical and background information were collected before the onset of the disease and were retrieved from the participant’s medical record. Furthermore, because the genotype was unknown, both for the participant and the researcher, there was limited risk of selection bias.
It is known that ethnicity might affect the distribution of the polymorphism as well as the rate of preeclampsia. However, both cases and controls originate from the same geographic area, thus limiting the influence of this factor. Furthermore, despite not knowing the exact ethnic origin of the population in our study, it is estimated that Caucasians constitute 90–95% of the population living in the area [
28]. Through the random selection of cases and controls from the same cohort, we have ensured a good match between the genetic background of cases and controls, which is adequate in preventing population stratification.
Because only 22% of women invited to participate in the BASIC cohort accepted, the results cannot be readily generalizable for the whole of the Uppsala population. Pregnant women with higher age, higher education and primiparas are slightly overrepresented among the BASIC study participants compared to the general population [
18]. The higher educational and social status of participants might be partly explained by the underrepresentation of women with foreign background due to the language communication requirement, as well as due to the well-known fact that higher educated women have higher participation rates in studies [
18]. The latter should be taken into account, as socioeconomic status may be related to the study’s main outcome. We have, however, no reason to believe that that would affect genotype distribution. It might, however, in part, reflect the lower incidence of severe or early onset preeclampsia among cases in our study. Late onset preeclampsia, occurring in > 75% of preeclamptic pregnant women [
23], appears to have a weaker placental component [
23] with maternal genetic, behavioural and environmental factors instead having a higher impact [
29]. Nonetheless, there is still evidence of increased incidence of placental pathology and impaired vascular remodelling compared with normal pregnancy [
30]. Based on the study findings, we cannot deduct any safe assumptions about early-onset preeclampsia, which is usually related to pronounced impaired angiogenesis and disturbed remodelling of the spiral arterioles [
23,
30]. We could only speculate that the results from a genetic study would have been more prominent if those women who suffered severe or early-onset preeclampsia were included compared to a mixed population.
One should, however, always be cautious when interpreting the genetic effects, as a positive association is not always proof of a direct causation, but might be sign of linkage disequilibrium, population stratification or might have occurred by random chance. In vitro studies might help to clarify the biological mechanisms behind the effects of HRG C633T SNP on the development of gestational hypertensive disorders.
Although the effect of the HRG SNP studied in gestational hypertensive disorders is considered moderate, it could be included together with other associated SNPs in larger studies with more complex design, such as meta-analysis and pathway analysis [
23], investigating the correlation between this relevant phenotype and genotype.