Preeclampsia affects approximately 2-8% of all pregnancies [1] and is clinically manifested as proteinuria and hypertension after 20 weeks of gestation [2]. Preeclampsia is recognized as a risk factor for hypertension and cardiovascular disease (CVD) later in life [3], with early-onset (with delivery prior to gestational week 34) and recurrent episodes of preeclampsia being even more strongly associated with later CVD [4]. The mechanistic links between preeclampsia and future health are incompletely elucidated. Common predisposition, genetic or environmental, that could contribute to the development of both preeclampsia and hypertension or CVD later in life, is a possible explanatory link. In line with this, classical CVD risk factors such as obesity, dyslipidemia, insulin resistance, family history and endothelial dysfunction are also associated with preeclampsia [3]. A pregnancy complicated by preeclampsia as a direct causal factor, or an “amplifying factor” for the association, can also not be ruled out.

The regulator of G protein signaling 2 (RGS2) acts as a GTP-ase activating protein for Gαq proteins and is thereby involved in controlling the termination of vasoconstrictor signaling initiated by ligands such as Angiotensin II (Ang II) [5]. Hercule et al. showed that RGS2 deletion promotes Ang II-dependent hypertension primarily through an increase of myogenic tone and vasoreactivity, probably by sensitization of Ang II type 1 receptor (AT1-receptor) [6]. The G allele of a polymorphism (rs4606) in the three prime untranslated region of RGS2 has previously been associated with reduced RGS2 expression [7]. This study, which included both men and women, also found that those with hypertension more often carried the G allele compared to normotensive controls [7]. We recently identified the CG or GG genotype to be more common among women with preeclampsia compared to women without preeclampsia [8]. This difference in the genotype distribution between cases and controls could possibly also contribute to the increased risk of hypertension later in life in women with a history of preeclampsia. Based on the majority of our previous study population, the population-based HUNT2 study, we wanted to explore whether rs4606 in RGS2 is associated with development of hypertension after pregnancy, which to our knowledge has not been investigated previously.

In 2011, The American Heart Association added pregnancy complications such as preeclampsia to the list of risk factors for developing CVD [12]. A mechanistic explanation for the relationship between preeclampsia and later CVD is lacking [13]. We recently reported that women with a pregnancy complicated by preeclampsia differed in the genotype distribution of the rs4606 in RGS2, with a slightly higher proportion of women having either the CG or GG genotype, compared to women who never had a preeclamptic pregnancy [8]. This finding, along with an association between the G allele of rs4606, reduced RGS2 expression and essential hypertension, as demonstrated by Semplicini et al. [7], led us to explore the possible association between this polymorphism and hypertension in women with a history of preeclampsia or uncomplicated pregnancy from our previously studied population. Our present findings support an association between the G allele of rs4606 and hypertension after pregnancy by showing that parous women have an increased risk of developing severe hypertension, in particular stage 2 systolic hypertension [11], if they have either the CG or GG genotype, compared to women with the CC genotype. Our data further showed that the association to current stage 2 systolic hypertension was strongest among individuals homozygous for the G allele, suggesting an allele dose–response relationship.

Merging of women with a history of preeclampsia and women without a history of preeclampsia is not without limitations, as these groups represent two highly selected populations, with preeclamptic women being at higher risk of later hypertension in general. Obstetric history was therefore accounted for in the regression model, and adjustment for a history of preeclampsia did not affect the significant relationship between genotype and hypertension. The magnitude of the association between the CG/GG genotypes and stage 2 hypertension was also similar in women with a history of preeclampsia and normotensive pregnancy if analyses were carried out separately for the two groups. The results did however not reach statistical significance in either group alone, which likely is due to limitations in statistical power resulting from the stratification of analyses. Our results also suggested a tendency towards a higher risk of milder hypertension defined as systolic blood pressure ≥140 mmHg and/or diastolic blood pressure ≥90 mmHg and/or current use of antihypertensive drugs in GG versus CC genotype carriers, with the strongest association in women with a history of preeclampsia. It is important to note that our study population was selected based on a previous diagnosis of preeclampsia or not, and not based on a retrospective case–control (hypertensives vs. normotensives) design. This could potentially have diluted an association between genotype and hypertension, especially in the control group. Our exclusion criteria, such as excluding women in the MBRN registered with pre-gestational chronic hypertension, renal disease, heart disease, any form of diabetes mellitus or non-proteinuric gestational hypertension, may also have diluted an association between genotype and hypertensive phenotype.

In our present study the association between the CG or GG genotype and stage 2 hypertension was detected only after multivariate adjustment and an interaction between rs4606 and physical activity was found. This is in line with remote hypertension after pregnancy being the result of multiple factors, with potential interaction effects. Previous studies have also suggested interactions between physical activity and genetics on blood pressure [14]. Moreover, it has been demonstrated that the blood pressure lowering effect of endurance training is greater in subjects with initial high blood pressures [14]. Results from our present study could therefore potentially suggest that in women with hypertension, and in particular severe, stage 2 hypertension, a more pronounced beneficial effect of physical activity on blood pressure reduction may be present in reference CC genotype carriers compared to CG or GG genotype carriers. An alternative interpretation is that despite being physically active, women carrying the CG or GG genotype will still be at a higher “baseline” (genetic) risk of developing hypertension compared to women carrying the CC genotype. In contrast, the elevated risk of hypertension associated with a sedentary lifestyle may dilute the difference in risk of hypertension that is seen between individuals with different genotypes, as those being “genetically protected” (i.e. CC genotype carriers) may significantly increase their risk of hypertension by being physically inactive, while physical inactivity may not add a similar risk in those already at risk (i.e. CG/GG genotype carriers). To our knowledge, no previous studies have investigated the potential interaction between this polymorphism (rs4606) and physical activity with respect to blood pressure reduction or the risk of hypertension, neither in women (whether parous or not), nor in men. Intervention studies are needed in order to elucidate this association. Nevertheless, all postpartum women should be encouraged to adhere to an active lifestyle, as physical activity in general is associated with a reduced risk of hypertension, in addition to being associated with a reduced risk of other lifestyle related conditions, such as obesity.

Several studies, including previous results from the HUNT study population, have shown that women with preeclampsia have more risk factors for CVD than women with uncomplicated pregnancies both pre-pregnancy and post-pregnancy [15, 16], the latter also supported by our present study (Table 1). A history of preeclampsia remained however a strong predictor of hypertension after adjustment for the rs4606 genotype, age and other variables classically found to modulate CVD risk. This finding suggests that the preeclamptic event per se, or the presence of other risk factors not explored in the present study, may additionally increase the risk of hypertension later in life, and thereby the risk of CVD.

The RGS2 is a candidate for regulation of signaling through the Ang II type 1 receptor (AT1-receptor) [17, 18]. Hercule et al. demonstrated that deletion of RGS2 in mice resulted in an augmented response to Ang II and altered blood pressure regulation, indicating that signaling by G protein-coupled receptors is abnormally prolonged [6]. Several studies have shown that increased Ang II sensitivity and activating autoantibodies against the AT1-receptor (AT1-AA) are present during and after preeclampsia [19–23]. Future studies should investigate if the presence of AT1-AA influences the association of hypertension and rs4606.

There is epidemiological support for early-onset preeclampsia and recurrent preeclampsia being more strongly associated with CVD later in life than a diagnosis of preeclampsia in general [4]. Even in our relatively healthy and young parous population, we found in crude and minimally adjusted analyses respectively, that women with a history of recurrent-preeclampsia and early-onset preeclampsia were at a significantly increased risk of later developing hypertension (blood pressure ≥140/90 mmHg, and/or current use of antihypertensive drugs) if they carried the CG or GG genotype compared to the CC genotype. Unfortunately, the very low numbers of women in the recurrent preeclampsia and early-onset preeclampsia sub-groups limited further analyses on these groups, such as exploring other cut-offs for hypertension or performing comprehensive logistic regression analyses.

We have previously shown an association between the CG/GG genotype and the presence of uteroplacental acute atherosis in women with a preeclamptic pregnancy, with the strongest association in the early-onset preeclampsia group [8]. Acute atherosis is characterized by foam cell accumulation and lipid depositions in decidual segments of spiral arteries, resembling early stages of atherosclerotic disease [24]. Whether presence of acute atherosis in pregnancy could reflect an increased risk of atherosclerotic CVD in the woman later in life is currently unknown, but we believe a common pathogenesis is plausible [24]. Interestingly, Kamide et al. showed an association between intima media thickening of carotid artery and polymorphisms in RGS2, including 1891-1892del TC which is in linkage disequilibrium with the functional variant rs4606 [25]. In the current study, based on the HUNT2 population, a diagnosis of uteroplacental acute atherosis and risk of future hypertension or CVD could not be explored, as this is not a routine clinical examination after delivery. We were also not able to explore an association between rs4606 and clinical CVD other than for hypertension, as our present study included very few clinical cardiovascular endpoints, likely due to a relatively young age group at HUNT2 inclusion, with a mean age of 41 years. Women with clinical CVD may also self selectively have been underrepresented in the HUNT2 study. Although a genetic predisposition for hypertension may be more evident for the younger age group, we cannot rule out the possibility of stronger associations in general, or according to pregnancy history, between our genotypes studied and hypertension in an older and less healthy (at parity) population.

Some caution is needed when interpreting our findings. Information on whether or not the study subjects were taking antihypertensive drugs was based on self-report. The definition of hypertension in those reporting no use of antihypertensive drugs was based on readings at only one time point (at enrollment in the HUNT2 study), using the average of the two last of three blood pressure readings. In a clinical setting, a diagnosis of hypertension, and the potential start of pharmacological treatment, is usually based on repeated measures of elevated blood pressure. Physical activity was also self-reported, based on a short questionnaire. A validation study of the physical activity questionnaire in the HUNT2 study has previously been carried out [26], and the “hard physical activity” question was found to be a useful measure of vigorous physical activity. In contrast, based on results from this report, we did not include the answer option “light” physical activity in our analyses, as this question was found to have poor reproducibility and to have poor correlation with most comparison measures [26]. Total physical activity level was therefore not accounted for in the present study. Additionally, for our study, several women (approx. 700) were lost in the regression analysis due to missing data for the hard physical activity question. Another limitation of our study may be the sampling of blood in the non-fasting state in the HUNT2 study. We decided to include the ratio of total cholesterol to HDL in our analyses, as it has been shown that serum concentrations of these lipid compounds are little influenced by recent food intake [27]. Inclusion (as in the multivariate model) or exclusion (as in the main effects model) of this variable did however not influence the association shown between the rs4606 and stage 2 hypertension. Despite these limitations, a major strength of our study is a well-described population-based cohort (HUNT2), with selection of a healthy parous subgroup (except for preeclampsia in the case group) and a large total sample size, enabling the adjustment for several important postpartum follow-up variables in the HUNT study, such as age, BMI, smoking, parities and postpartum diabetes.

In our present study we have shown that the rs4606 CG/GG genotypes of RGS2, in particular the GG genotype, are associated with hypertension in women after pregnancy, even after excluding women who had a pregnancy record of chronic hypertension, cardiac or renal disease or pregestational diabetes mellitus. Our study in a relatively healthy population therefore strengthens a previously reported association between the G allele of this polymorphism and essential hypertension, and for the first time demonstrates an association between this polymorphism and hypertension outside of pregnancy in a Norwegian population. As the CG or GG genotypes are more prevalent in women with a history of preeclampsia compared to women with a history of normotensive pregnancy, these genotypes may represent one piece of a puzzle for the epidemiologically reported higher prevalence of hypertension later life in women with preeclampsia. Dysregulated angiotensin signaling could be a potential mechanistic link that merits further investigation. Thus, this polymorphism in the RGS2 gene may be a common predisposing factor for the development of hypertension in pregnancy (preeclampsia) and hypertension outside of pregnancy. The present study also support a multifactorial relationship between preeclampsia and an increased risk of hypertension later in life, as a history of preeclampsia remained a strong and independent risk factor for hypertension after accounting for this polymorphism and classical risk factors for CVD, such as BMI, age and diabetes. The relative contributions from genetic factors, environmental factors and pregnancy factors on the long term risk of hypertension after preeclampsia needs further examination, and may also differ between populations as well as within populations.