Placental hypoxia-regulating network in relation to birth weight and ponderal index: the ENVIRONAGE Birth Cohort Study

Fetal growth is dependent on genetic, placental and maternal factors. The fetus is thought to have an inherent growth potential that, under normal circumstances, results in a healthy newborn of appropriate size. The maternal–placental–fetal units act in harmony to provide the needs of the fetus while supporting the physiologic changes of the mother. The most common cause of low birth weight (LBW) at term in Western societies is placental insufficiency [1]. LBW infants have a 10–20-fold increased risk of dying in the perinatal period [2] and are at increased risk of developing chronic diseases including type 2 diabetes, hypertension and heart disease in later life [3]. Fetal growth restriction is the second leading cause of perinatal morbidity and mortality [4]. The ponderal index has been installed as an indicator of fetal growth particularly able to identify normal weight newborns (birth weight above 2500 g) with higher probability to develop disease such as hypoglycemia and hyperbilirubinemia (lower ponderal index) [5]. Placental development is dependent on tight regulation of oxygen tension [6]. Hypoxia-inducible factor 1α (HIF1α) plays an important role in control of placental oxygenation. Normal pregnancy is hallmarked by a state of oxidative stress, in which placental mitochondria generate reactive oxygen species [7]. Feto-placental hypoxia, hallmarked by increased levels of HIF1α, has been associated with development of preeclampsia [8]. A placental mRNA/miRNA network involving microRNA (miR)-210, iron sulfur-cluster assembly enzyme (ISCU) and HIF1α was proposed to regulate mitochondrial function during preeclampsia [9]. HIF1α is at the core of the hypoxia-responsive network able to target mRNAs and miRNAs executing diverse cellular functions [10]. HIF1α not only induces expression of miRNAs-20a and -210, these miRNAs in turn were shown to downregulate HIF1α [11, 12]. Preeclampsia is one of the most common pregnancy complications and it is believed that an inadequate blood flow in the placenta is responsible for this [13]. Preeclampsia could lead to fetal growth restriction. Fetal growth restriction is associated with increased fetal and perinatal morbidity and mortality, even in the absence of preeclampsia [14]. Little is known on the relation between HIF1α expression in placental tissue from normal term pregnancies and fetal growth. As tight regulation of oxygen levels is essential for proper fetal-placental development, we hypothesize placental expression levels of the hypoxia-responsive network are associated with measures of fetal growth in normal pregnancies. We analyzed expression of 7 members from the hypoxia-regulation network in placental tissue (Fig. 1), by examining miR-20a and miR-210 and mRNA HIF1α, cytochrome C oxidase-10 (COX-10), ISCU, phosphatase and tensin homolog (PTEN), and DNA repair protein RAD52 expression levels.

We here report on the expression of the placental hypoxia-responsive network involving HIF1α and several of its target genes/miRNAs and its association with fetal growth parameters. The relationship between low birth weight or low ponderal index and chronic disease has been established in several publications [5, 19, 20]. As such, an inverse relationship between birth weight and blood pressure in childhood was already observed in the 1980s by David Barker [21], which can lead to development of cardiovascular disease (reviewed in [22]).

Regulation of oxygen tension is tightly controlled throughout gestation, and follows a distinct pattern. During the first trimester, the placenta develops under low oxygen tension, a physiologic form of hypoxia, oxygen tension, as well as HIF1α expression, increase towards the end of the first trimester (week 11 of gestation) [23, 24]. From this point onwards, the placenta is well-oxygenated throughout the remainder of gestation and a physiologic state of normoxia exists. If placental oxygen tension drops towards the end of pregnancy, a hypoxic or ischemic condition arises that can be detrimental for both mother and fetus. HIF1α expression achieves two peaks during the first trimester, one around week 7–10 of gestation, another one around week 14–18. Levels drop much lower around the time of birth [24]. PTEN expression on the other hand was reported higher during the first trimester pregnancy as compared to any time in the normal menstrual cycle [25]. PTEN expression of villous trophoblasts was decreasing as the pregnancy advanced. PTEN expression decreased parallel to the development of placenta. Expression of PTEN in decidual cells was significantly stronger in placental tissues of spontaneous abortion than placental tissues from normal pregnancies at the first and third trimester [26]). miR-210 expression remains at fairly constant levels throughout healthy pregnancies, whereas its expression is increased in pregnancies complicated by preeclampsia between 26 and 38 weeks of gestation, this association disappears after 38 weeks of gestation [27].

We observed significant negative associations between placental HIF1α, RAD52 and the ponderal index in newborn girls and boys. Moreover, we found a significant positive association between placental expression of miR-20a and birth weight in newborn boys only. Our findings support the importance of expression of these genes in the placenta for normal fetal growth. Decreased expression of HIF1α has previously been reported in placental tissue from pregnancies complicated by preeclampsia compared to normal pregnancies [8, 28]. Preeclampsia is hallmarked by hypertension and proteinuria and is associated with significant maternal morbidity and mortality [29]. Preeclampsia is characterized by placental hypoxia and/or ischemia, excessive oxidative stress, in association with endothelial dysfunction [30]. Whether HIF1α expression is associated with intrauterine growth restriction (IUGR) is not clear. Some studies report no detected differences in placental HIF1α expression between IUGR babies versus babies born from normal singleton pregnancies [31], whereas others have reported increased expression of HIF1α in maternal plasma [32]. Zhang et al. reported higher HIF1α expression in placental shares from twins with IUGR compared to normal twins, although no significant correlation with birth weight could be established [33]. These apparent differences could stem from differences in tissue type, study population and study size.

The incidence of childhood cancer is estimated to increase but the causes remain uncertain [34, 35]. Insults to the genome in the perinatal period are likely to contribute to carcinogenesis and may be more important relative to other life stages because of the higher probability that mutated and genomically unstable cells could populate the rapidly growing tissues of an infant [36]. Mutations of the genome may have particularly adverse consequences in early life, including developmental defects and immune system dysfunction [36, 37]. Gene expression related to DNA damage and immune response among children is observed to correlate with micronuclei frequency (MN) as a consequence of exposure to environmental pollutants [38, 39]. As for RAD52, this is the first time that expression of this DNA damage marker was reported to be associated with measures of fetal growth. RAD52 is a key player in DNA doublestrand break repair and homologous recombination. It forms a heptameric ring, catalyses DNA annealing and mediates Rad51-catalysed strand invasion [40]. The process of DNA repair has previously been linked with changes in birth weight, as a genome-wide methylation study in the Norwegian mother and child cohort identified increased levels of X-ray repair cross complementing 3 (XRCC3) methylation to be associated with increased birth weight [41]. They noted an inverse association between DNA damage gene expression (through increased methylation) and birth weight consistent with our findings. Furthermore, a positive association between DNA damage markers and birthweight was recently reported in the Australian DADHI cohort, where genome damage was measured by scoring MN as a biomarker of both chromosome breakage and/or loss; nucleoplasmic bridges as a biomarker of DNA mis-repair and/or telomere end fusions and nuclear buds were measured as biomarker of gene amplification and/or the removal of unresolved DNA repair complexes. Infant birth weight was associated positively with all the aforementioned chromosomal lesions in cord blood from term newborns [42].

We did not identify any associations between placental miR-210 expression and fetal growth characteristics, although miR-210 is a target of HIF1α [43] and has been shown to be increased in placental tissue from preeclamptic women [9]. As miR-210 itself is known to target HIF1α and thereby functions in a negative feedback loop, and other mRNAs such as Akt and p53 can also regulate miR-210 expression [44, 45], this could explain why we did not observe an association with fetal growth in our current analyses.

As such, our study involves a large sample size and expression profiling was performed in placental tissue, the main regulator of fetal development. The current findings must be interpreted within the context of its limitations. At this time we cannot make any predictions on the role of the hypoxia-regulating network in postnatal child growth and adult disease risk. As we only collect placentae at birth, we cannot make any conclusions on dynamics of expression throughout pregnancy.

We observed significant associations between hypoxia-network members HIF1, RAD52 and miR-20a expression and ponderal index as well as birth weight as measures of fetal growth in term newborns. The inverse relationship between fetal growth and expression of hypoxia-network members HIF1, RAD52 and miR-20a indicates that this network is important in potential intrauterine insults.