Mouse models for preeclampsia: disruption of redox-regulated signaling

Approximately 5–7% of pregnant women worldwide suffer from common hypertensive pregnancy disorders culminating in preeclampsia (PE), intrauterine growth restriction and premature child birth. PE is the major cause of maternal mortality (80%) in developing nations and in recent years, the perinatal mortality and morbidity in developed countries have increased by five-fold [1, 2]. Moreover, the incidence of PE has increased by 40% in the last 15 years [3]. The most widely accepted cause of pre-eclampsia is reduced utero-placental circulation (superficial implantation of the fetus) due to sub-optimal vascular remodeling of the decidual and the uterine arterioles, secondary to inadequate trophoblast invasion [4]. Increased oxidative stress and an altered immune response [5] at the fetal-maternal interface (Th1 bias) are likely effectors contributing to the development of systemic endothelial and renal dysfunctions in the later phase of the disease.

A series of recent discoveries, specifically the isolation and functional characterization of non-phagocytic NADPH oxidase-homologues in epithelial, endothelial, fibroblast and muscle cells, argue that reactive oxygen species (ROS) are indispensible to both physiological and patho-physiological conditions such as growth, differentiation, apoptosis and senescence [6‐8]. The transition from growth to degeneration is finely tuned by the relative concentration of oxidants. For example, in ambient conditions, low levels of H₂O₂ (nano-micromolar) are necessary for angiogenesis [9], while agonist-induced activation resulting in its excessive accumulation (> 150–200 μM) prompts endothelial damage [10]. The scheme shown in Fig. 1a depicts the essential role played by ROS in both early and late pregnancy together with the detrimental effect when ROS are produced in excess (oxidative stress).

During pregnancy, the maternal energy demand increases significantly to sustain the growing fetus. This demand is met via a substantial increase (30–50% compared to the non-gravid state) in the uterine blood flow in the pregnant women [11]. Consequently the 'shear stress' (dragging frictional force generated by the blood flow), is negotiated by endothelial cell (EC)-derived vasodilatory agonists, nitric oxide (NO^.) and prostaglandin I₂. The release of vasodilators from EC is controlled by mitochondrial Ca⁺²-influx (ATP-GPCR and ISP3-ER pathways) as well as NADPH oxidase-dependent mitochondrial O₂^.- production. The pregnancy-induced adaptation of ROS-regulated Ca⁺² signaling in the mitochondria of EC is essential for re-establishing physiological laminar flow in uterine vessels [12]. In extreme circumstances, dysregulation of mitochondrial Ca⁺² homeostasis due to high ROS, could lead to Ca⁺² overload of the matrix, triggering apoptosis. The first indication of a mitochondrial involvement in preeclampsia/eclampsia came from a case report where the frequency of the disease was high in a family with inborn mitochondrial defects [13]. A number of in vitro studies on a hypoxia-reoxygenation model of placental culture and plasma oxidant/antioxidant analyses further pointed to mitochondrial involvement by suggesting that ROS could influence trophoblast fusion, migration and apoptosis relevant to preeclampsia [14, 15]. In support of this, it is pertinent that homozygous deletion of SOD-2 (MnSOD) has the most severe effect on embryonic development in mouse pregnancy compared to that of SOD-1 and SOD-3 knockouts [16].

Reactive oxygen species (ROS) are key mediators of growth factor-dependent redox-regulated signaling in angiogenesis. While > 90% of O₂ is reduced in mitochondria, NADPH oxidases (Nox2, gp91phox) are the major source of ROS in endothelial cells [17, 18]. The O₂^.- generated at the outer surface of the plasma membrane is internalized through ion channels, a process which increases the intracellular Ca⁺² release, activating mitochondrial O₂^.- production. Thus, NADPH oxidases and mitochondria together perpetuate a cascade of O₂^.- production in vascular endothelial cells [19, 20].

Hypoxia inducible factor -1α(HIF-1α) is a dynamic partner of the heterodimeric transcription factor HIF-1 which is essential for angiogenesis [21]. Irrespective of the oxygen tension (hypoxic or normoxic), the stability of HIF-1α is determined by it ligation to von Hippel-Lindau tumor suppressor protein, VHL [22]. Once bound to VHL, HIF-1α undergoes ubiquitination prior to proteosomal degradation. A prerequisite for VHL binding is site-specific hydroxylation of HIF-1α by prolyl hydroxylases (PHD). The transfer of these hydroxyl moieties to HIF-1α by PHD requires O₂ and 2-oxogluterate as co-substrates, together with reduced iron (Fe⁺²) and ascorbate as cofactors [23, 24]. Therefore, prolyl hydroxylases act as negative regulators of HIF-1α since active PHD-Fe⁺² promote HIF-1α binding to VHL and subsequent degradation. It is apparent that the enzymatic activity of PHD could be modulated by the relative concentration of co-substrates as well as cofactors. Indeed, the stability of HIF-1α in angiotensin II-treated vascular smooth muscle cells under normoxic conditions is due to H₂O₂-mediated reduction in cellular ascorbate concentration and increased Fe⁺³ [25]. It is noteworthy that commonly encountered ROS such as O₂^.- and OH^. anions are short-lived and, owing to their limited diffusion capacity, fail to cross the plasma membrane except through ion channels. Therefore, the majority of cellular ROS-effects are mediated via relatively stable H₂O₂. This is consistent with genetic, molecular and pharmacological experiments [26‐29] suggesting that H₂O₂ plays a central role in stabilizing HIF-1α by converting PHD-Fe⁺² to PHD-Fe⁺³.

Clinical research on preeclampsia (PE) is mostly restricted to observations, in vitro correlative studies on placenta and patient serum samples originating from mid-to-late gestation when the disease is fully manifested, long after it initiates at early pregnancy. These problems have underlined the necessity for animal models of PE where the progression of pathogenesis can be traced through longitudinal studies from early pregnancy. Three recently developed mouse models for PE [30‐32] exhibit hallmarks of the human disease. In one of these models, the critical symptoms of PE (hypertension, proteinurea due to glomerulosclerosis and fetal resorption) were ameliorated by continued treatment with an SOD mimetic, tempol [31], suggesting that the PE-like symptoms of these model mice are precipitated by disruption of redox-regulated signaling during pregnancy. A further link between redox-regulated signaling and human pregnancy pathology has been provided by the most recently developed animal model for PE [32]. The homozygous deletion of Catechol-O-methyl transferase (Comt^-/-) in pregnant mice results in the loss of 2ME2 which has direct involvement in redox-regulated signaling.

Despite the advantage of targeted disruption of a specific gene (Comt) in the mouse system and similarities in placental cell types in mouse and human, this animal model might be insufficient in understanding all aspects of the development of human preeclampsia and fetal growth restriction. The limitations of the mouse model for preeclampsia stem from intrinsic differences in the physiology of mouse and human pregnancies [33]. The mode of implantation [34, 35] differs between mouse and human; there is shallow decidual invasion in mouse compared to extensive uterine remodeling by invasive trophoblasts in human. In addition, the uterine spiral arterioles in mice are remodeled by maternal factors rather than by invasive trophoblasts in human pregnancies [36], and the endocrine control of mouse pregnancy is mediated by a limited number of placental hormones compared to that of human [37]. Moreover, the short gestation period (3 weeks) in mice, the incomplete development and birth of altricial young are unlike human pregnancies [38]. All of these factors limit mouse pregnancy as a model system for human fetal growth restrictions.