The Clinical Value of Rodent Models in Understanding Preeclampsia Development and Progression

Purpose of Review Preeclampsia (PE) is a leading global cause of maternal and fetal morbidity and mortality. The heterogeneity of this disorder contributes to its elusive etiology. Due to the ethical constraints surrounding human studies, animal models provide a suitable alternative for investigation into PE pathogenesis and novel therapeutic strategies. The purpose of this review is to compare and contrast the various rodent models used to study PE, in order to demonstrate their value in investigating and identifying different characteristics of this disorder. Recent Findings Several approaches have been employed to create an appropriate animal model of PE, including surgical, genetic manipulation, and pharmacological methods in an attempt to mimic the maternal syndrome. Despite the absence of a model to completely model PE, these models have provided valuable information concerning various aspects of PE pathogenesis and novel therapeutic strategies and have led to the discovery of potential predictive markers of PE. Summary Rodent and murine models have contributed significantly to the study of the pathology associated with specific aspects of the disorder. As a single fully encompassing animal model of PE remains absent, the use of a combination of models has potential value in understanding its etiology as well as in new treatment and management strategies.


Introduction
Preeclampsia (PE) is a hypertensive disorder of pregnancy and a principal cause of maternal and fetal morbidity and mortality worldwide, resulting in approximately 46,000 maternal and 500,000 neonatal deaths per annum [1,2]. Clinically, PE manifests as new-onset hypertension developing after 20 weeks of gestation (defined as blood pressure ≥ 140/90 mm Hg) together with one or more of the following: proteinuria (≥ 300 mg/day), maternal organ dysfunction, or uteroplacental dysfunction [3, 4••]. Thus far the only effective treatment is premature delivery and consequent early placental delivery, which is associated with the risk of neonatal morbidities [5].
The development of PE is a complex process that involves a number of dysfunctional physiological processes. A key event is abnormal placentation [6], characterized by reduced trophoblast invasion, inadequate remodeling of the maternal spiral arteries, and consequent placental ischemia [7]. This reduction in placental perfusion leads to placental hypoxia and oxidative stress [6]. The hypoxic placenta consequently secretes increased levels of anti-angiogenic markers, including soluble fms-like tyrosine kinase (sFlt-1) and soluble endoglin (sEng), into the maternal circulation, thereby inhibiting the bioavailability of proangiogenic vascular endothelial growth factor (VEGF) and placental growth factor (PlGF) [8,9]. This angiogenic imbalance precedes the maternal syndrome of new-onset hypertension with or without the presence of proteinuria and with systemic endothelial dysfunction [10]. Preeclampsia is further categorized into early-onset PE, which can be diagnosed prior to 34 weeks of gestation and late-onset PE that is diagnosed from 34 weeks [11]. Early-onset PE is associated with abnormal placentation and fetal growth restriction, while late-onset PE is linked with maternal endothelial dysfunction [12].
While the precise etiology remains unclear, poor maternal and fetal outcomes continue to be exacerbated by inaccuracies in the identification and early diagnosis of women at high risk of PE development, especially in low-and middleincome countries. Therefore, identifying an appropriate animal model that mimics aspects of PE etiology will advance the current understanding of the conceptual framework underlying its development. However, since PE is a disorder of extreme heterogeneity, developing a gold standard model in this field remains challenging.
Current therapeutic interventions include anticoagulants such as aspirin, antihypertensive drugs such as labetalol, methyldopa, and nifedipine, as well as magnesium sulfate to prevent seizures [13]. Low-dose aspirin is reported to exert a prophylactic effect, by lowering the risk of early-onset PE development if administered before the 16th week of gestation [14]. However, it is ineffective in decreasing the risk of late-onset PE development [15]. Moreover, PE is identified by the American Heart Association, as a risk factor for impending circulatory disorders [16] and stroke [17]. Thus far, treatment strategies are only effective in managing the symptoms associated with PE and cannot be used to cure this disorder. With the absence of definitive treatment options and ethical limitations associated with the research in pregnancy, the development of animal models that cover the pathological aspects of PE is essential to increase our understanding of the disorder. Moreover, these models will enable the evaluation of novel treatment strategies in determining the safety and effectiveness of treatment interventions prior to clinical trial testing [18••].
The progress made towards understanding the development and progression of PE has seen a greater demand for the identification of novel agents for the treatment or prevention of PE [5]. Effective therapeutic interventions apart from delivery of the placenta would significantly improve maternal and neonatal health and pregnancy outcomes [5]. Despite the availability of extensive literature surrounding various models of PE development, this review will examine and provide a summary of the most extensively studied, as well as novel rodent models that have been used to study hypertension, proteinuria, maternal organ dysfunction, and fetal growth restriction in PE development.

Animal Models of PE
Animal models to study PE development have been implemented using various methods. Preeclampsia may be induced via surgical and [19][20][21], pharmacological intervention [22, 23••], genetic, and immunological modification [24, 25••, 26, 27] and via the use of animals with pre-existing hypertension that develop superimposed PE [28,29]. These models exhibit aspects of PE such as pregnancy-induced hypertension, elevated urinary protein levels, renal dysfunction, placental ischemia, and fetal growth restriction [18 ••]. However, the primary focus of several of these PE models lie in reproducing the maternal syndrome, and therefore, only a limited number of studies address initiating factors and primary stages of PE development [30].
Examining the initiation and progression of fetoplacental disorders in the first trimester in humans remains a challenge due to the potential danger associated with both maternal and fetal wellbeing [30]. Moreover, human clinical studies are associated with constraints that hinder a comprehensive investigation of the time-dependent processes that occur in PE development [31••]. In contrast, rodent models of PE support the study of the mechanisms that initiate development and progression of this disorder [23••, 25••, 32], since they have short gestations relative to humans. This permits the investigation of specific aspects of this multifactorial disorder, forming the preclinical basis for experimental testing and supporting the development of predictive tests and therapeutic strategies [6]. Moreover, in vivo, in vitro, and molecular methods may be explored to examine the underlying processes elicited by novel therapeutic interventions as well as effecting proof of concept experimental studies Animal models have provided valuable insights into our knowledge of PE, including the mechanisms of deficient trophoblast migration [24,33,34] and placental ischemia [19,35]. They have added to our understanding of endothelial dysfunction arising from the release of various factors from the hypoxic placenta [31••, 36]. Moreover, they have contributed significantly to the development of novel diagnostic strategies such as angiogenic screening platforms and immunoassays to aid in PE diagnosis [25••]. The use of Triage PlGF and the Elecsys immunoassay sFlt-1/PlGF ratio tests in conjunction with standard clinical assessment and subsequent clinical follow-up have been fully endorsed by the National Institute for Health and Care Excellence (NICE) [37]. Moreover, their clinical potential in helping to diagnose or rule out PE in women presenting with suspected symptoms between 20 and 34 weeks is undeniable. Albeit, the lack of sufficient evidence regarding their accuracy prevents their use as standard procedure for PE diagnosis [37].
Additionally, the accuracy of the BRAHMS Kryptor sFlt-1 and PlGF assays has been investigated, and the results have been found to be comparable to the Elecsys assays for sFlt1 and PlGF [38, 39 •]. When used in combination with standard clinical methods of evaluation, the KRYPTOR assays have also displayed their utility in predicting the risk for PE-associated short-term adverse maternal and perinatal outcomes occurring within 2 weeks of presentation in women with suspected PE [38]. A recent study has demonstrated the KRYPTOR assays ability to rule in or rule out PE within a week and also proposes the clinical implementation of a simpler single decision sFlt1/PlGF ratio threshold of 66 as opposed to the currently used gestation-specific dual thresholds [39•].

Rodent Models of PE
Rodent models have substantially contributed to the conceptual framework underlying the pathogenic mechanisms enabling a better understanding of the clinical manifestation of PE (Fig. 1). A brief overview of these models is displayed in Table 1. Since PE is a placental disorder, it is crucial for PE models to share common anatomical and functional features of human placentation [30]. Humans, rats, and mice are among the species that share hemochorial placentation. In rodents, placental development commences with the invasion of trophoblast cells into the maternal decidua which is followed by the remodeling of the maternal spiral arteries [40]. Humans and rodents share a similar profile of immune cells, including uterine natural killer cells in the maternal decidua [40]. While mice have enabled the study of placentation [41,42], they exhibit shallow intrauterine trophoblast invasion, and placentation is superficial [43]. This is in direct contrast to human and rat placentation which is characterized by extensive trophoblast invasion and uterine arterial remodeling [44]. Furthermore, humans and rats share a similar discoid placental shape and deep invasiveness of the placenta but do differ in histological structure, as humans have a hemomonochorial structure, while rats display a hemotrichorial placenta [44,45].
Additionally, both mice and rats mimic human pregnancy-induced cardiovascular changes such as hypotension in early pregnancy, reduced pressor responses to angiotensin II, and decreased hematocrit, as well as elevations in cardiac output, stroke volume, and plasma volume [46,47]. • New onset hypertension (Systolic BP t140 mmHg and diastolic BP t90 mmHg at t 20 weeks gestation) -Accompanied by one or more of the following: • Proteinuria (t300mg protein/24hr) • Maternal end-organ dysfunction such as: -Acute kidney injury -Liver abnormalities (ALT or AST levels t40 IU/L) -Neurological complications -Uteroplacental dysfunction -Pulmonary edema -Hematological complications (e.g. reduced platelet count, hemolysis)

Fig. 1
Various rodent models employed to study the pathogenic pathways associated with PE development 1 3 Thus, the extensive use of mice and rats to study hypertensive disorders of pregnancy such as PE is not surprising [45,48]. Numerous experimental models of PE are also associated with elevated tissue levels of prepro-endothelin-1 mRNA. These models have been used to investigate whether the inhibition of the endothelin pathway could improve Table 1 A brief overview of current rodent and murine models of PE Ad-sFlt-1 adenovirus expressing soluble fms-like tyrosine kinase-1, AT1-AA angiotensin II type 1 receptor autoantibody, BPH blood pressure high, CNS central nervous system, CXCL1 C-X-C motif chemokine ligand 1, GFR glomerular filtration rate, IFN interferon, IL interleukin, L-NAME L-NG-nitroarginine methyl ester, MCP-1 macrophage chemoattractant protein-1, NO nitric oxide, PlGF placental growth factor, ROS reactive oxygen species, sEng soluble endoglin, sFlt-1 soluble fms-like tyrosine kinase, STOX1 storkhead box-1, TLR toll-like receptor, TNF-a tumor necrosis factor-a, VEGF vascular endothelial growth factor hypertension. Additionally, the reduced uterine perfusion pressure model, soluble fms-like tyrosine kinase rat model, BPH5 mouse model, and the nitric synthase inhibition (L-NAME) model have been explored as models to mimic PE development. The mean arterial pressure is reduced in the RUPP, sFlt-1 infusion, TNF infusion, and AT1-AA infusion models by the administration of the endothelin type A receptor antagonist [19,[49][50][51]. This implies that endothelin-1 is a potential common pathway in which placental factors exert their effects on the maternal vasculature to induce vasoconstriction and hypertension [31••]. These rodent models are also used to test vitamins (D and B) and drug (e.g., statins) interventions. Vitamin D administration in the RUPP model lowers blood pressure, endothelin-1, sFlt-1, and AT1-AA levels; however, fetal outcomes were not improved [52][53][54]. In contrast, the L-NAME model demonstrated reduced levels of sFlt-1 and TNF in response to vitamin D treatment [55].

Reduced Uterine Perfusion Pressure Model
The reduced uterine artery perfusion (RUPP) rat model is the most widely characterized experimental model of placental ischemia. It reproduces the PE phenotype of endothelial dysfunction, glomerular endotheliosis, hemodynamic changes, and higher circulating levels of sFlt-1 and sEng [21]. This model is produced by clipping the aorta and uterine ovarian arteries on gestational day 14, resulting in a 40% decrease in uteroplacental perfusion along with a 20-30 mmHg increase in maternal mean arterial pressure on gestational day 19 in comparison to control groups [21]. The RUPP model has made substantial inroads in elucidating the role of the adaptive immune system in PE development [30]. Zenclussen and co-workers studied the role of inflammatory T cells in PE development and showed that the transference of activated T helper-1-like cells into healthy mice leads to the development of PE-like symptoms [56]. The RUPP model demonstrates elevated levels of inflammatory CD4 + T cells [57], along with reduced anti-inflammatory regulatory T cell levels [58]. This model also found that the adoptive transfer of CD 4 + T cells from RUPP rats to control rats induced hypertension, proteinuria, glomerular endotheliosis with concomitant elevated cytokines, and anti-angiogenic expression in circulation [59]. Albeit a major limitation of this model is its inability to replicate the immune mechanisms, deficient trophoblast invasion, and abnormal remodeling of the spiral arteries since the clipping of the lower abdominal aorta and uterine arteries is conducted mid-pregnancy (gestational day 14) [21]. Moreover, liver dysfunction and intrauterine growth restriction associated with human PE development are not replicated.
Moreover in the RUPP rat model, pravastatin treatment downregulates blood pressure and reactive oxygen species and improves angiogenic balance [60]; however, this was not observed in early human clinical trials [61,62]. Pravastatin administration in early-onset PE women reduces the incidence of poor fetal outcomes but does not influence the concentration of plasma sFlt-1 levels or the sFlt-1:PlGF ratio in comparison to the placebo group [61,62]. However, the focus still remains on angiogenic factors sFlt-1, PlGF, and VEGF due to the importance of angiogenic imbalance in PE pathogenesis. The administration of PlGF in RUPP rats has been reported to lower sFlt-1 levels, blood pressure, and proteinuria [63].

BPH/5 Mouse Model
This is a spontaneous or superimposed model of PE since an existing mild hypertension is present in the non-pregnant mice. Hypertension increases with pregnancy together with endothelial dysfunction, elevated uterine vascular resistance, placental dysfunction, and reduced litter size [29,64]. This model was created through the continued mating of inbred BPH (blood pressure high)/2 mice leading to the development of the BPH/5 mouse strain [65]. A study using the BPH/5 model found that the maternal phenotype of PE may be initiated by the increased decidual expression of cyclooxygenase-2 (COX-2) and interleukin-15 [6]. This model has also been used to determine the effectiveness of therapeutic intervention of proangiogenic factor, where adenoviral delivery of VEGF 121 inhibited the development of superimposed PE [66]. These findings highlight the therapeutic potential of early proangiogenic intervention in pregnancy-associated hypertensive disorders. Despite the pre-existing hypertension displayed by non-pregnant mice, the usefulness of this model in providing information of how pre-existing hypertension influences the pathogenesis of superimposed PE is valuable [30].

sFlt-1 Sprague Dawley Rat Model
The clinical importance of the sFlt-1/PlGF ratio encouraged research that evaluated its clinical value in the diagnosis and prediction of PE. This ratio has contributed substantially to the creation of automated angiogenic biomarker platforms (sFlt-1 and PlGF) to aid the diagnosis and prognosis of PE in high-income countries. Maynard and co-workers developed the novel experimental rat model which replicated clinical characteristics of PE [25••]. Their findings indicate that exogenous administration of sFlt-1 to pregnant rats produces hypertension, proteinuria, and glomerular endotheliosis, which are characteristic of PE [25••]. In contrast Thadhani and co-workers reported that sFlt-1 clearance by apheresis improve angiogenic balance in preeclamptic women and demonstrated lowered mean arterial pressure and extended gestation by up to 15 days [67,68].
Maynard's breakthrough study thus endorsed the development of the immunoassays, Alere Triage PlGF test and the Elecsys sFlt-1/PlGF from Roche, which have the potential to exclude PE diagnosis in women with suspected PE between 20 and 34 weeks [69]. The BRAHMS Kryptor sFlt-1 and PlGF assays have also demonstrated comparable accuracy to the Elecsys assays for sFlt1 and PlGF when ruling in or ruling out PE [38, 39•]. Additionally, this model was used to study the longterm effects of pregnancy-induced hypertension on maternal and fetal outcomes. The offspring of sFlt-1 treated mice exhibited elevated blood pressure levels and also reported that baseline maternal cardiovascular function was not adversely affected postpartum [70]. Off note, a shortcoming of this model is its inability to reproduce the liver dysfunction and intrauterine growth restriction associated with human PE development.

Nitric Oxide Synthase Inhibition (L-NAME)
The nitro-L-arginine methyl ester (L-NAME) model is a model of endothelial dysfunction associated with PE development. In PE, the nitric oxide pathway is defective and polymorphisms in nitric oxide synthase (NOS) exist [71,72]. Using a rodent model, the administration of L-NAME inhibits NOS and produces PE-like characteristics such as high blood pressure, proteinuria, decreased glomerular filtration rate, and intrauterine growth restriction [73][74][75][76]. Additionally, early-onset PE (EOPE) and late-onset PE (LOPE) phenotypes may also be produced by altering the timing of L-NAME administration in pregnant Sprague Dawley rats [32]. The L-NAME rat and mouse models of PE have been utilized in testing potential treatment and biomarker predictor tests for PE development. Administration of sildenafil improves hypertension, proteinuria, and fetal outcomes in early-and late-onset cases of PE [36,74,77,78] and downregulates plasma sFlt-1 and sEng levels [22]. Furthermore, urinary nephrin and podocin mRNA levels are significantly higher in both EOPE and LOPE L-NAME-treated rats in comparison to pregnant control rats, suggestive of the presence of podocyturia. A significant reduction in urinary mRNA levels of podocin in EOPE rats and nephrin levels in LOPE rats was demonstrated following treatment with sildenafil citrate [79].
In contrast, sildenafil treatment administered late in gestation to rats and pregnant women with PE is associated with poor outcome [80,81]. The uncertainty of NOS regulation in PE development raises concerns regarding the validity of the L-NAME model in PE investigations [ [71,72], inconsistencies are reported in studies involving the genetic deletion of NOS in pregnant mice [84][85][86]. Reduced blood pressure has been reported in pregnant NOS knockout mice [85], in contrast to higher blood pressure observed in non-pregnant NOS knockout mice [84]. Non-pregnant mice administered with L-NAME also show elevations in blood pressure with aortic vascular contraction, suggestive that the PE-like symptoms produced by this model may not be pregnancy-specific [86].

Tumor Necrosis Factor Alpha Model
Tumor necrosis factor alpha (TNF) is a pro-inflammatory cytokine involved in physiological processes such as cellular proliferation and differentiation, apoptosis, cell proliferation, differentiation, apoptosis, and inflammation [87]. Higher levels of TNF are reported in preeclamptic compared to normotensive pregnancies and hypertensive pregnancies uncomplicated by PE [88][89][90][91]. Women with pregnancies complicated by early-onset PE have significantly elevated levels of serum TNF in contrast to women with late-onset PE [92]. Normotensive pregnant rats administered with TNF (50 ng/d) between gestational days 14-19 develop hypertension and express elevated levels of renal, placental, and aortic prepro-endothelin-1 [93]. The administration of an endothelin type A receptor antagonist in pregnant rats ameliorates hypertension induced by TNF [51]. Additionally this model proposes that TNF-induced hypertension may emanate from the decline in renal NOS expression [93] and increase in AT1-AA production [94]. A major limitation observed in this model is that the PE phenotypes produced in response to chronic administration of TNF may be an exaggeration of the pro-inflammatory state of pregnancy [83 •].

Indoleamine 2,3-Dioxygenase Knockout Model
Indoleamine 2,3-dioxygenase (IDO) is a cytosolic hemeprotein which catalyzes the rate-limiting step in tryptophan breakdown [83•] and is essential in the T cell-mediated immune response [95]. The IDO mouse model is produced via IDO inhibition using 1-methyl-tryptophan [27]. These mice develop high blood pressure or proteinuria, and their placentae express edematous changes and fibrin deposits, and there is no fetal growth restriction [27]. Santillan and co-workers reported glomerular endotheliosis, intrauterine growth restriction, and proteinuria in IDO-knockout mice; no changes in placental morphology and blood pressure were observed [26]. Preeclamptic women display downregulated placental levels of IDO [96][97][98]; interestingly, this is not observed in placentae from pregnancies with fetal growth restriction but without hypertension [97][98][99]. Late-onset PE patients demonstrate significantly lower IDO expression on endothelial cells in comparison to women with early-onset PE [100]. A shortfall of this model is its inability to fully mimic the clinical characteristics of hypertension and placental abnormalities associated with human PE development; however, it supports investigations surrounding placental inadequacy and renal dysfunction in pregnancy.

Angiotensin II Type 1 Receptor Autoantibody Model
The angiotensin II type 1 receptor autoantibody (AT1-AA) mouse model was generated by administering AT1-AAs obtained from preeclamptic women into mice on day 13 of gestation [101]. These mice displayed significantly higher blood pressure, proteinuria, glomerular endotheliosis, elevated sFlt-1 and sEng levels, smaller placentae, and fetal growth restriction compared to the control group [101]. Normal pregnancy and PE are significantly affected by the renin-angiotensin system (RAS). Angiotensin II type 1 receptor autoantibodies (AT1-AA) are reported to be higher in some women with PE and are linked to other disorders such as systemic sclerosis, tissue fibrosis, hypertension, and reno-vascular disease [102]. These autoantibodies influence vasoconstriction and increase blood pressure through the stimulation of angiotensin II type 1 (AT1) receptors with transduction of signals via the MAPK/ERK pathway [102,103]. LaMarca and co-workers have reported that pregnant rats infused with purified AT1-AA between gestational days 12-19 demonstrated high blood pressure and high serum AT1-AA levels as well as dysregulated angiogenic factor levels and increased tissue levels of prepro-endothelin-1 in comparison to normotensive pregnant control rats [50]. The endothelin system may play a role in the elevation of blood pressure induced by AT1-AA administration; this theory is corroborated by the report of endothelin type A receptors inhibiting the blood pressure response in AT1-AA-infused rats [104]. These studies display a significant interaction between inflammatory and angiogenic markers that are released in response to placental ischemia. A disadvantage of the AT1-AA mouse model is that some aspects observed are not specific to pregnancy. The study by Zhou and coworkers also assessed the involvement of AT1-AAs independently of excess sFlt1; they reported elevations in blood pressure in non-pregnant mice following treatment with IgG isolated from preeclamptic women [101]. In these animals, renal injury and elevations in urinary protein and sFlt-1 levels were absent, suggesting that AT1-AA in non-pregnant women would chronically induce high blood pressure in an autoimmune manner. Furthermore, in the case of pregnancy, PE symptoms may not be resolved post-delivery [83 •].

Toll-Like Receptor Rat Model
The Toll-like receptor (TLR) rat model was induced by the activation of TLR3 in pregnant rats with a viral mimetic called polyinosinic:polycytidylic acid [105]. This results in elevated maternal blood pressure, systemic inflammation, proteinuria, and endothelial dysfunction [105]. This study was the first to demonstrate that activation of TLR signaling during pregnancy adversely effects maternal cardiovascular function. These receptors are a class of pattern recognition receptors which induce signaling cascades to elicit appropriate inflammatory responses to pathogen-and damage-associated molecular patterns [106]. These receptors are present at the maternal-fetal interface and function to ensure a successful pregnancy outcome; conversely, excessive TLR signaling may induce maternal systemic inflammation with adverse pregnancy outcomes [107]. This rat model has since been replicated in pregnant mice [108,109], and mice treated with synthetic ligands for TLR7/8 have produced similar outcomes as earlier reports [109]. Low-dose of unmethylated CpG DNA administered to female rats have been shown to stimulate TLR9 signaling in late gestation and increases blood pressure, vasoconstriction, vascular oxidative stress, and inflammation [110]. Earlier studies in mice showed that activation of TLR9 with fetal DNA or high doses of CpG DNA resulted in poor fetal outcome, including increased fetal resorption and malformations [111][112][113]. The findings of these studies imply that the TLR-induced initiation of the innate immune system plays a role in the development of hypertension in pregnancy.

STOX1 Mouse Model
Storkhead box-1 (STOX1) is a transcription factor found in column extravillous trophoblast cell populations and facilitates cytotrophoblast invasion during normal placentation by regulating α-T-catenin expression [114,115]. Alterations in STOX1 expression have been implicated in PE development [116,117]. Overexpression of STOX1 in pregnant mice elevates systolic blood pressure in early pregnancy, proteinuria, occlusion of renal capillaries, fibrin deposition, and increases sFlt-1 and sEng expression [24,33]. The increase in systolic blood pressure precedes placental development and indicates that abnormal placental development may not be responsible for the hypertension observed in this model. This finding also suggests that the placenta may not be the initiating organ of PE development [83•]. This model also reports increased renal artery resistance, cardiac hypertrophy, fetal growth restriction, and greater umbilical resistance [118,119] and demonstrates left ventricular hypertrophy, cardiac fibrosis, and markers of inflammation and cellular stress up to 8 months postpartum [120]. A benefit of this model is its ability to replicate the early pathogenic aspects of PE as well as the later systemic features. However, a limitation of the STOX1 mouse model is the observation of increase in blood pressure in early gestation and non-comparable to what is observed in preeclampsia. Therefore, the hypertension observed cannot be attributed to abnormal placentation.

Dahl Salt-Sensitive Rat Model
The Dahl salt-sensitive rat has pre-existing hypertension and during the course of gestation develops further PE-like symptoms such as increase in blood pressure, proteinuria, glomerulomegaly, placental hypoxia, fetal growth restriction, and higher circulating levels of TNF and sFlt-1 [28]. This model represents a spontaneous or superimposed rat model of PE. Gillis and co-workers used this model to validate the usefulness of sildenafil citrate treatment in PE [121]. Their data confirm that administration of sildenafil citrate between gestational day 10 and 20 alleviated further increase in blood pressure and proteinuria, decreased uterine artery resistance, and aided fetal growth [121]. Nonetheless, a disadvantage of this model is the pre-existing hypertension exhibited by pregnant rats [30].

Arginine Vasopressin Mouse Model
Hypertension induced by arginine vasopressin (AVP) is characterized by low circulating renin-angiotensin system activity, which is also found in PE compared to normotensive pregnant women [23 ••]. AVP exerts its physiological functions via V1a and V2 receptors [122]; the activation of these receptors has been implicated in proteinuria, renal glomerular endotheliosis, and intrauterine growth restriction, respectively. Additionally, the V1b receptor is a regulator of adrenocorticotropic hormone secretion, which can exert its effects on the immune system and blood pressure [123]; cullin-5 plays a role in angiogenesis [124], while the oxytocin receptor is involved in pregnancy and labor [125].
Studies by Santillan and co-workers highlight the role of AVP in PE development; however, chronic infusion of AVP does not reproduce placental hypoxia in mice which is a characteristic of human PE [122]. This model demonstrates the potential of AVP as both a predictive biomarker for PE development as well as an initiator of this disorder. Plasma copeptin levels, a biomarker of AVP, was reported to be significantly higher at 6 weeks of gestation in PE cases compared to normotensive pregnancy [23••]. The chronic administration of AVP in pregnant mice replicated pregnancy-specific hypertension, glomerular endotheliosis, proteinuria, and intrauterine growth restriction, thus supporting the role of AVP in PE progression and copeptin as an early biomarker for PE prediction [23••].
Santillan and co-workers have further expanded this work, in that AVP infusion into pregnant C57BL/6 J mice reduced the levels of placental expression of placental growth factor, altered placental morphology, placental oxidative stress, and placental gene expression consistent with the characteristic features of human PE [122]. They have also demonstrated that AVP infusion throughout gestation in mice promoted pro-inflammatory T H 1-associated interferon gamma in maternal plasma [126]. The effectiveness of AVP in inducing PE-like symptoms in a mouse model was therefore successfully demonstrated. This model proposes the concept that regulators of blood pressure are activated in the early stages of pregnancy,and could therefore be a potential new model for studying the origins of PE [127].
In addition, our laboratory has extrapolated the AVP mouse model of Santillan and co-workers to a Sprague Dawley rat model [128••]. Our findings demonstrate that chronic AVP infusion (150 ng/h) in pregnant rats over 18 days successfully reproduced the PE phenotype of elevated blood pressure (≥ 140/90 mmHg) [10], increased urinary protein levels, and fetal growth restriction. Albeit, our study models a mild case of PE development, and future studies should explore higher AVP dosages to induce more severe features of PE development. Additionally, our study did not confirm the levels of angiogenic markers such as PlGF and sFlt-1 which are commonly associated with PE.
Biochemical analysis revealed significantly upregulated serum alanine transaminase and triglyceride levels along with downregulated high-density lipoprotein levels in pregnant AVP-treated rats. We further demonstrate alterations in kidney morphology including a mild increase in mesangium, mild glomerular crescents, and reduced Bowman's space in AVP-treated rats. An earlier study performed by our lab assessed liver injury in the AVP rat model and found that serum expression of the liver injury enzymes arginase and 5′-nucleotidase, as well as transforming growth factor-2, was significantly higher in pregnant rats treated with AVP [129 ••]. Our findings are indicative of acute pregnancy-initiated liver dysfunction and support the utility of this model in the study of PE development.
The novel AVP mouse model highlights the potential use of AVP as a predictive biomarker for PE development. This model recapitulates phenotypes consistent with human PE, most notably pregnancy-specific hypertension. The activation of V1a and V2 receptors has been implicated in proteinuria and renal glomerular endotheliosis and intrauterine growth restriction, respectively, in this model. Future studies using the AVP model should explore the roles of the V1b, the oxytocin receptor, and cullin-5 (VACM-1) in the pathogenesis of this disorder.

Conclusion
Despite the advances made in understanding PE development, it continues to be a leading cause of maternal and fetal mortality and morbidity worldwide. This review provides a brief overview of various rodent and murine models that mimic PE development while also highlighting the associated limitations. We report that there is no ideal rodent model to date that fully epitomizes the phenotype of PE such as abnormal placentation, fetal growth restriction, pregnancy-specific hypertension, proteinuria, endothelial dysfunction, and an imbalance in angiogenic factors. Despite the invaluable contribution of the different models, they do not unravel the early events in PE development that precede abnormal placentation. We advocate that the combined use of different models is still required to enable novel developments regarding PE pathogenesis and treatment. However, as the advances made in this field of research continue to grow, the refinement of these models will undoubtedly occur, leading to the discovery of new aspects of this disorder.
Author Contribution All authors contributed to the conception and design of the manuscript. The first draft of the manuscript was written by Sapna Ramdin, and the work was critically revised and edited by Sooraj Baijnath, Thajasvarie Naicker, and Nalini Govender. All authors read and approved the final manuscript.
Funding Open access funding provided by Durban University of Technology. This work was supported by the National Research Foundation (grant numbers 107236 and 122014).

Compliance with Ethical Standards
Ethics Approval Ethical approval was not required as our study did not involve humans or animals.

Conflict of Interest
The authors have no competing interests to declare that are relevant to the content of this article.

Human and Animal Rights and Informed Consent
This article does not contain any studies with human or animal subjects performed by any of the authors.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.