Interventions for placental insufficiency and fetal growth restriction

Abstract

Pregnancy complications adversely impact both mother and/or fetus throughout the lifespan. Fetal growth restriction (FGR) occurs when a fetus fails to reach their intrauterine potential for growth, it is the second highest leading cause of infant mortality, and leads to increased risk of developing non-communicable diseases in later life due ‘fetal programming’. Abnormal placental development, growth and/or function underlies approximately 75% of FGR cases and there is currently no treatment save delivery, often prematurely. We previously demonstrated in a murine model of FGR that nanoparticle mediated, intra-placental human IGF-1 gene therapy maintains normal fetal growth. Multiple models of FGR currently exist reflecting the etiologies of human FGR and have been used by us and others to investigate the development of in utero therapeutics as discussed here. In addition to the in vivo models discussed herein, utilizing human models including in vitro (Choriocarcinoma cell lines and primary trophoblasts) and ex vivo (term villous fragments and placenta cotyledon perfusion) we have demonstrated robust nanoparticle uptake, transgene expression, nutrient transporter regulation without transfer to the fetus. For translational gene therapy application in the human placenta, there are multiple avenues that require investigation including syncytial uptake from the maternal circulation, transgene expression, functionality and longevity of treatment, impact of treatment on the mother and developing fetus. The potential impact of treating the placenta during gestation is high, wide-ranging across pregnancy complications, and may offer reduced risk of developing associated cardio-metabolic diseases in later life impacting at both an individual and societal level.

1. Introduction

Pregnancy complications can adversely affect both the mother and/or the fetus. In the United States, a 16.4% increase in pregnancy complications and a 31.5% increase in women who experienced pregnancy and childbirth complications between 2014 and 2018 has been reported [1]. Some of the most common pregnancy complications include fetal growth restriction (FGR), preeclampsia, gestational diabetes, preterm labor, and stillbirth/miscarriage, and are all associated with defective placentation and/or abnormal placental function. The placenta is the interface in which nutrient and waste exchange between maternal and fetal circulations occurs [2]. In many cases, defective placentation is caused by abnormal placental invasion of the uterine wall which leads to improper placenta perfusion and subsequent placenta hypoxia, and/or issues within the fetal villi [3,4].

2. Fetal growth restriction

FGR is defined as estimated fetal weight <10th percentile and/or abdominal circumference <10th percentile [5]. In the developed world this occurs in 5–10% of all live births, and this rate is up to 6x greater in developing countries with roughly 30 million cases annually [6,7]. An important distinction of FGR from constitutively small fetuses or those with low birth weight, is a pathological condition causing an impaired supply of nutrients and/or oxygen that prevents adequate growth. This can be due to pregnancy complications like those previously mentioned, or physical and/or functional anomalies of the placenta preventing nutrient transfer, or genetic abnormalities impeding transfer of nutrients/oxygen from the placenta [8]. Furthermore, FGR has been associated with the developmental origins of health and disease (DOHaD) hypothesis in which insults, like inadequate nutrient supply, during critical developmental periods result in “fetal programming” that have life-long effects on health [9]. Specifically, in childhood, those affected by FGR are at higher risk for continued poor growth and reduction in neurological function. Into adulthood, these individuals are predisposed to many noncommunicable diseases (NCD) including hypertension, type 2 diabetes, and coronary artery disease [8,10].

Diagnosis and differentiation of FGR from constitutively small fetuses has long been difficult without adequate tests and parameters. Current procedures for diagnosis and outcome predictors utilize Doppler ultrasonography to measure the Doppler cerebroplacental ratio (CPR), or the ratio of the middle artery pulsatility index (PI) to the umbilical artery doppler PI. This is a non-invasive procedure that assesses vascular resistance. This in combination with an estimated fetal weight has proven to more reliably diagnose FGR and predict poor outcomes in pregnancies [11,12]. However, there is currently no treatment for FGR during pregnancy. The potential for maternal supplementation with nutrients such as antioxidants, omega-three fatty acids and calcium/vitamin D have been trialed but with mixed success (see review [13]). The only true intervention is delivery, often prematurely which carries its own risks, and postnatal emphasis on nutrition intake [14].

FGR infants are significantly more likely to spend extended time in the NICU and demonstrate feeding difficulties leading to increased mental and financial stress for the parents [15]. Increased NICU stay alone has been associated with impaired cognitive and socioemotional development of children, and has even been equated to complex trauma from repeated exposure to stressors [16,17]. Hence, there is much need for therapeutic approaches to treat FGR in utero in order to prevent preterm birth and the complications associated with this form of intervention, as well as potentially alleviate or prevent the predisposition of many NCD’s and the lifelong burden on individuals and healthcare systems. Given impaired placental function is often an underlying cause of FGR, specifically targeting the deficiencies within the placenta, such as nutrient transport or placental vascular development, may allow for increased length of time in utero as well as continued, appropriate fetal growth and development. In order for such treatments to be developed, appropriate animal and human models must be utilized.

3. Animal models of fetal growth restriction

3.1. Mice and rats

The use of rodents to study FGR has many advantages for mechanistic understanding during restriction, especially at a cellular or molecular level [18]. Their small size and short gestation periods (mice: 19–21 days, and rats: 21–23 days) mean they are relatively inexpensive compared to many other animal models, and studies can be conducted much faster. However, placental development and structure does not reflect humans as wells as other laboratory species. Mouse and rat placentas are haemotrichorial with one layer of cytotrophoblasts and two layers of syncytiotrophoblast separating maternal and fetal circulation, and exhibit relatively shallow invasion into the uterine wall compared to humans [19,20]. Furthermore, unlike the human placenta in which the exchange region develops into cotyledons, the exchange region of mice and rats develops as a labyrinth [19]. In terms of fetal development, offspring are born altricial and many organs like the kidneys continue to develop after birth [18]. These developmental differences between rodents and humans limit their utilization for DoHAD related studies.

Despite differences in placental and fetal development, many studies using transgenic knockout mice have elucidated factors within the placenta that are necessary for normal fetal development and when gene expression is removed/inhibited, produce FGR (see review [20]). Some examples include the eNos^−/− [21,22], Vegf^+/− [23], Igf-1^+/− [24], and Egfl7^−/− [25] mouse models. Each of these knockout models demonstrates an FGR phenotype via different mechanisms. Endothelial nitric oxide synthase (eNos) knockout mice (global knockout) display hypertension, uterine artery constriction, and FGR with hypoxic placentas [21,22]. Vascular endothelial growth factor (Vegf) knockout embryos show abnormal blood vessel formation and defective extraembryonic vasculature illuminating the necessity of Vegf for the supply of blood from mother to placenta and the proper connection of yolk sac/placenta to fetus for nutrient and oxygen transport [23]. Homozygous insulin-like growth factor 1 (Igf-1) knockout fetuses, obtained by crossing heterozygous knockout parents, showed significantly decreased placental efficiency, fetal/placental weight ratios and FGR compared to wildtype litter mates [26]. Epidermal growth factor-like protein 7 (Egfl7) knockout fetuses show reduced placental vascularization and a deficiency in endothelial cell migration and cord formation when compared to wildtype litter mates [25]. In addition to models based on genetic modification, there are other models of FGR in rodents caused by external inducible factors which pregnant women are exposed to. For example, recent rat studies have examined the consequences of commonly inhaled substances that could lead to FGR such as tetrahydrocannabinol and ozone [27]. Overall, each of these different rodent models mentioned can be utilized to study the potential etiologies of FGR in humans and aids in identification of possible therapeutic targets.

3.2. Rabbits

While rabbits experience a short gestation period (31–33 days) like those of rodents mentioned previously, there are facets that make them more reflective of the human. The rabbit genome is more similar to humans [28], and they experience similar hemodynamic changes during pregnancy [29,30]. Some aspects of fetal development such as brain and lung development are similar to human fetal development [30]. However, the rabbit’s short gestation period leads to the occurrence of neonatal development/maturation of some organ systems instead of during gestation as they would in humans with much longer gestation time. Similar to rodents and humans, the rabbit placenta is hemochorial meaning maternal blood is in direct contact with trophoblast, however rabbits placentas are haemodichorial with two trophoblastic layers separating maternal and fetal circulation, and more reflective of human placenta structure in early pregnancy [30]. Nutrient restriction, caloric restriction, recreational drug administration/inhalant studies, and surgical studies have been very well characterized in rabbits to model and study FGR. Along with these induced models of FGR, rabbits have a natural propensity to develop a growth restricted runt in their litters [31]. Pups in the middle of the uterine horn are farther from the uterine arteries connecting blood supply and usually have FGR which makes for a natural model of FGR.

3.3. Guinea pigs

Unlike rodents and rabbits, the guinea pig offers a unique advantage in developing in utero treatments because they have similar developmental milestones to humans throughout gestation and following birth [32,33]. The guinea pig placenta, like humans, is haemomonochorial with one trophoblastic layer separating maternal and fetal circulation [34] and placental development involves deeper trophoblast invasion [35]. Furthermore, their precocial offspring are more developed at time of birth reflecting humans. Guinea pigs experience a longer gestation time (65–70 days) than other rodents and rabbits. While this does lead to more expensive animal costs both due to length of time and increase in size compared to that of rodents, it is beneficial for evaluation of interventions and therapeutics. The most common models of FGR in the guinea pig are uterine artery ligations and maternal nutrient/caloric restriction [18,36,37]. There are now many studies into the effects of FGR in the fetus during pregnancy, through adolescence, and during adulthood (reviewed in Ref. [32]). Recent studies have looked at the consequences of this condition on different fetal organs and long-term effects on the affected fetus. These guinea pig FGR studies have identified increased cardiovascular and renal disease [38], decreased maturation of myelin sheaths [39], and altered liver transcript levels [40]. In the brain alone, FGR has been linked to increased apoptosis of the white matter and hippocampus, reduced brain size [41], and decreased granule cell proliferation and purkinje cell development creating impaired cerebellar development [42].

3.4. Sheep

Sheep models are consistently used to study FGR because fetal developmental milestones are similar to those of human fetuses, and their longer gestation (144–151 days) is advantageous to performing serial measurements, for example imaging and/or blood sampling, in mother and fetus during pregnancy in live animals. Their propensity to have single or twin pregnancies/offspring also mimics that of humans, and they offer several different ways to model FGR [43]. FGR models include surgical removal of endometrial caruncles [44], umbilical artery ligation [45], maternal hyperthermia [46,47], and nutrient/diet change [48–50]. Each of these models have led to new findings in the adaptations and outcomes of growth restriction in the placenta and fetus. These studies have identified several factors that are decreased in times of restriction such as Igf-1 and 2, Vegf, prolactin, and suppressor of cytokine signaling 3 as well as overall decreases in myoblast function, reduced muscle mass, and protein accretion. Increases in cortisol levels and fat deposition have also been reported. Each of these deficient factors plays a key role in the growth and development of the fetus, so naturally this leads to many complications in utero and after birth. High cortisol levels are also unsurprising with high stress caused by the restriction [51]. While there are many advantages to the sheep model, placentation is significantly different from the human [18,43]. Sheep are classified to have synepitheliochorial placentation where there are multiple placentation sites, known as placentomes, throughout the uterus [52]. Each placentome is then comprised of trophoblast-rich cotyledon tissue in contact with a caruncle. Thus, their usefulness to the development of treatments for FGR that target the placenta is limited.

3.5. Non-human primates

Research utilizing non-human primates (NHP) is highly regulated, and expensive which leads to low replication numbers, but they are the closest model to humans in almost every aspect making them an excellent pre-clinical model. The most commonly used non-human primates in FGR studies are baboons and rhesus macaques [53]. While these studies are long due to gestation time (rhesus macaques: 166 days, baboons: 185 days) and life span, the translational relevance to humans’ aids understanding of short-term and long-term effects of FGR on mother, placenta, and affected offspring. NHP have a haemomonochorial placenta with cotyledons containing villous structures just like humans [54], setting them apart from all other species previously mentioned. Serial measurements like imaging, chorionic villous sampling and/or blood sampling can also easily be performed. Models of reduced amino acid transport, restricted placenta vasculature, and maternal nutrient restriction have all lead to FGR in NHP models [55–58]. Studies in these models have elucidated long term effects of FGR on hypothalamic arcuate nuclei which control appetite [59], impaired right ventricle [58], and cardiac remodeling [60]. These different models contribute to the understanding of FGR development in the human, all have a similar phenotype – FGR or fetal demise in extreme cases, and offer pre-clinical models of impaired placenta. Furthermore, preliminary studies utilizing NHP models are underway to determine the clinical relevance of placental gene therapies and potential adverse effects of such treatments [61,62].

4. Human models for therapeutic development

In addition to animal models, researchers often implement human ex vivo and in vitro models that can be used to mimic metabolism and transport within the human placenta. Furthermore, these techniques can aid in the development of treatments for FGR. Given the placenta is discarded at birth, it is relatively easy to obtain primary tissue to use for scientific purposes. The ex vivo human placenta perfusion model offers the ability to study placenta storage, vascularization, metabolism, and maternal-fetal transfer [63]. Furthermore, it is the only model that retains the full structure of the placenta [64]. However, experiments utilizing the ex vivo perfusion approach are limited by the model’s short-lasting performance and laborious set-up procedures [65]. In terms of in vitro approaches, several immortalized cell lines exist, including BeWo choriocarcinoma cells, Jeg3 and HTR8s, to study trophoblast biology. Furthermore, primary cells and/or tissue explants can be isolated and cultured for research purposes. More recently, the development of a ‘placenta-on-a-chip’ has been allowing for the in vitro study of human placenta physiology and functionality [66]. Such devices present the fetal endothelium and trophoblastic endothelium in a manner that replicates the placental barrier thus providing a cost-effective way of testing transfer of factors like nutrients and pharmaceuticals [63]. Finally, organoid techniques are offering methods of studying the human placenta in vitro in three-dimensions [67], although such technologies are still being optimized.

5. Therapeutic target identification

Human and animal studies have all led to identifying key factors that are altered in the placenta of FGR pregnancies. This knowledge in turn, is now being used to develop potential in utero treatments, and several are being tested to determine their relevance in alleviating the growth restricted phenotype.

EGF is an endogenous placental growth factor that increases proliferation and differentiation of cytotrophoblasts, and secretion of human chorionic gonadotropin [68]. In women with pregnancies complicated by FGR, urine EGF is lower [69]. Targeted delivery of EGF to normal human placenta explants has been shown to significantly enhance nutrient transport, and increased down-stream phosphorylation signaling [70], as well as protect against apoptosis induced by oxidative stress [71]. However, the ability for EGF to elicit similar responses in placental explants from FGR pregnancies is conflicting, possibly because of the multifactorial nature of FGR, and therefore hypothesized to be a good candidate for those FGR cases caused by abnormal trophoblast development, but not vascular defects or uteroplacental perfusion [70].

The IGF system regulates placental growth, regulation, invasion and angiogenesis [72]. Trophoblasts produce both IGF-1 and IGF-2 throughout pregnancy [73]. Transgenic mice with mutations in either IGF1 or IGF2 show a 60% reduction in size compared to wild-type litter mates [74], and it is well established that IGF-1, its receptor, and its binding partners are downregulated in FGR [75]. There is a growing body of evidence that targeted delivery of IGF-1 or IGF-2 successfully improves fetal growth in FGR models. We have demonstrated that placental IGF-1 gene therapy, through multiple mechanisms, maintains fetal weight in FGR fetuses [76], increases placental glucose and amino acid transporter expression [77–79], increases fetal glucose concentrations and reduces interhaemal distance in the placenta [40]. Similarly, targeted delivery of the IGF2 protein to the placenta in the P0 mouse model of FGR, led to an increase in placenta weight and a reduced number of FGR fetuses [80].

During pregnancy, VEGF is produced by the placenta trophoblast cells [81], and regulates vascular endothelial cell growth and proliferation in the placenta [82]. Additionally, maternal serum levels of VEGF have been shown to be lower in women with FGR pregnancies [83,84]. Targeted delivery of VEGF locally to uteroplacental circulation has been shown to increase fetal weight at term in models of FGR [85,86] as well as improve uterine blood flow [87–89]. Along with this, fetal lungs, livers, and brains were found to be significantly larger with VEGF delivery than without [85]. Uterine artery activity was also found to be positively altered with VEGF delivery with increased relaxation and decreased tension compared to untreated FGR models [85].

Nitric oxide is required for trophoblast function, nutrient transport, uterine artery relaxation, and proper growth of the fetus [21,90]. More recently, the use of nitric oxide donors as potential targets to improve fetal growth has been investigated as nitric oxide regulates placental expression of growth factors including EGFl7 [91]. EGFl7 is a secreted protein which regulates vasculogenesis and angiogenesis [92], as well as placental vascularization [25]. Nitric oxide induced increased EGFl7 expression in the placenta increases NOTCH1 expression which controls development of extravillous trophoblasts in the placenta, and notch target genes such as HEY1 and HEY2 [91]. Similarly, targeted delivery of the nitric oxide donor SE175 to the placenta has been shown to increase fetal weight and reduced placental oxidative stress in the eNOS^−/− mouse model of FGR [93]. In ovine fetoplacental artery endothelial cells another nitric oxide donor, sodium nitroprusside, s upregulates activation of the ERK pathway [94]. Overall, these studies highlight the potential to use nitric oxide donors or downstream targets for treatment of those FGR cases caused by abnormalities in placental and arterial remodeling.

6. Therapeutic development in humans

Several factors have now been identified as potential targets for the treatment of FGR, but delivering therapeutics in a safe and effective way with the desired outcomes is the next step to developing in utero therapy options. Research into several mechanisms to target the placenta and deliver these targets are currently in preliminary testing. These include targeted liposomes, adenoviral vectors and adeno-associated virus delivery, and polymer nanoparticle delivery.

Liposomes are fluid vesicular structures that can envelope many different types of molecules with low immune response at an inexpensive cost. Hence, there are several types of liposomes currently on the market and in clinical trials for a number of different uses: from anticancer and anti-fungal treatments to anesthetics [95]. Additionally, modifications such as poly-ethylene glycol can be made to the liposomes to help increase circulation time and decrease toxicity and immune response [96]. These liposomes can be coupled with targeting components such as a ligand with a receptor on a specific cell type [97]. In terms of placental targeting, Harris et al. have used this liposome technology to target the placenta through peptide sequences specific to the placenta and uterine vasculature [80]. These liposomes have been used to deliver IGF-2 [80], EGF [70], and SE175 [93] in various animal and human in vitro models to test their ability to correct FGR. In the animal studies, liposomes were administered intravenously in the female. Most importantly, these studies consistently showed effective delivery of the liposome and cargo it contained to the placenta with no adverse effects on maternal or fetal health, and highlights the therapeutic potential of liposomes with further studies.

Adenoviral vectors are increasingly popular delivery mechanisms for gene therapy. These viral vectors are single stranded DNA genomes surrounded by protein shells. In recombinant vectors the viral DNA is replaced with the transgene of interest under the control of tissue specific promoters [98,99]. In terms of targeting the maternal-fetal interface, David et al. have previously shown an increase in fetal weight with adenoviral vector delivery of VEGF in a guinea pig growth restriction model [85]. The adenoviral vector was applied in a thermosensitive gel directly to the exposed uterine and radial arteries. Additionally, using the same adenoviral vector expressing VEGF but injected directly into the proximal part of the uterine artery, they demonstrated increased uterine blood flow and vasodilation in a sheep model counteracting the abnormalities in FGR [100]. Direct placental injection of adenoviral vectors containing eight different growth factor transgenes have also been performed to examine the effects of overexpressing genes including IGF-1, placental growth factor (PGF) and angiopoietin 2 (ANG-2) on placental and fetal growth [101]. Following on from this study, we performed intraplacental delivery, through direct injection of the placenta, of adenoviral vectors containing IGF-1 have been performed in various animal [76,102] and in vitro human models [77,78] with positive effects on fetal growth and placental nutrient transport. Whilst these early studies clearly demonstrate efficient delivery of factors to the placenta and positive effects on placental function and fetal growth, there are major concerns regarding immunogenicity and off-target effects [103] that need to be considered.

Non-viral, polymer-based nanoparticles are widely studied in cancer therapeutics [104]. They offer a controlled delivery system with easy manipulation of release time and interactions. These can be natural or synthetic polymers, conjugated with targeting molecules to enhance specificity, and used for drug or gene delivery [105]. Synthetic polymer-based nanoparticles can serve as carriers for plasmid DNA, siRNAs, mRNAs, proteins and oligonucleotides [106–109] - improving transfection, therapeutic effects, and evoking less immune response than viral delivery [110,111]. Our laboratory has been using a PHPMA₁₁₅-b-PDMEAMA₁₁₅ co-polymer to deliver a plasmid containing the IGF-1 gene under a trophoblast specific promoter (CYP19A1 or PLAC1) in human placental cells and ex vivo placenta cotyledons [79], and intraplacentally in mice [76,90], and guinea pig studies [40]. These studies have shown effective delivery to the placenta, regulation of placental signaling mechanisms, maintenance of birth weight of offspring affected by FGR, and increased glucose transport in vitro and in vivo. Most importantly, we have achieved effective delivery to the placenta without plasmid passing to the fetus, while having significant effects on placental structure, IGF-1 signaling, fetal organ development and fetal growth. Future pre-clinical model studies have the potential to verify therapeutic potential and move towards in utero human placenta gene therapies.

With multiple target proteins and multiple modes of delivery being studied, gene therapy targeting the placenta for FGR treatment may soon be a reality. Additionally, the vast number of animal and in vitro human models contributes significantly towards the ability to effectively study potential treatments for FGR. Hence, with continual research support there is the potential for reducing prenatal mortality rates, NICU admission, and many long-term health effects in children and adults.