Placental development and function in trisomy 21 and mouse models of Down syndrome: Clues for studying mechanisms underlying atypical development

Abstract

Down syndrome (DS) is the most common genetic disorder leading to developmental disability. The phenotypes associated with DS are complex and vary between affected individuals. Placental abnormalities in DS include differences in cytotrophoblast fusion that affect subsequent conversion to syncytiotrophoblast, atypical oxidative stress/antioxidant balance, and increased expression of genes that are also upregulated in the brains of individuals with Alzheimer’s disease. Placentas in DS are prematurely senescent, showing atypical evidence of mineralization. Fetuses with DS are especially susceptible to adverse obstetric outcomes, including early in utero demise, stillbirth and growth restriction, all of which are related to placental function. The placenta, therefore, may provide key insights towards understanding the phenotypic variability observed in individuals with DS and aid in identifying biomarkers that can be used to evaluate phenotypic severity and prenatal treatments in real time. To address these issues, many different mouse models of DS have been generated to identify the mechanisms underlying developmental changes in many organ systems. Little is known, however, regarding placental development in the currently available mouse models of DS. Based upon the relative paucity of data on placental development in preclinical mouse models of DS, we recommend that future evaluation of new and existing models routinely include histologic and functional assessments of the placenta. In this paper we summarize studies performed in the placentas of both humans and mouse models with DS, highlighting gaps in knowledge and suggesting directions for future research.

1. Introduction

Down syndrome (DS) or trisomy 21 (T21) affects 1 in 700–1200 live births and is the most common genetic disorder leading to developmental disability [1,2]. This syndrome results from abnormalities in both gene dosage imbalance as well as globally dysregulated gene expression [3,4]. The phenotypes associated with DS are complex and vary significantly between affected individuals [3]. For example, intellectual disability is present in all individuals with DS, however IQ varies significantly [5]. Other phenotypic features, such as cardiac defects, are only seen in 40–50% of individuals with DS [6].

The placenta has many functions to ensure an optimal environment for offspring survival. Abnormalities of placental function are widely recognized as impacting pregnancy outcome and the long-term health of the child [7,8]. Multiple processes are dysregulated in the placentas of fetuses with DS [9–14]. Fetuses with DS are especially susceptible to adverse obstetric outcomes, including in utero demise, stillbirth and growth restriction, all of which can be a consequence of poor placental function [15–17]. The presence of aneuploid cells confined to the placenta is also associated with poor pregnancy outcomes [18–20].

During human pregnancies, the placenta is difficult to evaluate in real time due to the potential risks to the pregnant woman and her fetus. Also, given that there is generally only one fetus in each pregnancy and there is a relatively long gestational period, it is difficult to generate longitudinal data and control for natural genetic variation. Given what is known about the impact of placental function on human development, this organ may provide key insights towards understanding the phenotypic variability in individuals with DS. To address these issues, many different mouse models of DS have been generated to identify the mechanisms underlying developmental changes in many organ systems [4,21–23]. Little is known, however, regarding placental development in currently available mouse models of DS. This review will summarize what is currently known regarding the human placenta in DS, as well as in mouse models of DS, identifying knowledge gaps and future directions for research.

2. Comparisons between the typically developing human and murine placentas

Typical embryonic development is dependent upon having enough oxygen, nutrient and waste exchange through the placenta. This process, however, can vary among species [24]. The fetal portion of the placenta is derived from two basic elements: an outer epithelium derived from the trophoblast cell lineage and an underlying vascular network and stroma that are derived from embryonic mesoderm [25]. The other key portion of the placenta includes the maternal decidua and vasculature. Concerted interaction between the fetal and maternal elements are required for successful implantation [25]. Considerable differentiation in the trophoblast lineage produces different cell subtypes with specialized endocrine, vascular, immunological and transport functions [24]. Although human and rodent placentas have different organizational structure, both species have discoid hemochorial placentas and undergo the same fundamental steps during development (Supplemental Fig. 1). These include trophoblastic invasion, vascularization of the trophoblast to establish and maintain feto-placental vasculature and maternal vascular remodeling to establish utero-placental circulation [24]. Detailed anatomical comparisons of the mouse and human placentas have been published elsewhere [25–35] and are summarized in Table 1.

Table 1.

Comparison of human and murine placentas.

Feature	Euploid Human	T21 Human	Wild-type Murine	DS Mouse Models
Gross Shape	Discoid	Unchanged	Discoid	Unchanged
Transport Barrier	Hemochorial	Unchanged	Hemochorial	Unchanged
Trophoblast layers	Monochorial (1 STB layer)	Abnormal (2 layers due to defective cell fusion)	Trichorial (2 STB layers and 1 TGC layer)	Unknown
Site of Fetal-Maternal Exchange	Chorionic villi (via STB)	Abnormal (abnormal cell fusion and signaling)	Labyrinth (via STB)	Abnormal in T16 model
Syncytiotropho-blast Fusion	Syncytin 1 and 2	Abnormal (delayed or absent)	Syncytin A and B	Unknown
Hormone synthesis	STB	Abnormal	TGC, SpTb	Unknown
Decidual/Endo-vascular Invasion	Extravillous trophoblast (interstitial and endovascular)	Abnormal (increased apoptosis, abnormal cell signaling and decreased invasion)	Primary TGC (implantation), Glycogen cells (only interstitial invasion), Secondary TGC (endovascular invasion)	Decreased glycogen and giant cells in T16 model
Endometrial Decidualization	Day 14 of menstrual cycle	Unchanged	Begins at fertilization	Unknown
Maternal uterine natural killer cells	Secretion of VEGF and NO to promote invasion	Unknown	Secretion of VEGF and NO to promote invasion	Unknown
Selective MHC class I expression by trophoblast	HLA-C, -G, E	Abnormal (Decreased expression of HLA-G)	H2–K	Increased expression in T16 model
Oxidative stress	Increases throughout gestation	Abnormally elevated	Increases throughout gestation	Unknown
Gestational length	40 weeks	Unchanged	3 weeks	Unchanged
Embryos	1	Unchanged	4–12	Decreased
Vascular invasion	Myometrial involvement (First 1/3)	Abnormal	Limited to the decidua	Unknown
Regulators of angiogenesis	VEGF, PlGF, sFlt-1	Abnormal (Decreased levels of PlGF)	VEGF, Proliferin, sFlt-1	Unknown
Placental Hormones	hGH/hCS 5-member gene cluster, Prolactin, Adipokines, hCG, Progesterone	Abnormal (Abnormal hCG production and secretion)	Prl 22-member gene cluster, growth hormone, placental lactogens, Progesterone, Adipokines	Unknown
Spiral arteries	100–150	Unknown	5–30	Unknown
Transformation of Spiral Arteries	Trophoblast dependent (with uterine natural killer cell assistance)	Abnormal	Uterine natural killer cell dependent	Unknown
Choriovitelline Placenta	Transient role in early pregnancy	Unchanged	Functions throughout pregnancy	Unchanged
Stem cell differentiation	Mediated by low oxygen tension	Abnormal (Increased oxidative stress)	Mediated by low oxygen tension	Unknown
Progesterone Production	Corpus luteum until, then STB	Defective	Corpus luteum throughout gestation	Unknown
Development of the choriovitelline placenta (yolk sac)	Day 8	Unknown	Day 8	Unknown
Early nutrition	Histiotrophic from endometrial glandular secretions	Unchanged	Histiotrophic from endometrial glandular secretions	Unchanged
Establishment of chorioallantoic circulation	12 weeks	Unknown	Day 10.5	Unknown
Temporal relationship	First trimester	Unchanged	Mid trimester	Unchanged

TGC = Trophoblast giant cell, STB = syncytiotrophoblast, SpTb = spongiotrophoblast, VEGF = vascular endothelial growth factor, NO = nitric oxide, hGH = human growth factor, hCS = human chorionic sommatotropin, hCG = human chorionic gonadotropin Prl = prolactin, sFlt-1 = soluble fms-like tyrosine kinase.

While the mouse placenta is not identical to its human counterpart many studies have shown that many cell lineages are largely conserved and similar genes direct placental development in both species (Table 2 and Supplemental Table 1) [30,31]. Also, in both humans and mice there is differential expression of genes throughout gestation [30,33].

Table 2.

Genes involved in placental development.

Gene Name	Function	Cell/Tissue type in Human	Cell/Tissue type in Mouse
ID2/Id2	Stem cell marker	TSC	TSC
ASH2/Mash2	Transcription factor required for spongiotrophoblast development and proliferation	TSC	Chorion, ectoplacental cone, SpTb
GCM1/Gcm1	Transcription factor required for initiation and morphogenesis and syncytiotrophoblast differentiation	CTB, STB	CTB, STB
HAND1/Hand1	Transcription factor required for giant cell differentiation	Trophectoderm	Trophectoderm, TGC
TEAD3/Tead3	Transcription factor involved in regulation of placental lactogen expression	STB	STB
MMP9/Mmp9	Matrix degrading enzyme	EVT	TGC
Integrin α1β1	Cell adhesion molecule	EVT	TGC
BHLHE40/Stra13	Transcription factor	EVT	TGC
GLUT1 (SLC2A1)/Glut1 (Slc2a1)	Glucose transport	Mesenchyme or endothelium, CTB	Mesenchyme or endothelium, CTB
MEST	Exact role unknown. Abnormal imprinting associated with pregnancy loss.	Mesenchyme or endothelium	Mesenchyme or endothelium
PLAC1/Plac1	Trophoblast differentiation, FGF7 signaling	CTB, STB	All trophoblast cell types
IGF2/IgF2	Promotion of fetal and placental growth, as well as nutrient transfer. Imprinted.	EVT	CTB, Glycogen cells
PECAM 1(CD31)/Pecam1 (Cd31)	Adhesion molecule involved in angiogenesis	EVT	EVT, TGC
UPA (PLAU)/Upa (Plau)	Role in trophoblast invasion	EVT	TGC
HGF/Hgf	Role in angiogenesis and villous development	Mesenchyme or endothelium	Mesenchyme or endothelium
VEGF/Vegf	Regulation of angiogenesis	CTB, STB	TGC
VEGFR-2/FlK1	VEGF receptor	Mesenchyme or endothelium	Mesenchyme or endothelium
VEGFR1/Flt-1	VEGF receptor	CTB, STB, EVT, mesenchyme or endothelium	Glycogen cells, mesenchyme or endothelium
FGFR2/Fgfr2	Role in mitogenesis and differentiation	TSC	TSC
EOMES/Eomes	Transcription factor	TSC	TSC
TFAP2C/Ap-2gamma	Transcription Factor	TSC	TSC
NODAL/Nodal	Required for maintenance of human embryonic stem cell pluripotency and may play a role in human placental development	TSC	TSC

ID2 = inhibitor of DNA binding 2, ASH2 = Achaete-scute family bHLH transcription factor 2, GCM1/Gcm1 = glial cells missing, HAND1 = Heart and neural crest derivatives expressed 1, TEAD3 = TEA domain transcription factor 3, MMP9 = Matrix metalloproteinase 9, BHLHE40 = Basic helix-loop-helix family member e40, GLUT-1/Glut-1 = Glucose transporter 1, MEST = Mesoderm specific transcript, PLAC1 = Placental specific protein 1, IGF2 = Insulin-like growth factor 2, PECAM = platelet endothelial cell adhesion molecule-1, PLAU-plasminogen activator urokinase, HGF/Hgf = Hepatocyte growth factor, VEFGF/R = Vascular endothelial growth factor/Receptor s-Flt1 = soluble fms-like tyrosine kinase 1, FGFR2 = Fibroblast growth factor receptor 2, TFAP2C = Transcription factor AP-2 gamma, TGC = Trophoblast giant cell, STB = syncytiotrophoblast, SpTb = spongiotrophoblast TSC = Trophoblast stem cells, CTB = cytotrophblast

3. The human placenta in trisomy 21

Compared to euploid pregnancies, studies conducted on placental specimens and cell cultures from pregnancies affected by T21 have shown differences in histology, function and gene expression.

3.1. Histological appearance

Limited reports exist regarding the histological appearance of placentas from pregnancies affected by T21. Examination of 14 aneuploid placentas (including three cases of trisomy 21) found that villous mineralization was more common in placentas from aneuploid compared to euploid stillbirths [34]. In euploid pregnancies, as gestational age advances, basal plate calcification is common, but villous mineralization is not. This can be a sign of cessation of fetal artery perfusion and intrauterine cardiac failure and hydrops [34].

A pathological analysis of placentas from 10 cases of trisomy 21 (four liveborn, six stillborn) delivered between 16- and 40-weeks’ gestation, demonstrated hemorrhagic endovasculitis in approximately 80% of T21 placentas compared to only 5% in euploid placentas [35]. Hemorrhagic endovasculitis is associated with stillbirth and fetal growth restriction. Coexisting lesions are often present, including feto-placental vessel thrombi, and villous fibrosis. This may indicate that there is primary placental dysfunction, abnormal vascular development or an abnormal feto-maternal interface [36].

3.2. Villous morphology

Histological analysis of placental biopsies and amniocytes in unaffected and affected pregnancies demonstrated that there was an increase in invasive cytotrophoblast apoptosis, vascular abnormalities, inflammation, villous hypoplasia and intervillous fibrin deposition during the second and third trimester [9–12]. Roberts et al. examined cross sections of placental villi obtained at 10–14 weeks of gestation. Compared to euploid, T21 placentas had an increased percentage of double layer trophoblast and an increased proportion of villous capillaries with nucleated red cells and basophilic stippling, indicating delayed maturation, differentiation and abnormal erythropoiesis [37]. Another study of second and third trimester placentas showed varying amounts of large chorionic villi interspersed with normal appearing villi. Small muscular arteries and arterioles were also absent from some large villi, giving the villi a hypovascular appearance. As seen in previous studies, there was also an increased percentage of two-layer trophoblast observed in T21 placentas [12].

3.3. Cytotrophoblast function

Two differentiation pathways exist for cytotrophoblasts (CTB); formation of the syncytiotrophoblast (STB) and formation of invasive extravillous trophoblasts (EVT). Both pathways show striking differences between T21 and euploid placentas (Fig. 1). Studies conducted using CTB derived from first, second and third trimester euploid placentas found that there was in vitro aggregation and fusion within 48–72 h of forming the STB. However, in T21, aggregation occurred, but there was little or no fusion [9,14]. Interestingly, abnormal trophoblast fusion and differentiation in T21 placentas is reversible in vitro with the addition of exogenous activin-A, a paracrine factor involved in crosstalk between trophoblasts and mesenchymal cells [38,39]. In euploid placentas it is highly secreted by mesenchymal cells but is secreted at a significantly lower level in T21.

Fig. 1.

Cell types and functional abnormalities in T21 and euploid placentas.

Illustration shows placental layers and cell types. Boxes shows magnified view of the cell types in the chorionic villi. The cytotrophoblasts (yellow) are precursors to the syncytiotrophoblast (green), which remains double layered secondary to poor fusion and cell differentiation. The box labeled euploid development shows the syncytiotrophoblast as a multinucleated monolayer surround the underlying cytotrophoblast layer.

In one study, analysis of CTB isolated from euploid and trisomic placentas showed defective STB formation and function leading to abnormally hyperglycosylated beta human chorionic gonadotropin (βhCG). This abnormal and weakly bioactive βhCG molecule cannot correctly stimulate CTB differentiation. In vitro function can be restored by treatment with non-hyperglycosylated βhCG [40]. Another cell culture and RNA expression study performed on placental samples at 7–12 weeks of gestation demonstrated a significant increase in the Luteinizing Hormone/Chorionic Gonadotropin Receptor (LHCGR) mRNA transcription in T21 placentas compared to controls. Despite this, the synthesis of high molecular weight mature LHCGR proteins was significantly reduced in T21 compared to unaffected pregnancies, suggesting a lack of utilization of circulating βhCG [41,42].

To identify the membrane proteins required for human trophoblast fusion, one study found that in euploid placentas cell fusion proteins are highly expressed. For example, syncytin-1 is expressed at elevated levels in the EVT and syncytin-2 is expressed in villous CTB. Syncytin-2 and zona occludens-1 (ZO-1, tight junction protein) are also highly expressed in isolated CTB. Production of these proteins rapidly decreases during STB formation. Syncytin-1 and connexin 43 (Cx43, gap junction protein) increase with cell aggregation and fusion. The expected time-dependent variations in expression of these membrane proteins are not observed in T21 placentas [39]. Additionally, a more recent study found that interferon induced transmembrane proteins (IFITMs) impair STB formation and inhibit syncytin activity, which may explain these findings given that there increased interferon activity seen in Down syndrome [43,44]. In cell cultures from placentas at 18–22 weeks, T21 CTBs were more likely to undergo apoptosis and have abnormal expression of cell surface markers compared to euploid. Specific findings included decreased expression of VE-cadherin (cell-cell adhesion glycoprotein) and upregulation of matrix metallopeptidase 9 (MMP-9, extracellular matrix breakdown). These results imply that in T21, there are abnormalities in the differentiation pathways that lead to CTB invasion of the uterus, and that these cells may be eliminated by programmed cell death from the maternal-fetal interface. In addition, up-regulation of MMP may be a compensatory mechanism for poor uterine invasion [13]. CTBs are also involved in activating the MHC class Ib molecule HLA-G, which is thought to play a role in maternal tolerance of the fetal hemiallograft, which was found to have decreased expression in T21 placentas [13,45, 46].

Lastly, in trophoblasts the keratin cytoskeleton appears to be important in maintaining structural integrity. Keratin 8 (KRT8) also functions in signal transduction and potentially in apoptosis. Klugman et al. demonstrated that in trophoblasts from T21 placentas, TNF-related Apoptosis-I inducing ligand (TRAIL) was significantly overexpressed compared to euploid. There was also an inverse relationship between KRT8 and TRAIL in T21 [47].

3.4. Oxidative stress

STBs have high synthetic and metabolic activities and are therefore vulnerable to oxidative stress [7]. In euploid pregnancies, STBs can adapt to minimal increases in reactive oxidation species (ROS) by restoring the oxidant/antioxidant balance. If this response is inadequate it can lead to hypoxia and chronic oxidative stress, which increases the vascularity of the villi and inhibits the differentiation of CTB to STB [48].

SOD-1, the gene for intra-cellular Cu, Zn-superoxide dismutase (SOD), maps to human chromosome 21. In T21, there are three copies of this gene. A study examining the impact of oxidative stress on placental development showed that overexpression of SOD-1 impairs STB formation [49,50]. It has been shown that ROS, overexpression of SOD-1 and abnormal oxidative status, demonstrated by a significant increase in catalase activity, are present in T21 CTB. In turn, T21 trophoblasts cannot compensate for the oxidative imbalance [49,50]. Excess SOD in T21 placentas prevents differentiation of CTB to STB, thereby down-regulating STB-specific hCG, placental growth factor secretions and synthesis of syncytin [51].

3.5. Placental hormone expression

Prenatal screening for aneuploidy can be performed by measurement of placental derived cell-free DNA or biochemical analytes in maternal serum during the first and/or second trimester (Table 3). In the first trimester, maternal serum free βhCG or total βhCG, and placental derived pregnancy-associated plasma protein-A (PAPP-A) are measured. In the second trimester, alpha-fetoprotein (AFP), unconjugated estriol (uE3), inhibin-A (DIA), and βhCG are measured. Cell-free fetal DNA can be measured any time after 10 weeks’ gestation [52]. Serum analyte trends in pregnancies affected with Down syndrome show significantly decreased PAPP-A, mild decrease in AFP and uE3, and increased DIA, βhCG and cell-free DNA [53–55]. These abnormalities are thought to be secondary to abnormal trophoblast invasion, increased trophoblast apoptosis and delayed maturation of the feto-placental unit [56,57].

Table 3.

Markers used for prenatal screening in trisomy 21.

Hormone	Typical Expression	First Trimester in T21	Second Trimester in T21
βhCG	- Detectable at 4 weeks	↑↑	↑↑
	- Peaks at 10 weeks
	- Declines until 20 weeks
	- Constant until term
PAPP-A	- Detectable at 8 weeks	↓↓	↔
	- Steady increase until term
AFP	- Detectable at 12 weeks	↓	↓
	- Peaks at 32 weeks
	- Declines until term
uE3	- Detectable pre-implantation	↓	↓
	- Steady increase until term
DIA	- Detectable at 5 weeks	↑	↑↑
	- Peaks at 8 weeks
	- Steady decrease from 8 to 16 weeks
	- Steady increase until 36 weeks
Activin-A	- Detectable at 8 weeks	↔	↔
	- Constant level throughout 1st and 2nd trimester
	- Steady increase from 24 weeks to term
PlGF	- Increases steadily throughout gestation	↓	↓
	- Peaks at 30 weeks
	- Steady decrease until term
Cell-free DNA	- Increases throughout gestation	↑	↑
	– 0.1% increase/week from 10 to 20 weeks
	– 1% increase/week from 21 weeks to term

↓ or ↑ = 25% difference, ↓↓ or ↑↑ = 50% difference, ↔ distribution overlaps with normal population.

βhCG and its receptor are also important regulators of embryo implantation and pregnancy maintenance. The beta subunit is placenta specific and is primarily produced by the STB. In addition to its early role in the maintenance of progesterone levels, βhCG facilitates trophoblast differentiation, endometrial decidualization, invasion of maternal spiral arteries and angiogenesis. It also plays a small role in stimulating the thyroid to produce thyroid hormone, which is essential for brain and nervous system development [41,58].

Elevated levels of βhCG in mid-pregnancy are associated with pre-eclampsia, fetal growth restriction and T21 [41,42]. The etiology of maternal serum βhCG elevation in T21 is unclear. However, it has been proposed that it could be due to up-regulation of the βhCG gene, increased half-life of the hyperglycosylated βhCG hormone, or reduced availability of the LHCGR [41].

PAPP-A is a metalloproteinase that increases the bioavailability of insulin-like growth factor binding proteins. It is expressed by the STB, as well as in many other human tissues, but at lower levels than in the placenta [59]. PAPP-A levels are also low in serum from pregnant women carrying fetuses with T21. Extremely low levels are a risk factor for impaired fetal growth, preterm delivery, preeclampsia and stillbirth [60, 61]. A role for this hormone in fetal growth has also been shown in knockout mice [62].

Maternal serum AFP is decreased in the first and second trimester in women carrying fetuses with T21 [63,64]. Although not a placental product, AFP levels are significantly increased in placental tissue while in the fetal liver, AFP levels are only slightly lower than euploid controls. This indicates that there may be a transport defect leading to placental accumulation [57]. In contrast to T21, elevated maternal serum levels of AFP are associated with growth restriction (<5%), preterm delivery and stillbirth, indicating normal placental function, but abnormal fetal production [61].

Estriol (UE3) is a cholesterol derived hormone that regulates utero-placental blood flow, development of placental vascularization and maternal cardiovascular remodeling. It is synthesized by STB from the precursor dehydroepiandrosterone (DHEAS), which originates in the fetal adrenal gland. Estrogen levels remain low in the first trimester of pregnancy and progressively increase, remaining elevated until the onset of labor [58]. UE3 is low in T21, however it is unknown whether this is secondary to fetal production or placental dysfunction [65].

Another placental hormone impacted by T21 is inhibin A. βhCG secretion is in part controlled by inhibin A (inhibition) and activin-A (secretion). Maternal serum inhibin A is elevated in T21 secondary to increased placental production. The exact mechanism for this increased production is unclear, however, it has been postulated that increased activin A production leads to increased βhCG, which in turn leads to increased inhibin A [57]. Conflicting data exist regarding the relationship between βhCG levels and both inhibin A and activin A in T21, given that levels of these hormones are low in both amniotic fluid and mesenchymal cells [66]. Finally, placental growth factor (PlGF) is a member of the vascular endothelial growth factor (VEGF) family. It is an important local mediator of angiogenesis and regulator of maternal Insulin derived growth factor 1 (IGF-1). It is produced by villous CTB, STB and EVT. Although its utility for screening in T21 is currently unclear, decreased levels of PlGF are associated with fetal growth restriction. Compared to euploid placentas, PlGF levels are reduced during the first and second trimester in T21 [57,67–69].

3.6. Placental senescence

The placenta ages as the pregnancy advances, but premature placental aging is associated with adverse pregnancy outcomes. It is also known that DS has been associated with early onset and higher incidence of age-related conditions such as Alzheimer’s disease [3]. A few studies in T21 placentas have found evidence of premature aging, which could be due to increased oxidative stress, mitochondrial dysfunction or telomere shortening that is associated with aneuploidy [10,13,70]. For example, USP16 is overexpressed in DS. It has been shown that upregulation of USP16 in the Ts65Dn mouse model and primary human fibroblasts results in downregulation of the Wnt pathway and reduces stem cell renewal. Although it the role of USP16 in placental development has not been fully elucidated it has been shown that Cdkn2a knockout mice have abnormal placental development [71]. USP 16 regulates activation of Cdkn2a, which in turn acts as a negative regulator of the Wnt pathway. This is relevant since Wnt signaling pathways plays a crucial role in stem cell maintenance, aging and cellular senescence in various tissues including the placenta [72,73].

Additionally, in a study conducted by Wong et al. using samples from first, second and third trimester placentas with T21, a 2.3-fold increase in amyloid precursor protein (APP) expression was found. Results also showed that increased expression of APP was associated with increased apoptosis, which in turn, decreased cell growth and suppressed differentiation of the STB [74].

Rozovski et al. also showed that several key genes that were overexpressed in T21 trophoblast had similar expression patterns to those found in the brains of people with Alzheimer’s disease (Supplemental Table 1). For example, APP was increased by two-fold and LOX was over expressed by three-fold [75].

3.7. Placental gene expression and epigenetic modifications

Hsa21 is the smallest human chromosome, representing ~1% of the human genome [76]. Although small, there is significant enrichment for genes located in cytoskeletal structures compared to the rest of the genome. Importantly, these cytoskeletal proteins play an important role in the development of the placenta and nervous system. Hsa21 also contains 31 transcription factors [76]. Overexpression of some of these transcription factors has been seen in multiple studies (Supplemental Table 1) and may lead to altered transcription of non-Hsa21 targets [3]. However, further studies are needed to confirm these hypotheses.

Kipp et al. evaluated placental samples from second and third trimester pregnancies with T21. They demonstrated that in T21 samples there was a trend towards higher levels of inhibin alpha (INHA) mRNA levels and decreased levels of Wilms Tumor 1 (WT1) mRNA levels, a repressor of INHA. The pathological basis of this alteration is not completely understood [77].

Bianco et al. studied gene expression profiles of second trimester placental samples from pregnancies affected by trisomies 13, 18 and 21. T21 was associated with the highest number of dysregulated genes. Greater than 90% of the dysregulated genes did not map to chromosome 21. T21 also had more pleiotropic effects and affected a larger number of signaling pathways than T13 or T18. These signaling pathways, however, were less crucial for development. In T21 there was also differential expression of 160 genes when compared to euploid placentas (145 up-regulated and 15 down-regulated). Differential expression was most notable for genes encoded by chromosome 21 (ABCG, ITSN1, ADAMTS1 and PLAC4), as well as HGF, EGFR, S100A and RAC1 (Supplemental Table 1) [78].

Using microarrays, Rozovski et al. performed gene expression analyses of chorionic villus samples (CVS) from T21 and euploid controls. They found genes that mapped to chromosome 21 were over-represented by 4.5-fold, but most of the differentially expressed genes did not map to chromosome 21. Some had greater than 10-fold over-expression. The most significant difference was seen in MEST, which was increased 13.7-fold in T21. Another interesting finding in this study was the two-fold increase of ITM2B [75].

Gene expression can also be impacted by DNA methylation. Studies regarding the methylation status of DS placentas present conflicting data, which may be secondary to differences in gestational ages. Lim et al. examined first trimester CVS samples and found differential gene expression in trisomic placentas compared to euploid controls. They demonstrated that 47 genes on Hsa21 had increased expression in T21 placentas; these genes were globally hypomethylated [79]. Another study by Jin et al. analyzed CVS samples from the first and second trimesters and found global DNA hypermethylation. More specifically, genes involved in epigenetic regulation (Ten-eleven translocation methylcytosine dioxygenase 1 and 2; TET1, TET2 and RE1-Silencing Transcription factor; REST) were hypermethylated (Supplemental Table 1) [80].

Recently, investigators have become interested in miRNA expression and function due to their involvement in post-transcriptional regulation of gene expression and its potential impact on development, including in cases of T21. An analysis of miRNAs identified 34 that were significantly differentially expressed (16 upregulated, 18 downregulated) in T21 compared to controls, and 76 possible gene targets located on chromosome 21. Differences were not seen in chromosome 21 derived miRNA. Target genes on chromosome 21 were associated with T21-related complications such as intellectual disability, neurobehavioral manifestations, and congenital abnormalities [81].

3.8. Placental proteome dysregulation

The underlying mechanisms of phenotypic variability in T21 are poorly understood. To date, no prenatal biomarkers that predict phenotype in T21-have been identified. In this context, the placental proteome offers a unique opportunity to identify biomarkers associated with disease severity and to potentially monitor treatment response and predict outcomes [82,83]. Several studies have looked for protein biomarkers using maternal serum. Few however show significant placental expression (Supplemental Table 1) [84–91].

To our knowledge, only two studies have investigated global proteomic dysregulation in T21 placentas using two-dimensional gel electrophoresis coupled with tandem mass spectroscopy (Supplemental Table 1). In the first, Sun et al. compared the proteomic profiles of five T21 and five euploid second and third trimester placentas. They identified seven significantly up-regulated (ANXA2, ERp29, SOD1, PSMA2, PPIA and FGB) and three down-regulated proteins (ECHS1, PRDX6 and PDIA3) in T21 placentas [90]. In the second, Chen et al. found 80 unique dysregulated proteins in T21 placentas between 16 and 18 weeks of gestation. In addition to a few proteins in maternal plasma that have been described in other studies (A1AT, VIM), this study reported novel placental proteins (ATXN3, XRCC2, CSH1, GBP2 and SPRED2) that can be used in combination with other T21-specific biomarkers [92].

4. The placenta in mouse models of down syndrome

In the human, functional understanding of the human placenta is limited to postpartum studies and some in vitro systems. Because of this, animal models are critical for investigation of basic placental biology. There are, however, limitations to using mouse models, given that no single mouse model completely recapitulates the full phenotype observed in people with T21 [92]. There are also many developmental processes that occur during intrauterine human life that are postnatal events in mice. Another issue is that significant differences exist in the lengths of murine and human pregnancies. On the other hand, changes in the amount of oxidative stress within the placenta throughout gestation are similar in both species [30]. Lastly in mice, the persistent yolk sac provides a secondary circulation, which may protect against environmental insults. Recent data, however, indicate that the human yolk sac may play more than just a vestigial role [93].

To better understand the pathophysiology of DS, many mouse models of DS have been generated [23]. The Ts16 mouse was created in 1985 but was shown to be lethal during the embryonic period. This model harbors a full trisomy of Mmu16, including 119 protein-coding genes that are orthologous to human chromosome 21 (Hsa21). Due to its early lethality, this mouse is not useful for studying DS phenotypes. Information on its placental development is, however, available and summarized below [94]. Subsequently, efforts have focused on generating models with segmental Mmu10, Mmu16 and Mmu17 trisomies that carry three copies of Hsa21 orthologous genes. The Ts1Cje, Ts65Dn and Dp(16)1/Yey mice are the most commonly used models in current studies of DS. While all three have large segments of triplicated Hsa21-orthologous genes on Mmu16, each was engineered using a distinct methodology, resulting in different cytogenetic profiles, numbers of triplicated genes, and prenatal and postnatal phenotypes [21]. Currently only limited data on the placentas from Ts1Cje mice exist, and there are no data on placentas from Ts65Dn or Dp(16)1/Yey mice.

4.1. Trisomy 16 (T16) mouse model

In one study, App expression was analyzed in T16 embryos and placentas at days E15 and E17. Although localization of App mRNA and protein in the brain and placenta was unchanged, there was a marked increase in App expression in the placenta that did not correlate with obvious differences in morphology [94].

In another study, histological analysis at days E15 and E17 found that there were decreased numbers of glycogen cells, giant cells and spongiotrophoblast cells. This study also demonstrated labyrinth and placental weights to be smaller than in controls [95].

Lastly, a study by Kornguth et al. showed that T16 placentas had a less extensively developed labyrinth, causing decreased nutrient supplies. Additionally, the junctional zone showed no significant signs of normal differentiation throughout gestation. When comparing placental cells and neurons, there was an altered pattern of proteins and increased levels of class I MHC H-2Kk on their cell surfaces, potentially affecting cell-cell interactions [96].

4.2. Ts1Cje mouse model

The Ts1Cje mouse demonstrates mild to moderate phenotypic abnormalities. Pennings et al. performed a gene expression study that compared fetal livers and placentas of Ts1Cje and wild-type embryonic mice at day E15.5. Increased expression of genes in the trisomic segment was shown. Forty-eight differentially expressed genes were present in the placentas of affected mice. Most were ascribed to the gene dosage effects of the trisomic locus at Mmu16. Interestingly, the fetal liver of Ts1Cje embryos exhibited more differentially expressed genes (109 genes) with very little overlap with the placenta outside the Mmu16 trisomic genes [97].

5. Future research and conclusions

In this review we have shown that compared to euploid placentas, there are multiple differences in the placentas of human fetuses with T21. By contrast, scant attention has been paid to placental development in preclinical mouse models of DS. This is a lost opportunity for understanding clues to the mechanisms underlying atypical development in DS and associated disease.

A central issue is whether the abnormalities seen in human placentas in T21 have downstream consequences regarding fetal or infant survival, growth restriction, and brain or another organ development. For instance, cardiac anomalies are seen in 40% of individuals with DS [6]. Although the mechanisms between the placenta, fetal heart development and later cardiac dysfunction have not been fully elucidated, it is hypothesized that there are links in common genetic developmental pathways such as Wnt/beta catenin signaling, angiogenesis and cell adhesion. Other theories include placental insufficiency leading to growth restriction, poor nutrient transfer and oxidative stress [98].

Placental abnormalities in DS include differences in CTB fusion that affect subsequent conversion to STB, atypical oxidative stress/antioxidant balance, and increased expression of genes that are also upregulated in the brains of individuals with Alzheimer’s disease. There is also a suggestion that placentas in human T21 are prematurely senescent, showing atypical evidence of mineralization. Is there a connection between the premature aging of the placenta and the premature aging of people with DS?

Initiatives such as the Human Placenta Project, created to address the need for better real-time assessment of human placental structure and function, and the Placental Atlas Tool were developed to create a shared research platform for placental data. These resources will be important in establishing normal parameters for placental development and creating data repositories to distinguish normal from abnormally developing placentas [99,100].

Based upon the relative paucity of data on placental development in preclinical mouse models of DS, we recommend that future evaluation of new and existing models routinely include histologic and functional assessments of the placenta. It is crucial to understand the role of this important organ in subsequent development of brain and other structural and functional abnormalities.