Novel Insights from Fetal and Placental Phenotyping in Three Mouse Models of Down Syndrome

Abstract

Background:

In human fetuses with Down syndrome, placental pathology structural anomalies and growth restriction are present. There is currently a significant lack of information regarding the early lifespan in mouse models of Down syndrome.

Objective:

The objective of this study was to examine embryonic (E18.5) and placental genotypes in the three most common mouse models of Down syndrome (Ts65Dn, Dp(16)1/Yey, Ts1Cje). Based on prenatal and placental phenotyping in three mouse models of Down syndrome we hypothesized that one or more of them would have a similar phenotype to human fetuses with trisomy 21, which would make it the most suitable for in utero treatment studies.

Study Design:

C57BL6J/6 females were mated to Dp(16)1/Yey and Ts1Cje males and Ts65Dn females to C57BL/B6Eic3Sn.BLiAF1/J males. At E18.5, dams were euthanized. Embryos and placentas were examined blindly for weight and size. Embryos were characterized as euploid or trisomic, male or female by polymerase chain reaction. A subset (34 euploid, 34 trisomic) was examined for malformations.

Results:

The Ts65Dn model showed the largest difference in fetal growth, brain development and placental development when comparing euploid and trisomic embryos. For the Dp(16)1/Yey model genotype did not impact fetal growth, but there were differences in brain and placental development. For the Ts1Cje model no significant association was found between genotype and fetal growth, brain development or placental development. Euploid embryos had no congenital anomalies; one was demised. Hepatic necrosis was seen in 6/12 (50%) of Dp(16)1/Yey and 1/12 (8%) Ts1Cje embryos; hepatic congestion/inflammation was observed in 3/10 (30%) Ts65Dn embryos. Renal pelvis dilation was seen in 5/12 (42%) Dp(16)1/Yey, 5/10 (50%) Ts65Dn and 3/12 (25%) Ts1Cje embryos. One Ts65Dn and one Dp(16)1/Yey embryo had an aortic outflow abnormality. Two Ts1Cje embryos had ventricular septal defects. Ts65Dn placentas had increased spongiotrophoblast necrosis

Conclusions:

Fetal and placental growth showed varying trends across strains. Congenital anomalies were primarily seen in trisomic embryos. The presence of liver abnormalities in all three mouse models of Down syndrome (10/34) is a novel finding. Renal pelvis dilation was also common (13/34). Future research will examine human autopsy material to determine if these findings are relevant to infants with Down syndrome. Differences in placental histology were also observed between strains.

Introduction

Down syndrome (DS) affects 1 in 700–1200 live births.¹‘² In humans with DS, placental pathology, structural anomalies and growth restriction occurs prenatally.^3,4,5 In prior work from our laboratory, we identified significant phenotypic and gene expression differences between the three most commonly used mouse models of Down syndrome (Ts1Cje, Ts65Dn, and Dp(16)1/Yey).^6–13

While all three models have large segments of triplicated Hsa21-orthologous genes on Mmu16, each one was engineered using a different methodology, resulting in different cytogenetic profiles and numbers of triplicated genes.^6,7 The Ts1Cje mouse model was generated via a reciprocal translocation of the distal portion of Mmu16 onto the telomeric region of Mmu12⁸.This created an elongated Mmu12 carrying an additional dose of 71 Hsa21 orthologs, but also led to the monosomy of seven telomeric genes and loss of a functional copy of superoxide dismutase 1⁹. The Ts65Dn is the most widely used trisomic mouse model. This mouse was generated by cesium irradiation that induced a reciprocal translocation of the most distal portion of Mmu16 onto a separate marker chromosome containing the centromeric portion of Mmu17^10,11. Because this triplication is carried as an additional freely segregating chromosome, the Ts65Dn mouse uniquely models the aneuploidy observed in 95% of individuals with DS¹². The triplicated segment consists of ~ 104 HSA21 orthologs as well as 60 centromeric Mmu 17 genes that are not triplicated in humans with DS⁹. The Dp(16)1/Yey mouse model was generated using Cre-mediated recombination to duplicate the entire 23.3 Mb segment of Hsa21 orthologs (~119 genes), adding them onto the distal portion of one of the endogenous Mmu16 chromosomes¹³. As this model contains the largest number of triplicated Hsa21 orthologs and lacks any perturbation of unrelated genes compared with Ts1Cje and Ts65Dn mice, Dp(16)1/Yey mice should, in theory, have the most similar murine representation of the phenotypes seen in people with DS.

Our prior study included a comprehensive assessment of gross anatomy, brain histology, brain gene expression in embryonic (E15.5), neonatal and adult mice, as well as postnatal behavior. However, because our work is focused on identifying prenatal treatments that could be given to a pregnant woman carrying a fetus with DS¹⁴, a more detailed analysis of the antenatal phenotypes of the embryo and placenta in mouse models of DS is warranted. A significant lack of information regarding the early lifespan in mouse models has been described.¹⁵ Our objective was to examine embryonic (E18.5) and placental phenotypes in three mouse models of DS because previous studies in the Ts1Cje mouse model showed postnatal growth restriction associated with early neonatal mortality¹⁶. Based on prenatal and placental phenotyping in three mouse models of Down syndrome we hypothesized that one or more of them would be similar to the human fetal phenotype in trisomy 21, making it the most suitable for in utero treatment studies.

Material and Methods

Animal Housing and Breeding

This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. All experimental procedures [NHGRI Protocol G-17–1] were approved by the Institutional Animal Care and Use Committee (IACUC). Mice were housed in standard cages with food and water ad libitum and in a controlled environment (temperature = 20 °C; humidity = 60%; light/dark cycles of 12 hours).

Animal breeding was performed using standardized methodology as described in our prior studies.¹⁴ B6.Cg-T(12;16)1Cje/CjeDnJ mice (Ts1Cje; stock number 004838), B6EiC3Sn.BLiA-Ts(17¹⁶)65Dn/DnJ (Ts65Dn; stock number 005252) mice and B6129S-Dp(16Lipi-Zfp295)1Yey/J (Dp(16)1/Yey; stock number 013530) mice were purchased from Jackson Laboratory (Bar Harbor, ME). Ts65Dn female mice were bred with B6EiC3Sn.BLiAF1/J (F1 hybrid; stock number 003647) males. Ts1Cje and Dp(16)1/Yey males were bred with C57BL/6J (Jackson Laboratory)female mice. Matings were set up every day between 4:00 and 5: 00 PM and separated every morning (8:00–9:00 AM) after examination for the presence of vaginal plugs. The presence of vaginal plugs was defined as embryonic day 0.5 (E0.5). The pregnancy was confirmed by a 15–20% weight gain 10 days later.

E18.5 Embryo and Placenta Studies

At embryonic day 18.5 (E18.5), pregnant females (Litters: Ts1Cje =11, Dp16 =10, Ts65Dn = 10) were anesthetized with 2.5% isoflurane in a 3/7O2/N20 mixture and euthanized by decapitation. Embryos were extracted in ice cold 1X phosphate-buffered-saline (PBS) containing RNAlater® stabilization solution (Invitrogen). A tail snip from each embryo was used for genotyping. Each embryo was photographed and body weight, crown rump length (CRL), brain and placenta weights and diameter were measured (individual embryo data located in Supplemental Tables S1–3). The placenta was bisected at the umbilical cord into two pieces. Half of each placenta was fixed for 24 hours in 4% paraformaldehyde at 4 °C, washed three times with PBS and placed in 30% sucrose at 4 °C. The placentas were then frozen in Optimal Cutting Temperature Compound (OCT; Sakura, Torrance, CA). Tissue blocks were stored at −20 °C for future immunohistochemistry experiments. The other half of the placentas and the brain specimens were snap frozen in liquid nitrogen before storage at −80 °C for future transcriptome analysis.

In a separate round of breeding the entire embryo was kept intact for detailed fetal phenotyping. Pregnant females (Litters: Ts1Cje =6, Dp(16)1/YeY =6, Ts65Dn = 5) were euthanized and embryos and placentas were documented as above (individual embryo data located in Supplemental Tables S1–3, without brain weights). Embryos and placentas were fixed for 24 hours in formalin and the transferred to 70% ethanol solution for storage prior to gross and histologic examination.

Extraction and Genotyping

DNA extraction was performed using the NucleoSpin Tissue XS kit according to the manufacturer’s instructions (Macherey-Nagel, Bethelehem, PA). DNA concentrations were measured as absorbance at 260 nm on the Nanodrop instrument (Thermo Fisher, Scientific, Waltham, MA). Genotype and sex determination was performed via multiplex PCR amplification using primers specific for the Ts1Cje¹⁷, Dp(16)1/Yey¹⁸ or Ts65Dn¹⁹ translocation breakpoints and the SRY gene along with an internal positive control (Supplemental Table S4). Initial genotypes were independently confirmed by a third-party genotyping service (Transnetyx, Cordova, TN).

Gross Anatomic and Histological Examination

To assess for the presence of congenital anomalies and gross anatomical differences in placental and brain morphology a cross sectional sample of 2 euploid and 2 trisomic embryos was chosen at random from each litter in the second round of breeding (Embryos: Ts1Cje = 12, Eup=12; Dp16 =12, Eup=12, Ts65Dn = 10, Eup=10). Paraffin embedded embryos were serially sectioned (step size of 200 μm). For each embryo at least 30 serial sections (8 μm thick) were obtained. For each placenta, at least 10 serial sections (8 μm thick) were obtained and stained with hematoxylin and eosin (H&E) (Histoserv, Bethesda, MD). The presence of congenital anomalies and differences in placental morphology were analyzed by light microscopy by an experienced veterinary pathologist (VH) blinded to genotype and sex. After histological analysis, the genotype and sex were unblinded and the slides were reviewed a second time (Supplemental tables S5–7).

Statistical Analysis

In all experiments, trisomic mice were compared with their euploid littermates using the same methodology as described previously.⁶ This was done precisely to avoid comparing trisomic mice of one strain with euploid mice from another strain. Specifically, it is well known that baseline differences exist between the various sub-strains of C57BL/6 mice, as well as those mice on a C57BL/6XC3Sn hybrid background. Therefore, to ensure that any phenotypic differences arose only from the trisomy and not from genetic differences in background strains, we first compared trisomic animals with euploids of that strain, then evaluated the presence, absence or magnitude of phenotypic differences across strains, to compare and contrast the three mouse models. Congenital anomalies were quantified and presented as frequencies and percentages. Placental morphology was scored by the presence and appearance of common cell types. For each cell type a score of 0–3 (normal, mild, moderate, severe) was given. Means scores were calculated and used for comparison of euploid versus trisomic placentas.

Data were analyzed with Graph Pad Prism software. Means, medians, standard deviations and ranges were calculated for continuous variables, and frequencies and percentages were calculated for categorical variables. For the continuous variables, differences in the averages between the two groups were tested using the t-test when the data were normally distributed, and the non-parametric Wilcoxon rank sum test was used when the normality assumption was not satisfied. Categorical variables were compared using Chi-squares test. A p-value of <.05 was considered a statistically significant difference.

Linear models were fitted for each strain and generalized estimating equations with exchangeable correlational structure were used to account for inter-litter covariation. Mean differences due to genotype were assessed for four biometric variables: weight, length, placental weight and placental diameter; all fitted models controlled for sex and litter size.

Results

Embryo Genotype and Sex Distribution

Genotype and sex were determined for all recovered embryos for each strain at day E18.5. The number of litters, embryos, their sex, and percentage of euploid and trisomic mice for each strain is given in Table 1.

Table 1.

Breeding Results

Strain	# of litters	# of Embryos	Abnormal1,2	Embryos used in histological analysis	litter size (range)	Maternal Weight Gain (g) (range)	Fetal Sex			Genotype
							Female N(%)	Male N(%)	P-value	Euploid N(%)	Trisomic N(%)	P-value
Ts65Dn	15	92	20	10	7(3-9)	15.8 (8.1-21.7)	49 (53)	43 (47)	0.65	44 (48)	48 (52)	0.30
Dp(16)1/Yey	16	118	19	12	8(3-12)	14.8 (9.9-18.1)	52 (44)	66 (56)	0.07	62 (53)	56 (47)	0.36
Ts1Cje	17	126	17	12	8 (4-10)	15.5 (8.1-21.7)	60 (48)	66 (52)	0.53	51 (40)	75 (60)	<0.01

Data are median and range unless otherwise specified. P-value <0.05 is significant.

¹Excluded from analysis

²Includes resorbed embryos, hydropic or macerated embryos, anencephaly

Biometric Analysis

For each strain, comparisons were made between euploid and trisomic embryos with regard to total weight, CRL, placental weight and diameter, brain weight and fetal:placental weight ratio (F:P), stratified by sex (Tables 2A–C).

Table 2.

Embryo Characteristics

A. Ts65Dn
Sex (N)	Genotype	N (%)	Weight (g)	P-value	CRL (mm)	P-value	Placental Weight (mg)	P-value	Placental Diameter (mm)	P-value	Brain Weight (mg)	P-value	F:P1	P-value
Female (49)	Euploid	22 (45)	1.37 (1.14-1.80)	<0.01	21.76 (19.97-25.45)	0.02	107.9 (80.9-138.0)	0.57	8.38 (7.27-9.71)	0.10	78.36 (67.40-99.10)	<0.01	12.86 (8.870-16.81)	0.24
	Trisomic	27 (55)	1.24 (0.94-1.50)		20.74 (18.77-23.58)		105.2 (71.2-148.2)		8.08 (6.63-9.08)		72.19 (63.50-80.60)		12.13 (8.204-18.28)
Male (43)	Euploid	22 (51)	1.45 (1.09-1.77)	0.08	22.61 (20.10-26.33)	0.01	121.1 (90.1-159.4)	0.25	8.85 (7.88-9.93)	0.53	77.97 (67.20-97.60)	0.25	12.16 (8.586-16.02)	0.13
	Trisomic	21 (49)	1.37 (1.23-1.86)		21.59 (20.48-22.97)		114.9 (91.1-152.9)		8.69 (6.63-10.85)		75.16 (67.30-88.20)		12.06 (8.143-15.57)
All (92)	Euploid	44 (47)	1.41 (1.09-1.80)	<0.01	22.18 (19.97-26.33)	<0.01	114.5 (80.9-159.4)	0.17	8.62 (7.27-9.93)	0.09	78.17 (67.20-99.10)	<0.01	12.51 (8.586-16.81)	0.75
	Trisomic	48(52)	1.30 (0.94-1.86)		21.11 (18.77-23.58)		109.4 (71.2-152.9)		8.35 (6.63-10.85)		73.45 (63.50-88.20)		12.10 (8.143-18.28)
B. Dp (16)1/Yey
Sex (N)	Genotype	N (%)	Weight (g)	P-value	CRL (mm)	P-value	Placental Weight (mg)	P-value	Placental Diameter (mm)	P-value	Brain Weight (mg)	P-value	F:P1	P-value
Female (52)	Euploid	29	1.13 (0.92-1.37)	0.07	20.06 (17.79-22.94)	0.88	86.10 (63.50-118.6)	0.88	8.223 (7.070-10.12)	0.54	71.84 (62.0081.60)	0.02	13.34 (8.943-17.32)	0.52
	Trisomic	23	1.07 (0.86-1.21)		20.12 (18.24-22.22)		85.51 (62.50-115.9)		8,106 (6.970-9.60)		67.18 (56.60-74.00)		12.93 (7.17-17.04)
Male (66)	Euploid	33	1.16 (0.86-1.38)	0.04	20.16 (16.99-22.94)	0.17	100.0 (68.70-145.3)	0.10	8.441 (6.960-10.46)	0.19	74.46 (58.60-86.30)	<0.01	11.93 (6.470-15.78)	0.81
	Trisomic	33	1.09(0.74-1.28)		19.66 (17.17-22.59)		92.79 (57.00-137.6)		8.215 (6.410-10.00)		67.62 (47.00-80.40_		12.07 (8.096-17.87)
All (118)	Euploid	62 (58)	1.14 (0.86-1.38)	<0.01	20.12 (16.99-22.94)	0.30	93.52 (63.50-145.3)	0.24	8.339 (6.960-10.46)	0.19	73.29 (58.60-86.30)	<0.01	12.59 (6.470-17.32)	0.70
	Trisomic	56 (42)	1.08 (0.74-1.28)		19.85 (17.17-22.59)		89.80 (57.00-137.6)		8.170 (6.410-10.00)		67.46 (47.0-80.40)		12.42 (7.717-17.87)
C. Ts1Cje
Sex (N)	Genotype	N (%)	Weight (g)	P-value	CRL (mm)	P-value	Placental Weight (mg)	P-value	Placental Diameter (mm)	P-value	Brain Weight (mg)	P-value	F:P1	P-value
Female (60)	Euploid	25 (42)	1.05 (0.83-1.28)	0.94	19.8 (15.7-22.8)	0.38	79.9 (52.1-116.9	0.29	8.28 (6.90-9.29)	0.33	70.7 (58.4-85.2)	0.80	13.8 (9.31-16.8)	0.37
	Trisomic	35 (58)	1.04 (0.82-1.27)		19.5 (15.6-21.7)		75.8 (59.2-120.2)		8.12 (6.69-9.58)		71.2 (62.3-82.5)		14.3 (9.22-17.5)
Male (66)	Euploid	26 (39)	1.04 (0.61-1.27)	0.53	20.2 (15.1-23.8)	0.97	80.6 (59.6-110.5)	0.85	8.57 (7.31-9.71)	0.54	66.9 (54.9-79.5)	0.01	13.1 (8.04-17.1)	0.57
	Trisomic	40 (61)	1.06 (0.86-1.30)		20.2 (16.3-23.9)		81.4 (56.3-136.0)		8.65 (7.87-9.72)		72.3 (64.0-80.7)		13.4 (8.08-17.2)
All (126)	Euploid	51(40)	1.05 (0.61-1.28)	0.76	20.0 (15.1-23.8)	0.64	80.3 (52.1-116.9)	0.596	8.43 (6.90-9.71)	0.83	68.8 (54.9-85.2)	0.06	13.4 (8.04-17.1)	0.33
	Trisomic	75 (60)	1.05 (0.82-1.30)		19.9 (15.6-23.9)		78.8 (56.3-136.0)		8.41 (6.69-9.72)		71.7 (62.3-82.5)		13.8 (8.08-17.5)

¹F:P is fetal weight:placental weight ratio.

Data are mean and range unless otherwise specified. Brain weights do not include animals used for histological analysis.

P <0.05 is statistically significant.

Regression analysis demonstrated strain differences in biometric variables due to genotype and sex (Table 3). For Ts65Dn, the genotype differences remained for weight (p <.001) and length (p < 0.01), with a trisomic genotype resulting in lighter and shorter embryos, controlling for sex. Sex differences were also observed across all biometric assessments except placental weight, with female embryos significantly smaller than male embryos. For the Dp(16)1/Yey genotype differences were observed with placental weight (p = 0.02) and diameter (p < 0.01) and significant sex differences observed for placental weight (p < .01). No further genotype differences were seen for Ts1Cje mice, however, sex differences were observed with respect to length and placental diameter (p<.01 and p<.001, respectively).

Table 3.

Linear Regression Analysis

		Weight		Length		Placental Weight		Placental Diameter
	Variable	B (SE)	p-value	B (SE)	p-value	B(SE)	p-value	B(SE)	p-value
Ts65Dn	Sex (F)	−.052 (.026)	.041	−.548 (.265)	.039	−.006 (.005)	.201	−.411(.132)	.002
	Genotype (TS)	−.113 (.025)	<.001	−.941 (.353)	.008	−.003 (.002)	.086	−.170(.137)	.214
	Litter size	−.006(.025)	.816	−.175 (.178)	.328	.000 (.002)	.837	−.128 (.025)	<.001
Dp(16)1/Yey	Sex (F)	−.017 (.022)	.441	−.150 (.291)	.607	−.006 (.002)	.007	−.110(.102)	.280
	Genotype (TS)	−.033 (.024)	.166	−.026 (.195)	.892	−.004 (.002)	.019	−.269(.098)	.006
	Litter size	.000 (.010)	.997	.059 (.124)	.635	−.002 (.002)	.142	−.192 (.051)	<.001
Ts1Cje	Sex (F)	.002 (.019)	.925	−.521 (.188)	.006	−.012 (.008)	.119	−.400(.101)	<.001
	Genotype (TS)	.008 (.018)	.646	−.094 (.260)	.718	.007 (.008)	.366	−.062(.093)	.507
	Litter size	−.034(.016)	.030	−.275 (.139)	.048	.002 (.003)	.560	−.116 (.043)	.006

P <0.05 is statistically significant.

F= female TS = Trisomic

Of note, the Ts1Cje model was the most sensitive to a litter size effect with larger litters resulting in lighter (p=.03) and shorter (p=.05) embryos, controlling for genotype and sex. A smaller placental diameter, controlling for genotype and sex, was observed in larger litters across all strains (Ts65Dn: p<.001; Dp(16)1/Yey: p<.001; Ts1Cje: p<.01).

Fetal Phenotyping Studies

A subset of 34 euploid and 34 trisomic embryos were selected at random for detailed gross anatomic and histologic studies across all strains. The summary of fetal and placental phenotyping findings is found in Tables 4A and B (details of the histological analyses for individual embryos located in Tables S5–7).

Table 4.

Summary of Fetal Phenotyping Results

A. Proportion of Congenital Anomalies
Strain	Genotype1	Embryo Characteristics
		Cardiac Defects N (%)	Liver/Pancreas Findings N (%)	Brain Finding N (%)	Renal Pelvis Dilation/Hydronephrosis (N%)
Ts65Dn	Euploid	0 (0)	0 (0)	0 (0)	1 (10)
	Trisomic	1 (10)	3 (30)	0 (0)	5 (50)
Dp(16)1/Yey	Euploid	0 (0)	0 (0)	0 (0)	1 (8)
	Trisomic	1 (8)	6 (50)	0 (0)	5 (42)
Ts1Cje	Euploid	0 (0)	0 (0)	0 (0)	0(0)
	Trisomic	2 (16)	1 (8)	0 (0)	3 (25)

B. Placental Characteristics
Strain	Genotype1	Placenta Characteristics2
		Labyrinth Necrosis	SpTB Necrosis	Giant Cell (Decrease)	Glycogen Cell (Increase)	Fibrin (Increase)	Decidua Necrosis
Ts65Dn	Euploid	0	0	2	2	2	2
	Trisomic	0	0.5	3	3	1.5	1
	P-value	1	0.03	0.04	0.19	0.51	0.72
Dp(16)1/Yey	Euploid	0	1	1.5	1	1	1
	Trisomic	0.5	1	1	0	1	1
	P-value	0.68	1	0.99	0.14	0.57	0.81
Ts1Cje	Euploid	1	1	1.5	1	1	1
	Trisomic	1	0.5	1	0	0.5	1
	P-value	1	0.99	0.65	0.82	0.05	1

¹1Ts1Cje: 6 litters (12 euploid, 12 trisomic), Dp16(1)/Yey: 6 litters (12 euploid, 12 trisomic), Ts65Dn: 5 litters (10 euploid, 10 trisomic)

²Data are mean scores. Comparisons made using t-test with p<0.05 considered statistically significant.

³Score key (deviation from normal appearance): 0 = normal, 1 = mild, 2 = moderate, 3 = severe

Among the Ts65Dn mice (10 euploid/10 trisomic), congenital anomalies were primarily seen in trisomic embryos, but 1 euploid embryo had renal pelvis dilation. Three trisomic embryos had hepatic congestion/inflammation (Fig. 1C upper image). One trisomic embryo had focal hepatitis and pancreatic fibrosis (Fig. 1C lower image, Figure 1D). Renal anomalies were common; one had hydronephrosis and four had dilated renal pelvices (Fig. 2A–C). One trisomic embryo had an aortic outflow abnormality (Fig. 3A and B). Differences were also seen in the placentas, showing increased spongiotrophoblast necrosis and decreased giant cells in trisomic placentas (Fig. 4A –C).

Figure 1. Liver Pathology.

Hematoxylin and Eosin (H&E) stained sections of E18.5 embryonic liver. A sagittal section of normal liver in a euploid Dp16(1)/YeY embryo. Upper 2X magnification. Lower 10X magnification. B. Sagittal section of liver demonstrating hepatic necrosis in a Dp16(1)/Yey embryo. Upper 2X magnification. Lower 20X magnification. C. Ts65Dn embryo. Upper sagittal section of liver with hepatic congestion (2X magnification). Lower sagittal section of pancreatic fibrosis (10X magnification). Arrows indicate pathologic findings D. Sagittal section of liver with focal hepatitis in a Ts65Dn embryo. Upper 10X magnification. Lower 40X magnification.

The red box in the low magnification section represents the location of the high magnification image.

Figure 2. Renal Pathology.

H&E stained sections of embryonic kidneys (E18.5). A. Sagittal section of normal kidney from a euploid Ts65Dn embryo at 2X magnification. B. Sagittal section of kidney with dilation of the renal pelvis from a trisomic Ts65Dn embryo at 2X magnification. C. Sagittal section of kidney with severe hydronephrosis with dilation of the proximal ureter from a trisomic Ts65Dn embryo at 2X magnification.

C= Renal cortex, G = Glomeruli, RP = renal pelvis.

Figure 3. Cardiac Pathology.

H&E stained sections of embryonic hearts (E18.5). A. Sagittal section of a normal cardiac anatomy in a trisomic Ts65Dn embryo at 2X magnification. The normal orientation of the aortic valve and aorta can be seen arising from the left ventricle. B. Sagittal section of an aortic outflow abnormality in a trisomic Ts65Dn embryo at 2X magnification. The aorta and aortic valve are seen arising from the right ventricle. C. Sagittal section of normal interventricular septum in a trisomic Ts1Cje embryo at 2X magnification. The arrow points to the intact septum between the left and right ventricles. D. Sagittal section showing a ventricular septal defect in a trisomic Ts1Cje embryo. The arrow is pointing to a defect in the septum with blood flow between the left and right ventricles. The red box in the low magnification section represents the location of the high magnification image in E. Upper 2X magnification. Lower 10X magnification.

LV= Left ventricle. RV = Right ventricle. LA = Left atrium. RA = right atrium. IVS = interventricular septum. Ao = Aorta. Aov = Aortic valve. VSD = Ventricular septal defect.

Figure 4. Placental Abnormalities.

Hematoxylin and eosin- stained sections of embryonic placentas (embryonic day 18.5). A. Sagittal section of placenta from a euploid littermate control from a Ts65Dn mating. Upper 2X magnification shows layers of the placenta. Decidua basalis (maternal side), Junctional zone (location of spongiotrophoblast cells and glycogen cells, with giant cells lining the periphery), labyrinth zone (fetal vasculature lined by multinucleated progenitor cells (syncytiotrophoblast), chorionic plate (fetal side). Lower 10X magnification. The arrows indicate layers and cell types. B. Sagittal section of a Ts65Dn placenta at 10X magnification showing evidence of syncytiotrophoblast necrosis. The box indicates the area of necrosis. The arrows are pointing to cells with both nuclear and cellular swelling, pale, eosinophilic cytoplasm, cellular debris and nuclear fragmentation. C. Sagittal section of Ts65Dn placenta at 10X magnification showing decreased number of giant cells. The outlined area shows the location of the spongiotrophopblast layer and absence of giant cells in the periphery. Jz = Junctional zone. Lb = Labyrinth. Dc = Decidua Basalis. STB = Syncytiotrophoblast. SpTb = Spongiotrophoblast. GC = Giant Cell. GlyC = Glycogen cell

Histological analysis of the Dp(16)1/Yey mice (12 euploid/12 trisomic) revealed one euploid embryo with findings consistent with intrauterine demise, but no other congenital anomalies noted (Fig.1A) and another euploid embryo with renal pelvis dilation. Six trisomic embryos had liver necrosis (Fig. 1B) and renal pelvis dilation was noted in five trisomic embryos. One embryo was noted to have an aortic outflow abnormality. There were no differences in brain or placental histology.

Among the Ts1Cje mice (12 euploid/12 trisomic), congenital anomalies were also only seen in trisomic embryos. One trisomic embryo had liver necrosis and two trisomic embryos had ventricular septal defects (Fig. 3C and D). Placental histology was unremarkable.

When comparing Dp(16)1/Yey and Ts1Cje placentas, there were many similarities among both euploid and trisomic groups. The Ts65Dn placentas had more differences compared to the other strains for both euploid and trisomic placentas.

Discussion

Principal findings of the study:

In this study, we examined embryonic and placental phenotypes in three commonly used mouse models of DS. This early point in the lifespan is underemphasized in most animal model studies of DS. We demonstrated that several of the prenatal phenotypic features seen in humans with DS are also observed in all three mouse models. The specific findings, however, differed between strains. The different genetic backgrounds among strains likely influences the timing and severity of the prenatal and postnatal development of fetal growth abnormalities.

Fetal Growth

Biometric assessment of the brain and placenta showed that the Ts65Dn model has many features consistent with the fetal DS phenotype seen in humans including growth restriction, atypical brain development, abnormal placental development and a 60% rate of congenital anomalies. This model showed significantly impaired fetal and brain growth even when controlling for factors such as litter size and sex. Additionally, the highest percentage of abnormal placental histology was seen in this model.

For the Dp(16)1/Yey strain, after controlling for litter size and sex, genotype did not have a significant impact on fetal growth. However, this model exhibited lighter brain weights when compared to euploid littermates. Differences in placental weight were also found to be driven by genotype, but no impact was seen on placental histology. This model did, however, show the highest rate of congenital anomalies (75%).

The Ts1Cje model showed fewer similarities to the human prenatal phenotype. The trisomic embryos did not show prenatal growth restriction or abnormal placental development. This model also had the lowest percentage of congenital anomalies at 40%.

Fetal growth restriction is a heterogeneous condition caused by maternal, fetal and/or placental factors. Normal fetal growth is dependent on normal placental function, placental morphometry (size and shape) and structure. In humans, 2–D ultrasound examinations measuring placental diameter and thickness, have been used as an indicator of high-risk pregnancies and correlates with birth weight.^19,20 In many species, fetal body weight late in gestation correlates positively with placental weight.^21,22 The findings of differences in fetal and placental growth in our mouse models may indicate a risk for growth restriction and adverse outcomes, as it has been found that in humans with DS there is a higher rate of fetal growth restriction and histopathologic findings of placental insufficiency are more common in cases of non-reassuring fetal surveillance and stillbirths.³

Congenital Anomalies

In all three strains the presence of liver abnormalities was a novel finding (Fig. 1). The underlying mechanisms are presently unknown, but liver necrosis may be secondary to ischemia (placental insufficiency), infarction (thrombosis secondary to vascular malformation or hematologic abnormalities) or infection.^23,24 Given that this study was performed in a controlled environment and no liver abnormalities were seen in euploid littermates, infection seems less likely. Liver fibrosis has been seen in mouse models of intrauterine growth restriction (IUGR) secondary to uterine artery ligation, which may support a placental insufficiency etiology. This finding supports the theory that an environment with limited nutrient supply may be divert nutrients to favor survival of vital organs (such as brain) at the expense of other organs such as the liver or pancreas.²⁵

In humans with DS, liver abnormalities are not commonly recognized. However, there are reports of transient abnormal myelopoiesis (TAM) leading to liver abnormalities and perinatal death.^26–33 Antenatally this presents as hydrops fetalis, with or without cardiac defects, and hepatosplenomegaly. In these cases, fetal blood sampling revealed findings suggestive of a myeloproliferative disorder and placental pathology consistent with myeloproliferative infiltration in the chorionic villi.^28–33 TAM occurs in 10–20% of infants with DS.²⁷ Histologic examination reveals cholestasis, fibrosis, hepatocellular necrosis, and occasionally, pancreatic involvement.²⁶ In up to 15% of cases potentially fatal liver disease develops.²⁷ Patients with TAM may also be predisposed to a hypercoagulative state and thrombosis in fetal vessels.³² Extensive hepatic necrosis in utero is rare, but case reports show that it is possible to survive in utero with hepatic function that is incompatible with extrauterine survival.²⁹

Renal pelvis dilation was frequent in both the Ts65Dn and Dp(16)1/Yey models (50 and 42%, respectively) (Fig. 2). Dilation of the renal collecting system can range from mild to severe hydronephrosis. In humans this is seen as an isolated ultrasound anomaly in 0.6–4.5% of all pregnancies and often resolves prior to delivery. ^34,35 While commonly a transient finding, it can also represent renal pathology and is a marker for fetal DS. Renal anomalies are seen in 2–21% of fetuses and children with DS. Fetal pyelectasis in the presence of other congenital anomalies has a positive likelihood ratio for DS ranging from of 1.5–17.44. ^34,36–39 In the current study 2/17 euploid embryos (11.7%) had renal pelvis dilation, which is a known finding in C57Bl6 mice.⁴⁰ Although the physiologic basis for these findings is presently unclear, the increased percentage of renal anomalies seen in the trisomic embryos may indicate that genes that are triplicated due to DS are important for kidney development.

Heart defects are the most common congenital anomaly in people with DS. They are detected in 40–60% of individuals.^37,41–43 Our results showed that cardiac defects (Fig. 3) were the third most common finding in mice; the rate observed was similar to prior studies.^16,42–47 In a previous report in the Ts1Cje model, 21% of embryos at E15.5 had VSDs and no cardiac defects were seen in euploid animals.¹⁶ Additionally, pre and postnatal studies in the Ts65Dn model have demonstrated the presence of aortic arch and septal defects.^47,48 For the Dp(16)1/YeY mice the results in the literature are mixed, with some studies showing multiple cardiac malformations, including VSD, ASD, coarctation, DORV, TOF and mitral valve abnormalities. Other studies have shown similar rates in euploid animals.⁴⁶ Given that there is incomplete penetrance of cardiac defects in humans with DS, further research into the specific mechanisms and environmental modifiers underlying cardiac development is needed.

Placental Development

Placental studies in humans with DS have shown increased cytotrophoblast apoptosis, abnormal syncytiotrophoblast fusion, vascular abnormalities, inflammation, villous hypoplasia, and intervillous fibrin deposition.¹⁵ In the current study, histological analysis did not reveal any significant findings in the Dp(16)1/Yey or Ts1Cje placentas. Increased spongiotrophoblast necrosis and decreased giant cells was seen in the Ts65Dn placentas (Fig. 4). Spongiotrophoblast cells comprise the junctional zone between the outer trophoblast giant cells and the inner labyrinth layer. The function of the spongiotrophoblast layer is poorly understood but supports the development of the labyrinth layer.⁴⁹ These abnormal findings in Ts65Dn placentas may be indicative of abnormal cytotrophoblast function and increased apoptosis, which may subsequently impact fetal growth and survival.¹⁵ It should also be noted that labyrinth fibrosis and calcification was frequently observed in both trisomic and euploid embryos in the Ts65Dn and Ts1Cje mouse models, but less frequently in the Ts65Dn model. In both mouse and human placentas calcium deposition has been found to be a marker of placental ageing, but it may also be associated with adverse outcomes when seen in early gestation. The lack of this finding in the Ts65Dn model may indicated a temporal difference in placental maturation.^15,50,51

Sex Differences

Although studies specifically examining sex differences in DS are limited, it has been shown that there are sex differences in the occurrence of CHD.^52,53 Heart disease is more common in females, with the exception of TOF, which is a more severe cardiac lesion.^52,53 One possible explanation for this could be a higher survival rate for females versus males.⁵² There are currently limited data on the impact of fetal sex on outcomes in DS, however, this is an important consideration when attempting to predict fetal outcomes. Future study designs should therefore take fetal sex into consideration.

Clinical Implications:

A central issue is whether the abnormalities seen in human placentas in DS have downstream consequences for fetal/infant survival, growth restriction, and brain or another organ development, and whether those relationships can be exploited to predict prognosis and develop treatments. Although each mouse model demonstrates a different aspect of the DS phenotype, taken together, our data provide insights into prenatal phenotypes not previously described and clues regarding the relationship between fetal and placental growth and organ development.

Research Implications:

Although day E18.5 is at the end of gestation in the mouse, it reflects a developmental time point that is consistent with the end of the first trimester in humans, a key opportunity for diagnosis and intervention. DS affects many cellular processes and signaling pathways, resulting in oxidative stress, and abnormalities in ion transport, cell stress response and the cell cycle. These abnormalities result from gene dosage imbalance as well as globally dysregulated gene expression.¹⁵ These processes are also dysregulated in placentas in humans with DS and include differences in cytotrophoblast fusion that affect subsequent conversion to syncytiotrophoblasts, atypical oxidative stress/anti-oxidant balance, atypical mineralization and increased expression of genes associated with premature senescence.¹⁵ Future research needs to focus on the underlying mechanisms for the observed differences between strains, including gene expression profiling in the tissues of interest, as well as the roles of oxidative stress and inflammation to better correlate these findings with human data⁵⁴.

Strengths and limitations:

This is a novel study that sheds light on embryonic and placental development in three mouse models of DS. The strengths of this study include a large sample size with multiple controls (including sex effects) within each litter. We also used methods such as evaluation of prenatal growth and histological analysis that can be eventually extrapolated to human data.

There are, however, limitations to using mouse models, given that no single mouse model recapitulates the full phenotype observed in people with DS.⁶ Also, most models have similar genetic backgrounds (C57/Bl/6J and F1) that do not reflect human diversity and may reduce complex genetic interactions that influence the expression of features of DS.⁷ Traditional light microscopy along with hematoxylin and eosin staining may not be sensitive enough to detect all congenital anomalies. Additionally, we did not record the presence of blood in the umbilical cord and therefore we could not accurately evaluate for possible in-utero demise.

Another complication that arose during this study was the presence of resorbed and abnormal embryos that were unable to accurately genotyped and were therefore excluded from analysis (Table 1). Resorption occurs due to developmental failure, however no relationship between trisomy and embryonic degeneration could be tested.

Consideration should also be given to the fact that trisomic fathers were used only for the Ts1Cje and Dp(16)1/Yey matings. The study design was intended to mimic transmission in humans in which the mother is euploid. Due to male sterility in Ts65Dn mice, the trisomy was propagated through the maternal line in which the mother was trisomic. A poor reproductive history has been documented in female Ts65Dn mice.^48,55,56 Although prior studies have shown that embryonic trisomy and not the trisomic maternal uterine environment has a larger impact on development, we cannot exclude the possibility that parent-of-origin effects influenced the phenotype.⁵⁵ The role of parent-of origin effects can be studied in Dp(16)1/Yey and Ts1Cje matings because in both of these strains the males and females are fertile.

Conclusions:

Animal models are crucial to our understanding of the pathogenesis of complex syndromes. Our findings shows that the Ts65Dn model best recapitulates the human prenatal DS phenotype, but it is not without limitations because of the significant number of triplicated genes that are not orthologous to Hsa21. A need still exists for a new mouse model that better mimics both the human genotype and phenotype in DS. Importantly, this study has identified fetal abnormalities that may serve as potential endpoints for the evaluation of therapeutic responses to in utero treatment⁵⁴.

Supplementary Material

1

S1–3. Individual embryo and placental biometric data

S4. Primer sequences for genotyping

S5–7. Individual embryo and placental histological data

Condensation:

Different mouse models of Down syndrome (DS) demonstrate variable phenotypes. Our data provide novel insights into embryonic and placental pathology in commonly studied strains.

AJOG at a Glance:

A. Why was this study conducted? Previously, we observed a significant correlation between growth and behavioral delays in the Ts1Cje mouse model of DS. It is known that human fetuses with DS are especially susceptible to adverse obstetric outcomes, but there is a lack of existing information at the early end of the lifespan in equivalent mouse models.

B. What are the key findings? For Ts65Dn and Dp16 strains, embryo weights were smaller for trisomic embryos compared to their euploid littermates. Congenital anomalies were more common in trisomic embryos compared to their euploid littermates, including cardiac defects and obstructive renal anomalies, which are seen frequently in humans with DS. There was also the novel finding of hepatic abnormalities observed only in trisomic embryos. Differences in placental histology were observed between strains and this has not been previously reported.

C. What does the study add to what is already known? Although each mouse model demonstrates a different aspect of the DS phenotype, taken together, our data provide insights into prenatal phenotypes not previously described and offers clues regarding the relationship between fetal and placental growth and organ development.