Embryonic and not maternal trisomy causes developmental attenuation in the Ts65Dn mouse model for Down syndrome


  • Joshua D. Blazek,

    1. Department of Biology, Indiana University-Purdue University Indianapolis and Indiana University Center for Regenerative Biology and Medicine, Indianapolis, Indiana
    Search for more papers by this author
  • Cherie N. Billingsley,

    1. Department of Biology, Indiana University-Purdue University Indianapolis and Indiana University Center for Regenerative Biology and Medicine, Indianapolis, Indiana
    Search for more papers by this author
  • Abby Newbauer,

    1. Department of Biology, Indiana University-Purdue University Indianapolis and Indiana University Center for Regenerative Biology and Medicine, Indianapolis, Indiana
    Search for more papers by this author
  • Randall J. Roper

    Corresponding author
    1. Department of Biology, Indiana University-Purdue University Indianapolis and Indiana University Center for Regenerative Biology and Medicine, Indianapolis, Indiana
    • Department of Biology, Indiana University-Purdue University Indianapolis, 723 W. Michigan Street, SL 306, Indianapolis, IN 46202
    Search for more papers by this author


Trisomy 21 results in Down syndrome (DS) and causes phenotypes that may result from alterations of developmental processes. The Ts65Dn mouse is the most widely used genetic and phenotypic model for DS. We used over 1,500 offspring from Ts65Dn and two nontrisomic genetically similar control strains to investigate the influence of trisomy on developmental alterations and number of offspring. For the first time, we demonstrate gross developmental attenuation of Ts65Dn trisomic offspring at embryonic day (E) 9.5 and E13.5 and show that the major determinant of the developmental changes is segmental trisomy of the embryo and not the trisomic maternal uterine environment. Maternal alleles of nontrisomic genes linked to Pde6b may also influence the development of Ts65Dn offspring. Both developmental attenuation and the contribution of trisomic and nontrisomic genes are important components in the genesis of DS phenotypes. Developmental Dynamics 239:1645–1653, 2010. © 2010 Wiley-Liss, Inc.


Down syndrome (DS) is the outcome of trisomy for human chromosome 21 (Hsa 21) and causes a constellation of pre- and postnatal abnormalities. The prevalence of DS is 1 in 700–800 live births (Centers for Disease Control and Prevention,2006; Christianson et al.,2006; Schieve et al.,2009). Fetuses with trisomy 21 also exhibit a high rate of spontaneous termination in the first and second trimesters (Spencer,2001). Additionally, neonatal and infant mortality are, respectively, five and eight times higher in individuals with DS compared with the general population (Weijerman et al.,2008). Individuals born with trisomy 21 exhibit subsets of nearly 80 clinically defined phenotypes with varying severity, including malformations of the heart, craniofacial region, as well as central nervous and gastrointestinal systems (Epstein,2001; Van Cleve et al.,2006; Van Cleve and Cohen,2006; Cleves et al.,2007). These congenital defects likely reflect changes during pre- and postnatal developmental processes (Pinter et al.,2001).

Trisomy 21 causes alterations in functional as well as physical milestones that are apparent during development. Postnatal functional landmarks including language, motor, and sensory milestones are often delayed (Epstein,2001; Nadel,2003; Chen and Fang,2005; Vicari,2006), and physical features including brain size, birth weight, stature and head circumference may be reduced in individuals with DS (Leonard et al.,2000; Smith,2006). Prenatal measurements in fetuses with Trisomy 21 have shown microcephaly, craniofacial abnormalities, and restricted neuroanatomy (Pinter et al.,2001; Vicari,2006). Prenatal screening often identifies cardiac and gastrointestinal anomalies, shortened humerus and femur bones, and hypoplasic or absent nasal bones (Guihard-Costa et al.,2006; Driscoll and Gross,2009). Concomitantly, the studies indicate that prenatal changes affect the development of physical structures and may predispose individuals with DS to functional abnormalities.

Mouse models have been developed to understand the correlation of genetic and phenotypic changes observed as a result of aneuploidy (Patterson,2009; Wiseman et al.,2009). Additionally, they offer the potential to examine embryos and tissues at any developmental stage. The Ts(1716)65Dn mouse model (Ts65Dn) contains a small translocation chromosome resulting in triplication of approximately half of the gene orthologs found on human chromosome 21 (Reeves et al.,1995; Hattori et al.,2001; Gardiner et al.,2003). Ts65Dn mice exhibit several phenotypes observed in humans with DS, including functional and physical abnormalities, cognitive impairment, craniofacial alterations, cardiovascular anomalies, and neurologic structural deformities (Holtzman et al.,1996; Moore and Roper,2007; Wiseman et al.,2009). These parallel phenotypes allow for the study of the cellular and molecular mechanisms affected by the presence of additional chromosomal material as well as changes in gene expression (Patterson,2009). Many gene expression studies have shown expression differences in both trisomic and euploid genes in the genome of humans and DS mouse models, and their interaction has been hypothesized to affect numerous phenotypes (Roper and Reeves,2006; Patterson,2009).

Accumulating evidence supports prenatal attenuation in developmental processes in humans with DS and DS mouse models. Alterations in functional and physical features in mice have been linked to developmental abnormalities in cardiac, neurological, and craniofacial precursors. Prenatally, the first pharyngeal arch is significantly smaller in Ts65Dn embryos at embryonic day (E) 9.5 compared with euploid littermates at the same developmental stage (Roper et al.,2009). Ts65Dn mice also exhibit altered prenatal growth of the cerebral cortex and hippocampus due to abnormal proliferation of embryonic neural precursor cells (Chakrabarti et al.,2007), and it is postulated that the foramen ovale exhibits divergent development in the heart (Williams et al.,2008). Postnatal dysmorphology in Ts65Dn mice include altered development of the cerebellum (Baxter et al.,2000), hippocampus (Contestabile et al.,2007), and basal forebrain cholinergic neurons (Holtzman et al.,1996). Accurate gross and tissue-specific measurements are necessary to find cellular and molecular bases for trisomic phenotypes because dysregulation in development of trisomic as compared to euploid mice and embryos may overshadow or preclude finding subtle tissue or cellular-specific alterations related to DS phenotypes.

In this study, we examined gross embryonic developmental attenuation in trisomic Ts65Dn and genetically similar control mice. Although Ts65Dn is the most widely used model for DS, little is known about how the maternal trisomy or the uterine environment and genome affect development of both trisomic and euploid embryos. We used 500 E9.5 embryos from 86 litters, 278 E13.5 embryos from 41 litters, and 794 postnatal mice from 173 litters to investigate the influence of trisomic and nontrisomic genes on gross developmental alteration and number of offspring. The characterization and influence of genetic and developmental changes in mouse models provides a foundation that is important when examining the genetics determining the specific tissue and cellular precursors linked to DS-like phenotypes.


Transmission Frequency of Marker Chromosome Slightly Altered During Prenatal Development

It has been previously reported that transmission frequency of the small T(1716)65Dn marker chromosome is slightly reduced at birth (44–48%) and is significantly different from Mendelian expectations at 6 days of age (∼40%) and weaning (36%; Moore,2006; Roper et al.,2006). We have observed a similar reduction from the expected Mendelian ratio of trisomic offspring from Ts65Dn mothers in our colony (33.6%; n = 794; P < 0.001). Additionally, it has been shown that Ts65Dn mice weigh less than euploid littermates at birth until at least 5 months of age (Roper et al.,2006). Beginning at birth, this difference in weight becomes more apparent within litters as the Ts65Dn offspring grow older. These data suggest that trisomy adversely affects both survival and development of Ts65Dn offspring. Although it has been postulated that poor mothering by Ts65Dn dams contributes to the loss of both trisomic and euploid offspring (Moore,2006; Roper et al.,2006), information about the prenatal influence of trisomy on development and survival is limited. Our data show a small yet significant deviation from Mendelian ratios in trisomic frequency during development at E9.5 (44.4% trisomic, (χ2 = 4.00; P = 0.05; n = 324) and a similar (but not significant) frequency at E13.5 (44.6% trisomic [χ2 = 1.29; P = 0.26], n = 112).

Ts65Dn Trisomic Embryos Exhibit Fewer Average Somites at E9.5

Ts65Dn mice cannot be inbred, and the marker chromosome is maintained on an advanced intercross genomic background composed of (on average) 50% alleles each from C57BL/6J (B6) and C3H/HeJ (C3H) inbred mice. Ts65Dn females are almost exclusively used to generate trisomic offspring. To understand the influence of trisomy (mother and/or offspring) as well as genomic background on development in trisomic mice, we have examined offspring from Ts65Dn, B6C3Fn euploid littermate controls (hereafter referred to as “euploid controls”) and B6C3F1 mothers. We quantified somite numbers in embryos removed from Ts65Dn and both types of control mothers during a 4-hr time period at midgestation (gestational age in mice is ∼19–21 days). Somite numbers were used to quantify developmental stage, as these have previously been shown to typify Theiler stage development (Kaufman and Bard,1999; Fan et al.,2003; Roper et al.,2009). Average somite numbers of E9.5 (midgestation or Theiler stage 15) embryos from Ts65Dn, euploid control, and B6C3F1 mothers were significantly different from each other (F(3,496) = 35.07; P < 0.0001). We then separated Ts65Dn trisomic and euploid embryos in the analysis. From Ts65Dn mothers, E9.5 trisomic embryos had an average of 19.12 somites (SEM = ± 0.29; n = 117) and had significantly fewer average somites than euploid littermates (20.41 ± 0.24 somites; n = 158; Fig. 1). E9.5 embryos from euploid control mothers averaged 20.67 somites (± 0.31; n = 120) and were not significantly different from euploid embryos from Ts65Dn mothers. Embryos from B6C3F1 mothers had more somites than all other embryos in the study (23.10 ± 0.18 somites, n = 105; Fig. 1). Although genetic background caused differences in average somite number between different strains, our data suggest the trisomic portion of the Ts65Dn embryo's genome was a significant determinant of fewer average somites in embryos from mothers with a similar genetic background.

Figure 1.

Developmental attenuation in trisomic embryonic day (E) 9.5 Ts65Dn embryos. Average somite numbers in E9.5 embryos from Ts65Dn and control mothers. A: Euploid (Ts65Dn) and trisomic (Ts65Dn) indicates these embryos are offspring of Ts65Dn mothers. (*) Trisomic embryos exhibit significantly fewer somites compared with their euploid littermates and control embryos. (**) There is no significant difference between Ts65Dn euploid somite number and euploid controls; however, both have significantly fewer somites than B6C3F1 control embryos (***). Stars represent significantly different groupings from analysis of variance and post hoc, least significant difference multiple comparison analysis (F(3,496) = 30.57; P < 0.0001; α = 0.05). Error bars are calculated as standard error of the mean. B–E: Representative E9.5 offspring from a B6C3F1 mother (B), a euploid control mother (C), and a Ts65Dn mother (D,E). D is a euploid embryo, and E is a trisomic embryo from a Ts65Dn mother. Scale bar = 500 μm.

Ts65Dn Embryos Exhibit a Reduction in Size at E13.5

Developmental size and relative growth of E13.5 embryos has been commonly determined by crown–rump length (CRL; Brown et al.,2006; Mu et al.,2008), and the average CRL at this time point is suggested to be between 9.5 and 11.5 mm (Kauffman,1992). Volumetric analysis of the embryo to determine an accurate measurement of developmental size and/or relative growth is possible but usually requires embryo processing, exhaustive sectioning, and is difficult and expensive to do. Additionally, alteration or removal of an embryonic structure often prohibits accurate volumetric measurements because the embryo must be kept intact (before sectioning) for the volumetric studies. As CRL is a one-dimensional linear measurement, we hypothesized that the area occupied by the embryo (calculated from an image captured as embryos are dissected from the mother) would be a better determinant of changes in embryonic size and could be used to detect differences in trisomic growth.

To examine methods of determining the size of E13.5 embryos, we established embryonic volume in a subset of 49 E13.5 embryos (34 nontrisomic and 15 trisomic with a similar average genetic background) and compared it with both CRL and area measurements. In this subset of E13.5 embryos, the average CRL of all classes of embryos fell within the 9.5- to 11.5 mm range specified for this developmental stage. With the exception of a trisomic embryo, these embryos displayed handplate digits and a straightened tail (two morphometric markers of E13.5 embryos; Kauffman,1992), although these qualitative parameters were difficult to discern in captured images and did not generally differentiate trisomic from euploid embryos. Volumes were calculated using exhaustively sectioned embryos through the use of unbiased stereology (see the Experimental Procedures section). We found embryo volume had a higher correlation to embryo area (r = 0.43; P = 0.0023) than CRL (r = 0.26; P = 0.0770), implicating embryo area as a more accurate measurement for size than CRL.

Practically, an estimation of size of an E13.5 embryo can be easily accomplished by quantifying the area of the embryo on a single image taken before other procedures are done. Thus, we used embryo area to determine the size of E13.5 embryos from Ts65Dn, euploid control, and B6C3F1 mothers. At E13.5 (Theiler stage 22), significant differences were seen between the size of all of the offspring classes as estimated by embryonic area (F(3,153) = 45.40; P < 0.0001). Trisomic embryos (49.26 ± 0.96 mm2; n = 14) from Ts65Dn mothers are significantly smaller than their euploid littermates (52.00 ± 0.59 mm2; n = 20). As with E9.5 embryos, the euploid E13.5 embryos were not significantly different than the embryos from the euploid control mothers (53.47 ± 0.83 mm2; n = 34). Embryos from B6C3F1 mothers (59.08 ± 0.38 mm2; n = 89), however, were significantly larger than all other embryos (Fig. 2).

Figure 2.

Developmental size alterations at embryonic day (E) 13.5 in Ts65Dn trisomic mice. A–E: Average area of E13.5 embryos a from B6C3F1 mother (A), a euploid control mother (B), and a Ts65Dn mother (C); euploid embryo (D) and trisomic embryo (E). D,E: Euploid (Ts65Dn) and trisomic (Ts65Dn) indicates these embryos are offspring of Ts65Dn mothers. Average area of E13.5 trisomic embryos from Ts65Dn mothers is significantly less than euploid littermates and Ts65Dn background control and B6C3F1 mothers. There is no statistical difference in the area of euploid embryos from Ts65Dn trisomic mothers or background control mothers. The E13.5 embryos from B6C3F1 mothers are larger than all other embryos. Stars represent significantly different groupings from analysis of variance and post hoc, least significant difference multiple comparison analysis (F(3,153) = 45.40; P < 0.0001; α = 0.05). Error bars are calculated as standard error of the mean. Scale bar = 2500 μm.

From our data at E9.5 and E13.5, we conclude that, although genetic background may contribute to developmental variability, embryonic trisomy is the largest influence on Ts65Dn size and, through comparison of animals with a similar genetic background, growth differences during embryogenesis. Additionally, because euploid E9.5 and E13.5 embryos from euploid control and Ts65Dn mothers did not show significant differences in embryonic size, the trisomic uterine environment does not appear to influence embryonic size or growth during development. Our data also suggest that the developmental attenuation of Ts65Dn neonatal mice begins early in gestation and that size differences are apparent throughout gestation.

Small Number of Embryos From Trisomic Mothers Influenced by Genetic Differences in Ts65Dn Mothers

Ts65Dn mothers have small litters with less than expected Mendelian ratios of trisomic offspring and the ratio of trisomic offspring becomes smaller and more significant as the offspring develop from birth to weaning (Roper et al.,2006). At E9.5, there was a difference in the number of embryos observed from the mothers of the three different genetic backgrounds (F(2,90) = 11.41; P < 0.0001). An average of 9.05 (± 0.22) embryos was observed in B6C3F1 mothers (n = 21 litters) and this was significantly greater than the average number of embryos from euploid control mothers (7.84 embryos ± 0.56; n = 19) or Ts65Dn mothers (6.83 embryos ± 0.29; n = 53). The average number of embryos per litter from Ts65Dn mothers was significantly less than euploid control mothers. Differences were also shown among litter sizes at E13.5 (F(2,32) = 6.20; P = 0.0053). Litters from Ts65Dn mothers at E13.5 (5.43 embryos ± 0.72; n = 14) were significantly smaller than those from B6C3F1 (8.61 embryos ± 0.54; n = 13) but not euploid control mothers (6.63 embryos ± 0.76; n = 8). Embryonic litter size was significantly reduced in Ts65Dn mothers as compared to B6C3F1 controls at both E9.5 and E13.5. Additionally, there does not appear to be subsequent reduction in Ts65Dn litter size after E13.5 as previous data have shown that the average litter size at birth is 5.4 pups from Ts65Dn mothers (Roper et al.,2006). Although the number of litters in our sample was small (especially at E13.5), our data suggest a reduction in Ts65Dn postnatal litter size originates during embryonic development. Taken together, these results suggest that the reduction in number of embryos may be due to a loss of trisomic embryos and influenced by trisomy and nontrisomic genes from the Ts65Dn mother.

Homozygosity for Pde6brd in Ts65Dn Mothers May Contribute to Reduced Number of Postnatal Mice

It has been hypothesized that nontrisomic background genes of the mother play a role in the survival of Ts65Dn pups, including alleles at the H2 locus (Paz-Miguel et al.,2001; Roper et al.,2006). The additional small marker chromosome is carried in Ts65Dn mice on an advanced intercross genome with alleles from B6 and C3H strains. A gene related to retinal degeneration, Pde6b is found on mouse chromosome 5. The C3H allele of this gene carries the rodless (rd) mutation caused by a murine retroviral insertion and a second nonsense mutation in exon 7 (Pde6brd; Bowes et al.,1993). Mice homozygous for the rd mutation (rd/rd) are blind by weaning age and show deficient cognition, increased rates of activity, and reduced anxiety when compared with mice with normal vision (Cook et al.,2001; Hoelter et al.,2008). Due to their advanced intercross genetic background, Ts65Dn mice may carry 0, 1, or 2 copies of the Pde6brd allele in their genome and are routinely screened to ensure that blindness does not influence the testing of some phenotypes. It has also been suggested that homozygosity for the Pde6brd gene has a negative impact on Ts65Dn colony maintenance and a new Ts65Dn model without the detrimental allele has been created (Costa et al.,2010). We hypothesized that one or more copies of this mutation or a closely linked gene would contribute to the poor mothering observed in Ts65Dn females (Moore,2006; Roper et al.,2006) and may be responsible for a reduction in trisomic Ts65Dn offspring at weaning.

To test our hypothesis, we determined the Pde6brd alleles of 54 Ts65Dn mothers that produced 173 total litters in our mouse colony. When comparing the three maternal Pde6b genotypes together in a single analysis, there was not a significant difference in the total number of pups per litter at postnatal day (P) 10 based on maternal rd genotype (F(2,162) = 2.08; P = 0.13; Table 1). Furthermore, no difference was found in the number of trisomic pups or the percent trisomic pups per litter at P10 (data not shown). When comparing only Ts65Dn mothers homozygous for the mutation (Pde6brd/rd) with mothers homozygous for the wild-type allele (Pde6b+/+), we found Pde6brd/rd Ts65Dn mothers have significantly fewer total pups per litter surviving at P10 (P = 0.015, see also Table 1). When comparing offspring from mothers affected by blindness (two mutant alleles) to those that are not (combining offspring data from Ts65Dn mothers with one or two wild-type Pde6b alleles), Ts65Dn Pde6brd/rd mothers had significantly fewer offspring at P10 than phenotypic nonblind Ts65Dn mothers (P = 0.045).

Table 1. Litter Size From Ts65Dn Pde6b+/+, Pde6b+/rd, and Pde6brd/rd Mothers at Embryonic Day (E) 9.5 and Postnatal Day (P) 10
Maternal rd genotypeP10 litters (n)Total pups per litteraE9.5 litters (n)Total embryos per litterb
  • a

    Analysis of variance (F(2,170) = 2.08, P = 0.13).

  • b

    Analysis of variance (F(2,30) = 1.27, P = 0.29).

  • c

    Pde6brd/rd Ts65Dn mothers have significantly fewer pups at P10 compared to Pde6b+/+ Ts65Dn mothers (P = 0.015).

  • d

    Pde6brd/rd Ts65Dn mothers (blind) have significantly fewer pups at P10 than Pde6b+/+ and Pde6b+/rd Ts65Dn mothers (not blind) combined (P = 0.045).


To determine if the rd mutation of the mother affected the number of embryos at midgestation, we analyzed E9.5 embryos from Pde6b+/+, Pde6b+/rd, and Pde6brd/rd Ts65Dn mothers. We found no significant difference in average litter size between the different mothers (F(2,30) = 1.27; P = 0.29, Table 1), or in the number of trisomic embryos or percent trisomic embryos per litter (data not shown). Although the numbers for embryos and offspring in each separate classification were small (especially from Ts65Dn Pde6brd/rd mothers), our data suggest that homozygosity for Pde6brd (or linked) alleles in Ts65Dn mothers have a small effect on the number of pups at P10.

Interaction of Retinal Degeneration With Trisomic Genes in Developmental Attenuation

Our analyses indicate that embryonic trisomy is the most important influence on the developmental attenuation observed in Ts65Dn embryos. Because homozygosity for Pde6brd may slightly affect the number of Ts65Dn offspring, we wondered if the rd genotype of the mother added to the reduction in size of trisomic embryos. To test this premise, we examined the contributions of ploidy and rd genotype of mothers and ploidy of E9.5 embryos on the developmental stage at midgestation. Multiple linear regression revealed that the ploidy (trisomy or euploid) of the embryo as well as ploidy and rd genotype of the mother were significant factors in determining somite number and developmental stage at E9.5 (F(4,208) = 9.51; P < 0.0001). When embryos were classified (binned) according to both mother's ploidy and rd genotype as well as embryo ploidy, the analyses also revealed a significant difference between embryos (F(4,208) = 4.881; P < 0.0001) (Table 2). Because Ts65Dn mothers have both trisomic and euploid offspring, the significance of the mother's trisomy in our regression analysis appears to be due to the inclusion of the trisomic embryos. In a post hoc, multiple comparisons analysis, trisomic embryos from Pde6brd/rd Ts65Dn mothers had significantly fewer somites than trisomic embryos from Ts65Dn Pde6b+/+ but not Pde6b+/rd mothers. Although not significant, a trend of fewer somites was seen in embryos from either euploid control or Ts65Dn mothers as increasing numbers of rd alleles were found in the maternal genotype. (The effect of the fetal Pde6b genotype was not accounted for and unknown in our study.) Our results indicate that maternal homozygosity for Pde6brd or a linked gene may play a role in embryonic development.

Table 2. Somite Number of Embryonic Day (E) 9.5 Offspring From Ts65Dn and Euploid Control Pde6b+/+, Pde6b+/rd, and Pde6brd/rd Mothers
Maternal rd genotypeMaternal ploidyEmbryo ploidyNAverage somite Number*
  • *

    Binned genotypes are statistically significantly different (F(4,208) = 4.881, P <0.0001). Numbers sharing at least one letter (a–d) in common are not statistically different by least significant difference multiple comparisons on binned genotypes.



The Ts65Dn mouse is the most widely used mouse model in the study of DS. It is a powerful model because these mice include approximately half of the Hsa21 genes in three copies, exhibit several DS-like phenotypes, and can be used to predict previously unknown phenotypes in individuals with DS (Baxter et al.,2000). Discerning the developmental origins of structural dysmorphologies in mice and humans may clarify corresponding functional dysregulation. Conceivably, these developmental alterations could occur because of gross or tissue-specific morphological changes during prenatal or postnatal development or a combination of these processes. The majority of studies on dysregulation in DS mouse models have focused on postnatal time periods, due to easier access to tissue samples, and both structural and functional alterations have been defined in adult mice. Less is known about changes occurring during embryonic development in Ts65Dn mice, with much of the prenatal evidence limited to tissue specific attenuation in the first pharyngeal arch and brain (Chakrabarti et al.,2007; Roper et al.,2009). Our results demonstrate significant developmental attenuation in trisomic embryos as early as midgestation in the Ts65Dn model of DS. Because Ts65Dn males are usually sterile and offspring from Ts65Dn trisomic mothers are used to study phenotypes associated with DS, the effect of the maternal trisomic uterine environment on both trisomic and euploid embryos had been questioned (Arron et al.,2006; Moore,2006).

By comparing embryonic offspring from three genetically similar mothers, we have determined the developmental attenuation seen in trisomic offspring of Ts65Dn mothers is largely due to the segmental trisomy of the mouse embryos. Similar average developmental stage and size in E9.5 and E13.5 euploid embryos, respectively, from Ts65Dn and euploid control mothers suggest the trisomic uterine environment does not alter gross development. These data support the use of embryos from Ts65Dn mothers as a model for DS in humans. Embryos from a single litter will vary in developmental stage, and overlap exists in embryonic developmental stages between litters. Developmental differences between trisomic and euploid embryos from a single Ts65Dn mother demonstrate the need to account for growth and size differences, as gestational age is not indicative of developmental stage or size when comparing trisomic and euploid embryos. Additionally, we conclude that at E13.5, embryonic area is a more accurate and a relatively easy way to estimate embryonic size compared with crown–rump length.

This and previous reports suggest a reduction in the number of mice born to Ts65Dn mothers (Moore,2006; Roper et al.,2006). We hypothesized that this may be due to variation in background genes in the mothers. In addition to the influence of H2 alleles investigated previously (Paz-Miguel et al.,2001; Roper et al.,2006), our data suggest that homozygosity for Pde6brd, or closely linked alleles segregating in Ts65Dn mothers, may have an influence on the number of pups from Ts65Dn litters at P10 and slightly alter development at E9.5. Based on our results, we hypothesize that the blindness of Ts65Dn Pde6brd/rd mothers impacts the number of offspring born to these mothers that survive to P10. A recent study evaluating Ts65Dn mice that have been bred without the rd mutation suggests no significant differences on behavioral or size phenotypes at 4–5 months compared with Ts65Dn mice with a segregating rd mutation (Costa et al.,2010). Although Costa et al. did not focus on the effects of the maternal Pde6b genotype on prenatal development or postnatal numbers of Ts65Dn mice, our results generally agree that Ts65Dn mice without the Pde6brd mutation may allow for more efficient production of trisomic animals and reduce laboratory costs, as well as eliminate some variability in developmental phenotypes.

Variation in developmental stage within a single litter has been reported in both prenatal and postnatal Ts65Dn offspring (Roper et al.,2006,2009). To reduce some of the variability between litters, we placed females and males together each night at approximately the same time. We then focused on E9.5 and E13.5 embryos dissected within a 4-hr window between 8 AM and 12 PM. A similar upper limit in embryo stage was observed from Ts65Dn (range = 11–28 somites), euploid control (10–27 somites), and B6C3F1 (18–26 somites) mothers. Furthermore, we saw no evidence of a more advanced developmental stage due to a small litter size in any of our crosses. Our embryos had, on average, a 75% B6 25% C3H background as in a previous analysis (Roper et al.,2006) in contrast to the average 50% frequency of both B6 and C3H alleles usually found in postnatal Ts65Dn mice. B6 mice have been shown to have a larger average litter size than C3H mice (6.60 vs. 5.20, respectively; Mouse Phenome Database, www.jax.org/phenome) and our mating strategy with increased average B6 alleles in the offspring may have slightly affected the phenotypes observed in this study. Some variability in trisomic development may be influenced by allelic differences in trisomic and other nontrisomic genes in addition to H2 and Pde6b.

Our previous studies showed no differences in somite number between trisomic embryos and their euploid littermates at E9.5 (Roper et al.,2006), contrary to what we observed in the present study. We examined both the previous and current data and found these differences were likely due to previously harvesting embryos in a larger window of time because we were attempting to somite match embryos at two different developmental stages. Other reasons could be different ages of the Ts65Dn mothers at the time of embryo dissection (average age of this study was 3 months old, 5 months younger than the average age in the previous study), rd genotype of the mothers of the previous study (data are unknown), or different environments in which the mice were generated. Based on this and previous work, we conclude it is essential to match the developmental stage of trisomic and euploid embryos to make an accurate assessment of tissue-specific developmental changes (Roper et al.,2009).

Understanding prenatal developmental alterations caused by trisomy is critical to determining the presence and etiology of trisomic phenotypes. Recent studies in mouse models of DS suggest cellular differences are a result of changes in cell cycle regulation and proliferation. The effects of trisomy on cell proliferation have been documented in the hippocampus of human DS fetuses and Ts65Dn mice (Chakrabarti et al.,2007; Contestabile et al.,2007), fibroblasts and keratinocytes of postnatal Ts65Dn mice (Contestabile et al.,2009), and in the first pharyngeal arch (Roper et al.,2009) of E9.5 mice. Furthermore, each of these studies has shown deficiencies in cell number result from reduced rates of cell division because apoptosis levels remained consistent in all samples. It is possible that reduced cell proliferation leading to tissue-specific or gross developmental changes may cause functional alterations in Ts65Dn mice and individuals with DS. Interactions between trisomic genes and nontrisomic genes throughout the genome may be important in determining the phenotypic variability in other developmentally related abnormalities. The Ts65Dn model provides an important tool in determining the etiology of developmental phenotypes associated with DS.



Female B6EiC3Sn a/A-Ts(1716)65Dn (Ts65Dn) and female and male B6CBA-Tg(Wnt1-lacZ)206Amc/J (Wnt1-lacZ), B6.129S4-Gt(ROSA)26Sortm1Sor/J (B6.R26R) and C3H/HeJ (C3H) mice were purchased from the Jackson Laboratory (Bar Harbor, ME). Wnt1-Cre mice were obtained from Dr. Yang Chai of the University of Southern California and backcrossed more than six generations to C57BL/6J (B6) mice. Wnt1-lacZ mice were brother–sister mated and mice homozygous for the Wnt1-lacZ transgene were identified and maintained in our colonies on an approximate B6 background. B6(R26R)C3F1 mice were bred by crossing B6.R26R females with C3H males. Ts65Dn females used as mothers in this study were generated at Indiana University-Purdue University Indianapolis (IUPUI) by crossing Ts65Dn females with B6(R26R)C3F1 males and identified by fluorescent in situ hybridization (FISH) genotyping (Moore et al.,1999). Three types of mice were used as mothers in the study: Ts65Dn (approximate 50% B6 and 50% C3H background with small marker [trisomic] chromosome), B6C3Fn euploid littermates “euploid controls” (similar background as Ts65Dn trisomic mice, but without small marker chromosome), and B6(R26R)C3F1 “B6C3F1” (background: 50% B6 and 50% C3H). All animal use and protocols were approved by the IACUC committee at IUPUI.

Dissection and Processing of Embryos

Ts65Dn, Ts65Dn euploid background controls, and B6C3F1 females were bred to either Wnt1LacZ (E9.5) or Wnt1-Cre (E13.5) males on a B6 background, and mothers were checked for vaginal plugs the morning after mating with E0.5 defined as 12:00 pm on the day the female was plugged. All embryos had an approximately equivalent 75% B6 25% C3H background. At 9.5 or 13.5 days after the plug was identified, Ts65Dn mothers were euthanized within a 4-hr window of time and embryos were removed. Somite pairs were counted in each E9.5 embryo, and a picture of each 9.5 and E13.5 embryo was taken using a Nikon SMZ1500 microscope with a DS-L2 Digital Sight camera. The E13.5 embryos were washed, stained in X-gal, pre- and post-fixed in 4% paraformaldehyde, dehydrated, cleared with xylenes, and embedded in paraplast according to commonly used protocols. Embryos were sectioned at 22 μm and stained using 0.1% eosin. The expression of lacZ in the embryos was used for other experiments in our laboratory, and all embryos used in this study went through similar processing. For this study, we used 500 E9.5 embryos (105 B6C3F1, 120 B6C3Fn euploid controls, and 275 Ts65Dn embryos [158 euploid and 117 trisomic] from a total of 86 mothers [21 B6C3F1, 19 B6C3Fn euploid controls, and 53 Ts65Dn]) and 278 E13.5 embryos (112 B6C3F1, 54 B6C3Fn euploid controls, and 112 Ts65Dn embryos [62 euploid and 50 trisomic]) from a total of 41 mothers (13 B6C3F1, 8 B6C3Fn euploid controls, and 20 Ts65Dn).


Embryonic offspring generated from Ts65Dn crosses were genotyped by FISH on yolk sac cells of Ts65Dn E9.5 (8% of the embryos were resorbed and not genotyped by FISH and not included in the averages reported; Roper et al.,2009). The P10 offspring (n = 794) were genotyped by a polymerase chain reaction (PCR) prescreen (Lorenzi et al.,2010), and 743 were confirmed by FISH on blood cells taken from weaned mice (Moore et al.,1999). Fifty-one mice either died or were PCR negative and euthanized before FISH typing. Some Ts65Dn female mice (n = 54, with 173 total litters) were genotyped for the presence of the retinal degeneration gene rodless (Pde6brd or rd) mutation using forward (CATCCCACCTGAGCTCACAGAAAG) and reverse (GCCTACAACAGAGGAG) primers purchased from Invitrogen (Carlsbad, CA). These primers amplify exon 7 of Pde6b contacting the mutation site causing retinal degeneration. For the PCR reaction, 2 μl of DNA (∼100 ng/μl) was placed into 23 μl of reaction mixture consisting of sterile Millipore water, 1× KCl reaction buffer, 3 mM MgCl2, 0.25 mM dNTPs, 0.02 U/μl Taq polymerase, and 0.4 μM each of forward and reverse primers. The DNA was amplified by PCR as follows: 95°C for 3′, 94°C for 30″, 60°C for 30″, 72°C for 30″, repeat steps 2–4 10×, −1/2 degree per cycle, 94°C for 30″, 55°C for 40″, 72°C for 30″, repeat 5–7 25×, 72°C for 10′, and 4°C hold. PCR products were then digested using DdeI enzyme (NEB, Ipswich, MA) and run on a 3% gel. Homozygous wild-type (Pde6b+/+) mice have one band at 300; heterozygous (Pde6b+/rd) mice have bands at 298, 137, and 104; and homozygous mutants (Pde6brd/rd) have bands at 137 and 104 bps.

Quantification of E13.5 Embryo CRL, Area, and Volume

Crown–rump length was calculated using the software provided with the Nikon DS-L2 Digital Sight camera. The program was calibrated using a hemocytometer at the magnification (×.75) the embryo pictures were taken. A line was drawn from the top of the crown to the bottom of the rump of the E13.5 embryo, and the distance between the two points was given in micrometers. Area was calculated using the ImageJ computer software program (National Institutes of Health, Bethesda, MD). Embryo pictures were imported onto the computer and area was calculated by tracing the embryo using the polygon function. To determine total embryo volume, unbiased stereology was performed on 22-μm embryo sections in accordance with established principles (Mouton,2002). Images were viewed using the Stereologer system and software (Stereology Resource Center, Chester, MD) with systematic random sampling using the Calvalieri-point counting method was used to obtain the volumes. Beginning at a random start point, every 15th section was analyzed with dissectors representing 100,000 μm2 per point. Average coefficient of error (CE) for all volumes was < 0.10. Forty-nine E13.5 embryos (12 offspring from B6C3F1, 7 euploid control, 15 Ts65Dn euploid, and 15 Ts65Dn offspring from trisomic mothers) were used for correlation analysis.

Statistical Analyses

A chi-square goodness of fit test was used to determine if trisomic and euploid offspring were different from Mendelian ratios. Differences between offspring from B6C3F1, euploid control, Ts65Dn euploid, and Ts65Dn trisomic mothers were determined using analysis of variance in PROC GLM (SAS, Cary, NC). Least significant difference post hoc comparisons (contrasts) were used to determine differences between strains for individual phenotypes. A significance level of α = 0.05 was used in all multiple comparison tests. Correlation between E13.5 CRL, area and volume was determined using PROC CORR in SAS. A one-tailed t-test was conducted to test for genotypic and phenotypic differences between two classifications of offspring.


We thank Emily Merkel Thomas, Danny Carney, Sashana Gordon, Nicole Shepherd, Brady Harman, Nikki Duvall, Jared Allen, and Justin VanHorn for assisting with the genotyping, dissecting, processing, and quantifying developmental stages of E9.5 and E13.5 embryos. We also thank Grady Chism and Roger Reeves for critical readings of the manuscript. We also appreciate helpful comments from two anonymous reviewers. R.J.R. was funded by a Research Support Funds Grant from IUPUI.