Advertisement
  • Loading metrics

Transcription factor ASCL2 is required for development of the glycogen trophoblast cell lineage

  • Aaron B. Bogutz ,

    Contributed equally to this work with: Aaron B. Bogutz, Rosemary Oh-McGinnis

    Roles Data curation, Formal analysis, Investigation, Methodology, Writing – review & editing

    Affiliation Department of Medical Genetics, Molecular Epigenetics Group, University of British Columbia, Vancouver, BC, Canada

  • Rosemary Oh-McGinnis ,

    Contributed equally to this work with: Aaron B. Bogutz, Rosemary Oh-McGinnis

    Roles Formal analysis, Investigation, Writing – original draft

    Current address: Department of Biology, Douglas College, New Westminster, BC, Canada

    Affiliation Department of Medical Genetics, Molecular Epigenetics Group, University of British Columbia, Vancouver, BC, Canada

  • Karen J. Jacob,

    Roles Investigation

    Affiliation Department of Medical Genetics, Molecular Epigenetics Group, University of British Columbia, Vancouver, BC, Canada

  • Rita Ho-Lau,

    Roles Investigation

    Affiliations Department of Medical Genetics, Molecular Epigenetics Group, University of British Columbia, Vancouver, BC, Canada, Lunenfeld-Tanenbaum Research Institute, Mount Sinai Hospital, Toronto, ON, Canada

  • Ting Gu,

    Roles Investigation

    Current address: College of Animal Science, South China Agricultural University, Guangzhou, China.

    Affiliation Department of Medical Genetics, Molecular Epigenetics Group, University of British Columbia, Vancouver, BC, Canada

  • Marina Gertsenstein,

    Roles Investigation

    Affiliation Lunenfeld-Tanenbaum Research Institute, Mount Sinai Hospital, Toronto, ON, Canada

  • Andras Nagy,

    Roles Conceptualization, Supervision

    Affiliation Lunenfeld-Tanenbaum Research Institute, Mount Sinai Hospital, Toronto, ON, Canada

  • Louis Lefebvre

    Roles Conceptualization, Funding acquisition, Investigation, Methodology, Project administration, Supervision, Writing – original draft, Writing – review & editing

    louis.lefebvre@ubc.ca

    Affiliation Department of Medical Genetics, Molecular Epigenetics Group, University of British Columbia, Vancouver, BC, Canada

Transcription factor ASCL2 is required for development of the glycogen trophoblast cell lineage

  • Aaron B. Bogutz, 
  • Rosemary Oh-McGinnis, 
  • Karen J. Jacob, 
  • Rita Ho-Lau, 
  • Ting Gu, 
  • Marina Gertsenstein, 
  • Andras Nagy, 
  • Louis Lefebvre
PLOS
x
?

This is an uncorrected proof.

Abstract

The basic helix-loop-helix (bHLH) transcription factor ASCL2 plays essential roles in diploid multipotent trophoblast progenitors, intestinal stem cells, follicular T-helper cells, as well as during epidermal development and myogenesis. During early development, Ascl2 expression is regulated by genomic imprinting and only the maternally inherited allele is transcriptionally active in trophoblast. The paternal allele-specific silencing of Ascl2 requires expression of the long non-coding RNA Kcnq1ot1 in cis and the deposition of repressive histone marks. Here we show that Del7AI, a 280-kb deletion allele neighboring Ascl2, interferes with this process in cis and leads to a partial loss of silencing at Ascl2. Genetic rescue experiments show that the low level of Ascl2 expression from the paternal Del7AI allele can rescue the embryonic lethality associated with maternally inherited Ascl2 mutations, in a level-dependent manner. Despite their ability to support development to term, the rescued placentae have a pronounced phenotype characterized by severe hypoplasia of the junctional zone, expansion of the parietal trophoblast giant cell layer, and complete absence of invasive glycogen trophoblast cells. Transcriptome analysis of ectoplacental cones at E7.5 and differentiation assays of Ascl2 mutant trophoblast stem cells show that ASCL2 is required for the emergence or early maintenance of glycogen trophoblast cells during development. Our work identifies a new cis-acting mutation interfering with Kcnq1ot1 silencing function and establishes a novel critical developmental role for the transcription factor ASCL2.

Author summary

By controlling precise networks of target genes, transcription factors play important roles in cell fate determination during development. The Ascl2 gene codes for a transcription factor essential for the maintenance of progenitor cell populations able to differentiate into specialized cell types in the intestine and in the extra-embryonic trophoblast lineage. The trophoblast is an essential component of the placenta, an organ required for development of the embryo in placental mammals. Ascl2 belongs to a group of unusual genes, called imprinted genes, which are expressed from only a single parental copy. Ascl2 is only expressed from the maternally inherited copy in the trophoblast, the paternal copy being kept silent. Here, we describe an engineered deletion neighboring Ascl2 that interferes with the complete silencing of the paternal copy of the gene. We show that the low amount of ASCL2 produced from this deletion can rescue the embryonic lethality associated with non-functional maternal copies of Ascl2. Although the rescued embryos can often survive to term, their placenta is highly disorganized and lacks members of a specific cell lineage, the trophoblast glycogen cells. By analyzing the transcriptional profile of mutant trophoblast progenitors in vivo and of differentiated trophoblast stem cells, we show that ASCL2 plays a very early role in the formation of this cell lineage.

Introduction

The Ascl2 gene—previously known as Mash2—was originally cloned as a mammalian homologue of Drosophila achaete-scute genes and codes for a group A bHLH transcription factor [1,2]. In mice, Ascl2 is required for development. Homozygous Ascl2 knock-out mice die at ~E10 due to an extra-embryonic phenotype which can be rescued by providing a functional tetraploid trophoblast lineage [3]. More recently, Ascl2 was shown to play critical roles in a number of other cell types and developmental processes, including intestinal stem cells [4], epidermal development [5], myogenesis [6], and follicular T-helper cells [7]. ASCL2 acts by hetero-dimerization with ubiquitous bHLH factors of the E protein family, such as TCF3/E12-E47, TCF4/E2-2/ITF2, and TCF12/HEB/ALF1 [4,6,8]. These studies have shown that ASCL2-E-protein complexes promote transcriptional activation, although during myogenesis ASCL2 inhibits the action of bHLH myogenic factors by sequestration of E proteins [6], suggesting context-dependent mechanisms for ASCL2 function.

Ascl2 is an imprinted gene in the mouse, expressed only from the maternal allele in trophoblast cells [9]. Consequently, embryos inheriting a maternal null allele (denoted Ascl2KO/+) fail to develop past mid-gestation, whereas paternal heterozygotes (Ascl2+/KO) are fully viable. Imprinting at Ascl2 is tissue-specific: it is biallelically expressed in LGR5-positive intestinal stem cells [4], although the mechanism involved in this tissue-specific loss of imprinting (LOI) is still unknown. In trophoblast, the paternal allele of Ascl2 is silenced by the paternally expressed long non-coding RNA (lncRNA) Kcnq1ot1, which bidirectionally establishes a repressive epigenetic domain in cis, covering more than 600 kb on distal Chr7 [1015]. As a consequence, Ascl2 is only expressed from the maternal allele, as are seven other protein-coding genes in the region [16].

Ascl2 transcripts are deposited in the oocyte and zygotic transcription starts at the late two-cell stage [17]. Despite this early expression, Ascl2 is first required only postimplantation, since maternal and zygotic mutants are both embryonic lethal at ~E10 [3,17]. Following implantation, the polar trophectoderm (TE) proliferates under FGF signaling from the underlaying epiblast [18]. These few cells will form a population of extra-embryonic ectoderm cells from which all the trophoblast cell types of the placenta will differentiate, with the exception of primary parietal trophoblast giant cells (P-TGCs), which are derived from the mural trophectoderm [19]. As it expands in the proximal-distal axis, this mass of diploid trophoblast cells forms two morphologically distinct structures: (i) the ectoplacental cone (EPC), surrounded by secondary P-TGCs, derived from precursors in the EPC by endoreduplication, that are in direct contact with the decidua basalis; and (ii) the extra-embryonic ectoderm (ExE), extending from the EPC down to the epiblast. From E6.5 to E7.5, high levels of Ascl2 are mostly restricted to diploid cells of the EPC, as FGF4 and Nodal signaling from the epiblast prevents high levels of expression in the stem cell compartment of the ExE and its derivative, the chorionic ectoderm (ChE) [2022]. With the occlusion of the ectoplacental cavity and the concomitant loss of stem cell potential in ChE between E8.0 and 8.5 [23], Ascl2 remains highly expressed in the EPC but is also detected throughout most of the ChE [8,17]. Subsequently, from E9.0 to E12.5, Ascl2 is detected in specific derivatives of the EPC and the ChE, namely spongiotrophoblast (SpT) cells of the junctional zone (Jz) and some labyrinth trophoblast of unknown cell type, respectively [8,17,24].

The phenotype of Ascl2-deficient conceptuses is characterized by an inability to maintain diploid precursors within the EPC and a consequent absence of Tpbpa-positive Jz cells at E9.5 [3]. In the absence of Ascl2, multipotent EPC cells appear to default to the P-TGC lineage, resulting in an expansion of the P-TGC layer. Consistent with a transient role in specific multipotent diploid EPC and SpT cells, Ascl2 levels decline past E12.5 and no expression is detected in differentiated derivatives of those cells, such as glycogen trophoblast (GlyT) cells and three giant cell lineages: P-TGCs, canal TGCs (C-TGCs), and spiral artery-associated TGCs (SpA-TGCs) [2527].

We previously described the generation of a ~280-kb deletion spanning the Ascl2-Ins2 intergenic region and encompassing the gene for tyrosine hydroxylase, Th (Fig 1A) [28]. This deletion—called Del7AI—does not interfere with the establishment or maintenance of differential DNA methylation at the two neighbouring imprinting centers (ICs) regulating, amongst other genes, the lncRNAs H19 and Kcnq1ot1. However, maternal transmission of Del7AI, recovered at the expected Mendelian ratio, leads to an intrauterine growth restriction phenotype [24]. Del7AI/+ placentae express Ascl2 at ~60% of wild-type levels, are highly disorganized, but can still support development to term. As in the full Ascl2 KO, these Ascl2 hypomorphs show an expanded P-TGC layer but here a very thin Jz is maintained throughout gestation. Consequently, the labyrinth layer in these mutant placentae show an important proximal expansion of the fetal vasculature, suggesting a role for the Jz in restricting growth of the placental vessels [24]. In the mature placenta, the Jz is normally composed of two distinct diploid cell types, SpT and GlyT cells, both derived from the EPC [27]. Although Del7AI/+ placentae maintained a small layer of SpT cells, we could not detect GlyT cells in these mutants, suggesting a role for ASCL2 in the differentiation and/or maintenance of this trophoblast lineage [24]. The GlyT cell type was previously thought to be established past mid-gestation due to the emergence of vacuolated cells within the Jz at ~E12 that also express junctional zone-specific genes such as Tpbpa [2931]. However, specific markers of these mature GlyT cells, notably Pcdh12 and Aldh1a3, have since been shown to be expressed in a subset of EPC cells at E7.5 and E8.5, respectively, suggesting an earlier origin for this lineage [32,33].

thumbnail
Fig 1. Loss of imprinting at Ascl2 in +/Del7AI placentae.

(A) Simplified map of the imprinted region on distal mouse Chr7. Maternally expressed genes are shown in red, paternally expressed genes in blue. The gametic differentially methylated regions acting as imprinting centres (IC) in the H19 and Kcnq1ot1 sub-domains are labelled as IC1 and IC2, respectively. Location of the Del7AI deletion is shown on the left, transcriptional orientation of each gene (arrows), on the right. (B) Allele-specific expression analysis for Ascl2, Tssc4, Cdkn1c, and Phlda2, on E13.5 placental cDNA. M and P show the position of the maternal and paternal bands, respectively. The 2M+P band contains two maternal bands and 1 paternal band; M+P, co-migrating maternal and paternal bands. Informative SNPs are boxed. For the Cdkn1c SNP, CAST = C and 129 = T; and for Phlda2, CAST = C and the 129 = A. C: WT CAST allele; +: WT 129 allele. (C) Maternal to paternal allele expression ratio (M to P ratio) for wild-type (C/+) and paternal deletion mutants (C/Del7AI). ∞: no paternal expression detected. (D) Total expression levels in E15.5 placentae determined by RT-qPCR for three biological replicates per genotype. Error bars: SD for biological replicates. Expression is relative to Ppia levels. *p<0.005.

https://doi.org/10.1371/journal.pgen.1007587.g001

Here we report that paternal inheritance of Del7AI interferes with silencing at Ascl2 in cis, causing partial LOI and providing enough Ascl2 mRNA to rescue maternally inherited Ascl2 mutations. The rescued conceptuses share several phenotypic characteristics with our previously described Del7AI/+ hypomorphs, but show an even more pronounced placental phenotype, associated with lower rates of postnatal survival. Our analysis of these mutant placentae, together with RNA-seq profiling of E7.5 EPCs and differentiation assays of Ascl2-deficient trophoblast stem cells (TSCs), show that ASCL2 plays an essential role in the early specification or maintenance of GlyT cell precursors in the EPC during development, establishing a novel important function for this transcription factor in the trophoblast lineage.

Results

Partial loss of imprinting at Ascl2 in paternal +/Del7AI heterozygotes

Paternal transmission of Del7AI is not associated with any obvious developmental or postnatal phenotypes [28]. However, expression studies suggested that total Ascl2 levels might be increased in +/Del7AI placentae at E9.5 compared to their wild-type littermates [24]. We first asked whether Del7AI interferes with the epigenetic silencing of Ascl2 and other Kcnq1ot1-regulated genes when paternally inherited (Fig 1A). For these allele-specific studies, we crossed +/Del7AI males with females homozygous for a Mus mus. castaneus (C) distal Chr7 haplotype on the C57BL/6J background to obtain C/Del7AI and C/+ embryos. We analyzed the ratios of expressed SNPs at four Kcnq1ot1 targets—Ascl2, Tssc4, Cdkn1c, and Phlda2—from E13.5 placental RNA. Our results provide direct evidence of abnormal paternal Ascl2 expression in C/Del7AI placentae, indicating that Ascl2 is not properly silenced from the paternal Del7AI allele and exhibits partial loss of imprinting (LOI) (Fig 1B). We also noticed slight paternal expression for the neighboring gene Tssc4, although in wild-type conceptuses this gene is not tightly imprinted, as also shown by others [34]. In contrast, the more distal imprinted genes Cdkn1c and Phlda2 showed no indication of LOI in the mutant placentae (Fig 1B). Maternal to paternal expression ratios confirmed that partial LOI was occurring at Ascl2 and Tssc4, but not at Cdkn1c or Phlda2 in paternal heterozygous Del7AI placentae (Fig 1C). We also assessed total levels of gene expression by RT-qPCR and found corresponding significant increases in total Ascl2 (1.4-fold) and Tssc4 (1.6-fold) mRNA levels in E15.5 +/Del7AI placentae compared to wild types, but not for the two other genes analyzed (Fig 1D). LOI at Ascl2 was also quantified by blotting of restriction fragment length polymorphism analysis and confirmed using a strain-specific primer (S1 Fig). Together, our results show that the paternal allele of Ascl2 is expressed at ~30% of the maternal allele levels in +/Del7AI heterozygotes, representing ~23% of the total Ascl2 levels in the mutants.

We previously showed that Kcnq1ot1 is expressed and normally imprinted in +/Del7AI embryos [28]. However, our analysis only documented transcription close (1.3 kb) to its transcription start site (TSS). The main stable and polyadenylated isoform of Kcnq1ot1 spans ~83 kb and covers parts of introns 11 and 10 of Kcnq1 [35]. However, transcripts extending all the way to ~121 kb and ~470 kb downstream of the TSS and requiring a functional Kcnq1ot1 promoter have also been reported, although their abundance relative to the shorter isoform is conflicting [35,36]. Nevertheless, since the polyadenylated 470-kb isoform extends 130 kb within the region deleted in Del7AI (S2A Fig), this raised the possibility that its production from the deletion allele might be perturbed in +/Del7AI mutants showing partial LOI at Ascl2. Using primer pairs amplifying regions at 0.3, 202, and 307 kb from the Kcnq1ot1 TSS, we were able to replicate the published data [36] and confirm transcription up to 21.3 kb upstream of Ascl2 (307k PCR reaction) in wild-type E13.5 placenta (S2B Fig). Expression of the longer Kcnq1ot1 isoform was also detected in +/Del7AI placentae, arguing against a major interference with the production of the lncRNA up to the deletion breakpoint (S2B Fig).

Importantly, Igf2 mRNA levels and expression pattern are not perturbed in the mutants, arguing against a spreading of Kcnq1ot1 silencing effects to the paternally expressed Igf2 gene in cis, across the deletion breakpoints (Fig 1A and S3 Fig). This result is consistent with the observation that +/Del7AI mice do not exhibit an abnormal growth phenotype [28].

Elevated Ascl2 levels in differentiated +/Del7AI trophoblast stem cells

To develop a cell culture-based system for further analyses of this deletion, we established trophoblast stem cell (TSC) lines from blastocysts carrying a maternally or paternally inherited Del7AI allele. Ascl2 levels peak between days 1 and 3 of TSC differentiation induced by FGF4 withdrawal [18,21,3740]. Furthermore, RNA-seq profiling of hybrid TSC lines showed that Ascl2 is already imprinted in undifferentiated TSCs, with more than 90% of expression coming from the maternal allele [41]. We used RT-qPCR to quantitate Ascl2 mRNA levels in wild-type and Del7AI heterozygous TSCs after 2 days of differentiation by culture in the absence of FGF4 and conditioned medium (S4A Fig). The paternal Del7AI allele causes a significant increase in total Ascl2 levels in these differentiated mutant cell lines compared to wild-type cells (~1.6-fold), in support of our observations in vivo (Fig 1D), and consistent with LOI at Ascl2 in +/Del7AI mutant TSCs.

Rescue of Ascl2 deficiency by paternal inheritance of the Del7AI allele

Maternal transmission of the Ascl2 knock-in allele Ascl2lacZ (official name Ascl2tm1.1Nagy) results in embryonic lethality at ~E10.0, as seen for the null allele [3], even though the bicistronic Ascl2-IRES-lacZ mRNA of this allele should produce a functional ASCL2 protein [42]. We found that the Ascl2 mRNA from this allele is produced at only ~11% of wild-type levels in mutant Ascl2lacZ/+ ectoplacental cones at E7.5 (see below). We crossed Ascl2+/lacZ females with +/Del7AI males to determine if the small amount of paternal Ascl2 expression from Del7AI could rescue the embryonic lethality seen in Ascl2lacZ/+ mutants. We focused on in utero stages beyond the E10 lethality (E12.5–18.5) as well as on live pups the day following delivery (P0, Fig 2). Since Ascl2lacZ is considered a weak, hypomorphic mutation [42], we also performed similar crosses with the original Ascl2 null allele (Ascl2KO, official name Ascl2tm1Alj) to assess the effects of total Ascl2 levels on our results [3]. First, we observed that both mutant Ascl2 alleles give very few Ascl2-/+ live revertants when crossed to wild-type males, a phenotypic rescue that would require activation of the normally silent paternal allele of Ascl2. No live Ascl2KO/+ embryos or pups were obtained from 20 litters (0/91.5, observed/expected). For the Ascl2lacZ allele, 4 Ascl2lacZ/+ embryos were recovered beyond E12.5 from 13 litters from wild-type males, but 3 of these were small and necrotic, suggesting incomplete resorption. Furthermore, two live Ascl2lacZ/+ pups were recovered from 43 litters implicating wild-type and +/Del7AI males, leading to an overall epigenetic reversion rate of ~0.9% (3/333) for Ascl2lacZ/+ progeny (Fig 2).

thumbnail
Fig 2. Phenotypic rescue of the mid-gestational lethality of Ascl2 mutants by paternal inheritance of Del7AI.

(A) Number of rescued embryos at E12.5 to E18.5 from crosses between Ascl2+/lacZ or Ascl2+/KO females and +/+ or +/Del7AI males. : includes 3 small and necrotic embryos. The data for individual litters are presented in S1 Table. (B) Number of live Ascl2 mutant pups at birth obtained from similar crosses. (C) Bar charts showing the percentage observed over expected values from crosses analyzed in utero (left) and at birth (right). X-axis: male genotypes. Values from crosses involving Ascl2+/lacZ and Ascl2+/KO females are shown in grey and black, respectively. The probability values were obtained by the chi-square test of the null hypothesis of Del7AI = WT, where the rescue frequency of the wild-type allele, 0.901%, is based on all crosses involving the Ascl2lacZ allele (3/333 Ascl2lacZ/+ progeny). * p<1×10−10.

https://doi.org/10.1371/journal.pgen.1007587.g002

Our results show that the paternal Del7AI allele acts as a strong suppressor of the Ascl2lacZ phenotype, with more than 75% of the expected Ascl2lacZ/Del7AI compound heterozygotes surviving past E15 (Fig 2A and S1.1 Table). This phenotypic rescue is sensitive to the overall levels of Ascl2 mRNA since we observed a marked decrease in the developmental rescue of the null allele Ascl2KO by Del7AI, which plummets to ~7%. We also found that for live pups, just over 22% of the expected Ascl2lacZ/Del7AI mice survived to term, down from ~75% rescue in utero, suggesting that although they can develop through late gestation, several of these conceptuses do not survive perinatally. Consistent with this possibility, Ascl2lacZ/Del7AI rescued embryos appeared normal, although growth retarded (see below), as late as E17.5 in gestation (S1.1 Table). Nevertheless, the observed frequencies of phenotypic rescue of the Ascl2lacZ allele, but not of Ascl2KO, by Del7AI are highly significant, both in utero and at birth (Fig 2, and S1 Table).

Live rescued Ascl2lacZ/Del7AI pups are growth-retarded and this growth phenotype largely persists until weaning age (Fig 3A). Since this growth retardation was already observed at birth, we measured the placental and embryonic weights of rescued Ascl2lacZ/Del7AI conceptuses at E15.5. We noticed a significant (~29%) reduction in placental weight compared to their wild-type littermates (Fig 3B). Embryonic weights of the mutants were also smaller than wild-type littermates at this stage (~21%) (Fig 3C). We have previously reported that there is no difference in placental and embryonic weights in +/Del7AI hemizygotes compared to wild-type littermates [24]. These rescue experiments provide genetic evidence that Del7AI causes LOI at Ascl2 and reveal a novel hypomorphic Ascl2 phenotype in rescued, growth retarded conceptuses.

thumbnail
Fig 3. Growth retardation of rescued Ascl2lacZ/Del7AI heterozygotes.

(A) Scatterplots showing weights of rescued live Ascl2lacZ/Del7AI pups (–/Del7AI, n = 7) and their wild-type littermates (n = 14) at postnatal days 0 (P0, birth), 7, and 21. Scatterplots of E15.5 placental (B) and embryonic (C) weights of Ascl2lacZ/Del7AI conceptuses (–/Del7AI, n = 10) and wild type littermates (n = 25). In all graphs, the bars show the average weight ± SD.

https://doi.org/10.1371/journal.pgen.1007587.g003

Ascl2lacZ/Del7AI placentae exhibit an expanded giant cell layer, a hypoplastic spongiotrophoblast, and lack glycogen trophoblast cells

The results presented above suggest that Ascl2 levels are suboptimal and lead to a placental phenotype in rescued conceptuses. This possibility was first studied by histological analyses of placental sections at E15.5 (Fig 4). As described for the growth-retarded Del7AI/+ conceptuses [24], we observed an expansion of the parietal trophoblast giant cell (P-TGC) layer in rescued Ascl2lacZ/Del7AI placentae. The histology data also suggest a reduction in number or absence of spongiotrophoblast and glycogen trophoblast (GlyT) cells in these mutants, the latter being present in both the Jz and decidua at this developmental stage in wild-type placentae [24,43]. These abnormalities were not observed in the paternal +/Del7AI heterozygotes, which were indistinguishable from wild-type placentae (Fig 4).

thumbnail
Fig 4. Hyperplastic parietal trophoblast giant cell layer in Ascl2lacZ/Del7AI placentae at E15.5.

Heamatoxylin and eosin (H&E, top) and DAPI (bottom) staining of wild-type, +/Del7AI and rescued Ascl2lacZ/Del7AI placental sections at E15.5. DAPI staining readily identifies the large polyploid nuclei of the cells stacked at the giant cell layer in the mutants. dec, decidua; P-TGC, parietal trophoblast giant cells; SpT, spongiotrophoblast cells; GlyT, glycogen trophoblast cells; lab, labyrinthine layer. Dashed lines indicate the P-TGC layer boundary. Scale bar: 0.5 mm.

https://doi.org/10.1371/journal.pgen.1007587.g004

To describe this phenotype in more detail we conducted a marker analysis by in situ hybridization on E15.5 placentae of wild-type, +/Del7AI, and rescued Ascl2lacZ/Del7AI genotypes. First, our results confirmed higher levels of Ascl2 expression in +/Del7AI compared to wild-type spongiotrophoblast and further suggested a near-absence of Ascl2-expressing spongiotrophoblast cells in rescued placentae (Fig 5). Both of these observations are supported by RT-qPCR analyses (Fig 1D and S4B Fig). The near absence of a Jz was also confirmed by the severe reduction in Tpbpa-positive cells in rescued placentae (Fig 5). In addition to strong expression in SpT cells, Tpbpa is also expressed at lower levels in GlyT cells at this stage [30,31]. These cells represent an important fraction of Tpbpa-positive cells in wild-type and +/Del7AI placentae at E15.5, readily seen in the decidua, but are absent in rescued Ascl2lacZ/Del7AI conceptuses, which only show a small layer of SpT cells. Loss of signal for Pcdh12 and Cdkn1c expression in these compound heterozygotes provided further support for the absence of GlyT cells (Fig 5). Pcdh12 is exclusively expressed in GlyT cells in the placenta [32] whereas Cdkn1c labels both the labyrinth and GlyT cells at E15.5 [43,44]. Histological staining of glycogen content by periodic acid-Schiff (PAS) showed that GlyT cells are absent in E15.5 Ascl2lacZ/Del7AI placentae (Fig 5) as well as in the single E16.5 Ascl2KO/Del7AI rescued conceptus recovered (Fig 6A). At this later stage, the mutant placenta also shows little sign of SpT cells and an extended labyrinth layer reaching the hyperplastic P-TGC layer, with no sign of Jz. The presence of a dense and disorganized labyrinth reaching the P-TGC layer was also evident in the expression patterns of Prl3b1, Gcm1, and laminin (S5 Fig). Together our results show that as a consequence of reduced Ascl2 levels the rescued mature placentae are severely hypoplastic for spongiotrophoblast cells and lack GlyT cells.

thumbnail
Fig 5. Abnormal spongiotrophoblast and glycogen trophoblast cell development in Ascl2lacZ/Del7AI placentae at E15.5.

Frozen sections of E15.5 placentae were analyzed for the expression of Ascl2, Tpbpa, Pcdh12, and Cdkn1c by in situ hybridization. Periodic acid-Schiff (PAS) staining of glycogen was performed on paraffin sections of E15.5 placentae. Multiple sections from two placentae of each genotype, labeled at the top, were assessed and representative pictures are shown. SpT, spongiotrophoblast cells; lab, labyrinthine layer; GlyT, glycogen trophoblast cells; P-TGC, parietal trophoblast giant cells; dec, maternal decidua. Scale bars: 0.5 mm.

https://doi.org/10.1371/journal.pgen.1007587.g005

thumbnail
Fig 6. Absence of glycogen cell precursors in Ascl2 mutant placentae.

(A) Histological sections of wild-type and rescued Ascl2KO/Del7AI placentae at E16.5 were stained with heamatoxylin and eosin (H&E, top) or PAS with heamatoxylin nuclear counter-stain (bottom). SpT, spongiotrophoblast cells; lab, labyrinthine layer; GlyT, glycogen trophoblast cells; P-TGC, parietal trophoblast giant cells. Scale bar: 0.5 mm. (B) Immunofluorescence staining with PCDH12 polyclonal antibodies (red) and phalloidin (green) on cryosections of E8.5 wild-type and Ascl2lacZ/+ conceptuses, counterstained with DAPI (blue). Scale bar: 0.1 mm.

https://doi.org/10.1371/journal.pgen.1007587.g006

ASCL2 is required for the emergence of glycogen trophoblast cells in the ectoplacental cone

Although the histological detection of glycogen has long been used to identify mature GlyT cells in the placenta past mid-gestation, few marker genes, such as Pcdh12 and Aldh1a3, have been shown to label presumptive GlyT cell progenitors within the ectoplacental cone (EPC) at E7.5-E8.5 [32,33]. Since Ascl2 is expressed at high levels in the EPC at those stages [3,42], we studied the requirement for ASCL2 in the emergence of this progenitor population. Using polyclonal anti-PCDH12 antibodies [45], the presumptive GlyT precursors are readily detected in the EPC of wild-type conceptuses at E8.5 (Fig 6B). As previously reported, we found that the EPC is already hypoplastic in Ascl2lacZ/+ mutants at this stage [3,42]. More importantly, the few remaining EPC cells seen in Ascl2lacZ/+ mutants are not positive for PCDH12 (Fig 6B), suggesting a failure to differentiate or maintain GlyT precursors in the absence of full Ascl2 levels. Although PCDH12-positive cells were previously reported in the uterine crypt at E7.5 [32], we found that such staining, visible notably in the Ascl2lacZ/+ mutants, represent a primary antibody-independent background (S6 Fig).

To analyze global gene expression profiles prior to a morphologically visible phenotype, we mined RNA-seq datasets for wild-type and Ascl2lacZ/+ EPCs at E7.5 [46]. Overall, expression levels for most of the 25,071 autosomal RefSeq genes queried are highly correlated between the two genotypes (r = 0.9739, Fig 7A). We identified 217 significantly downregulated genes in the mutant EPCs (Z-score < −1) and 72 upregulated genes (Z-score > 1, Fig 7A and S2 Table). Ascl2 itself is expressed at only ~11% of wild-type levels from the Ascl2 hypomorphic allele in the mutant EPCs. Other significantly downregulated genes include markers of GlyT precursors in the EPC such as Pcdh12 [32], Aldh1a3 [33], and Tpbpa [31], as well as placental prolactin-related genes expressed in both E8.5 EPC and E14.5 GlyT cells, such as Prl6a1, and Prl7c1 (Figs 7B and 7C, and S3 Table) [47]. Other genes significantly downregulated in the Ascl2 mutant EPC and expressed in GlyT cells of the mature placenta include Car2, Plau, Ncam1, and Prdm1/Blimp1 (Fig 7A) [4850]. The RNA-seq data show that relatively few genes are affected at this early stage by the loss of more than 90% of Ascl2 mRNA in the E7.5 EPC, but also that, significantly, several of the affected genes are known markers of the GlyT cell lineage.

thumbnail
Fig 7. RNA-seq analysis of wild-type and Ascl2-deficient E7.5 ectoplacental cones.

(A) Scatter plot showing average RPKM values for 25,071 autosomal RefSeq genes, from the RNA-seq analysis of dissected E7.5 EPCs from three wild-type (Ascl2+/+) and three mutant (Ascl2lacZ/+) littermates. Significantly upregulated genes (72 genes, Z > 1) and downregulated genes (217 genes, Z < −1) are shown in blue and red, respectively. r: Pearson correlation coefficient. (B) Average RPKM values for selected genes known to be expressed in spongiotrophoblast and glycogen trophoblast derivatives of EPC progenitors, from the RNA-seq analysis of E7.5 EPCs of given genotypes. Ppia is included as a housekeeping gene control. * p<0.05; ** p<0.01; *** p<0.005. (C) Fold change in RPKM values (mutant/wild-type) for the genes analyzed in B. Graphs show average and SD for three biological replicates per genotype.

https://doi.org/10.1371/journal.pgen.1007587.g007

Markers of GlyT cell precursors are not induced upon differentiation of mutant Ascl2lacZ/+ TSCs

The observation that Ascl2lacZ/+ EPCs lack PCDH12-positive cells in vivo and show broad defects in the expression of several GlyT cell markers suggest that ASCL2 might be required as early as E7.5 for the differentiation of this trophoblast lineage. To look at these early events, we established TSC lines from Ascl2+/+ and Ascl2lacZ/+ littermate blastocysts. Cell lines of both genotypes were recovered, with no visible differences in morphology or growth when maintained under undifferentiated growth conditions. Ascl2 is expressed at ~14% of wild-type levels on average in the mutant TSC lines (Fig 8A), similar to the results obtained in EPCs (Fig 7B). This suggests that Ascl2 is not required for the derivation or self-renewal of undifferentiated TSCs, consistent with the absence of morphological defects prior to E8.5 in Ascl2-null conceptuses [3]. Wild-type and mutant TSC lines were differentiated by FGF4 and fibroblast conditioned medium withdrawal for 8 days and lineage markers were analyzed by RT-qPCR. Within 24 hours of differentiation, lines of both genotypes showed a drastic decrease in the levels of Cdx2 mRNA, a marker of undifferentiated TSCs (Fig 8A) [18]. Whereas Pcdh12 [32], Tpbpa [30], and Car2 [48] are induced within ~4 days of differentiation of wild-type TSCs, the mutant cells failed to activate these markers of GlyT (Pcdh12) and Jz cells (Tpbpa, Car2; Fig 8A). Although levels of the P-TGC marker Prl3d1 [47] were reduced ~8-fold in the differentiated mutant TSCs, endoreduplication was not affected (Fig 8B and S7 Fig). The results show that Ascl2 is required for the differentiation of GlyT precursors from undifferentiated TSCs, revealing a novel function for this bHLH transcription factor in early trophoblast development.

thumbnail
Fig 8. Differentiation defects of Ascl2lacZ/+ trophoblast stem cells.

(A) Wild-type (Ascl2+/+) and mutant (Ascl2lacZ/+) TSCs were differentiated by FGF4 and conditioned medium withdrawal for 8 days and RNA samples were collected at the time points shown. Expression levels for Ascl2, Cdx2, Pcdh12, Tpbpa, Car2, and Prl3d1 were analyzed by RT-qPCR and expressed as levels relative to those of the house-keeping gene Ppia. The graphs show the average ± SD for technical triplicates. (B) Flow cytometric analysis of ploidy during TSC differentiation, performed by DNA staining with propidium iodide. Cells of higher ploidy (8n and 16n) are seen in differentiated cells of both genotypes starting at day 6 of differentiation. FACS profiles are shown in S7 Fig.

https://doi.org/10.1371/journal.pgen.1007587.g008

Discussion

Glycogen trophoblast cells represent an abundant lineage of the mature mouse placenta although both the origin and function of these cells remain unknown. Originally characterized by their accumulation of large cytoplasmic stores of glycogen past mid-gestation, they were first detected as cell clusters within the Jz from ~E12.5 [29]. GlyT cells are migratory and starting at ~E14.5 they cross the P-TGC layer and invade the decidua, a property they share with SpA-TGCs [26,29]. Between E12.5 and E16.5 they are highly proliferative and expand more than 250-fold, while SpT cells increase by less than 4-fold [43]. More than 50% of these cells lose their glycogen content after E15.5. Consequently, it was proposed that GlyT cells might provide energy for embryonic growth late in gestation or for triggering parturition [29,43,51].

A number of observations suggest that GlyT precursor cells are already specified in the EPC. The most compelling evidence comes from the expression patterns of markers specific to GlyT cells in mature placentae and also expressed in specific cells of the EPC at E7.5 to 8.5, such as Pcdh12 and Aldh1a3 [32,33]. Several other markers are expressed in the EPC and later in GlyT cells of the mature placenta, but unlike Pchd12 and Aldh1a3, these are also expressed in other trophoblast cell types. Cre/loxP-based cell lineage tracing experiments have been performed for two of those genes, Tpbpa and Prdm1, and confirmed their expression in GlyT amongst other trophoblast cell types in mature placentae [26,49]. However, Cre recombination and activation of the conditional reporters used were not documented in EPC, therefore leaving the possibility that GlyT cells were labelled only past mid-gestation, within the Jz. Another line of evidence comes from the direct detection of few glycogen-containing cells within the middle top portion of the EPC at E6.5–7.5 using histological PAS staining [32,52]. Together, these results suggest that the EPC contains different diploid populations, including precursors of GlyT cells initially found in the proximal portion of the EPC.

Here we show that the bHLH transcription factor ASCL2, which is broadly expressed in the EPC, is required for the presence of these GlyT cell precursors. Since heterodimerization with one of the E proteins is required for ASCL2 function [8], the subset of EPC cells developing as GlyT cell precursors may therefore be at least in part dictated by restricted expression of E protein genes. Both Tcf4 and Tcf12, coding for the E proteins E2-2/ITF2 and HEB/ALF1, but not Tcf3, are expressed in the EPC at E8.5. However the available data suggest that both are mostly expressed in the distal portion of the EPC, closer to the chorion [8]. In contrast to those results, the RNA-seq data on individual wild-type EPC at E7.5 show that those three genes, including Tcf3, are expressed at this stage, with the following RPKM values: Tcf3, 6.74±0.31; Tcf4, 9.80±0.22; and Tcf12, 15.68±1.25 [46]. Mice deficient for any of these three E protein genes die before weaning, with no report of abnormalities suggesting a developmental phenotype, except perhaps for the Tcf4-/- mice, which are recovered at low frequency at birth [5355]. Whether the development of GlyT precursors is affected in any of those mutants, with the possibility of redundancy, still needs to be addressed.

Mature placentae developing from EPCs expressing Ascl2 at suboptimal levels show a severe reduction in SpT cells and an absence of GlyT cells, as seen in Del7AI/+ placentae (~60% of wild-type Ascl2 levels) [24] and in the current description of Ascl2lacZ/Del7AI rescued placentae (~41% of wild-type levels). Our observations that several GlyT cell markers are downregulated in mutant EPCs before the manifestation of a visible phenotype and that Ascl2-deficient TSCs fail to upregulate Pcdh12 and Tpbpa upon differentiation both support the conclusion that ASCL2 is required in diploid EPC cells for the emergence or early maintenance of the GlyT cell lineage. Together with previous observations [3,24,42] our results show that Ascl2 function is level-sensitive in the trophoblast. Total Ascl2 mRNA expression levels between ~60% and ~40% of wild-type levels are compatible with development to term but are associated with a severe placental phenotype leading to embryonic growth retardation. Although GlyT cells are deficient in these models and could contribute to the observed embryonic growth phenotype, we also described several other associated placental abnormalities which could contribute to an insufficient placenta. Specific ablation of the GlyT cell lineage using a conditional approach previously applied to other trophoblast lineages would be required to establish the developmental function of this cell population [56,57].

As Ascl2 mRNA levels fall below ~60% of wild-type levels, the placenta fails to develop normally although it can still support the development of most growth restricted embryos to term (S8 Fig). Although the placentae of Ascl2lacZ/Del7AI and Del7AI/+ mutants shares similarities [24], the phenotype described here for Ascl2lacZ/Del7AI mutants is more pronounced, with fewer persisting SpT cells, consistent with the lower overall Ascl2 levels. The postnatal survival rate of Ascl2lacZ/Del7AI mutant pups is consequently much lower (22% vs. 100%), although the cause of the late lethality is currently unknown [28]. At the molecular level, there are two additional distinctions between these two models; i) the maternally expressed form of placental tyrosine hydroxylase, Th [36,58], is deleted in Del7AI/+ but not in Ascl2lacZ/Del7AI mutants; ii) whereas Phlda2 is upregulated in the Del7AI/+ model [24], we found that Phlda2 mRNA levels are similar in wild-type and rescued Ascl2lacZ/Del7AI placentae at E13.5 (S4B Fig). The RNA-seq analysis of E7.5 EPCs also failed to detect a significant increase in Phlda2 levels in the Ascl2lacZ/+ mutants, with RPKM values of 107.0±7.0 and 115.2±6.9 in the wild-types and mutants, respectively (Z-score = 0.551; Fig 7B and 7C). Phlda2, like Ascl2, is also maternally expressed and regulated by Kcnq1ot1 [10,59]. Although overexpression of Phlda2 (2.3- to 4-fold) from BAC transgenes causes a reduction in the volume of the junctional zone, as seen in Ascl2 hypomorphs, the phenotype described here is much more pronounced and is also associated with an expansion of the P-TGC layer, not seen with increased Phlda2 levels [60].

A recent study also described the consequences of overexpressing Ascl2 by ~6-fold in the trophoblast via BAC transgenesis [61]. Although the authors reported a decrease in the average Jz volume at E14.5, when Ascl2 levels are normally lower, the spongiotrophoblast and GlyT cell populations were not affected at earlier stages (E10.5–12.5). Importantly, the transgene leads to higher levels of Ascl2 in the labyrinth, where it is normally expressed at low levels in a few scattered cells, and this overexpression is accompanied by ectopic clusters of PAS-positive, Tpbpa-positive cells within the labyrinth [61]. We propose that some of the expressing labyrinthine cells are able to acquire a GlyT cell phenotype at this ectopic location due to higher Ascl2 levels. Those results thus provide additional evidence that Ascl2 plays a critical role in the emergence of GlyT cells within the trophoblast. It would be important to precisely define the nature of the cell types expressing Ascl2 in wild-type labyrinth and study their function in this cell population using a conditional mutagenesis approach [4].

Another important aspect of our work is the mechanism by which Del7AI leads to partial LOI at Ascl2. Through our analyses in vivo and in mutant TSCs, we estimated that the paternal Del7AI allele allows an abnormal expression of Ascl2 representing ~30% of the wild-type mRNA levels transcribed from the maternal allele. This could imply that all +/Del7AI ASCL2-positive cells express the paternal Ascl2 allele at low levels (constitutive partial LOI) or that only ~30% of those cells experience full LOI and express Ascl2 biallelically (mosaic full LOI). Our observations that +/Del7AI placentae show broad overexpression of Ascl2 in SpT cells and that much fewer than 30% of Tpbpa-positive SpT cells persist in Ascl2lacZ/Del7AI rescued placentae (Fig 5) strongly support the former scenario. More importantly, unlike what is observed for the Ascl2lacZ allele, the paternal Del7AI allele essentially fails to rescue the KO allele (Fig 2). If the observed Ascl2lacZ/Del7AI rescue was attributable to 30% of the cells expressing Ascl2 at wild-type levels from the paternal Del7AI allele (mosaic full LOI), we would expect similar rescue frequencies and phenotypes in both Ascl2lacZ/Del7AI and Ascl2KO/Del7AI conceptuses. Since our deletion can only rescue the Ascl2lacZ allele, our results are consistent with the frequency of rescue being dependent on the overall levels of Ascl2 mRNA per cell, with the paternal Del7AI allele allowing Ascl2 expression at ~30% of wild-type levels by constitutive partial LOI.

Silencing of the paternal allele of Ascl2 is dependent on expression of the lncRNA Kcnq1ot1 in cis, which is itself silenced on the maternal allele by a DNA methylation mark directly inherited from oocytes [10,13,6264]. A conditional deletion of the Kcnq1ot1 promoter showed that continuous expression of the lncRNA is required to maintain silencing of the paternal Ascl2 allele in vivo [65]. The mechanism by which Kcnq1ot1 silences Ascl2 and other maternally expressed genes in the region is mediated in part by recruitment of histone modifying enzymes such as EHMT2/G9A, EZH2 (PRC2 component), and RNF2 (PRC1 component), and cis deposition of the repressive histone marks H3K9me2, H3K27me3, and H2AK119ub1 on the paternal allele of the repressed genes [11,14,15,66]. The most parsimonious explanation for the observed LOI at Ascl2 is that the Pgk-loxP-neo-pA cassette present at the deletion junction, 4.3 kb downstream of Ascl2, interferes with some aspect of Ascl2 silencing [28]. However, we have not detected any paternal Ascl2 expression from the Pgk-loxP-neo-pA-loxP insertion allele (Ascl2M2) used to define the deletion breakpoint [28] when we analyzed Ascl2CAST/M2 placental RNA (S1A–S1C Fig). Furthermore, the Ascl2M2 allele could not rescue the embryonic lethality of the maternal Ascl2lacZ mutation above the rate of reversion seen with a wild-type paternal allele (S1.3 Table).

Another possibility is that the Del7AI deletion removes sequences required for spreading or stable maintenance of the Kcnq1ot1 silencing domain all the way to Ascl2. A number of observations lend support to this hypothesis. First, RT-PCR evidence supported by RNA knock-down and promoter deletion suggests that the lncRNA Kcnq1ot1 can extend up to 470 kb from its promoter, past Ascl2 and Th (S2A and S9A Figs) [36]. Here, we have shown that the longer isoform of Kcnq1ot1, up to 307 kb, appears normally produced from the paternal Del7AI allele, but obviously a large fraction of this isoform is missing because of the deletion (S9 Fig). Second, we and others showed that the Th gene, located in between Ascl2 and Ins2, is expressed at low levels and only from the maternal allele in placenta and TSCs [36,58,67]. We found that placental Th is produced as a maternally-expressed chimeric transcript, initiating within an endogenous retroviral LTR element (RMER19A) located in between Ascl2 and Th, and at least partially silenced via a Kcnq1ot1-dependent mechanism on the paternal allele [58]. Third, we have previously shown that a ubiquitous EGFP transgene inserted upstream of Ins2, at the site defining the proximal breakpoint of Del7AI, is maternally expressed in a Kcnq1ot1-dependent mechanism (S9B Fig) [58]. Fourth, an 800-kb YAC transgene of the Kcnq1ot1 imprinted domain, extending from Nap1l4 to past Th [68], can recapitulate appropriate imprinted expression of most genes present on the transgene when integrated at an ectopic genomic location [69]. Strikingly, unlike the other protein-coding imprinted genes on the construct, both Ascl2 and Tssc4 failed to be silenced upon paternal transmission of this YAC transgene. Since silencing of the same two genes is perturbed by Del7AI, and based on the observations described above, we propose that sequences in between Th and Ins2 are required to establish a full functional silencing domain dependent on Kcnq1ot1 transcription (S9D Fig). Further experimental support for such an extended domain of Kcnq1ot1 silencing would require the analysis of Kcnq1ot1-dependent deposition of repressive histone marks in the region or the description of intra-chromosomal interactions between Kcnq1ot1 lncRNA and sequences around the Th locus, as described for Kcnq1 [66], and could lead to further understanding of the mechanism of silencing and spreading in this imprinted domain. Alternatively, 3’ end sequences missing from the longer Kcnq1ot1 isoform expressed from Del7AI–and possibly from the YAC transgene as well–could be critical for establishment of a fully silenced domain on the paternal allele. Kcnq1ot1 truncation alleles could potentially be exploited to delimit those putative critical elements of the lncRNA, although the available evidence suggests that transcription of this lncRNA can somehow bypass the termination signals of at least three genes: Trpm5, Ascl2, and Th (S9 Fig).

Materials and methods

Ethics statement

Mice were bred and maintained in the Centre for Disease Modelling, Life Sciences Institute, University of British Columbia, under pathogen-free conditions. All animal experiments were performed under certificates A07-0160, A11-0293, and A15-0181 from the UBC Animal Care Committee and complied with the national Canadian Council on Animal Care guidelines for the ethical care and use of experimental animals.

Mice and genotyping

For all heterozygous genotypes, the maternal allele is always presented first. For simplicity, Ascl2–/+ +/Del7AI compound heterozygotes are referred to as Ascl2/Del7AI. The generation and genotyping of the Del7AI (Del(7Ascl2-Ins2)1Lef; MGI ID:3662901) and M2 (Ascl2tm2Nagy; MGI ID: 4399141) alleles were previously described [28]. The Del7AI deletion spans a ~280 kb region, from a SpeI site 2.7 kb upstream of Ins2 exon 1 to a NcoI site 4.3 kb downstream of Ascl2 exon 3 (positions 142,682,271–142,962,499 on GRCm38/mm10) and deletes a single known protein-coding gene, Th. The junction contains a Pgk-loxP-neo-pA cassette transcribed on the (-) strand, like Ascl2 and Ins2 [28]. Two different Ascl2 (Mash2) mutant alleles were used in this study. The original Ascl2KO allele [3] deletes most of the Ascl2 open reading frame and is therefore a null allele (Ascl2tm1Alj; MGI ID: 2153832). In the lacZ knock-in allele Ascl2lacZ (Ascl2tm1.1Nagy; MGI ID:2155757) an IRES-lacZ-pA reporter inserted in the 3‘UTR of Ascl2 destabilizes the bicistronic mRNA leading to embryonic lethality at E10.5, as seen in Ascl2KO/+ heterozygotes [42]. The animals in this study were all on the CD-1 outbred mouse background except for allele-specific studies, in which one parent is an incipient congenic for a Mus mus. castaneus (CAST/EiJ) distal chromosome 7 haplotype on a mixed CAST/EiJ × C57BL/6J background. Since they were originally derived from R1 ES cells [70], the mutant alleles are all on strain 129S1 or 129X1 Chr7 haplotypes. Weights of placentae and embryos were taken immediately upon dissection with as much of the liquid removed as possible before weighing.

Allele-specific expression analysis and quantitative reverse transcriptase (RT-qPCR)

For allele-specific expression analyses of IC1- and IC2-regulated genes, random-primed cDNA was generated from E9.5 or E13.5 placental RNA as described previously [24]. For Ascl2 the SNP analyzed is a C to T transition within a HpaII site in the 3’ UTR (position 1471 of RefSeq mRNA NM_008554.3). RT-PCR with primers 1148F and 726R yielded a 725-bp fragment which upon HpaII digestion gives a CAST-specific 316-bp band and 129-specific 218- and 98-bp bands, in addition to common bands of 99 and 310 bp (S1 Fig). For Tssc4, RT-PCR with primers F1 and R35 followed by HaeIII digest give a CAST-specific 247-bp band and a 129-specific 159-bp band. For Cdkn1c, primers p57S and p574 give a 364-bp product and the CAST>129 SNP is C>T at position 316. For Phlda2, primers Ipl 1 and Ipl R2 yield a 578-bp product and the CAST>129 SNP is A>C at position 209. For Cdkn1c and Phlda2, direct sequencing of the PCR products followed by Phred analysis [71] was performed to determine allelic expression ratios. For total expression level determination, cDNAs were analyzed by RT-qPCR. Total RNA was purified using Trizol (Invitrogen) according to manufacturer’s directions and DNase-treated (RQ1, Promega) for at least 1 h to remove contaminating gDNA. RNA was reverse-transcribed with SSII (Invitrogen), using random primers (N15) according to the SSII protocol. Quantitative PCR was performed on a Step-One Plus Real time PCR system (Applied Biosystems) using Eva Green (Biotium). Ct values of three biological replicates, obtained by the LinReg PCR program [72], were averaged and used to calculate relative amounts of transcripts, normalized to levels of the housekeeping gene Ppia [73]. All primer sequences are available in S4 Table.

In situ hybridization (ISH) and immunohistochemistry (IHC)

E13.5 and E15.5 placentae were dissected in PBS and fixed in fresh 4% paraformaldehyde/1×PBS (RNase-free) overnight at 4°C. For E9.5 placentae, entire conceptuses were fixed to ensure integrity of cryostat sections during the ISH procedure. Antisense and sense strand probes for Ascl2, Tpbpa/4311, Pcdh12, Cdkn1c, and Igf2 were DIG-labeled and used for ISH on 10-μm cryostat sections from E13.5 and E15.5 placentae as previously described [25]. Nuclear fast red was used as the counterstain. The Igf2 probe sequence was obtained from www.genepaint.org (Accession number NM_010514; probe 486) and amplified from E13.5 placenta cDNA. IHC for laminin was done as previously described [24]. Slides were treated to remove paraffin, hydrated, and blocked with 0.3% hydrogen peroxide for 30 minutes and then with 5% goat normal serum, 0.5% BSA in PBS-T. Slides were incubated with rabbit polyclonal anti-laminin (Sigma L9393, at a 1/50 dilution in serum blocking solution) overnight at 4°C. The next day, the secondary biotinylated goat anti-rabbit antibody (Jackson ImmunoResearch) was added at a dilution of 1/500 and incubated for 30 minutes. ABC (Vector) was added for 30 minutes and DAB for 1 minute. Sections were counterstained with hematoxylin and washed in Scott’s Tap water solution (2% MgSO4, 0.35% NaHCO3 in distilled water) to help sharpen the contrast. Sections were then dehydrated and mounted with Entellan mounting medium (EM Science) under glass coverslips.

Periodic acid-Schiff (PAS) staining

Placental paraffin sections were treated to remove paraffin, hydrated, and oxidized in 0.5% periodic acid solution for 5 minutes. Slides were placed in Schiff Reagent for 15 minutes, counterstained with hematoxylin, and washed in Scott’s tap water solution to help sharpen the contrast. Sections were then dehydrated and mounted with Entellan mounting medium (EM Science) under glass coverslips. Consecutive sections of one placenta per genotype were examined.

Immunofluorescence

For PCDH12 localization analyses, entire E8.5 conceptuses were processed for cryostat frozen sections as described [74]. The affinity-purified anti-PCDH12 polyclonal antibodies were prepared and used as described [45]. For F-actin staining, sections were washed with water, counterstained with phalloidin (1/400, Invitrogen) for 20 minutes at room temperature, washed again with water, and counterstained with 4’,6-diamindino-2-phenylindole (DAPI, 2 μg/ml, Sigma) for 5 minutes at room temperature. At early developmental stages, we found that staining for F-actin was much stronger in maternal cells, allowing a clear demarcation of the conceptus-uterus boundary. Sections were washed with water and mounted on glass slides under coverslips with Vectashield (Vector Labs). The samples were analyzed on a Leica DMI6000B inverted fluorescent microscope and images were captured and processed with a Q-imaging Retiga 4000R monochrome camera and Openlab (Improvision).

Analysis of RNA-seq data

The NGS datasets for the EPC RNA-seq analysis [46] were downloaded from GEO (accessions GSM1613552-7) and aligned to the mouse genome (mm9) using STAR two-pass mapping for de novo splice-junction generation with default parameters [75]. Data were input into SeqMonk (http://www.bioinformatics.babraham.ac.uk/projects/seqmonk/) for further analysis, removing reads with a mapping quality score of less than 5. RPKM values were generated using SeqMonk’s RNA-Seq Quantitation Pipeline, normalizing read count by gene length.

Trophoblast stem cells

All the TSC lines used in this study were established from E4.0 blastocysts, expanded, and maintained as described [18]. TSCs were grown in RPMI 1640 (Invitrogen) supplemented with 50 μg/ml penicillin/streptomycin, 20% fetal bovine serum (Wisent), 1 mM sodium pyruvate, 100 μM β-mercaptoethanol, 2 mM L-glutamine, 70% 3-day feeder-conditioned medium, 25 ng/ml human recombinant FGF4 (Sigma F2278) and 1 μg/ml heparin (Sigma H3149). Cells were collected for genotyping and sexing by PCR; only XY male cells were selected for expression studies. Differentiation of TSCs was induced by omission of FGF4 and fibroblast-conditioned medium in the culture. The flow cytometric analysis of ploidy during TSC differentiation was performed by DNA staining with propidium iodide as described [76]. All data were obtained on a BD LSRII running BD FACS DIVA and analyzed in FlowJo 9.5.

Statistical analysis

For RT-qPCR, weight comparisons, and RNA-seq expression data, statistical significance was determined using the Student’s t-test (unpaired, two tailed distribution), and data were presented as mean ± SD. For the rescue experiments, the probability values were calculated using the chi-square test of statistical significance against the null hypothesis of Del7AI(or M2) = WT allele in its ability to rescue the embryonic lethality of maternal Ascl2 mutations. The rescue frequency of the wild-type allele, 0.901%, is based on all crosses involving the Ascl2lacZ allele (3/333 Ascl2lacZ/+ progeny) and was used to calculate the expected numbers of Ascl2/Del7AI or Ascl2/M2 progeny (S1.3 Table).

Supporting information

S1 Fig. Allele-specific expression of Ascl2 in wild-type and +/Del7AI placentae.

(A) Diagram of the Ascl2 725-bp exons 2–3 RT-PCR product (primers 1 + 4) showing the positions of HpaII sites for the 129 and CAST (C) alleles with sizes of each fragment given in base pairs. The polymorphic HpaII site within the 3’ UTR is marked by an asterisk. (B) Non-radioactive blot of the HpaII-digested RT-PCR products, hybridized with a DIG-labelled probe from the 218-bp 129 HpaII band (shown in A). The RNA samples analysed are from E9.5 placentae of the given genotypes, where C is the maternal CAST allele and M2, the targeted PGK-loxP-neopA-loxP insertion (Ascl2tm2Nagy) used to define the distal breakpoint of the Del7AI deletion. M, maternal; P, paternal. (C) Allelic ratios (maternal/paternal) for each sample, as determined by ImageJ analysis of the data presented in B. (D) Diagram of the Ascl2 genome from exon 2 to 3, showing the positions of PCR primers for genomic (E) and RT-PCR (F) analyses. The asterisk marks the position of the polymorphic HpaII site. The reverse primer 3 (129R) is 129-specific at its 3’ terminal nucleotide. (E) Intron 2 to exon 3 PCR on genomic DNA from pure 129 and CAST mice as well as a CAST/Del7AI embryo (C/D). Lanes–and M are water controls and a 100-bp marker. (F) Exon 2 to exon 3, 129-specific RT-PCR on cDNA from wild type (C/+) and mutant (C/D) placentae. Lanes–, + and M are water control, a 129 Ascl2 cDNA clone, and a 100-bp marker, respectively. PCR primers: 1, 1148F; 2, in2F1; 3, 129R (129-specific); 4, 726R. PCR primers used are shown at the bottom of each gel figure. Their sequences are given in S4 Table.

https://doi.org/10.1371/journal.pgen.1007587.s001

(PDF)

S2 Fig. Expression of Kcnq1ot1 in +/Del7AI placentae.

(A) UCSC Genome Browser screenshot for the Kcnq1ot1 imprinted domain. From the top, the tracks show: (i) The positions of PCR reactions used by Golding (2011) to define the longest Kcnq1ot1 isoform. (ii) The three PCR reactions used in this study. (iii) The extent of the Del7AI deletion. (iv) The longer Kcnq1ot1 isoforms reported by Golding (2011, ~470 kb)) and Redrup (2009, ~121 kb), as well as the more stable and annotated transcript of ~83 kb. All are transcribed on the (-) strand, from a transcriptional start site (TSS) within intron 11 of Kcnq1. Note that ~130 kilobases of the longest isoform are deleted in Del7AI. (v) UCSC genes annotated in the region, including Ascl2, located immediately distal of the Del7AI breakpoint. (B) RT-PCR detection of Kcnq1ot1 at 0.3, 202, and 307 kb downstream of the TSS, on E13.5 placental RNA from two +/Del7AI and one wild-type control conceptuses. PCRs were performed on total RNA samples, with (+) or without (-) reverse transcriptase (RT) priming of cDNA with random primers (N15). C-: water control. C+: genomic DNA. The molecular weight ladder is the exACTGene 100bp ladder (Fisher Scientific).

https://doi.org/10.1371/journal.pgen.1007587.s002

(PDF)

S3 Fig. Paternal Igf2 expression is unaffected in +/Del7AI placentae at E13.5.

(A) Igf2 RT-qPCR on wild type and +/Del7AI E13.5 placental cDNA. Expression is relative to Ppia. Three biological replicates for each genotype were analysed. Graphs show mean ± SD.

(B) Igf2 ISH on frozen sections of wild type and +/Del7AI E13.5 placentae. Multiple sections from two placentae of each genotype were assessed and representative pictures are shown. The sense probe gave no signal (not shown). The blue stain shows Igf2 expression, mostly in the junctional zone and GlyT cells in the decidua. Scale bar: 0.5 mm. jz, junctional zone; lab: labyrinth; dec, decidua.

https://doi.org/10.1371/journal.pgen.1007587.s003

(PDF)

S4 Fig. Effect of Del7AI on Ascl2 mRNA levels in differentiated TSCs and rescued placentae.

(A) Trophoblast stem cell (TSC) lines of the given genotypes were differentiated for 2 days by FGF4 withdrawal and Ascl2 levels, normalized to Ppia levels, were measured by RT-qPCR. In paternal deletion mutants (+/Del7AI), total Ascl2 levels are increased by 1.6-fold over wild-type TSCs (*, p<0.05). Graphs show mean + SD. The numbers of independent TSC lines of each genotype analysed (biological replicates) are given at the bottom (n =). (B) Relative levels of Ascl2 and Phlda2 in E13.5 wild-type and Ascl2lacZ/Del7AI rescued placentae, determined as described in A. Three samples of each genotype were analysed and graphs show mean ± SD of biological triplicates (**, p = 0.0003).

https://doi.org/10.1371/journal.pgen.1007587.s004

(PDF)

S5 Fig. Abnormal labyrinth development in Ascl2lacZ/Del7AI placentae at E15.5.

Frozen sections of E15.5 placentae of the given genotypes were analysed for the expression of Prl3b1 and Gcm1 by ISH. The basement membrane marker laminin was detected by IHC on paraffin sections. Scale bar: 0.5 mm. Spt, spongiotrophoblast cells; dec, decidua; P-TGC, parietal trophoblast giant cells; lab, labyrinthine layer.

https://doi.org/10.1371/journal.pgen.1007587.s005

(PDF)

S6 Fig. Primary antibody-independent staining in the decidua.

Adjacent sections of the Ascl2lacZ/+ E8.5 conceptuses analysed in Fig 6B were treated as described in this figure but without incubation with the anti-PCDH12 primary antibodies. Punctate staining for the secondary antibody (arrow) is still visible above the giant cell layer, within the decidua. P-TGC, parietal trophoblast giant cells; dec, decidua; ch, chorion.

https://doi.org/10.1371/journal.pgen.1007587.s006

(PDF)

S7 Fig. Endoreduplication of differentiating wild-type and Ascl2lacZ/+ TSCs.

(A) Cell-cycle distribution of wild-type and Ascl2lacZ/+ mutant TSCs as monitored by flow cytometry using propidium iodide staining. Profiles were generated on the indicated days following FGF4 and conditioned medium withdrawal. 2n marks diploid cells in G1 phase, whereas 4n represents a mixture of G2-phase diploid and G1-phase tetraploid cells. Endoreduplication is clearly seen at higher ploidies. (B) Images of wild-type and Ascl2lacZ/+ mutant TSCs at d0 and d4 of differentiation. Scale bar, 100 μm. Refers to data presented in Fig 8B.

https://doi.org/10.1371/journal.pgen.1007587.s007

(PDF)

S8 Fig. Dosage-sensitive effects of Ascl2 mRNA levels on placental phenotype.

For each genotype, the approximate total Ascl2 mRNA levels are presented as a percentage of the wild-type levels, set at 100%.

https://doi.org/10.1371/journal.pgen.1007587.s008

(PDF)

S9 Fig. Potential critical region for extended Kcnq1ot1 silencing.

(A) Structure of the IC1-IC2 imprinted domains on distal mouse Chr7, drawn to scale. Paternally expressed genes are in black, maternally expressed genes in white. Two isoforms of Kcnq1ot1 have been described; the more stable form terminates within intron 10 of Kcnq1, whereas a longer form has been detected, extending all the way past Th, a gene maternally expressed in placenta from the LTR RMER19A (Jones, 2011). (B) The Tel7KI allele carrying a pCAGGS-EGFP reporter inserted upstream of Ins2. The EGFP is imprinted and maternally expressed in the embryo in a Kcnq1ot1-dependent manner (Jones, 2011). (C) Del7AI allele, showing partial LOI at Ascl2 and Tssc4 upon paternal transmission. (D) YAC transgene showing appropriate imprinting of the IC2 domain, except at Ascl2 and Tssc4 (Cerrato, 2005). In both C and D, the 3’ end structure of the longer Kcnq1ot1 isoform is unknown (question marks).

https://doi.org/10.1371/journal.pgen.1007587.s009

(PDF)

S1 Table.

S1.1 Table. Litters collected to assess the rescue of the Ascl2lacZ mutation in utero.

S1.2 Table. Litters collected to assess the rescue of the Ascl2KO mutation in utero.

S1.3 Table. Chi-square tests of rescue data.

https://doi.org/10.1371/journal.pgen.1007587.s010

(XLSX)

S2 Table.

S2.1 Table. Significantly down-regulated autosomal genes (Z-score < -1) identified by RNA-seq analysis of wild-type and Ascl2lacZ/+ mutant EPC at E7.5.

S2.2 Table. Significantly up-regulated autosomal genes (Z-score > 1) identified by RNA-seq analysis of wild-type and Ascl2lacZ/+ mutant EPC at E7.5.

https://doi.org/10.1371/journal.pgen.1007587.s011

(XLSX)

S3 Table. RPKM values for all expressed Prl genes, from RNA-seq analysis of wild-type and Ascl2lacZ/+ mutant EPC at E7.5, as well as their reported expression pattern in the placenta.

https://doi.org/10.1371/journal.pgen.1007587.s012

(XLSX)

Acknowledgments

The authors thank Jay Cross, Rosalind John and Janet Rossant for cDNA clones used as ISH probes, Ken Harpal for placental histology and PAS staining, and John Hsien for help with dissections. We also thank Philippe Huber for the gift of anti-PCDH12 polyclonal antibodies.

References

  1. 1. Johnson JE, Birren SJ, Anderson DJ. Two rat homologues of Drosophila achaete-scute specifically expressed in neuronal precursors. Nature. 1990; 346: 858–861. pmid:2392153
  2. 2. Ledent V, Paquet O, Vervoort M. Phylogenetic analysis of the human basic helix-loop-helix proteins. Genome Biol. 2002; 3: 1–18.
  3. 3. Guillemot F, Nagy A, Auerbach A, Rossant J, Joyner AL. Essential role of Mash-2 in extraembryonic development. Nature. 1994; 371: 333–336. pmid:8090202
  4. 4. van der Flier LG, van Gijn ME, Hatzis P, Kujala P, Haegebarth A, Stange DE, et al. Transcription factor achaete scute-like 2 controls intestinal stem cell fate. Cell. 2009; 136: 903–912. pmid:19269367
  5. 5. Moriyama M, Durham A-D, Moriyama H, Hasegawa K, Nishikawa S-I, Radtke F, et al. Multiple roles of Notch signaling in the regulation of epidermal development. Dev Cell. 2008; 14: 594–604. pmid:18410734
  6. 6. Wang C, Wang M, Arrington J, Shan T, Yue F, Nie Y, et al. Ascl2 inhibits myogenesis by antagonizing the transcriptional activity of myogenic regulatory factors. Development. 2017; 144: 235–247. pmid:27993983
  7. 7. Liu X, Chen X, Zhong B, Wang A, Wang X, Chu F, et al. Transcription factor achaete-scute homologue 2 initiates follicular T-helper-cell development. Nature. 2014; 507: 513–518. pmid:24463518
  8. 8. Scott IC, Anson-Cartwright L, Riley P, Reda D, Cross JC. The HAND1 basic helix-loop-helix transcription factor regulates trophoblast differentiation via multiple mechanisms. Mol Cell Biol. 2000; 20: 530–541. pmid:10611232
  9. 9. Guillemot F, Caspary T, Tilghman SM, Copeland NG, Gilbert DJ, Jenkins NA, et al. Genomic imprinting of Mash2, a mouse gene required for trophoblast development. Nat Genet. 1995; 9: 235–242. pmid:7773285
  10. 10. Fitzpatrick GV, Soloway PD, Higgins MJ. Regional loss of imprinting and growth deficiency in mice with a targeted deletion of KvDMR1. Nat Genet. 2002; 32: 426–431. pmid:12410230
  11. 11. Umlauf D, Goto Y, Cao R, Cerqueira F, Wagschal A, Zhang Y, et al. Imprinting along the Kcnq1 domain on mouse chromosome 7 involves repressive histone methylation and recruitment of Polycomb group complexes. Nat Genet. 2004; 36: 1296–1300. pmid:15516932
  12. 12. Lewis A, Mitsuya K, Umlauf D, Smith P, Dean W, Walter J, et al. Imprinting on distal chromosome 7 in the placenta involves repressive histone methylation independent of DNA methylation. Nat Genet. 2004; 36: 1291–1295. pmid:15516931
  13. 13. Mancini-DiNardo D, Steele SJS, Levorse JM, Ingram RS, Tilghman SM. Elongation of the Kcnq1ot1 transcript is required for genomic imprinting of neighboring genes. Genes Dev. 2006; 20: 1268–1282. pmid:16702402
  14. 14. Lewis A, Green K, Dawson C, Redrup L, Huynh KD, Lee JT, et al. Epigenetic dynamics of the Kcnq1 imprinted domain in the early embryo. Development. 2006; 133: 4203–4210. pmid:17021040
  15. 15. Pandey RR, Mondal T, Mohammad F, Enroth S, Redrup L, Komorowski J, et al. Kcnq1ot1 antisense noncoding RNA mediates lineage-specific transcriptional silencing through chromatin-level regulation. Mol Cell. 2008; 32: 232–246. pmid:18951091
  16. 16. Lefebvre L. The placental imprintome and imprinted gene function in the trophoblast glycogen cell lineage. Reproductive BioMedicine Online. 2012; 25: 44–57. pmid:22560119
  17. 17. Rossant J, Guillemot F, Tanaka M, Latham K, Gertenstein M, Nagy A. Mash2 is expressed in oogenesis and preimplantation development but is not required for blastocyst formation. Mech Dev. 1998; 73: 183–191. pmid:9622625
  18. 18. Tanaka S, Kunath T, Hadjantonakis AK, Nagy A, Rossant J. Promotion of trophoblast stem cell proliferation by FGF4. Science. 1998; 282: 2072–2075. pmid:9851926
  19. 19. Kunath T, Strumpf D, Rossant J. Early trophoblast determination and stem cell maintenance in the mouse—a review. Placenta. 2004; 25 Suppl A: S32–8. pmid:15033304
  20. 20. Georgiades P, Rossant J. Ets2 is necessary in trophoblast for normal embryonic anteroposterior axis development. Development. 2006; 133: 1059–1068. pmid:16481355
  21. 21. Mugford JW, Yee D, Magnuson T. Failure of extra-embryonic progenitor maintenance in the absence of dosage compensation. Development. 2012; 139: 2130–2138. pmid:22573614
  22. 22. Guzman-Ayala M, Ben-Haim N, Beck S, Constam DB. Nodal protein processing and fibroblast growth factor 4 synergize to maintain a trophoblast stem cell microenvironment. Proc Natl Acad Sci USA. 2004; 101: 15656–15660. pmid:15505202
  23. 23. Uy GD, Downs KM, Gardner RL. Inhibition of trophoblast stem cell potential in chorionic ectoderm coincides with occlusion of the ectoplacental cavity in the mouse. Development. 2002; 129: 3913–3924. pmid:12135928
  24. 24. Oh-McGinnis R, Bogutz AB, Lefebvre L. Partial loss of Ascl2 function affects all three layers of the mature placenta and causes intrauterine growth restriction. Dev Biol. 2011; 351: 277–286. pmid:21238448
  25. 25. Oh-McGinnis R, Bogutz AB, Lee KY, Higgins MJ, Lefebvre L. Rescue of placental phenotype in a mechanistic model of Beckwith-Wiedemann syndrome. BMC Dev Biol. 2010; 10: 50–62. pmid:20459838
  26. 26. Simmons DG, Fortier AL, Cross JC. Diverse subtypes and developmental origins of trophoblast giant cells in the mouse placenta. Dev Biol. 2007; 304: 567–578. pmid:17289015
  27. 27. Latos PA, Hemberger M. From the stem of the placental tree: trophoblast stem cells and their progeny. Development. 2016; 143: 3650–3660. pmid:27802134
  28. 28. Lefebvre L, Mar L, Bogutz A, Oh-McGinnis R, Mandegar MA, Paderova J, et al. The interval between Ins2 and Ascl2 is dispensable for imprinting centre function in the murine Beckwith-Wiedemann region. Hum Mol Genet. 2009; 18: 4255–4267. pmid:19684026
  29. 29. Redline RW, Chernicky CL, Tan HQ, Ilan J. Differential expression of insulin-like growth factor-II in specific regions of the late (post day 9.5) murine placenta. Mol Reprod Dev. 1993; 36: 121–129. pmid:8257562
  30. 30. Adamson SL, Lu Y, Whiteley KJ, Holmyard D, Hemberger M, Pfarrer C, et al. Interactions between trophoblast cells and the maternal and fetal circulation in the mouse placenta. Dev Biol. 2002; 250: 358–373. pmid:12376109
  31. 31. Lescisin KR, Varmuza S, Rossant J. Isolation and characterization of a novel trophoblast-specific cDNA in the mouse. Genes Dev. 1988; 2: 1639–1646. pmid:3215514
  32. 32. Bouillot S, Rampon C, Tillet E, Huber P. Tracing the glycogen cells with protocadherin 12 during mouse placenta development. Placenta. 2006; 27: 882–888. pmid:16269175
  33. 33. Outhwaite JE, Natale BV, Natale DRC, Simmons DG. Expression of aldehyde dehydrogenase family 1, member A3 in glycogen trophoblast cells of the murine placenta. Placenta. 2015; 36: 304–311. pmid:25577283
  34. 34. Paulsen M, El-Maarri O, Engemann S, Strödicke M, Franck O, Davies K, et al. Sequence conservation and variability of imprinting in the Beckwith-Wiedemann syndrome gene cluster in human and mouse. Hum Mol Genet. 2000; 9: 1829–1841. pmid:10915772
  35. 35. Redrup L, Branco MR, Perdeaux ER, Krueger C, Lewis A, Santos F, et al. The long noncoding RNA Kcnq1ot1 organises a lineage-specific nuclear domain for epigenetic gene silencing. Development. 2009; 136: 525–530. pmid:19144718
  36. 36. Golding MC, Magri LS, Zhang L, Lalone SA, Higgins MJ, Mann MRW. Depletion of Kcnq1ot1 non-coding RNA does not affect imprinting maintenance in stem cells. Development. 2011; 138: 3667–3678. pmid:21775415
  37. 37. Hughes M, Dobric N, Scott IC, Su L, Starovic M, St-Pierre B, et al. The Hand1, Stra13 and Gcm1 transcription factors override FGF signaling to promote terminal differentiation of trophoblast stem cells. Dev Biol. 2004; 271: 26–37. pmid:15196947
  38. 38. Kidder BL, Palmer S. Examination of transcriptional networks reveals an important role for TCFAP2C, SMARCA4, and EOMES in trophoblast stem cell maintenance. Genome Res. 2010; 20: 458–472. pmid:20176728
  39. 39. Takao T, Asanoma K, Tsunematsu R, Kato K, Wake N. The Maternally Expressed Gene Tssc3 Regulates the Expression of MASH2 Transcription Factor in Mouse Trophoblast Stem Cells through the AKT-Sp1 Signaling Pathway. Journal of Biological Chemistry. 2012; 287: 42685–42694. pmid:23071113
  40. 40. Kaiser S, Koch Y, Kühnel E, Sharma N, Gellhaus A, Kuckenberg P, et al. Reduced Gene Dosage of Tfap2c Impairs Trophoblast Lineage Differentiation and Alters Maternal Blood Spaces in the Mouse Placenta. Biol Reprod. 2015; 93: 31. pmid:26063869
  41. 41. Calabrese JM, Starmer J, Schertzer MD, Yee D, Magnuson T. A survey of imprinted gene expression in mouse trophoblast stem cells. G3 (Bethesda). 2015; 5: 751–759. pmid:25711832
  42. 42. Tanaka M, Puchyr M, Gertsenstein M, Harpal K, Jaenisch R, Rossant J, et al. Parental origin-specific expression of Mash2 is established at the time of implantation with its imprinting mechanism highly resistant to genome-wide demethylation. Mech Dev. 1999; 87: 129–142. pmid:10495277
  43. 43. Coan PM, Conroy N, Burton GJ, Ferguson-Smith AC. Origin and characteristics of glycogen cells in the developing murine placenta. Dev Dyn. 2006; 235: 3280–3294. pmid:17039549
  44. 44. Georgiades P, Ferguson-Smith AC, Burton GJ. Comparative developmental anatomy of the murine and human definitive placentae. Placenta. 2002; 23: 3–19. pmid:11869088
  45. 45. Rampon C, Prandini M-H, Bouillot S, Pointu H, Tillet E, Frank R, et al. Protocadherin 12 (VE-cadherin 2) is expressed in endothelial, trophoblast, and mesangial cells. Exp Cell Res. 2005; 302: 48–60. pmid:15541725
  46. 46. Branco MR, King M, Perez-Garcia V, Bogutz AB, Caley M, Fineberg E, et al. Maternal DNA Methylation Regulates Early Trophoblast Development. Dev Cell. 2016; 36: 152–163. pmid:26812015
  47. 47. Simmons DG, Rawn S, Davies A, Hughes M, Cross JC. Spatial and temporal expression of the 23 murine Prolactin/Placental Lactogen-related genes is not associated with their position in the locus. BMC Genomics. 2008; 9: 352. pmid:18662396
  48. 48. Singh U, Sun T, Shi W, Schulz R, Nuber UA, Varanou A, et al. Expression and functional analysis of genes deregulated in mouse placental overgrowth models: Car2 and Ncam1. Dev Dyn. 2005; 234: 1034–1045. pmid:16247769
  49. 49. Mould A, Morgan MAJ, Li L, Bikoff EK, Robertson EJ. Blimp1/Prdm1 governs terminal differentiation of endovascular trophoblast giant cells and defines multipotent progenitors in the developing placenta. Genes Dev. 2012; 26: 2063–2074. pmid:22987638
  50. 50. Teesalu T, Blasi F, Talarico D. Expression and function of the urokinase type plasminogen activator during mouse hemochorial placental development. Dev Dyn. 1998; 213: 27–38. pmid:9733098
  51. 51. Lopez MF, Dikkes P, Zurakowski D, Villa-Komaroff L. Insulin-like growth factor II affects the appearance and glycogen content of glycogen cells in the murine placenta. Endocrinology. 1996; 137: 2100–2108. pmid:8612553
  52. 52. Tesser RB, Scherholz PLA, do Nascimento L, Katz SG. Trophoblast glycogen cells differentiate early in the mouse ectoplacental cone: putative role during placentation. Histochem Cell Biol. 2010; 134: 83–92. pmid:20544215
  53. 53. Zhuang Y, Cheng P, Weintraub H. B-lymphocyte development is regulated by the combined dosage of three basic helix-loop-helix genes, E2A, E2-2, and HEB. Mol Cell Biol. 1996; 16: 2898–2905. pmid:8649400
  54. 54. Zhuang Y, Soriano P, Weintraub H. The helix-loop-helix gene E2A is required for B cell formation. Cell. 1994; 79: 875–884. pmid:8001124
  55. 55. Bain G, Maandag EC, Izon DJ, Amsen D, Kruisbeek AM, Weintraub BC, et al. E2A proteins are required for proper B cell development and initiation of immunoglobulin gene rearrangements. Cell. 1994; 79: 885–892. pmid:8001125
  56. 56. Outhwaite JE, McGuire V, Simmons DG. Genetic ablation of placental sinusoidal trophoblast giant cells causes fetal growth restriction and embryonic lethality. Placenta. 2015; 36: 951–955. pmid:26091829
  57. 57. Hu D, Cross JC. Ablation of Tpbpa-positive trophoblast precursors leads to defects in maternal spiral artery remodeling in the mouse placenta. Dev Biol. 2011; 358: 231–239. pmid:21839735
  58. 58. Jones MJ, Bogutz AB, Lefebvre L. An extended domain of Kcnq1ot1 silencing revealed by an imprinted fluorescent reporter. Mol Cell Biol. 2011; 31: 2827–2837. pmid:21576366
  59. 59. Qian N, Frank D, O'Keefe D, Dao D, Zhao L, Yuan L, et al. The IPL gene on chromosome 11p15.5 is imprinted in humans and mice and is similar to TDAG51, implicated in Fas expression and apoptosis. Hum Mol Genet. 1997; 6: 2021–2029. pmid:9328465
  60. 60. Tunster SJ, Tycko B, John RM. The imprinted Phlda2 gene regulates extraembryonic energy stores. Mol Cell Biol. 2010; 30: 295–306. pmid:19884348
  61. 61. Tunster SJ, McNamara GI, Creeth HDJ, John RM. Increased dosage of the imprinted Ascl2 gene restrains two key endocrine lineages of the mouse placenta. Dev Biol. 2016; 418: 55–65. pmid:27542691
  62. 62. Smilinich NJ, Day CD, Fitzpatrick GV, Caldwell GM, Lossie AC, Cooper PR, et al. A maternally methylated CpG island in KvLQT1 is associated with an antisense paternal transcript and loss of imprinting in Beckwith-Wiedemann syndrome. Proc Natl Acad Sci USA. 1999; 96: 8064–8069. pmid:10393948
  63. 63. Shin J-Y, Fitzpatrick GV, Higgins MJ. Two distinct mechanisms of silencing by the KvDMR1 imprinting control region. EMBO J. 2008; 27: 168–178. pmid:18079696
  64. 64. Yatsuki H, Joh K, Higashimoto K, Soejima H, Arai Y, Wang Y, et al. Domain regulation of imprinting cluster in Kip2/Lit1 subdomain on mouse chromosome 7F4/F5: large-scale DNA methylation analysis reveals that DMR-Lit1 is a putative imprinting control region. Genome Res. Cold Spring Harbor Lab; 2002; 12: 1860–1870. pmid:12466290
  65. 65. Mohammad F, Pandey GK, Mondal T, Enroth S, Redrup L, Gyllensten U, et al. Long noncoding RNA-mediated maintenance of DNA methylation and transcriptional gene silencing. Development. 2012; 139: 2792–2803. pmid:22721776
  66. 66. Zhang H, Zeitz MJ, Wang H, Niu B, Ge S, Li W, et al. Long noncoding RNA-mediated intrachromosomal interactions promote imprinting at the Kcnq1 locus. J Cell Biol. 2014; 204: 61–75. pmid:24395636
  67. 67. Schulz R, Menheniott TR, Woodfine K, Wood AJ, Choi JD, Oakey RJ. Chromosome-wide identification of novel imprinted genes using microarrays and uniparental disomies. Nucleic Acids Res. 2006; 34: e88. pmid:16855283
  68. 68. Kato R, Shirohzu H, Yokomine T, Mizuno S, Mukai T, Sasaki H. Sequence-ready 1-Mb YAC, BAC and cosmid contigs covering the distal imprinted region of mouse chromosome 7. DNA Res. 1999; 6: 401–405. pmid:10691133
  69. 69. Cerrato F, Sparago A, Di Matteo I, Zou X, Dean W, Sasaki H, et al. The two-domain hypothesis in Beckwith-Wiedemann syndrome: autonomous imprinting of the telomeric domain of the distal chromosome 7 cluster. Hum Mol Genet. 2005; 14: 503–511. pmid:15640248
  70. 70. Nagy A, Rossant J, Nagy R, Abramow-Newerly W, Roder JC. Derivation of completely cell culture-derived mice from early-passage embryonic stem cells. Proc Natl Acad Sci USA. 1993; 90: 8424–8428. pmid:8378314
  71. 71. Ewing B, Hillier L, Wendl MC, Green P. Base-calling of automated sequencer traces using phred. I. Accuracy assessment. Genome Res. 1998; 8: 175–185. pmid:9521921
  72. 72. Ramakers C, Ruijter JM, Deprez RHL, Moorman AFM. Assumption-free analysis of quantitative real-time polymerase chain reaction (PCR) data. Neurosci Lett. 2003; 339: 62–66. pmid:12618301
  73. 73. Mamo S, Gal AB, Bodo S, Dinnyes A. Quantitative evaluation and selection of reference genes in mouse oocytes and embryos cultured in vivo and in vitro. BMC Dev Biol. 2007; 7: 14. pmid:17341302
  74. 74. MacIsaac JL, Bogutz AB, Morrissy AS, Lefebvre L. Tissue-specific alternative polyadenylation at the imprinted gene Mest regulates allelic usage at Copg2. Nucleic Acids Res. 2012; 40: 1523–1535. pmid:22053079
  75. 75. Dobin A, Davis CA, Schlesinger F, Drenkow J, Zaleski C, Jha S, et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics. 2012; 29: 15–21. pmid:23104886
  76. 76. Quinn J, Kunath T, Rossant J. Mouse trophoblast stem cells. In: Soares MJ, Hunt JS, editors. Placenta and trophoblast. Methods in molecular medicine; 2006. pp. 125–148.