Loss of DNA methylation disrupts syncytiotrophoblast development: Proposed consequences of aberrant germline gene activation

DNA methylation is a repressive epigenetic modification that is essential for development and its disruption is widely implicated in disease. Yet, remarkably, ablation of DNA methylation in transgenic mouse models has limited impact on transcriptional states. Across multiple tissues and developmental contexts, the predominant transcriptional signature upon loss of DNA methylation is the de‐repression of a subset of germline genes, normally expressed in gametogenesis. We recently reported loss of de novo DNA methyltransferase DNMT3B resulted in up‐regulation of germline genes and impaired syncytiotrophoblast formation in the murine placenta. This defect led to embryonic lethality. We hypothesize that de‐repression of germline genes in the Dnmt3b knockout underpins aspects of the placental phenotype by interfering with normal developmental processes. Specifically, we discuss molecular mechanisms by which aberrant expression of the piRNA pathway, meiotic proteins or germline transcriptional regulators may disrupt syncytiotrophoblast development.


INTRODUCTION
DNA methylation is predominantly a repressive epigenetic modification, most commonly observed in mammals as the addition of a methyl group onto a cytosine in a CpG context, although non-CpG methylation is also present at low levels.DNA methylation exerts its repressive function at promoters and regulatory regions through the recruitment of epigenetic "readers," such as methyl-binding proteins, or through its ability to sterically interfere with transcription factor binding. [1][2][3] In early mammalian embryogenesis, DNA methylation undergoes widespread reprogramming, largely erasing the DNA methylation patterns inherited from gametes to reset the genome-wide landscape This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.© 2023 The Authors.BioEssays published by Wiley Periodicals LLC. after implantation. [4,5]During the post-implantation programming event, DNA methylation is established de novo, through complementary activity of DNA methyltransferase (DNMT) enzymes DNMT3A and DNMT3B, as well as DNMT3L in trophoblast cells, [6][7][8] which are specifically upregulated during this phase of development. [9][9] Embryonic cell types become highly methylated throughout the genome with the exception of regulatory elements and CpG islands.10] This feature appears BioEssays.2024;46:2300140.
wileyonlinelibrary.com/journal/bies 1 of 13 https://doi.org/10.1002/bies.202300140F I G U R E 1 (A) DNA methylation in embryonic day (E)7.5 embryonic epiblast is highly methylated across the genome, with the exception of regulatory regions such as promoters and CpG islands.Extra-embryonic trophoblast cells exhibit a partially methylated state throughout much of the genome with intermittent highly methylated regions and unmethylated regulatory regions.Dnmt3b knockout (KO) trophoblast shows a global partial depletion of DNA methylation.Genes are annotated at the bottom in grey, CpG islands in green, with promoters denoted by vertical grey shaded bars.(B) Upon loss of DNMT3B, trophoblast cells exhibit a de-repression of a subset of germline genes.Dnmt3b KO placentas at E12.5 show impaired placental labyrinth development, with reduced vascular branching and aberrant formation of the adjacent syncytiotrophoblast layer 2 (SynT-II).This placental defect leads impaired embryo development and embryonic lethality by E13.5. [8] be conserved across mammals and has been observed in some species of marsupials. [11,12]Hence, evidence supports that the unique DNA methylation state of extra-embryonic lineages is developmentally important, but its function remains poorly understood.
Characterization of knockout (KO) mouse models for Dnmt3a, Dnmt3b and Dnmt3a/3b have demonstrated that post-implantation programming of DNA methylation is essential for embryo survival.
Dnmt3a KO embryos are born but cannot survive past weaning, while Dnmt3b KO embryos die at around E13.5 with impaired growth and developmental abnormalities. [6]Yet, single cell RNA-sequencing studies have identified few defects in differentiation or cell identity in these embryos at E8.5. [13,14]Embryos lacking both DNMT3A and DNMT3B exhibit more severe morphological abnormalities and die by E10.5. [6]cently, we revisited these models, showing that Dnmt3b KOs fail to properly form the placental labyrinth, the interface between maternal and fetal circulations for nutrient, waste and gas exchange, and that this placental defect underpins the embryonic lethality observed in this model (Figure 1B). [8]Together, these findings demonstrate that de novo DNA methylation is essential for the development of both embryonic and extra-embryonic lineages, but the precise causative molecular changes that lead to these phenotypes remains elusive.
][16][17] Notably, while some models also show de-repression of transposons and/or imprinted genes, which can also be regulated by DNA methylation, these domains were not significantly dysregulated in the Dnmt3b KO placental trophoblast cells. [8,15]Comparing studies utilizing bulk RNAsequencing datasets, there appears to be a "core" set of methylationsensitive germline genes, including key components of the piRNA pathway, meiosis and gametic transcript regulators from across the stages of gametogenesis (Figure 2A and 2B, Table S1).
During embryogenesis, the majority of germline gene promoters appear to uniquely require a variant form of polycomb repressive complex 1 (PRC1.6)to target these domains in the pre-implantation embryo, which then permits targeting of de novo DNMTs in the postimplantation embryo (Figure 2D).The mechanisms underpinning this requirement are still not fully understood, but involve the deposition of repressive marks H2AK119ub and H3K9me3 by RING1B and SETDB1, respectively, and the maintenance of transcriptional repression and exclusion of H3K4me3 by ZMYM2. [18,19]PRC1.6 is targeted to germline promoters by DNA-binding proteins, E2F6 and MGA-MAX heterodimer, that bind CpG-containing motifs enriched within these regions. [20]][26][27][28] Although there are varying severities observed across PRC1.6 KO models, many complex components have additional roles and/or functional redundancies that could explain some of this variability.
Few studies have explored the cellular consequences of derepression of germline genes, except in cancers where they are widely expressed and are considered a cancer hallmark. [29]In the context of development, there are many key processes that could be negatively impacted by these gene products.Thus, we hypothesize that the de-repression of germline genes contributes to the developmental phenotypes observed in Dnmt and PRC1.6 KOs.In this Hypothesis article, we discuss developmental processes particular to placental formation that could be perturbed by the aberrant de-repression of germline genes seen in the Dnmt3b KO, as this phenotype underpins the embryonic lethality observed in this model [8] (Figure 1B).We propose three potential molecular mechanisms that could individually or combinatorially drive impair placental labyrinth development (Figure 3), presenting evidence from the literature and experiments to test these mechanisms.

MECHANISM 1: piRNA SILENCING OF PLACENTAL ERVs
Endogenous retroviruses (ERVs) are remnants of retroviral genetic material found throughout the genome.ERVs contain long terminal repeats (LTRs) that act to transcribe viral genes, but over time the associated viral genes of ERVs are often lost through recombination and accumulating mutations.The remaining LTRs can still contain sites for transcription factor (TF) binding, transcription initiation, splicing and termination, and hence can be readily incorporated into the gene regulatory network in a variety of ways. [30][35] Tissue-specific co-optivity of LTRs in the placental lineage has been linked to the presence of binding sites for the core trophoblast transcription factors ELF5, CDX2, and EOMES within certain classes of ERVs that are highly expressed in trophoblast stem cells. [31]Targeted deletion of a subset of these LTRs has been shown to disrupt expression levels of nearby genes, [32,36] demonstrating their functional importance in the trophoblast gene regulatory network.
The retroviral-derived Syncytin genes have convergently evolved to be an essential part of placental development in primates [37][38][39] and rodents. [40,41]Syncytins mediate the fusion and formation of the multinucleated syncytiotrophoblast and these single-pass transmembrane proteins have retained an almost identical structure to their viral envelope protein (env) origins.In mouse, Syncytin genes are expressed in the trophoblast compartment in a distinct niche of syncytialisationcompetent precursor cells giving rise to SynT-I and SynT-II in the placenta.The SynT-I layer, adjacent to maternal blood in the mature labyrinth, is derived from Syna-expressing precursors, while the SynT-II layer adjacent to the fetal vasculature, arises from Synb-expressing precursors. [40,41]While both Syna and Synb mutant placentae show minimal disruption to the gross labyrinthine morphology, the SynT-I or SynT-II cells are unable to fuse, respectively. [41,42]Syna mutant embryos do not survive past mid-gestation and Synb mutant neonates show impaired growth and are recovered at sub-mendelian ratios. [41]nce, retroviral-mediated fusion of the syncytium is necessary for placental function and healthy pregnancy.
In the germline, mobility of ERVs could cause deleterious mutations to be transmitted to offspring, hence there are layers of mechanisms to repress ERVs to protect the germline genome.One of these mechanisms in spermatogenesis is the piRNA pathway.[45][46][47][48][49] In mouse, there are three PIWI proteins (PIWIL1, PIWIL2, and PIWIL4) that bind to short (24-30 bp) PIWI-interacting RNAs (piRNAs).Although piRNA biogenesis varies across fetal and adult spermatogenesis, it can be generalized into two main stages.Primary piRNA biogenesis involves the processing of piRNAs from larger transcripts, mediated by PIWIL2 and other cofactors (e.g., MOV10L1 and ASZ1) (Figure 3).Secondary piRNA biogenesis requires the assembly and function of piRNA-PIWI protein complexes (detailed in [50,51] ) to repress complementary genes or transposons post-transcriptionally ('ping-pong cycle') or through epigenetic silencing of these genomic regions.There is a wide diversity of piRNAs detectable during spermatogenesis, and although a large subset arise from repetitive elements, many are from unique genomic regions as well. [50]Beyond silencing of ERVs, aberrant activity of piRNAs in placenta may have functional consequences, as it has been reported that genes important in placental function are among the putative targets of eutherian piRNA clusters. [52] the Dnmt3b KO trophoblast cells, several primary piRNA biogenesis components, including Piwil2, Gpat2, Asz1, and Mov10l, exhibit detectable gains in expression (Figure 2C). [8]][17] A core set of 7 germline genes were common amongst all datasets, with 57 methylation-sensitive genes identified in at least 4 studies (Table S1).(B) Gene ontology analysis, using DAVID, [113,114] of the 57 methylation-sensitive genes revealed highly significant enrichment for spermatogenesis, oogenesis and germline pathways.The top 25 gene ontology terms are listed in order of significance based on Benjamini-Hochberg corrected p-value.(C) Heatmap shows the relative expression of selected methylation-sensitive germline genes involved in the piRNA pathway meiosis, and gametic transcript regulation, grouped by function.Relevant methylation-sensitive germline genes were identified using gene ontology terms reported in DAVID [114] for combined gene lists from Table S1 and those upregulated in Dnmt3a/b DKO trophoblast cells. [8]Gene expression was quantified for Dnmt3a/b DKO, Dnmt3b KO, Dnmt3a KO and wildtype (WT) E7.5 extra-embryonic ectoderm as normalized reads per million (RPM), using published RNA-sequencing data. [8]Genes on the X chromosome were excluded as samples were not sex-matched.Abbreviations: chromatin -chromatin modifiers, TFs -transcription factors.(D) Promoters of germline genes are targeted for silencing during early embryogenesis by two mechanisms.In the pre-implantation embryo (left), when the majority of the genome is demethylated, a variant form of polycomb repressive complex 1 (PRC1.6)targets germline promoters through its interaction with CpG-containing DNA motifs.Components of this complex then modify adjacent histone tails, including the addition of H2AK119ub and H3K9me2/3.In the post-implantation embryo (right), DNMT3A and DNMT3B target DNA methylation to germline gene promoters.The localization and activity of PRC1.6 is necessary for subsequent targeting of DNMTs.Fkbp6 (Figure 2C).Together with high baseline expression of many of the other primary piRNA biogenesis components in trophoblast cells, [8] these data suggest that the piRNA pathway may indeed be active in Dnmt3b KO placentas.Small RNA sequencing in Dnmt3b KO placentas would shed light on whether there is bona fide piRNA biogenesis and identify potential targets.In principle, the mis-expression of the piRNA pathway could interfere with the placental gene regulatory program through silencing of genomic ERVs or targeted degradation of ERV-derived or other transcripts (Figure 3).Assaying enhancer histone modifications at key ERVs would establish whether there is altered ERV function at the genomic level in Dnmt3b KO placenta.The induction of SynT-II formation in Dnmt3b KO trophoblast stem cells using canonical WNT agonist will inform whether the impairment in SynT-II formation in Dnmt3b KO placentas is related to altered Synb expression and disrupted cell fusion. [53]These investigations would provide an initial indication as to whether PIWI-mediated repression may be important in the Dnmt3b KO phenotype.

MECHANISM 2: DISRUPTION OF CELL CYCLE BY MEIOTIC PROTEINS
Cell proliferation in early embryogenesis is linked to key morphogenic events and hence is highly regulated.Aberrant expression of meiosisspecific proteins during mitosis has been proposed to interfere with its normal progression due to the fundamental differences in alignments of sister chromatids versus homologous pairs along the metaphase plate and recombination events in mitosis versus meiosis I (Figure 3).Therefore, de-repression of meiotic proteins in Dnmt KOs could result in impaired embryogenesis due to delays in cell division or excess cell death due to DNA damage or genomic instability.Through such a mechanism, the loss of DNA methylation could lead to developmental defects with relatively little effect on cell identity. [8,13,14]Developmental processes, such as induction of branching morphogenesis, may be particularly susceptible to the depletion of progenitors, which critically underpins the formation of the placental labyrinth.
In cancer studies, it is evident that de-repression of synaptonemal complex (SC) proteins can contribute to DNA damage, genomic instability and delays in cell cycle progression.SC proteins mediate the pairing of homologous chromosomes for recombination during meiosis prophase I and are consistently de-repressed upon loss of DNA methylation in mitotic cells (Figure 2).Accumulation of SC components, such as SYCP1 and SYCP3, can form aggregates without other members of the SC present. [54,55]In cancer cells, expression of SYCP3 has been shown interfere with double-strand break repair by homologous repair [56] and induced expression of SYCP3 in HeLa cells after irradiation resulted in synaptonemal-like structures, impacting proliferation. [57][60] Similarly, another SC member, TEX12, when aberrantly expressed in mitotic cancer cells has been shown to lead to genomic instability through centrosome dysfunction. [61]e placental labyrinth defect observed in the Dnmt3b KO was characterized by disrupted formation of syncytiotrophoblast layer II (SynT-II). [8]SynT-II itself does not divide, rather labyrinth progenitor cells exit the cell cycle, induce Synb expression in response to chorioallantoic fusion, and fuse to form a multi-nucleated syncytium (Figure 3). [62]Nuclei within the syncytium are thought to be stably in G0. [63] The disruption to the SynT-II layer in Dnmt3b KO placentae could suggest that these progenitors may be acutely sensitive to disruption in cell cycle progression or depletion of cell number.It has been shown that cycling cells aberrantly expressing Syncytins causes disorganized, unstable syncytia to form in vitro. [62]This mechanism could be explored by staining for markers of DNA damage (e.g., H2A.X) and cell division (e.g., Ki-67) in DNMT3B null placentae to assess both labyrinth progenitor populations and the forming SynT-II layer. [64,65]storation of the placental methylome in the Dnmt3b conditional KO (cKO), using Sox2-cre, which results in a genetically wildtype placenta, rescues the SynT-II defect and consequently the mid-gestation lethality of Dnmt3b null embryos [8] (Figure 4A).The corresponding Dnmt3b cKO embryos were consistently found to be ∼70% weight of their WT littermates at E18.5, being proportionally smaller with no obvious gross morphological differences.Although not attributed to de-repression of SC proteins, DNMT3A gain-of-function mutations in humans cause microcephalic dwarfism accompanied by altered DNA methylation patterns; the proposed mechanism is the depletion of progenitor pools in development due to reduced rates of cell division. [66][69][70] Together, these studies support that disruption of cell cycle progression in development impacts the abundance of progenitor cell populations in the embryo.With the rescue of placental morphology and apparent function in the Dnmt3b cKO model, reduced organism size potentially due to DNMT3B-dependent disrupted proliferation, can be further explored.

MECHANISM 3: GERMLINE TRANSCRIPTIONAL REGULATORS DISRUPT INDUCTION OF PLACENTAL BRANCHING MORPHOGENESIS
In murine development, the morphogenesis and differentiation of syncytiotrophoblast in the placental labyrinth is coordinated by localized and temporal expression of TFs in response to several developmental signaling pathways, including WNT and FGF.The formation of the placental labyrinth is initiated by chorioallantoic attachment (fusion).The allantois, a projection of extra-embryonic mesoderm cells, attaches to the chorion, resulting in the involution of the trophoblast in a wavelike pattern at the interface between the two cell types (Figure 4A).
At the tips of the involuting trophoblast, GCM1 is expressed, [71] which directly regulates expression of fusion protein SYNB in SynT-II progenitor cells.
4][75][76] Several components of the WNT pathway are expressed during The formation of the placental labyrinth in murine development initiates with chorioallantoic attachment.The allantois is a projection of epiblast-derived mesodermal cells that comes into contact with the chorionic trophoblast cells.The allantoic mesoderm branches into the chorion at defined regions, resulting in enfolding of the trophoblast as it begins to differentiate and syncytialise in response to allantoic signaling queues.The allantoic mesoderm then differentiates into the fetal placental vasculature, as the closing of the ectoplacental cavity results in crosstalk between the SynT-II layer and the adjacent differentiating of SynT-I and sinusoidal trophoblast giant cells, forming the labyrinth zone.Dnmt3b KO and conditional KO embryos and placental phenotypes are shown in the greyed panels, with darkened colors depicting the tissues lacking DNMT3B in each model.In the Dnmt3b KO, there is impaired formation of the SynT-II layer and the underlying fetal vasculature, resulting in defects in embryo growth and development, and embryonic lethality by E13.5.In the Dnmt3b conditional KO using Sox2-Cre, all embryonic cell types derived from the epiblast are lacking DNMT3B, while the placenta is genetically wildtype.Dnmt3b conditional KO placentas are phenotypically normal, resulting in an improved embryo morphology and survival until E18.5. [8](B) Knockout mouse models have been instrumental in identifying key regulators of chorioallantoic attachment, branching morphogenesis and SynT-II formation, [115,116] some of the key regulators are highlighted here.Chorionic trophoblast cells express some trophoblast stem cell markers, indicative of their progenitor state.WNT and FGF signaling from the allantois to the chorionic trophoblast upon chorioallantoic attachment are critical for labyrinth formation.For example, labyrinth defects of varying severities have been observed in knockouts for Wnt7b and its receptor Fzd5, FGF receptor Fgfr2 and downstream effectors in SynT-II formation, such as Gcm1, Ctnnb1, and Tcf4.chorioallantoic fusion (E8.0-E8.5),with Wnt7b showing trophoblastspecific expression in the chorion with sustained expression in the labyrinth to at least E10.5. [73]WNT7B is required for chorioallantoic attachment and subsequent development of the labyrinth. [73,77]Targeted disruption to WNT2 during gestation causes reduction in size of the labyrinth with enlarged lacunae of maternal blood and reduced fetal vasculature. [72]Conversely, knockout of the WNT receptor Fzd5 revealed a positive feedback loop with GCM1, in which the expression of one enforces the expression of the other during this window of chorioallantoic fusion. [75,76]Fzd5 KO placenta show a complete loss of branching morphogenesis upon chorioallantoic attachment [75] with no induction of Gcm1. [76]Furthermore, WNT targets, BCL9L and sumoylated TCF4 along with active CTNNB1 (β-catenin) regulate expression of GCM1 and are also required for both branching mor-phogenesis and syncytialisation, particularly of SynT-II, in the murine placenta. [74]Fusion of the human choriocarcinoma cell line, BeWo, can be induced by the cAMP agonist, Forskolin, [78] via activation of βcatenin/BCL9L/TCF. [74]This signaling cascade can be recapitulated in vitro using murine trophoblast stem cells, with canonical WNT agonist, CHIR99021, used to induce differentiation of SynT-II. [53]In turn, the modulation of WNT signaling is critical for the differentiation of the junctional zone, with WNT inhibitors SFRP1 and SFRP5 playing a critical role in its formation. [77]These results highlight that there is a highly regulated response to and fine-tuning of WNT signaling during placental development that is essential for proper morphogenesis and cellular differentiation.
There is evidence that loss of DNA methylation impairs branching morphogenesis in contexts beyond the placenta.Lung endoderm-specific deletion of DNMT1 leads to disruption of branching morphogenesis in this tissue with loss of cell polarization and precocious differentiation. [16]81][82][83] Similar to what was observed upon loss of the de novo DNMTs, loss of DNA methylation maintenance enzyme, DNMT1 leads to upregulation of germline genes (Figure 2A). [16]And similar to the placenta, disrupting WNT and FGF signaling in the lung impairs branching morphogenesis.Conditional knockout of the WNT ligand receptor, FZD2, in the lung endothelium results in defective branching of the tissue. [79]While expression of a dominant-negative form of FGFR2 in the lungs also leads to a block in branching and perinatal lethality due to asphyxiation in mice. [84]Expression of a truncated form of FGFR2 leads to embryonic lethality by E11.5 with a failure of chorioallantoic fusion, supporting that some key regulators of branching morphogenesis are common between these two tissues. [85]The transcriptional response to WNT signaling may be regulated directly by DNA methylation, but it is highly likely that aberrant expression of several classes of transcriptional regulators from the germline may impair normal processes of regulating transcript abundance, stability or translation.
Germline transcription factors and chromatin remodelers are among those genes deregulated by the loss of DNA methylation across studies.Such genes include PHD finger protein 20 (Phf20) and spermatogenesis-and oogenesis-specific basic helix-loop-helix (bHLH) 2 (Sohlh2) (Figure 2C).PHF20 is a member of the chromatin modifying complex, NSL (nonspecific lethal), [86] and has been demonstrated to be a chromatin reader of H3K4me2, leading to deposition of H4K16ac and hence is thought to be a transcriptional activator. [87][90] The aberrant expression of such genes could be reinforcing aspects of a germline transcriptional program in affected cells.Such changes could have a wider impact on the chromatin architecture of effected cells leading to not only changes in the transcriptional output but also higher order chromatin conformation, potentially altering loci interactions, enhancer contacts and domain structures.Consequently, these changes could have a knock-on effect on whether cells can respond either epigenetically or transcriptionally to stimulating signals, such as during chorioallantoic fusion (Figure 3).Several genes associated with RNA regulation and translation were sensitive to the loss of DNA methylation and were among those dysregulated in Dnmt3b KO trophoblast, such as ribosomal protein L10-like (Rpl10l) (Figure 2C).Rpl10l and Rpl39l are paralogues of X-linked ribosomal protein genes, Rpl10 and Rpl39, respectively, and normally exhibit testis-specific expression. [91,92]Changes to the expression of ribosomal proteins could impact the fine-tuning in the translational capacity of effected cells as the ribosomal repertoire is altered [93,94] (Figure 3).Conversely, post-transcriptional regulation of mRNAs via direct RNA binding may be impacted.Both DAZL and ZAR1 have been shown to be RNA binding proteins with a role in meiosis.[100] ZAR1 is a translational activator, binding mRNAs in the cytoplasm of mature GV mouse oocyte via its zinc-finger domain, [101] and is also required for completion of meiosis and fertilization. [101,102]Thus, aberrant expression of germline RNA-binding proteins may alter the mRNA stability and repertoire in affected cells (Figure 3).While these germline transcriptional regulators are almost unanimously dysregulated upon loss of DNA methylation, Dnmt3b KO trophoblast cells showed de-repression of subset (Figure 2C).To directly test whether placental branching morphogenesis is impacted by any of these classes of germline transcriptional regulators, these genes could be ectopically expressed together or individually in trophoblast stem cells to determine whether these cells can then appropriately differentiate to SynT-II in response to canonical WNT induction. [53]

CONCLUSIONS AND PERSPECTIVES
][14][15][16] Due to the clear role for DNA methylation at imprinted domains and on the silent X chromosome [103,104] and its cell type-specific patterning, [105] it is anticipated that DNA methylation plays a role in regulating gene expression across the genome during differentiation.Thus, it is surprising that there would be so little impact on differentiation in embryos lacking the DNMTs. [13,14]DNMTs interact with modified histones when targeting the genome [106,107] with chromatin state being a strong predictor of DNA methylation, [108] supporting that there may be redundancies between repressive chromatin marks and DNA methylation in regulating the majority of transcriptional programs during development.It is evident from gene expression data from KO models that DNA methylation plays a non-redundant role in repressing germline genes across cell types throughout post-implantation development.We find there is substantial evidence to support that aberrant expression of germline genes can disrupt cellular function from cancer models.Thus, de-repression of these methylation-sensitive germline genes upon loss of DNA methylation in the Dnmt, or PRC1.6 KO models, could have profound consequences in embryogenesis.
In the context of placental development, we propose that aberrant expression of three classes of germline genes, including the piRNA pathway, meiotic proteins and/or transcript regulators, underpins aspects of the placental phenotype observed in the Dnmt3b KO model.Notably, gene expression analysis supports that the pathogenic effect of germline genes would be far more profound in the Dnmt3a/b DKO and may contribute to the increased severity of embryonic lethality at E10.5. [6,104]A detailed characterization of early labyrinth formation and in vitro trophoblast differentiation in the Dnmt3a/b DKO may provide additional mechanistic insights.
Investigating molecular mechanisms underpinning impaired placental development in general currently presents some challenges.
Functional investigations and gene targeting approaches can be performed using cultured trophoblast stem cells. [109]However, an important consideration in evaluating de novo DNMTs is that the methylome is already established in cultured trophoblast stem cells, so to obtain cells that reflect the DNA methylation defects acquired in in vivo KO models, newly derived cells would need to be obtained from postimplantation KO embryos.Conversely, to exclude pleiotropic effects contributing to placental phenotypes in vivo, a trophoblast-specific Cre would be invaluable in evaluating placentation in embryonic lethal models.However, while the Sox2-Cre is highly efficient in generating embryo-specific conditional KO models, [8,110] the field is currently lacking a pan-trophoblast Cre that is (1) highly penetrant, (2) exclusively expressed in trophoblast throughout development and (3) robustly activated well before the wave of post-implantation epigenetic programming.Finally, advances in single-cell and single-nuclei RNA-sequencing methods have recently provided new insights into gene regulation and cell identity in placental development. [111,112]wever, none of the current methods allows for a comprehensive assessment of all cell types in the mouse placenta, due to the presence of multi-nucleated syncytiotrophoblast and several populations of large, highly polyploid trophoblast giant cells.Innovations in these areas will be important for future mechanistic studies in placental development.

F I G U R E 3
We hypothesize that de-repression of germline genes upon loss of DNA methylation in the Dnmt3b KO may underpin aspects of the impaired placental labyrinth formation by interfering with normal developmental processes.The schematic summarizes three potential molecular mechanisms by which aberrant expression of the germline piRNA pathway, meiotic proteins or transcript regulators may disrupt syncytiotrophoblast development.(1) Spurious activation of the piRNA pathway could interfere with the placental gene regulatory program through silencing of genomic ERVs or targeted degradation of ERV-derived or genic transcripts.(2) De-repression of meiotic proteins in mitotic progenitors populations necessary for branching morphogenesis and SynT-II formation may result in an aberrantly prolonged cell cycle or excess cell death due to DNA damage or genomic instability.(3) Placental branching morphogenesis is regulated by transcription factor upregulation in response to key signaling pathways, including WNT and FGF, among others.Expression of germline transcript regulators, such as ribosomal proteins, RNA-binding proteins, transcription factors and chromatin modifiers, could result in transcriptional changes that impair the response to signaling events directing placental branching morphogenesis.