OVOL2 reinforces epithelial identity and promotes SynT-I differentiation in the mouse placenta

Abstract

The placenta establishes the maternal-fetal interface during pregnancy, enabling the selective exchange of nutrients and gases between maternal and fetal circulations. In humans and mice, this exchange surface is lined by multinucleated epithelial layers called the syncytiotrophoblast (SynT), which in the mouse placenta is organized into two distinct layers, SynT-I and SynT-II. SynT formation from trophoblast precursors requires epithelial plasticity to permit cell differentiation, fusion, and morphogenesis, while preserving epithelial integrity to maintain the maternal-fetal exchange barrier. However, the mechanisms that safeguard epithelial features during SynT development remain unclear. OVO-like 2 (OVOL2), a transcriptional repressor of mesenchymal-associated programs and key regulator of epithelial identity, is highly expressed in the mouse placenta and essential for its development. We hypothesized that OVOL2 promotes SynT lineage formation by restraining mesenchymal-associated transcriptional programs during trophoblast differentiation. To test this, placental development was examined following Ovol2^+/− matings, and wild-type and Ovol2-deficient trophoblast stem (TS) cells were analyzed under stem conditions or differentiated with CHIR99021 to enrich for SynT lineages. SynT-I lineage development was disrupted in both Ovol2-deficient placentas and differentiating TS cells, whereas SynT-II-associated differentiation appeared less severely affected. Chromatin profiling identified OVOL2 binding near genes associated with epithelial-to-mesenchymal transition, including Id1, Zeb1, and Vim, which were upregulated in Ovol2-deficient trophoblasts. Consistent with these observations, Ovol2-deficient cells showed elevated levels of mesenchymal markers such as ZEB1 and Vimentin and reduced levels of epithelial markers including E-cadherin. These findings identify OVOL2 as a critical regulator of SynT-I lineage formation and epithelial identity in the mouse placenta.

OVOL2 safeguards epithelial identity during trophoblast differentiation by restraining mesenchymal-associated gene programs, thereby supporting SynT-I development and formation of the maternal-fetal exchange surface.

Graphical Abstract

Graphical Abstract.

Introduction

The placenta is a specialized organ that evolved in viviparous mammals to serve as the interface between mother and fetus during pregnancy. One of its primary roles is to establish an epithelial barrier that separates maternal and fetal circulations, enabling the controlled exchange of oxygen and nutrients required for proper fetal growth and development. Disruption of the tightly regulated processes underlying placental formation can compromise both fetal and maternal health, and in severe cases lead to pregnancy loss [1]. Despite its importance, the regulatory mechanisms that control placental development and function remain poorly understood.

Hemochorial placentation is characterized by the direct apposition of maternal blood with the chorionic epithelium (trophoblast) and occurs in humans and many rodent species, including mice. In both humans and mice, several distinct trophoblast subtypes arise through a multilineage differentiation pathway, which can be modeled in vitro using trophoblast stem (TS) cells [2]. In mice, these cells differentiate into multiple lineages populating two anatomically and functionally distinct placental regions: the junctional and labyrinth zones. Cells within the junctional zone redirect nutrients to the conceptus by producing hormones that modulate maternal cardiovascular physiology and metabolism, and by promoting expansion of the uterine vasculature [3]. In contrast, the labyrinth zone contains two layers of syncytiotrophoblast (SynT-I and SynT-II), multinucleated syncytia formed through cell–cell fusion that establish the maternal-fetal exchange interface. Sinusoidal trophoblast giant cells (s-TGCs) are also present in the labyrinth zone and secrete prolactins that support pro-angiogenic signaling conducive to placental and embryonic development [4, 5].

When cultured in the presence of fibroblast growth factor 4 (FGF4), heparin, and activin A in medium conditioned by mouse embryonic fibroblasts (MEFs), mouse TS cells can be continuously maintained in a self-renewing state [6]. Under these conditions, they exhibit high expression of stem-associated genes required for their specification and multipotency, such as caudal type homeobox 2 (Cdx2), eomesodermin (Eomes), and estrogen-related receptor beta (Esrrb). Withdrawal of growth factors and MEF conditioned medium stimulates TS cells to differentiate primarily into junctional zone lineages, defined by high expression of placental lactogens (Prl3b1 and Prl3d1), achaete-scute family bHLH transcription factor 2 (Ascl2), and trophoblast-specific protein alpha (Tpbpa) [7, 8]. Formation of labyrinth zone lineages is generally less robust under these conditions; however, supplementing differentiation medium with the glycogen synthase kinase 3β (GSK3β) inhibitor CHIR99021 (CHIR) promotes differentiation toward labyrinthine trophoblasts, particularly SynT-II, characterized by high expression of syncytin b (Synb) and glial cells missing homolog 1 (Gcm1) [9]. No culture condition is currently known to specifically enrich for SynT-I or s-TGC lineages, which express syncytin a (Syna) and cathepsin Q (Ctsq), respectively, although these genes are induced to varying degrees in both protocols.

OVO-like 2 (OVOL2) is a member of the Ovo family of zinc finger transcription factors that serve as key regulators of epithelial differentiation and maintenance across diverse tissues. For example, in the epidermis, OVOL2 promotes keratinocyte differentiation and preserves epithelial identity by repressing epithelial-mesenchymal transition (EMT)-associated factors such as Zeb1 and Snai1 [10]. In the mammary gland, where branching morphogenesis involves transient epithelial plasticity, OVOL2 restricts full mesenchymal transformation during ductal morphogenesis by suppressing EMT-associated programs [11]. In thymic epithelial cells, OVOL2 epigenetically enforces postnatal epithelial identity and is required to support proper T cell maturation; its deficiency leads to lymphopenia and features associated with thymic involution [12]. OVOL2 also facilitates mesenchymal-to-epithelial transition during fibroblast reprogramming [13], and is frequently downregulated in metastatic carcinomas, correlating with mesenchymal switching and invasive behavior [14, 15]. Together, these studies highlight OVOL2 as a gatekeeper of epithelial differentiation, morphogenesis, and identity. In the mouse placenta, we and others have shown that Ovol2 is prominently expressed in the labyrinth zone, and that OVOL2 plays a critical role in trophoblast differentiation and placental development. Embryos lacking OVOL2 die around embryonic day (E)10.5 due to failed placental development and a breakdown of the placenta-heart axis [16–18]. Although OVOL2 is unequivocally important for placental development, the molecular mechanisms underlying its function in this context remain unclear.

Like several other epithelial differentiation paradigms, formation of SynT layers requires epithelial plasticity. On the one hand, trophoblast fusion is associated with partial EMT-like remodeling, including changes in junctional complexes and cytoskeletal organization that facilitate cell–cell fusion. On the other hand, preservation of epithelial gene expression programs is essential for establishing and maintaining the maternal-fetal barrier. Thus, given the key role of OVOL2 as a regulator of epithelial identity and differentiation, we investigated its function during trophoblast differentiation, with a focus on SynT lineage formation. We found that Ovol2-deficient placentas fail to efficiently form SynT-I lineages prior to embryonic lethality, while SynT-II-associated differentiation appears less severely affected. Moreover, whereas wild-type (WT) TS cells reinforce an epithelial phenotype when induced to differentiate, Ovol2-deficient TS cells upregulate mesenchymal-associated gene programs and display features consistent with impaired progression through a transitional differentiation state. Our findings identify OVOL2 as a key regulator of SynT-I differentiation and epithelial identity, highlighting the importance of epithelial plasticity during formation of the maternal-fetal interface.

Materials and methods

Animals

WT C57BL/6 and CD-1 mice were purchased from Charles River Laboratories (Senneville, QC, Canada). Ovol2^+/− mice, maintained on a C57BL/6 background, were generously provided by Seiji Ito through the RIKEN Bioresource Research Centre (Tsukuba, Ibaraki Prefecture, Japan) based on a Materials Transfer Agreement [16]. Mice were housed in a 12-h light–dark cycle with ad libitum access to food and water. Female and male mice (8–16 weeks old) were housed in the same cage overnight and the presence of a copulatory plug the following morning was defined as embryonic day (E) 0.5. CD-1 mice were used to generate MEFs needed for TS cell culture. Animal use was approved by the University of Western Ontario Animal Care and Use Committee in accordance with the guidelines of the Canadian Council on Animal Care.

Tissue collection

Following Ovol2^+/− × Ovol2^+/− matings, dams were sacrificed at E9.5 and whole conceptuses (including the embryo, extraembryonic tissues, and surrounding decidua) were collected. E11.5, E14.5, and E17.5 placentas were collected from WT matings. All tissues were fixed in 4% paraformaldehyde (Sigma-Aldrich, Oakville, ON, Canada) for 24 h at room temperature. Tissues were stored in 70% ethanol until standard histological processing, which included dehydration, clearing, and paraffin wax infiltration using a Leica TP1020 tissue processor (Leica Microsystems, Wetzlar, Germany). Tissues were then serially sectioned at 5 μm for subsequent analysis. Genotyping could not be performed on E9.5 tissues prior to fixation. Therefore, in situ hybridization was used to identify Ovol2-expressing (Ovol2⁺) and Ovol2^−/− conceptuses. Conceptuses exhibiting Ovol2⁺ signal could represent either Ovol2^+/+ or Ovol2^+/− genotypes, which develop comparably [17].

In situ hybridization

Tissue sections were stained for RNAScope in situ hybridization as per the manufacturer’s recommendations (Advanced Cell Diagnostics, Newark, CA, USA). Briefly, sections were deparaffinized with xylene, washed inanhydrous ethanol, treated with hydrogen peroxide, and rehydrated in distilled water. Sections were then incubated at 95°C in a citrate-based antigen retrieval solution, treated with protease, and hybridized with mouse-specific Ovol2 (558501, Advanced Cell Diagnostics), Syna (446701-C3, Advanced Cell Diagnostics), and Synb (57631-C2, Advanced Cell Diagnostics) probes. Sections were then subjected to a series of amplification steps, developed using RNAscope Multiplex FL v2 HRP, and counterstained with 4′,6-diamidino-2-phenylindole (DAPI). Finally, sections were dehydrated, mounted with Fluoromount G (Southern Biotech, Birmingham, AL, USA), and imaged using a Nikon Ti2e spinning disk confocal microscope (Nikon Instruments Inc., Minato City, Tokyo, Japan).

Cell culture

WT mouse TS cells (F4 line) were a generous gift from Janet Rossant (University of Toronto, Toronto, ON, Canada). Derivation of Ovol2^−/− TS cells was described previously [17]. Cells were maintained at 37°C with 5% CO₂ in TS cell media, which consisted of Roswell Park Memorial Institute (RPMI)-1640 media (Sigma-Aldrich) supplemented with 20% fetal bovine serum (Wisent, Saint-Jean-Baptiste, QC, Canada), 25 ng/ml FGF4 (R&D Systems, Minneapolis, MN, USA, 235-F4-025), 1 μg/ml heparin (Sigma-Aldrich, H3149), 10 ng/ml activin A (R&D Systems, 338-AC-10), 100 μM 2-mercaptoethanol (ThermoFisher Scientific), 1 mM sodium pyruvate (ThermoFisher Scientific), 100 units/ml penicillin, and 100 μM streptomycin (ThermoFisher Scientific). 70% of the media was preconditioned by mitomycin C-treated MEFs. Cells were passaged with 0.25% trypsin prior to confluency and routinely monitored to ensure the absence of mycoplasma contamination using MycoAlert (NC0529908, ThermoFisher Scientific). To induce differentiation, cells were cultured in medium lacking FGF4, activin A, heparin, and MEF preconditioning. Differentiating WT trophoblast cells were assigned to three groups: one treated with 3 μM CHIR99021 (CHIR; Stemgent, Cambridge, MA, USA), one treated with dimethyl sulfoxide (DMSO; 1:1000, ThermoFisher Scientific) as a vehicle control, and one left untreated. Cells were exposed to differentiation media for 3 to 7 days. All culture conditions included treatment-matched stem-state controls for statistical analysis. In subsequent experiments, WT and Ovol2^−/− TS cells were cultured under stem conditions or CHIR-treated differentiation conditions for up to 72 h.

Ectopic OVOL2 expression

To ectopically express Ovol2, mouse Ovol2 was subcloned into the pEF-GFP plasmid as described previously [17], and the resulting construct was transfected into Ovol2^−/− TS cells using jetPRIME transfection reagent (Polyplus Transfection, New York, NY, USA, 101000015). Briefly, 1.5 μl jetPRIME transfection reagent was diluted to 25 μl with JetBuffer (Polyplus Transfection, 201000003), and 1 μg pEF-OVOL2 was diluted to 25 μl in OptiMEM media (ThermoFisher Scientific), with solutions mixed after a 5-min incubation at room temperature. Plasmid mixtures were added to 200 μl of serum-free RPMI media, and complete RPMI media with 3 μM CHIR was added after 4 h. Transfected cells and untreated controls were cultured with CHIR-treated RPMI media for up to 72 h.

Immunohistochemistry

Serially-sectioned E9.5 conceptuses and E11.5, E14.5, and E17.5 placentas were deparaffinized using Histo-Clear (National Diagnostics, Atlanta, GA, USA), rehydrated through a graded series of ethanol washes, and incubated at 95°C in a citrate-based antigen retrieval solution. Sections were then blocked in 10% normal goat serum and incubated overnight at 4°C with antibodies targeting monocarboxylate acid transporter 1 (MCT1; 1:200, Sigma-Aldrich, AB1286-I), monocarboxylate acid transporter 4 (MCT4; 1:200, Santa Cruz Biotechnology, sc-376140, Dallas, TX, USA), E-Cadherin (1:400, Cell Signaling Technology, Danvers, MA, USA, 3195), and Vimentin (1:100, Cell Signaling Technology, 5741). Sections were then washed and incubated with species-specific secondary antibodies conjugated to Alexa Fluor-488 (1:250), Alexa Fluor-555 (1:250), and Alexa Fluor-647 (1:250) and counterstained with DAPI (1:10,000, ThermoFisher Scientific). After washing, sections were mounted using Fluoromount G, and slides imaged using a Nikon Ti2e spinning disk confocal microscope.

Immunocytochemistry

WT and Ovol2^−/− TS cells were cultured for 72 h under stem or CHIR-treated differentiation conditions, then fixed for 15 min in 4% paraformaldehyde and blocked in 10% normal goat serum. Cells were then incubated overnight at 4°C with antibodies targeting zonula occludens-1 (ZO-1; 1:100, ThermoFisher Scientific, 33-9100), ID1 (1:200, BioCheck, San Francisco, CA, USA, BCH-1/37–2), and ZEB1 (1:1000, Cell Signaling Technology, 70512). The next day, cells were incubated for 1 h with anti-mouse secondary antibodies conjugated to Alexa Fluor-488 (1:250), then counterstained with Dylight 554-conjugated phalloidin (Cell Signaling Technology, 13054) to visualize actin filaments, and DAPI to identify nuclei. Images were acquired using a Nikon Ti2e spinning disk confocal microscope. The fusion index was calculated using ImageJ software by quantifying the number of nuclei within ZO-1-positive cell borders [19]. Fused cells were defined as those containing two or more nuclei enclosed within a continuous cell border. For each biological replicate, three random fields of view containing at least 100 nuclei were analyzed.

Quantitative RT-PCR

TRA-zol (GeneBio Systems, Burlington, ON, Canada) was used to extract and purify RNA from cells following the manufacturer’s protocol. RNA was reverse transcribed into complementary DNA (cDNA) using a High-Capacity cDNA Reverse Transcription Kit (ThermoFisher Scientific). Samples were then mixed with Sensifast SYBR Green PCR Master Mix (FroggaBio, Toronto, ON, Canada) and primers detailed in Table 1. Quantitative PCR was performed on a CFX96 instrument (Bio-Rad Laboratories, Mississauga, ON, Canada), with an initial denaturation at 95°C for 10 min, followed by 40 cycles of 95°C for 15 s and 60°C for 1 min. A melt curve was performed at the end of the run to confirm product specificity. Relative transcript levels were calculated using the ΔΔCt method, as previously described [20]. The geometric mean of 2–4 constitutively expressed genes that showed no differences between conditions were used as reference RNAs (Table 2).

Table 1.

Primer sequences used for quantitative RT-PCR

Gene	Forward	Reverse	Accession number
Ascl2	GCCTATCGTCGGAGGAAAG	TAGCCCCCTAACCAACTGGA	NM_008554.3
B2m	ACGTAACACAGTTCCACCCG	TCTCGATCCCAGTAGACGGT	NM_009735.3
Cdh1	CGATTACGAGGGCAGTGGTT	TGTCCGCCAGCTTCTTGAAT	NM_009864.3
Cdh2	CGCAGTCTTACCGAAGGATG	TCACATTGAGAAGGGGCTGTC	NM_007664.5
Cdh3	TATACCGCACAGGACCCAGA	CCAGTTGGCTGGATCTCTCG	NM_007665.3
Cdh4	AACCAGGTCTACAACGGCTC	CACGATGTCCCCTGTTTCAC	NM_009867.4
Cdh5	ACTGCACATCTACGGCTACG	GTCGGAGGAATTGGTGCTCA	NM_009868.4
Cdx2	CCAGCTCTTTGCCTCTCTGT	TGCCTCTGGCTCCTGTAGTT	NM_007673.3
Ctsq	GGCTTTTCCTGTGACTGGTG	ACCCCAACGACAGCCTCTAT	NM_029636.3
Eef2	AAGCGCAAGGGCCTGAAAGA	AGCAAGTGAGTGCTGAGTGATG	NM_007907.3
Eomes	AACATGCAGGGCAATAAGATG	AGCCTCGGTTGGTATTTGTG	NM_010136.4
Esrrb	GGGAGCTTGTGTTCCTCATC	CTACCAGGCGAGAGTGTTCC	NM_011934.4
Gcm1	AACACCAACAACCACAACTCC	CAGCTTTTCCTCTGCTGCTT	NM_008103.3
Gjb3	TGTACGTTCTCCACACGCTC	ACTTGCCCTTGCTTATGCCT	NM_008126.3
Hprt	TGCTGACCTGCTGGATTACA	CATTTTGGGGCTGTACTGCTT	NM_013556.2
Id1	AAAGACACTACCGCAGCCA	GTCTCTGGAGGCTGAAAGGT	NM_010495.3
Krt7	ACCAGCGTGCCAAGTTAGAG	GAGCATCCTTGATTGCCAGC	NM_033073.3
Ovol2	CTTCGACCTGAAGAGGCACG	GTGTAGCCGCAATCCTCACA	NM_152947.3
Ppia	TGGCAAGCATGTGGTCTTTGG	TAGAGCTGTCCACAGTCGGAA	NM_008907.2
Prl3b1	CCAACGTGTGATTGTGGTGT	TCTTCCGATGTTGTCTGGTG	NM_008865.4
Prl3d1	TGGTGTCAAGCCTACTCCTTT	CAGGGGAAGTGTTCTGTCTGT	NM_172156.3
Syna	CCCTTGTTCCTCTGCCTACTC	TCATGGGTGTCTCTGTCCAA	NM_001013751.2
Synb	TGAACACCCCAACTGAGCAA	GGACGAAACAGAGGACCCAA	NM_173420.3
Tpbpa	CGGAAGGCTCCAACATAGAA	TCAAATTCAGGGTCATCAACAA	NM_009411.5
Ubc	GACGTACCTTCCTCACCACAG	TCACACCCAAGAACAAGCACA	NM_019639.4
Vim	CAGCAGTATGAAAGCGTGGC	GCTCCAGGGACTCGTTAGTG	NM_011701.4
Ywhaz	CCCTCTTGGCAGCTAATGG	GGAGTAAGGGGAAAATGTGGG	NM_011740.3
Zeb1	AAACCGATGAGCAGCCTCAA	ACACAAGAGTAACCCTGCGG	NM_011546.3

Table 2.

Reference genes used for quantitative RT-PCR

Experiment	Reference genes
3–7 day differentiation	B2m, Eef2, Ppia, Ubc
24–72 h CHIR differentiation	Hprt, Ubc, Ywhaz
Stem and 72 h CHIR differentiation	B2m, Ubc
Ovol2 overexpression	Eef2, Ubc

Western blotting

Cells were lysed using radioimmunoprecipitation assay buffer (50 mM Tris, 150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate, and 0.1% sodium dodecyl sulfate [SDS]) supplemented with a protease inhibitor cocktail (Sigma-Aldrich). Cell lysates were sheared with a 27G needle, and protein concentrations quantified using a bicinchoninic acid assay (Bio-Rad Laboratories). Approximately 25 μg of total protein was diluted in 4× loading buffer (0.25 M Tris, 8% SDS, 30% glycerol, 0.02% bromophenol blue, 0.3 M dithiothreitol), incubated at 60°C for 10 min, and separated with SDS-polyacrylamide gel electrophoresis. Proteins were transferred to polyvinylidene difluoride membranes, and blocked in 3% bovine serum albumin diluted in tris-buffered saline with 1% Tween-20 (TBST) for 1 h with agitation. Blots were then probed for ZEB1 (1:1000, Cell Signaling Technology, 70512), E-Cadherin (1:1000, Cell Signaling Technology, 14472), Vimentin (1:1000, Cell Signaling Technology, 5741), and β-actin (1:1000, Santa Cruz Biotechnology, Dallas, TX, USA, sc-47778) overnight with agitation at 4°C. The following day, membranes were washed with TBST and incubated with species-specific infrared-conjugated secondary antibodies (1:10 000) for 1 h at room temperature with agitation. After washing, protein abundance was measured using a LI-COR Odyssey imaging system (LI-COR Biosciences, Lincoln, NE, USA).

Chromatin immunoprecipitation sequencing

WT mouse TS cells were cultured under differentiation conditions with 3 μM CHIR for 3 days. Cells were then fixed for 10 min in 0.7% formaldehyde, lysed, and sonicated with a Bioruptor (Diagenode, Sparta, NJ, USA). An aliquot corresponding to 1% of the total sonicated chromatin was set aside as an input control, representing total chromatin prior to immunoprecipitation. This input sample was processed in parallel and used to normalize enrichment in subsequent analyses. The remaining lysate was evenly split and incubated overnight with either 5 μg OVOL2 antibody (GeneTex, Irvine, CA, USA, GTX120220) or 5 μg rabbit IgG negative control (Cell Signaling Technology, 2729) at 4°C with agitation. Immunoprecipitated lysates were collected using Sepharose G beads (Sigma-Aldrich, 17061801), DNA fragments eluted, and replicates from 3 independent experiments were submitted to the London Regional Genomics Centre for sequencing (London, ON, Canada). DNA integrity and concentration were assessed using a Qubit fluorometer (ThermoFisher Scientific), and libraries generated with a NEBNext Ultra II DNA Library Prep Kit (New England Biolabs, Whitby, ON, Canada). Sequencing was performed on an Illumina NextSeq 500 (Illumina, San Diego, CA, USA) using 75 bp paired-end reads, generating at least 30 million reads per sample. Adapter sequences, overrepresented sequences, and sequences with a phred score below 30 were removed using TrimGalore-0.6.10 [21]. Sequences were aligned to the GRCm39 mouse genome using Bowtie2 (version 2.5.4), and peaks were called using MACS3 (version 3.0.3) with an adjusted P < 0.05 cutoff [22, 23]. Sequences from each input sample were merged together, as were OVOL2 fractions prior to data visualization, which was performed using deepTools (version 3.5.6) for heatmap generation and the MEME Suite (version 5.5.9) for motif analysis [24, 25]. All analyses were performed in R (version 4.5) using packages from Bioconductor (version 3.22). The packages ChIPseeker (version 1.46.1) and ClusterProfiler (version 4.18.4) were used to annotate genomic peak locations and gene ontology (GO) analysis, respectively; GO analyses were visualized using enrichplot (version 1.30.4) [26, 27]. Ontologies were identified using the top 100 peaks by fold enrichment (excluding peaks within intergenic regions), with an adjusted P < 0.05 considered statistically significant. Raw and processed data have been deposited in the NCBI Gene Expression Omnibus under accession number GSE324788.

RNA sequencing

WT and Ovol2^−/− mouse TS cells were cultured under differentiation conditions with 3 μM CHIR for 3 days. Cells were lysed using TRA-zol, and the aqueous phase was placed on RNeasy columns (Qiagen, Toronto, ON, Canada). RNA was treated with DNase I to remove residual DNA contamination and subjected to column purification based on the protocol provided by the manufacturer. Four samples from each group were submitted to Génome Québec (Montreal, QC, Canada). RNA integrity was assessed using a Bioanalyzer (Agilent Technologies, Santa Clara, California, USA), with all samples achieving an RNA integrity number greater than 9.0. Libraries were generated using the NEBNEXT Ultra RNA Library Prep Kit for Illumina (New England Biolabs) and sequenced using a NovaSeq 600 sequencer (Illumina, San Diego, CA, USA) with 100 bp paired-end reads, to obtain at least 35 million reads per sample. The first three bases from the 3′ end, as well as adapter sequences, overrepresented sequences, and sequences with a phred score below 30 were removed and the remaining reads aligned to the GRCm39 mouse genome with Bowtie2. Read counts were normalized using the DESeq2 (version 1.50.2) package in R/Bioconductor, and differential exon usage was assessed with DEXSeq (version 1.56.0), where an adjusted P < 0.01 was considered significant [28, 29]. Differential gene expression analysis and visualization was performed using DESeq2, and results were visualized using ComplexHeatmap (version 2.26.1). Genes with an adjusted P < 0.01 and log₂(fold change) > 1 were considered differentially expressed. Differentially expressed genes were used for downstream ontology analyses, including GO and gene set enrichment analysis, conducted with the ClusterProfiler and enrichplot packages. Ontology terms with an adjusted P < 0.01 were considered statistically significant. Raw and processed data have been deposited in the NCBI Gene Expression Omnibus under accession number GSE324789.

Statistical analysis

Statistical analyses for chromatin immunoprecipitation sequencing (ChIP-seq) and RNA sequencing (RNA-seq) are described above. For all other experiments, a minimum of three biological replicates were used. Statistical comparisons with one independent variable and two groups were performed using a Student t test, comparisons with more than two groups were performed using one-way ANOVA with Dunnett multiple comparisons test. Experiments involving two independent variables were analyzed using two-way ANOVA with Dunnett or Sidak multiple comparisons post hoc test, as appropriate. The specific statistical test used for each experiment is indicated in the corresponding figure legend. Differences were considered statistically significant when P < 0.05 and denoted with an asterisk (*). Main effect differences between two groups are shown peripheral to the graph, whereas multiple comparisons differences are shown directly between graph elements where appropriate. Bar graphs show the means ± SEM.

Results

Ovol2 is prominently expressed in the labyrinth zone

Embryos lacking Ovol2 are not viable beyond E10.5. Therefore, to determine Ovol2 localization in the conceptus just prior to embryonic lethality, male and female Ovol2^+/− mice were mated, and E9.5 conceptuses were collected and serially sectioned. Ovol2 was readily observed in the placental labyrinth zone, as evidenced by its proximity to Syna and Synb positive regions, markers of SynT-I and SynT-II, respectively [30, 31]. There was little to no Ovol2 expression in the junctional zone, embryonic tissue, or surrounding decidua. Ovol2 was undetectable in null conceptuses (Figure 1A). Interestingly, in contrast to Ovol2⁺ placentas, Syna-positive cells were not readily visible in Ovol2^−/− placentas, whereas Synb-positive cells were present (Figure 1A).

Figure 1.

Ovol2 localizes to the labyrinth zone of the mouse placenta. Fluorescent in situ hybridization showing Ovol2, Syna, and Synb in (A) Ovol2⁺ and Ovol2^−/− mouse placentas at E9.5, and (B) WT mouse placentas at E11.5, E14.5, and E17.5. DAPI was used as a counterstain to visualize nuclei. Ovol2⁺ represents either a WT or Ovol2^+/− conceptus. Low magnification panels in (B) are labeled with dashed lines to denote the decidua (D), junctional zone (JZ), and labyrinth zone (LZ). In high magnification panels, arrowheads denote Ovol2 and Syna co-localization, whereas arrows denote Ovol2 and Synb co-localization. Scale bar: 100 μm (A).

To further assess Ovol2 localization in the mouse placenta across gestation, we mated WT mice and collected placentas at E11.5, E14.5, and E17.5. Similar to E9.5 placentas, Ovol2 consistently localized to the labyrinth zone and was detected in regions containing Syna-positive cells, although its expression was not restricted to these cells and likely includes additional labyrinth-associated trophoblast populations (Figure 1B).

Given the altered labyrinth zone organization in Ovol2^−/− placentas, we next sought to determine how loss of Ovol2 affects SynT formation in vitro. WT mouse TS cells were cultured under stem conditions for 3 days, or differentiation conditions (removal of MEF-conditioned media, FGF4, and activin A) for 3, 5, or 7 days. During the differentiation time course, cells were either left untreated or cultured in differentiation media supplemented with DMSO (vehicle) or CHIR, a GSK3β inhibitor that promotes differentiation toward labyrinth zone lineages [9]. All differentiation protocols resulted in significantly reduced expression of the stem-associated markers Cdx2, Eomes, and Esrrb, and increased expression of Ovol2 relative to the stem state (Supplemental Figure 1A). Only CHIR-treated cells showed elevated expression of Synb and Gcm1 relative to the stem state, whereas the SynT-I marker Syna and the s-TGC marker Ctsq were upregulated in all differentiation protocols (Supplemental Figure 1B). In CHIR-treated cells, the induction of Syna occurred more gradually and in a stepwise manner compared to untreated and vehicle-treated cells, which exhibited high Syna expression by day 3 that persisted through days 5 and 7. Conversely, a modest increase in Ctsq was observed by day 5 in all conditions (Supplemental Figure 1B). Untreated and vehicle-treated cells showed increased expression of the junctional zone markers Tpbpa and Prl3b1 by 7 days, unlike CHIR-treated cells (Supplemental Figure 1C). Prl3d1 expression also increased by day 5 in untreated and vehicle-treated cells. A modest upregulation of Prl3d1 was observed on day 7 following CHIR treatment, although to a lesser extent than in untreated or vehicle-treated cells. The junctional zone progenitor marker Ascl2 was upregulated by day 3 in all conditions. However, this upregulation was more modest and transient in CHIR-treated cells, whereas untreated and vehicle-treated cells sustained high Ascl2 expression to day 7 (Supplemental Figure 1C). These findings suggest that culturing mouse TS cells in differentiation conditions with CHIR preferentially enriches for labyrinth zone trophoblasts, particularly SynT.

Ovol2-deficiency preferentially disrupts SynT-I differentiation

Given that CHIR treatment drives mouse TS cell differentiation mostly toward labyrinth zone-associated lineages, and since Ovol2 localizes to the labyrinth zone in vivo, we next used this protocol to investigate the role of OVOL2 during differentiation of WT and Ovol2^−/− mouse TS cells. Under stem conditions, Ovol2^−/− TS cells formed densely packed colonies, whereas CHIR treatment produced more loosely organized clusters, with some cells displaying an elongated, spindle-shaped morphology. In contrast, WT TS cells formed compact colonies with well-defined cell borders in the stem state, but CHIR treatment led to an apparent loss of border integrity (Figure 2A). Using immunofluorescence, we evaluated changes in cell border organization following 3 days of CHIR treatment using antibodies targeting ZO-1 to demarcate cell borders and DAPI to stain nuclei. Compared with cells cultured under stem conditions, WT cells exhibited a fourfold increase in the percentage of nuclei located within regions lacking clear ZO-1-defined cell borders following CHIR treatment, whereas Ovol2^−/− cells showed only a small, non-significant increase (Figure 2B, P < 0.05). These data suggest that loss of OVOL2 impairs the cell–cell remodeling events associated with SynT lineage formation.

Figure 2.

Ovol2 ^−/− trophoblasts exhibit impaired SynT lineage formation. WT and Ovol2^−/− mouse TS cells were cultured under stem or CHIR-treated differentiation conditions for up to 72 h. (A) Phase contrast images of WT and Ovol2^−/− mouse TS cells. (B) Immunofluorescence for ZO-1 was used to demarcate cell borders in mouse TS cells, DAPI was used to visualize nuclei. Syncytialized cells, defined as two or more nuclei contained within one contiguous ZO-1 border, were quantified and analyzed by two-way ANOVA with Sidak’s multiple comparisons test (N = 3). (C–E) Quantitative RT-PCR was used to evaluate the expression of (C) Ovol2, and stem-associated markers Cdx2, Eomes, and Esrrb; (D) labyrinth zone-associated markers Synb, Gcm1, Syna, and Ctsq; and (E) junctional zone-associated markers Tpbpa, Prl3d1, Gjb3 and Ascl2. Ovol2 expression in WT samples was analyzed by one-way ANOVA with Dunnett multiple comparisons test; Ovol2^−/− samples were excluded from statistical analysis due to values falling below the threshold of detection. All other gene expression data were analyzed by two-way ANOVA with Dunnett multiple comparisons test (N = 4). Statistical differences are denoted with an asterisk (*, P < 0.05). Scale bar: 100 μm.

To determine whether loss of OVOL2 alters the expression of genes associated with SynT formation, WT and Ovol2^−/− mouse TS cells were cultured under differentiation conditions with CHIR for up to 72 h, and gene expression was assessed by quantitative RT-PCR. First, we verified that Ovol2^−/− TS cells cultured in stem and CHIR-treated conditions did not express Ovol2 (Figure 2C). We then measured stem-associated markers Cdx2, Eomes, and Esrrb and found reduced expression in both WT and Ovol2^−/− TS cells upon exposure to CHIR treatment (Figure 2C, P < 0.05). Consistent with earlier observations, culturing WT TS cells under differentiation conditions with CHIR significantly increased expression of Synb and Gcm1 (SynT-II markers), Syna (SynT-I marker), and Ctsq (s-TGC marker; Figure 2D, P < 0.05). In contrast, Ovol2^−/− cells exposed to the same conditions showed attenuated induction of Gcm1 and Synb, indicating that aspects of SynT-II differentiation were initiated but less robust than in WT cells. In comparison, induction of Syna and Ctsq was completely abolished. Collectively, these findings suggest that loss of OVOL2 most strongly disrupts SynT-I-associated differentiation, with parallel loss of s-TGC marker induction, whereas SynT-II-associated differentiation is initiated but attenuated.

To investigate whether Ovol2^−/− cells were instead adopting junctional zone fates, we evaluated the expression of the junctional zone markers Tpbpa, Prl3d1, Gjb3, and Ascl2. Expression of Tpbpa, Prl3d1, and Gjb3 did not differ from the stem state in either WT or Ovol2^−/− cells, although Prl3d1 expression was reduced under all conditions in Ovol2^−/− cells relative to WT. In contrast, expression of Ascl2, a marker associated with junctional zone progenitors, was increased in Ovol2^−/− trophoblasts cultured under differentiation conditions (Figure 2E, P < 0.05). Together with the downregulation of stem-associated markers Cdx2, Eomes, and Esrrb, and the attenuated induction of SynT-II-associated markers Gcm1 and Synb, these findings suggest that Ovol2^−/− TS cells initiate aspects of trophoblast differentiation but show a more profound impairment in SynT-I-associated differentiation. The increased expression of Ascl2 may indicate impaired progression through a progenitor-like or transitional differentiation state rather than a diversion toward mature junctional zone fates.

OVOL2 binds to the promoter proximal regions of genes associated with epithelial-to-mesenchymal transition and represses their expression

We used ChIP-seq to map OVOL2 binding sites across the genome in differentiating trophoblasts. Chromatin was isolated from WT mouse TS cells cultured under differentiation conditions with CHIR for 72 h. Using the MEME Suite for motif analysis, OVOL2 showed a strong preference for binding to the hexameric sequence CCGTTA (Figure 3A), consistent with the consensus motif established for OVO-like transcription factors [25, 32]. A heatmap of all peaks showed strong signal enrichment near transcription start sites, which was not observed in input samples (Figure 3B). Approximately 53% of peaks localized to promoter proximal regions within 1 kb upstream of transcriptional start sites, 4% of peaks were within 1–3 kb upstream of transcriptional start sites, 16% of peaks were within actively transcribed genes, and 27% of peaks localized to distal intergenic regions (Figure 3C). We used ChIPseeker to perform a GO analysis of the top 100 peaks by fold enrichment, excluding those found in distal intergenic regions [26]. Among the top enriched pathways were those associated with mesenchymal development and dysregulated epithelial formation, including “mesenchyme cell differentiation”, “mesenchyme development”, “epithelial to mesenchymal transition”, “regulation of epithelial cell differentiation”, and “negative regulation of epithelial cell differentiation” (Figure 3D, adjusted P < 0.05). Given the strong representation of EMT-associated ontologies, we verified OVOL2 binding at regions proximal to genes associated with several of these pathways (Supplemental File 1). Consistent with previous findings in other epithelial systems [33], robust OVOL2 binding was identified in differentiating trophoblasts upstream of Zeb1, Vim, and Id1 (Supplemental Figure 2A). OVOL2 binding was also detected upstream of the Ovol2 gene, suggesting autoregulation of its own transcription (Supplemental Figure 2A). Given the prominent binding of OVOL2 proximal to the EMT-associated genes Id1, Zeb1, and Vim, we assessed their expression in WT and Ovol2^−/− trophoblasts cultured under differentiation conditions for up to 72 h. Expression of all three genes was higher in Ovol2^−/− trophoblasts than in WT cells (Figure 3E, P < 0.05). Id1 levels increased significantly (3.7-fold by 72 h) relative to the stem state in CHIR-treated Ovol2^−/− trophoblasts whereas Zeb1 and Vim levels decreased approximately 2-fold relative to the stem state but still maintained more than 5-fold higher expression across all time points compared to WT. WT trophoblasts showed similar trends with increased Id1 and reduced Zeb1 and Vim expression with CHIR treatment; however, these trends were modest and none reached statistical significance (Figure 3E). Consistent with gene expression data, ID1 and ZEB1 staining appeared qualitatively stronger in Ovol2^−/− trophoblasts compared with WT controls, though this was not quantified (Figure 3F). Together these data indicate that OVOL2 contributes to the repression of EMT-associated processes, consistent with a role in supporting epithelial identity during trophoblast differentiation.

Figure 3.

OVOL2 binds regulatory regions of EMT-associated genes and represses their expression. ChIP-seq was performed on WT trophoblasts cultured under CHIR-treated differentiation conditions for 72 h. (A) Motif analysis of OVOL2 binding sites. (B) Heatmap showing enrichment of OVOL2 peaks centered on transcription start sites (TSS), compared to input samples. (C) Genomic distribution of OVOL2 peak binding across annotated regions. (D) GO analysis for annotated biological processes of the top 100 peaks (excluding distal intergenic peaks) by fold enrichment (adjusted P < 0.05). (E) Quantitative RT-PCR analysis of EMT-associated genes (Id1, Zeb1, and Vim) in WT and Ovol2^−/− mouse TS cells cultured under stem and CHIR-treated differentiation conditions for up to 72 h. (F) Immunofluorescence for ID1 and ZEB1 in WT and Ovol2^−/− mouse TS cells cultured under stem or CHIR-treated differentiation conditions for 72 h. Dylight 554-conjugated phalloidin was used to label actin filaments, and DAPI was used to visualize nuclei. Data in (E) were analyzed by two-way ANOVA with Dunnett multiple comparisons test. Statistical differences are denoted with an asterisk (*, P < 0.05, N = 4). Scale bar: 50 μm.

OVOL2 loss disrupts trophoblast differentiation and epithelial gene programs

To compare the global transcriptional responses of WT and Ovol2^−/− TS cells under differentiation conditions with CHIR, RNA-seq was performed (Supplemental File 1). Principal component analysis was used to assess overall similarity between the gene expression profiles of WT and Ovol2^−/− trophoblasts, revealing that genotype accounted for most of the variation between groups (Figure 4A). There were 1744 upregulated and 1111 downregulated genes in WT relative to Ovol2^−/− cells (Figure 4B, Supplemental Figure 2B). Among genes associated with mature labyrinth zone lineages, expression of the SynT-I markers, Syna and Stra6, and the s-TGC markers, Ctsq and Hand1, was increased in WT relative to Ovol2^−/− trophoblasts (Supplemental Figure 3A). In contrast, Synb and Gcm1 did not meet the threshold for differential expression (data not shown), which is consistent with our earlier observations that OVOL2 has a more prominent role in SynT-I and s-TGC formation. Surprisingly, Ovol2 also did not meet this threshold. Therefore, DEXSeq was used to further interrogate the dataset by assessing differential exon usage of Gcm1, Synb, and Ovol2 [29]. No differences in exon usage were detected for Synb and Gcm1, consistent with the absence of significant gene-level expression differences between groups. In contrast, exon 3 of Ovol2 (which contains the targeted mutation) showed significantly greater abundance in WT compared to Ovol2^−/− trophoblasts (Supplemental Figure 2C and D, adjusted P < 0.01). Several genes associated with progenitor cell identity and early placental development were among those upregulated in Ovol2^−/− trophoblasts, including Ascl2, Ly6a, Cdx1, and Msx2 (Supplemental Figure 3B) [34–36]. Gene set enrichment analysis of the “Placenta development” ontology revealed significant enrichment in WT compared to Ovol2^−/− trophoblasts, with genes in this set upregulated in WT and reduced in Ovol2^−/− (NES = 1.68, Figure 4C, adjusted P < 0.01). These data further support the notion that WT trophoblasts are driven towards labyrinth zone lineages, whereas Ovol2^−/− trophoblasts initiate differentiation but fail to efficiently progress beyond a progenitor-like or transitional state, with a particularly pronounced impairment in SynT-I-associated differentiation.

Figure 4.

OVOL2 loss disrupts trophoblast differentiation and epithelial gene programs. RNA-seq was performed to compare gene expression profiles of WT and Ovol2^−/− trophoblasts cultured under CHIR-treated differentiation conditions for 72 h. (A) Principal component analysis showing sample clustering and variance. (B) Volcano plot depicting differentially expressed genes. (C) Gene set enrichment analysis for the GO term “Placenta development”. Differentially expressed genes are denoted by vertical lines, and the normalized enrichment score (NES) is shown above the lines. (D) GO analysis of the top 10 upregulated (left) and downregulated (right) biological processes by adjusted P value.

We next determined if genes associated with epithelial and mesenchymal identity were differentially expressed between WT and Ovol2^−/− trophoblasts. Normalized read counts of EMT and mesenchymal identity-associated genes, including Cdh2, Cdh4, Zeb1, and Vim, were decreased in WT compared to Ovol2^−/− cells, whereas epithelial-associated genes (Cdh1, Cdh3, Cdh5, and Krt7) were upregulated (Supplemental Figure 3C and D). GO analysis was used to further characterize differences between WT and Ovol2^−/− trophoblasts. Ontologies associated with epithelial maturation and development, including “epithelial tube morphogenesis” and “morphogenesis of branching epithelium” were among the top pathways enriched in genes upregulated in WT trophoblasts (Figure 4D, adjusted P < 0.01). Conversely, “mesenchyme development” was among the top pathways enriched in Ovol2^−/− trophoblasts (Figure 4D, adjusted P < 0.01). Together, these findings suggest that OVOL2-mediated regulation of epithelial identity is important for placental development and trophoblast differentiation toward mature labyrinth zone lineages, consistent with our earlier ChIP-seq data implicating OVOL2 in the regulation of epithelial-mesenchymal-associated gene programs.

Ovol2-deficient trophoblasts exhibit diminished epithelial identity

Using quantitative RT-PCR, we measured relative transcript levels of epithelial and mesenchymal markers in WT and Ovol2^−/− TS cells cultured under stem and differentiation conditions with CHIR for 72 h. Compared to WT trophoblasts, Ovol2^−/− trophoblasts exhibited significantly reduced expression of genes encoding epithelial-associated cadherins (Cdh1, Cdh3, and Cdh5), and keratin 7 (Krt7) under both stem and differentiation conditions (Figure 5A, P < 0.05). Conversely, expression of EMT-associated and mesenchymal markers Cdh2, Cdh4, Zeb1, and Vim was significantly increased in Ovol2^−/− trophoblasts relative to WT under both stem and CHIR-treated conditions (Figure 5B, P < 0.05). We then used immunoblotting to measure the protein levels of a subset of these markers. We found a significant 2-fold decrease in E-cadherin levels in Ovol2^−/− trophoblasts compared to WT when cells were cultured under differentiation conditions (Figure 5C, P < 0.05). In both stem and differentiating conditions, there was a significant 9- and 18-fold increase in ZEB1 levels relative to WT, respectively (Figure 5C, P < 0.05). Vimentin was strongly observed in Ovol2^−/− trophoblasts in both stem and differentiation conditions, but not detectable in WT trophoblasts under either condition (Figure 5C). Collectively, our data strongly suggest that a key function of OVOL2 in trophoblasts is to reinforce epithelial identity, a process that appears to be essential for the development of labyrinth trophoblast lineages.

Figure 5.

Ovol2-deficient trophoblasts exhibit mesenchymal-associated characteristics and reduced epithelial features. WT and Ovol2^−/− TS cells were cultured under stem and CHIR-treated differentiation conditions for 72 h. Expression of (A) epithelial identity-associated markers (Cdh1, Cdh3, Cdh5, and Krt7) and (B) mesenchymal identity-associated markers (Cdh2, Cdh4, Zeb1, and Vim). (C) Protein levels of ZEB1, E-cadherin, and Vimentin. β-actin was used as a loading control. Band intensities of E-cadherin and ZEB1 normalized to β-actin are shown to the right. Vimentin was not quantified because no signal was detectable in WT samples. All blot images were cropped to show the labeled molecular weight. Data were analyzed by two-way ANOVA with Sidak multiple comparisons test. Statistical differences are denoted with an asterisk (*, P < 0.05, N = 4–6).

Given our data supporting a role for Ovol2 in TS cell differentiation and epithelial identity, we next asked whether ectopic Ovol2 expression could partially rescue the Ovol2^−/− phenotype. Ovol2^−/− TS cells were cultured under CHIR-treated differentiation conditions for 3 days, and transfected with a plasmid encoding Ovol2, followed by RNA extraction and quantitative RT-PCR analysis. As expected, cells transfected with the Ovol2 plasmid showed robust expression of Ovol2 (Supplemental Figure 4A, P < 0.05). Ectopic Ovol2 did not alter differentiation-associated repression of stem-associated markers including Cdx2, Eomes, and Esrrb (not shown). Interestingly, Syna expression significantly increased in cells transfected with the Ovol2 plasmid by approximately 2-fold, indicating a partial rescue in SynT-I differentiation capacity; however, there was no change in Ctsq expression suggesting no effect on s-TGC populations (Supplemental Figure 4A, P < 0.05 for Syna). Additionally, because Ovol2^−/− cells exhibit reduced epithelial and enhanced mesenchymal-associated features, we next assessed whether ectopic Ovol2 could reverse these changes. Compared with controls, cells with ectopic Ovol2 expression showed significantly increased expression of the epithelial markers Cdh1, Cdh3, and Krt7, and decreased expression of the mesenchymal-associated markers Cdh4, Zeb1, and Vim (Supplemental Figure 4B and C, P < 0.05). Collectively, these data suggest that ectopic Ovol2 expression partially improves SynT-I marker expression and supports maintenance of epithelial identity in Ovol2^−/− TS cells.

Ovol2-deficient placentas show a reduced epithelial phenotype in the labyrinth zone

To determine whether altered epithelial-mesenchymal marker expression is evident in placentas lacking OVOL2 in vivo, we performed immunofluorescence for E-cadherin and Vimentin in E9.5 placentas with and without OVOL2, as well as in E11.5, E14.5, and E17.5 WT placentas. Antibodies targeting MCT1 and MCT4 were used in parallel to identify SynT-associated regions, although their staining patterns at E9.5 were less well resolved than at later gestational stages. At E9.5, Ovol2⁺ placentas showed prominent E-cadherin staining, whereas E-cadherin signal was not readily detectable in Ovol2^−/− placentas. Vimentin was primarily restricted to mesodermal cells within the chorionic plate, with strong expression also observed in the decidua (Figure 6A). We did not observe clear Vimentin signal in trophoblast compartments at this stage in either Ovol2⁺ or Ovol2^−/− placentas. Although these data should be interpreted cautiously given the complexity of the developing labyrinth zone and limitations of marker resolution at E9.5, they suggest that loss of OVOL2 is associated with reduced epithelial marker expression in vivo.

Figure 6.

OVOL2 supports epithelial identity in the developing placenta. Immunofluorescence was performed for E-cadherin (left panels) and Vimentin (right panels). In both panels, MCT1 and MCT4 were used to identify SynT-I and SynT-II layers, respectively, whereas DAPI was used to visualize nuclei (A and B). (A) E9.5 Ovol2⁺ and Ovol2^−/− mouse placentas. Ovol2⁺ represents either a WT or Ovol2^+/− conceptus. Low magnification merged panels are labeled to denote the decidua (D) and placenta (P). (B) E11.5, E14.5, and E17.5⁻ WT mouse placentas. Labels for the decidua (D), junctional zone (JZ), and labyrinth zone (LZ) are indicated in the low magnification merged panels. Arrowheads in high magnification panels indicate E-cadherin or Vimentin-positive staining proximal to MCT1-positive staining. Scale bar = 100 μm.

We next examined E-cadherin and Vimentin localization across later stages of placental development in WT mice. From E11.5 to E17.5, E-cadherin staining remained prominent within the labyrinth zone, including regions closely associated with MCT1- and MCT4-positive SynT layers (Figure 6B). In contrast, Vimentin localized primarily to mesenchyme-associated regions of the labyrinth zone and decidua (Figure 6B). The persistence of strong E-cadherin staining during continued branching morphogenesis supports the notion that epithelial identity is maintained at the maternal-fetal interface throughout placental development.

Discussion

OVOL2 has long been recognized as essential for placental development in mice, as OVOL2 deficiency leads to embryonic lethality by E10.5 with severe placental defects [16, 37]. However, its specific role remains poorly defined. In this study, we show that Ovol2 localizes to the labyrinth zone throughout gestation and that Syna is not detectable in Ovol2^−/− placentas, suggesting a profound defect in SynT-I layer formation prior to embryonic lethality. Consistent with these findings, Ovol2^−/− trophoblasts exhibit a preferential disruption of SynT-I-associated differentiation, whereas SynT-II-associated differentiation is initiated but attenuated. These defects correlate with increased mesenchymal-associated features and reduced epithelial characteristics, contrasting with WT trophoblasts, which reinforce epithelial polarity during differentiation. These data position OVOL2 as an important regulator of epithelial identity during SynT-I differentiation and labyrinth zone development.

In this study, we subjected mouse TS cells to two distinct differentiation strategies to enrich for specific trophoblast subtypes. Withdrawal of stem-promoting growth factors and MEF conditioned medium promoted differentiation primarily toward junctional zone lineages, as indicated by high expression of Tpbpa, Prl3b1, Prl3d1, and Ascl2 [38]. This differentiation paradigm also induced expression of genes associated with SynT-I (Syna, Stra6) and s-TGCs (Ctsq, Hand1). Conversely, differentiation toward the SynT-II lineage was apparent only following CHIR treatment, as indicated by elevated expression of Synb and Gcm1. Notably, increased expression of Syna, Ctsq, and Ovol2 was observed across all differentiation conditions. Because both SynT-I and SynT-II markers were elevated in CHIR-treated cultures, these findings suggest that CHIR treatment supports the emergence of multiple labyrinth zone lineages, potentially through induction of labyrinth trophoblast progenitor populations [36]. However, the modest increase in junctional zone markers Prl3d1 and Ascl2 in CHIR-treated TS cells also indicates that CHIR treatment does not fully abolish the capacity of TS cells to generate other trophoblast lineages, though at lower levels relative to CHIR-free differentiation conditions. These findings are consistent with previous reports showing that differentiation of TS cells in the presence of CHIR leads to increased Syna and Ctsq expression beginning approximately 4 days after differentiation, albeit at lower levels than those observed under CHIR-free conditions [9]. Thus, the CHIR treatment paradigm provides a useful model for enriching SynT-associated differentiation, although it should be interpreted as a lineage-enriched differentiation system rather than as a model that exclusively forms SynT-II or fully reconstructs labyrinth zone development. This is particularly important given that in vitro TS cell differentiation lacks the spatial organization, cellular interactions, and microenvironment of the developing placenta. Nonetheless, given the prominent localization of Ovol2 to the labyrinth zone in vivo, the CHIR treatment paradigm provides a suitable model for investigating the role of OVOL2 in labyrinth trophoblast development.

Mouse TS cells are multipotent and generate several major progenitor populations that ultimately form the junctional and labyrinth zones of the mature placenta [36]. Progenitors of the junctional zone persist well into late gestation, supporting continued growth of this compartment up to approximately E16.5 [35]. Labyrinth zone progenitors remain present until around E14.5, sustaining ongoing development of the maternal-fetal exchange interface [39]. Under differentiation conditions with CHIR, Ovol2^−/− trophoblasts acquire a transcriptional profile characteristic of a bipotential progenitor state capable of producing both junctional zone and labyrinth zone trophoblasts as evidenced by elevated Ascl2, Ly6a, and Cdx1 expression, and maintain expression of the SynT lineage progenitor marker Msx2 [36, 40]. Although Ovol2^−/− trophoblasts, like WT cells, rapidly downregulate stem-associated factors such as Cdx2, Eomes, and Esrrb upon exposure to CHIR-containing differentiation conditions, they do not efficiently upregulate factors associated with terminal trophoblast lineages, most notably SynT-I. Instead, Ovol2-deficient trophoblasts may remain in, or fail to efficiently progress beyond, an earlier progenitor-like state. A similar block has been observed following conditional deletion of GATA2 and GATA3, which disrupts SynT maturation and leads to an accumulation of labyrinth zone progenitors [41]. Collectively, these findings suggest that impaired activity of certain transcription factors required for terminal trophoblast differentiation, like OVOL2, may not prevent exit from the stem state, but instead may limit progression beyond a transitional progenitor-like state. Failure to progress beyond this point would be expected to impair mature lineage formation, compromise placental function, and contribute to mid-gestation embryonic lethality [41, 42].

The identity of the trophoblast populations that persist in Ovol2^−/− placentas in vivo remains unresolved. Although our in vitro data suggest that Ovol2^−/− TS cells initiate aspects of differentiation but fail to efficiently progress toward mature SynT-I lineages, our findings in E9.5 placental tissue indicate that Synb-positive cells remain detectable in Ovol2^−/− placentas. Thus, the persisting trophoblast populations may include SynT-II-associated cells, undifferentiated trophoblasts, or aberrant transitional states. Determining the relative contribution and identity of these populations will require additional approaches such as lineage tracing and comprehensive spatial profiling. In particular, single-cell or spatial transcriptomic approaches combined with trajectory analysis could help determine whether OVOL2-deficient trophoblasts are delayed in SynT-I lineage progression, diverted toward alternative trophoblast states, or maintained in a progenitor-like transitional state.

The emergence of mature SynT lineages represents a pivotal step in establishing the labyrinth zone. Our findings demonstrate that OVOL2 supports this process, with a particularly pronounced requirement for SynT-I differentiation, likely contributing to the midgestation embryonic lethality observed in conceptuses lacking OVOL2. This conclusion aligns with prior reports localizing Ovol2 to the labyrinth zone [18, 36], and with broader evidence that mid-gestation lethality often stems from placental defects [43]. Consistent with this, we and others have linked OVOL2 deficiency to embryonic lethality driven by impaired placentation, accompanied by abnormal heart development and disrupted placental vascularization [16, 18]. Using CHIR-treated mouse TS cells, we further demonstrate that the emergence of SynT-I populations requires OVOL2 activity, reinforcing the notion that OVOL2 deficiency disrupts labyrinth zone development in vivo. Notably, differentiation toward labyrinth zone lineages was associated with a shift toward epithelial polarity as evidenced by elevated E-cadherin levels with CHIR treatment. Conversely, Ovol2^−/− trophoblasts exhibit features consistent with a mesenchymal-associated state, including increased ZEB1 and Vimentin levels and reduced epithelial marker expression compared to WT trophoblasts. This phenotype was partially reversed by ectopic OVOL2 expression. Collectively, these observations suggest that OVOL2 regulates labyrinth zone trophoblast differentiation, at least in part, by modulating EMT-associated signaling and maintaining epithelial identity. These findings provide a foundation for future studies aimed at elucidating the roles of individual epithelial and mesenchymal factors in trophoblast differentiation.

OVOL2-mediated repression of EMT-associated programs is a conserved mechanism that supports epithelial differentiation across multiple tissues, including the skin and mammary gland [10, 33]. Our findings show that OVOL2 binds to genomic regions near key EMT-associated regulators, including Zeb1, Id1, and Vim. Although additional functional validation will be required to confirm direct transcriptional repression of these targets by OVOL2, these findings support a role for OVOL2 in restraining mesenchymal-associated gene expression and preserving epithelial identity during trophoblast differentiation. In TS cells, EMT-associated processes are dynamically regulated during differentiation, contributing both to labyrinth trophoblast morphogenesis and the generation of invasive trophoblast lineages [44, 45]. This highlights the need for precise control of epithelial plasticity across trophoblast fates. In many epithelial progenitor systems, transient acquisition of partial mesenchymal-associated features enables cellular plasticity during differentiation, but epithelial programs must subsequently be reinforced to stabilize differentiated epithelial states. Consistent with this regulatory role, WT TS cells exhibit higher expression of epithelial-associated cadherins (Cdh1, Cdh3, and Cdh5) and reduced expression of mesenchymal markers (Cdh2, Cdh4, Zeb1, and Vim) compared to Ovol2^−/− cells. These observations align with previous reports demonstrating increased E-cadherin expression during CHIR-treated labyrinth trophoblast differentiation, as well as evidence that E-cadherin is required for trophoblast fusion in both humans and mice [9, 46]. In vivo, E-cadherin was strongly expressed within the labyrinth zone across gestation and was less readily detectable in E9.5 Ovol2-deficient placentas, supporting altered epithelial marker expression following OVOL2 loss. Conversely, Vimentin localized primarily to mesenchymal compartments, as expected, and we did not observe a clear increase in trophoblast-associated Vimentin staining in E9.5 Ovol2-deficient placentas. Thus, while our in vitro data indicate increased mesenchymal-associated marker expression in Ovol2^−/− TS cells, the corresponding in vivo phenotype is most clearly reflected by reduced epithelial marker expression at this stage. This difference may indicate transient or stage-specific mesenchymal marker expression, or differences between the use of TS cells in vitro and the in vivo placental environment.

A limitation of the current study is that we cannot determine whether the reduced epithelial marker expression and increased mesenchymal-associated transcriptional program observed in vitro are directly responsible for impaired SynT-I differentiation, or whether they are consequences of disrupted trophoblast lineage progression following Ovol2 loss. Despite extensive branching morphogenesis in the labyrinth zone across gestation, E-cadherin staining remained strongly detectable in regions adjacent to SynT layers, suggesting that epithelial identity is actively reinforced during formation and maintenance of the maternal-fetal exchange interface. The reduced E-cadherin signal observed in E9.5 Ovol2-deficient placentas is consistent with a role for OVOL2 in supporting this epithelial program during early labyrinth zone development. Future studies examining expression of epithelial and mesenchymal-associated markers at earlier developmental stages, including in Ovol2-deficient conceptuses, will be important for determining whether the mesenchymal-associated features observed following Ovol2 loss arise broadly across trophoblast populations or reflect lineage-specific alterations at certain stages of placental development.

In conclusion, our findings identify OVOL2 as a critical transcriptional regulator of SynT-I differentiation in the developing mouse placenta. We further demonstrate that OVOL2 restrains an inappropriate mesenchymal-associated transcriptional program during trophoblast differentiation, consistent with a role in preserving the epithelial state that supports labyrinth zone development. These findings position OVOL2 as a key regulator of epithelial identity within the trophoblast differentiation hierarchy and reveal a previously unrecognized requirement for OVOL2 in SynT-I differentiation when establishing the maternal-fetal interface.

Author contributions

S.J.R and V.S.P. designed the study. V.S.P., M.J.J., A.M.R., and G.J.B. performed experiments and analyzed data. The manuscript was written and data analyzed by S.J.R. and V.S.P. The manuscript was approved by all co-authors.

Conflict of Interest: The authors have declared that no conflict of interest exists.

Data availability

Data from this article have been deposited in the NCBI Gene Expression Omnibus (GEO) under accession number GSE324788 and GSE324789.