SEARCH

SEARCH BY CITATION

Keywords:

  • PDGF signaling;
  • Primitive endoderm;
  • Blastocyst;
  • Mouse embryo;
  • Apoptosis

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

At the end of the preimplantation period, the inner cell mass (ICM) of the mouse blastocyst is composed of two distinct cell lineages, the pluripotent epiblast (EPI) and the primitive endoderm (PrE). The current model for their formation involves initial co-expression of lineage-specific markers followed by mutual-exclusive expression resulting in a salt-and-pepper distribution of lineage precursors within the ICM. Subsequent to lineage commitment, cell rearrangements and selective apoptosis are thought to be key processes driving and refining the emergence of two spatially distinct compartments. Here, we have addressed a role for Platelet Derived Growth Factor (PDGF) signaling in the regulation of programmed cell death during early mouse embryonic development. By combining genetic and pharmacological approaches, we demonstrate that embryos lacking PDGF activity exhibited caspase-dependent selective apoptosis of PrE cells. Modulating PDGF activity did not affect lineage commitment or cell sorting, suggesting that PDGF is involved in the fine-tuning of patterning information. Our results also indicate that PDGF and fibroblast growth factor (FGF) tyrosine kinase receptors exert distinct and non-overlapping functions in PrE formation. Taken together, these data uncover an early role of PDGF signaling in PrE cell survival at the time when PrE and EPI cells are segregated. Stem Cells 2013;31:1932-1941


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

Apoptosis or programmed cell death is an essential process for a developing organism. Besides its critical function in removing damaged cells, apoptosis plays important roles during embryogenesis under normal physiological conditions (reviewed in ref. [1-3]). However, the temporal and spatial coordination between selective cell death and developmental signals remains an open question. To address this question, we have focused on the first apoptotic events during development, which occur at the end of the preimplantation period in the mouse blastocyst stage embryo [4-6]. At this stage, the embryo comprises three molecularly and spatially distinct cell lineages: an outer layer of trophectoderm (TE) cells and an inner cell mass (ICM), which is composed of the epiblast (EPI) and the primitive endoderm (PrE). Later on, the pluripotent EPI will predominantly form the embryo-proper, the TE will give rise to the fetal portion of the placenta and the PrE will contribute to the parietal and visceral yolk sacs and the endoderm of the fetus [7-11]. Formation of these embryonic and extraembryonic cell lineages results from two sequential cell fate decisions (reviewed in [12, 13]). The first concerns the specification of the TE and the ICM, while the second allows the divergence between EPI and PrE cell lineages within the ICM. The current model of EPI versus PrE specification involves three successive steps. Initially, key lineage-specific transcription factors are expressed at various levels in most cells (blastomeres) until the early blastocyst stage [14]. In mid-blastocyst embryos (approximately 64-cell stage), these markers become progressively restricted such that the ICM is a mosaic of cells expressing either PrE or EPI markers that are organized in an apparent salt-and-pepper pattern [14, 15]. It is thought that this stage corresponds to the period when ICM cells become committed to either a PrE or a EPI fate, a process driven by fibroblast growth factor (FGF) signaling (reviewed in [16]). The final step concerns the segregation of these two lineages into two distinct tissue layers. Several mechanisms have been shown to regulate the sorting of ICM lineage precursors including actin-dependent cell movements, retention of positional information by sorted PrE cells and epithelialization [14, 17, 18]. It has been proposed that apoptosis could also be involved at these stages where it functions to remove cells that are not properly fated or are mispositioned [14, 17, 19, 20]. However, the molecular mechanisms involved in this selective cell death remain unknown.

Here, we provide evidence that the Platelet Derived Growth Factor (PDGF) signaling pathway is involved in the regulation of this process of selective apoptosis. By combining genetic and pharmacological approaches, we demonstrate that the absence of PDGF signaling specifically affects the survival of PrE cells in a caspase-dependent manner. Using a null knock-in PdgfraH2B-GFP allele [21], which is expressed in the PrE [14, 22], we observed that in mutant embryos cell death affected PrE cells independently from their position within the ICM.

Taken together, this study uncovers a role for PDGF signaling in the survival of the PrE lineage at the time when a subset of ICM cells have committed to a PrE fate but before their segregation into a distinct layer. Furthermore, we show that FGF and PDGF, two prominent classes of receptor tyrosine kinase (RTK) signaling that regulate similar transduction pathways within the ICM of the mouse blastocyst, exert distinct roles in cell lineage specification and survival during cell sorting, respectively.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

Mouse Husbandry

Mice were maintained under a 12-hour light–dark cycle. Mouse strains used were PdgfraH2B-GFP/+ [21] and Fgf4+/− strains [23, 24] and maintained on a CD1 genetic background, wild-type CD1 (Charles River, Wilmington, MA, http://www.criver.com) and ICR (Taconic, Petersburgh, NY, www.taconic.com).

Tamoxifen Injection

Female mice aged 2–3 months were injected intraperitoneally with a mixture of 10 µg Tamoxifen (Sigma-Aldrich, St. Louis, MO, http://www.sigmaaldrich.com) and 2–3 mg progesterone (Abraxis, Schaumburg, IL, http://www.apppharma.com]) diluted into corn oil (Sigma-Aldrich, St. Louis, MO, http://www.sigmaaldrich.com) at 2.5 days post coitum (dpc). Embryos were recovered 1-day later (referred to as embryonic day (E)3.5 + 1-day).

Embryo Recovery and Culture

Embryos were recovered by flushing uteri or oviducts in M2 (Millipore, Temecula, CA, www.millipore.com). Embryos were cultured in 10 µl drops of KSOM (Millipore, Temecula, CA, www.millipore.com) under mineral oil (Sigma) for up to 50 hours at 37°C, 5% CO2. Inhibitors used were Gleevec (gift of P. Besmer, Sloan-Kettering Institute, NY) at concentrations ranging from 1 to 10 µM and Z-VAD-FMK (R&D systems, Minneapolis, http://www.rndsystems.com) at 20 µM. Recombinant human PDGF-AA (R&D systems) was used at 500 ng/ml.

Electroporation of Blastocyst Embryos

Electroporation was performed according to Frankenberg et al. [25]. Embryos were recovered around noon at 3.5 dpc. The zona pellucida was removed using acidic Tyrode's (Sigma-Aldrich, St. Louis, MO, http://www.sigmaaldrich.com). Embryos were washed in M2 and then G2 (Vitrolife, Kungsbacka, Sweden, http://www.vitrolife.com) buffered with 20 mM HEPES (hG2). Embryos were then transferred into 50 µl hG2 with 60 µg pCAGGS [26] or pCAG::Pdgfa vectors and 40 µg pCAG::H2B-mCherry [27]. Electroporation was performed in 1 mm gap cuvettes with an ECM 830 (BTX, Harvard Bioscience, Holliston, MA, http://www.btxonline.com) apparatus. Settings were four pulses of 33 V and 1 millisecond duration each with a 100-millisecond interval. Embryos were subsequently washed through several drops of hG2 and cultured for 25–30 hours in 10 µl G2 drops covered by mineral oil at 37°C, 5% CO2.

Live Imaging

For live imaging, a group of three to four embryos was transferred into glass-bottom microwell dishes (MatTek Corp., Ashland, MA, http://www.glass-bottom-dishes.com/) coated with 2% Agar, 0.9% NaCl. Three-dimensional (3D) time-lapse data were acquired using Volocity 5.5.1 acquisition software (Perkin-Elmer, Gaithersburg, MD, www.perkinelmer.com) and a Perkin-Elmer RS3 Nipkow-type scanhead mounted on a Zeiss Axiovert 200M with Hamamatsu C4742-80-12AG camera. Green fluorescent protein (GFP) was excited using a 488-nm Argon laser. Images were acquired using a Zeiss plan-Neofluar 25×/0.8 DIC korr objective; 10–20 xy planes were acquired, separated by 2–4 µm every 15 minutes. Embryos were subsequently genotyped as described previously [22].

Immunostaining

Embryos were fixed for 10 minutes in 4% paraformaldehyde at room temperature, washed in phosphate-buffered saline (PBS) 10 mg/ml bovine serum albumin (BSA) and permeabilized in 0.25% Triton PBS for 10 minutes at room temperature. After several washes in PBS-BSA, embryos were preincubated 10 minutes in PBS with 0.1% Tween 20 and 10% fetal bovine serum and then incubated with primary antibodies overnight at 4°C. The following primary antibodies were used: anti-GATA4 (1/300, Santa Cruz, Santa Cruz, CA, http://www.scbt.com), anti-cleaved Caspase-3 (1/100, Cell Signaling, Hitchin, UK, www.cellsignal.com), anti-Nanog (1/700, Cosmo Bio, Oxford, UK, www.cosmobio.co.jp), and anti-SOX17 (1/300, R&D Systems). The next day, embryos were washed in PBS-BSA then incubated in presence of secondary antibodies overnight at 4°C. Secondary AlexaFluor-conjugated antibodies (Invitrogen) were used at a dilution of 1/500. DNA was counterstained with Hoechst 33342 (Molecular Probes, Carlsbad, CA, http://probes.invitrogen.com).

Image Data Acquisition and Processing

Image data of immunostained embryos were acquired on a Zeiss LSM510 META confocal microscope. Fluorescence was excited with a 405-nm laser diode (Hoechst), a 488-nm Argon laser (GFP, Alexa Fluor 488), a 543-nm HeNe laser (Alexa Fluor 546, 555, 568) and a 633-nm HeNe laser (Alexa Fluor 633 and 647). Images were acquired using a Plan-apochromat 20×/NA 0.75 objective with optical section thickness ranged from 1 to 2 µm. Raw data were processed using Zeiss AIM software (Carl Zeiss Microsystems, Oberkochen, Germany, http://www.zeiss.com), IMARIS 7.2.2 software (Bitplane AG, Zurich, Switzerland, www.bitplane.com) and Volocity 5 software (Perkin-Elmer).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

Embryos Lacking PDGF Signaling Exhibit Increased Cell Death Within the PrE Lineage

We reported previously that in the absence of PDGF signaling, E4.5 embryos and tamoxifen-induced diapause embryos exhibited a reduction in the number of PrE cells [22]. During the course of this analysis, we noted the presence of picnotic nuclei in mutant embryos raising the possibility that selective cell death may contribute to the reduction in the size of the PrE. To test this possibility, we quantified the number of fragmented nuclei in E3.5 embryos recovered from females 1 day after tamoxifen injection (E2.5 + 1-day) (Fig. 1A). We noted a significant increase in the number of fragmented nuclei in mutant embryos (3.9 ± 0.8 nuclei in mutant vs. 1.1 ± 0.3 in heterozygotes and 1.4 ± 0.6 in wild type). To determine whether cell death was a result of apoptosis, we analyzed the number of cleaved Caspase-3-positive cells in embryos. We noted a significant increase in the number of active Caspase-3-positive cells in mutant embryos (3.6 ± 0.9 apoptotic cells in mutant vs. 0.7 ± 0.4 in heterozygotes and 1.4 ± 0.8 in wild type).

image

Figure 1. Increased apoptosis in blastocyst embryos lacking Pdgfra. Analysis of cell death in E2.5 embryos recovered 1 day after tamoxifen injection (A) and in E3.5 embryos (B). Embryos were recovered from PdgfraH2B-GFP/+ intercrosses, processed for immunostaining, imaged, and subsequently genotyped, as described previously [22]. Upper panels represent three-dimensional rendering of a PdgfraH2B-GFP/H2B-GFP embryo. Lower panels represent distribution of fragmented nuclei and Caspase-3 positive cells in Pdgfra+/+ (wild type), PdgfraH2B-GFP/+ (heterozygous), and PdgfraH2B-GFP/H2B-GFP (mutant) embryos at E2.5 + 1-day after tamoxifen injection and E3.5. Only E3.5 blastocysts with a total cell number ranging from 60 to 120 cells were analyzed. This corresponds to the time when inner mass cells are committed toward primitive endoderm and pluripotent epiblast cell fates. GFP, green; active Caspase-3, red; Hoechst, blue. Scale bar = 20 µm. Statistical Mann–Whitney tests are indicated when significant (*, p < .05; **, p < .01). Error bars indicate SEM; n, number of embryos analyzed. Abbreviations: GFP, green fluorescent protein; HET, heterozygous; HOM, mutant; WT, wild type.

Download figure to PowerPoint

To exclude the possibility that the increase in apoptosis in Pdgfra-deficient embryos could result from the injection of tamoxifen, we conducted similar analyses on E3.5 embryos recovered from heterozygous intercrosses. We observed a significant increase in fragmented nuclei and Caspase-3-positive cells in mutant embryos (3.6 ± 0.7 nuclei for mutant vs. 1.4 ± 0.3 for heterozygotes and 1.3 ± 0.4 for wild type; 2.3 ± 0.8 apoptotic cells vs. 0.4 ± 0.1 for heterozygotes and 0.5 ± 0.4 for wild type) (Fig. 1B). Interestingly, combining these two sets of experiments, we observed that 82% Caspase-3-positive cells (66 of 80) in mutants were GFP-positive. We found a comparable situation in heterozygous embryos with 72% Caspase-3-positive cells being GFP-positive (21 of 29). During this developmental stage, which is before ICM lineage sorting, and when EPI/PrE cells have acquired mutually exclusive expression of markers, Pdgfra is expressed predominantly in PrE-biased cells [14]. Although EPI or PrE-biased cells retain their developmental plasticity until peri-implantation [28], high PdgfraH2B-GFP expression appears to be an accurate predictor of PrE lineage commitment. Therefore, our current findings strongly suggest that in the absence of PDGF signaling PrE cells exhibit increased levels of apoptosis. However, to conclusively rule out that we are not following Pdgfra-GFP-positive cells that eventually acquire an EPI fate through downregulation of the reporter, we would need to develop an EPI-specific reporter and repeat the analysis. Although concurrent visualization of an EPI reporter would be desirable, it is outside the scope of this study.

Localization of Dying PrE Cells in Embryos Lacking PDGF Signaling

The transition between the mid-blastocyst (where PrE and EPI have been specified and organized in a salt-and-pepper distribution) and the late blastocyst (where ICM cell lineages are segregated) is achieved through coordination of several mechanisms including selective apoptosis of PrE cells located internally within the ICM and not having a major contact with the blastocoel cavity [14, 17]. Therefore, we hypothesized that absence of PDGF signaling might preferentially affect the viability of inner PrE cells. Unfortunately, our statistical analyses of markers only provided a snapshot of the events occurring within the ICM, and we could not determine the position of Caspase-3-positive cells. Therefore, we used the PdgfraH2B-GFP strain of mice, which functions as a PrE live imaging reporter and in which the nuclear-localized fluorescence allows the tracking of individual cells, as well as the visualization of cell death by nuclear fragmentation (Fig. 2A; Supporting Information Movie 1). 3D time-lapse imaging revealed a significant increase in GFP+ (GFP-positive) cell death in mutant embryos (8.2 ± 2.0 dead cells in mutants vs. 2.5 ± 0.3 in heterozygotes) (Fig. 2B). Then, we assigned the position of GFP+ cells before their death: inside cells (located inside the ICM) are indicated in red, outside cells (located at the surface of the ICM in contact to the blastocoel cavity) in blue (Fig. 2C, 2D; Supporting Information Movie 1). In heterozygous embryos, we determined that 54% (42 of 78) were inner cells and 41% (32 of 78) were outside. We found a similar situation in mutant embryos in which 56% (37 of 66) were inner cells and 45% outside (25 of 66). From these data, we conclude that absence of PDGF signaling affects the survival of GFP+ cells independently from their position within the ICM.

image

Figure 2. Live imaging reveals increased cell death of inner primitive endoderm (PrE) cells in Pdgfra-deficient embryos. (A): Single frames from three-dimensional time-lapse sequence of inner cell mass (ICM) of a PdgfraH2B-GFP/H2B-GFP E3.5 mid-blastocyst embryo. Cell death events occurring in green fluorescent protein (GFP)-positive cells are depicted either with a red (inner PrE cell) or blue arrowhead (outer PrE cell). Annotated movie sequence is provided as Supporting Information Movie 1. GFP, green. Scale bar = 10 µm. (B): Numbers of cell death events in heterozygous (HET) and mutant (HOM) embryos. Statistical Mann–Whitney test is indicated when significant (***, p < .001). Error bars indicate SEM, n, number of analyzed embryos. (C, D): Localization of PrE cells before death within ICM of HET (C) or HOM (D) embryos. n, number of cells analyzed. Abbreviations: GFP, green fluorescent protein; ICM, inner cell mass; IN, inside; ND, not determined; OUT, outside.

Download figure to PowerPoint

Gleevec, a PDGF Signaling Inhibitor, Specifically Affects the PrE Lineage

Our results suggested that PDGF signaling might be required at the time of PrE cell sorting. However, in Pdgfra-null embryos, it was not possible to determine the precise time when PDGF signaling was required during PrE formation. To address this question, we used the pharmacological inhibitor Gleevec (also known as Imatinib), a potent inhibitor of PDGF receptors [22, 29]. We recovered wild-type embryos at various developmental stages (E2.5, early and mid-blastocyst stages) and cultured them until E4.5 in the presence of various concentrations of Gleevec (Fig. 3A). Embryos were then immunostained with lineage-specific markers (Supporting Information Fig. S1). We first noticed that Gleevec used at a 10 µM concentration strongly affected the development of E2.5 and early blastocyst stage embryos (Fig. 3B). Indeed, all (5 of 5) E2.5 embryos died, while cultured early blastocysts were smaller and had a reduced blastocoel cavity. However, at 1 and 5 µM concentrations, we did not observe any notable morphological differences in inhibitor treated embryos as compared to control embryos (Fig. 3B).

image

Figure 3. Gleevec treatment affects preimplantation development. (A): Schematic representation of timeline of Gleevec treatment. (B): Single optical sections of embryos cultured at indicated concentrations of Gleevec. Hoechst, blue. Scale bar = 20 µm; n, number of analyzed embryos. (C): Distribution of trophectoderm (green), primitive endoderm (blue), and pluripotent epiblast (red) cells in E2.5, early and mid E3.5 embryos cultured with or without Gleevec until E4.5. Statistical Mann–Whitney tests are indicated when significant (*, p < .05; **, p < .01; ***, p <.001). Error bars indicate SEM. Abbreviations: bf, bright-field; EPI, pluripotent epiblast; PrE, primitive endoderm; TE: trophectoderm.

Download figure to PowerPoint

We analyzed the number of cells for each of the three blastocyst cell lineages (TE, EPI, and PrE) in embryos treated with concentrations 0, 1, and 5 µM Gleevec (Fig. 3C; Supporting Information Fig. S1). Interestingly, while the numbers of TE and EPI cells were unaffected compared with untreated embryos, the number of PrE cells was severely reduced in E2.5, early and mid blastocyst stage embryos that had been treated with Gleevec. Strikingly, the reduction in PrE cell number was of a similar magnitude in the three categories of embryos at 1 and 5 µM concentrations (fold reduction compare with control embryos: 1.9 and 2.2 in E2.5; 1.7 and 1.9 in early E3.5; 1.7 and 1.8 in mid E3.5). Thus, at these concentrations, Gleevec exerted a comparable effect at all stages analyzed, but a treatment between E3.5 and E4.5 was sufficient to reduce the number of PrE cells. This indicates that PrE cells are sensitive to Gleevec treatment between E3.5 and E4.5, which corresponds to the period when they are already specified and become sorted to form the PrE layer on the surface of the ICM.

Increased Cell Death in PrE Cells Treated with Gleevec

To determine whether the reduction in PrE cell number observed upon Gleevec treatment was due to cell death, we live imaged the development of early- and mid-blastocyst stage Pdgfra+/H2B-GFP embryos cultured with or without 5 µM Gleevec (Fig. 4A; Supporting Information Movies 2–4). We first observed that early-blastocyst embryos exhibited a significantly higher proportion of cell death when cultured in the presence of Gleevec (8.1 ± 0.6 dead cells, n = 7 vs. control 2.9 ± 0.7 dead cells, n = 10). We next analyzed the position of GFP+ cells (GFP-positive) before their death (Fig. 4A–4D). As in control embryos (55% dying cells were localized inside the ICM), cell death was observed both in inner and outer GFP+ cells of treated embryos (51% and 48% dying cells localized inside the ICM at early and mid-blastocyst stages, respectively). These data further suggest that as with the genetic ablation of Pdgfra, Gleevec treatment affects the survival of PrE cells independently of their position within the ICM.

image

Figure 4. Gleevec treatment phenocopies Pdgfra inactivation. (A): Single frames from three-dimensional time-lapse sequence of inner cell mass (ICM) of a Pdgfra+/H2B-GFP mid E3.5 blastocyst embryo cultured in 5 µM Gleevec. Cell death events occurring in green fluorescent protein (GFP)-positive cells are depicted either with a red (inner primitive endoderm [PrE] cell), a blue (outer PrE cell), or a gray arrowhead (undetermined localization). Annotated movie sequence is provided as Supporting Information Movie 4. GFP, green. Scale bar = 10 µm. (B–D): Localization of PrE cells before cell death within the ICM of control early E3.5 (B), Gleevec-treated early (C), and mid (D) E3.5 blastocyst embryos. n, number of cells analyzed. Abbreviations: GFP, green fluorescent protein; ICM, inner cell mass; IN, inside; ND, not determined; OUT, outside.

Download figure to PowerPoint

Caspase Inhibition Rescues the Phenotype Resulting from Loss of PDGF Signaling

As the absence of PDGF signaling correlates with increased apoptosis of PrE cells, we tested whether we could rescue the phenotype in mutant embryos by culturing them in presence of a pan-caspase inhibitor (Z-VAD [30]) (Fig. 5). Z-VAD treatment did not significantly affect the number of TE cells or EPI cells in wild type, heterozygous and mutant embryos (Fig. 5C, 5D). It also did not affect the number of PrE cells in wild type and heterozygous embryos (Fig. 5E). However in mutants, the number of PrE cells was significantly increased following Z-VAD treatment (4.8 ± 1.0 PrE cells in control vs. 10.6 ± 1.6 in treated mutant embryos) (Fig. 5E). Taken together, these data suggest that Pdgfra-deficient PrE cells undergo caspase-dependent apoptosis.

image

Figure 5. Caspase inhibition rescues the phenotype resulting from the loss of Pdgfra. (A): Three-dimensional rendering of Pdgfra+/H2B-GFP (heterozygous [HET]) and PdgfraH2B-GFP/H2B-GFP (mutant [HOM]) early blastocyst embryos cultured 24 hours in dimethyl sulfoxide (DMSO) or 20 µM pan-caspase inhibitor (Z-VAD). GFP, green; GATA4, red; Hoechst, blue. Scale bar = 20 µm. (B–E): Distribution of total (B), trophectoderm (C), pluripotent epiblast (D), and primitive endoderm (E) cells in wild type, HET, and HOM embryos cultured in DMSO or 20 µM Z-VAD. n, number of embryos analyzed. Statistical Mann–Whitney tests are indicated when significant (*, p < .05; **, p < .01). Error bars indicate SEM. Abbreviations: DMSO, dimethyl sulfoxide; EPI, pluripotent epiblast; GFP, green fluorescent protein; HET, heterozygous; HOM, mutant; PrE, primitive endoderm; TE: trophectoderm; WT, wild type.

Download figure to PowerPoint

Manipulating PDGF Activity Does Not Affect Cell Sorting

Several hypotheses could explain why in the absence of PDGF signaling, PrE cells exhibited cell survival defects. We favor two, non mutually exclusive scenarios. First, PDGF may provide survival signals operating during PrE specification. Alternatively, PDGF signaling might be involved in PrE cell sorting, and so in mutant embryos, PrE cells failing to sort consequently undergo caspase-dependent cell death as proposed previously [14, 17].

To test these two possibilities, we first analyzed the localization of PrE cells in embryos treated with or without Gleevec. We did not observe any major differences in their localization, such that PrE cells were usually found in contact with the blastocoel cavity (Supporting Information Fig. S1). Thus, Gleevec treatment did not affect PrE cell localization at the time of cell sorting. We next analyzed the localization of PrE cells in embryos treated with Z-VAD [30]. We did not observe notable defects in PrE cell sorting. Taken together, these data suggest that, rather than being involved in lineage segregation, PDGF signaling provides the necessary survival signals during PrE specification.

To exclude the possibility that PDGF signaling might be involved in PrE segregation, we analyzed the localization of PrE cells in embryos that ectopically expressed the ligand PDGF-A. To do this, we coelectroporated wild type early blastocysts with two plasmids, one driving constitutive expression of a histone H2B-mCHERRY fusion and a second driving constitutive expression of the PDGF-A ligand. We cultured electroporated embryos for 24 hours (Fig. 6A). We detected mCHERRY fluorescence in most of the cohort of electroporated embryos (29 out of 32 embryos) with fluorescence restricted to TE cells (Fig. 6C). We observed an average of 11.2 ± 1.8 and 9.5 ± 1.0 CHERRY-positive cells per embryo (Fig. 6B). We noted that in all cases, PrE cells were in contact with the blastocoel cavity indicating that cell sorting had proceeded normally. Thus, an ectopic source of PDGF-A did not affect PrE segregation, confirming the idea that PDGF is not involved in lineage segregation.

image

Figure 6. Overexpression of PDGF-A ligand does not affect sorting of primitive endoderm cells. (A): Schematic representation of plasmid DNA electroporation. (B): Average number of mCHERRY-positive cells per embryo. n, number of electroporated embryos. Error bars indicate SEM (C) Three-dimensional -rendering of blastocyst embryos coelectroporated with pCAG::H2B-mCherry and pCAGGS or pCAG::Pdgfa vectors. GATA4, green; mCHERRY, red; Hoechst, blue. Scale bar = 20 µm.

Download figure to PowerPoint

Manipulating PDGF Activity Does Not Affect Cell Lineage Specification

Cell fate decisions within the ICM are dependent on MAPK/ERK activity such that high levels of activity are required to promote a PrE identity, while acquisition of an EPI identity correlates with low ERK activity [31, 32]. Several lines of evidence demonstrate that this activity is controlled by FGF RTK signaling in a non-cell autonomous manner [25, 33, 34]. Whether PDGF RTK signaling could also contribute to regulate MAPK/ERK activity by the time of PrE specification remains an open question.

Previous studies did not report major defects in the distribution of blastocyst cell lineages in Pdgfra mutant embryos [22]. However, we noticed that early blastocyst stage embryos cultured with 10 µM Gleevec usually exhibited a reduced level of GATA4 (Supporting Information Fig. S1H). This observation could suggest defects in the kinetics of the PrE specification program in presence of a high-dose of Gleevec, resulting in delayed development and an embryo with an overall decrease in size. Therefore, we determined whether increasing PDGFRα activity, by providing an excess of its cognate PDGF-A ligand, would affect EPI/PrE specification. We cultured wild type E2.5 embryos for 2 days in the presence of 500 ng/ml PDGFA and then determined the distribution of cells within the different lineages (Supporting Information Fig. S2A, S2B). We failed to detect any significant changes in the proportion of TE (77.4 ± 3.6 cells in control vs. 67.6 ± 3.4 in treated conditions), PrE (15.9 ± 1.7 vs. 16.4 ± 1.3), and EPI cells (10.8 ± 1.3 vs. 9.8 ± 1.4). Thus, the exogenous addition of PDGF-A ligand does not affect the PrE/EPI ratio within the ICM.

We next reasoned that under normal conditions a high level of FGF signaling activity may mask a possible role of the PDGF pathway in PrE specification. As Fgf4 mutant embryos lack a PrE endoderm layer [24, 35, 36], we compared the distribution of ICM lineages in E4.5 implanting Pdgfra+/+, Pdgfra+/H2B-GFP, and PdgfraH2B-GFP/H2B-GFP blastocyst stage embryos having one or two functional Fgf4 alleles (Supporting Information Fig. S3). We first noted that, as we reported previously [24], removing one Fgf4 functional allele significantly reduced the number of PrE cells and slightly increased the number of EPI cells. The reduction in PrE cell number in Fgf4−/+ compared with Fgf4+/+ was of the same order of magnitude in Pdgfra+/+, Pdgfra+/H2B-GFP, and PdgfraH2B-GFP/H2B-GFP embryos (fold changes: −1.8 in Pdgfra+/+, −1.5 in Pdgfra+/H2B-GFP, and −1.9 in PdgfraH2B-GFP/H2B-GFP). Similarly, the absence of PDGF signaling led to a comparable reduction in the number of PrE cells in Fgf4+/+and Fgf4−/+ embryos (fold changes: −2.7 and −2.9, respectively). Taken together, these data demonstrate that PDGF is not involved in PrE specification and suggest that FGF and PDGF RTK signaling pathways exert distinct functions during PrE formation.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

In this study, we have investigated the role of PDGF RTK signaling during the specification and segregation of embryonic and extraembryonic lineages within the early mouse embryo. Ex vivo, PDGF signaling is required for XEN cell maintenance as well as during the conversion of mouse ES cells to cXEN cells [22, 37]. In vivo, we reported previously a significant reduction in the number of PrE cells in implanting mutant embryos, a defect that was exacerbated when implantation was artificially delayed through induction of a period of diapause [22]. In this study, we have identified an important role for PDGF signaling in PrE cell survival. We show that genetic inactivation of Pdgfra and pharmacological inhibition of PDGF activity specifically induce apoptosis of PrE cells. This process of programmed cell death is mediated by caspase activity, as we noted an increase in the number of Caspase-3-positive cells in Pdgfra-deficient embryos. We further demonstrated that inhibition of caspase activity rescued the defect associated with a reduction in PrE cell number observed in Pdgfra mutant embryos.

As apoptosis has been proposed as a mechanism for the removal of cells that are not properly specified and/or positioned within the ICM at the time when EPI and PrE cells are segregating into distinct tissue layers [14, 17], it was possible that in the absence of PDGF signaling apoptosis might affect a subset of PrE cells based on their position within the ICM. Using live imaging, we noted that both genetic and pharmacological inhibition of PDGF signaling activity affected the survival of PrE cells independently from their position within the ICM. As Gleevec inhibits not only PDGF receptor activity, but also additional RTKs including KIT and ABL [29], whether these RTKs participate in early cell lineage specification remains an open question.

Using a pharmacological approach, we determined more precisely the period when PDGF signaling is critical for PrE cell survival. We first noted that the effect was similar when Gleevec was applied to E2.5 and mid E3.5 embryos. In addition, using live imaging, we observed that PrE cell death occurred earlier in the movies of Gleevec-treated mid-blastocyst stage embryos compared with treated early-blastocyst embryos. Therefore, these data suggest that PDGF signaling is required for cell survival at the time when committed PrE cells segregate from EPI cells. Interestingly, a previous study used computational modeling to simulate the process of PrE/EPI segregation and noted that while cell sorting and cell position parameters increased the probability score of the model, changes in the parameter for apoptosis did not improve it [17]. Therefore, it is possible that apoptosis does not function as a key mechanism driving cell sorting per se, but that it might act to refine existing patterning information, thereby resulting in the formation of two spatially distinct ICM compartments, a process that we propose is dependent on PDGF signaling.

In addition to its role in cell survival, PDGF signaling has been implicated in chemotaxis and cell migration in various cellular contexts [38-44]. Therefore, it was possible that the reduction in PrE cell number was a secondary consequence of an impairment of ICM cells to properly sort. Indeed, in a model where PDGF signaling could be involved in PrE cell sorting, absence of PDGF signaling would impair proper PrE segregation and, as a consequence, mislocalized PrE cells would undergo apoptosis. In early mouse development three observations argue against this being the case : (i) the presence of exogenous PDGFA proteins, or (ii) ectopic source of PDGFA ligand in TE cells do not affect PrE cell sorting, and (iii) rescued PrE cells lacking PDGF signaling by inhibition of caspase activity properly locate at the surface of the ICM. Therefore, we propose, that PDGF signaling is involved in PrE cell survival at the time of cell sorting, but it does not play a role in the process of cell sorting itself.

It is interesting to note that Pdgfra is, together with Gata6, an early marker of PrE identity but seems required only when ICM cells have committed to a PrE fate. As FGF RTK signaling is a prominent determinant in ICM cell fate decisions (for review, see ref. [ [16]), it was possible that an earlier role of PDGF RTK in PrE specification could have been masked by FGF. Our results indicate that this is unlikely to be the case as embryos cultured in the presence of high doses of exogenous PDGF-A develop similarly to control embryos. In addition, we observed a similar reduction in PrE cells in embryos lacking one functional allele of Fgf4 in wild type and Pdgfra mutant embryos. Taken together, we propose that the successive steps in PrE formation (cell lineage specification and survival, sorting and expansion) are dependent on MAPK activity which in turn is controlled by FGF and PDGF RTK signaling activities respectively. It remains to be determined whether additional RTKs operate in these successive events.

We and others reported previously an increase in apoptosis at the transition from mid-to-late blastocyst stages that has been proposed to participate to the formation of two distinct EPI and PrE lineages within the ICM [6, 14, 17, 45]. Interestingly, Pdgfra is a target of GATA transcription factors and we have proposed previously that initiation and maintenance of Pdgfra expression might be regulated by GATA6 and GATA4 respectively [22, 46]. As such, it is tempting to speculate that failure to properly specify a PrE fate might directly affect the levels of PDGF activity leading to cell death. This mechanism would facilitate the removal not only of cells that failed to properly activate the PrE program, but also of PrE cells that are incorrectly positioned within the ICM. Therefore, it will be important to test this model in various situations where the PrE program is initiated, but fails to progress, and where committed PrE cells fail to properly sort.

Conclusion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

In this study, we have uncovered a role for PDGF signaling in the survival of PrE cells at the time when they segregate from to the pluripotent epiblast to form a distinct compartment in contact with the blastocoel cavity. Absence of PDGF signaling activity using genetic and pharmacological approaches specifically affects the PrE lineage and induces a caspase-dependent selective apoptosis. In addition, we demonstrate that FGF and PDGF RTK signaling exert distinct roles in PrE lineage specification and survival, respectively.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

We thank Peter Besmer for Gleevec and Claire Chazaud for technical advice on embryo electroporation. We thank Sigolène Meilhac for critical reading of the manuscript. Work in A.-K.H.'s laboratory is supported by the Human Frontier Science Program, NIH (R01-HD052115 and R01-DK084391) and NYSTEM. J.A. is supported in M.C.-T.'s laboratory by the European program Marie Curie (International Incoming Fellowship, Seventh European Community Framework Programme). M.C.-T.'s laboratory is supported by the Institut Pasteur, the CNRS and the ANR “Laboratoire d'Excellence” program (REVIVE, ANR-10-LABX-73-01).

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Acknowledgments
  9. Disclosure of Potential Conflicts of Interest
  10. References
  11. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
stem1442-sup-0001-suppfig1.tiff5525KSupporting Information Figure 1
stem1442-sup-0002-suppfig2.tiff8901KSupporting Information Figure 2
stem1442-sup-0003-suppfig3.tiff10613KSupporting Information Figure 3
stem1442-sup-0004-suppinfo1.mov1585KSupporting Information
stem1442-sup-0005-suppinfo2.mov1456KSupporting Information
stem1442-sup-0006-suppinfo3.mov1987KSupporting Information
stem1442-sup-0007-suppinfo4.mov2282KSupporting Information

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.