Visualization of the Epiblast and Visceral Endodermal Cells Using Fgf5-P2A-Venus BAC Transgenic Mice and Epiblast Stem Cells

Fibroblast growth factor 5 (Fgf5) has been widely used as a marker for the epiblast in the postimplantation embryo and epiblast stem cells (mEpiSCs) in the mouse, making it valuable for study of differentiation of various tissues and epiblast cells in vivo and in vitro. Here, we report for the first time the generation of Fgf5-P2A-Venus BAC transgenic (Tg) mice and show that the BAC Tg can recapitulate endogenous Fgf5 expression in epiblast and visceral endodermal cells of E6.5 and 7.5 embryos. We also show that Fgf5-P2A-Venus BAC Tg mEpiSCs in the undifferentiated state expressed abundant Venus, and upon reprogramming into naïve state, Venus was suppressed. Furthermore, while most Tg mEpiSCs expressed Venus abundantly, surprisingly the Tg mEpiSCs contained a minor subpopulation of Venus-negative cells that were capable of conversion to Venus-positive cells, indicating that even Fgf5 expression shows dynamic heterogeneity in mEpiSCs. Taken together, Fgf5-P2A-Venus BAC Tg mice and mEpiSCs generated in this study will be useful for developmental biology as well as stem cell biology research.


Introduction
Mouse embryonic stem cells (mESCs) are the first pluripotent stem cell type that was derived from the inner cell mass of the developing blastocyst [1,2]. Self-renewal and pluripotency are the defining features of mESCs, meaning that these cells can be maintained indefinitely in culture while retaining their ability to differentiate into all cell lineages of an adult organism. It is well-known that the core pluripotency transcription factor network formed by Oct3/4, Sox2, and Nanog is connected with extracellular signaling pathways, such as leukemia inhibitory factor (LIF), bone morphogenetic protein, and Wnt, which shields mESCs from differentiating stimuli [3][4][5]. mESCs can be grown either in conventional medium supplemented with LIF and serum or in serum-free medium containing dual inhibitors (known as 2i) for mitogen activated protein kinase (Mapk) and glycogen synthase kinase-3 (Gsk3) [6].
Subsequent studies led to the establishment of another pluripotent stem cell type, termed epiblast stem cells (mEpiSCs), which are isolated from the postimplantation mouse epiblast [7,8]. Unlike mESCs whose pluripotency relies on LIF/Janus-associated kinase-signal transducer and activator of transcription 3 (Jak-Stat3) signaling, mEpiSC self-renewal is dependent on basic fibroblast growth factor (bFGF) and Activin/transforming growth factor beta (TGFβ) signaling. In addition, when injected back into the host blastocyst, mESCs highly contribute to chimera formation while only a very small fraction of mEpiSCs analogous to the early postimplantation epiblast can do so [9]. However, a recent study reported that mEpiSCs could readily form chimeras including germ cell lineage provided they were grafted to gastrulating embryos that retained pluripotency of the postimplantation epiblast [10]. Thus, the inherent discrepancies in colony morphology, molecular and epigenetic status and chimera formation support the notion that mESCs and mEpiSCs are representatives of distinct pluripotent states termed naïve and primed pluripotency, respectively [11]. Interestingly, these naïve and primed pluripotent states can be interconverted in defined culture conditions. Naïve mESCs can achieve a primedlike state by stimulating bFGF and Activin/TGFβ signaling while mEpiSCs can be reprogrammed back into a naïve-like state by a combination of 2i/LIF and forced expression of pluripotency-related factors, such as Nanog, Esrrb, Klf2, Klf4 or Klf5 [12][13][14][15][16].
Heterogeneity is an inherent feature of mESCs when grown in the conventional culture condition containing LIF and serum [17][18][19][20]. mEpiSCs also exhibit heterogeneous expression of Oct3/4 and T (also known as Brachyury), resulting in differences in differentiation potential of subpopulations [9,21]. Oct3/4-negative mEpiSCs could not incorporate into the host blastocyst for chimera contribution, whereas a very small fraction of Oct3/4-positive mEpiSCs harboring distal enhancer activity of Oct3/4 could efficiently form chimeras [9]. Furthermore, while Tpositive mEpiSCs were prone to differentiate towards mesoderm and endoderm fates, a feature similar to that of in vivo epiblast cells that ingress through the primitive streak during gastrulation process, T-negative mEpiSCs had a propensity to give rise to the neuroectoderm cell lineage [21].
Fibroblast growth factors (FGFs) are structurally related proteins comprising 22 members in mammals [22]. The interactions between FGFs and FGF receptors (FGFRs) play important roles in regulating a wide variety of biological processes, ranging from modulation of tissue repair, inflammation, cell proliferation, survival and differentiation [23] to pluripotency and lineage specification [24] and regulation of energy expenditure [25]. Among the FGF family members, Fgf5 is transiently expressed at different stages of the developing embryo [26]. Subsequent studies proposed a potential role of Fgf5 in the process of gastrulation through stably maintaining the mobility of cells subjected to become the prospective embryonic germ layers [27][28][29][30]. Fgf5 has since been used as a marker for epiblasts in pre-streak and streak stages of mouse embryos [31][32][33]. Fgf5 is also strongly expressed in mEpiSCs [7,15,34,35], whereas it is hardly detectable in mESCs [36]. These findings implicated Fgf5 as a valuable marker for differentiation study of various tissues and epiblast cells in vivo and in vitro. Therefore, generation of an animal model mimicking Fgf5 expression in vivo and in vitro would be useful for a better understanding of epiblast cells and other biological events occurring during development as well as cell fate decision made by mEpiSCs.
Here we report for the first time the generation of Fgf5-P2A-Venus BAC (bacterial artificial chromosome) transgenic (Tg) mice to trace Fgf5 expression during early embryonic development. Our results show the recapitulation of endogenous Fgf5 expression governed by Fgf5-P2A-Venus BAC Tg in the postimplatation epiblast and visceral endodermal layer of E6.5 and E7.5 embryos as well as in mEpiSCs. Furthermore, surprisingly, while most Tg mEpiSCs expressed Venus abundantly, Tg mEpiSCs contained a minor subpopulation of Venus-negative cells that were capable of conversion to Venus-positive cells. This observation indicates that even Fgf5 expression shows dynamic heterogeneity in mEpiSCs. These will serve as valuable tools for marking Fgf5-expressing cells during development and studying lineage commitment initiated by mEpiSCs.

Material and Methods
Construction of the Fgf5-P2A-Venus BAC Tg The BAC clone (RP23-153I24) harboring the Fgf5 gene was purchased from Invitrogen (Carlsbad, CA, USA). For generation of the reporter cassette, the PGK-gb2-neo sequence flanked by FRT sites was ligated to a DNA fragment encoding P2A-Venus-ipacpA (Gene Bridges, Heidelberg, Germany). The cassette was then inserted into the first or third exon of Fgf5 by PCR amplification. The resulting BAC targeting vector and RED/ET expression plasmid (Gene Bridges) were co-transformed into Escherichia coli. After screening with Kanamycin, recombinants were identified by PCR analysis.

Generation of Tg mice
Recombinant BAC DNAs were purified with a NucleobondXtra BAC Kits (Macherey-Nagel, Düren, Germany) and then linearized by Pl-SceI digestion. Pronuclear injection was performed in fertilized eggs isolated from C57B6/J females, followed by transplantation into pseudo-pregnant ICR females (SLC Inc., Shizuoka, Japan). Tg mice were confirmed by PCR genotyping with the following primer sequences: 5 0 -TTCAAGGACGACGGCAACTACAAGAC-3 0 and 5 0 -GCTTCTCGTTGGGGTCTTTCTCAG-3 0 . The Tg mice were maintained on an ICR or B6 genetic background. This study was approved and conducted in accordance with the Regulations for Animal Experimentation of Shiga University of Medical Sciences.
Immunohistochemical analysis for mESCs and mEpiSCs was performed as previously described [16].
For reprogramming experiments, 2-4 × 10 4 cells were seeded onto fibronectin-coated 6-well plates in EpiSC culture conditions. After 24 h, the medium was switched to 2i/LIF conditions; the 2i inhibitors included 1 μM Mek inhibitor PD0325901 (Wako Pure Chemical Industries) and 3 μM Gsk3 inhibitor CHIR99021 (Wako Pure Chemical Industries). After 7 days, immunofluorescence analysis was performed. To check the characteristics of the resulting miPSCs, several miPSC colonies were picked up, expanded, and then cultured in 2i/LIF conditions for further experiments.

Reverse transcription (RT)-quantitative (q) PCR analysis
Total RNAs were extracted using the RNeasy Micro Kit (Qiagen, Hilden, Germany), followed by cDNA synthesis using the ReverTra Ace (TOYOBO CO., LTD, Osaka, Japan) according to the manufacturer's instructions. Real-time PCR was performed with the Thermal Cycler Dice Real Time System (Takara Bio Inc., Otsu, Shiga, Japan) and SYBR Premix EX Taq II (Takara Bio Inc.). Data were normalized against the expression of β-actin gene. Primer sequences are listed in S1 Table. Cell sorting Cells were dissociated by 0.25% trypsin-EDTA and resuspended in DMEM supplemented with 10% FBS. To exclude dead cells, the single cell suspension was incubated with propidium iodide (Cat #P3566, Molecular Probes Inc.) for 10 min on ice. Flow cytometry analysis was performed with FACSCalibur (Becton Dickinson Biosciences, San Jose, CA, USA). Cell sorting was performed with a FACSAria Fusion (Becton Dickinson Biosciences).

Statistical analysis
Student's t-test was applied for statistical analysis. Data are presented as means and standard errors. Statistical significance was determined at P < 0.05.

Generation of Fgf5-P2A-Venus BAC Tg mice
To generate transgenic (Tg) mice recapitulating endogenous Fgf5 expression, we took advantage of the enhanced yellow fluorescence protein Venus, which possesses valuable features for visualization, such as quick maturation and resistance to acidosis [40]. As the first trial, a BAC clone was used to cover the entire genomic region of Fgf5 with a modification at the first exon in which an in-frame fusion Venus was implemented right after the start codon (S1 Fig). Although we obtained six Tg lines, Venus expression was not found in the epiblast of E6.5 and 7.5 embryos (data not shown). Because we anticipated that the reporter cassette perturbed potential regulatory regions located around the first exon and intron, we inserted the P2A (porcine teschovirus-1 self-cleaving peptide)-Venus reporter cassette into the third exon of Fgf5 in the BAC clone (Fig 1A), and generated six lines of Fgf5-P2A-Venus BAC Tg mice. To check Venus expression, we dissected embryos at E6.5 and 7.5 from 6 Tg lines, and found strong Venus expression in the epiblast of all Tg lines (Fig 1B and 1C, data not shown), indicating that Fgf5-P2A-Venus BAC Tg construct can efficiently direct expression in the epiblast.

Venus expression in Fgf5-P2A-Venus BAC Tg embryos
The observation that Venus expression was detectable in the postimplantation epiblast of Tg embryos prompted us to examine detailed expression of Venus directed by the Tg. Embryos collected at E6.5 and E7.5 were stained for the epiblast marker Oct3/4 and endodermal marker Gata4 or mesodermal marker T (Figs 2 and 3). We found uniform expression of Venus in the epiblast of the Tg embryo at E6.5 (Fig 2A), consistent with previous reports [27][28][29][30]. Venus was also weakly seen in the visceral endodermal layer, in accordance with previous observations [27,28,30]. Higher magnification confirmed our observation that Venus and Oct3/4 expression were expressed uniformly throughout the epiblast, whereas Gata4 protein was only discernible in the visceral endoderm (Fig 2A). Importantly, fluorescent mRNA labeling by cytoplasmic FISH revealed the presence of endogenous Fgf5 mRNA signals in the epiblast and visceral endodermal layer of Tg embryos at E6.5 (Fig 2B). These results indicated that the Fgf5-P2A-Venus BAC Tg construct recapitulated endogenous Fgf5 expression in Tg embryos at E6.5. In Tg embryo at E7.5, Venus and Oct3/4 expression was found to be overlapping in the epiblast regions while Venus was also detected abundantly in the anterior visceral endoderm layer (Fig 3A), consistent with previous reports [28,41]. We also confirmed endogenous Fgf5 mRNA expression in the epiblast and visceral endodermal layer of Tg embryos at E7.5 (Fig 3B). Collectively, these results demonstrated that Fgf5-P2A-Venus BAC Tg is capable of recapitulating endogenous Fgf5 expression in the postimplantation epiblast and visceral endodermal layer.
While epiblast cells that ingress through the primitive streak can form the mesoderm and endoderm, epiblast cells that do not traverse the primitive streak can give rise to the ectoderm [42,43]. It is of note that at E8.25, we observed Venus expression in the neuroepithelium of the forebrain (S2 Fig), consistent with Venus expression in the anterior epiblast at E7.5.
Derivation and characterization of Fgf5-P2A-Venus BAC Tg mEpiSCs mEpiSCs represent a primed pluripotent state that can be utilized as a useful model for studying biological events that take place during the transition from the primed to naïve state and vice versa. To confirm the Fgf5-P2A-Venus BAC Tg expression in vitro, we established mEpiSCs, by culturing the epiblast of E6.5 Fgf5-P2A-Venus BAC Tg embryos in NDiff227 medium supplemented with bFGF and Activin. Immunofluorescence analysis revealed that the Tg mEpiSCs expressed pluripotency markers Oct3/4 and Nanog (Fig 4A). RT-qPCR analysis showed that Oct3/4 expression is significantly higher in the Tg mEpiSCs than in mESCs, while expression levels of Nanog and Sox2 were lower in the Tg mEpiSCs than in mESCs. Notably, Fgf5 and Sox1 expression were detected at much higher levels in the Tg mEpiSCs than in mESCs (Fig 4B). Flow cytometric analysis showed that most Tg mEpiSCs, if not all, expressed Venus, consistent with uniform Venus expression in the epiblast of the E6.5 Tg embryo ( Fig  4C). Taken together, these results indicated that Fgf5-P2A-Venus BAC Tg mEpiSCs have similar properties to bona fide mEpiSCs. Reprogramming of Fgf5-P2A-Venus BAC Tg mEpiSCs into miPSCs Reprogramming of mEpiSCs into miPSCs can be accomplished by ectopic expression of Nanog, Esrrb, Klf2, Klf4 or Klf5 cultured in the presence of 2i/LIF. [12][13][14][15][16]. The reprogramming process upregulates many naïve state-associated markers containing Nanog, Esrrb, Tfcp2l1, Cd31, Rex1, Stella, Nr0b1, Prdm14, Nr5a2, Tbx3, Klf2, Klf4 and Klf5, with a parallel rapid reduction in the expression of primed state-associated markers such as Fgf5. Next, we asked whether Venus fluorescence is suppressed during the reprogramming process toward miPSC state. After ectopic expression of Nanog or Klf5 in the Tg mEpiSCs, the culture medium was changed Higher magnification of optical sections is shown in panels X and Y. Note that Oct3/4 and Venus were coexpressed in the epiblast of the Tg embryo, while T expression was observed in the mesodermal layer of the Tg embryo. Venus expression was also seen in visceral endodermal layer. Open and closed arrowheads indicate mesodermal and epiblast cells, respectively. Ex: extraembryonic region; Em: embryonic region; Epi: epiblast; Me: mesoderm; T: Brachyury/T. Scale bar: 50 μm. All images were captured by a Leica TCS-SP8 confocal microscope using a 20x/0.7 dry objective lens (projection images) and 40x/1. from bFGF and Activin to 2i/LIF (Fig 5A). We found that ES-like colonies emerged in the Klf5and Nanog-overexpressing Tg mEpiSCs within 5-7 days after addition of reprogramming medium. Immunofluorescence analysis revealed that overexpression of Klf5 or Nanog could reactivate the expression of CD31 (also known as PECAM-1: platelet endothelial cell adhesion molecule-1), a useful marker of inner cell mass cells, and increase Nanog expression in the miPSCs (Fig 5B). Importantly, Venus expression in the Tg mEpiSCs was markedly reduced in the miPSCs (Fig 5B). To further validate the characteristics of the miPSCs, several miPSC colonies were randomly picked to generate miPSC lines. These miPSC lines were maintained in 2i/ LIF conditions and used for further experiments. RT-qPCR analysis showed upregulation of pluripotency factors Nanog, Rex1, Esrrb, Tfcp2l1, Klf2, Klf4, and Klf5, and downregulation of lineage commitment factor Fgf5 in miPSCs (Fig 5C). Taken together, these results demonstrated that Venus expression can be used as an indicator when the Tg mEpiSCs are forced to reprogram into miPSCs.
Dynamic heterogeneity of Fgf5 expression in Fgf5-P2A-Venus BAC Tg mEpiSCs mEpiSCs consist of several subpopulations: T-positive and -negative populations, and also Sox1-positive and -negative populations; these positive/negative populations are interconverted [21]. Although we found that Fgf5-P2A-Venus BAC Tg embryo showed uniform Venus expression in the epiblast and most Tg mEpiSCs expressed Venus abundantly, we investigated whether the Tg mEpiSCs contained a Venus-negative population. Careful examination of flow cytometry revealed that a minor fraction of the Tg mEpiSCs was Venus-negative (about 4%) (Fig 4C). To further explore this phenomenon, we purified Venus-positive and -negative mEpiSCs by cell sorting (Fig 6A) and cultured each cell fraction in mEpiSC growth conditions to investigate the ability to re-establish heterogeneity from each subpopulation. Interestingly, Venus-positive cells quickly emerged from the sorted Venus-negative cells, and the sorted Venus-negative cells could re-establish the original cell state within 4 days in culture (Fig 6A). Similarly, the purified Venus-positive cells also generated Venus-negative cells, although the re-establishment of heterogeneity from Venus-positive cells occurred more slowly compared with that from Venus-negative cells (Fig 6A). These results indicated that the Tg mEpiSCs contain at least two subpopulations that can be interconverted. We also examined gene expression patterns in Venus-positive and -negative mEpiSCs by RT-qPCR analysis (Fig 6B). We confirmed that Fgf5 mRNA expression was enriched in Venus-positive cells (Fig 6Ba). Importantly, we found that the expression levels of other epiblast markers, Oct3/4, Nanog, and Sox2, were predominantly detectable in Venus-positive cells (Fig 6Ba). Furthermore, both subpopulations expressed very low levels of naïve pluripotency markers Rex1, Esrrb, Tfcp2l1, and Klf2 compared with mESCs (Fig 6Bd). Interestingly, we found that mesendodermal markers T, Sox17 and Foxa2 were expressed at significantly higher levels in Venus-positive cells than in Venus-negative cells (Fig 6Bb). On the other hand, among the tested ectodermal markers, only Sox1 expression was enriched in Venus-negative cells relative to Venus-positive cells (Fig 6Bc). Taken together, while it was thought that Fgf5 marks mEpiSCs uniformly, our experiments clearly demonstrated that even Fgf5 expression shows dynamic heterogeneity in mEpiSCs.

Discussion
In this study, we demonstrated that Fgf5-P2A-Venus BAC Tg mice can recapitulate endogenous Fgf5 expression in the postimplantation epiblast and visceral endoderm. To the best of our knowledge, this is the first Tg mouse model allowing for the visualization of endogenous Fgf5 expression during early embryonic development. Initial Tg lines carrying Venus at the first exon failed to drive epiblast Venus expression while Tg lines carrying P2A-Venus in the place of the stop codon drove strong Venus expression in the epiblast of all Tg lines. The exact reason underlying these observations is currently not clear. Because many genes harbor important regulatory elements around the first intron, the insertion of the Venus-pA reporter cassette into the first exon could have potentially abolished transcription by perturbing the epiblast promoter/enhancer elements.
Previous studies showed that primitive streak formation in the posterior portion of epiblast is a crucial event through which the body plan is established during the gastrulation process [44]. More specifically, while epiblast cells that ingress through the primitive streak can form the mesoderm and endoderm, epiblast cells that do not traverse the primitive streak can give rise to the ectoderm [42,43]. Future work will delineate the actual cell fates of Fgf5-positive epiblast cells, which would provide important insights into how the ectoderm lineage is established and regulated in the gastrulating mouse embryo.
Gastrulation begins when a population of epiblast cells is triggered to move to the primitive streak located in the posterior epiblast while the other cells remain in the epiblast. This cell movement leads to the formation of the primary germ layers, namely the ectoderm, mesoderm, and endoderm [43,44]. However, how epiblast cell movement is regulated and which factors stimulate and determine the fates of cell populations at the onset of gastrulation is not fully understood. Therefore, a detailed analysis of Fgf5 and T expression pattern will be required to understand the molecular basis of epiblast cell behavior during gastrulation. mESCs consist of several subpopulations; the subpopulations expressing either PECAM1, Rex1 or Stella efficiently form chimeric animals when injected into blastocysts. mEpiSCs are also heterogeneous in terms of gene expression: T-positive cells are primed to differentiate into mesoderm and endoderm lineages, while T-negative cells are primed to ectoderm [21]. Our data showed that Fgf5 overall uniformly marks mEpiSCs, consistent with the previous report [21]. However, surprisingly, there is a small subpopulation of Fgf5-negative cells in mEpiSCs. Our gene expression analysis revealed that, while Fgf5-positive cells predominantly expressed important mesendodermal markers such as T, Sox17 and Foxa2, Fgf5-negative cells exhibited a high expression level of the ectodermal marker Sox1. Currently, the actual cell type of Fgf5-Venus-negative cells is unknown, but these results suggest the possibility that the heterogeneous expression of Fgf5 observed in our study may be the foundation for the distinct differentiation biases of subpopulations in mEpiSCs. Thus, investigation of the differentiation propensity of Fgf5-negative and -positive mEpiSCs in response to differentiation stimuli will be of a great interest in future studies.
Previous studies indicated potential roles of Fgf5 in the progression of hepatic fibrosis [45], the process of gastrulation [27,28] and hair growth cycle [46], but the molecular basis of how Fgf5 manifests its functions has not been clearly understood. In addition, the impact of biological events on lineage commitment initiated by mEpiSCs is not known. Taken together, Fgf5-P2A-Venus BAC Tg mice and mEpiSCs established in our study may be used to investigate novel functions of Fgf5 as well as to unravel molecular mechanisms underlying lineage specification in vivo and in vitro. Immunofluorescence staining of the Tg embryo at E8.25 for Venus (green). Note that Venus was expressed in the neuroepithelium of the Tg embryo. Ne: neuroepithelium. Scale bar: 100 μm. (TIF) S1 Table. Primers used for RT-qPCR analysis.