Bovine trophectoderm cells induced from bovine fibroblasts with induced pluripotent stem cell reprogramming factors

Thirteen independent induced bovine trophectroderm (iBT) cell lines were established by reprogramming bovine fetal liver‐derived fibroblasts after viral‐vector transduction with either six or eight factors, including POU5F1 (OCT4), KLF4, SOX2, MYC, NANOG, LIN28, SV40 large T antigen, and hTERT. Light‐ and electron‐microscopy analysis showed that the iBT cells had epithelial cell morphology typical of bovine trophectoderm cells. Reverse‐transcription‐PCR assays indicated that all of the cell lines expressed interferon‐tau (IFNT) at passages 1 or 2. At later passages (≥ passage 8), however, immunoblot and antiviral activity assays revealed that more than half of the iBT cell lines had stopped expressing IFNT. Messenger RNAs specific to trophectoderm differentiation and function were found in the iBT cell lines, and 2‐dimensional‐gel analysis for cellular proteins showed an expression pattern similar to that of trophectoderm cell lines derived from bovine blastocysts. Integration of some of the human reprogramming factors, including POU5F1, KLF4, SOX2, MYC, NANOG, and LIN28, were detected by PCR, but their transcription was mostly absent in the iBT cell lines. Gene expression assessment of endogenous bovine reprogramming factor orthologs revealed endogenous bLIN28 and bMYC transcripts in all; bSOX2 and bNANOG in none; and bKLF4 and bPOU5F1 in less than half of the iBT cell lines. These results demonstrate that bovine trophectoderm can be induced via reprogramming factor expression from bovine liver‐derived fibroblasts, although other fibroblast populations—e.g., derived from fetal thigh tissue—may produce similar results, albeit at lower frequencies.


| INTRODUCTION
The "reprogramming" of somatic cells by specific transgene expression to convert them into induced pluripotent stem cells (iPSCs), or  (Takahashi & Yamanaka, 2006). Soon thereafter came the reported creation of human iPSCs through the use Mention of trade name, proprietary product or vendor does not constitute a guarantee or warranty of the product by the U.S. Department of Agriculture or imply its approval to the exclusion of other products or vendors that also may be suitable. of these and other so-called reprogramming factors, and, more importantly, the ability of mouse iPSCs to give rise to chimeric embryos with germ-line transmission in the resulting mice (Okita, Ichisaka, & Yamanaka, 2007;Takahashi et al., 2007;Wernig et al., 2007;Yu et al., 2007).
Reports describing a failure to fully reprogram ungulate fibroblasts using OSKM plus other factors are rare. Thomson et al. described OSKM-mediated reprogramming of pig fibroblasts that yielded iPSClike cells, which were apparently not fully reprogrammed (Thomson et al., 2012). Despite the iPSC-like cells being positive for alkaline phosphatase (AP), a universal marker of pluripotency in embryonal carcinoma cells, ESC, and pig, sheep, and bovine epiblast cells (Nicolas et al., 1976;Talbot, Powell, & Rexroad, 1995;Talbot, Rexroad, Pursel, & Powell, 1993;Wobus, Holzhausen, Jäkel, & Schöneich, 1984), in vitro testing of these pig iPSC-like cell lines only demonstrated their differentiation into cells expressing neuronal markers or cells positive for alpha-fetoprotein expression, which could be indicative of either yolk-sac endoderm cells or hepatocytes (Thomson et al., 2012).
Indeed, injection of these porcine iPSC-like cells into immunocompromized mice, which tests for teratoma formation, resulted in tumors containing tissue of a "predominantly glandular phenotype" (Thomson et al., 2012). Another recent report documented the apparently incomplete OSKM-mediated reprogramming of goat fibroblasts into goat iPSC-like cells that differentiated into neuronal cells, yolk-sac endoderm, and trophectoderm (Sandmaier et al., 2015).

Induction of other cell types besides ESC-like cells was noted
during OSKM-mediated reprogramming of pig fibroblasts to putative pig iPSCs (Ezashi et al., 2009). Ezashi et al. characterized the nature of this commonly occurring, non-ESC-like reprogrammed cell type, and showed that it was indistinguishable from the trophectoderm cells of the pig embryo (Ezashi, Matsuyama, Telugu, & Roberts 2011). These cells, designated as pig induced trophoblast (iTR), were flat, epithelial cells connected to one another by tight-junctions and associated desmosomes. Their epithelial nature and tight-junction attribute was evidenced by the cells' colony-outgrowths forming domes indicative of the transport of culture fluid basolaterally underneath the cell monolayer. These pig iTR cells expressed several developmental and differentiation genes associated with trophectoderm, including cytokeratin 7 (KRT7), GATA2, PPARG, MSX2, DLX3, HAND1, GCM1, CDX2, ID2, ELF5, TCFAP2C, and TEAD4. Furthermore, the cells expressed trophectoderm-related functional genes of secreted proteins (i.e., PAG6 and PAG10), steroid synthesis (HSD17B1, CYP11A1, and STAR), mediators of inflammation (IFNG and IL1B) and prostaglandin synthesis (PTGES, COX1, and COX2). Thus, porcine fibroblasts clearly could be reprogrammed into cell types other than pluripotent cells by ectopic expression of OSKM-in this case, into extraembryonic trophectoderm cells.
Outlined in this report is our attempt to create bovine iPSC, as previously reported by others (Cao et al., 2012;Cravero et al., 2015;Han et al., 2011;Kawaguchi et al., 2015;Nong et al., 2015;Sumer et al., 2011;Talluri et al., 2015), using similar reprogramming factors, vectors, and culture methods. While bovine iPSC-like cells were produced in our experiments, these reprogrammed cells did not express high levels of AP. A second epithelial cell type was observed in some of the reprogramming experiments, many of which were AP-positive but did not resemble iPSCs in morphology. Several of these non-iPSC-like cells were independently colony-cloned and factor-induced colonies, whose cell and colony morphology were "ESC-like," were also observed, but these were invariably negative for AP activity, so their characterization was not pursued further.
The JF3 liver-derived fibroblasts produced the greatest number of trophectoderm-like colonies (Figure 1). JF3 liver-derived fibroblasts produced numerous colonies, whereas the other bovine fetal fibroblasts, derived from fetal thigh tissue, and endothelium-like cells, derived from Wharton's jelly or amniotic fluid, produced few or no such colonies after 3 weeks of culture. For example, the 6-factor induction of JF3 liver-derived fibroblasts gave 14 trophectoderm-like colonies after 30 days of culture, whereas the same induction paradigm of JF3 Wharton's jelly-derived and amniotic fluid-derived cells resulted in only one or no induced colonies, respectively. JF2 cells TALBOT ET AL. | 469 derived from the liver, bone marrow, thigh tissue, and Wharton's jelly all failed to produce any trophectoderm-like or ESC-like colonies after transduction with either six or eight reprogramming factors. Comparing JF3 liver-derived fibroblasts to JF2 Wharton's jelly derived endothelium-like cells, the 8-factor induction paradigm resulted in a total of 125 trophectoderm-like colonies in the two primary induction plates of the JF3 liver-derived fibroblasts by Day 17 of culture, whereas the JF2 Wharton's Jelly-derived cells produced no colonies of any kind after 4 weeks of culture.
Fourteen independent colony isolations of trophectoderm-like cells from the JF3 fibroblast transductions were performed from the six-and eight-factor experiments to establish induced bovine trophectoderm-like (iBT) cell lines, and each colony was expanded to cultures of millions of cells on STO (Sandoz inbred strain, thioguanine-and ouabain-resistance) feeder cells in 10% Dulbecco's Modified Eagle Medium (DMEM), as previously described for bovine embryo-derived trophectoderm cell lines (Talbot et al., 2000). Thirteen independent iBT cell lines were established from the 14 colonycloning isolations. The remaining cells in the primary culture plates were stained for AP activity, revealing that most of the trophectoderm-like colonies were AP-positive (Figure 1).
At passage 1 (P1), and subsequent secondary passages on STO feeder cells, the iBT cell lines displayed the typical morphology of bovine trophectoderm cells: the cells grow as tightly knit, epithelial monolayers with the boundary of their expanding colonies pushing many of the feeder cells out of their path ( Figure 2) (Talbot et al., 2000).
The iBT cell lines lost their AP activity after secondary culture began (not shown). As observed in some of the primary colonies (Figure 1 (Talbot et al., 2000).

| Interferon-tau antiviral activity assay and protein expression in iBT cell lines
Interferon-tau (IFNT) antiviral activity was assessed in conditioned medium of the iBT cell lines (Table 1); the CT-1 bovine trophectoderm FIGURE 1 Low-magnification light micrograph of primary colonies of bovine trophectoderm induced from bovine fibroblasts. Dissection-scope image of primary reprogramming factor-induced colonies stained for AP activity (red color). Arrows indicate AP-positive colonies with dome formation that is indicative of basolateral fluid transport. Arrowheads indicate AP-negative colonies. Scale bar, 10 mm FIGURE 2 Light micrographs of iBT-6 cells cultured on STO feeder cells at passage 11. a, Phase-contrast micrograph showing the cells' roughly cuboidal shape, granular cytoplasm, and occasional prominent lipid droplets (arrows). Scale bar, 25 μm. b, Pseudodark-field micrograph showing the coalescing colony outgrowths of recently passaged cells as they grow out across the STO feeder layer. Arrows indicate the iBT-6 monolayer boundary with the STO feeder cells. Scale bar, 100 μm cell line, established from an in vitro-fertilized and cultured bovine blastocyst (Talbot et al., 2000), acted as a positive control. Assessment of the conditioned medium from the STO feeder cells alone, which all the iBT cell lines and the CT-1 cells are grown on, did not contribute to the measured IFNT antiviral activity ( Table 1B). All of the iBT cell lines, however, showed some IFNT activity at an early passage (P2) (Table 1A), although some progressively lost IFNT expression with increasing passages (e.g., iBT-1, -7, -9, and -11) (Table 1B). While the levels of IFNT in the conditioned medium may reflect relative quantitative secretion of IFNT by the different iBT cell lines in some cases, such conclusions cannot be drawn from the data presented (Table 1)   Two consistent and prominent differences in protein spots were found comparing the proteins of iBT cells to those of FT and VIVOT cells (circled in Figure 5); however, both proteins were identified as bovine serum proteins-albumin and alpha-2-HS glycoprotein (fetuin-A). This differential expression may be an artifact of the culture conditions since these early secondary-passage iBT cells were maintained in iPSC medium; indeed, culturing the iBT-6 cell line in 10% fetal bovine serum in DMEM instead of iPSC medium, which contained 7.5% fetal bovine serum and 10% knockout serum replacement, resulted in the disappearance of the albumin and fetuin spots in the 2-dimensional gels of the crude iBT-6 lysate (not shown).
The FT19-1 and VIVOT-199 lines were cultured in 10% fetal bovine serum in DMEM prior to analysis, so this difference may be an artifact resulting from the inability to remove all of the serum-proteins bound to the cells by simple washing of the cell monolayers with serum-free medium prior to extraction with the urea/thiourea lysis buffer. The presumptive identity of many of the remaining protein spots can be ascertained by consulting our previously published two-dimensional gel analysis of bovine trophectoderm cell lines (Talbot, Powell, Caperna, & Garrett, 2010).
Expression of endogenous bovine POU5F1, SOX2, KLF4, MYC, LIN28, and NANOG orthologs was also assessed by semi-quantitative reverse-transcription PCR using bovine-specific primers (Table 3; Figure 9). Endogenous bPOU5F1 expression was not detected or showed relatively low abundance in the iBT cell lines, which is similar to the CT-1 cell line. Expression of bSOX2 was not detected in any of the iBT cell lines (Figure 9). Bovine KLF4 expression was found in most of the iBT cell lines, but, again, its expression was relatively low, with the exception of iBT-6; CT-1 cells also had relatively low bKLF4 expression. The iBT cell lines did not express bNANOG or bTERT ( Figure 9; the positive signals in the CT-1 sample lane were generated from CT-1 genomic DNA, not cDNA, to verify PCR primer specificity and function). bMYC and bLIN28 were expressed in all the iBT cell lines, as in CT-1 bovine trophectoderm cells (Figure 9).

| DISCUSSION
The most significant finding of the study was that stable trophectoderm cell lines were readily derived from the ectopic expression of reprogramming factors in bovine fibroblasts, particularly fibroblasts derived from fetal bovine liver tissue.
These iBT cell lines had characteristics similar to those of the bovine blastocyst-derived trophectoderm cell line, CT-1, including cell morphology, IFNT secretion (Talbot et al., 2000), and the expression of several trophectoderm marker genes, such as CDX2 (Berg et al., 2011;Schiffmacher & Keefer, 2013), ASCL2 (Guillemot, Nagy, Auerbach, Rossant, & Joyner, 1994), ELF5 (Donnison et al., 2005), GATA2, and GATA3 (Bai et al., 2011;Home et al., 2009). The iBT cell lines also expressed endogenous reprogramming factor FIGURE 7 PCR detection of viral vector gene integration of human reprogramming factors in the iBT lines. The CT-1 bovine trophectoderm cell line (Talbot et al., 2000) was included as a negative control. NTC, no-template control (total RNA from the iBT-1 cell line that was not treated with reverse transcriptase) FIGURE 8 Semi-quantitative reverse-transcription PCR for human reprogramming gene expression in iBT lines. The CT-1 bovine trophectoderm cell line (Talbot et al., 2000) was included as a negative control. The passages at which the cells were assayed were: iBT-1 Semi-quantitative reverse-transcription PCR for endogenous bovine reprogramming factor ortholog gene expression in iBT lines. The CT-1 bovine trophectoderm cell line (Talbot et al., 2000), derived from a bovine blastocyst, was included as a comparative control or positive control, either as cDNA or as genomic DNA (gDNA), respectively. cDNA of the CICM-3 cell line, a bovine epiblast-derived neuronal cell line (Talbot, Powell, & Garrett, 2002), was used as a positive control for bovine SOX2 expression. -RT, total RNA from the iBT-1 cell line that was not treated with reverse transcriptase TALBOT ET AL.

| 475
NANOG transcription was reportedly present in bovine trophectoderm/yolk sac-endoderm tissue from later-stage ovoid and elongated bovine blastocysts (Degrelle et al., 2005). At the protein level, however, NANOG was undetectable in the trophectoderm of laterstage bovine embryos (Maruotti et al., 2012). Ectopic hNANOG expression observed in our iBT-1 and iBT-3 lines did not appear to change the essential trophectoderm phenotype, although ASCL2 expression was absent in both lines, which correlates with the reported suppression of Ascl2 expression by Nanog in a mouse ESC loss-of-function system (Ivanova et al., 2006). As with POU5F1 gene expression, the relative abundance of NANOG compared to CDX2 in our iBT cell lines may be of most importance (Chen et al., 2009;Hyslop et al., 2005)-although species differences in NANOG expression among preimplantation embryo have been noted, and could complicate interpretations (Kuijk et al., 2008).
Exogenous hSOX2  Similarly, SOX2 expression is down-regulated in the trophectoderm of elongating bovine blastocysts (Degrelle et al., 2005). Therefore, the absence of SOX2 expression, ectopic or endogeneous, in the iBT cell lines is consistent with known ungulate trophectoderm gene expression.
KLF4 expression does not appear to be required for trophectoderm induction (Segre, Bauer, & Fuchs, 1999), whereas KLF5 is vital for its formation (Lin, Wani, Whitsett, & Wells, 2010). Expression of Klf4 so we cannot be certain that they occurred.
In our experiments, colonies of induced epithelial cells in primary culture usually acquired trophectoderm cell morphology and colony characteristics-e.g., the dome-formation clearly visible in AP-positive colonies [high AP expression was shown in the trophectoderm of early bovine blastocysts (Talbot et al., 1995)]. Thus, our major results contrast those of other published bovine iPSC-derivation reports, even given the possibility that rare AP-positive iPSC colonies were present, but somehow missed, in our primary transduction plates. By comparison, during the derivation of OSKM-induced pig trophectoderm cell lines, nascent trophectoderm colonies comprised a relatively minor proportion (17% of those picked) of the reprogrammed colonies . Similar to our observations with bovine fetal fibroblast reprogramming, the occurrence of colonies that were trophectoderm in nature depended on the cell type being reprogrammed-that is, when pig umbilical cord-derived fibroblasts rather than whole-body pig fetal fibroblasts were transduced with OSKM, 79% of the resulting reprogrammed colonies were judged to be trophectoderm-like .
Reprogramming results may vary according to the culture methods and medium employed. Variations in growth factor supplementation, usually involving leukemia inhibitory factor (LIF) and basic fibroblast growth factor (bFGF), and the choice of feeder cells employed are common among the existing bovine iPSC reports.
Exposure to compounds, such as inhibitors of glycogen synthase kinase three beta (GSK3B) and MAPK/ERK kinase (MEK) (i.e., 2i medium conditions), were also claimed to enhance bovine cell reprogramming (Heo et al., 2015;Kawaguchi et al., 2015). The various culture methods, growth factors, and small molecules we employed were expected to improve the chances of generating bovine iPSC cells (Gafni et al., 2013;Huangfu et al., 2008;Silva et al., 2008;Telugu et al., 2010) and to inhibit the formation and growth of trophectoderm cells (Kwiecińska, Wiśniewska, & Gregoraszczuk, 2011;Lu et al., 2008). necessary for an internal Nanog-GFP reporter to be activated and iPSC characteristics to be acquired (Kim et al., 2011). Somatic cell fate resulting from short-term expression of OSKM appeared to be dependent on extrinsic factors since, in related work from the same laboratory, exposure of cells to alternate medium constituents gave rise to cardiac muscle cells instead of neuronal progenitor cells (Efe et al., 2011). Another group also demonstrated direct induction of neural progenitor cell lines by limiting Pou5f1 expression in the initial phase of the mouse fibroblast OSKM reprogramming (Thier et al., 2012). Similarly, short-term OSKM exposure (4 days from which to learn more about the molecular interactions between reprogramming factors and trophectoderm-differentiation-related transcription factors, including CDX2, ELF5, and EOMES (Pfeffer & Pearton, 2012;Niwa et al., 2005;Schulz et al., 2008), and trophectoderm differentiation-related signal transduction factors, such as bone morphogenetic protein-4 (BMP4), nodal (NODAL), activin A (INHBA), and Fibroblast growth factors (Bernardo et al., 2011;Lee et al., 2011;Schulz et al., 2008;Tanaka, Kunath, Hadjantonakis, Nagy, & Rossant, 1998;Xu et al., 2002).

| Cell culture
Bovine fetal fibroblasts were established from two Jersey cow pregnancies, as previously described . The preparation of feeder-cells was as previously described (Talbot, Sparks, Powell, Kahl, & Caperna, 2012) Alkaline phosphatase staining of cells was performed as previously described (Talbot et al., 1995). For reprogramming experiments from which the iBT cell lines were established, medium conditioned for 24 hr by CF-1 feeder-cells was used. The iPSC medium employed was similar to those previously published Ying et al., 2008), and

| Interferon-tau antiviral assay
Forty-eight-hour conditioned medium (10% DMEM) was collected from T12.5 cultures of iBT lines at ≥50% confluency. The conditionedmedium samples were stored frozen at −75°C prior to the antiviral assay, which was performed as previously described (Roberts et al., 1989)  4.6 | Two-dimensional gel electrophoresis and tandem mass spectrometry The cellular proteins of the iBT lines were separated by twodimensional gel electrophoresis and identified by mass spectrometry, as previously described (Talbot et al., 2010). Cellular proteins were extracted from a single T25 tissue culture flask of each cell line, at 75-100% confluency.   following the manufacturer's suggestions for each, in a final reaction volume of 50 μl. The amplification conditions for human and bovine transcript analysis were the same as described above for genomic PCR.

| Confirmation of viral vector integration by PCR
The reverse-transcription PCR products from the DreamTaq protocol were mixed with 10 μl of 6X Purple Gel Loading Dye (New England BioLabs, Ipswich, MA) before separation on a 1.5% agarose Trisacetate-EDTA gel with 10 μg/ml ethidium bromide (Sigma-Aldrich), and then visualized on an ultraviolet transilluminator (Vilber Lourmat, Marne La Vallée, France). Gene-specific primers for bovine trophectoderm markers and reprogramming factors (below) were synthesized according to bovine-and human-specific sequences obtained from GenBank  (Table 3), were examined.

| Transmission electron microscopy
Transmission electron microscopy sample preparation and microscopy were performed with assistance of JFE Enterprises (Brookville, MD).
T12.5 flasks were washed with phosphate-buffered saline, fixed for 1 hr with 2.5% glutaraldehyde, and washed with and stored under Millonig's buffer at 4°C. Cells were post-fixed with 1% osmium tetroxide, and stained with 2% uranyl acetate. Samples were dehydrated in ethanol, and placed in propylene oxide prior to embedding in Epon 812. Plastic sections were prepared and stained with lead citrate for examination with a Zeiss EM10 CA transmission electron microscope (Zeiss Corporation).

ACKNOWLEDGMENTS
The authors thank Dr. Bhanu P. Telugu for guidance in the design and preparation of reprogramming vectors. We also thank Ms. Amy Shannon and Mr. Paul Graninger for technical assistance in proteomic analysis, and Ms. Tammy Putmon for assistance with reverse-transcription PCR assays.

FUNDING
This research did not receive any specific grant from any funding agency in the public, commercial or not-for-profit sector.