Technical note: improving the efficiency of generating bovine extraembryonic endoderm cells

Abstract

The formation of extraembryonic endoderm (XEN) occurs early in embryonic development. The cell types that develop from the XEN remain poorly studied in ruminant species because of the lack of suitable cell culture model systems. The goal of this work was to establish a protocol for producing XEN cell cultures from bovine blastocysts. Previous work identified fibroblast growth factor 2 (FGF2) as a facilitator of bovine XEN development. Further refinements in culture conditions studied here included exposure to 20% fetal bovine serum and FGF2 replenishment. These modifications yielded an endoderm outgrowth formation incidence of 81.6% ± 5.5% compared with 33.3% ± 5.5% in bovine serum albumin (BSA)-supplemented controls. These cells resembled XEN when examined morphologically and contained XEN transcripts (GATA binding protein 4 [GATA4] and GATA binding protein 6 [GATA6]) as well as transcripts present in visceral (BCL2 interacting protein 1 [BNIP1] and vascular endothelial growth factor A [VEGFA]) and parietal (C-X-C motif chemokine receptor 4 [CXCR4], thrombomodulin [THBD], and hematopoietically expressed homeobox [HHEX]) XEN. Two XEN cell lines were maintained for prolonged culture. Both lines continued to proliferate for approximately 6 wk before becoming senescent. These cultures maintained an XEN-like state and continued to express GATA4 and GATA6 until senescence. An increase in the abundance of visceral and parietal XEN transcripts was observed with continued culture, suggesting that these cells either undergo spontaneous differentiation or retain the ability to form various XEN cell types. Stocks of cultured cells exposed to a freeze-thaw procedure possessed similar phenotypic and genotypic behaviors as nonfrozen cells. To conclude, a procedure for efficient production of primary bovine XEN cell cultures was developed. This new protocol may assist researchers in exploring this overlooked cell type for its roles in nutrient supply during embryogenesis.

Introduction

Early embryogenesis contains two cell specification events. The first is the segregation of trophectoderm (TE) and inner cell mass (ICM). This process begins at the morula stage in cattle (Berg et al., 2011). The second cell fate decision is the differentiation of the ICM into extraembryonic endoderm (XEN) and epiblast (EPI). The XEN will migrate along the inner blastocoel border to underlie the TE layer. The XEN differentiates into visceral endoderm (VE) or parietal endoderm (PE) and gives rise to distal and proximal portions of the yolk sac (YS), respectively (Moerkamp et al., 2013). The YS develops into a vascularized membrane over the next few weeks, reaching maximal function on day 20. During this time, the YS is responsible for gas and nutrient transfer, synthesis of several developmentally relevant proteins, and is the site of initial blood vessel formation and primordial germ cell migration during early embryogenesis (Shi et al., 1985; Freyer and Renfree, 2009). This time coincides with substantial pregnancy loss in cattle (Wiltbank et al., 2016; Reese et al., 2020). By day 30, the YS begins to recede but remains functional until days 50 to 60 of gestation (Galdos-Riveros et al., 2015).

A common misconception is that the YS is rudimentary in mammals. In mice, early embryonic lethality occurs in the absence of a YS, and YS abnormalities appear to be an underlying cause of pregnancy failure for in vitro-produced bovine embryos (Morin-Kensicki et al., 2006; Mess et al., 2017). The implications of insufficient YS development on pregnancy loss in cattle require further investigation.

Arguably, the biggest limitation with learning more about the functions of XEN and the YS in cattle is the lack of suitable cell culture models. Several XEN cell lines exist for the human and mouse, and they have been used extensively to study embryogenesis in these species (D’Amour et al., 2005; Kunath et al., 2005; Brown et al., 2010; Drukker et al., 2012). However, only a few reports exist describing bovine XEN cell lines (Talbot et al., 2000; Yang et al., 2011). Moreover, there was a low efficiency of generating XEN cultures in previous attempts. This report describes work aimed at 1) developing a protocol for improved rates of XEN outgrowth formation from bovine embryos and 2) delineating the various types of XEN represented in the resulting cultures.

Materials and Methods

No animals were used for this work. Materials were collected from a commercial slaughterhouse that followed humane slaughter practices according to USDA guidelines.

In vitro embryo production

In vitro bovine embryos were produced as previously described (Rivera and Hansen, 2001; Wooldridge and Ealy, 2019). Cumulus oocyte complexes (COCs) collected from ovaries were obtained from Brown Packing Co. (Gaffney, SC). The COCs were fertilized with semen from four Holstein bulls (Select Sires, Inc.) for 14 to 18 h at 38.5 °C in 5% CO₂ in humidified air. Presumptive zygotes were cultured in groups of 20 to 30 in 50 μL drops of synthetic oviduct fluid-BE1 at 38.5 °C in 5% O₂, 5% CO₂, and 90% N₂ (Fields et al., 2011).

Endoderm outgrowth cultures

At day 8 post-fertilization, individual embryos were transferred to 12-well plates (3.8 cm²; Corning, Tewksbury, MA) coated with Matrigel Basement Membrane Matrix (BD Biosciences, San Jose, CA). Embryos were cultured in Dulbecco’s Modified Eagle’s Medium containing 5.5 mM glucose, 20% [v/v] fetal bovine serum (FBS), antibiotic/antimycotic mix (5 IU Penicillin G and 50 μg/mL streptomycin sulfate; Thermo Fisher Scientific), and 10 ng/mL recombinant bovine fibroblast growth factor 2 (FGF2) or bovine serum albumin (BSA) control at 38.5 °C in 5% CO₂ in the air. Additional FGF2 or BSA (10 ng/mL final concentration) was added on day 10. On day 12, half of the medium was replaced with fresh medium containing 10 ng/mL FGF2 or BSA. On day 15, the presence or absence of TE and XEN outgrowths was determined via visual assessment (see Figure 1). The medium was replaced with fresh medium containing 10% FBS. Cultures were maintained at 38.5 °C in 5% CO₂ in air. The medium was replaced every 2 d thereafter. Upon reaching 80% to 90% confluence, initial outgrowths were passed into CELLBind tissue culture plates (Corning Inc; Corning, NY) using trypsin-ethylenediaminetetraacetic acid (EDTA) (0.25%) as they reached 80% to 90% confluence (approximately every 7 d). Cells were frozen in 92% [v/v] FBS and 8% [v/v] dimethyl sulfoxide (DMSO) (Sigma-Aldrich).

Figure 1.

Outgrowth formation from bovine blastocysts. Bovine blastocysts at day 8 post-fertilization were transferred to 12-well plates (3.8 cm²) coated with Matrigel Basement Membrane Matrix in culture medium with 10 ng/mL recombinant bovine FGF2. Panel (A): Day 8 blastocyst. Panel (B): Day 15 embryo with TE and XEN outgrowths. Panel (C): Day 19 TE and XEN outgrowths at 100× magnification. Panel (D): Day 19 TE and XEN outgrowths at 200× magnification. Panel (E): DAPI staining observed at 10× magnification. Panel (F): TE-specific CDX2 (green) and XEN-specific GATA6 (red) localization at 10× magnification.

Culture with STO feeder layer

The STO-SNL murine cell line (ATCC #SCRC-1050) was inactivated with 10 µg/mL mitomycin C for 3 h and plated at a density of 75,000 cells/cm². Passage (P) 1 XEN cells were seeded onto the STO feeder cells in the culture medium described above. Cells were passed upon reaching 80% to 90% confluence. After 6 to 8 wk, RNA was isolated from cultures.

Immunofluorescence

Immunofluorescence staining was completed as previously described, with modifications (Negrón-Pérez et al., 2017). Cells were fixed with 4% (w/v) paraformaldehyde in 0.01 M PBS (pH 7.4) and probed with antibodies targeting CDX2 (Biogenex) and GATA6 (Cell Signaling Technology). Outgrowths were also stained with DAPI (ThermoFisher Scientific). Secondary antibodies included donkey anti-mouse Alexa Fluor 488 (Thermo Fisher Scientific) and donkey anti-rabbit Alexa Fluor555 (ThermoFisher Scientific). Outgrowths were visualized using an Eclipse Ti-E inverted microscope equipped with an X-Cite 120 epifluorescence illumination system and DS-L3 digital camera (Nikon Instruments Inc.).

Quantitative RT-PCR analysis

RNA extraction was completed using TRIzol reagent (Thermo Fisher Scientific) and the PureLink RNA mini kit (Thermo Fisher Scientific). Samples were incubated with RNase-free DNase (Thermo Fisher Scientific) and were reverse transcribed using a High Capacity cDNA Reverse Transcription Kit (Thermo Fisher Scientific). A polymerase chain reaction (PCR) was performed using the SybrGreen detection system (Thermo Fisher Scientific) in combination with primers for XEN, TE, and EPI-specific transcripts (Table 1). Primers were designed using BLAST from the National Center for Biotechnology Information and synthesized by Life Technologies. Efficiency was determined for each primer set by running a standard curve analysis on pooled samples (efficiencies ranged from 83% to 100%). The internal control was RPS9 as it contained minimal sample-to-sample variation and was not altered by time in culture. Each sample was run in triplicate. A fourth sample lacking the reverse transcriptase was included as a negative control. The average threshold cycle (C_T) value for each gene was determined and applied to the following equation: (2^{−CT gene of interest})/(2^{−CT RPS9}).

Table 1.

List of primers used for qRT-PCR

Gene	Primer sequence (5′ to 3′)1
GATA4	F2: ATGAAGCTCCATGGCGTCCC R: CGCTGCTGGAGCTGCTGGAA
GATA6	F: ATACTTCCCCCACCACACAA R: AGCCCGTCTTGACCTGAGTA
CXCR4	F: ACTTGAGTAGCCGGTAGCCC R: CGTTGCCCACTATGCCAGTC
THBD	F: CACTGCGACACTGGCTATGA R: GCAGATGGTCGGGTAGTGAG
HHEX	F: AGAAATACCTCTCCCCGCCC R: CAAGTCTTGCCTCTGGTCGC
BNIP1	F: TCAGACCTCATGGAGGAAGGC R: AGCTTCCGTCCCAACTGGAT
VEGFA	F: GTCTACCAGCGCAGCTTCTG R: TGCTGGCTTTGGTGAGGTT
CDX2	F: GGCAGCCAAGTGAAAACCAG R: GCTTTCCTCCGGATGGTGAT
IFNT	F: GCCCGAATGAACAGACTCTC R: CCATCTCCTGAGGAAGACCA
NANOG	F: GACACCCTCGACACGGACAC R: CTTGACCGGGACCGTCTCTT
RPS9	F: GAGCTGGGTTTGTCGCAAAA R: GGTCGAGGCGGGACTTCT

¹ All primers were developed using the NCBI Basic Local Alignment Search Tool (BLAST).

²F, forward primer; R, reverse primer.

Statistical analyses

Outgrowth formation efficiency was analyzed using the chi-square analysis in GraphPad Prism (n = 5 replicate studies). Changes in the relative transcript abundance were determined in log-transformed values by least-squares analysis of variance using the general linear model (SAS). Pair-wise comparisons were completed to further examine differences (PDIFF analysis in SAS). Significance was set at P < 0.05.

Results

Outgrowth production efficiency

Free-floating blastocysts (Figure 1A) attached to plates on day 10 to 12 post-fertilization. Initial outgrowths contained a darkened, central region that presumably represented ICM, an inner ring of TE cells, and XEN cells localized around the outer perimeter of the outgrowth (Figure 1B). As time progressed, the ICM cluster washed away, but the TE and XEN persisted (Figure 1C and D).

Based on visual evaluations, XEN was observed in 81.6% ± 5.5% of the blastocyst outgrowths treated with FGF2 and 33.3% ± 5.5% of control cultures treated with carrier (BSA) and 10% FBS (P = 0.0002). No additional outgrowths formed after day 15 post-fertilization. A second study determined that XEN outgrowths developed when using regular, expanded, or hatched blastocysts selected at day 7 or 8 post-in vitro fertilization (IVF); however, outgrowths were rarely observed when using blastocysts after day 8 post-fertilization (data not shown).

Endoderm verification

Outgrowths were collected for qRT-PCR between days 16 and 19 or days 21 and 23 post-fertilization (Table 2). All outgrowths contained transcript markers for XEN (GATA4 and GATA6), PE (CXCR4, HHEX, and THBD), and VE (BNIP1 and VEGFA). The TE-specific marker, CDX2, was detected in 4/6 samples collected at days 16 to 19 and 9/9 on days 21 to 23, though TE could not be visualized after the first passage. Transcripts for the pluripotency marker, NANOG, were detected in 3/6 samples collected at days 16 to 19 and in all samples collected at days 21 to 23. Total nuclei were labeled with DAPI (Figure 1E), and definitive zones of TE and XEN cells were visualized by targeting CDX2 and GATA6, respectively (Figure 1F).

Table 2.

The presence (+) or absence (−) of cell-specific transcripts in bovine embryo outgrowths at days 16 to 19 or days 21 to 23 post-IVF

Gene of interest	Cell lineage	Days 16 to 191,2	Days 21 to 231,2
GATA4	XEN	+	+
GATA6	XEN	+	+
CXCR4	PE	+	+
THBD	PE	+	+
HHEX	PE	+	+
BNIP1	VE	+	+
VEGFA	VE	+	+
CDX2	TE	+/−	+
IFNT	TE	+	+
NANOG	ICM	+/−	+

¹ n = 6 samples for days 16 to 19 and n = 7 for days 21 to 23.

²+ indicates detection in all samples, − indicates no detection, and +/− indicates detection in some of the samples.

XEN cell line propagation

Two outgrowths were maintained beyond day 23 post-fertilization (XEN1 and XEN2). Cells divided every 12 to 18 h before becoming senescent at 6 wk in culture. Contact inhibition was observed as cells required passage prior to 100% confluence. Cells not plated at the time of passage were frozen. Upon thawing, cells maintained their XEN morphology and mitotic potential for another 4 to 5 wk before ceasing to proliferate.

Cell specification of extended XEN cultures

Cell marker profiles were examined in XEN1 and XEN2 cultures to describe the changes in expression profiles as cells were actively proliferating (P2; ~day 30), beginning to undergo mitotic arrest (P5; ~day 60) or incurring replicative senescence (P6; days 70 to 80) (Figure 2). The XEN markers, GATA4 and GATA6, were expressed at all stages of culture (Figure 2A). The VE-specific BNIP1 was detected in both cultures at P2 and P5 but were absent from both cultures at P6 (Figure 2B). The VE-specific VEGFA was observed in both cultures at P2 and P5, but only in one culture at P6 (Figure 2B). Three PE-specific transcripts were also examined (Figure 2C). Transcripts for CXCR4 were detected in both cultures at P2, but only in one culture at P5 and P6. Transcripts for HHEX were absent from both cultures at P2 but present in both at P5 and P6. Transcripts for THBD were detected in both cultures in all passages. Transcripts demarking TE (CDX2 and IFNT) and EPI (NANOG) were examined (Figure 2D). Transcript abundance for CDX2 was increased (P < 0.05) from P2 to P5 and from P5 to P6, while the abundance of IFNT and NANOG was increased (P < 0.05) only at P6 (Figure 2D).

Figure 2.

Relative abundance of cell-specific transcripts in bovine embryo outgrowths. Outgrowth RNA was collected at passages (P) 2, 5, and 6 (n = 3 to 4/passage/culture). P2 is representative of actively dividing cells, P5 represents cells entering senescence, and P6 represents cells that are likely senescent. Samples were separated between the XEN1 and XEN2 cultures. Panel (A): XEN-specific transcripts; Panel (B): Visceral XEN (VE)-specific transcripts; Panel (C): Parietal XEN (PE)-specific transcripts; Panel (D): TE-specific (CDX2 and IFNT) and ICM-specific (NANOG) transcripts. Differing superscripts and asterisk (*) denote significant differences in transcript expression between passages (P < 0.05).

A final study was completed to determine if the proliferative lifespan of XEN cultures could be prolonged by culture with a STO feeder layer. Early passage XEN cells proliferated rapidly, but all such cultures senesced by day 70 to 80 (data not shown).

Discussion

The development and function of bovine XEN during embryogenesis remain understudied because of the low efficiency of attaining bovine XEN cell lines. This work set out to describe a technique to generate bovine embryo outgrowths that contain a proliferative, feeder cell-independent XEN cell population. Recently, an improvement in XEN outgrowth efficiency was achieved by supplementing FGF2 during blastocyst outgrowth formation (Yang et al., 2011; Kuijk et al., 2012). In both studies, FGF2 supplementation promoted XEN specification, while blocking FGF2/4 signaling prevented the formation. Two refinements were made to the previously reported bovine XEN outgrowth approach (Yang et al., 2011). First, embryos were cultured from day 8 to 15 in medium containing 20% (v/v) FBS. Second, embryo cultures received multiple treatments with FGF2. Previous work reported that outgrowth rates improved from 1.2% to 23.5% when providing 5 ng/mL FGF2 at day 8 post-fertilization (Yang et al., 2011). The protocol in the present study increased the outgrowth rate to 81.6%. Concentrations greater than 10 ng/mL FGF2 were not examined, as previous work showed no improvement in XEN formation when embryos were treated with 50 ng/mL (Yang et al., 2011).

Providing increased amounts of FBS during XEN development likely contributed to improvements in XEN formation. In mice, 15% to 20% FBS is traditionally added to culture media to generate TE and XEN outgrowths (Kang et al., 2013; Wu et al., 2017). Previous work utilized 5% to 10% FBS to generate XEN cultures from bovine embryos (Talbot et al., 2000; Yang et al., 2011). The mechanism of FBS activity on XEN development is unclear. No studies have been completed to examine if FBS alone is sufficient to induce XEN formation. Nonetheless, it is apparent that utilizing FGF2 in medium containing 20% FBS maximizes the incidence of XEN outgrowth formation in bovine embryos. Another possible reason for the increased rates of XEN outgrowth formation was the use of Matrigel, which contains several extracellular matrices factors and growth factors, all of which may facilitate XEN development (Vukicevic et al., 1992).

Early cultures contained transcripts for XEN, VE, and PE. This may indicate a plasticity to develop into XEN subtypes. Authors are not comfortable definitively calling these cells XEN stem cells, as it is unclear if individual cells express transcripts necessary for this designation or if the numerous transcript markers are representative of multiple endoderm populations within the cultures (Brown et al., 2010; Niakan et al., 2013). The multitude of XEN markers inspired an investigation into the longevity of these cells in culture and the type of XEN present. Extended cultures were propagated for 6 to 8 wk before becoming senescent. Senescence is common for primary cell cultures, though this outcome may also have a developmental basis. The YS has a definitive life span, and perhaps these cells are programmed to have limited viability. The cultures also showed indications that PE and VE specification occurs over time in vitro. Cultures continued to express markers for XEN, PE, and VE, though this was dependent on culture. Selectivity in XEN cell type may exist with extended culture, but the loss in specific transcripts does not indicate that cells were preferentially forming PE or VE. Rather, the presence of all of these markers suggests that the initial XEN outgrowths have multipotent potential.

An attempt to improve the longevity of XEN lines was completed by coculturing XEN cultures with a STO feeder layer. No extension in longevity was evident. The lines from previous reports were likely self-immortalized, and the culture of numerous outgrowths was needed to identify the few that grew indefinitely (Talbot et al., 2000, 2005). Sufficient numbers of extended XEN cultures were not completed to determine if presumptive self-immortalizing XEN lines would emerge using these culture techniques.

Initial outgrowths contained TE-specific transcripts. TE is common in such cultures, but its poor attachment to non-Matrigel surfaces and low viability after Trypsin-EDTA treatment minimized TE growth after a few passages. The EPI-dependent transcript, NANOG, was also detected in early passage cultures. It remains unclear if these cells were present in cultures or if low-level NANOG expression exists in TE and/or XEN. Moreover, TE- and EPI-specific transcripts were detected in extended cultures, and their relative abundance increased with each passage. Though neither cell type was detectable by phase-contrast microscopy, we cannot discount that they were present. Alternatively, non-committed cells may be present in these cultures. Thus, the cultures can only be considered XEN-enriched.

To conclude, this work established a procedure for generating bovine embryonic outgrowths enriched with XEN cell types. Several XEN subtypes may exist in these cultures, and they undergo senescence after 6 wk. These findings are anticipated to provide a new tool for studying XEN and YS development in ruminants. Such work is necessary to better understand how XEN functions to promote embryogenesis.

Glossary

Abbreviations

BSA

bovine serum albumin

COC

cumulus oocyte complex

EPI

epiblast

FBS

fetal bovine serum

FGF2

fibroblast growth factor 2

ICM

inner cell mass

PE

parietal endoderm

TE

trophectoderm

VE

visceral endoderm

XEN

extraembryonic endoderm

YS

yolk sac

Conflict of interest statement

The authors have no conflict of interest to declare.