|Home | About | Journals | Submit | Contact Us | Français|
The objective of this study was to identify transcription factors associated with differentiation of the chorionic girdle, the invasive form of equine trophoblast. The expression patterns of five transcription factors were determined on a panel of conceptus tissues from early horse pregnancy. Tissues from Days 15 through 46 were tested. Eomesodermin (EOMES), glial cells missing homologue 1 (GCM1), heart and neural crest derivatives expressed transcript 1 (HAND1), caudal type homeobox 2 (CDX2), and distal-less homeobox 3 (DLX3) were detected in horse trophoblast, but the expression patterns for these genes varied. EOMES had the most restricted distribution, while DLX3 CDX2, and HAND1 were widely expressed. GCM1 seemed to increase in the developing chorionic girdle, and this was confirmed by quantitative RT-PCR assays. GCM1 expression preceded a striking increase in expression of equine chorionic gonadotropin beta (CGB) in the chorionic girdle, and binding sites for GCM1 were discovered in the promoter region of the CGB gene. GCM1, CGB, and CGA mRNA were expressed preferentially in binucleate cells as opposed to uninucleate cells of the chorionic girdle. Based on these findings, it is likely that GCM1 has a role in differentiation and function of the invasive trophoblast of the equine chorionic girdle and endometrial cups. The equine binucleate chorionic girdle (CG) secreting trophoblast shares molecular, morphological, and functional characteristics with human syncytiotrophoblast and represents a model for studies of human placental function.
During the early stages of pregnancy, the trophoblast cells of the placenta differentiate into subtypes with distinct molecular and functional phenotypes [1, 2]. A few transcription factors have been found to be key determinants of trophoblast cell fate during this critical period of development. These include eomesodermin (EOMES) , glial cells missing homologue 1 (GCM1) [4–6], heart and neural crest derivatives expressed transcript 1 (HAND1) , caudal type homeobox 2 (CDX2) , and distal-less homeobox 3 (DLX3) . Thus far, most studies of the genes that regulate trophoblast differentiation have been conducted in mice. Studies in humans have produced confirmatory [1, 10, 11] and disparate [1, 11–13] results, but such studies are limited by the availability of materials. Gene expression by subtypes of trophoblasts in other species remains largely unexplored. Equine pregnancy has features that make the horse a good model species for comparative studies of placentation. The late attachment of the equine placenta to the uterine wall allows nonsurgical conceptus recoveries between Days 6 and 36 of gestation, a period that covers many stages of trophoblast and fetal development. In addition, only equids and primates produce chorionic gonadotropin.
The following four main subtypes of trophoblast cells have been described in the horse using morphological techniques: chorion, allantochorion, chorionic girdle, and binucleate (endometrial cup) trophoblast cells . These subtypes begin to differentiate following the loss of the equine glycoprotein capsule, which surrounds the equine trophectoderm until Days 22 through 23 of pregnancy . At approximately Days 24 through 26 of pregnancy, the expanding allantois fuses with the chorion, giving rise to the allantochorion trophoblast . These cells form the continuous epithelium-like surface of the equine epitheliochorial placenta and have characteristics of human villous cytotrophoblast. The chorionic girdle is a discrete annular structure of the equine placenta that forms between Days 25 and 35 at the junction of the enlarging allantois and regressing yolk sac. It is initially composed of rapidly proliferating uninucleate trophoblast cells, which eventually differentiate into nondividing binucleate cells. At Days 36 through 38, the binucleate cells gain an invasive phenotype and migrate into the endometrium to form specialized structures called endometrial cups. The binucleate endometrial cup trophoblast cells share many features with the syncytiotrophoblast of the human placenta. These terminally differentiated trophoblast cells are binucleate (equine) or multinucleate (human); they downregulate major histocompatibility complex class I and class II genes at the transcriptional level [15–17], and they secrete chorionic gonadotropin [14, 18].
Morphological studies [17, 19–23] of the chorionic girdle, together with investigations that have characterized the temporal expression of chorionic girdle (CG) , have indicated that terminal differentiation of binucleate cells begins at around Day 32 of pregnancy. The proportion of binucleate cells in the chorionic girdle then rapidly increases until Days 36 through 38. Findings from morphological and expression studies [25, 26] of equine conceptuses obtained between Days 29 and 34 of pregnancy suggested that mesenchymal-derived products such as hepatocyte growth factor may regulate processes such as chorionic girdle trophoblast proliferation and migration, although this remains to be demonstrated functionally. The factors that regulate terminal differentiation of binucleate cells are unknown. This study was designed to identify key transcription factors involved in equine trophoblast development, with a focus on the invasive trophoblast cells of the chorionic girdle.
Horses of mixed breeds and ages were used in this research. Horses were maintained at the Baker Institute for Animal Health, College of Veterinary Medicine, Cornell University. Animal care was performed in accord with the guidelines set forth by the Institutional Animal Care and Use Committee of Cornell University. Pregnancies were established as previously described .
Conceptuses were obtained by established methods . Duplicate equine conceptuses were obtained by nonsurgical uterine lavage on Days 15, 21, 25, 30, 31, 32, 33, and 34 of pregnancy, and samples of endometrial cups and gravid endometrium were obtained using surgical techniques on Days 43 and 46 of pregnancy (Fig. 1). The duplicate conceptuses were used for qualitative and quantitative assays. With the aid of a dissecting microscope, the capsule was removed from Days 15 and 21 conceptuses, and the tissue was separated into trophectoderm and early fetus. Days 25 through 33 conceptuses were microdissected into chorionic girdle, allantochorion, chorion, and fetus, while Day 34 conceptuses were microdissected into these four tissues plus two additional tissues, yolk sac and bilaminar omphalopleure. Days 43 and 46 equine conceptus tissues, including endometrial cups and samples of gravid endometrium, were obtained surgically immediately following euthanasia of pregnant mares. Peripheral blood lymphocytes were isolated from heparinized samples of venous jugular blood using methods described previously . Total RNA was isolated from snap-frozen conceptus tissues, following homogenization by QIAshredder (Qiagen, Valencia, CA), using an RNeasy kit (Qiagen) as directed by the manufacturer. Five hundred nanograms of tRNA was treated with DNase I (Invitrogen, Carlsbad, CA), and then first-strand cDNA synthesis was performed using M-MLV Reverse Transcriptase (USB, Cleveland, OH) as per the manufacturer's guidelines.
Chorionic girdle trophoblast cells were cultured using established methods [20, 23]. Briefly, Days 34 through 36 chorionic girdle tissue was dissected free of adjacent conceptus tissues and placed into Dulbecco modified Eagle medium (DMEM). Girdle cells were gently scraped off the basement membrane and underlying mesodermal cell layer using a scalpel blade, ensuring that all nontrophoblast tissue was discarded. Cells were transferred to a 15-ml tube, allowed to settle on ice, and then cultured in DMEM enriched with 100 U/ml of penicillin-streptomycin, 2 μM l-glutamine (Gibco, Carlsbad, CA), 0.5 μg/ml of vitamin C, and 0.4 μg/ml of insulin (Sigma, St. Louis, MO) and 10% fetal calf serum.
Chorionic girdle cells were passaged using standard trypsinization procedures, with a slight modification for first-passage cells. These cultures were composed of uninucleate and binucleate girdle cells. The binucleate cells were very susceptible to trypsin treatment and were dislodged easily from the flasks. This fraction is referred to as “binucleate cell enriched.” The cells remaining on the bottom of the flask were more resistant to trypsinization. They were removed using plastic cell scrapers. This fraction is referred to as “uninucleate cell enriched.”.
Horse sequences for EOMES, DLX3, GCM1, HAND1, CDX2, and succinate dehydrogenase complex subunit A (SDHA) cDNA were identified using the equine whole-genome sequence database . Primers were designed using Primer3 . Primer sequences are given in online Supplemental Table 1 (all supplemental data are available at www.biolreprod.org). Amplification of 15 ng of cDNA was performed in a 25-μl reaction using standard PCR conditions. Products were separated by electrophoresis on 1% agarose gels, stained with ethidium bromide, and visualized under a Spectroline UV Transilluminator, (Spectronic, Westbury, NY). The PCR products were purified, cloned, and sequenced to confirm the specificity of the PCR product.
SYBR Green (Applied Biosystems, Shelton, CO) real-time PCR reactions for amplification of equine GCM1, CG/LH/FSH/TSHA (thyroid-stimulating hormone alpha) (CGA), CG/LHB, HAND1, DLX3, CDX2, or the housekeeper gene equine ubiquitin-conjugating enzyme E2D 2 (UBE2D2)  mRNA were performed using an ABI PRISM 7700 sequence detector (PerkinElmer Life and Analytical Sciences, Waltham, MA) in a total volume of 20 μl. Primers were designed over intron/exon boundaries to prevent amplification of genomic DNA. A dissociation curve was performed after each experiment to confirm that a single product was amplified. A standard curve was generated using known copy numbers of a plasmid that contained the cDNA specific to the gene. Each sample was normalized to 7000 copies of UBE2D2. Relative expression of HAND1 was calculated using the 2(−delta delta C[T]) method and was normalized to UBE2D2 expression. The sequences of the oligonucleotides are given in Supplemental Table 2. For statistical comparison of spatial gene expression by Day 34 conceptus tissues, Dunnett multiple comparison one-way ANOVA tests were performed using chorionic girdle as the control tissue and GraphPad Prism software (GraphPad Software, Inc., La Jolla, CA). For statistical comparison of temporal gene expression, Tukey one-way ANOVA tests were used with an alpha error = 5% using GraphPad Prism software.
Following a review of the human and mouse literature, we selected five transcription factors (DLX3, CDX2, HAND1, EOMES, and GCM1) that are hypothesized to have a role in regulating different aspects of equine trophoblast differentiation and function. Homologues for all of the genes were identified in the horse whole-genome sequence database. A high degree of nucleic acid identity (84%–91%) between horse and human and horse and mouse (78%–87%) sequences was observed (Supplemental Table 1). Equine conceptuses were obtained at Days 15, 21, 25, 30, 31, 32, 33, and 34 of pregnancy, and endometrial cups and gravid endometrium were obtained on Days 43 and 46 of pregnancy (Fig. 1). Each conceptus was microdissected into two or more components, varying with the maturity of the conceptus (see Materials and Methods). Tissues were screened by qualitative RT-PCR for DLX3, CDX2, HAND1, EOMES, and GCM1 mRNA. As a loading control, we determined the expression of the placental housekeeper gene SDHA . SDHA was constitutively expressed in all tissues tested (Fig. 2). All five transcription factors were detected in trophoblast cells, but the expression patterns varied by time and spatial distribution. Only EOMES was detected in adult tissues. DLX3 and CDX2 were widely expressed in all trophoblast tissues tested between Days 15 and 34, with weak expression in endometrial cups. HAND1 was expressed in trophoblast tissues between Days 25 and 34, together with Days 43 and 46 endometrial cups. EOMES had the most restricted pattern of expression of any of the genes tested. EOMES was detected in Days 15 and 21 trophectoderm samples. EOMES was also detected in endometrial cups, endometrium, and lymphocytes, suggesting that expression by endometrial cup tissue may not be specific to the trophoblast cells in that tissue. GCM1 was detected only in trophoblast tissues, with expression primarily in chorion and chorionic girdle. The pattern suggested increasing expression in developing chorionic girdle. Therefore, GCM1 was selected for further investigation as a candidate molecule regulating cell fate determination, binucleate differentiation, and function in chorionic girdle and endometrial cup trophoblast cells.
To determine if GCM1 expression was specific to the chorionic girdle, we quantified GCM1 mRNA in six different Day 34 equine conceptus tissues obtained from two independent conceptuses using absolute quantitative RT-PCR. As a comparison, we also determined the relative expression of the more constitutively expressed gene HAND1. There was a statistically significant increase in GCM1 expression compared with all other Day 34 conceptus tissues (P < 0.01). Chorionic girdle expressed 31258 and 21188 copies of GCM1 mRNA per 10 ng of RNA (Fig. 3, upper left panel), which was 10- to 20-fold higher than expression in the corresponding allantochorion (2900 and 1013 copies per 10 ng of RNA). Chorion expressed 429 and 649 copies of mRNA per 10 ng of RNA, with little or no GCM1 mRNA detectable in fetus, yolk sac, or bilaminar omphalopleure. Using relative expression, we detected no significant difference in HAND1 mRNA in chorionic girdle compared with all other Day 34 conceptus tissues (Fig. 3, upper right panel).
We also quantified mRNA for equine CG in Day 34 conceptuses tissues. CG, the principal hormone produced by the mature endometrial cup cells that are derived from chorionic girdle trophoblast, is composed of an alpha subunit encoded by a gene common to all the gonadotropins , referred to as CGA, and a beta subunit encoded by a gene common to LH, referred to as chorionic gonadotropin beta (CG/LHB [official symbol, CGB]) . Similarly to GCM1, CGA and CGB mRNA were expressed by chorionic girdle at significantly higher levels compared with all other Day 34 conceptus tissues (P < 0.01). Average CGA mRNA levels in Day 34 chorionic girdle tissues were 1.4 million copies per 10 ng of RNA (Fig. 3, lower left panel). The alpha subunit was also expressed, albeit at significantly lower levels, in allantochorion (174445 and 20293 copies) and chorion (15966 and 5672 copies). CGB mRNA was highly expressed in chorionic girdle (Fig. 3, lower right panel), with an average of 119000 copies per 10 ng of RNA. It was also expressed in an allantochorion tissue sample (10303 copies per 10 ng of RNA). CGB mRNA was virtually undetectable in all other Day 34 tissues tested.
Terminal differentiation of chorionic girdle trophoblast to binucleate cells begins at about Day 32 of pregnancy and rapidly progresses until Days 36 through 38. Experiments were performed to determine expression of GCM1 mRNA within this small window of development. Using quantitative RT-PCR, GCM1 mRNA expression was determined in chorionic girdle tissues obtained from two individual conceptuses recovered on Days 28, 30, 31, 32, 33, and 34, as well as endometrial cups and corresponding gravid endometrium from Days 43 and 46 of pregnancy. GCM1 mRNA was rapidly induced beginning at Day 31 of pregnancy, with maximal expression at Day 34 (Fig. 4A, upper panel). There was a 16.5-fold difference in GCM1 expression between Days 30 and 34 of pregnancy. GCM1 mRNA was minimally expressed in Day 28 chorionic girdle, a time outside the window of binucleate differentiation. Endometrial cups expressed an average of 8590 copies of mRNA per 10 ng of RNA, while the corresponding gravid endometrium did not express GCM1. As controls, relative expression of HAND1 mRNA was determined in the tissues already described. HAND1 mRNA expression varied by less than 2-fold across Days 28 through 34 chorionic girdle, changes that were not statistically significant. HAND1 was minimally expressed by the endometrial cups and was absent from the gravid endometrium (Fig. 4A, upper right panel).
To ascertain the relationship between GCM1 and CG expression, the expression of CGA and CGB mRNA in differentiating chorionic girdle was also determined. CGA mRNA expression was rapidly increased in chorionic girdle from Day 32 (4.0 × 106 copies per 10 ng of RNA), and expression was sustained until Day 34 (Fig. 4A, middle left panel). CGA mRNA was expressed at lower levels in Days 28 through 31 chorionic girdle. CGA mRNA expression was highest in endometrial cups (average, 15722307 copies per 10 ng of RNA), while negligible expression was detected in gravid endometrium. CGB mRNA was first detected in Day 32 chorionic girdle, and expression increased in chorionic girdle until Day 34 (Fig. 4A, middle right panel). Similar to the alpha subunit, endometrial cups expressed significantly higher levels of CGB mRNA (539321 copies per 10 ng of RNA) compared with all other tissues tested. Days 28 through 31 chorionic girdle expressed very little CGB mRNA. Finally, we mapped the average GCM1 and CGB mRNA expression in Days 28 through 34 chorionic girdle (Fig. 4B). GCM1 mRNA began to rise at 24 h before the rise in CGB.
To investigate which trophoblast cells of the chorionic girdle expressed GCM1, we determined GCM1 mRNA expression in different populations of cultured chorionic girdle cells. First-passage Day 36 chorionic girdle trophoblast cells were harvested using the following two different conditions: 1) mild trypsinization (cell population enriched for binucleate cells) and 2) mild trypsinization followed by scraping (cell population enriched for uninucleate cells). GCM1 mRNA was preferentially expressed by the binucleate-enriched fraction, with little GCM1 mRNA detectable in the uninucleate-enriched fraction (Fig. 5, left panel). We also determined absolute HAND1 and DLX3 expression. The binucleate-enriched fraction expressed an average of 500 copies of HAND1 and 382 copies of DLX3, while the uninucleate-enriched fraction expressed an average of 1688 copies of HAND1 and 132 copies of DLX3 mRNA. In contrast, the binucleate cells expressed an average of 19018 copies of GCM1, and the uninucleate fraction expressed 597 copies of GCM1. Next, we quantified CG expression in each of the cell fractions. GCA and GCB mRNA were highly expressed by the binucleate-enriched cell fraction, with minimal expression in the uninucleate-enriched fraction (Fig. 5, right panel).
Finally, we quantified GCM1, DLX3, HAND1, GCA, and GCB mRNA expression in cultured Day 34 chorionic girdle that had undergone two to five passages and was composed of primarily uninucleate cells. Fewer than 50 copies of GCM1, DLX3, HAND1, and GCB mRNA were detectable in all samples tested (data not shown). Between 50000 and 70000 copies of GCA were expressed by these primarily uninucleate cell populations (data not shown).
The natural variation in placental structure in mammals has made it difficult to discern conserved cell types, function, and mechanisms of differentiation and development across species. However, the recent explosion in whole-genome sequencing of several species has facilitated large-scale gene expression studies [1, 11, 36, 37] that have revealed common molecular phenotypes, as well as important mechanistic differences between species. The expression pattern of equine GCM1 highlights an example of a conserved molecular marker. GCM1 is a member of a family of transcription factors that share a GCM motif and the ability to regulate a number of cellular differentiation processes . In human placenta, GCM1 can be detected in syncytiotrophoblast cells and their progenitor cells of the cytotrophoblast until Week 37 of pregnancy, at which time expression rapidly decreases [5, 38, 39]. Murine Gcm1 has been localized to labyrinth trophoblast cells and to terminally differentiated layer II syncytiotrophoblast cells . Using quantitative RT-PCR assays, we demonstrated that GCM1 expression was rapidly induced between Days 31 and 34 of equine pregnancy during the short window when binucleate cell differentiation occurs. Furthermore, GCM1 was highly expressed by binucleate cells of chorionic girdle and endometrial cups, with minimal expression in other Day 34 extraembryonic and embryonic tissues. These molecular findings, together with other known functional and morphological characteristics of equine binucleate trophoblast and human syncytiotrophoblast, suggest that the equine chorionic girdle and endometrial cup trophoblast cells may be a useful model for the study of human syncytiotrophoblast differentiation and function.
Investigations in human cytotrophoblast cell lines support a role for GCM1 in human syncytiotrophoblast differentiation via regulation of the syncytin gene and subsequent syncytin-mediated cell fusion . To our knowledge, it has not yet been determined whether equine binucleate cells form following cell fusion or as a result of endoreduplication. Equine GCM1 has a high level of identity with human GCM1 (85%), particularly in the GCM1 motif (data not shown), suggesting that functional roles for GCM1 may be conserved, at least in part, in the horse. We propose that GCM1 will be a determinant of equine trophoblast cell fate. We initially hypothesized that GCM1 regulation of syncytin-mediated cell fusion may explain binucleatism in the horse. Using bioinformatic approaches, we have been unable to identify an equine syncytin gene to date (data not shown). Future studies are needed that investigate the functional role of GCM1 in equine binucleate cell differentiation. In contrast to the results reported herein, Thway and colleagues  failed to detect GCM1 expression in endometrial cup cells and immortalized chorionic girdle cell lines. This may be due to a less sensitive Northern blot assay used to detect GCM1 mRNA in that study.
The restricted expression of GCM1 to equine binucleate trophoblast cells led us to hypothesize whether GCM1 may also regulate other genes expressed by binucleate cells such as the alpha and beta subunits of CG. The N-terminus of the human and murine GCM1 protein contains a zinc-containing DNA-binding domain referred to as the GCM1 motif [43–45], and the C-terminus contains two transactivation domains [43, 45]. These DNA-binding domains have been shown to bind and subsequently regulate the expression of syncytiotrophoblast-specific genes such as syncytin and aromatase [41, 46]. In the case of aromatase, GCM1 binds to an enhancer element and confers placenta-specific expression of the gene. Analysis of a 3000-bp fragment of the equine CGB promoter revealed five exact match consensus sites available for GCM1 binding, two of which are located within 200 bp of the transcription start site (data not shown). In these studies, we showed that GCM1 expression immediately preceded the expression of CGB mRNA and that expression of the two genes was correlated. Whether GCM1 binds and activates the predicted GCM1-binding sites within the equine CGB promoter has not been demonstrated (to our knowledge), but our expression and bioinformatics data suggest that this warrants further investigation.
We observed a disparity in CGA and CGB mRNA expression levels in equine chorionic girdle and endometrial cups. In Day 34 chorionic girdle and in mature endometrial cups tissue, CGA mRNA was expressed at 10- to 30-fold higher levels compared with CGB. In addition, CGA expression temporally preceded CGB expression. As already discussed, CG is composed of two subunits encoded by the genes CGA and CGB. Binucleate cells have been shown to secrete CG but not LH, FSH, or thyrotropin, suggesting that if these genes are expressed they encode CG. Alternatively, a free alpha subunit may exist, a finding that has been described in human pregnancy and for the beta subunit of horse CG [47, 48].
Based on the trophoblast lineages identified in the mouse , we propose differentiation pathways that give rise to the four main subtypes of equine trophoblast cells (Fig. 6). Using a qualitative RT-PCR screening strategy, we identified subsets of genes expressed by these subtypes of trophoblast cells (Figs. 2 and and6).6). The qualitative data illustrated in Figure 6 are supplemented for some genes and stages with absolute copy numbers of mRNA detected in this study (Figs. 3–5). EOMES is a T-box transcription factor expressed by early murine trophectoderm . It regulates early trophoblast differentiation in the mouse and is required for murine stem cell maintenance . We detected EOMES in undifferentiated Days 15 through 21 trophectoderm, suggesting that the functional role of EOMES in regulating early differentiation events may be conserved in the horse. CDX2 is a member of the caudal-related gene family of transcription factors and, similarly to EOMES, regulates early trophectoderm differentiation . Investigations of murine placentation indicated that Cdx2 is first expressed by Embryonic Day 3.5 trophectoderm, with expression also present in chorion and spongiotrophoblast . It has also been shown to be expressed by human trophectoderm and by Days 15 through 17 ovine conceptuses [37, 51]. Consistent with the continuing expression pattern in mouse trophoblast cells, CDX2 was widely expressed in equine trophectoderm, chorion, chorionic girdle, and allantochorion trophoblast. HAND1 is a basic helix-loop-helix transcription factor that shows variable expression patterns across species [1, 36]. In human pregnancy, HAND1 is restricted to the trophectoderm of the blastocyst stage , while in murine pregnancy Hand1 is expressed by trophectoderm and by giant cells . We observed a different expression pattern for HAND1 in horse trophoblast cells. HAND1 was first detected in Day 25 chorion, chorionic girdle, and allantochorion, suggesting that HAND1 is specific for the chorion and cells derived from chorion. DLX3 is a member of the distal-less family of homeobox genes expressed by the mouse ectoplacental cone cells, chorionic plate, and labyrinth layer. Dlx3 has a critical role in regulation of branching morphogenesis of the murine placenta . Equine DLX3 was widely expressed by horse trophoblast subtypes between Days 15 and 34, similar to equine CDX2.
In conclusion, we have identified five transcription factors expressed by equine trophoblast cells at different stages of development, including molecular markers specific for binucleate cells (GCM1) and early trophectoderm (EOMES). The expression of GCM1 by human syncytiotrophoblast and equine binucleate trophoblast cells supports previous observations of the morphological and functional similarities of these two cell types. Further studies of the role and regulation of GCM1 in equine binucleate cells, together with the discovery of additional molecular determinants of equine binucleate differentiation, may provide new insights into the mechanisms that regulate syncytiotrophoblast differentiation and GCM1 expression in human pregnancy.
Many thanks to Scott Hoffay, Emily Silvela, and Meleana Hinchman for assistance with horse breeding. Thanks to Dr. Leela Noronha for assistance with conceptus recoveries and to Christina Costa for technical support.
1Supported by the Dorothy Russell Havemeyer Foundation, Inc; by the Zweig Memorial Fund for Equine Research; and by grant RO1-HD049545 from the National Institutes of Health.