|Home | About | Journals | Submit | Contact Us | Français|
The mechanism by which the inner cell mass (ICM) and trophectoderm (TE) become specified is poorly understood. Considerable species variation is evident in the expression of lineage-specific and embryonic stem cell (ESC) regulatory markers. We sought to investigate localization patterns of these markers in rhesus macaque compact morulae and blastocysts. NANOG protein was restricted to the ICM of blastocysts. In contrast to a previous report, the expression of CDX2 was detected in the primate blastocyst, localized specifically to the TE. Unlike the mouse embryo, OCT4 protein was detected using two different antibodies in both the ICM and TE. The ubiquitous pattern of OCT4 expression is consistent with observations in human, cow, and pig embryos. Significantly, lack of restricted OCT4 protein, and ICM localization of NANOG in primate blastocysts, suggests that NANOG may determine inner cell mass fate more specifically during primate development or may be less susceptible to culture artifacts. These results contrast markedly with current mechanistic hypotheses, although other factors may lie upstream of NANOG to constitute a complex interactive network. This difference may also underlie observations that regulatory mechanisms in ESC differ between mice and primates.
Maintenance of pluripotency in embryonic stem cells (ESC) is regulated by specific transcription factors that are activated during preimplantation embryonic development. Following fertilization, the cleaving zygote undergoes the first lineage decision, forming the outer trophectoderm (TE) cells that enclose the inner cell mass (ICM). Long-standing models of how the embryo regulates the differentiation of the ICM and TE propose that cell position drives cell fate, the “inside outside” hypothesis ; or, conversely, that cell fate drives cell position, the cell polarity hypothesis  (reviewed by ).
The prevailing molecular model of lineage specification (Fig. 1A; ) highlights the importance of the POU domain transcription factor OCT4 (also known as OCT3/4 and POU5F1). OCT4 is expressed throughout the early embryo until the blastocyst stage, when its expression becomes restricted to the ICM in the mouse . While OCT4 null mouse embryos appear to form normal blastocysts, with both TE- and ICM-like cell compartments, the embryos die around the time of implantation, possessing only TE-like cells . These results suggest that OCT4 is required for ICM maintenance, but is not essential for initial specification. In contrast, CDX2 (caudal-related transcription factor 2) is restricted to the TE by the late morula stage in the mouse [7,8]. In the absence of CDX2, the blastocyst forms, but a functional TE is not established and the embryo dies prior to implantation , with OCT4 and NANOG expression detected throughout the embryo. These data suggest that CDX2 plays a role in overriding the ICM fate, but is not required for TE specification. Reciprocal inhibition of OCT4 and CDX2 was evident in a stem cell model of early development. Specifically, an increase in OCT4 lead to decreased CDX2 expression, while overexpression of CDX2 reduced OCT4 expression . These data have supported the model in which OCT4 and CDX2 act as “selector genes” for ICM and TE fates and negatively regulate each other to promote the segregation of the two lineages. However, recent studies report CDX2 expression and TE specification appear to be regulated by the transcriptional regulator TEAD4 [9,10].
Stem cells derived from either lineage likewise express the respective markers. The expression of OCT4 is commonly used as a measure of ESC pluripotency [11,12], while CDX2 is a marker for trophoblast stem cells (TSC) . NANOG is a second ICM-specific transcription factor identified in ESC [14,15]. Loss of NANOG expression in ESCs is associated with loss of pluripotency, and differentiation toward primitive endoderm . Mutant embryos fail to develop an ICM  and form only TE and primitive endoderm, supporting a role of NANOG in regulating epiblast cell fate. KLF4 has also been identified as a necessary regulator of stem cell maintenance , as have a number other factors; however, its regulation during preimplantation development has not been investigated. Evidence suggests that OCT4 expression in non-murine embryos, including the human, is not restricted to the ICM [17–19] in vitro or in vivo, possibly reflecting differences in the mechanism responsible for formation of the ICM. Interestingly, a recent report  described the derivation of TSC from rhesus macaque blastocysts that lack CDX2 expression, which is surprising, as human embryos express CDX2 [21,22], and other species of embryos display a similar TE-specific localization . However, a more recent review  suggests that CDX2 is likely localized to the TE in rhesus blastocysts.
Little is known about the expression of lineage-specific markers in rhesus blastocysts, with the exception of OCT3/4 , which appears to display localization similar to that of the mouse. Previous studies have not combined detection of multiple markers of pluripotency and TE-specific CDX2 expression in primates. Therefore, we sought to examine the expression patterns of markers of lineage specification and/or ESC maintenance in rhesus macaque morulae and blastocysts during the period of lineage divergence.
Unless otherwise stated, chemicals were obtained from Sigma-Aldrich (St. Louis, MO).
All procedures were performed according to the Institutional Animal Care and Animal Use Committee protocols approved at the Oregon National Primate Research Center (ONPRC) and the Caribbean Primate Research Center (CPRC), Puerto Rico. Ovarian stimulation was carried out as previously described . In brief, rhesus macaques received a sequential regimen of recombinant human gonadotropins to support follicular growth. Female macaques at ONPRC received intramuscular (IM) injections of 30 IU recombinant human follicule-stimulating hormone (FSH) (Organon, Oss, The Netherlands) twice daily for 7 days, followed by 2 days of 30 IU each of recombinant human FSH and recombinant human luteinizing hormone (LH) (EMD Serono, Rockland, MA) twice daily. During the final 3 days of the recombinant human FSH treatment, animals also received the GnRH antagonist acycline (0.075 mg/kg animal body weight) to prevent a spontaneous LH surge. Ovarian stimulation of rhesus macaques situated at the CPRC was performed by administration of recombinant human FSH (37.5 IU per injection, twice daily, IM; Organon, Roseland, NY) for 8 days, followed by 2 days of recombinant human FSH and recombinant hCG EMD, Serono; 37.5 IU recombinant human FSH and 100 IU recombinant hCG per injection, twice daily, IM). At both primate centers, on the final day of recombinant human FSH treatment, 32–33 h before follicular aspiration, oocyte maturation was induced with a single injection of recombinant hCG (750–1,000 IU, IM).
Procedures for oocyte recovery, sperm collection, insemination, and embryo culture have been described previously [27–29]. Briefly, follicular fluid aspirates were collected in TALP-HEPES containing 0.3% bovine serum albumin (BSA). The aspirates were sifted through a cell strainer (Becton-Dickinson, Franklin Lakes, NJ), rinsed, and oocytes were collected from the resulting suspension. After cumulus removal using hyaluronidase (0.03%), oocytes were rinsed and placed in culture drops of TALP medium supplemented with BSA  and incubated in 5% CO2 in air at 37°C until insemination. Semen was obtained by electroejaculation . Seminal plasma was removed according to standard protocols and obtained spermatozoa were activated with 1 mM each of cyclic AMP and caffeine. Sperm were used for insemination at a final concentration of 1.5–2.0 × 105 sperm per milliliter [28,32]. The presence of pronuclei was assessed 18–20 h after insemination. Presumptive zygotes were transferred to culture drops of amino acid-supplemented Hamster Embryo Culture Medium 6 (HECM-6aa)  and incubated at 5% CO2 in air at 37°C. The culture medium was refreshed every 48 h with HECM-9aa supplemented with 5% fetal bovine serum (FBS).
To confirm the specificity of the antibodies used, rhesus macaque ESC were cultured as previously described  and stained as outlined below for each marker. Briefly, Ormes 22 ESCs cultured as previously described were grown on mitotically inactivated mouse embryonic fibroblast (MEF) feeder cells in Dulbecco’s Modified Eagle Medium (DMEM/F12) (Invitrogen, Grand Island, NY) supplemented with 15% FBS (Hyclone, Logan, UT), 0.1 mM β-mercaptoethanol, 1% nonessential amino acids (Invitrogen, Grand Island, NY), 2 mM l-glutamine (Invitrogen, Grand Island, NY), and 4 ng/mL FGF2 (Sigma), at 37°C under a 5% CO2-balanced air atmosphere.
Mouse TSC isolated from E3.5 blastocysts or E6.5 extraembryonic ectoderm provided as a gift by Dr. Janet Rossant were cultured as previously described  and plated on slides for immunofluorescent localization of CDX2 to confirm antibody specificity.
Immunofluorescence of compact morulae and various blastocyst stages was carried out as described elsewhere . At least 10 embryos (from a minimum of five replicates) were analyzed from each stage. Compact morulae were collected at 120 h post-insemination and blastocysts were collected at 164 and 214 h post-insemination, fixed in 4% paraformaldehyde (PFA), and stained immediately, or stored in 0.4% PFA at 4°C for no longer than 1 week prior to staining. Primary antibodies were obtained from R&D Systems (NANOG, OCT3/4, KLF4; Minneapolis, MN), BioGenex (CDX2; San Ramon, CA), and Santa Cruz Biotechnology (OCT4; Santa Cruz, CA). Embryos were subsequently incubated with Cy-3-conjugated donkey anti-goat IgG, or donkey anti-mouse IgG. Embryos were counterstained for 1 min with 10 µg/mL DAPI (Calbiochem, La Jolla, CA) and mounted in Vectoshield (Vector Laboratories, Burlingame, CA). Mounted embryos were examined using an Olympus BX41 fluorescence microscope equipped with a DP71 color camera. Controls were performed by omission of the primary antibody and use of the species-specific IgG.
Maintenance of pluripotency in ESC is regulated by a core group of transcription factors whose expression is activated during preimplantation embryonic development. While the interplay between OCT4 and CDX2 is thought to regulate ICM and TE fate, the mechanism of lineage specification in mammalian embryos remains poorly understood. A large proportion of our understanding comes from studies of mouse development, leading to the current model of lineage specification regulated in part by the transcription factors OCT4 and CDX2 (Fig. 1A). Genetic ablation of OCT4 or CDX2 prevents implantation, resulting from the lack of the establishment of a functional ICM or TE, respectively. However, in both cases nonfunctional ICM or TE formation is initiated with deficient expression of the opposing transcription factor, suggesting that each factor is required for subsequent maintenance rather than initial specification.
Consistent with this view, OCT4 expression was not exclusively localized to the ICM of primate blastocysts at 164 h post-insemination (Fig. 2D). To determine whether OCT4 localization became more specific at a later stage, embryos were cultured to 214 h post-insemination. OCT4 localization became more restricted at 214 h (Fig. 2G), although weak staining was still present in the TE in all embryos examined. Blastocysts examined at 164 and 214 h post-insemination displayed an ICM-specific localization pattern for NANOG (Fig. 3). CDX2 protein was localized to the TE at both 164 and 214 h post-insemination (Fig. 4). KLF4 protein expression was not specific to either lineage at any stage (Fig. 5). OCT4, NANOG, CDX2, and KLF4 were detected ubiquitously in compact morula stage embryos (Figs. 2–5). Examination of protein expression during cleavage stages demonstrated nuclear localization of OCT4 only in 16+ cell embryos (Supplementary Fig. 1; Supplementary materials are available online at http://www.liebertpub.com). Antibody specificity was confirmed by analyzing monkey ESC (Figs. 2, ,3,3, and and5),5), which displayed nuclear localization of OCT4, NANOG, and KLF4, respectively. Non-pluripotent MEFs were not labeled by these antibodies. Mouse TSC expressed CDX2 in some, but not all, nuclei (Fig. 4). Negative controls stained without primary antibody, or using the corresponding species of non-immune IgG, were negative for staining (Supplementary Fig. 1).
Previous studies in several species have documented both TE and ICM localization of OCT4, including early stages of mouse preimplantation embryo development , and in human preimplantation embryos [22,36]. Table 1 summarizes the distribution of OCT4, NANOG, CDX2, and KLF4 in preimplantation embryos of several mammalian species found in this and prior reports. TE and ICM localization of OCT4 has been demonstrated in both bovine and porcine blastocysts derived in vivo or cultured in vitro [19,37], as well as in human blastocysts [17,18]. However, these results contrast with those of Mitalipov et al.  where OCT4 expression in rhesus macaque expanded and hatched blastocysts, cultured to 164 h post-insemination, displayed an ICM-specific localization. This result was never observed in the present study, despite use of the same culture methods and antibody (Santa Cruz, CA), although Mitalipov et al.  utilized intracytoplasmic sperm injection for fertilization, while this study relied on standard in vitro fertilization methods. Unrestricted expression of OCT4 protein during in vitro embryo culture may reflect an inability to respond to appropriate environmental stimuli, including growth factors, although this remains to be examined. Slight alterations in the number of embryos expressing OCT4 (and other) transcripts has been shown in human embryos associated with differences in growth factor supplementation of culture media . Results of the present study suggest that OCT4 may be more sensitive to culture conditions, while NANOG is less affected, localizing specifically to the ICM regardless of the culture environment. Furthermore, all embryos analyzed may have exhibited inappropriate OCT4 expression predictive of poor developmental ability. This may be reflected in the rates of ESC derivation reported for the monkey (27%) , and has been correlated with ESC derivation success in the mouse . It should be noted that mouse in vitro-cultured embryos can display an unrestricted pattern of OCT4 expression (TE and ICM) during blastocyst formation, compared with equivalent in vivo-collected embryos , perhaps indicating suboptimal culture conditions. This may be the case in the rhesus macaque where culture to the blastocyst stage still relies on the use of serum. However, in vivo produced porcine embryos also display OCT4 expression throughout both the ICM and TE compartments . The ratio of OCT4:CDX2 may be important for the proper delineation of the TE and maintenance of pluripotency within the ICM, but may not be essential for the initial specification of the ICM. In mice, TEAD4 appears to provide the signal that specifies TE prior to blastocyst formation upstream of CDX2 [9,10]. A comparable protein has not been identified for ICM specification, perhaps indicating that ICM is the default state.
Unlike OCT4, NANOG expression was restricted to the ICM, suggesting that localization of NANOG precedes that of OCT4 in the primate. A recent study of human preimplantation embryos by Cauffman et al.  documented similar, although variable, ICM-specific localization of NANOG at the expanded blastocyst stage. Kimber et al.  also demonstrate ICM-specific localization of NANOG in human blastocysts. Expression of NANOG was not detected prior to blastocyst expansion , contrasting that observed in rhesus compact morulae in the present study. SOX2 expression in human embryos was localized largely to ICM nuclei of expanded human blastocysts , but was detected in all blastomeres in early blastocysts [22,36] and in early stages . Considering previous localization of OCT4 protein expression in human embryos to both the TE and ICM, results suggest that both SOX2 and NANOG are specified in the ICM upstream of OCT4 in primates. However, a cooperative interaction between OCT4 and SOX2 has been proposed to regulate NANOG , based on a study performed in mouse ESC. Other studies show cooperation between all three transcription factors ; however, OCT4 appears not to initiate the signaling cascade in primate embryos. The regulatory importance of NANOG is supported by mouse knockout studies, in which OCT4 ablation results in the formation of an ICM, albeit with TE properties , while NANOG ablation results in the complete absence of an ICM . However, Dietrich and Hiiragi  have reported mosaic NANOG expression in mouse embryos. ESC deficient in NANOG fail to maintain pluripotency and undergo differentiation . Additionally, selection of induced pluripotent stem cells was improved by the use of NANOG instead of Fbx15 [42,43]. Recently, upstream regulators of NANOG have been found in ESC, including Med12 (mediator complex 12)  and SMAD signaling . Whether these factors play a role in regulating lineage specification in mammalian embryos requires investigation. Recent studies of early lineage specification have largely focused on the involvement of cell polarity in regulating cell position , as a potential activator of transcription factor responses.
Niwa et al.  demonstrated that the ratio of OCT4 to CDX2 was critical in mice for determining the differentiation of TE and ICM. Whether this is also true for non-human primates requires further examination. Primate embryo lineage specification may instead rely on a reciprocal relationship between NANOG and CDX2, with upstream effectors up-regulating these factors. In contrast to a previous report , CDX2 was detected in the TE of developing primate embryos in the present study (as cited by ). Significantly, Vandevoort et al.  reported the derivation of rhesus TSC lacking CDX2 at both the transcript and protein level. While the authors tested two antibodies, which also failed in the present study (data not shown), the antibody that successfully detected CDX2 has been used extensively in other reports [7,41,46,47]. A recent review by Vandevoort et al. supports our current findings, documenting CDX2 localization in the TE of rhesus macaque blastocysts . Therefore, of the transcription factors associated with lineage specification analyzed in the present study, only NANOG and CDX2 displayed lineage-specific localization in rhesus blastocysts.
KLF4 expression has been associated with the maintenance of pluripotency in human ES cells , and has been used as one of the four transcription factors required to induce pluripotency in somatic cells . Its localization during early embryo development has not been reported previously. Here we demonstrate that KLF4 does not localize to a specific compartment within the primate embryo. KLF4 cooperates with OCT4 and SOX2 to regulate pluripotency . However, the function of KLF4 during early development and its specific role in regulating pluripotency are less clear. Based on its lack of cell specificity, KLF4 may play a role in regulating both embryonic and TSC populations, or it could participate in the potentiation of genes that regulate pluripotency. Whether KLF4 controls later differentiation steps, or in combination with other KLFs regulates the core transcription factors, should be explored with the use of knockout mouse models. Rather than an activating role in reprogramming, KLF4 may act to repress transcription  in support of a nonspecific localization pattern in preimplantation embryos. Examination of additional transcription factors, including GATA6 and SOX2, could delineate the regulatory mechanisms that control lineage specification in the primate.
The dynamic process of cell fate specification is modulated by regulatory gene networks. Results of this study demonstrate, for the first time, TE-specific localization of CDX2 in primate blastocysts. Restricted ICM-specific localization of NANOG suggests that it is a potent indicator of ICM pluripotency. The unrestricted expression pattern of OCT4 at the blastocyst stage in primates may represent a key regulatory distinction between the human and mouse that is also reflected by differences in ESC maintenance, whereby mouse ESC are dependent on leukemia inhibitory factor (LIF) and bone morphogenic protein (BMP) , while these factors are not sufficient to maintain pluripotency, or result in differentiation, of human ESC . Recent analysis of mouse epiblast stem cells (EpiSC) isolated under human ESC culture conditions (in the presence of bFGF and activin A) has documented higher expression levels of both NANOG and OCT4  than observed in mouse ESC, as well as an ability to generate TE that has not been achieved from mouse ESC . Likewise, EpiSC lack expression of Rex1 and alkaline phosphatase, known markers of ESC and ICM. Whether these markers are expressed in rhesus ICM is yet to be determined, but raises the question as to whether primate ESC truly represent their ICM counterparts or are epiblast in nature.
Our data suggest that the murine model of lineage determination may not accurately mirror other species. NANOG and CDX2, rather than OCT4 and CDX2, may interact to repress each other in each respective lineage in the primate (Fig. 1B). This is supported by a lack of ICM-specific localization of OCT4, preceded by ICM- and TE-specific localization of NANOG and CDX2, respectively. In this respect, NANOG may antagonize the activity of CDX2, and vice versa, to maintain ICM and TE lineages. Downstream specification of primitive endoderm and epiblast may be mediated by GATA6, although this was not examined in the present study. Further molecular analysis is required to elucidate the interactive network and upstream mediators of this first differentiation step.
A.J. Harvey, Department of Physiology, Wayne State University, Detroit, Michigan. Department of Obstetrics and Gynecology, Wayne State University, Detroit, Michigan.
D.R. Armant, Department of Obstetrics and Gynecology, Wayne State University, Detroit, Michigan. Department of Anatomy and Cell Biology, Wayne State University, Detroit, Michigan. National Institutes of Health, NICHD, Reproductive Biology and Medicine Branch, Bethesda, Maryland.
B.D. Bavister, Department of Obstetrics and Gynecology, Wayne State University, Detroit, Michigan. Caribbean Primate Research Center, Sabana Seca, Puerto Rico.
S.M. Nichols, Reproductive Biology Program, Caribbean Primate Research Center, Unit of Comparative Medicine, University of Puerto Rico, Sabana Seca, Puerto Rico.
C.A. Brenner, Department of Physiology, Wayne State University, Detroit, Michigan. Department of Obstetrics and Gynecology, Wayne State University, Detroit, Michigan.
Preliminary data relating to this study were presented at the International Society for Stem Cell Research, Philadelphia, PA, 2008, and the Society for Gynecological Investigation, San Diego, CA, 2008.
We thank Dr. Janet Rossant at the Hospital for Sick Children and University of Toronto for generously providing the TSC line used in this study. Supported, in part, by the intramural research program of the Eunice Kennedy Shriver National Institute of Child Health and Human Development, National Institutes of Health, DHHS and NIH Grants HD045966, RR015395, RR021881, and HD046553. Support was also provided by Wayne State University through the President’s Research Enhancement Program.
There are no conflicts of interest relating to this work for any of the authors.