Pluripotency is defined as the capacity of a cell to differentiate into all lineages composing the adult organism, including the germ line. At least in the mouse, the transcription factors Oct4 (Pou5f1), Sox2, and Nanog have been identified as a triad that stabilises pluripotency both in the embryo and in embryonic stem cells (ESCs). In the murine embryo, Oct4 is essential for the formation and maintenance of the inner cell mass (ICM), and down-regulation of Oct4 below a certain threshold results in trophectoderm differentiation both in vivo and in vitro (Nichols et al., 1998). Oct4 regulates pluripotency-related gene expression in partnership with the transcription factor Sox2 (Yuan et al., 1995). Another transcription factor, Nanog, seems to be essential during the subsequent process when the ICM differentiates into the epiblast and hypoblast, since in Nanog-null embryos the epiblast does not form properly and hypoblast is either not formed or quickly degenerates (Mitsui et al., 2003; Silva et al., 2009).
The expression pattern of NANOG and OCT4 has been reported to differ markedly between embryos from different mammalian species indicating that these transcription factors may have distinct species-specific roles during development (Kirchhof et al., 2000; Kuijk et al., 2008). In the porcine embryo, NANOG transcripts have been detected in 4- and 8-cell stages, morulae, blastocysts, epiblasts, and several organs such as brain, liver, and lung (Brevini et al., 2007; Kumar et al., 2007; Blomberg et al., 2008; Hall et al., 2009). NANOG protein, on the other hand, has not been localised to porcine embryos until after hatching of the blastocyst (Kuijk et al., 2008), when it becomes localised to the epiblast at day 9 and 11 of development (Hall et al., 2009).
OCT4 transcripts have been detected in porcine oocytes and throughout pre-implantation development until establishment of the epiblast (Magnani and Cabot, 2008; Hall et al., 2009). OCT4 protein has been localised to the ICM and trophectoderm of porcine blastocysts before hatching (Kirchhof et al., 2000; Kuijk et al., 2008), whereas after hatching the localisation becomes restricted to the ICM (Vejlsted et al., 2006). Later, the OCT4 localisation is restricted to the epiblast, and during gastrulation the mesoderm retains some OCT4 expression. During late gastrulation and neurulation, OCT4 expression seems to be gradually lost in an anterior-to-posterior direction, in order, finally, to be retained exclusively in the presumptive primordial germ cells (PGC) (Vejlsted et al., 2006). One group has, however, reported detection of OCT4 protein as well as mRNA in the extra-embryonic tissue in elongating porcine embryos as late as day 11 (Keefer et al., 2007).
E(pithelial)-CADHERIN is a cell-to-cell adhesion molecule expressed in all mammalian epithelia and it is implicated in many developmental processes. In the mouse, E-Cadherin is necessary for a proper establishment of the blastocyst since E-cadherin (Cdh1) null embryos fail to form a trophectoderm and are unable to generate the blastocyst cavity (Larue et al., 1994; Ohsugi et al., 1997). When the embryo gastrulates, E-Cadherin is down-regulated in the primitive streak to allow for an epithelial-mesenchymal transition and ingression through the streak (Damjanov et al., 1986; Ciruna and Rossant, 2001). E-Cadherin is expressed in ectoderm and endoderm, but not in mesoderm during gastrulation (Damjanov et al., 1986; Burdsal et al., 1993; Ciruna and Rossant, 2001). The specific expression pattern and role of E-CADHERIN has, to our knowledge, not previously been described in porcine embryos.
Double labelling for NANOG and OCT4 has been reported in porcine pre-gastrulation embryos (Hall et al., 2009) but during primitive streak formation the simultaneous expression pattern of these pluripotency markers is still not known and, thus, in this study, we investigated the three-dimensional expression pattern of NANOG, OCT4, and E-CADHERIN in porcine pre- and peri-implantation embryos by triple-labeling immunohistochemistry and following confocal microscopy. A refinement of immunohistochemical protocols has allowed for higher resolution analysis than previous studies and we show here that the localisation of NANOG in the porcine embryo is remarkably different from not only the mouse embryo, but also from any other species studied so far with respect to onset of appearance as well as expression pattern. Further, we have identified a surprising territorial and mutually exclusive localisation for NANOG and OCT4 as well as an intriguing localisation of NANOG and E-CADHERIN to the same cell populations.
In hatched, spherical blastocysts, where the epiblast was still covered by a thin layer of trophectoderm, i.e., the Rauber's layer, at day 9 to early day10 of gestation, OCT4 was exclusively observed in nuclei of the epiblast (Wolf et al., 2010) (Fig. 1A,B,D; n=4). E-CADHERIN was localised to the cell periphery in epiblast as well as trophectoderm. NANOG was not detected in any compartment (Fig. 1A,C; n=4). The red colour in Figure B–D is non-nuclear, unspecific background staining, which was also seen when the primary antibody to NANOG was preabsorbed with recombinant NANOG. In these negative controls, no nuclear background staining was seen and, therefore, specific, nuclear staining was easily distinguished from unspecific staining.
At day 10 of gestation, the trophectoderm had expanded and the embryos had developed into larger, spherical blastocysts (inset in Fig. 2B), where Rauber's layer had more or less completely disappeared and where the epiblast had established the embryonic disc. No signs of polarity were visible in the epiblast at this stage of development. However, from the closer apposition of the hypoblast cells in a region around the anterior pole of the embryonic disc (data not shown), the anterior-posterior axis could be determined according to previous descriptions by Hassoun et al. (2009). NANOG was localised to nuclei in the entire epiblast and to individual, but not all, nuclei of the hypoblast underlying the epiblast and the trophectoderm (Fig. 2A,D,G,I; n=4). The hypoblast underlying the epiblast and the trophectoderm around the posterior half of the epiblast stained more intensively for NANOG as compared to the remaining hypoblast (Fig. 2A,D,G,I; n=3). To confirm the hypoblast nature of the underlying NANOG-positive cells, we co-stained for the hypoblast marker HNF4α (Duncan et al., 1994) and found co-localisation of this marker and NANOG (Fig. 2G–I; n=2). In the epiblast, NANOG was co-localised with OCT4, which was expressed in the entire epiblast (Fig. 2B,C,F). E-CADHERIN was localised to the cell periphery in epiblast as well as trophectoderm (Fig. 2C,F,H,I).
In ovoid blastocysts, at late day 10–11 of gestation, which presented a fully developed embryonic disc, NANOG was still localised to all nuclei of the epiblast (n=4), but in two embryos the staining was slightly more intense in the posterior end of the epiblast (Fig. 3A). NANOG was detected in most nuclei in the hypoblast underlying the epiblast and the trophectoderm (Fig. 3A,D,F) and staining was still more intense in the hypoblast surrounding the posterior half of the disc (Fig. 3A). In the epiblast, NANOG was co-localised with OCT4 (Fig. 3B,C,E), and a restricted number of epiblast cells displayed markedly stronger NANOG-staining than the rest (Fig. 3A,D,F). Interestingly, some of the cells that stained very intensely for NANOG also displayed a particularly intense staining for E-CADHERIN (Fig. 3C, arrow in F).
During day 12–14, the porcine trophectoderm and hypoblast expand dramatically forming an up to one-meter-long, filamentous embryo. In embryos at around day 12 of gestation, which had assumed a tubular or filamentous appearance, but did not present a primitive streak, the anterior-posterior axis now became evident by a higher density of cells in the posterior rim of the epiblast and the appearance of extra-embryonic mesodermal cells that occupied a position between the posterior and lateral trophectoderm and hypoblast (Oestrup et al., 2009) and that were clearly distinguishable in the stereomicroscope (data not shown). In these embryos, NANOG was localised to nuclei in a minor portion of the epiblast forming small territories (n=3). In one case, the NANOG territories formed lines converging towards the posterior pole of the embryonic disc (Fig. 4A), whereas in another two cases only a few cells here and there in the epiblast stained for NANOG. OCT4, on the other hand, was localised to the remaining nuclei of the epiblast in a mutually exclusive, non-overlapping pattern (n=3) (Fig. 4B,D,F), except for a few cells at the posterior pole of the embryonic disc, which co-expressed NANOG and OCT4 (n=3) (arrowhead in Fig. 4D). These cells are believed to represent the primordial germ cells (PGCs). Moreover, OCT4 was also localised to nuclei of the extra-embryonic mesoderm (Fig. 4B,C,D), which had formed before the appearance of a visible primitive streak. The hypoblast no longer exhibited NANOG-staining. Interestingly, the NANOG-positive epiblast cells again displayed a particularly intense staining for E-CADHERIN (Fig. 4C,E,G; n=3).
In filamentous embryos, in which the embryonic disc had become more ovoid and slightly pointed at the posterior pole (arrow in Fig. 5B), at day 12–13 of gestation, and in which the primitive streak had become visible based on the OCT4-staining pattern (see later in this section), NANOG localisation had become restricted to a few nuclei in the most posterior, pointed portion of the epiblast (Fig. 5A,D; n=2) Some embryos (n=2) at this stage of development still expressed NANOG, but not OCT4 in a few cells in the epiblast as previously described. OCT4 was localised to almost all nuclei of the epiblast, but a notably weaker staining was evident in a band of the epiblast stretching anteriorly (n=4), from the posterior pole to about halfway through the epiblast identifying the primitive streak (arrowheads Fig. 5B). Moreover, OCT4 was localised to nuclei of the intra- (Fig. 5E) and extra-embryonic mesoderm (Fig. 5B; n=4).
In filamentous embryos, in which the neural groove had formed, at around day 15 of gestation, NANOG was exclusively localised to the presumptive PGCs. These cells were scattered in the caudo-ventral aspect of the yolk sac (arrow in Fig. 6A; n=4) and formed a cell cluster (arrowhead in Fig. 6A) attached to the caudo-ventral wall of the yolk sac extending into the extra-embryonic coelom (Fig. 6D). OCT4 was localised to nuclei in the epiblast and mesoderm as well as being co-localised with NANOG to nuclei of the presumptive PGCs (Fig. 6B,C,E,G; n=4). However, in two embryos a particular down-regulation of OCT4 was evident in the most posterior part of the ectoderm, corresponding to the posterior primitive streak region (Fig. 6B–D). E-CADHERIN was localised to the developing ectoderm, endoderm, and trophectoderm as well as to the cluster of potential PGCs (Fig. 6D).
To understand and artificially control pluripotency in vivo and in vitro, detailed knowledge on the expression patterns of pluripotency-associated factors in different species is a prerequisite. Based on a growing body of evidence, it is likely that the pluripotent cell population is established by different mechanisms in different mammalian species. In the present study, we examined the 3D-localisation of the pluripotency-associated transcription factors NANOG and OCT4 together with the epithelial marker E-CADHERIN in porcine pre- and peri-implantation embryos by immunohistochemistry.
Molecular knowledge on the regulation of cellular pluripotency has so far mainly been generated in the mouse as well as from additional studies on the signalling pathways in murine and human ESCs. In the mouse, it is evident that Oct4 and Nanog expression play crucial roles in early lineage segregation: Oct4 is important for ICM-formation and maintenance (Nichols et al., 1998), whereas Nanog is important for establishment of the epiblast and hypoblast (Mitsui et al., 2003; Silva et al., 2009). It has been stated several times that the maintenance of the pluripotent cell population must be regulated by alternative mechanisms in other mammalian species, such as man, goat, cow, and pig. One fact that has sparked this opinion is controversy that OCT4, in addition to being expressed in the ICM, is also expressed in the trophectoderm of the pre-hatching blastocyst in these non-rodent species (Cauffman et al., 2006; He et al., 2006). Interestingly, some data indicate that trophectoderm expression of this transcription factor is actually also found in the mouse (Palmieri et al., 1994; Silva et al., 2009). Accordingly, we also detected Oct4 in the trophectoderm in pre-hatching mouse blastocysts (unpublished data) using the same antibody as we have used for detection of porcine OCT4 in this study. Moreover, analysis of OCT4 mRNA in bovine and mouse blastocysts has revealed similar patterns of expression in both the ICM and trophectoderm, though at a lower level in the latter compartment (Kurosaka et al., 2004). These controversies have sparked the notion that differences are likely to exist between OCT4 protein localisation and its mRNA expression. In this study, we clearly demonstrated that OCT4 protein expression is restricted to the epiblast in the hatched porcine blastocyst. This finding is contrasted by other reports on expression of OCT4 protein and mRNA in the trophectoderm at this stage of development in both porcine and caprine blastocysts (He et al., 2006; Keefer et al., 2007). This contradiction may be explained by the possible existence of two isoforms of OCT4. In man, two OCT4 isoforms, OCT4A and OCT4B, have been reported, but only OCT4A is associated with regulation of pluripotency (Cauffman et al., 2006). Antibodies are available that recognise either the individual isoforms (N-terminal) or both isoforms (C-terminal). The antibody used in the present study recognises exclusively the human isoform of OCT4A, whereas the antibody used by Keefer et al. (2007) recognises a human 265-aa isoform of OCT4, which corresponds to isoform B, and the antibody used by He et al. (2006) is likely to recognise both isoforms, since it is raised against the C-terminal of OCT4, which is shared by the two isoforms. In the present study, we clearly show that OCT4 is expressed exclusively in the epiblast until the first OCT4-expressing mesoderm is formed, just prior to the appearance of the primitive streak. Further, great care has to be taken when examining extra-embryonic tissue at this early stage of development, especially by PCR, where the exact localisation of the transcript is illusive, since this region consists of at least two and, at later developmental stages, three different cell types: trophectoderm, hypoblast, and, later, extra-embryonic mesoderm. Our data indicate that observed OCT4 mRNA in the regions outside the epiblast is likely to originate from extra-embryonic mesoderm and with our data in mind and the possible existence of two isoforms of OCT4, it is likely that the role of OCT4 in the hatched, porcine blastocyst resembles that in the mouse. We have earlier reported that OCT4 is gradually lost in an anterior-to-posterior direction in porcine embryos with a neural groove (Vejlsted et al., 2006). We show here that in some embryos at this developmental stage, however, OCT4 is not detected in the most posterior part of the ectoderm, while OCT4 is still expressed in the more anterior ectoderm. The loss of OCT4 in the posterior ectoderm may be explained by a further regression of the primitive streak as compared to the embryos still expressing OCT4 in this posterior region. Analysis of this phenomenon would require staining of embryos with a primitive streak marker, which, however, is not within the scope of this report.
In the porcine embryo, NANOG protein has not been detected in the ICM (Kuijk et al., 2008; Hall et al., 2009) and, furthermore, in this study we did not detect expression of NANOG before the disintegration of Rauber's layer, and thus the role of this transcription factor for early lineage segregation in the porcine embryo must be questioned. In one study, a weak staining for NANOG was reported at day 9 of development, i.e., immediately before the loss of Rauber's layer (Hall et al., 2009). We were not able to repeat this observation using the same antibody and signal amplification by TSA. This discrepancy may be due either to actual variations between individual embryos or to the difficulties in distinguishing between weak specific antibody staining and unspecific background staining. Hence, in this study we did notice a significant unspecific staining after labelling with the NANOG antibody (see Fig. 1). At day 11 of development, NANOG mRNA expression levels in porcine epiblasts has shown to vary remarkably between embryos (Hall et al., 2009). However, no variation in protein expression at this stage of development was reported (Hall et al., 2009). In this study, we found that the expression of NANOG at day 10 to 11 of development varies between embryos since in some embryos all epiblast cells stained equally for NANOG whereas in other embryos NANOG staining was more intense at the posterior pole. The fluctuation in NANOG mRNA at day 11 of development could also be a predecessor for the territorial expression of NANOG, we show here, in a variable, but minor, proportion of the epiblast as the embryos elongates at day 12 of development. The highly dynamic expression pattern of NANOG and OCT4 in the porcine early embryo seems to be very complex and different from what has been reported in other species. The expression of NANOG and OCT4 seems to be delicately regulated relative to each other in the pig, and in contrast to other species studied so far OCT4 is not always expressed in cells expressing NANOG. Also the overlapping expression of NANOG and an abundance of E-CADHERIN in the porcine epiblast just prior to gastrulation are findings that, to our knowledge, have not been reported in any other species studied so far. Further studies are definitely needed to shed light on what developmental significance this apparently mutually related expression pattern has. In the mouse blastocyst, NANOG is co-expressed with OCT4 and seems to be important for ICM cells to properly segregate into epiblast and hypoblast (Silva et al., 2009). The expression profile of NANOG in the porcine embryo indicates another, and later, role of this transcription factor. In the bovine embryo, which exhibits developmental features comparable with its porcine counterpart, the expression profile of NANOG is, however, more similar to that in the mouse. Hence, NANOG has been localised to the ICM of the zona pellucida–enclosed blastocyst by in situ hybridisation (Kuijk et al., 2008), to the epiblast of ovoid blastocysts, and to the epiblast of early filamentous embryos (Degrelle et al., 2005). However, NANOG has also been detected by PCR in regions localised outside the epiblast in ovoid and filamentous bovine embryos (Degrelle et al., 2005). The authors referred to these regions as being trophectoderm, but according to their protocol, these tissues are likely also to include hypoblast, which, based on our findings, is a potential site of NANOG expression.
A factor that should be kept in mind when comparing expression patterns of pluripotency factors in different species is the differences in the timing of key developmental events. For example, the porcine ICM and epiblast are maintained for an extended period of time (6–7 days) compared with their murine (1 day) and human (3 days) counterparts, a fact that complicates a direct comparison between the three species.
In conclusion, we find that the expression pattern of OCT4 in porcine blastocysts resembles that seen in the mouse blastocyst, but that the dynamic expression pattern of NANOG and to some extent OCT4 in porcine embryos differs from what has been reported in other species. This may indicate a different developmental role for NANOG in this species. However, NANOG is down-regulated in the epiblast as the primitive streak develops, indicating a link between NANOG expression and pluripotency in the porcine embryo. The relationship between the NANOG, OCT4, and E-CADHERIN expression domains is a notable finding and calls for further investigation. Initially, OCT 4 but not NANOG is localised to the ICM, later NANOG and OCT4 become co-localised to the epiblast, with particularly strong NANOG-staining in cells also staining intensively for E-CADHERIN. Later, a mutually exclusive localisation of NANOG and OCT4 is established in the epiblast and at the same time NANOG and OCT4 are co-expressed in presumptive PGCs at the posterior pole of the epiblast. NANOG is localised to small E-CADHERIN-positive, epiblast territories and OCT4 to the remaining epiblast cells and from around the time where the primitive streak becomes visible and at least until the formation of the neural groove, NANOG and OCT4 are co-expressed in the presumptive PGCs only.
Embryo Collection and Immunohistochemistry
Uterine horns from slaughtered sows were flushed with 150 ml embryo flushing medium (LIFE Pharmacy, KU, Frederiksberg, Denmark) with 1% FBS 8 to 16 days after insemination. Embryos (Danish landrace × Yorkshire crosses) were collected by filtration of flushing fluid through a filter and embryos were collected under a stereomicroscope. In filamentous embryos, the embryonic disc and the nearest trophectoderm were carefully separated from the other embryos using forceps and the main part of the trophectoderm was cut off with a scalpel, thereby releasing the embryo from the entangled trophectoderm. Embryos were fixed in 4% paraformaldehyde in 0.1 M PBS for 30 min at room temperature (RT) and transferred to methanol for storage at −20°C. In total, 26 embryos were collected for this study. At least three embryos of each developmental stage were analysed.
For immunohistochemistry, embryos were incubated in 0.5% TNB blocking reagent (PerkinElmer, Waltham, MA) with 0.1% Triton X-100-100 (Sigma Chemicals Co, St. Louis, MO) and 3% H2O2 (Merck, Darmstadt, Germany) for 2 hr at RT to promote antibody penetration, block unspecific antigen binding, and quench endogenous peroxidases. Following this, embryos were incubated with a mix of three primary antibodies diluted in 0.5% TNB and 0.1% Triton X-100 overnight at 4°C. The antibodies used in this study are Polyclonal Rabbit-anti-NANOG 1:10,000 (PeproTech, Rocky Hill, NJ), Polyclonal Goat-anti-OCT4 SC8628 1:1000 (Santa Cruz Biotechnology, Santa Cruz, CA), monoclonal mouse anti-E-CADHERIN 1:1,000 (cat. no. 610181; BD Biosciences, San Jose, CA), and polyclonal Goat-anti-HNF4α SC6556 1:500 (Santa Cruz). Embryos were washed for at least 3 times 20 min in 0.1 M PBS with 0.1% Triton X-100 at RT and incubated with a mix of three secondary antibodies diluted 1:500 in 0.5% TNB and 0.1% Triton X-100 for at least 2 hr at RT. The secondary antibodies used are whole IgG biotin anti-rabbit (cat. no. 711-065-152; for TSA amplification), Cy2-anti-goat (cat. no. 705-225-147), Cy5-anti-mouse (cat. no. 715-175-151) all from Jackson ImmunoResearch Laboratories Inc. (West Groove, PA). Embryos were washed for at least 3 times 20 min in 0.1 M PBS with 0.1% Triton X-100 at RT. To TSA amplify the biotin anti-rabbit secondary antibody, embryos were incubated with peroxidase conjugated streptavidin (Zymed histotation kit, Zymed, San Francisco, CA) for 1 hr at RT and washed as previously described. The signal was developed by incubation for 15 min at RT with Cy3-TSA substrate 1:100 (Cy3-TSA kit, PerkinElmer) in amplification diluent (PerkinElmer). Embryos were washed at least 3 times 20 min in 0.1 M PBS with 0.1% Triton X-100 at RT and transferred to Methanol where they were allowed to equilibrate for at least 1 hr at 4°C, transferred to fresh methanol, and stored at −20°C until examination by microscopy.
When primary antibody was omitted, no specific staining was seen with any of the secondary antibodies used in this study and the secondary antibodies did not cross-react with primary antibodies from other species. The specificity of the NANOG antibody was tested by blocking with recombinant human NANOG, which abolished nuclear specific staining. Confocal microscopy embryos were mounted in BABB (1:2 Benzyl Alcohol to Benzyl Benzoate) to clear the embryos.
We thank Ragna Jørgensen for excellent technical assistance.