Embryonic expression of murine 5T4 oncofoetal antigen is associated with morphogenetic events at implantation and in developing epithelia


  • Katie M. Barrow,

    1. CR UK Immunology Group, Paterson Institute for Cancer Research, Christie Hospital NHS Trust, Manchester, United Kingdom
    Search for more papers by this author
  • Christopher M. Ward,

    1. CR UK Immunology Group, Paterson Institute for Cancer Research, Christie Hospital NHS Trust, Manchester, United Kingdom
    Current affiliation:
    1. Centre for Molecular Medicine, Faculty of Medical and Human Sciences, The University of Manchester, M13 9PT, UK
    Search for more papers by this author
  • Jennifer Rutter,

    1. CR UK Immunology Group, Paterson Institute for Cancer Research, Christie Hospital NHS Trust, Manchester, United Kingdom
    Search for more papers by this author
  • Sumia Ali,

    1. CR UK Immunology Group, Paterson Institute for Cancer Research, Christie Hospital NHS Trust, Manchester, United Kingdom
    Search for more papers by this author
  • Peter L. Stern

    Corresponding author
    1. CR UK Immunology Group, Paterson Institute for Cancer Research, Christie Hospital NHS Trust, Manchester, United Kingdom
    • CR UK Immunology Group, Paterson Institute for Cancer Research, Christie Hospital NHS Trust, Wilmslow Road, Manchester, M20 4BX, UK
    Search for more papers by this author


Overexpression of 5T4 oncofoetal antigen, an early marker of ES cell differentiation, in vitro increases cellular motility and decreases adhesion, properties relevant to development and cancer. Embryonic expression of m5T4 antigen is first detected on trophectoderm at implantation and is restricted to extra-embryonic tissues to embryonic day (E) 11.5. In the embryo, significant m5T4 expression is detected at E12.5 in hindbrain roofplate and in various epithelia derived from all germ layers. In keratin 14-expressing epithelia, there is a congruent 5T4 expression pattern with many of these cells being Ki-67 positive. In brain, expression is observed in roofplate, ependymal layers, choroid plexus, and subventricular zones of lateral ventricles at E14.5. By E17.5, expression is decreased in the subventricular zone with further restriction to choroid plexus in adult brain. Our data demonstrate a limited 5T4 expression profile during embryogenesis associated with actively cycling, undifferentiated epithelial progenitor cells that may contribute to their migration. Developmental Dynamics 233:1535–1545, 2005. © 2005 Wiley-Liss, Inc.


Malignant cells of primary tumors often acquire an invasive and migratory phenotype. Placentation may share aspects of such mechanisms where fetal trophoblast cells must invade the maternal host tissue to establish the embryo within the uterus. Cell surface molecules of invading trophoblast that are also expressed by tumor cells may be agents of these processes. One example is the 5T4 oncofetal antigen originally defined by a monoclonal antibody raised against human trophoblast (Hole and Stern, 1988). 5T4 antigen expression in normal non-pregnant adult tissue is restricted to some epithelia at low levels, but is also found on many transformed cell lines and carcinomas of diverse origin (Hole and Stern, 1988; Southall et al., 1990). In colorectal, gastric, and ovarian carcinomas, 5T4 antigen expression correlates with poorer clinical outcome consistent with metastatic spread (Starzynska et al., 1992, 1994, 1998; Wrigley et al., 1995; Mulder et al., 1997; Naganuma et al., 2002). This has promoted the use of 5T4 as a target for immunotherapies (Shaw et al., 2000; Forsberg et al., 2001; Mulryan et al., 2002; Myers et al., 2002).

The 5T4 protein is a 72-kDa highly N-glycosylated transmembrane glycoprotein. The cytoplasmic region contains a potential tyrosine phosphorylation sequence and the extracellular portion contains leucine-rich repeats (LRR) (Hole and Stern, 1990; Myers et al. 1994), motifs involved in protein-protein interactions (reviewed in Kobe and Kajava, 2001). The protein is concentrated at microvillus projections in the plasma membrane and overexpression in fibroblast and epithelial cell lines in vitro causes a decrease in cellular adhesion and more dendritic cell morphology. This change in morphology is related to changes in the actin cytoskeleton, abrogation of actin/E-cadherin containing junctions, and a decrease in cell surface E-cadherin (Carsberg et al., 1995, 1996). The C-terminal end of h5T4 protein is a class 1 PDZ-binding motif that interacts with TIP-2/GIPC (Awan et al., 2002). Since TIP-2/GIPC interacts with the actin cytoskeleton through α-actinin-1 (Bunn et al., 1999), it may be through this interaction that 5T4 exerts effects on morphology and motility.

To determine more about the biological function of 5T4 oncofoetal antigen during development and upregulation in many cancers, reagents to the mouse homologue were developed (King et al., 1999; Woods et al., 2002). The murine 5T4 amino acid sequence is 81% identical to the human sequence including complete conservation of the transmembrane and the cytoplasmic tail domains (King et al., 1999). Similar to h5T4, in vitro overexpression of m5T4 causes alterations in cell adhesion, motility, and morphology. In the adult mouse, the m5T4 antigen is expressed in various epithelia at low levels including choroid plexus of the brain and outer epithelium of the ovary (Woods et al., 2002). Only the latter tissues have significant levels of 5T4 transcripts but preliminary studies detected high levels of transcripts in embryos (King et al., 1999). In addition, embryonic stem cells are 5T4 negative but rapidly upregulate expression when allowed to differentiate, with cells giving rise to all three germ layers becoming positive (Ward et al., 2003). One possibility is that 5T4 plays a role in early embryogenesis where migration or motility of cells and cell-cell interactions are pivotal.

This study is focused on examining the expression of m5T4 during murine embryogenesis using immunolabeling. The expression pattern is associated with actively cycling, undifferentiated epithelial progenitor cells and previous data suggest a possible role in their migration.


5T4 Antigen Is Expressed Following Hatching of Blastocyst From Zona Pellucida

Embryos were examined for 5T4 antigen expression at all stages of preimplantation development from fertilized oocytes to blastocysts by immunofluorescence using rat mAb 9A7 recognizing m5T4, E-cadherin-positive control or rat IgG-negative control. Numbers of embryos labeled at each stage of pre- and peri-implantation development with results are shown in Table 1. No labeling was seen with rat IgG control at any stage and, in contrast, E-cadherin was detected on embryos from the 2-cell to the hatched blastocyst stage, principally at the cell surface (not shown). No m5T4 expression was observed from fertilized oocyte to E3.5 blastocyst stage (Fig 1Ai–iv, and corresponding phase images shown in Fig 1Bi–iv). However, at E4, about half of the blastocysts that had hatched from the zona pellucida in utero expressed 5T4 antigen, and labeling was mostly cell surface and punctate (Fig. 1C; Table 1). Dual labeling of hatched blastocysts (Fig. 1D) using OCT-4 (red) as a marker of ICM cells showed that 5T4 antigen (green) is expressed only at the cell surface of trophectoderm and not by ICM cells.

Table 1. Numbers of Embryos Examined at Each Stage of Pre- and Peri-Implantation Development and Results of Labelling With Each Antibody
StageAnti-m5T4Isotype negative controlE-cadherin
Fertilised oocyte15/15 −ve14/14 −ve15/15 +ve
2-cell8/8 −ve6/6 −ve7/7 +ve
4-cell17/17 −ve15/15 −ve16/16 +ve
8-cell10/10 −ve5/5 −ve10/10 +ve
Morula5/5 −ve2/2 −ve4/4 +ve
E3.5 blastocyst27/27 −ve28/28 −ve27/27 +ve
Hatched E4 blastocyst14/31 (9/16 single; 4/15 dual +ve)21/21 −ve27/27 +ve
Figure 1.

5T4 antigen is first detected in preimplantation embryos on acquisition of attachment competence as detected by indirect immunofluorescence and confocal microscopy at different stages of development. A: mAb 9A7, B: corresponding phase image (i) fertilized oocyte; (ii) 2-cell stage; (iii) 8-cell stage; (iv) unhatched E3.5 blastocyst. Images shown are at 600x magnification and are representative of all embryos in that group. C: (i) blastocyst hatched in utero labeled with mAb 9A7 with corresponding phase image shown in (ii); (iii) and (iv) ∼35 fold enlargement of boxed area in (i) in same focal plane; 600x magnification. D: 5T4 antigen (green) is expressed only by trophectoderm cells in the late hatched blastocyst as determined by dual labeling with OCT-4 (red) as a marker of ICM cells. Partial z-stack at 600x magnification. At least 7 embryos per group were pooled from at least 5 different females. Scale bar = 10 μm.

Expression of 5T4 Antigen in Blastocyst Outgrowths

Blastocysts were cultured in vitro for 3 days until the trophectoderm layer of cells surrounding the blastocoel began to show morphological evidence of differentiation (flattened cells with enlarged nuclei, indicative of endoreduplication that is characteristic of these cells). The primitive endoderm-covered ICM has developed into a spheroid-like cell mass sitting on top of the sheet of adherent primary trophoblast cells (Fig. 2A). Outgrowths were positive for 5T4 antigen expression in both the ICM-derived cells (Fig. 2B ii) and trophoblast (Fig. 2B iii) cells (phase images shown in Fig. 2C). Expression appeared to be principally cytoplasmic, and since these cells were fixed and permeabilized in the same way as the hatched blastocysts, this pattern of labeling is unlikely to be due to the methodology used.

Figure 2.

5T4 antigen expression in blastocyst outgrowths in vitro. A: Inverted phase image at 200x magnification of blastocyst cultured on a fibronectin-coated chamber slide for 3 days. B: Outgrowths labeled by indirect immunofluorescence and visualized by confocal microscopy with (i) negative control rat IgG, (ii) mAb 9A7 (ICM-derived region of blastocyst outgrowth), (iii) mAb 9A7 (trophectoderm-derived region of blastocyst outgrowth), and (iv) positive control β1 integrin. C: Corresponding phase images. Images shown at 600x magnification and are representative of 10 outgrowths labeled.

The expression of β1 integrin (positive control) in blastocyst outgrowths was also principally cytoplasmic (Fig. 2Biv) whereas no detectable fluorescence with rat IgG (Fig. 2Bi) was observed.

5T4 Localization in ES Cells

To further analyze 5T4 expression in embryonic cells at a post blastocyst stage, in vitro maintained murine embryonic stem (mES) cells, derived from the epiblast, were investigated. mES cells were maintained in an undifferentiated state in a defined medium with leukemia inhibitory factor (LIF) or cultured in medium with fetal calf serum (FCS) with or without LIF. Figure 3Ai shows that cells grown in medium containing LIF supplemented with Knockout™ Serum Replacement (KOSR) did not express 5T4 antigen. However, when cells were cultured in medium containing LIF and FCS, cytoplasmic 5T4 expression was observed (Fig. 3Aii) and this localization can be maintained on passaging in the same culture medium. When LIF was withdrawn, antigen expression was primarily cell surface (Fig. 3Aiii), suggesting that mES cells have three distinct states with respect to 5T4 antigen expression and localization.

Figure 3.

5T4 can be retained in the cytoplasm of ES cells and colocalizes with calreticulin, an endoplasmic reticulum marker. A: 5T4 localization in 129 ES cells cultured in KO DMEM supplemented with (i) LIF and KOSR, (ii) LIF and ES cell tested FCS, (iii) FCS without LIF for 6 days. B: ES cells cultured in medium supplemented with LIF and FCS were labeled with rat 9A7, rabbit anti-calreticulin, and mouse anti-OCT4 and detected with goat anti-rat 633 AlexaFluor, goat anti-rabbit 546 AlexaFluor, and goat anti-mouse 488 AlexaFluor, respectively. 5T4 is blue, calreticulin is red, and OCT-4 is green. Laser scanning confocal images shown at 600x magnification. Scale bar = 10 μm.

The intracellular compartment appears to be the endoplasmic reticulum as judged by dual labeling with an antibody to calreticulin (Nakamura et al., 2001). Figure 3B shows ES cells with cytoplasmic expression of 5T4, colocalization of calreticulin, but not nuclear OCT-4.

5T4 Expression in Post-Implantation Embryos

Previous work has shown that there are significant levels of 5T4 transcripts in late postimplantation mouse embryos, as compared to adult tissues (King et al., 1999). In order to determine the cell types that express 5T4 oncofoetal antigen, cryostat embryo sections from E4.5 to E16.5 were labeled by indirect immunohistochemistry and visualized using di-aminobenzidine (DAB). No labeling was observed with rat IgG-negative control. Skin at E16.5 was found to label strongly positive and in all experiments this was used as an internal positive control.

5T4 Expression Is Restricted to Extra-Embryonic Tissues in Early Stage Postimplantation Embryos

Figure 4 shows expression of 5T4 expression at the point of implantation and immediately post-implantation. At E4.5, the uterine luminal epithelium strongly expressed 5T4 antigen (Fig. 4B) and this was a pregnancy-related change as non-pregnant uterine epithelium was negative for the antigen (Fig. 4A). Some cells of the flattened trophectoderm of the embryo were positive for 5T4 at E4.5 (not shown). At E5.5 and E6, polar trophectoderm expressed 5T4 (TE, Fig. 4C at E5.5), and at later stages, the ectoplacental cone that is derived from polar trophectoderm was also positive until E8.5, after which point, expression was not examined further in extra-embryonic tissues. Additionally at E5.5, extra-embryonic endoderm was 5T4-positive (arrows; Fig. 4C) and this was not seen after E6. The endodermal cell type was not determined by use of other markers; however, from the morphology, it is likely to be parietal endoderm that has collapsed onto the surface of the embryo. Expression was also detected in some trophoblast giant cells at these stages.

Figure 4.

5T4 expression is confined to extra-embryonic tissues immediately post-implantation. Sections through uterine luminal epithelium from (A) non-pregnant and (B) pregnant mice shown at 100x (i) and 600x (ii) magnification. C: E5.5 embryo within maternal stroma. TE, polar trophectoderm; arrows, endoderm; E, epiblast.

5T4 Expression in the Embryo

The first stage at which significant 5T4 expression could be detected in the embryo itself (outside the nervous system, discussed later), was at E12.5 in lateral and ventral but not dorsal areas of the periderm (Fig. 5A); labeling was relatively weak and punctate on the basal surface (arrows). Labeling of the skin persisted throughout development as the periderm develops from a single layered ectoderm to a multilayered epidermis at E14.5; labeling intensity increased and was predominately confined to the basal layer (arrows, Fig. 5B). The mature stratified epidermis forms at around E16.5, with less ordered outer differentiated cell layers, and here 5T4 expression remains primarily in the basal layer of cells (arrows, Fig. 5C). It must be noted that 5T4 expression in the skin was not uniform at any given stage. Skin receives inductive cues from the underlying mesenchyme that has arisen from different origins in different parts of the embryo. Hence, at E13.5, for example, ventral skin is further differentiated and different layers can be seen, whereas periderm still covers the dorsal surface of the embryo. Therefore, at E13.5, expression of 5T4 was strong and restricted to the basal layer ventrally, while dorsally, expression in the periderm was relatively weak and punctate.

Figure 5.

5T4 antigen expression is associated with many different epithelia in the developing murine embryo. Cryostat embryo sections were labeled with mAb 9A7 and an HRP-conjugated secondary antibody and DAB. Images shown are at 600x magnification. A: E12.5 lateral periderm. B: E14.5 layered epidermis. C: E16.5 stratified epidermis. D: E13.5 mesothelium. E: E13.5 bronchus. F: E13.5 oral epithelium. G: E14.5 tongue. H: E14.5 hair follicle (vibrissae). I: E15.5 stomach epithelium. J: E15.5 kidney glomerulus. K: E16.5 intestine. L: E16.5 tongue and palate. Arrows indicate examples of positive cells.

In other tissues, at E12.5, weak labeling was seen in epithelia of the gut, tongue, mesothelium, bronchi, and also in the condensing mesenchyme of the kidney primordium (not shown). At E13.5, epithelia of the gut, tongue, mesothelium (Fig, 5D), bronchus (Fig, 5E), esophagus, olfactory and oral cavities (Fig, 5F), as well as the vibrissae hair follicles in both the inner and outer root sheaths showed 5T4 expression. Expression was also detected at these sites at E14.5. In the lung, the bronchi were labeled but not all bronchioles. Similarly, in the kidney, expression could be detected in some but not all glomeruli, with epithelial labeling evenly distributed throughout. Additionally at this stage, expression was detected in the eyelid epithelium and in the condensing mesenchyme of the splenic primordium. Examples of expression at E14.5 are tongue epithelium (Fig. 5G) and a hair follicle expressing 5T4 antigen in both the inner and outer root sheath (Fig. 5H). 5T4 expression was seen in the outer root sheath only of hair follicles at a later stage of development. At E15.5, expression was still detectable in all the above-mentioned epithelial tissues. Stomach epithelial lining is shown in Figure 5I, and a representative positive kidney glomerulus in Figure 5J. Labeling was also detected in the some epithelial structures of the submandibular gland and in the pancreatic epithelium at this stage of development. E16.5 embryos showed a very similar pattern of labeling to E15.5 embryos. Intestinal epithelium is shown in Figure 5K; Figure 5L shows tongue (more differentiated than in Fig. 5G) and palate epithelium. Through to E16.5, high levels of 5T4 expression were increasingly confined to the basal layer of all stratified epithelia and also detected in most non-stratified epithelia.

5T4 Expression in the Mouse Embryo: Localization to Basal Undifferentiated Keratinocytes

From immunohistochemistry, 5T4 was apparently expressed at high levels by the basal layer of the epidermis during murine embryogenesis. In order to confirm expression by this layer of the developing skin, keratin 14 (K14) was used as a marker of basal undifferentiated keratinocytes. K14 is first expressed as periderm starts to form and as this develops into a multi-layered epidermis; expression becomes confined to the basal cell layer with a lag in detectable K14 expression dorsally due to underlying mesenchymal interactions during skin development (Byrne et al., 1994). In this study, K14 was detected in all developing stratified epithelia, such as the tongue, palate, submandibular gland, olfactory epithelial lining of the nasal cavity and nares, and the oesophagus (not shown). The mesothelium was also K14 positive. The timing of the expression pattern of K14 was very similar to that of 5T4. Dual immunofluorescence microscopy shows a vibrissae follicle expressing both K14 and 5T4 in the same cells at E15.5 (Fig. 6A), and the dorsal anterior periderm also positive for both antigens (Fig. 6B). It appears that K14-positive cells are also 5T4-positive. K14 is commonly expressed in mitotically active epithelia (Fuchs and Cleveland, 1998). Mitotic cells in different tissues of E15.5 embryos were identified by Ki67 (Endl and Gerdes, 2000) and their 5T4 status determined using 9A7 immunohistochemistry on the adjacent cryostat sections. Not all 5T4-positive cells are Ki67-positive but in all 5T4-expressing epithelia at least a proportion are actively dividing.

Figure 6.

Antibodies to 5T4 and K14 label identical cell populations in developing embryos. Cryostat E15.5 embryo sections were indirectly immunolabeled with mAb 9A7 (green) and α-K14 (red). A: Vibrissae hair follicle. B: Epidermis. Sections were photographed using an Olympus BX-51 fluorescent microscope at 600x magnification and images were overlaid using Adobe Photoshop 6.

5T4 Expression in the Developing Nervous System

5T4 transcripts and protein have been detected in adult mouse brain previously (King et al., 1999; Woods et al., 2002). To study 5T4 expression in relation to brain differentiation, horizontal serial sections of mouse brain were obtained at E14.5, E17.5, and P0 (birth), and compared to that of fully differentiated adult brain (∼3 months). To document the differentiation of neural tissue at these stages, other commonly used markers of various neural cell types were also examined. Nestin, an intermediate filament in neural stem and both early and late progenitor cells (Lendahl et al., 1990) was used to identify undifferentiated progenitor regions. Glial fibrillar acidic protein (GFAP) is the major intermediate filament in astrocytes (Eng et al., 1971), and details the process of astrogliogenesis. The neuronal nuclei (NeuN) marker is expressed by the majority of postmitotic neurons (Mullen et al., 1992) and was used to examine neurogenesis and neural cell proliferation was investigated using Ki-67.

At E14.5, significant 5T4 expression was detected in a stripe near the lateral ventricles in the region of the future stria terminalis (Fig. 7A and higher magnification of boxed region in Fig. 7B), midbrain roofplate (Fig. 7G), and in the cuboidal epithelial ependymal layer of the lateral and third ventricles and choroid plexus (Fig. 7H) of these ventricles. The lateral ventricles in the future stria terminalis region were also highly positive at E14.5 (Fig. 7A and B) and Figure 7B–F shows serial sections in this region labeled for expression of different markers at 400x magnification. This 5T4-positive region of the lateral ventricles consists mainly of undifferentiated progenitor cells (Malatesta et al., 2000) as evidenced by significant nestin expression (Fig. 7C). High levels of nestin expression were also detectable throughout the brain at this stage. NeuN-positive cells could be seen in much of the brain in non-proliferative areas such as the neopallial cortex, and around the ventricles away from the nuclear-dense SVZ (arrows; Fig. 7D), suggesting there may be some NeuN positive cells in the strongly 5T4-positive region of the lateral ventricle but the majority of neurons have migrated away from this region. Few GFAP-positive glial cells (Fig. 7E) were detectable at this stage, consistent with differentiation of progenitor cells to glial cells occurring later in development (Anderson, 2001). Those that were present were found in the 5T4-positive region of the lateral ventricle. Ki67-positive nuclei could be detected in the 5T4-positive zone (arrows; Fig. 7F), which suggests the active division of progenitor cells. Scattered Ki67-positive nuclei were also present in other regions, possibly representing the division of immature neurons during their migration. Some dividing cells were also labeled in the choroid plexus.

Figure 7.

5T4 antigen expression is associated with regions of progenitor cells in the developing brain. E14.5 horizontal brain sections labeled immunohistochemically with antibodies for various markers. A: 100x magnification of lateral ventricle region including choroid plexus (CPX) labeled with 9A7. B–F: The same region of lateral ventricle in serial sections labeled with different markers at 400x magnification. B: 5T4 positive boxed region of lateral ventricle shown in A at 400x magnification. C: Nestin. D: NeuN. E: GFAP. F: Ki67. G,H: Other 5T4-positive regions at E14.5. G: Roofplate. H: CPX of third ventricle. Arrows indicate examples of positive cells.

Expression at E17.5 was similar to P0 (Fig. 8). At P0, the roof plate and the ependymal cells and choroid plexus were still positive (not shown) and labeling round the lateral ventricles was still intense but restricted to a smaller region (Fig. 8A and higher magnification of boxed region shown in Fig. 8B). The ependymal cells and choroid plexus of the lateral ventricles (arrows, Fig. 8B) also expressed antigen. Nestin-positive filaments were still relatively widespread throughout the brain with increased expression around the ventricular regions consistent with a large progenitor cell population here (Fig. 8C). Lack of NeuN positivity in the lateral ventricle region is shown in Figure 8D, but significant NeuN expression was detected away from this region. Many more GFAP-positive astrocytes were detected primarily around the ventricles in the 5T4 positive region and this increased from E17.5 to P0 (Fig. 8E) with no gross change in localization, indicative of an increase in progenitor differentiation to this cell type. Neural cell proliferation is still active but at lower levels than at E14.5, although at P0, Ki67-positive nuclei (arrows, Fig. 8F) did not appear to be confined to any one region of the brain, rather a more scattered random distribution including some cells in the choroid plexus and in the 5T4-positive region of the lateral ventricle.

Figure 8.

5T4 antigen expression is decreased at P0, coincident with increased CNS differentiation. P0 horizontal brain sections labeled immunohistochemically with antibodies for various markers. A: 100x magnification of lateral ventricle region including CPX of P0 brain section labeled with 9A7. B–F: The same region of lateral ventricle in serial sections labeled with different markers at 400x magnification. B: 5T4 positive boxed region of lateral ventricle shown in A at 400x magnification. C: Nestin. D: NeuN. E: GFAP. F: Ki67. CPX: choroid plexus. Arrows indicate examples of positive cells.

In an adult mouse brain (3 months), neural cell proliferation, migration, and differentiation is effectively complete. Thus, there was no detectable nestin or Ki67 expression and, in contrast, GFAP and NeuN were expressed at very high levels in an almost reciprocal pattern, e.g., in the granule cell layer of the dentate gyrus (NeuN+ve) and the polymorph cell layer of the dentate gyrus (GFAP+ve). 5T4 expression was found to be restricted to the choroid plexus (not shown; Woods et al., 2002).


This study describes the expression pattern of murine 5T4 oncofoetal antigen during embryogenesis. The first time point at which 5T4 protein is expressed is at E4 upon hatching of the blastocyst from the zona pellucida, an event that occurs immediately prior to embryo implantation. Trophoblast is formed from differentiated trophectoderm and it is these cells of the hatched blastocyst that primarily express 5T4. Furthermore, during murine postimplantation development, 5T4 exhibits restricted expression to different epithelial cell types derived from all three germ layers, and in some regions of the brain. These results are consistent with the original isolation of h5T4 from human trophoblast (Hole and Stern, 1988) and detection of 5T4 protein and/or transcripts in murine placenta and embryos (King et al., 1999; Woods et al., 2002).

The first 5T4 expression is associated with early morphogenetic events in murine embryogenesis, the delineation of trophectoderm, and implantation. During implantation, the process of embryo attachment has been called a cell biological paradox in which two normally non-adherent epithelial surfaces become adherent for a limited period, the so-called “window of receptivity.” It is interesting that both the trophectoderm and the receptive epithelial lining of the uterine lumen, but not the non-pregnant uterus, express the 5T4 molecules. Overexpression of 5T4 has been shown previously to be associated with anti-adhesive properties and increased cell migration in vitro (Carsberg et al., 1996; Woods et al., 2002; Ward et al., 2003). At the time of implantation, it is possible that 5T4 expression may act to facilitate migration of the extra-embryonic cells away from the embryo into the uterine tissue, and/or guide invasion of trophectoderm cells through the luminal epithelial cell layer in a regulated fashion. Trophectoderm to trophoblast (Sutherland, 2003) and the generation of parietal endoderm (Veltmaat et al., 2002) have both been described as forms of epitheliomesenchymal transition (EMT), a process also associated with cancer. EMT is a fundamental process governing morphogenesis in multicellular organisms. This process is also reactivated in a variety of diseases including fibrosis and in the progression of carcinoma, and various studies of epithelial cells have identified changes in cellular polarity and adhesion as important in the acquisition of a variety of mesenchymal phenotypic traits (Thiery, 2003). A common role for 5T4 in EMT in development and malignant transformation is an interesting possibility.

Blastocyst culture to outgrowth can provide a model for early embryonic differentiation events, and here 5T4 protein was detected principally in the cytoplasm of primary trophoblast cells of such blastocyst outgrowths and also in the cells of the differentiating ICM. In contrast, 5T4 antigen expression could only be detected by immunohistochemistry in some giant trophoblast cells from E4.5 to E9.5, in the trophectoderm cell layer at E4.5, and in the parietal endoderm at E5.5. It is likely that the positive cells associated with the ICM of in vitro cultured embryos are endoderm rather than epiblast (negative by immunohistochemistry). Furthermore, upregulation of cell surface 5T4 expression is an early event in ES cell differentiation (Ward et al., 2003). The ES cell type is derived from epiblast cells (Brook and Gardner, 1997), which might predict that the antigen would be present in embryonic tissues prior to E12.5. This may reflect a lack of a tissue context for embryonic cells cultured in vitro compared to those in their natural environment. Alternatively, if 5T4 is expressed at low levels by a small subpopulation of cells at an early stage of development, expression could have gone undetected.

Studying expression of 5T4 antigen in ES cells showed that whilst undifferentiated cells are negative for protein expression, and definitively differentiating cells (Ward et al., 2003) have cell surface expression, ES cells can be maintained in an apparently intermediate state with cytoplasmic 5T4 localisation by culture in LIF and FCS. This may be significant in light of the recent findings (Yamamoto et al., 2005) where the FGF and the Frizzled receptors were shown to be retained in the endoplasmic reticulum during Xenopus embryogenesis by interaction with the novel protein, Shisa. It is possible that the timing of 5T4 localization to the cell membrane may also act in regulating of function during development by subcellular sequestration.

In postimplantation embryos, 5T4 antigen is next detected at E12.5. Expression becomes relatively widespread through development and is found at high levels in the basal layer of all stratified epithelia (e.g., skin, tongue, esophagus) and in many types of simple epithelia such as the gut and mesothelium and those organs undergoing branching morphogenesis (e.g., lung and kidney). In adult tissues, 5T4 expression is found at lower levels and is restricted to the basal layers of stratified epithelia, but in other positive epithelia the pattern is much more restricted (Woods et al., 2002; Southall et al., 1990; Barrow et al., unpublished data). For example, in the intestinal crypt, all cells of the undifferentiated embryonic epithelium express 5T4, but in the adult, expression is restricted to a subset of cells at the base of the crypt (Barrow et al., unpublished data). Tissue formation and its homeostasis, require epithelial cells that can reorganize their cytoskeleton in order to proliferate, migrate, and differentiate. 5T4 can induce cellular motility (Carsberg et al., 1996; Woods et al., 2002; Ward et al., 2003) through interactions with the actin cytoskeleton and, thus, may facilitate migration of such cells as they differentiate. A more detailed study of the natural history of 5T4-positive crypt cells is required to properly test this hypothesis.

5T4 expression during brain differentiation did not correlate exactly with any of the markers for neurons, astrocytes, or progenitor cell types at any time point examined. However, 5T4 may be associated with an intermediate cell type, i.e., a subtype or subtypes of committed progenitor cells. In the neuroepithelial SVZ, 5T4 is associated with regions of proliferating progenitor cells shown by high nestin and Ki67 expression, although these markers are also expressed in other regions. Furthermore, in the 5T4-positive SVZ, some cells expressed NeuN or GFAP and could represent migrating committed progenitor cell types, e.g., neuroblasts (NeuN-positive), or glioblasts (GFAP-positive). Alternatively, some GFAP-positive cells associated with the 5T4-positive region may be radial glia, cells known to give rise to both neurons and astrocytes and that also provide migratory guidance for neurons (Malatesta et al., 2000; Noctor et al., 2002). Downregulation of 5T4 when neural cells have ceased to migrate and are differentiated supports this putative association with migratory progenitor cells. In addition, epithelial ependymal cells of the ventricles and of the choroid plexus (both 5T4-positive) have also been shown to be progenitor cells from neurosphere formation in vitro with a lack of passaging ability (Chiasson et al., 1999).

5T4 overexpression in vitro results in altered cell morphology by abrogation of adherens junctions (AJs) through downregulation of E-cadherin and reorganization of the actin cytoskeleton and these effects may be mediated by the interaction of the h5T4 cytoplasmic tail with the PDZ-domain protein TIP-2/GIPC (Carsberg et al., 1996; Awan et al., 2002). From in vitro overexpression data, it might be expected that regions of high 5T4 expression would be regions of lowered E-cadherin expression, consistent with the decreased adhesion and motile phenotype and this is true in some 5T4-positive epithelia, e.g., hatched blastocysts. We also find decreased expression of E-cadherin in the embryonic basal epidermis and other stratified epithelia (unpublished data) and, in addition, basal epidermal cells with a high proliferative potential express lower levels of E-cadherin (Molès and Watt, 1997) with fewer AJs found in this cell layer (Ishiko et al., 2003). Therefore, 5T4 may affect cell motility through similar mechanisms to those seen in vitro. However, subtle alterations in levels of E-cadherin and 5T4 are difficult to detect using immunohistochemistry. Remodeling of epithelial adhesion systems is a general feature of organogenesis (Nanba et al., 2000, 2001; and others) so the presence of 5T4 in many developing epithelia and the putative negative relationship with E-cadherin indicates that 5T4 could have a role in this remodeling. Furthermore, the actin cytoskeleton has recently been shown to be important in epithelial development (Vasioukhin et al., 2000; Vaezi et al., 2002) and 5T4 may contribute to epithelial remodeling by interacting with the actin cytoskeleton through TIP-2.

Taken together, these data show that 5T4 oncofetal antigen is expressed by many different types of epithelia derived from all three germ layers, and one commonality between these regions of expression is an association with regions of undifferentiated progenitor cells and/or migration of differentiating cells. A transgenic mouse has now been created with the LacZ gene under the control of the 5T4 promoter. A similar analysis of tissue expression in the transgenic embryos will provide further information on the putative relationship of 5T4 antigen expression with progenitor cell types in the brain and other organs and their migratory ability. Initial studies suggest that the pattern of expression of LacZ is similar to that of the antigen.


Preimplantation Embryo Collection and Culture

BDF1 hybrid female mice were maintained in a 12L:12D cycle, with water and food provided ad libitum. Two female mice were paired overnight with a BDF1 male and checked the next morning for a vaginal plug. The day of plug observation was termed embryonic day (E) 0.5. Plugged female mice were removed to individual holding cages and were killed by cervical dislocation at specified time points. Oviducts and/or uterine horns were dissected and trimmed of excess fat and mesentery tissue and embryos of desired stage flushed out with M2 medium (HEPES buffered, containing 0.4% bovine serum albumin; BSA; Sigma, Dorset, UK). In some cases, preimplantation embryos were cultured in 40-μl droplets of M16 medium (0.4% BSA; Sigma) under embryo grade mineral oil (Sigma) in a humidified atmosphere in 5% CO2 in air at 37°C until the desired developmental stage was reached.

Indirect Immunofluorescence of Preimplantation Embryos

Embryos were pooled from at least 5 BDF1 females and at least 4 embryos of each stage were labeled with each antibody. All incubations were performed in M2 medium. Embryos were incubated in 0.5% pronase and observed under a dissecting microscope until zona pellucidae had been removed (∼10 min). Zona-free embryos were then washed through 20-μl droplets of M2 and then fixed for 20 min in 1% paraformaldehyde. Fixed embryos were washed 3 times in M2 and then blocked and permeabilized in 10% normal rabbit serum containing 0.05% Triton X-100 for 15 min. After blocking, embryos were washed once in M2, and incubated in 20-μl droplets of primary antibody for 1 hr at RT under mineral oil. Primary antibodies were used as follows: rat IgG negative control (eBiosciences, San Diego, CA) at 10 μg/ml; monoclonal rat IgG anti-mouse 5T4 (9A7) (Woods et al., 2002) at 10 μg/ml; monoclonal rat IgG anti-mouse uvomorulin (E-cadherin) positive control (Sigma) at 1:75 dilution. After incubation in primary antibody, embryos were washed 3 times in M2 and incubated in 20-μl droplets of rabbit anti-rat FITC conjugated secondary antibody (Dako, Carpinteria, CA) for 1 hr at RT under oil. Embryos were then washed in M2 and mounted in Vectashield mounting medium with DAPI (Vector Laboratories, Burlingame, CA) on a glass slide. Mounted embryos were left at 4°C overnight and the following day visualized by laser scanning confocal microscopy. Confocal settings were identical for all groups within an experiment, although settings between experiments sometimes varied.

Methodology used was verified by FACS analysis of a B16 mouse melanoma cell line stably transfected with mouse 5T4 and a control transfected cell line that were labeled in an identical fashion to the preimplantation embryos with the same antibody control and were found to be positive for 5T4 protein expression but not the control cell line (data not shown).

For dual labeling of in utero hatched blastocysts, embryos were blocked and permeabilized as above, and incubated in 10 μg/ml anti-5T4 and monoclonal mouse anti-mouse OCT4 (Santa Cruz Biotech, Santa Cruz, CA) at 1:100 for 1 hr at RT under oil. Embryos were washed as previously, and incubated in rabbit anti-rat FITC and rabbit anti-mouse TRITC conjugated secondary antibodies (Dako, 1:30) as previously. Embryos were then mounted and visualized as above.

Blastocyst Culture, Outgrowth, and Labeling

Blastocysts were flushed from the uterine horn at E3.5 and zona pellucidae removed using 0.5% pronase as previously. Blastocysts were cultured on fibronectin-coated chamber slides in DMEM with FCS for 3 days at 37°C in a humidified atmosphere containing 5% CO2. Outgrowths were then fixed in situ with 4% paraformaldehyde, blocked and permeabilized with 10% rabbit serum containing 0.05% Triton X-100 for 15 min and antibodies used as described below. Embryos were mounted in Vectashield with DAPI and visualized by laser scanning confocal microscopy.

ES Cell Culture and Differentiation

E14TG2a ES cells (American Type Culture Collection, Rockville, MD) were cultured on gelatin-treated tissue culture grade plates in Knockout™ DMEM (Invitrogen Life Technologies, Paisley, UK) containing Knockout™ Serum Replacement (KOSR) (Invitrogen) or 10% Hyclone tested FCS (Fisher Scientific, Pittsburgh, PA) and in the presence of LIF (ESGRO; Chemicon Int., Middlesex, UK). Cells were passaged every 2 days. For monolayer differentiation, LIF was omitted from the medium containing 10% FCS.

ES Cell Immunofluorescence

ES cells grown on gelatin-treated tissue culture grade plates were fixed in situ with 4% paraformaldehyde. For labeling, fixed cells were blocked and permeabilized by incubation in 10% normal rabbit serum and 0.05% Triton X-100 diluted in PBS for 20 min. Antibodies were used as described previously and rat anti-mouse β1 integrin (Chemicon Int.) was used at 15 μg/ml for 2 hr at RT. After incubation in primary antibody, cells were washed in PBS, and then incubated in fluorescein conjugated secondary antibody (FITC or TRITC) (Dako Cytomation, Ely, UK) at 1:30 dilution in PBS for 1 hr at RT in the dark. Cells were washed a further 3 times in PBS for 5 min each in the dark.

Alternatively, cells were blocked and permeabilized with 0.1% BSA, 0.1% Triton X-100, and 1% normal goat serum (Dako Cytomation) (blocking buffer) for 30 min at RT. Primary antibodies were used diluted in blocking buffer for 2 hr at RT at the following concentrations: 9A7 10 μg/ml; mouse mAb anti-mouse OCT4 (Santa Cruz Biotech) at 1:100; rabbit anti-calreticulin (Chemicon Int.) at 1:100. Cells were then washed 3 times in blocking buffer for 5 min each. AlexaFluor (488, 546, 633) conjugated secondary antibodies (Molecular Probes Inc., Eugene, OR; Invitrogen) were used at 1:500 in blocking buffer for 1 hr at RT in the dark. Cells were then washed as previously with blocking buffer in the dark.

After labeling by either method, cells were mounted in 2 drops of Vectashield containing DAPI, coverslipped and sealed with nail polish. Coverslipped “slides” were then excised from plates using a hot knife and visualized using a laser scanning confocal microscope. Confocal settings were identical for all groups within an experiment.

Immunohistochemistry of Postimplantation Embryo Cryostat Sections

For E4.5–6, tail veins of pregnant mice were injected with 1% Chicago Sky Blue (Sigma) in PBS 5 min before culling. Uteri were then removed and implantation sites as determined by blue labeling were excised. Embryos from E6.5–E9.5 had decidual swellings visible and these were identified and dissected; from E10.5–E16.5, the embryos were removed from the uterine horn. All embryos were washed in PBS and embedded in Tissue-Tek (Sakura Finetek U.S.A., Inc., Torrance, CA), and then frozen quickly in solid CO2 and isopentane. Eight-micrometer sections were cut at −20°C using a Reichert Jung 2800 Frigocut E (Leica Microsystems UK, Bucks., UK), fixed in ice-cold acetone for 5 min, air dried for 10 min, and then rehydrated in Tris buffered saline (TBS: 50 mM tris pH 7.6, 140 mM NaCl). Endogenous peroxidase activity was blocked by incubation in Peroxidase Blocking reagent (Dako) at RT for 10 min. Sections were then blocked with 10% normal rabbit or goat serum diluted in TBS for 20 min to prevent non-specific antibody binding. Primary antibodies were diluted in TBS and incubated for 2 hr at 37°C. Primary antibodies were used at concentrations as follows: mAb 9A7 and rat IgG at 10 μg/ml; polyclonal rabbit anti-mouse keratin 14 antibody (Covance Research Products, Denver, PA) at 1:1,000 dilution; monoclonal rat anti-mouse Ki-67 (Dako) at 1:25 dilution. HRP-conjugated secondary antibodies raised against the primary species (Dako) were used at 1:75 dilution for 1 hr at RT. Sections were washed after incubation with either antibody in TBS containing 0.1% Tween (TBST). The following antibodies were used in conjunction with a M.O.M. kit (Vector Laboratories); monoclonal mouse anti-nestin (Chemicon) at 10 μg/ml; monoclonal mouse anti-GFAP (Chemicon) at 1 μg/ml; monoclonal mouse anti-NeuN (Chemicon) at 1 μg/ml. Antibody labeling was visualized with di-aminobenzidine (DAB) and slides counter-stained, cleared, fixed, and mounted as described by Southall et al. (1990). Slides were viewed using an Olympus BX-51 microscope (Olympus, West Midlands, UK).

Immunofluorescence of Postimplantation Embryo Cryostat Sections

The method was as described above with the omission of the peroxidase-blocking step, and substitution of the HRP-conjugated secondary with either a FITC- or a TRITC-conjugated antibody raised against the primary species (diluted 1:30; Dako). For mAb 9A7 and K14 dual labeling, sections were incubated with primary antibodies together, further blocked with 10% normal rabbit serum in TBS followed by the secondary antibodies of swine anti-rabbit TRITC conjugate and rabbit anti-rat FITC conjugate. Sections were mounted in Vectashield mounting medium with DAPI and visualized using a fluorescent Olympus BX-51 microscope or a laser scanning confocal microscope.


We thank Garry Aston for aspects of histology expertise and Mark Willington and other staff of the Biological Resources Unit, PICR, for mouse husbandry and for teaching preimplantation embryo manipulation. This work was supported in part by a MRC studentship, University of Manchester and PICR (K.M.B.), Cancer Research UK (C.M.W., S.A., P.L.S.), and a Wellcome Trust Vacation Studentship (J.R.).