Plasminogen activators are believed to play an important role in tissue remodeling and cell migration. During mouse embryogenesis, visceral endoderm secretes urokinase-type plasminogen activator (uPA) whereas parietal endoderm secretes tissue-type plasminogen activator (tPA). Visceral endoderm from F9 embryoid bodies can transdifferentiate into parietal endoderm under the appropriate culture conditions. We have examined at the protein and mRNA levels the type of plasminogen activator expressed in whole embryoid bodies, visceral endoderm and its parietal endoderm derivatives. Our experiments show that the visceral endoderm on F9 embryoid bodies synthesizes and secretes substantial amounts of both tPA and uPA. In contrast, the parietal endoderm derived directly from the visceral endoderm secretes dramatically increased levels of tPA and decreases production of uPA to low or below detectable levels. These data support the finding that visceral endoderm can transdifferentiate to parietal endoderm. In addition, this transition provides an excellent model for studying the molecular basis of the coincident down- and upregulation of the two plasminogen activators as well as their potential function during embryogenesis.

Cell-cell and cell-matrix interactions influence the pathway of differentiation during mammalian embryogenesis, although the precise nature of these interactions and the mechanisms by which they mediate changes in the pattern of gene expression remain obscure. It has been suggested that plasminogen activators and proteases in general play an important role in the cell movements associated with morphogenesis during early development.

We are using the mouse F9 teratocarcinoma embryoid body system (Hogan et al. 1981) as a model to study the environmental distinctions that direct primitive endoderm precursors in the peri-implantation mouse embryo to differentiate into two separate endodermal lineages, parietal and visceral (Gardner, 1983). Parietal endoderm cells migrate along the inner surface of the trophectoderm and secrete the components of the Reichert’s membrane. Visceral endoderm cells form a tight columnar epithelium covering the inner cell mass. When F9 teratocarcinoma cells are treated with retinoic acid and grown in suspension they form embryoid bodies consisting of a layer of visceral endo derm surrounding a core of stem cells (Hogan et al. 1981). If the aggregates are treated with dibutyryl cyclic AMP (dbcAMP) in addition to retinoic acid or the cells are treated in monolayer with retinoic acid, they will differentiate into parietal endoderm (Strickland & Mahdavi, 1978; Strickland et al. 1980). We have demonstrated that the visceral endoderm from the F9 embryoid bodies can transdifferentiate directly into parietal endoderm if it is provided with an appropriate environment. In one instance, when whole embryoid bodies that have an outer layer of visceral endoderm are plated on a fibronectin containing substrate, parietal endoderm migrates away from the embryoid body onto the substrate (Grabel & Casanova, 1986; Grabel & Watts, 1986). Thus, providing a migration-promoting substrate can support the translocation and differentiation of parietal endoderm. This in vitro outgrowth system reflects the spatial segregation of endoderm types observed in the peri-implantation mouse embryo. In addition, the fact that fibronectin is present on the inner surface of the trophectoderm at the time that the parietal endoderm cells begin their migration (Wartio-vaara et al. 1979) further suggests that this model system is relevant to parietal endoderm migration and difieren-dation in vivo. Transdifferentiation is also initiated if visceral endoderm is proteolytically isolated from the embryoid bodies and plated on gelatin-coated dishes; the isolated visceral endoderm undergoes an apparent transition to parietal endoderm, based on morphology, the downregulation of α-fetoprotein (AFP) and an increase in the level of the parietal-specific cytokeratin Endo C (Casanova & Grabel, 1988).

In the mouse embryo, visceral and parietal endoderm are characterized by the synthesis and secretion of different forms of the extracellular protease plasminogen activator. Plasminogen activators cleave plasminogen into the active serine protease plasmin. Tissuetype plasminogen activator (tPA) and urokinase-type plasminogen activator (uPA) are functionally similar but arise from two distinct and differentially regulated genes. It has been proposed that these enzymes play an important role in invasive growth, cell migration and morphogenesis (reviewed by Dan0 et al. 1985), such as the involution of mammary glands (Ossowski et al. 1979), the invasive growth of Lewis lung carcinoma into normal tissue (Skriver et al. 1984) and the migration of capillary endothelial cells (Pepper et al. 1987). It is therefore not surprising that the first migratory cell type of the postimplantation mouse embryo, parietal endoderm, is characterized by high levels of tPA. In addition, it has been demonstrated that visceral endoderm makes uPA at the same stage (Marotti et al. 1982). It seemed, therefore, that the transition of visceral endoderm to parietal endoderm in vitro would provide an excellent system to study the control of expression of the two plasminogen activators as well as their potential function during embryogenesis.

We have examined the synthesis and secretion of plasminogen activators in retinoic-acid-treated F9 aggregates and during the transition of the visceral endoderm from F9 embryoid bodies into parietal endoderm. We report here that the visceral endoderm on the F9 embryoid bodies secretes both tPA and uPA. When embryoid bodies are plated onto collagen-coated dishes giving rise to parietal endoderm outgrowth, the embryoid bodies continue to secrete both tPA and uPA whereas the parietal outgrowth makes large amounts of tPA almost exclusively. When the visceral endoderm is isolated from the embryoid body and plated on gelatin-coated dishes, the cells increase tPA secretion and the level of tPA mRNA while dramatically decreasing their uPA secretion and the level of uPA mRNA. The value of using this model system to study the transition of a stationary cell to a migratory cell during embryogenesis and the potential role plasminogen activators may play in this process are discussed.

Cell culture and sample preparation

F9 cells were cultured and treated with 7·5×10−7M-retinoic acid and 0·5mm-dibutyryl cyclic AMP (dbcAMP; Sigma) as previously described (Hogan et al. 1981). Visceral endoderm was proteolytically isolated and plated onto gelatin-coated dishes as described (Casanova & Grabel, 1988). To obtain outgrowth cultures, embryoid bodies formed after seven days of culture in retinoic acid were plated onto tissue culture dishes coated with 0·1 mg ml-1 acid-soluble calf skin collagen as described (Grabel & Martin, 1983). After two to three days, the embryoid bodies were removed from the dish by pipetting vigorously. Most of the embryoid bodies detach and the outgrowth remains on the plate. Embryoid bodies that remained on the plate were picked off individually with a pulled Pasteur pipette under a dissecting microscope.

Conditioned medium was collected and cell extracts were prepared as described (Casanova & Grabel, 1988). All samples were spun in an Eppendorf microfuge for 3 min to rid the sample of cell debris. Protein concentrations in conditioned medium and cell extracts were determined by the method of Bradford (1976).

SDS-PAGE

Samples were electrophoresed through a 1·5 mm thick 10% polyacrylamide gel with a 4% stacking gel as described by Laemmli (1970). Samples were mixed with ×10 solubilizing buffer to a final concentration of 2·5 % SDS and 1 % sucrose. Bromophenol blue was added to the samples to serve as a tracking dye. Gels were run at 4°C at constant current until the dye front reached the bottom of the gel. Gels were removed and washed 3x20 min and 1 × 10 min in 2·5 % Triton X-100 to remove the SDS from the gel. Gels were washed again in Ca2+ and Mg2+-free phosphate-buffered saline (PBS) 3×5 min to remove the Triton X-100. Gels were then placed on a fibrin-agar underlay indicator gel.

Fibrin-agar underlay

Fibrin-agar underlays were prepared according to Granelli-Pipemo and Reich (1978). Underlays measured 11x14cm and contained a total volume of 25 · 6 ml and 200 μl penicillin/ streptomycin (Gibco, as supplied) to prevent microbial growth. The washed acrylamide gel was placed gently onto the fibrin-agar underlay, sealed in a plastic container and incubated at 37°C for 12−48 h. Underlay gels were occasionally viewed via dark-field illumination. When lytic zones were apparent against a cloudy background and had cleared to the desired extent, gels were photographed under dark-field illumination. Underlays were stained in I mg ml-1 Amido Black in 2% acetic acid and destained with 2% acetic acid.

To determine the extent of inhibition with antibody to human melanoma tPA or amiloride, these reagents were added to the underlay as described (Pepper et al. 1987; Vassalli & Belin, 1987). The gels were washed an additional 10min in PBS containing 10 μg ml-1 tPA antibody or 1·2mm-amiloride prior to being placed on the underlay, and the underlay contained either 10 μg ml-1 tPA antibody or 1·2 mm-amiloride accordingly.

RNA collection and analysis

2−20 ×106 cells were lysed and total nucleic acid was collected as described by Clark et al. (1977) and modified by Bernstein & Donady (1980). The pellet was resuspended in water and the RNA was isolated by LiCl precipitation as described (LeBlanc & Infante, 1988).

RNA was electrophoresed and blotted onto nitrocellulose as described (Maniatis et al. 1987). cDNA probes to tPA (Rickles et al. 1988) and uPA (Belin et al. 1985) (generous gifts of Dr Sidney Strickland) were radiolabeled with P32 by the random priming technique (Feinberg & Vogelstein, 1983). Filters were incubated with 2−4×107ctsmin-1 of labeled DNA at 42 °C for 12-36 h. The filters were washed with 2 × SSC, 0·1% SDS (once briefly, once for 15 min at room temperature, and once for 30min at 68°C), with 1 x SSC, 0·1% SDS (once for 30min at 68°C), and with 0·2 × SSC, 0×1 % SDS (once for 30min at 68°C). Filters were exposed to Kodak X-OMAT AR film.

Densitometry measurements of autoradiograms were performed and the tracings were integrated using a Zeineh Soft Laser Densitometer model SLR-2D/1D (Biomed Instruments Inc., Fullerton, CA).

Determination of plasminogen activator types and levels in F9 embryoid bodies

To determine the type of plasminogen activator produced by differentiating F9 embryoid bodies, conditioned medium and cell extracts were collected and analyzed by the fibrin-agar underlay technique. Figs 1A and 1B show a fibrin-agar analysis of conditioned medium (Fig. 1A) and cell extracts (Fig. 1B) isolated from F9 cells and embryoid bodies three to ten days (D3-D10) after addition of retinoic acid. All lytic zones were plasminogen dependent. Lane 1 represents material isolated from STO cells, a fibroblast cell line that we have shown makes relatively large amounts of uPA; lane 2 from PYS cells, a parietal yolk sac carcinoma cell line that makes tPA; and lane 3 from untreated F9 cells. D1 and D2 embryoid bodies demonstrated little or no detectable plasminogen activator activity (data not shown). In conditioned medium samples isolated from D3−D10 retinoic-acid-treated embryoid bodies (Fig. 1A, lanes 4−11), a plasminogen activator with a relative molecular mass of 75K is observed starting at D4 and slowly and steadily increasing to D10. This band comigrates with the tPA band from PYS cells and is close to the previously reported 79K relative molecular mass of murine tPA (Marotti et al. 1981). A second plasminogen activator with a relative molecular mass of 39K is apparent starting at D6 after induction and increases through D10. This band is close to the reported relative molecular mass of both uPA, 48K, and a minor species of tPA of 45K (Marotti et al. 1981). In the cell extracts (Fig. IB), a plasminogen activator of 75K is apparent beginning at D3 and steadily increases to D10. As tPA is predominantly a secreted enzyme it is surprising to find substantial activity in cell extracts, and this may be due to its association with extracellular matrix components. Unlike the conditioned medium, little or no 39K plasminogen activator activity was detected in the cell extracts. The differential temporal regulation of the two plasminogen activators suggests that they represent distinct gene products. In addition, they suggest that the higher relative molecular mass plasminogen activator is tPA and the lower relative molecular mass plasminogen activator is uPA and not a minor form of tPA. Furthermore, since the embryoid bodies develop a layer of visceral endoderm, it is expected that uPA would be present, and the timing of the appearance of uPA correlates with that of AFP (Casanova & Grabel, 1988). Since uPA is often found in association with the plasma membrane, the absence of activity in cell extracts suggests that either the uPA exists exclusively in its secreted form or the enzyme is present in the cell extracts as an inactive proenzyme. It is perhaps surprising that the embryoid bodies make tPA, although the cell type responsible for tPA production is unspecified. Therefore, an effort was made to identify the bands more precisely and to determine the cell type responsible for each band.

Fig. 1.

Fibrin-agar underlay analysis of D3-D10 conditioned medium and cell extracts. A is conditioned medium and B is cell extracts. Analysis was done on samples from STO, lane 1; PYS, lane 2; untreated F9 stem cells, lane 3; and embryoid bodies three to ten days after treatment with retinoic acid, lanes 4−11. Lanes 3−11 contain 1·5μg protein.

Fig. 1.

Fibrin-agar underlay analysis of D3-D10 conditioned medium and cell extracts. A is conditioned medium and B is cell extracts. Analysis was done on samples from STO, lane 1; PYS, lane 2; untreated F9 stem cells, lane 3; and embryoid bodies three to ten days after treatment with retinoic acid, lanes 4−11. Lanes 3−11 contain 1·5μg protein.

To verify the identity of the two bands, anti-tPA antibody or amiloride was added to the fibrin-agar underlay to inhibit the activity of tPA and uPA, respectively. Fig. 2 shows that when D7 embryoid body conditioned medium (lane 1) was analyzed on an underlay in the presence of 10 μg ml−1 tPA antibody (lane 2), the lytic zone of the higher relative molecular mass plasminogen activator was significantly inhibited. When the same sample was analyzed on an underlay containing l×2mm-amiloride (lane 3), the lytic zone of the lower relative molecular mass plasminogen activator was almost completely inhibited. The data show that the higher relative molecular mass lytic zone represents tPA activity and the lower relative molecular mass lytic zone represents uPA activity.

Fig. 2.

Inhibition analysis of plasminogen activator activity from D7 embryoid bodies with antibody to human melanoma tPA or amiloride. A and B are underlays before and after staining with amido black, respectively. Lane 1 contains conditioned medium analyzed on an underlay without an inhibitor, lane 2 with 10 μg ml−1 tPA antibody and lane 3 with l·2mm-amiloride. All lanes contain 1·5 μg protein and were incubated for 15 to 17 h.

Fig. 2.

Inhibition analysis of plasminogen activator activity from D7 embryoid bodies with antibody to human melanoma tPA or amiloride. A and B are underlays before and after staining with amido black, respectively. Lane 1 contains conditioned medium analyzed on an underlay without an inhibitor, lane 2 with 10 μg ml−1 tPA antibody and lane 3 with l·2mm-amiloride. All lanes contain 1·5 μg protein and were incubated for 15 to 17 h.

To further substantiate the finding that the embryoid bodies were making both tPA and uPA, mRNA levels of tPA and uPA were examined using Northern blot analysis. Total RNA isolated from STO, PYS, D7 dbcAMP-treated embryoid bodies, D7 embryoid bodies and F9 cells was probed with cDNA for mouse tPA (Fig. 3A) and mouse uPA (Fig. 3B). Figs 3A and 3B show that relatively low but definite levels of tPA and uPA mRNA were detected in the D7 embryoid bodies (lane 4). Relatively high levels of tPA message are present in PYS (lane 2) and dbcAMP treated embryoid bodies (lane 3) as would be expected. A lower level of tPA is present in the STO cells (lane 1) and retinoic acid treated embryoid bodies (lane 4) and no detectable tPA message is present in F9 cells (lane 5) prior to treatment with retinoic acid. Analysis for uPA message shows high levels in the STO cells (lane 1), low levels in the D7 embryoid bodies (lane 4) and no detectable levels in PYS cells (lane 2), dbcAMP treated embryoid bodies (lane 3), or F9 cells prior to treatment (lane 5).

Fig. 3.

Northern analysis for tPA and uPA. A represents a blot hybridized with tPA probe and B with uPA probe. Lane 1 contains total RNA from STO, lane 2 from PYS, lane 3 from dbcAMP treated D7 embryoid bodies, lane 4 from D7 embryoid bodies and lane 5 from untreated F9 stem cells. The tPA blot was exposed to film for 1·5 days and the uPA blot for 1 week. All lanes contain 20 μg total RNA. Reprobing blots with mammalian ribosomal cDNA probe confirmed that the lanes were loaded equally (data not shown).

Fig. 3.

Northern analysis for tPA and uPA. A represents a blot hybridized with tPA probe and B with uPA probe. Lane 1 contains total RNA from STO, lane 2 from PYS, lane 3 from dbcAMP treated D7 embryoid bodies, lane 4 from D7 embryoid bodies and lane 5 from untreated F9 stem cells. The tPA blot was exposed to film for 1·5 days and the uPA blot for 1 week. All lanes contain 20 μg total RNA. Reprobing blots with mammalian ribosomal cDNA probe confirmed that the lanes were loaded equally (data not shown).

As the embryoid bodies consist of two apparent cell types, stem cells and endoderm, it becomes critical to determine which cell type is responsible for the synthesis of tPA and uPA. Experiments that address this problem are described below.

Plasminogen activator production in outgrowth cultures from F9 embryoid bodies

When F9 embryoid bodies are plated onto collagen-coated dishes, endoderm migrates away from the embryoid body onto the substrate. The endoderm was identified as parietal based on morphology and the absence of AFP (Grabel & Casanova, 1986). Plasminogen activator production of the plated embryoid bodies and the outgrowth was examined. Fig. 4 shows a fibrin-agar underlay analysis of the outgrowth culture. Analysis of total secreted protein shows that while the resuspended embryoid bodies continue to make levels of tPA and uPA with approximately equal activity to that observed prior to plating (lane 4), the parietal outgrowth makes primarily tPA (lane 5) in large amounts. Lanes 1 and 2 represent STO and PYS samples, lane 3 D7 embryoid bodies before plating and lane 6 the whole outgrowth culture. Thus during the transition from visceral endoderm to parietal endoderm, elicited by a migration-promoting substrate, the visceral endoderm transforms to parietal endoderm as characterized by its tPA production.

Fig. 4.

Fibrin-agar underlay analysis of conditioned medium from D7 embryoid bodies and outgrowth cultures. Lane 1 contains conditioned medium isolated from STO, lane 2 PYS, lane 3 D7 embryoid bodies, lane 4 embryoid bodies isolated from an outgrowth culture, lane 5 the outgrown parietal endoderm after the embryoid bodies were removed, and lane 6 the whole outgrowth culture containing the plated embryoid bodies and the parietal outgrowth. Lanes 3−6 contain 1−5μg protein.

Fig. 4.

Fibrin-agar underlay analysis of conditioned medium from D7 embryoid bodies and outgrowth cultures. Lane 1 contains conditioned medium isolated from STO, lane 2 PYS, lane 3 D7 embryoid bodies, lane 4 embryoid bodies isolated from an outgrowth culture, lane 5 the outgrown parietal endoderm after the embryoid bodies were removed, and lane 6 the whole outgrowth culture containing the plated embryoid bodies and the parietal outgrowth. Lanes 3−6 contain 1−5μg protein.

Plasminogen activator production in isolated visceral endoderm and during the transdifferentiation from visceral endoderm to parietal endoderm

Using a number of morphological and biochemical markers, we have definitively identified the endodermal cells isolated from retinoic-acid-treated embryoid bodies as visceral endoderm. We have also determined that, following monolayer culture, the cells become parietal-like based on their morphology, the downregulation of AFP, and an increase in the parietal-specific cytokeratin Endo C (Casanova & Grabel, 1988). Initial plasminogen activator studies showed a 7-fold increase in total plasminogen activator activity and an increase in activity that was inhibitable by tPA antibody (Casanova & Grabel, 1988). The regulation of expression of both tPA and uPA during the phenotypic transition from visceral to parietal endoderm was examined here in more detail. Visceral endoderm was isolated from D7 embryoid bodies and plated on gelatin-coated dishes for varying lengths of time, and conditioned medium was analyzed on a fibrin-agar underlay. Fig. 5 shows that by 24 h after the harvested endoderm is plated, there is a dramatic increase in tPA activity consistent with our previous report. Over the same time, uPA activity has decreased to low levels. Thus the transition of visceral endoderm to parietal endoderm is accompanied by a significant increase in tPA production and the downregulation of uPA. Due to the use of proteolytic enzymes in the visceral endoderm isolation procedure, and the rapid transition of the cells into parietal endoderm, it is impossible to obtain conditioned medium directly from freshly harvested visceral endoderm, and the cell extracts contain no uPA activity. Therefore, in order to determine if these cells synthesize both types of plasminogen activators, we examined message levels for the two genes.

Fig. 5.

Fibrin-agar underlay analysis of conditioned medium from D7 embryoid bodies and isolated endoderm 24 and 48 h after plating. Lane 1 contains conditioned medium isolated from STO, lane 2 PYS, lane 3 D7 embryoid bodies, lane 4 endoderm after 24 h in monolayer and lane 5 endoderm after 48 h in monolayer. Lytic zones represent tPA and uPA activity as indicated. Lanes 3−5 contain 1·5 ug protein.

Fig. 5.

Fibrin-agar underlay analysis of conditioned medium from D7 embryoid bodies and isolated endoderm 24 and 48 h after plating. Lane 1 contains conditioned medium isolated from STO, lane 2 PYS, lane 3 D7 embryoid bodies, lane 4 endoderm after 24 h in monolayer and lane 5 endoderm after 48 h in monolayer. Lytic zones represent tPA and uPA activity as indicated. Lanes 3−5 contain 1·5 ug protein.

Northern analysis of D7, D8 and D9 embryoid bodies, freshly harvested endoderm, and endoderm 24, 48 and 72 h after plating show a similar trend in the message levels for tPA and uPA as that seen for the protein. Figs 6A and 6B show that within 24 h after plating tPA message levels increase (6A) and uPA message levels decrease sharply (6B). Densitometry measurements show that the tPA mRNA level has increased 3·1-fold over the freshly harvested cells after 24h, 3·6-fold after 48h, and 4·0-fold after 72h. It should be noted that the harvested endoderm contains both tPA and uPA mRNA suggesting visceral endoderm can produce both plasminogen activator types. In addition, the enrichment for both messages seen in freshly harvested endoderm (lane 7) versus the total D7 embryoid body (lane 4) suggests that the visceral endoderm is certainly the primary source and may be the exclusive source for embryoid body plasminogen activator activity. The observed downregulation of uPA levels parallels the previously reported decrease in AFP protein and message levels (Casanova & Grabel, 1988). The increase in the parietal tPA message and the decrease of the visceral uPA message strongly support the idea that visceral endoderm from F9 embryoid bodies transdifferentiates into parietal endoderm when plated in monolayer.

Fig. 6.

Northern analysis of embryoid bodies and plated isolated endoderm. A represents a blot probed with tPA cDNA and B with tPA cDNA. Lane 1 contains STO RNA, lane 2 PYS, lane 3 untreated F9 stem cells, lane 4 D7 embryoid bodies, lane 5 D8 embryoid bodies, lane 6 D9 embryoid bodies, lane 7 freshly isolated visceral endoderm, lane 8 isolated endoderm after 24 h in monolayer, lane 9 isolated endoderm after 48 h in monolayer and lane 10 isolated endoderm after 72 h in monolayer. A was exposed to film for 1 day and B for 1 week. All lanes contain 20 μg total RNA, and reprobing the blots with mammalian ribosomal cDNA confirmed that the lanes were loaded equally (data not shown).

Fig. 6.

Northern analysis of embryoid bodies and plated isolated endoderm. A represents a blot probed with tPA cDNA and B with tPA cDNA. Lane 1 contains STO RNA, lane 2 PYS, lane 3 untreated F9 stem cells, lane 4 D7 embryoid bodies, lane 5 D8 embryoid bodies, lane 6 D9 embryoid bodies, lane 7 freshly isolated visceral endoderm, lane 8 isolated endoderm after 24 h in monolayer, lane 9 isolated endoderm after 48 h in monolayer and lane 10 isolated endoderm after 72 h in monolayer. A was exposed to film for 1 day and B for 1 week. All lanes contain 20 μg total RNA, and reprobing the blots with mammalian ribosomal cDNA confirmed that the lanes were loaded equally (data not shown).

Visceral endoderm from F9 embryoid bodies synthesizes and secretes both tPA and uPA, and when the visceral endoderm undergoes a transition to parietal endoderm, the cells dramatically upregulate their production of tPA and downregulate their production of uPA at the protein and mRNA levels. The transition from visceral endoderm to parietal endoderm is induced by plating whole embryoid bodies on a migration-promoting substrate so that the parietal cells undergoing differentiation can migrate away from the embryoid body or by growing the isolated visceral endoderm on a flat surface. The end result in both cases is similar: the cells lose close contact with the undifferentiated core cells and with each other and grow on a flat surface rather than a convex one. The shift by the differentiating cells to the synthesis of high levels of primarily tPA supports the evidence demonstrating that visceral endoderm can transdifferentiate directly into parietal endoderm. Both tPA and uPA appear to be made by the visceral endoderm layer and not the stem cells since the mRNA for both tPA and uPA is present in the isolated endoderm cells. The presence of substantial amounts of tPA in the visceral endoderm is an interesting distinction between these cells and visceral endoderm in vivo, and may be the result of treatment with retinoic acid or other in vitro conditions. Alternatively, the ability to synthesize tPA and the ease with which these cells differentiate into parietal endoderm may indicate that they are a form of pluripotential endoderm cell. In addition, we can not eliminate the possibility that a small subpopulation of primitive endoderm cells is present in the embryoid bodies and harvested visceral endoderm preparations. As the primitive endoderm cells are generally characterized by parietal-endoderm-specific markers they could be responsible for the tPA activity and tPA mRNA ob served. Our previous observations that the harvested visceral endoderm contains barely detectable levels of the parietal endoderm marker Endo C argues against the presence of such a cell population (Casanova & Grabel, 1988).

It is intriguing that the visceral endoderm cells undergo such an elaborate alteration in plasminogen activator gene expression given the observation that tPA and uPA ultimately demonstrate the same enzyme specificity. In this light, it is interesting that different species sometimes use different plasminogen activator types in the same tissue to achieve the same function. For example, Canipari et al. (1987) have shown that mouse ovarian granulosa cells produce uPA to weaken the follicle wall at the time of ovulation whereas rat ovarian granulosa cells secrete tPA for the same function. Human milk contains primarily tPA (Marshall et al. 1986) whereas bovine milk contains predominately uPA (Navaeetha Rao, personal communications). The enzymes appear to have interchangeable functions in these cases. The simultaneous increase of tPA and decrease of uPA activity during the transdifferentiation of visceral endoderm to parietal endoderm suggests that the two enzymes may have different specificities for an as yet unidentified substrate. Alternatively, the differential gene expression may reflect the fact that the two enzymes can be regulated by alternate mechanisms. For example, uPA can be present in an active or inactive form and is often membrane bound (reviewed by Blasi et al. 1987). Analyses designed to identify specific inhibitors or receptors for tPA or uPA in embryoid body derived materials should clarify this issue.

We are particularly interested in using the in vitro outgrowth and subsequent differentiation of parietal endoderm from embryoid bodies to evaluate the purported role of tPA in parietal endoderm migration (Strickland et al. 1978). Plasminogen activators have been reported to be involved in the migration of capillary endothelial cells in vitro (Pepper et al. 1987), human tumor metastasis (Ossowski & Reich, 1983), and several other migratory cell types. The increase in tPA levels correlates with the migration of parietal endoderm on fibronectin substrates and previous data have suggested that the cells can clear the fibronectin from the dish, perhaps via proteolysis, as they migrate over it (Grabel & Watts, 1986). It has also been demonstrated that proteases can digest fibronectin (Chen & Chen, 1987; Fairbairn et al. 1985; Chen et al. 1985) and tPA may demonstrate such activity in a plasminogen dependent or independent manner (Quigley et al. 1987). Thus, this system may also prove useful in identifying a biological role for plasminogen activators during embryogenesis.

We would like to thank Dr Sidney Strickland for his generous gifts of the tissue-type plasminogen activator and urokinase cDNA probes and Sandy Becker for essential technical assistance. This work was supported by grant #CD-314 from the American Cancer Society. L. Grabel is the recipient of Research Career Development Award CA 1065 from the National Institute of Health.

Beers
,
W. H.
,
Strickland
,
S.
&
Reich
,
E.
(
1975
).
Ovarian plasminogen activator: Relationship to ovulation and hormonal regulation
.
Cell
6
,
387
394
.
Belin
,
D.
,
Vassalli
,
J-D.
,
Combepine
,
C.
,
Godeau
,
F.
,
Nagamine
,
Y.
,
Reich
,
E.
,
Kocher
,
H. P.
&
Duvoisin
,
R. M.
(
1985
)
Cloning, nucleotide sequencing and expression of cDNAs encoding mouse urokinase-type plasminogen activator
.
Eur. J. Biochem.
148
,
225
232
.
Bernstein
,
S. I.
&
Donady
,
J. D.
(
1980
).
RNA synthesis and coding capacity of polyadenylated and nonpolyadenylated mRNA from cultures of differentiating Drosophola Melanogaster myoblasts
.
Devi Biol.
79
,
388
398
.
Blasi
,
F.
,
Vassalli
,
J.D.
& DAN0, K
. (
1987
).
Urokinase-type plasminogen activator: Proenzyme, receptor, and inhibitors
.
J. Cell Biol.
104
,
801
804
.
Bradford
,
M. M.
(
1976
).
A rapid and sensitive method for quantitation of microgram quantities of protein utilizing the principle of protein-dye binding
.
Analyt. Biochem.
102
,
196
202
.
Casanova
,
J. E.
&
Grabel
,
L. B.
(
1988
).
The Role of Cell Interactions in the Differentiation of Teratocarcinoma-Derived Parietal and Visceral Endoderm
.
Devi Biol.
129
,
124
139
.
Chen
,
J.M.
&
Chen
,
W-T.
(
1987
).
Fibronectin-degrading proteases from the membranes of transformed cells
.
Cell
48
,
193
203
.
Chen
,
W-T.
,
Chen
,
J-M.
,
Parsons
,
S. J.
&
Parsons
,
J. T.
(
1985
).
Local degradation of fibronectin at sites of expression of the transforming gene product pp60src
.
Nature, Lond.
316
,
156
158
.
Clark
,
S. H.
,
Strausbaugh
,
L. D.
&
Kiefer
,
B. I.
(
1977
).
Genetic modulation of RNA metabolism in Drosophila. I. Increased rate of ribosomal RNA synthesis
.
Genetics
86
,
789
800
.
DAN0
,
K.
,
Andreasen
,
P. A.
,
Grondahl-Hansen
,
J.
,
Kristensen
,
P.
,
Nielsen
,
L. S.
&
Skriver
,
L.
(
1985
).
Plasminogen activators, tissue degradation and cancer
.
Adv. Cancer Res.
44
,
139
266
.
Fairbairn
,
S.
,
Gilbert
,
R.
,
Ojakian
,
G.
,
Schwimmer
,
R.
&
Quigley
,
J. P.
(
1985
).
The extracellular matrix of normal chick embryo fibroblasts: Its effect on transformed chick fibroblasts and its proteolytic degradation by the transformants
.
J. Cell Biol.
101
,
1790
1798
.
Feinberg
,
A. P.
&
Vogelstein
,
B.
(
1983
).
A technique for radiolabeling DNA restriction endonuclease fragments to high specific activity
.
Analyt. Biochem.
132
,
6
13
.
Gardner
,
G. L.
(
1983
).
Origin and differentiation of extraembryonic tissue in the mouse
. In
International Review of Experimental Pathology
(ed.
G. W.
Richter
&
M. A.
Epstein
),
Academic
Press
,
New
York
, pp. 63-133
.
Grabel
,
L. B.
&
Casanova
,
J. E.
(
1986
).
The outgrowth of parietal endoderm from mouse teratocarcinoma stem-cell embryoid bodies
.
Diff.
32
,
67
73
.
Grabel
,
L. B.
&
Martin
,
G. R.
(
1983
).
Tunicamycin reversibly inhibits the terminal differentiation of teratocarcinoma stem cells to endoderm
.
Devi Biol.
95
,
115
125
.
Grabel
,
L. B.
&
Watts
,
T. D.
(
1987
).
The role of extracellular matrix in the migration and differentiation of parietal endoderm from teratocarcinoma embryoid bodies
.
J. Cell Biol.
105
,
441
447
.
Granelli-Piperno
,
A.
&
Reich
,
E.
(
1978
).
A study of proteases and protease inhibitor complexes in biological fluids
.
J. exp. Med.
148
,
223
234
.
Hogan
,
B. L. M.
,
Taylor
,
A.
&
Adamson
,
E.
(
1981
).
Cell interactions modulate embryonic carcinoma cell differentiation into parietal or visceral endoderm
.
Nature, Lond.
291
,
235
237
.
Laemmli
,
U. K.
(
1970
).
Cleavage of structural proteins during the assembly of the head of bacteriophage T4
.
Nature, Lond.
227
,
680
685
.
Leblanc
,
J.
&
Infante
,
A. A.
(
1988
).
Association of 7SL RNA and an SRP-like particle with polysomes and endoplasmic reticulum in the developing sea urchin embryo
.
Devi Biol, (in press)
.
Maniatis
,
T.
,
Fritsch
,
E. F.
&
Sambrook
,
J.
(
1987
).
Molecular Cloning, A Laboratory Manual.
Cold Spring Harbor
,
New York
, pp. 202-203,
382
386
.
Marotti
,
K. R.
,
Belin
,
D.
&
Strickland
,
S.
(
1981
).
The Production of Distinct Forms of plasminogen activator by Mouse Embryonic Cells
.
Devi Biol.
90
,
154
159
.
Marshall
,
J. M.
,
Rees
,
M. C. P.
&
Cederholm-Williams
,
S. A.
(
1986
).
Identification of t-PA as the major active plasminogen activator in human milk
.
Thrombosis and Haemostasis
55
,
279
281
.
Ossowski
,
L.
,
Biegel
,
O.
&
Reich
,
E.
(
1979
).
Mammary plasminogen activator, Correlation with involution, hormonal modulation and comparison between normal and neoplastic tissue
.
Cell
16
,
929
940
.
Ossowski
,
L.
&
Reich
,
E.
(
1983
).
Antibodies to plasminogen activator inhibit human tumor metastasis
.
Cell
35
,
611
619
.
Pepper
,
M. S.
,
Vassalli
,
J-D.
,
Montesano
,
R.
&
Orci
,
L.
(
1987
).
Urokinase-type Plasminogen Activator Is Induced in Migrating Capillary Endothelial Cells
.
J. Cell Biol.
105
,
2535
2541
.
Quigley
,
J. P.
,
Gold
,
L. L
,
Schwimmer
,
R.
&
Sullivan
,
L. M.
(
1987
).
Limited cleavage of cellular fibronectin by plasminogen activator purified from transformed cells
.
Proc. natn. Acad. Sci., U.S.A.
84
,
2776
2780
.
Rickles
,
R. J.
,
Darrow
,
A. L.
&
Strickland
,
D.
(
1988
).
Molecular cloning of complementary DNA to mouse tissue plasminogen activator mRNA and its expression during F9 teratocarcinoma cell differentiation
.
J. biol. Chem.
263
,
1563
1569
.
Strickland
,
S.
&
Mahdavi
,
V.
(
1978
).
The induction of differentiation in teratocarcinoma stem cells by retinoic acid
.
Cell
15
,
393
403
.
Strickland
,
S.
,
Smith
,
K.
&
Marotti
,
K. R.
(
1980
).
Hormonal induction of differentiation in teratocarcinoma stem cells: Generation of parietal endoderm by retinoic acid and dibutyryl cAMP
.
Cell
21
,
347
355
.
Vassalli
,
J.D.
&
Belin
,
D.
(
1987
).
Amiloride selectively inhibits the urokinase-type plasminogen activator
.
FEBS Lett.
214
,
187
191
.
Wartiovaara
,
J.
,
Leivo
,
I.
&
Vaheri
,
A.
(
1979
).
Expression of the cell surface-associated glycoprotein, fibronectin, in the early mouse embryo
.
Devi Biol.
69
,
247
257
.