A P19 embryonal carcinoma stem cell line carrying an insertion of the E. coli LacZ gene in an endogenous copy of the Pax-3 gene was identified. Expression of the Pax-3/LacZ fusion gene in neuroectodermal and mesodermal lineages following induction of differentiation by chemical treatments (retinoic acid and dimethylsulfoxide) was characterized using this line and is consistent with the previous localization of Pax-3 expression in the embryo to mitotically active cells of the dorsal neuroectoderm and the adjacent segmented dermomyotome. Pax-3/LacZ marked stem cells were also utilized as target cells in mixing experiments with unmarked P19 cells that had been differentiated by pretreatment with chemical inducers. Induction of -galactosidase and neuroectodermal markers in the target cells demonstrates that: (1) some differentiated P19 cell derivatives transiently express endogenous Pax-3- and neuroectoderm-inducing activities, (2) undifferentiated target stem cells respond to these activities even in the presence of leukemia inhibitory factor and (3) the endogenous activities can be distinguished from, and are more potent than, retinoic acid treatment in inducing neuroectoderm. These observations demonstrate that P19 embryonal carcinoma cells provide a useful in vitro system for analysis of the cellular interactions responsible for neuroectoderm induction in mammals.

In vertebrates, neuroectoderm formation occurs in regions of the primitive ectoderm that are located in close proximity to the underlying invaginated mesoderm (reviewed in Beddington, 1986; Hamburger, 1988). In amphibian embryos, the ability to reconstitute neuroectoderm induction in vitro has allowed the direct demonstration of an essential role of mesoderm in providing signals that redirect primitive ectoderm towards a neuroectodermal fate (reviewed in Nieuwkoop et al., 1985). Such reconstitution experiments are possible in amphibians since ectodermal and endodermal lineages are determined early in development as a consequence of localized determinants and mesoderm is then induced from primitive ectoderm due to interactions of ectoderm with endoderm (reviewed in Smith, 1989; Smith et al., 1991; Cho et al., 1991). Isolated primitive ectoderm, cultured in the absence of endoderm, fails to differentiate either mesodermal or neuroectodermal lineages but, rather, differentiates towards epiderm (e.g. Akers et al., 1986). Addition of mesoderm to primitive ectoderm cultures provides the signals necessary for induction of neuroectodermal lineages from the primitive ectoderm when assayed both morphologically and by the induction of neuroectoderm-specific markers (e.g. Nieuwkoop et al., 1985; Gurdon, 1987; Sive et al., 1989; Dixon and Kinter, 1989; Jones and Woodland, 1989; Ruiz i Altaba, 1990; Mitani and Okamoto, 1991; Servetnick and Grainger, 1991).

Although it is likely that induction of neuroectoderm in mammalian embryos also depends on signals from the underlying mesoderm, this induction has not been demonstrated directly. Similar to amphibian embryos, mesoderm in mammals forms initially by differentiation of primitive ectoderm at one side of the egg cylinder in a region called the primitive streak. This mesoderm then migrates between the remaining primitive ectoderm and extraembryonic endoderm. Neuroectoderm induction occurs in the region of the primitive ectoderm that is adjacent to the invaginated mesoderm (reviewed in Beddington, 1986). However, mammalian embryos differ from amphibian embryos in that ectodermal and endodermal cell lineages are not determined at the onset of embryogenesis. Rather, the primitive ectoderm is the sole founder tissue of all ectodermal, mesodermal and endodermal lineages of the fetus and most regions of the primitive ectoderm are pluripotent (reviewed in Beddington, 1983a; 1986). Further, primitive ectoderm will give rise to endodermal and mesodermal lineages spontaneously when cultured in vitro (Beddington, 1986; Robertson, 1987). The inability to isolate primitive ectoderm in the absence of mesodermal lineages prevents the use of in vitro reconstitution experiments to demonstrate directly the induction of neuroectoderm by mesoderm.

Embryonal carcinoma (EC) stem cells arise from primitive ectoderm at 7 or 8 days of gestation (Diwan and Stevens, 1976; Beddington, 1983b). Many of these cell lines remain pluripotent and will differentiate to form neuroectodermal, mesodermal and extraembryonic endodermal cell lineages in vitro either spontaneously or following chemical induction. The differentiation of P19 EC cells (McBurney and Rogers, 1982) has been extensively characterized. These cells have a very low rate of spontaneous differentiation when grown in monolayer cultures in the absence of chemical inducers but can be induced to differentiate efficiently towards different cell lineages when allowed to aggregate in the presence of specific inducers. Predominantly neuroectodermal lineages, including neurons and astroglial cells, are formed when aggregates are grown in the presence of high doses of RA (1×10−6 M under the conditions used here; McBurney et al., 1982; Jones-Villeneuve et al., 1982, 1983) and predominantly mesodermal lineages (including cardiac and skeletal muscle) are formed when aggregates are grown in the presence of DMSO (0.5 - 1.0%) or lower doses of RA (McBurney et al., 1982; Edwards et al., 1983; Smith et al., 1987). Aggregation in the absence of chemical inducers results in predominantly endodermal lineages (McBurney et al., 1982). The cytokine leukemia inhibitory factor (LIF) inhibits differentiation of P19 EC cells towards endodermal or mesodermal lineages under all of these conditions but does not inhibit differentiation of neuroectodermal lineages following treatment of aggregates with high doses of RA (Pruitt and Natoli, 1992).

In the present study, a P19 EC stem cell line (O1A1pT1IIEH3, referred to as EH3 cells) which is marked by the integration of the E. coli LacZ gene into the endogenous Pax-3 gene was isolated. This cell line was used in cell mixing experiments, in the presence of LIF, to demonstrate the transient expression of Pax-3- and neuroectoderm-inducing activities during differentiation of unmarked P19 cells.

Cell culture and induction of differentiation

The cell line P19S18O1A1 (McBurney et al., 1982) or the derivative O1A1pT1IIEH3 (EH3 cells, described below) were used in all of the studies reported here and maintained as described previously (Pruitt and Natoli, 1992). RA, DMSO and LIF treatments were also performed using reagents as described previously (Pruitt and Natoli, 1992).

Isolation of O1A1pT1IIEH3 cells

P19S18O1A1 cells were transfected using calcium-phosphatemediated transfection (Wigler et al., 1979) with the gene trap vector pT1 (Gossler et al., 1989). Transfectants were selected by growth in the presence of 400 μg/ml G418 (Sigma) for 12 days. Approximately 2,000 resistant colonies were obtained in each of five culture dishes (15 cm diameter). Each dish of transfectants was trypsinized, and cells were counted and plated to tissue-culture-grade microtitre dishes at concentrations that resulted in approximately 10 colonies per microtitre well. Colonies were amplified to approximately 100 –500 cells each, trypsinized and replated to microtitre dishes under three conditions: monolayer in the absence of inducer, aggregates in the presence of 1.0% DMSO and aggregates in the presence of 1×10 −6 M RA. For aggregate cultures, uncoated microtitre dishes were used. Approximately 5,000 independent transfectants were assayed. DMSO- and RA-treated cultures were fixed and stained for β-galactosidase activity as described below following 3 days of induction. Wells containing cells that expressed β-galactosidase under one induction condition but not the other were identified and cells exhibiting the expression were isolated by additional sib-selection. The EH3 cell line was one of several that exhibited β-galactosidase activity following RA but not DMSO induction and was characterized further as described below and in the Results section.

RNA extraction, cDNA synthesis and inverse PCR

RNA was isolated from EH3 cells that had been induced to differentiate as aggregates by 1×10−6 M RA treatment on day 6 following initiation of the experiment using acid guanidinium thiocyanate-phenol-chloroform extraction (Chomczynski and Sacchi, 1987; Puissant and Houdebine, 1990). Polyadenylated RNA was isolated and used for first strand cDNA synthesis essentially as described previously (Pruitt, 1988) except that synthesis was primed using oligo(dT). Second strand synthesis was performed essentially as described in Gubler and Hoffman (1983). Following second strand synthesis, ClaI linkers were ligated to the cDNAs and cDNAs were digested with an excess of ClaI, recovered and ligated at a concentration of 1 μg/ml for 2 hours at 23°C. cDNAs containing LacZ sequences were then amplified by inverse PCR using LacZ-specific primers (5′ ATTAAGTTGGGTAAC-GCCAG 3′ and 5′ TATGTGGCGGATGAGCGGCA 3′) and Taq polymerase (Cetus) as recommended by the manufacturer using 1.0 mM MgCl2 and 26 cycles of 94°C for 1 minute, 45°C for 1 minute and 72°C for 2 minutes. PCR products were analyzed on a 1.0% agarose gel and the single resulting band of approximately 600 bp was isolated and cloned. Sequence analysis of the resulting clone was performed as described previously (Mielnicki and Pruitt, 1991). The resulting sequence was used in a FASTA homology search of Genbank (Pearson and Lipman, 1988).

β-galactosidase and protein concentration assays

In situ β-galactosidase assays were performed on cells fixed by treatment with 2.0% glutaraldehyde (Sigma) in phosphate-buffered saline (PBS) for 5 minutes followed by washing with PBS. Staining for β-galactosidase activity was performed as described in Goring et al. (1987) except that the substrate for the reaction was Bluo-gal (BRL). Reactions were stopped after 6 to 48 hours by washing with PBS. In cases where immunofluorescence was also performed, cells were fixed using −20°C methanol fixation as described by Rudnicki and McBurney (1987).

For β-galactosidase solution assays, cell extracts were prepared essentially as for use in chloramphenicol-acetyl-transferase assays (Gorman et al., 1982). Protein concentrations were determined using a Coomassie protein assay (Bio Rad) as described by the manufacturer. β-galactosidase assays were performed by addition of 25 to 50 μl of cell extract to 1.2 ml of 60 mM Na2HPO4, 40 mM NH2PO4, 10 mM KCl, 1 mM MgCl2, 50 mM β-mercaptoethanol and 0.33 mg/ml o-nitrophenyl-β-galactoside (pH 7.0) at 37°C for periods of 2 to 5 hours. Reactions were stopped by addition of 0.5 ml of 1 M Na2CO3 and absorbance at 420 nm was determined. One unit of activity is defined as the change in absorbance at 420 nm.

Northern blot analysis

Total RNAs (10 μg), isolated as described above, were electrophoresed, transferred to nylon membranes and probed as described previously (Pruitt and Natoli, 1992). The following DNA fragments were used as hybridization probes for specific mRNAs: wnt-1, a 2.5 kbp EcoRI fragment from pS621; Oct-3, a 450 bp EcoRI to HindIII fragment from pG2F9 (Rosner et al., 1990) and TPI, a 1.3 kbp BamHI fragment from pcDmTPI (Cheng et al., 1990).

Immunofluorescence

Cells were fixed and stained for β-galactosidase activity as described above. Immunofluorescence was performed as described previously (Pruitt and Natoli, 1992) using the following Abs: anti-NF-M (Oncogene Sciences), anti-GFAP (Oncogene Sciences) and fluorescein-conjugated goat anti-mouse IgG/IgM (Boeringer-Mannheim).

Isolation of a P19 EC stem cell line containing a Pax-3/LacZ fusion

To identify genes that are expressed in specific lineages following induction of P19 EC cell differentiation, P19S18O1A1 cells were transfected with the gene trap vector pT1 (Gossler et al., 1989). Integration of this vector has previously been shown not to alter the normal pattern of expression from endogenous genes in which it is integrated (Skarnes et al., 1992). Stable transfectants, which express β-galactosidase activity following induction as aggregates with either 1×10−6 M RA or 1.0% DMSO, but not constitutively, were isolated by sib-selection. Several lines that express β-galactosidase activity under only one induction condition were isolated and the present study utilizes one of these lines, O1A1pT1IIEH3 (EH3). This line was initially selected based on detection of β-galactosidase activity in aggregates at 3 days following RA, but not DMSO, treatment (Fig. 1A, D). Further characterization of this line (described below) demonstrated that β-galactosidase activity is maximal 6 days following RA treatment and RNA was isolated from RA-treated cultures at this time. Inverse PCR (e.g. Ochman et al., 1990) performed on cDNA synthesized from the isolated RNA, utilizing primers from the LacZ gene, resulted in the amplification of a single DNA fragment. Sequence analysis demonstrated that the LacZ sequence was fused to Pax-3 cDNA at position 882, placing the gene trap vector in the fourth intron of one copy of this gene. Further, analysis of RNA from untreated and 6 day RA-treated EH3 cells by northern blotting and hybridization with a LacZ probe fragment demonstrates the induction of an approximately 4 kb RNA following RA treatment, which is the predicted size of the Pax-3/LacZ fusion message (data not shown).

Expression of Pax-3/LacZ during chemically induced EH3 cell differentiation

EH3 cells treated as aggregates with either 1×10 −6 M RA or 1.0% DMSO were assayed in situ for the expression of β-galactosidase activity at various times after treatment. Chemical treatments were for a period of 4 days followed by growth in monolayer cultures in the absence of chemicals as described previously (McBurney et al., 1982). Following treatment with RA, β-galactosidase activity is first detected in some cells of day 3 aggregates (Fig. 1A). By day 6, nearly all of the cells express β-galactosidase activity (Fig. 1B, C). At later times, the more fully differentiated, mitotically inactive, cells near the periphery of the aggregates lose β-galactosidase activity (Fig. 2A, B) consistent with the localization of Pax-3 expression to mitotically active cells in the embryonic neuroepithelium (Goulding et al., 1991).

Fig. 1. Expression of β-galactosidase at early times in RA- and DMSO-treated EH3 cells. EH3 stem cells were treated as aggregates with l ×10−6 M RA (A-C) or 1.0% DMSO (D-F) as described in the text. A and D show aggregates, fixed and stained for β-galactosidase activity, on day 3 of induction (photomicrographs were taken at a magnification of 40 ×). B-C and E-F show aggregates, fixed and stained for β-galactosidase activity, on day 6 of the experiment following repating to monolayer cultures as described in the text (photomicrographs were taken at a magnification of 100×). Panels B and E were taken using phase-contrast optics. C and F are the same fieids respectively, taken using bright-field optics.

Fig. 2.

Expression of β-galactosidase activity at later times in RA- and DMSO-treated EH3 cells. EH3 stem cells were induced to differentiate and stained for β-galactosidase activity as described in Fig. 1. A (phase-contrast) and B (bright-field) show a culture on day 8 following 1×10 −6 M RA treatment and C (phase-contrast) and D (bright-field) show a culture on day 12 following 1.0% DMSO treatment (photomicrographs were taken at a magnification of 66×).

Fig. 2.

Expression of β-galactosidase activity at later times in RA- and DMSO-treated EH3 cells. EH3 stem cells were induced to differentiate and stained for β-galactosidase activity as described in Fig. 1. A (phase-contrast) and B (bright-field) show a culture on day 8 following 1×10 −6 M RA treatment and C (phase-contrast) and D (bright-field) show a culture on day 12 following 1.0% DMSO treatment (photomicrographs were taken at a magnification of 66×).

Cells exhibiting each of the morphologies that are typically formed following 1×10−6 M RA treatment of P19 EC cells transiently express β-galactosidase activity including those exhibiting neuronal morphologies (arrow in Fig. 2B). To confirm the identity of β-galactosidase-expressing cells resembling neurons, expression of two neuronal markers, neuro-filament M (NF-M, Fig. 3A) and HNK-1 (Fig. 3B; Abo and Balch, 1981; McBurney et al., 1988), and the astroglial cell marker glial fibrillary acidic protein (GFAP; Fig. 3C) were assayed by immunofluorescence and compared to β-galactosidase expression. Virtually all aggregates contained cells expressing NF-M, HNK-1 and β-galactosidase on days 6 or 8 following initiation of the experiment. No cells expressing GFAP were observed at these times, consistent with previous studies that have demonstrated that mature astroglial cells, expressing GFAP, do not appear until days 9 or 10 in RA-treated cultures (e.g. Edwards and McBurney, 1983). Quantitation demonstrated that each of 105 aggregates scored contained HNK-1-positive cells in close association with β-galactosidase-expressing cells. Additionally, individual cells expressing β-galactosidase and either NF-M or HNK-1 were found. However, as suggested by morphology, many of the less differentiated cells towards the center of the aggregates express β-galactosidase, but not NF-M or HNK-1, and many fully differentiated NF-M- or HNK-1-expressing cells near the periphery of the aggregates no longer express β-galactosidase. Quantitation of the fraction of cells expressing both HNK-1 and β-galactosidase was performed on individual cells following brief trypsinization of aggregates and replating to poly-L-lysine-coated dishes. These studies demonstrated that only approximately 4% (n = 725) of cells expressing β-galactosidase were also positive for HNK-1 expression. Additionally, in cells that expressed both markers, expression of each was weak relative to cells expressing a single marker consistent with the down regulation of Pax-3/LacZ expression in differentiated cells. Similar results were obtained regardless of whether β-galactosidase expression was detected using the histochemical stain for β-galactosidase activity shown in Fig. 3 or double label immunofluorescence using an anti-β-galactosidase Ab (data not shown). Although not proven, it is likely that many of the Pax-3-expressing cells present following induction under these conditions are neuroectodermal derivatives which have not yet differentiated sufficiently to express neural or glial cell-specific markers.

Fig. 3. Expression of neuroectodermal differentiation markers in RA- and DMSO-treated EH3 cells. EH3 cells were treated, as aggregates, with 1×10−6 M RA (A-C) or 1.0% DMSO (D-F) as previously described. On day 8 following initiation of the experiment, all cultures were stained in situ for β-galactosidase activity and expression of NF-M (A and D), HNK-1 (B and E), and GFAP (C and F) were assayed by immunofluorescence. Photomicrographs were taken at a magnification of 100×.

Following treatment with DMSO, β-galactosidase expression is not detected in day 3 aggregates (Fig. 1D) and is never expressed in a large proportion of DMSO-induced cells. However, by day 6, small foci of β-galactosidase-expressing cells are found (Fig. 1E, F). At later times, cells exhibiting a variety of differentiated morphologies including skeletal myoblasts (arrow in Fig. 2E) are formed and do not express β-galactosidase activity although foci of expressing cells are present (Fig. 2E, F). To determine if cells expressing the Pax-3/LacZ gene following DMSO treatment were neuroectodermal derivatives, expression of NF-M (Fig. 3D), HNK-1 (Fig. 3E) and GFAP (Fig. 3F) were assayed by immunofluorescence and compared to the expression of β-galactosidase activity. Typically, these markers were not expressed in DMSO-induced cultures, even in those cells expressing β-galactosidase. On rare occasions, a few cells expressing NF-M were present (no cases of HNK-1 or GFAP expression were found) and these cells were invariably in close association with foci of β-galactosidase-expressing cells. Quantitation demonstrated that no HNK-1-expressing cells were found in 107 β-galactosidase-expressing foci assayed. As expected on the basis of observations in intact aggregates, analysis of individual cells following trypsinization demonstrated that HNK-1 was not detectable in any of the β-galactosidase-expressing cells assayed (n = 504). These observations suggest that the Pax-3-expressing cells present in cultures treated with DMSO are predominantely not neuroectodermal derivatives and expression under this condition may reflect the in vivo expression of Pax-3 in the dermomyotome.

Quantitation of the level of β-galactosidase activity by solution assays at different times following treatment with RA or DMSO is shown in Fig. 4B and correlates well with observations from in situ analysis.

Fig. 4.

Expression and quantitation of β-galactosidase activity at various times following treatment of EH3 cells with RA or DMSO in the presence or absence of LIF. (A) EH3 cells were treated, as aggregates, with no chemical treatment (plates in upper left and right), 1.0% DMSO (plates in middle left and right), or 1×10 −6 M RA (plates in lower left and right) in either the absence (left) or presence (right) of 100 units/ml LIF as described previously. Cultures were fixed and stained for β-galactosidase activity on day 8 following initiation of the experiment. (B) β-galactosidase activity was quantitated from cell extracts (prepared as described in the text) of cultures treated for various times under the following conditions: 1×10 −6 M RA, no LIF (•); 1×10 −6 M RA, 100 units/ml LIF (●); 1.0% DMSO, no LIF (▪); or 1.0% DMSO, 100 units/ml LIF (▫).

Fig. 4.

Expression and quantitation of β-galactosidase activity at various times following treatment of EH3 cells with RA or DMSO in the presence or absence of LIF. (A) EH3 cells were treated, as aggregates, with no chemical treatment (plates in upper left and right), 1.0% DMSO (plates in middle left and right), or 1×10 −6 M RA (plates in lower left and right) in either the absence (left) or presence (right) of 100 units/ml LIF as described previously. Cultures were fixed and stained for β-galactosidase activity on day 8 following initiation of the experiment. (B) β-galactosidase activity was quantitated from cell extracts (prepared as described in the text) of cultures treated for various times under the following conditions: 1×10 −6 M RA, no LIF (•); 1×10 −6 M RA, 100 units/ml LIF (●); 1.0% DMSO, no LIF (▪); or 1.0% DMSO, 100 units/ml LIF (▫).

LIF inhibits Pax-3/LacZ expression in DMSO-but not RA-treated EH3 cells

Previous studies have demonstrated that the cytokine leukemia inhibitory factor (LIF) prevents differentiation of endodermal and mesodermal, but not neuroectodermal lineages, following chemical treatments of P19 EC cells (Pruitt and Natoli, 1992). The ability of LIF to block Pax-3/LacZ expression following treatment of EH3 cell aggregates with either RA or DMSO was determined by including 100 units/ml of LIF during the 4-day chemical treatment. Fig. 4A shows plates of cells that were fixed and stained for β-galactosidase activity on day 8 following initiation of the experiment and demonstrates that LIF efficiently blocks expression of both the low level of activity observed in cells which are aggregated in the absence of chemical treatment and the higher levels observed in cells aggregated in the presence of 1.0% DMSO. In contrast, LIF fails to inhibit expression of β-galactosidase activity following 1×10−6 M RA treatment. These results are confirmed quantitatively in Fig. 4B using a β-galactosidase solution assay for various times following treatment with either DMSO or RA in the presence or absence of LIF.

Endogenous factors expressed by differentiated P19EC cell derivatives induce Pax-3/LacZ expression in EH3 stem cells

To assay for the presence of cells that express a factor capable of inducing Pax-3/LacZ expression in EH3 stem cells, cell mixing experiments were performed as shown in Fig. 5. First, unmarked P19 EC stem cells were pretreated as aggregates with either 1.0% DMSO, LIF, 1×10 −6 M RA, or 1 ×10 −6 M RA plus LIF. At various times, cells cultured under each of these conditions were trypsinized, washed repeatedly and mixed in a 5:1 ratio with target EH3 stem cells. These mixtures were cultured for an additional 4 days as aggregates in the presence of LIF, plated to monolayer cultures and cultured for an additional two days in the continued presence of LIF. Cells were then either fixed and stained for β-galactosidase activity in situ or harvested for solution assays. Plates of cells from mixtures containing unmarked P19 cells pretreated using different conditions for 4 days prior to mixing are shown in Fig. 6A.

Fig. 5.

Cell mixing assay for endogenous Pax-3/LacZ-inducing activity. To assay for the expression of an endogenous activity capable of inducing expression from the Pax-3/LacZ fusion gene present in EH3 cells, unmarked P19 EC cells were, first, differentiated by chemical treatments (RA, RA plus LIF, DMSO) as aggregates. Undifferentiated P19 cells grown as aggregates in the presence of LIF were also used as a control. At various times following initiation of the experiment, chemical treatments were removed by repeated washing and the pretreated unmarked P19 cells were mixed with target EH3 stem cells in the presence of LIF. Cell mixtures were then grown as aggregates for 4 days followed by replating to monolayer culture for an additional 2 or 4 days in the continuous presence of LIF (to prevent any spontaneous differentiation). If an activity capable of inducing Pax-3 expression was expressed in the pretreated unmarked P19 cells, and this activity continues to be expressed following removal of the inducer, its expression would be detected in mixed aggregates by induction of β-galactosidase activity from the Pax-3/LacZ gene present in the target EH3 stem cells.

Fig. 5.

Cell mixing assay for endogenous Pax-3/LacZ-inducing activity. To assay for the expression of an endogenous activity capable of inducing expression from the Pax-3/LacZ fusion gene present in EH3 cells, unmarked P19 EC cells were, first, differentiated by chemical treatments (RA, RA plus LIF, DMSO) as aggregates. Undifferentiated P19 cells grown as aggregates in the presence of LIF were also used as a control. At various times following initiation of the experiment, chemical treatments were removed by repeated washing and the pretreated unmarked P19 cells were mixed with target EH3 stem cells in the presence of LIF. Cell mixtures were then grown as aggregates for 4 days followed by replating to monolayer culture for an additional 2 or 4 days in the continuous presence of LIF (to prevent any spontaneous differentiation). If an activity capable of inducing Pax-3 expression was expressed in the pretreated unmarked P19 cells, and this activity continues to be expressed following removal of the inducer, its expression would be detected in mixed aggregates by induction of β-galactosidase activity from the Pax-3/LacZ gene present in the target EH3 stem cells.

Fig. 6.

Induction and quantitation of β-galactosidase activity following mixing of target EH3 stem cells with unmarked P19 EC cells pretreated under different conditions. (A) Unmarked P19 EC cells were pretreated as aggregates with different combinations of chemicals and LIF for four days, trypsinized and washed (as described in the text) and mixed in a 5:1 ratio with target EH3 stem cells. Mixtures were incubated as aggregates in the presence of 100 units/ml LIF, but no chemical treatment, for a period of 4 days and replated to monolayer cultures in the continued presence of LIF. On day 6 following mixing, cultures were fixed and stained for the presence of β-galactosidase activity. Unmarked P19 EC cells were pretreated as follows: upper left, 1.0% DMSO; upper right 1×10 −6 M RA; lower left, 100 units/ml LIF; lower right 1×10 −6 M RA plus 100 units/ml LIF. (B) Mixed cultures of target EH3 stem cells and pretreated unmarked P19 cells were prepared as described in A except that P19 cells were pretreated for different lengths of time prior to mixing. β-galactosidase activity was quantitated from cell extracts on day 6 following mixing as described in the text. Unmarked P19 cells were pretreated under the following conditions: 1×10 −6 M RA, no LIF (•); 1×10 −6 M RA, 100 units/ml LIF (○); 1.0% DMSO, no LIF (▪); and no chemical treatment, no LIF (▵).

Fig. 6.

Induction and quantitation of β-galactosidase activity following mixing of target EH3 stem cells with unmarked P19 EC cells pretreated under different conditions. (A) Unmarked P19 EC cells were pretreated as aggregates with different combinations of chemicals and LIF for four days, trypsinized and washed (as described in the text) and mixed in a 5:1 ratio with target EH3 stem cells. Mixtures were incubated as aggregates in the presence of 100 units/ml LIF, but no chemical treatment, for a period of 4 days and replated to monolayer cultures in the continued presence of LIF. On day 6 following mixing, cultures were fixed and stained for the presence of β-galactosidase activity. Unmarked P19 EC cells were pretreated as follows: upper left, 1.0% DMSO; upper right 1×10 −6 M RA; lower left, 100 units/ml LIF; lower right 1×10 −6 M RA plus 100 units/ml LIF. (B) Mixed cultures of target EH3 stem cells and pretreated unmarked P19 cells were prepared as described in A except that P19 cells were pretreated for different lengths of time prior to mixing. β-galactosidase activity was quantitated from cell extracts on day 6 following mixing as described in the text. Unmarked P19 cells were pretreated under the following conditions: 1×10 −6 M RA, no LIF (•); 1×10 −6 M RA, 100 units/ml LIF (○); 1.0% DMSO, no LIF (▪); and no chemical treatment, no LIF (▵).

In cultures containing DMSO-induced P19 cells, over 80% of the aggregates contained EH3 target cells in which β-galactosidase activity from the Pax-3/LacZ gene is expressed even though the mixed aggregates were maintained in the continuous presence of LIF without chemical treatment (Fig. 6A, upper left). This contrasts with the very low levels of β-galactosidase activity in EH3 cell aggregates grown in the presence of LIF (Fig. 4A, upper right), DMSO plus LIF (Fig. 4A, middle right), or when uninduced P19 EC stem cells (grown in the presence of LIF) are mixed with EH3 target cells under the same conditions (Fig. 6A, lower left). Quantitation of the efficiency of β-galactosidase induction when DMSO-pretreated unmarked P19 cells are allowed to differentiate for increasing lengths of time before mixing with EH3 stem cells is shown in Fig. 6B where β-galactosidase activity was measured using a solution assay. This assay confirms the results from in situ studies and demonstrates that cells expressing the Pax-3/LacZ-inducing activity are present transiently, peaking on day four, following treatment of unmarked P19 cells with DMSO.

RA-treated P19 cells also contain a population of cells that is capable of inducing Pax-3/LacZ expression in target EH3 stem cells even when the mixed aggregates are cultured in the presence of LIF (Fig. 6A, upper right, and Fig. 6B). However, when LIF is present during the pretreatment of unmarked P19 cells in 1×10−6 M RA, expression of the Pax-3/LacZ-inducing activity is inhibited (Fig. 6A, lower right, and Fig. 6B). This inhibition occurs despite the fact that direct treatment of EH3 cells with 1×10−6 M RA in the presence of LIF results in both Pax-3/LacZ expression and neuroectodermal differentiation as discussed previously. Additionally, inspection of the aggregates resulting from mixtures of target EH3 stem cells and RA plus LIF-pre-treated unmarked P19 cells demonstrates that each aggregate contains a high proportion of morphologically differentiated cells (including neurons). This observation is consistent with the presence of neuroectodermal cell types derived from the pretreated unmarked P19 cells in each aggregate (data not shown) and suggests that these differentiated cell types do not express the endogenous Pax-3-inducing activity. Hence, in 1×10−6 M RA-treated cultures, two mechanisms can apparently result in Pax-3/LacZ expression: a LIF-insensitive induction (possibly occurring as a direct effect of RA), and a LIF-sensitive mechanism (requiring an endogenous factor that is not expressed by the neuroectodermal cell lineages formed following treatment with RA in the presence of LIF).

Endogenous factors expressed by differentiated P19 EC cell derivatives induce neuroectoderm differentiation in EH3 stem cells

Experiments described above demonstrate that both DMSO- and RA-pretreated P19 EC cell cultures contain cells that express a factor capable of inducing Pax-3/LacZ expression in target EH3 stem cells. In many cases, the morphology of individual β-galactosidase-expressing cells present in mixed aggregates could be observed. In these cases, at least some of the β-galactosidase-expressing cells in each of the staining foci exhibited a neuronal morphology regardless of whether the unmarked P19 cells used to induce expression of the Pax-3/LacZ gene had been pre-treated with DMSO or RA (Fig. 7). More generally, it is possible to detect NF-M-, HNK-1- and GFAP-expressing cells in close association with most of the Pax-3/LacZ-expressing foci resulting from mixing experiments with either DMSO- or RA-pretreated unmarked P19 cells. Fig. 8 shows examples of β-galactosidase-staining foci associated with NF-M- (Panel A), HNK-1- (Panel B) and GFAP- (Panel C) expressing cells in mixing experiments using DMSO-treated, unmarked, P19 cells. The frequent association of NF-M- and HNK-1-expressing cells with β-galactosidase expressing foci in these cultures is in sharp contrast to the near absence of these cells when EH3 cells are treated directly with DMSO. Additionally, in mixing experiments, cells expressing GFAP are found in association with most β-galactosidase-expressing foci on days 6 and 8 following mixing even though cells expressing this marker were never observed in cultures treated directly with DMSO and appeared only at later times, days 9-10, in cultures treated directly with RA. Quantitation of individual, trypsinized, cells suggests that a suprisingly high proportion of cells that express β-galactosidase also express HNK-1 (20% in mixing experiments using RA pretreated P19 cells and 35% in mixing experiments using DMSO-pre-treated P19 cells). However, as observed previously in cells that co-express both markers, the level of expression of each is low (data not shown). Quantitation of the extent to which neuroectodermal markers are tightly associated with β-galactosidase-expressing foci in undisrupted aggregates is more conclusive. In mixing experiments using unmarked P19 cells pretreated with either RA or DMSO, 92% and 90% respectively of β-galactosidase-expressing foci were closely associated with HNK-1-expressing cells. In the same experiment, approximately 30% of HNK-1-positive cells occurred in the absence of β-galactosidase-expressing foci in the case where RA-pretreated P19 cells were used, consistent with the high proportion of neurons in the RA-treated P19 cell population. However, in the case where DMSO-pretreated P19 cells were used, nearly all (96%) of the HNK-1-expressing cells were closely associated with β-galactosidase-expressing foci, similar to the example shown in Fig. 8B, confirming that the DMSO-pretreated P19 cells contribute very few HNK-1-positive cells to the mixed cultures. These observations demonstrate that most of the neuroectodermal differentiation occurring in mixing experiments using DMSO-pretreated, unmarked, P19 cells results from induction of target EH3 stem cells by an endogenous activity and is closely associated with the expression of Pax-3.

Fig. 7.

Neuronal morphologies of β-galactosidase-expressing cells present in cell mixing experiments. Examples of β-galactosidase-expressing cells exhibiting a neuronal morphology in mixed cultures of target EH3 stem cells and P19 EC cells, pretreated with either DMSO (A) or RA (C) for 4 days prior to mixing, on day 8 following mixing as described previously. The cells shown in A were also assayed for expression NF-M by immunofluorescence in B. Photomicrographs were taken at a magnification of 54×.

Fig. 7.

Neuronal morphologies of β-galactosidase-expressing cells present in cell mixing experiments. Examples of β-galactosidase-expressing cells exhibiting a neuronal morphology in mixed cultures of target EH3 stem cells and P19 EC cells, pretreated with either DMSO (A) or RA (C) for 4 days prior to mixing, on day 8 following mixing as described previously. The cells shown in A were also assayed for expression NF-M by immunofluorescence in B. Photomicrographs were taken at a magnification of 54×.

Fig. 8. Co-localization of β-galactosidase activity and NF-M-, HNK-1-, or GFAP-specific immunofluorescence in mixed aggregates between target EH3 stem cells and DMSO-pretreated P19 cells. Mixed aggregates of target EH3 stem cells and unmarked P19 EC cells pretreated with 1.0% DMSO for 4 days were prepared and Cultured in the presence of 100 units/ml LTF as described in Fig. 5. Cultures were fixed on day 8 following mixing and assayed for the presence of β-galactosidase activity as described above and the presence of NF-M (A), HNk-1 (B), or GFAP (C) by immunofluorescence. Photomicrographs were taken at a magnification of 100×.

Wnt-1 expression correlates with expression of Pax-3-and neuroectoderm-inducing activities but not RA treatment

Previous studies have demonstrated expression of the neuroectodermal marker, wnt-1, following induction of P19 EC cells with 1×10 −6 M RA (St-Arnaud et al., 1989; Schuuring et al., 1989). The effect of LIF on RA-induced wnt-1 expression was determined by treating EH3 cells, as aggregates, with different concentrations of RA either in the presence or absence of LIF as previously. On day 6 following initiation of the experiment, cultures were harvested and used for preparation of RNAs for northern blot analysis. Fig. 9 shows RNAs probed for the presence of the stem-cell-specific marker Oct-3 (to assess the level of differentiation; Rosner et al., 1989; Scholer et al., 1989; Pruitt and Natoli, 1992), wnt-1 and triose-phosphate isomerase (TPI; a constitutively expressed control). Densitometric analysis, corrected for the level of signal obtained for the constitutively expressed control TPI, demonstrates that Oct-3 mRNA expression is down regulated by a factor of greater than 100 in cells treated with 1×10 −6 M RA and approximately 20 in cells treated with 1×10 −6 M RA plus 100 units/ml LIF (Panel A). (Induction of differentiation following treatment of cultures with 1×10−7 M RA, or less, occurs more slowly and Oct-3 expression is not down regulated until between 7 and 10 days under these conditions.) These results suggest that a high proportion (approximately 90 –95%) of cells present in cultures treated with 1×10−6 M RA were induced to differentiate even in the presence of LIF. However, wnt-1 expression (Panel B) which, relative to untreated cultures, is induced by a factor of greater than 20 in 1×10 −6 M RA-treated cultures fails to be induced in cultures treated with 1×10 −6 M RA in the presence of LIF. This result demonstrates that LIF inhibits wnt-1 expression following treatment with RA. Experiments described above have shown that LIF also prevents expression of the Pax-3- and neuroectoderm-inducing activities observed in RA-treated cultures as assayed by mixing experiments. It is possible that wnt-1 expression is not directly dependent on RA for induction but, rather, results from the endogenous neuroectoderm-inducing activity demonstrated here.

Fig. 9.

Effect of LIF on wnt-1 expression in RA-treated EH3 cells. RNAs extracted from EH3 cells on day 6 following treatment of aggregates (as described previously) with different concentrations of RA (none, lanes 1 and 6; 1×10 −8 M, lanes 2 and 7; 1×10 −7 M, lanes 3 and 8; 1×10 −6 M, lanes 4 and 9; and 1×10 −5 M, lanes 5 and 10) in either the absence (lanes 1 –5) or presence (lanes 6 –10) of 100 units/ml LIF were analyzed for the expression of different mRNA species by northern blot analysis as described in the text. RNAs were probed for: (A) Oct-3 (12 hour exposure); (B) wnt-1 (7 day exposure); and (C) TPI (3 hour exposure).

Fig. 9.

Effect of LIF on wnt-1 expression in RA-treated EH3 cells. RNAs extracted from EH3 cells on day 6 following treatment of aggregates (as described previously) with different concentrations of RA (none, lanes 1 and 6; 1×10 −8 M, lanes 2 and 7; 1×10 −7 M, lanes 3 and 8; 1×10 −6 M, lanes 4 and 9; and 1×10 −5 M, lanes 5 and 10) in either the absence (lanes 1 –5) or presence (lanes 6 –10) of 100 units/ml LIF were analyzed for the expression of different mRNA species by northern blot analysis as described in the text. RNAs were probed for: (A) Oct-3 (12 hour exposure); (B) wnt-1 (7 day exposure); and (C) TPI (3 hour exposure).

In vitro expression of endogenous Pax-3- and neuroectoderm-inducing activities

An in vitro system that accurately reflects the inductive interactions occurring in the mammalian embryo during gastrulation and neurulation would facilitate analysis of early cell lineage decisions. The possibility that inductive interactions that mimic those in the early embryo occur in EC and ES cells in vitro is suggested by previous observations. Many ES, and some EC, cell lines differentiate spontaneously to form a variety of cell types, including some neuroectodermal lineages, when grown as aggregates in the absence of chemical inducers (reviewed in Robertson, 1987). In these cases, formation of neuroectodermal lineages is preceded by differentiation of endodermal and mesodermal cell types. Although differentiation of neuroectodermal lineages seems likely to result from specific inductive interactions within such aggregates, a stochastic partitioning of cells to this lineage is not eliminated. In the case of P19 EC cells, both RA treatment and aggregation are necessary for efficient neuroectoderm formation suggesting that neuroectoderm formation is an indirect consequence of RA treatment and requires secondary interactions between differentiating cells. The present study directly demonstrates the expression, during differentiation of P19 EC cells, of endogenous activities that are capable of inducing both Pax-3 expression and neuroectodermal differentiation in undifferentiated stem cells. These inductions may reflect mechanisms responsible for induction of Pax-3 expression and neuroectoderm differentiation in the embryo.

Pax-3- and neuroectoderm-inducing activities act on pluripotent stem cells

The ability of the endogenous Pax-3- and neuroectoderminducing activities to act on undifferentiated EH3 cells even in the presence of LIF, which is known to inhibit spontaneous differention in a variety of EC and ES cell lines, suggests that stem cells respond directly to these activities. This observation is consistent with studies in amphibians, which demonstrate that a factor expressed in mesoderm acts directly on primitive ectoderm to induce neuroectoderm. Embryonal carcinoma stem cells are known to derive specifically from the primitive ectoderm (epiblast) of the mouse embryo on days 7 or 8 of gestation (Diwan and Stevens, 1976; Beddington, 1983b). Many EC cell lines, including P19 EC cells, continue to exhibit characteristics of primitive ectoderm including pluripotency (Beddington, 1983b; 1986) and expression of primitive ectoderm-specific cell markers (e.g. Oct-3; Rosner et al., 1989; Scholer et al., 1989). Hence, the ability of P19 EC cells to respond to Pax-3- and neuroectoderm-inducing signals is likely to reflect the mechanism by which neuroectoderm is induced in vivo.

Effect of LIF on endogenous neuroectoderm-inducing activity

Previous work has established that differentiation of neuroectoderm following treatment of P19 cells with RA is not affected by the presence of LIF, even though differentiation of mesodermal lineages is inhibited (Pruitt and Natoli, 1992). The present study extends this observation and demonstrates that the endogenous Pax-3- and neuroectoderm-inducing activities produced by differentiating P19 EC cells are also capable of acting in the presence of LIF. Although expression of LIF mRNA has not been detected in the early embryo by in situ hybridization, the uterine endometrium expresses high levels of LIF mRNA just prior to blastocyst implantation (Bhatt et al., 1991). The possibility that LIF expressed in the endometrium is translocated to the inner cell mass of the embryo via the polar trophectoderm has been suggested (Bhatt et al., 1991). If LIF is present in the early embryo, the ability of endogenous neuroectoderm-inducing factors to act in its presence could have a role in allowing neuroectoderm differentiation in regions where mesoderm differentiation is inhibited.

Endogenous Pax-3- and neuroectoderm-inducing activities are not expressed in all differentiated cell lineages

The cell type, or types, that produce the endogenous Pax-3- and neuroectoderm-inducing activities have not been identified. However, not all differentiated lineages express this factor. Unmarked P19 cells pretreated with RA in the presence of LIF differentiate towards neurons and astroglial cells but fail to express Pax-3/LacZ-inducing activity when mixed with target EH3 stem cells. This observation suggests that neuroectodermal lineages do not express the inducing activities. In contrast, the most efficient induction of Pax-3/LacZ in mixing experiments occurs when the unmarked differentiated cells are derived from 1.0% DMSO-pre-treated cultures, a condition that is known to induce mesodermal lineages efficiently (McBurney et al., 1982; Edwards et al., 1983). Further, it is possible that the ability of P19 cells pretreated with 1×10 −6 M RA in the absence of LIF to induce neuroectoderm results from the presence of at least some mesodermal lineages in these cultures. Slightly lower RA concentrations (1×10 −7 M) are known to induce mesodermal lineages efficiently (including many of the same lineages formed following 1.0% DMSO treatment) and this differentiation is blocked by LIF (Pruitt and Natoli, 1992). The observation that LIF inhibits expression of wnt-1 mRNA, but not differentiation of neurons or astroglial cells, in 1×10−6 M RA-treated cultures suggests that LIF is inhibiting differentiation of a specific population of cells in these cultures. These observations are consistent with the possibility that, in both DMSO- and RA-treated cultures, cells differentiating along mesodermal lineages express the Pax-3- and neuroectoderm-inducing activities.

Comparison of RA treatment and endogenous neuroectoderm-inducing activity

The observation that neurons and astroglial cells differentiate when stem cell aggregates are treated directly with 1×10 −6 M RA in the presence of LIF, even though LIF inhibits expression of the endogenous neuroectoderm-inducing activity under these conditions, suggests that RA can act directly on stem cells to induce at least some neuroectodermal lineages. This conclusion is also consistent with the observation that single cells in suspension culture can be induced towards neuroectodermal differentiation by RA (Berg and McBurney, 1990). It is likely that both the endogenous neuroectoderm-inducing activity and RA ultimately act on the same differentiation pathway. One possibility is that the endogenous factor is RA and the activation of both Pax-3 expression and the differentiation of cells exhibiting a neuronal morphology by both of these inducers is consistent with this possibility. However, some observations suggest that RA and the endogenous neuroectoderm-inducing activity act at different points in the neuroectoderm differentiation pathway. First, cells expressing GFAP appear at earlier times when induced by the endogenous factor than following treatment with RA. Second, although expression of wnt-1, a neuroectodermal marker which is localized to the roof plate of the neural tube (McMahon and Bradley, 1990; Thomas and Capecchi et al., 1990), is induced by RA treatment, this induction is blocked in the presence of LIF. This block correlates with the effect of LIF in blocking expression of the endogenous neuroectoderm-inducing activity in unmarked P19 cells treated with both RA and LIF. One explanation for these results is that, in 1×10 −6 M RA-induced cultures, RA does not induce wnt-1 expression directly but, rather, its expression is entirely dependent on the endogenous factor.

Relationship between endogenous Pax-3- and neuroectoderm-inducing activities

The experiments described here suggest that, in differentiating P19 cells, expression of the Pax-3-inducing activity is closely associated with the neuroectoderm-inducing activity since greater than 90% of Pax-3-expressing cells were found closely associated with neuroectodermal derivatives in cell mixing experiments. This observation is particularly clear in mixing experiments where DMSO-pre-treated, unmarked, P19 cells were used since treating cells directly with DMSO results in virtually no neuroectodermal lineages. Additionally, in these cultures, nearly all of the neuroectodermal derivatives present were within or adjacent to Pax-3-expressing foci. Further, in each case where the morphology of individual Pax-3-expressing cells within the β-galactosidase-expressing foci present in mixed aggregate cultures was observed, at least some of these cells exhibited a neuronal morphology and, where assayed, expressed the neuronal markers NF-M or HNK-1. These results suggest that expression of the endogenous neuroectoderm-inducing activity is tightly coupled to the Pax-3-inducing activity in differentiating P19 cells. One possible explanation for the close association of Pax-3- and neuroectoderm-inducing activities in mixed aggregate experiments is that each activity results from expression of a different factor but that both factors are expressed in the same cell type. Alternatively, the same factor may be responsible for both Pax-3 and neuroectoderm induction. In the developing embryo, Pax-3 transcripts are restricted to dorsal regions of the neuroepithelium and neural tube in positions that are not adjacent to the underlying mesoderm. Nonetheless, transplantation experiments using chick embryos demonstrate that elimination of the notochord and floor plate results in expression of dorsal markers throughout the neural tube, suggesting that differentiation towards dorsal lineages may be the constitutive fate of neuroepithelial cells (Yamada et al., 1991; Placzek et al., 1991). If this is the case, induction of Pax-3 (and wnt-1) expression could be a direct consequence of the induction of neuroectodermal lineages.

In addition to expression in the neuroepithelium, Pax-3 message is detected in the adjacent segmented dermomyotome in 8.5 day embryos (Goulding et al., 1991). It is likely that expression of Pax-3 is also occurring in mesodermal lineages formed during differentiation of P19 EC cells since very little expression of the neuroectodermal markers NF-M, HNK-1 or GFAP could be detected following direct treatment of EH3 cells with DMSO even though substantial levels of expression from the Pax-3 gene occurred in these cultures. A likely explanation for this observation is that the same signal transduction system acts to induce Pax-3 expression both in cultures treated directly with DMSO and in mixed aggregate experiments but that different target cells are present in these different cases. In mixing experiments, the target for the Pax-3- and neuroectodermal-inducing activities are primitive ectoderm-like, undifferentiated, stem cells while, in the case of direct treatment with DMSO, many of the cells that are exposed to the Pax-3-inducing activity may already have become committed to mesodermal lineages. The possibility that a single signal is responsible for Pax-3 expression in both neuroectodermal and mesodermal lineages is also consistent with the close proximity of the neuroectodermal and mesodermal lineages which express Pax-3 in the 8.5 day embryo.

The present results suggest that, similar to amphibian embryos, primitive ectoderm in the mammalian embryo can respond to signals which direct it toward a neuroectodermal fate. Although not directly demonstrated, these results are consistent with this signal originating from mesodermal lineages. Preliminary experiments using embryonic cells in the EH3-cell/LIF mixing assay are also consistent with a mesodermal origin of the signal since embryos at the early neural plate stage, but not at the egg-cylinder stage, contain cells that express a factor capable of inducing Pax-3/LacZ expression in EH3 cells (data not shown) and this approach should allow localization of these cells in the early mouse embryo. However, an important difference exists between lineage decisions in the primitive ectoderm of amphibians and mammals. In amphibians, primitive ectoderm that fails to receive any additional signal within a narrow window of time becomes committed to an epidermal fate (e.g. Grainger and Gurdon, 1989; Servetnick and Grainger, 1991). In contrast, the ability to isolate and propagate murine EC and ES cell lines from primitive ectoderm at high efficiency (e.g. Diwan and Stevens; 1976; Beddington, 1983b), and the fact that these lines remain pluripotent and, at least in the case of P19 cells, retain the ability to respond to endogenous neuroectoderm-inducing signals suggest that primitive ectoderm of mammalian embryos does not become committed to an epidermal fate unless an additional signal is received. In mammals, rather than redirecting the fate of primitive ectoderm, the neuroectoderm-inducing signal demonstrated here may act to commit primitive ectoderm to a neuroectodermal fate prior to the activity of an epiderm-inducing factor.

The author is grateful to M. McBurney for providing the P19S18O1A1 cell line; H. Baumann for providing LIF; J. Rossant for providing the plasmid pT1; H. Varmus for providing the plasmid pS621 containing the wnt-1 cDNA; and M. Rosner for providing the plasmid pG2F9 containing the Oct-3 cDNA. This work was funded by Public Health Research Grant HD25419 from the National Institutes of Health to S.C.P.

Abo
,
T.
and
Balch
,
C. M.
(
1981
).
A differentiation antigen of human NK and K cells identified by a monoclonal antibody (HNK-1
).
J. Immunol
.
127
,
1024
1029
.
Ackers
,
R. M.
,
Phillips
,
C. R.
and
Wessells
,
N. K.
(
1986
).
Expression of an epidermal antigen used to study tissue induction in the early Xenopus laevis embryo
.
Science
231
,
613
616
.
Beddington
,
R. S. P.
(
1983a
). The origin of the foetal tissues during gastrulation in the rodent. In
Development in Mammals, vol. 5
(ed.
M. H.
Johnson
) pp
1
-
32
. Amsterdam:
Elsevier
.
Beddington
,
R. S. P.
(
1983b
).
Histogenic and neoplastic potential of different regions of the mouse embryonic egg cylinder
.
J. Embryol. Exp. Morph
.
75
,
189
204
.
Beddington
,
R.
(
1986
). Analysis of tissue fate and prospective potency in the egg cylinder. In
Experimental Approaches to Mammalian Embryonic Development
(ed.
J.
Rossant
and
R. A.
Pedersen
) pp
121
-
147
.
Cambridge
:
Cambridge University Press
.
Berg
,
R. W.
and
McBurney
,
M. W.
(
1990
).
Cell density and cell cycle effects on retinoic acid-induced embryonal carcinoma cell differentiation
.
Dev. Biol
.
138
,
123
135
.
Bhatt
,
H.
,
Brunet
,
L. J.
and
Stewart
,
C. L.
(
1991
).
Uterine expression of leukemia inhibitory factor coincides with the onset of blastocyst implantation
.
Proc. Natl. Acad. Sci. USA
88
,
11408
11412
.
Cheng
,
J.
,
Mielnicki
,
L. M.
,
Pruitt
,
S. C.
and
Maquat
,
L. E.
(
1990
).
Nucleotide sequence of murine triosephosphate isomerase cDNA
.
Nuc. Acids Res
.
18
,
4261
.
Cho
,
K. W. Y.
,
Blumberg
,
B.
,
Steinbeisser
,
H.
and
De Robertis
,
E. M.
(
1991
).
Molecular nature of Spemann’s organizer: the role of the Xenopus homeobox gene goosecoid
.
Cell
67
,
1111
1120
.
Chomczynski
,
P.
and
Sacchi
,
N.
(
1987
).
Single-step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction
.
Anal. Biochem
.
162
,
156
159
.
Diwan
,
S. B.
and
Stevens
,
L. C.
(
1976
).
Development of teratomas from ectoderm of mouse egg cylinders
.
J. Natl. Cancer Inst
.
57
,
937
942
.
Dixon
,
J. E.
and
Kintner
,
C. R.
(
1989
).
Cellular contacts required for neural induction in Xenopus embryos: evidence for two signals
.
Development
106
,
749
757
.
Edwards
,
M. K. S.
,
Harris
,
J. F.
and
McBurney
,
M. W.
(
1983
).
Induced muscle differentiation in an embryonal carcinoma cell line
.
Mol. Cell. Biol
.
3
,
2280
2286
.
Edwards
,
M. K. S.
and
McBurney
,
M. W.
(
1983
).
The concentration of retinoic acid determines the differentiated cell types formed by a teratocarcinoma cell line
.
Dev. Biol
.
98
,
187
191
.
Goring
,
D. R.
,
Rossant
,
J.
,
Clapoff
,
S.
,
Breitman
,
M. L.
and
Tsui
,
L.-C.
(
1987
).
In situ detection of β-galactosidase in lenses of transgeneic mice with α-crystallin/lacZ gene
.
Science
235
,
456
458
.
Gorman
,
C. M.
,
Moffat
,
L. F.
and
Howard
,
B. H.
(
1982
).
Recombinant genomes which express chloramphenicol acetyltransferase in mammalian cells
.
Mol. Cell. Biol
.
2
,
1044
1051
.
Gossler
,
A.
,
Joyner
,
A. L.
,
Rossant
,
J.
and
Skarnes
,
W. C.
(
1989
).
Mouse embryonic stem cells and reporter constructs to detect developmentally regulated genes
.
Science
244
,
463
465
.
Goulding
,
M. D.
,
Chalepakis
,
G.
,
Deutsch
,
U.
,
Erselius
,
J. R.
and
Gruss
,
P.
(
1991
).
Pax-3, a novel murine DNA binding protein expressed during early neurogenesis
.
EMBO J
.
10
,
1135
1147
.
Grainger
,
R. M.
and
Gurdon
,
J. B.
(
1989
).
Loss of competence in amphibian induction can take place in single nondividing cells
.
Proc. Natl. Acad. Sci. USA
86
,
1900
1904
.
Gurdon
,
J. B.
(
1987
).
Embryonic induction - molecular prospects
.
Development
99
,
285
306
.
Gubler
,
U.
and
Hoffman
,
B. J.
(
1983
).
A simple and very efficient method for generating cDNA libraries
.
Gene
25
,
263
269
.
Hamburger
,
V.
(
1988
).
The Heritage of Experimental Embryology. Hans Spemann and the Organizer
.
Oxford
:
Oxford University Press
.
Jones
,
E. A.
and
Woodland
,
H. R.
(
1989
).
Spatial aspects of neural induction in Xenopuslaevis
.
Development
107
,
785
791
.
Jones-Villeneuve
,
E. M. V.
,
McBurney
,
M. W.
,
Rogers
,
K. A.
and
Kalnins
,
V. I.
(
1982
).
Retinoic acid induces embryonal carcinoma cells to differentiate into neurons and glial cells
.
J. Cell Biol
.
94
,
253
262
.
Jones-Villeneuve
,
E. M. V.
,
Rudnicki
,
M. A.
,
Harris
,
J. F.
and
McBurney
,
M. W.
(
1983
).
Retinoic acid-induced neural differentiation of embryonal carcinoma cells
.
Mol. Cell Biol
.
3
,
2271
2279
.
McBurney
,
M. W.
,
Jones-Villeneuve
,
E. M. V.
,
Edwards
,
M. K. S.
and
Anderson
,
P. J.
(
1982
).
Control of muscle and neuronal differentiation in a cultured embryonal carcinoma cell line
.
Nature
299
,
165
167
.
McBurney
,
M. W.
and
Rogers
,
B. J.
(
1982
).
Isolation of male embryonal carcinoma cells and their chromosome replication patterns
.
Dev. Biol
.
89
,
503
508
.
McBurney
,
M. W.
,
Reuhl
,
K. R.
,
Ally
,
A. J.
,
Nasipuri
,
S.
,
Bell
,
J. C.
and
Craig
,
J.
(
1988
).
Differentiation and maturation of embryonal carcinoma-derived neurons in cell culture
.
J. Neurosci
.
8
,
1063
1073
.
McMahon
,
A. P.
and
Bradley
,
A.
(
1990
).
The wnt-1 (int-1) protooncogene is required for development of a large region of the mouse brain
.
Cell
62
,
1073
1085
.
Mielnicki
,
L. M.
and
Pruitt
,
S. C.
(
1991
).
Isolation and nucleotide sequence of a murine cDNA homologous to human activating transcription factor 4
.
Nuc. Acids Res
.
19
,
6332
.
Mitani
,
S.
and
Okamoto
,
H.
(
1991
).
Inductive differentiation of two neural lineages reconstituted in a microculture system from Xenopus early gastrula cells
.
Development
112
,
21
31
.
Nieuwkoop
,
P. D.
,
Johnen
,
A. G.
and
Albers
,
B.
(
1985
). The induction of the neural plate. In
The Epigenetic Nature of Early Chordate Development
pp
150
-
162
. Cambridge: Cambridge University Press.
Ochman
,
H.
,
Medhora
,
M. M.
,
Garza
,
D.
and
Hartl
,
D. L.
(
1990
). Amplification of flanking sequences by inverse PCR. In
PCR Protocols: A Guide to Methods and Applications
(ed.
M. A.
Innis
,
D. A.
Gelfand
,
J. J.
Sninsky
and
T. J.
White
) pp
219
-
227
. San Diego: Academic Press.
Pearson
,
W. R.
and
Lipman
,
D. J.
(
1988
).
Improved tools for biological sequence comparison
.
Proc. Natl. Acad. Sci. USA
85
,
2444
2448
.
Placzek
,
M.
,
Yamada
,
T.
,
Tessier-Lavigne
,
M.
,
Jessell
,
T.
and
Dodd
,
J.
(
1991
).
Control of dorsoventral pattern in vertebrate neural development: induction and polarizing properties of the floor plate
.
Development
Supplement
2
,
105
122
.
Pruitt
,
S. C.
(
1988
).
Expression vectors permitting cDNA cloning and enrichment for specific sequences by hybridization/selection
.
Gene
66
,
121
134
.
Pruitt
,
S. C.
and
Natoli
,
T. A.
(
1992
).
Inhibition of differentiation by leukemia inhibitory factor distinguishes two induction pathways in P19 embryonal carcinoma cells
.
Differentiation
50
,
57
65
.
Puissant
,
C.
and
Houdebine
,
L.
(
1990
).
An improvement of the single-step method of RNA isolation by acid guanidinium thiocyante-phenol-chloroform extraction
.
BioTech
.
8
,
148
149
.
Robertson
,
E. J.
(
1987
). Embryo-derived stem cell lines. In
Teratocarcinomas and Embryonic Stem Cells a Practical Approach
(ed.
E. J.
Robertson
) pp
71
-
112
. Oxford:
IRL Press
.
Rosner
,
M. H.
,
Vigano
,
M. A.
,
Ozato
,
K.
,
Timmons
,
P. M.
,
Poirier
,
F.
,
Rigby
,
P. W. J.
and
Staudt
,
L. M.
(
1990
).
A Pou-domain transcription factor in early stem cells and germ cells of the mammalian embryo
.
Nature
345
,
686
692
.
Rudnicki
,
M. A.
and
McBurney
,
M. W.
(
1987
). Cell culture methods and induction of differentiation of embryonal carcinoma cell lines. In
Teratocarcinomas and Embryonic Stem Cells, A Practical Approach
(ed.
E.
Robertson
) pp
19
-
49
. Oxford:
IRL Press
.
Ruiz i Altaba,
A.
(
1990
).
Neural expression of the Xenopus homeobox gene Xhox3: evidence for a patterning neural signal that spreads through the ectoderm
.
Development
108
,
595
604
.
Scholer
,
H. R.
,
Hatzopoulos
,
A. K.
,
Balling
,
R.
,
Suzuki
,
N.
and
Gruss
,
P.
(
1989
).
A family of octamer-specific proteins present during mouse embryogenesis: evidence for germline specific expression of an Oct factor
.
EMBO J
.
8
,
2543
2550
.
Schuuring
,
E.
,
van Deemter
,
L.
,
Roelink
,
H.
and
Nusse
,
R.
(
1989
).
Transient expression of the proto-oncogene int-1 during differentiation of P19 embryonal carcinoma cells
.
Mol. Cell. Biol
.
9
,
1357
1361
.
Servetnick
,
M.
and
Grainger
,
R. M.
(
1991
).
Changes in neural and lens competence in Xenopus ectoderm: evidence for an autonomous developmental timer
.
Development
112
,
177
188
.
Sive
,
H. L.
,
Hattori
,
K.
and
Weintraub
,
H.
(
1989
).
Progressive determination during formation of the antereoposterior axis in Xenopus laevis
.
Cell
58
,
171
180
.
Skarnes
,
W.C.
,
Auerbach
,
B.A.
and
Joyner
,
A.L.
(
1992
).
A gene trap approach in mouse embryonic stem cells: the lacZ reporter is activated by splicing, reflects endogenous gene expression, and is mutagenic in mice
.
Genes Dev
. in press.
Smith
,
J. C.
(
1989
).
Mesoderm induction and mesoderm-inducing factors in early amphibian development
.
Development
105
,
665
677
.
Smith
,
J. C.
,
Price
,
B. M. J.
,
Green
,
J. B. A.
,
Weigel
,
D.
and
Herrmann
,
B. G.
(
1991
).
Expression of a Xenopus homolog of Brachyury (T) is an immediate-early response to mesoderm induction
.
Cell
67
,
79
87
.
Smith
,
S. C.
,
Reuhl
,
K. R.
,
Craig
,
J.
and
McBurney
,
M.
(
1987
).
The role of aggregation in embryonal carcinoma cell differentiation
.
J. Cell Physiology
131
,
74
84
.
St-Arnaud
,
R.
,
Craig
,
J.
,
McBurney
,
M. W.
and
Papkoff
,
J.
(
1989
).
The int-1 proto-oncogene is transcriptionally activated during neuroectodermal differentiation of P19 mouse embryonal carcinoma cells
.
Oncogene
4
,
1077
1080
.
Thomas
,
K. R.
and
Capecchi
,
M. R.
(
1990
).
Targeted disruption of the murine int-1 proto-oncogene resulting in severe abnormalities in midbrain and cerebellar development
.
Nature
346
,
847
850
.
Wigler
,
M.
,
Pellicer
,
A.
,
Silverstein
,
S.
,
Axel
,
R.
,
Urlaub
,
G.
and
Chasin
,
L.
(
1979
).
DNA-mediated transfer of adenine phosphoribosyltransferase locus into mammalian cells
.
Proc. Natl. Acad. Sci. USA
76
,
1373
1376
.
Yamada
,
T.
,
Placzek
,
M.
,
Tanaka
,
H.
,
Dodd
,
J.
and
Jessell
,
T. M.
(
1991
).
Control of cell pattern in the developing nervous system: polarizing activity of the floor plate and notochord
.
Cell
64
,
635
647
.