Establishment of a differentiated mesodermal line from P19 EC cells expressing functional PDGF and EGF receptors

Experimental

Cell Research

Establishment of a Differentiated PI9 EC Cells Expressing Functional

165 (1986) 229-242

Mesodermal Line from PDGF and EGF Receptors

C. L. MUMMERY,* A. FEIJEN, W. H. MOOLENAAR, C. E. van den BRINK and S. W. de LAAT Hubrechl

Laboratory, International Embryological 3584 CT Utrechr. The Netherlands

Institute,

Aggregation of pluripotent PI9 embryonal carcinoma (EC) cells in the presence of DMSO induces differentiation to various mesodermal cell types, including spontaneously contracting muscle [9]. We have established clonal cell lines from these cultures and characterized one (MES-1) in particular for its response to growth factors. In contrast to the undifferentiated stem cells, but as a number of myoblast and muscle cell lines, MES-1 cells respond to both carbachol and bradykinin by the rapid release of Ca*+ from intracellular stores. In addition, MES-1 express receptors for and respond mitogenically to epidermal growth factor (EGF) and platelet-derived growth factor (PDGF). Isolated membranes from these cells retain the capacity to bind both ligands; addition of EGF to membranes induces endogenous phosphorylation of several proteins, including the EGF receptor itself and a 38 kD protein, while addition of PDGF specifically induces phosphorylation of the PDGF receptor. By contrast, other derivatives of P19, isolated from retinoic acid (RA)-treated aggregates and resembling neuroectodermal or endodermal cell types [lo] respond only to EGF; PDGF neither binds nor induces phosphorylation and a mitogenic response in these cells. During differentiation from EC cells therefore MES-I cells developed a combination of growth factor receptor characteristics typical of somatic mesodermal cells and indicate that such receptors on EC-derived mesodermal cells are also functional. @ 1986 Academic Press, Inc.

The development of a multicellular organism requires that cellular proliferation and diversification be strictly regulated both in time and in space. In early murine development, interactions between individual cells and groups of cells are thought to play an important role in this regulation. Among those interactions already identified are (1) the formation of specialized structures between cells such as gap junctions [I, 21; (2) the production of specific diffusible signal molecules, such as polypeptide growth factors, by certain cell types in the developing embryo [3-51; and (3) short-range interactions mediated by the production of extracellular matrix components [6]. Mouse embryonal carcinoma (EC) cells, the undifferentiated stem cells of teratocarcinoma, are frequently used as a model system to study early murine development because of the many properties they have in common with the pluripotent cells of the embryo itself [7, 81. EC cells can frequently be induced * To whom offprint requests should be sent. Address: Hubrecht Laboratory, International logical Institute, Uppsalalaan 8, 3585 CT Utrecht, The Netherlands.

Embryo-

Copyright @ 1986 by Academic Press, Inc. All rights of reproduction in any form reserved 0014~4827iE6 $0303.00

230

Mummery

et al.

under controlled conditions in culture to form derivatives of all three germ layers. The P19 EC line for example can be induced to form neuroectodermal and endodermal derivatives upon aggregation with retinoic acid (RA) and mesodermal derivatives, including spontaneously contracting muscle cells upon aggregation in the presence of dimethylsulfoxide (DMSO) [93. We have used these cells as a model to investigate differential growth regulation during differentiation in early development, and in particular to establish whether growth factor receptor expression is dependent on cell lineage. We have isolated various clonal, differentiated derivatives of P19 EC cells and characterized growth factor receptor expression and response, in particular that of EGF, PDGF and insulin. Previously we have shown that two clones derived from RA-treated aggregates expressed functional receptors for EGF, with induction of phosphorylation on tyrosine residues following EGF binding and a subsequent mitogenic response [lo]. These clones did not, however, respond mitogenically to PDGF. In the present study we have isolated a clonal derivative from aggregates of P19 EC cells treated with DMSO. This cell line (MES-1) exhibits several general properties of mesodermal cells, including expression of receptors for and response to PDGF, in addition to EGF. Furthermore, the apparent presence of muscarinic and bradykinin receptors on its surface suggests this line may represent an ECderived myoblast. As such, MES-1 is the first derivative of an EC line in which phosphorylation of the PDGF receptor has been described. The implications of differential growth factor receptor expression for growth regulation during early development are further discussed.

MATERIALS Cell Culture

Isolation

of Differentiated

AND

METHODS

(Mesodermal)

Clones

P19 EC cells were cultured as described previously [lo, 121 on gelatinized flasks in a 1 : 1 mixture of Dulbecco’s Minimum Essential Medium and Ham’s F12 medium (DF) containing 7.5% fetal calf serum (FCS) (Flow Laboratories) and buffered with NaHC03 (44 mM) in a 7.5 % CO2 atmosphere. Cells were subcultured after treatment with 0.125% w/v trypsin, 50 mM ethylenediaminetetraacetic acid (EDTA) in Ca’+-, Mg*+-free phosphate-buffered saline (PBS). To induce differentiation to mesodermal tissues, cells were aggregated (by plating 2x 10’ cells/2 ml in bacteriological Petri dishes) in the presence of 1% DMSO for 5 days, then replated in tissue-culture grade plastic dishes [9]. After 7 days, cultures consisted of a heterogeneous mixture of fibroblast-like cells with areas of rhythmically contracting muscle; no EC cells were visible. These cultures were cloned by diluting cell suspensions to a density of 1 cell/ml then plating in a 96-well plate (Costar; 200 @/well). Wells were checked for the presence of single cells. Three weeks later, nine wells contained confluent cultures which were subcultured to increasingly larger culture flasks. Subsequently cells were routinely cultured using a split ratio 1 : 5. One of these clonal lines (MES-1) was selected for detailed study in view of its morphological resemblance to tibroblasts at high cell density, with the cells tending to form linear arrays at confluence. Isolation and characterization of the EPI-7 and END-2 cell lines has been described in detail previously [lo]. In brief, these differentiated clones were obtained from a mixture of neuroectodermal and endodermal derivatives of P19, formed 2 weeks after aggregation of EC cells in the presence of all-trans-retinoic acid (RA) (5x IO-’ M) for 3 days and replating in tissue-culture grade plastic. Exp Cell

Res 165 (1986)

PDGF Cellular

DNA

and EGF receptor

Synthesis

expression

in differentiated

EC cells

231

Assay

Cells were plated at 5x lo4 cells per 1 cm diameter well in DF+7.5 % FCS. After 24 h, medium was replaced by DF+0.5 % FCS. EGF (receptor grade; BRL), insulin (bovine; Sigma), PDGF (purified from human platelets; a gift from C.-H. Heldin) or 10% FCS were added to quiescent cultures 72 h later without change of medium, as previously described [IO]. After 8 h, [3H]TdR (0.5 l&i/ml, 55 Ci/mmole; Amersham) was added and cells incubated a further 16 hat 37°C.

Soft Agar Growth The percentage of cells able to form colonies in soft agar was determined by plating 1x lo4 cells in DF+7.5 % FCS with 0.375 % agar onto a base layer of 0.5 % agar in 60 mm Costar Petri dishes, as described by Todaro et al. [13]. Colonies larger than eight cells were scored after 14 days by counting 20 random fields corresponding to 2.3 % of the total dish area. Data are expressed as the percentage of cells plated able to form colonies.

Measurement of [Ca’+]i Confluent MES-1 cells attached to rectangular coverslips coated with gelatine were loaded with the fluorescent Ca*’ indicator quin-2 by incubating them in HEPES-buffered DMEM (pH 7.4) containing 15 uM quin-2 acetoxymethylester for 45 min at 37°C. Fluorescence measurements and calibration procedures were essentially similar to those described previously [33, 341. Quin-2 acetoxymethylester was from Calbiochem and was stored as a 20 mM stock in DMSO at -20°C.

Growth Factor Binding Assays To cells in monolayer. 0.5x IO6 cells were plated in 3.5 cm gelatinized dishes (Costar) and grown for 24 h before use. The binding assay buffer contained DMEM, HEPES (25 mM, pH 7.4) and 0. I % w/v bovine serum albumin (BSA), as previously [lo, 141. For EGF binding, cells were incubated with [“‘I]EGF (1 @ml; 250000 cpm; Amersham) in 1 ml binding buffer and varying amounts of unlabelled EGF (from zero to 200 @ml). For PDGF binding, cells were incubated with [‘251]PDGF (4 @ml; 100000 cpm) in 0.5 ml binding buffer and varying amounts of unlabelled PDGF (from zero to 130 rig/ml). Labelled and unlabelled PDGF was a gift from C.-H. Heldin, Sweden. Cells were incubated for 2 h at room temperature (ca 2O”C), the binding medium aspirated and dishes washed live times with ice-cold phosphate-buffered saline (PBS). Cells were solubilized with 0.5 M NaOH and extracts counted in a y-counter. The data points shown are the mean of values in 3-fold. For EGF, non-specific binding was determined by measuring the bound counts in the presence of a large excess (200-fold over [‘*‘I]EGF of unlabelled EGF. This was less than 10% of the specifically bound counts. Binding kinetics were analysed initially by standard Scatchard analysis, as previously [lo]. For PDGF, the availability of unlabelled PDGF was limiting so that the maximum excess possible was only 30.fold over [‘2SI]PDGF. Binding kinetics were therefore analysed by a modification of standard Scatchard analysis (van Zoelen, to be published elsewhere). Cross-linking ro isolated cell membranes. Membranes isolated according to Thorn et al. [15] from MES-1 cells (100 ng protein per sample) were incubated with [‘251]EGF (50 @ml; 1~10~ cpm) or [‘*‘I]PDGF (40 rig/ml; 25x IO5 cpm) for 60 min at 37°C as described previously [IO]. Excess EGF (5 pg/ml) or PDGF (2 ug/ml) was added as required. Labelled growth factor then bound was cross-linked to its receptor with ethyl-3(3-dimethylaminopropyl-carbodiimide HCl (EDAC; 1.0 mM) [30] for 15 min at 15°C; the cross-linker was quenched with 900 ul of 0.01 M Tris-HCl/l mM EDTA (pH 7.4). the sample centrifugated for 45 min at 100000 g, resuspended in SDS sample buffer and incubated for 3 min at 100°C. Samples were then subjected to SDS gel electrophoresis under non-reducing conditions (5-15 % polyacrylamide gel) and autoradiography.

Endogenous Phosphorylation

of Membrane Proteins

Membranes (50 pg protein/sample) isolated from PI9 EC cells, MES-I and the differentiated line END-2 derived previously [lo] were preincubated with EGF and PDGF (1.67 kg/ml) for 10 min at 0” or 37°C. as indicated. Phosphorylation was started by adding membranes to a mixture containing Exp

Cell

Res

165 (1986)

232 Mummery

et al.

1 . n orphology of a mesodermal derivative of P19 EC cells (MES-I) compared with the undiffe :rer tiated stem cell. (a) MES-1 cells at confluence. Formation of typical fibroblas ;tic Ii1near arrays. (b Undifferentiated P19 EC cells. Growth in ‘islands’; large nuclear: cytoplasm rat io. Fig.

Exp Ce 11 Res 165 (1986)

PDGF

Table 1. Growth

and EGF receptor properties

expression

in differentiated

of P19 EC cells compared

EC cells

with differentiated

233

deriva-

tives

Differentiation Generation time (h) Soft agar growth (% colonies) EGF binding (receptors/cell) PDGF binding (receptors/cell) Growth factor requirement Mitogenic response to” FCS EGF Insulin PDGF

EC

END-2

EPI-7

MES-I

14.3 29.0 900

Aggr. + RA 25.0 <2 6CL78 000

Aggr + RA 19.3 <2 3942 000

-

-t

-

Aggr. + DMSO 24.0 <2 12 000 (high affinity) 180 000 (low affinity) 115 000 +

ND 1.0 ND ND

1.7 4.0 2.5 1.0

1.3 2.6 I.0 1.0

8.6 6.3 2.1 1.4

” Cumulative [3H]TdR incorporation-units relative to control. ND, not determined. Characteristics of MES-I cells are derived from data of the present study, while those of P19 EC, END-2 and EPI-7 are largely summarized from Mummery et al. [lo].

HEPES (20 mM, pH 7.4), MnClz (2 mM), [Y-~*P]ATP (10 PM; 5000 cpm/pmol; Amersham), BSA (0.125 mg/ml), NajV04 (50 FM) and PNPP (p-nitrophenylphosphate, 10 mM) in a total volume of 60 pl. After 10 min at O”C, the reaction was stopped by addition of SDS sample buffer and raising the temperature to 100°C for 5 min. Samples were then subjected to SDS gel electrophoresis under reducing conditions (S-15 % polyacrylamide) and autoradiography.

RESULTS Growth Properties of Mesodermal P19 Derivatives

Aggregates of PI9 EC cells treated with DMSO then replated on an adhesive substrate differentiate to a variety of mesodermal cell types including cardiac and skeletal muscle tissue [16]. We have derived homogenous cell lines from single cells cloned from these mixed cultures and characterized one (MES-I) in detail for its growth properties and response to exogenous growth factors and hormones. Morphologically the cells resemble fibroblasts, becoming contact-inhibited at confluency and forming typical fibroblastic linear arrays at high densities (fig. 1). The generation time (24.2k2.1 h) determined by cell counting is significantly longer than that of the undifferentiated parent cell (14.3* 1.0 h) [lo] but comparable with that of other differentiated derivatives of EC cells (table 1) [lo, 121. The ability to grow in semi-solid media in vitro has been highly correlated with the ability to form tumours; the failure of MES-1 cells to grow in soft agar, in contrast to the undifferentiated EC cells (table 1) indicates the loss of tumorigenic potential during differentiation to mesoderm. MES-1 cells are, however, immorExp

Cell

Res 165 (1986)

234

Mummery

et al.

brk

Fig. 2. Effects of carbachol and bradykinin on [Ca2’li as measured by quin-2 fluorescence. Response of quin-2-loaded monolayers to additions of carbachol (100 PM) and bradykinin (2 PM). [Ca’+]; was calculated as described previously [32, 331. Prior to quin-2 loading, the cells were incubated in serum-free DMEM for 1 h.

tal in culture and have been passaged for more than 9 months without phenotypic changes. Response to Neuropeptides

A number of known Ca*+ mobilizing neurotransmitters were tested for their ability to elicit a [Ca21i rise in quin-2-loaded monolayers. A rapid but transient Ca*+ signal was evoked by carbochol (maximum response at 100 PM) and by bradykinin (maximum response at 2 uM) (fig. 2). These Ca*+ transients were not prevented by removal of external Ca*+ using EGTA (not shown), indicating that carbachol and bradykinin trigger the release of Ca*+ from internal stores. The carbachol response appears to be mediated by the muscarinic receptor, as atropine was found to be a potent inhibitor (not shown). In the apparent presence of muscarinic and bradykinin receptors, characteristic of myogenic cells on their surface, MES-1 cells therefore strongly resemble myoblasts rather than tibroblasts. However, under conditions where myotube formation has been observed in both teratocarcinoma-derived and primary myoblast culture [34, 351 we do not observe fusion in MES-1 cells. Response to Growth Factors: PDGF

Addition of PDGF to cultures made quiescent at high density under reduced serum concentrations (0.5 % FCS) induces an increase in [3H]TdR incorporation approx. 6-fold at a concentration of 10 rig/ml (fig. 3). In the addition presence of insulin (10 rig/ml), itself only marginally mitogenic at this concentration, a 9-fold stimulation above control is induced; synergism between PDGF and insulin is particularly striking at suboptimal PDGF concentrations (fig. 3; 0.5 r&ml PDGF and 10 rig/ml insulin). The 9-fold stimulation is equivalent to that induced by 7.5% FCS (fig. 3) and may represent the maximum attainable under these assay conditions. Under slightly different assay conditions, where quiescence is induced by maintenance of cultures at confluence in medium with 7.5% FCS for several days prior to transfer to serum-free conditions, and growth factors are added for 24 h with [3H]TdR present only in the last 6 h, significantly greater stimulation is observed. FCS then stimulates approx. 70-fold and, in order of potency, FCS>EGF (35-fold)>PDGF (1Zfold) (not shown). However, in the complete absence of serum as under these assay conditions, MES-1 cells acquire a more rounded morphology; since the significance of these changes in terms of Exp Cell

Res 165 (1986)

PDGF

and EGF receptor expression in differentiated

r IES-11

EC cells

235

l-

J

Fig. 3. Effect of growth factors on DNA synthesis in MES-1 cells. Confluent cultures were preincubated in DF+0.5% FCS for 72 h before growth factor addition for 24 h with L3H]TdR present for the last 16 h, as described in Materials and Methods. Increases in cell number paralleled those in DNA synthesis (not shown). Concentrations are given in &ml.

the state of differentiation is at present unclear and the serum factor(s) involved unidentified, the data illustrated in fig. 3 were considered more representative and relevant to subsequent characterization. In all cases, stimulation of [3HlTdR incorporation was followed by cell division (not shown). The response of MES-1 cells to PDGF contrasts with that found previously in the other differentiated derivatives of P19, END-2 and EPI-7, where under similar conditions PDGF has no effect on [3H]TdR incorporation even in the presence of insulin [lo]. Further, it may partially explain why FCS, which contains approx. 10-15 rig/ml PDGF (see ref. [ 171) is a relatively poor mitogen for these cell lines while inducing a maximum response in MES-1 cells. Response

to Growth Factors:

EGF

DNA synthesis can also be stimulated in MES-1 cells by EGF; as for PDGF, insulin acts synergistically increasing [3H]TdR incorporation &9-fold, equivalent to the maximum induced by 7.5% FCS under the same conditions (fig. 3). Mitogenic activity of EGF in various differentiated EC cells has been reported previously [5, 10, 111 but, as standard assay systems for growth factors, these cells are inconvenient as they have to be obtained on each occasion by stem cell differentiation. Since MES-1 are easily induced to quiescence by serum deprivaExp

Cell

Res 165 (1986)

236 Mummery

et al.

LO IPDGFlo

60

I-J’

80

‘“g/m”

t:

0 t

P,

‘;

a w

s

Fig. 4. PDGF binding to P19 EC cells and their differentiated derivatives.

(A) Specific PDGF binding as a function of external PDGF concentration to MES-1 cells at equilibrium (2 h at 20°C; time curves under which these conditions were established are not shown). Inset shows data expressed as in Scatchard plot; the experiment illustrated is one of three giving essentially similar results. (E) Specific PDGF binding to MES-1 cells compared with the undifferentiated EC cells and the differentiated clones END-2 and EPI-7. Confluent cultures were incubated for 2 h at 20°C with [‘251]PDGF (4 rig/ml) in 0.5 ml binding buffer. Non-specific binding was determined as that in the presence of excess unlabelled PDGF (400 @ml).

tion at confluence and are sensitive to EGF, PDGF and insulin, they may provide a more suitable target system for embryonic growth factors. Growth Factor Binding

to Cells in Monolayer:

PDGF

Specific receptors for PDGF could be detected on MES-1 and the cells exhibited saturation binding kinetics, as shown in fig. 4A. Analysis of binding data (fig. 4A, inset, shows data expressed in the form of a Scatchard plot), indicated 1.15~ 10’ receptors per cell with a dissociation constant kd of 0.3 nM. This compares favourably with receptor levels in several somatic cell lines of mesodermal origin in culture such as NRK which express 0.4-0.8~ lo5 receptors/cell [15, 191. In the undifferentiated EC cells and the differentiated lines END-2 and EPI-7 [lo] specific PDGF binding was low relative to MES-1 or undetectable (fig. 4B). Growth Factor Binding

to Cells in Monolayer:

EGF

Like EPI-7 and END-2, MES-1 cells showed specific EGF binding with saturable binding kinetics, as shown in fig. 5. Scatchard analysis (fig. 5, inset), however, revealed two classes of binding site, in contrast to one site observed in EPI-7 and END-2 [lo]. Binding in each cell type was measured under identical conditions, suggesting that the two classes were not an artefact of the methodology, e.g. a difference in the affinities of labelled and unlabelled EGF for the receptor. The two receptor classes were resolved using the LIGAND programme written by Munson & Rodbard [20] modified to Applesoft by M. H. Teicher. Approx. 12000 high-affinity receptors per cell were detected (kd, 0.03 nM) and 180000 Exp

Cell

Res 165 (1986)

PDGF

and EGF receptor expression in differentiated

? 20

60

40 [EGFI,

1 rig/ml

EC cells

237

Fig. 5. EGF binding to MES-I cells. Specific EGF binding as a function of external EGF concentration to MES-1 cells at equilibrium (2 h at 20°C). The experiment illustrated in one of three giving essentially similar results. Inset shows data expressed as a Scatchard plot.

200

I

low-affinity sites (kd, 12.3 nM). Thus P19 EC cells, which do not bind detectable amounts of EGF [lo], develop significant receptor expression following differentiation to both mesodermal and neuroectodermal/endodermal derivatives. Growth Factor Binding. to Isolated

Cetl Membranes:

PDGF

and EGF

Membranes isolated from MES-1 retain the capacity to bind both PDGF and EGF specifically, as shown in fig. 6. ‘251-labelled growth factor bound to membranes in the absence and presence of excess unlabelled was cross-linked to the receptor prior to SDS gel electrophoresis. Bands representing the EGF receptor appear as a doublet at 155 and 165 kD, while the PDGF appears at a higher molecular weight (MW) as a single band at 185 kD. The two receptors are therefore clearly distinguishable by MW, although labelled PDGF receptor appears at slightly lower MW than expected (the PDGF receptor has a MW of approx. 180 kD and together with the weight of 35 kD for PDGF the cross-linked receptor would be expected at 210 kD). PDGF-

and EGF-induced

Phosphorylation

In order to establish functional consequences of growth factor receptor binding, MES-1 cells were compared with P19 EC and END-2 for the effect of PDGF and EGF in protein phosphorylation. Preincubation of membranes isolated from MES-1 cells with PDGF or EGF followed by [Y-~~P]ATP induced distinctive changes in the phosphorylation of endogenous substrates (fig. 7). With EGF at 0°C autophosphorylation of the EGF receptor (180 kD) and phosphorylation of a 38 kD protein were particularly significant, with a slight increase in phosphorylation of a 23 kD band (fig. 7 c). Similar changes were observed in membranes from END-2 (fig. 7 b), but not from the undifferentiated EC cells (fig. 7a), in agreeExp Cell Res 165 (1986)

238

Mummery

et al.

6

6. [‘2SI]PDGF and [“‘I]EGF binding to isolated plasma membranes from MES-1. Lanes: I, 2, iZ51-PDGF in the absence and presence of excess unlabelled PDGF respectively; 3, 4, [‘251]EGF in the absence and presence of excess unlabelled EGF respectively. Bound ‘251-labelled ligand was crosslinked to its receptors using EDAC prior to SDS gel electrophoresis (5-15% gradient under nonreducing conditions), as described under Materials and Methods. Fig. 7. Protein phosphorylation of isolated cell membranes by PDGF and EGF. Isolated membranes were preincubated with EGF or PDGF (1.67 &ml) at 0” or 37°C as indicated. (a) P19 EC, 0°C; (b) END-2, 1°C; (c) MES-1, 0°C. In each case, lane I is the control without growth factor addition, lane 2 with EGF and lane 3 with PDGF. (6) MES-1,37”C without and with PDGF (lanes 1 and 2 respectively). SDS gel electrophoresis was carried out under reducing conditions (5-15 % polyacrylamide). Fig.

ment with previous results [lo]. However, there are distinct differences between the various cell types in the basic pattern of phosphorylation prior to induction. Particularly noticeable is a band at 36 kD found exclusively in END-2 cells. In contrast to EGF, PDGF induced phosphorylation of only one band, at 180 kD, detectable following PDGF preincubation at 0°C (fig. 7c), but optimal following preincubation at 37°C (fig. 74. This band is most probably the PDGF receptor. Others have shown that PDGF induces phosphorylation of a band at the same level as the EGF receptor in human tibroblasts [21] and that this indeed represents autophosphorylation of the PDGF receptor. No changes in phosphorylation were induced by PDGF in membranes isolated from the undifferentiated EC cells (fig. 7 a) or END-2 cells (fig. 7 6) either at 0°C as illustrated or at 37°C (not shown). This was expected on the basis of binding studies to cells in monolayer and provides indirect evidence against any possibility that PDGF is contaminated by EGF or that non-specific phosphorylation of the EGF receptor is induced. Direct evidence that the receptors are indeed distinct in MES-1 cells Exp Cell

Res 165 (1986)

PDGF

and EGF receptor expression in differentiated

had already been provided by the cross-linking isolated plasma membranes.

EC cells

239

of [‘251]EGF and PDGF directly to

DISCUSSION In the present study we isolated differentiated cell lines from a heterogeneous culture of PI9 cells induced to form mesodermal derivatives by aggregation with DMSO. At the time of cloning, the cultures consisted of a mixture of fibroblastlike cells and areas of spontaneously beating muscle, as described by McBurney et al. [9]; one of these cell lines (MES-1) was characterized in detail. As many cells of somatic mesodermal origin, MES-I cells express specific PDGF receptors which are phosphorylated upon ligand binding; subsequently PDGF induces a mitogenic response, evidenced by an increase in DNA synthesis. This contrasts with the effects of PDGF on the undifferentiated EC cells and several of the differentiated derivatives of other germ layers (END-2 and EPI-7), where there is neither mitogenic response to PDGF nor significant levels of binding. The biological action spectrum found in PI9 EC derivatives thus reflects that found in somatic cells both in culture, where PDGF is mitogenic for fibroblasts and other connective tissue cells (although also for glial cells) but not for normal epithelioid cells and lymphocytes [31], and in the tissue distribution of PDGF receptors in vivo [22, 231. Relatively few examples of PDGF receptor expression in the differentiated derivatives of EC cells have been reported in the literature [24, 251 and there are, to our knowledge, no descriptions of phosphorylation induced by PDGF in these cells. The present study thus demonstrates that the PDGF receptors present on mesodermal derivatives of an EC line are functional and induce mitogenic stimulation of the target cell at quiescence. PDGF is a heat-stable polypeptide secreted by platelets and generally puritied from blood [26]. Its presence would therefore not be expected in an embryo during early development. However, evidence has been presented that at least some EC cells produce a PDGF-like growth factor [24, 271. This factor competes with [“‘I]PDGF for receptor binding and is mitogenic for appropriate target cells such as BALB/3T3 fibroblasts [24]. It may not necessarily be identical to PDGF however; Rizzino & Bowen-Pope [27] have shown, for example that significantly greater concentrations of anti-PDGF antibody were required to inhibit the PDGFlike activity produced by F9 cells than PDGF from platelets suggesting that, although similar, there are significant differences antigenically. The factor may, for example represent an embryonic form of PDGF, analogous to transforming growth factor a (TGFa), a possible, embryonic form of EGF [13]. Indeed, embryos themselves have been shown at different developmental stages to produce growth factor-like substances with properties similar to those of TGFs produced by a variety of chemically and virus-transformed cells [3-51. To what extent production of a PDGF-like factor by P19 EC cells, EPI-7 or END2 may contribute to the failure to detect signiticant [“‘I]PDGF binding in these 16-868337

Exp

Cell

Res 165 (1986)

240

Mummery

et al.

cells or may lead to an underestimate of the number of receptors on ME&l cells, should they produce as well as bind PDGF-like factors, is at present not known. In addition to functional PDGF receptor expression, ME&l cells also have functional EGF receptor expression. Others have shown that EGF binding is not detectable on undifferentiated EC cells but that after differentiation with RA there is significant and irreversible EGF receptor expression [5, 10, 111. It is known, however, that RA itself can induce a reversible increase of EGF binding in non-EC cells which do not differentiate as a result of RA treatment [28]. Even in EC cells differentiated by low-density plating, where EGF receptors are already detectable, binding can still be further increased by RA [ 111. The relatively high levels of functional EGF receptors in MES-1 cells, which have not been exposed to RA, provide additional evidence that increased EGF binding observed after differentiation of EC cells is not a direct effect of RA. The exact nature of MES-1 cells in terms of the cell type they might represent during normal development has not as yet been clearly established. At low density they resemble fibroblasts morphologically and linear arrays typical of these cells are formed at confluence. However, the presence of acetylcholine receptors on the cell surface cells is implicated by the rapid internal Ca*+ release induced by various neurotransmitters. This would be compatible with a myogenic cell type and strongly suggest that MES-1 cells represent an EC-derived myoblast cell line similar to that derived from C17SI EC cells [34]. The morphology acquired in defined, serum-free medium in preliminary studies again has supported this view. To date we have been unable to induce fusion of MES-1 cells under a variety of conditions; this may imply that, as a variety of myoblast cell lines [36], MES-1 have failed to retain the capacity for terminal differentiation in vitro. Fusion is not a prerequisite for the expression of several muscle-specific differentiated functions including acetylcholine receptors [35]. The expression of musclespecific markers including characterization of electrical properties is currently being investigated in order to clarify these points. Further, since the present study has indicated the growth factor receptor expression which might be expected with particular cell lineages, we are at present characterizing their relative rate of expression during the differentiation process itself and their localization in cell aggregates. Taken together, the results of the present and a previous study and (table 1) [ 101 have shown that depending on the nature of the differentiation step, EC cells can develop differential growth factor receptor expression and response to exogenous growth factors. In themselves, EC-derived differentiated cell lines such as P19 EPI-7, P19 END-2, PSA-5 and PYS-2 described previously [lo, 37,381 and MES1 described in the present study, may provide a useful alternative to conventional assay systems in the characterization of EC cell and embryo-derived growth factors. Specific embryonal growth factors may not be detected by fibroblasts normally used as targets for screening for growth-promoting activity. A mechanism of differential growth factor receptor expression and response, if present in a Exp Cell Res 16.5 (1986)

PDGF

and EGF receptor

expression

in differentiated

EC cells

241

developing embryo, would provide a means by which growth of specific cell populations could be selectively regulated. Further not only undifferentiated EC cells (e.g. [29]) but also their differentiated derivatives [lo, 37, 381 appear to produce growth factors, so that dynamic interaction and cooperativity between cells may form an essential regulatory mechanism throughout early development. We are greatly indebted to Dr C.-H. Heldin, University of Uppsala, Sweden, for providing both the ‘*‘I-labelled and unlabelled PDGF. We also thank Drs E. J. J. van Zoelen and P. T. van der Saag for critical reading of the manuscript and Dr J. Boonstra for use of the LIGAND programme.

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33.

34.

Ducibella, T, Albertini, D F, Anderson, E & Bigger% J D, Dev biol 45 (1975) 231. Goodall, H & Johnson, M H, Nature 295 (1982) 524. Nexo, E, Hollenberg, M D, Figueroa, A & Pratt, R M, Proc natl acad sci US 77 (1980) 2782. Proper, J A, Bjornson, C L & Moses, H L, J cell physiol 110 (1982) 169. Rizzino, A, Orme, L S & De Larco, J E, Exp cell res 143 (1983) 143. Leivo, I, Vaheri, A, Timpl, R & Wartiovaara, J, Dev biol 76 (1980) 100. Graham, C F, Concepts in mammalian embryogenesis (ed M I Sherman & C F Graham) p. 315. MIT, Cambridge, Mass. (1977). Martin, G, Science 209 (1980) 768. McBumey, M W, Jones-Villeneuve E M V, Edwards, M K S & Anderson, P J, Nature 299 (1982b) 165. Mummery, C L, Feijen, A, van der Saag, P T, van den Brink, C E & de Laat, S W, Dev biol 109 (1985) 402. Rees, A R, Adamson, E D & Graham, C F, Nature 281 (1979) 309. Mummery, C L, van den Brink, C E, van der Saag, P T & de Laat, S W, Dev biol 104 (1984) 297. Todaro, G J, De Larco, J E, Fryling, C, Johnson, PA & Spom, M B, J supramol str cell biochem 15 (1981) 287. Mummery, C L, van der Saag, P T Jr de Laat, S W, J cell biochem 21 (1983) 63. Thorn, D, Powell, A J, Lloyd, C W & Rees, D A, Biochem j 168 (1977) 187. Jones-Villeneuve, E M, McBumey, M W, Rogers, K A & Kalnins, V I, J cell bioI94 (1982) 253. Josephs, S F, Wong-Staal, F & Gallo, R L, Cancer surveys 3 (1984) 265. Owen, A J, Pantazis, P & Antoniades, H N, Science 225 (1984) 54. Huang, J J, Huang, S S & Deuel, T F, Cell 39 (1984) 79. Munson, P J & Rodbard, D, Anal biochem 107 (1980) 220. Heldin, C-H, Ek, B & Ronnstrand, L, J biol them 258 (1983) 10054. Heldin, C-H, Wasteson, A & Westermark, B, J biol them 257 (1982) 4216. Bowen-Pope, D F 62 Ross, R, J biol them 257 (1982) 5161. Gudas, L J, Singh, J P & Stiles, C D, Teratocarcinoma stem cells (ed L Silver, G R Martin & S Strickland) vol. 10, p. 229. Cold Spring Harbor conferences on cell proliferation (1983). Levine, R A, La Rosa, G J & Gudas, L J, Mol cell biol4 (1984) 2142. Heldin, C-H, Westermark, B & Wasteson, A, Proc natl acad sci US 76 (1979) 3722. Rizzino, A & Bowen-Pope, D F, Dev biol 110 (1985) 15. Jetten, A M, Jetten, M E R, Shapiro, S & Poon, J, Exp cell res 119 (1979) 289. Heath, J K & Isacke, C M, EMBO j 3 (1984) 2957. Grob, P M, Berlot, C H & Bothwell, M A, Proc natl acad sci US 80 (1983) 6819. Stiles, C D, Pledger, W J, van Wijk, J J, Antoniades, H N & Scher, C D, Cold Spring Harbor conferences on cell proliferation 6 (1983) 425. Moolenaar, W H, Tertoolen, L G J & de Laat, S W, J biol them 259 (1984) 8066. Moolenaar, W H, Aerts, R J, Tertoolen, L G J & de Laat, S W, J biol them 261 (1986) 279. Jakob, H, Buckingham, M E, Cohen, A, DuPont, L, Fiszman, M & Jacob, F, Exp cell res 114 (1978) 403. Exp Cell

Res

165 (1986)

242 35. 36. 37. 38.

Mummery

et al.

Merlie, J P & Gross, F, Exp cell res 97 (1976) 406. Richler, C & Yaffe, D, Dev biol 23 (1970) 1. Isacke, C M & Deller, M J, J cell physiol 117 (1983) 407. van Zoelen, E J J et al. In preparation.

Received September 5, 1985 Revised version received January 9, 1986

Exp Cell

Res 165 (1986)

F’rinted

in Sweden