[CANCER RESEARCH 46, 332-340. January 1986]

Clonal Variation in the Production of a Platelet-derived Growth Factor-like

Protein and Expression of Corresponding Receptors in a HumanMalignant Glioma1

Monica Nister,2 Carl-Henrik Heldin, and Bengt Westermark

Departments of Pathology [M. N., B. W.] and Medical and Physiological Chemistry [C-H. H.], University of Uppsala, Uppsala, Sweden

ABSTRACT

Two cell lines (U-343 MG and U-343 MGa) with different

phenotypic characteristics were established from the same human glioblastoma multiforme biopsy. Previous studies haveshown that a clonal derivative (CI 2) of the U-343 MGa lineproduces a PDGF-like growth factor. In the present investigationglioma PDGF production and 125I-PDGFbinding were found to

be differently expressed in U-343 MG, U-343 MGa, and U-343

MGa CI 2 cultures, providing evidence for a clonal variation inthese properties. In order to investigate this point further, severalclones were derived from low (23 clones) and high (30 clones)passage U-343 MGa cultures, as well as from U-343 MGa CI 2

cells (30 clones). The clones could be divided into 4 groupsaccording to morphology and growth pattern. A determinationof the amount of PDGF receptor competing activity in serum-

free conditioned media gave evidence for a clonal variation in theproduction of glioma PDGF, corresponding to 0-87 ng of au

thentic PDGF per ml. There was also a considerable range in125I-PDGF binding (0-44 fmol of tracer bound per 106 cells).

Scatchard plots performed on two clones confirmed the presence of saturable, high affinity PDGF receptors. High passagecultures were found to give rise to a higher number of highproducing clones than did low passage cultures. There appearedto be a negative correlation between production of glioma PDGFand binding of 125I-PDGF,probably due to the receptor blocking

activity of the endogenous growth factor. However, the presenceof clones, apparently devoid of both glioma PDGF productionand 125I-PDGFbinding, suggests a true clonal variation in these

two parameters. The growth rate in serum-free medium was

found to correlate fairly well to the extent of glioma PDGFproduction. Production of glioma PDGF was found to have amorphological correlate and be most prominent among clonesof "immature" looking, tightly growing cells. Clones that had large

star-shaped cells with some resemblance to normal glia-like cellsin culture were found to have a low production and a high 125I-

PDGF binding capacity.

INTRODUCTION

In recent years much attention has been focused on theputative role of tumor cell-derived growth factors as effectors in

neoplastic transformation. Strong support for this notion was

Received 5/17/85; revised 9/10/85; accepted 9/16/85.1This investigation was supported by grants from the Swedish Cancer Society

and the Swedish Society of Medical Sciences.1To whom requests for reprints should be addressed, at Department of Pathol

ogy, University Hospital, S-751 85 Uppsala, Sweden.

obtained from the amino acid sequence analysis of PDGF3 whichrevealed a virtual identity between the B-chain of PDGF and partof the protein product of the v-s/s oncogene, i.e., the transforminggene of simian sarcoma virus (1-3). More recently, experimentalevidence for the production and release of a PDGF-like factor bysimian sarcoma virus-transformed cells has been presented (4-7). The cellular homologue to the viral sis gene (c-s/s) encodes aprecursor polypeptide of the PDGF B-chain (8).

An involvement of the c-s/s gene in human neoplasia is suggested by the finding of a PDGF-like factor produced by a humanosteosarcoma cell line (U-2 OS) (9, 10), a cloned human glioma

line (11,12), and a human rhabdomyosarcoma cell line (13). Theexpression of c-s/s at a high frequency in human sarcoma and

glioma cell lines but not in normal human embryonic fibroblastsor carcinoma and melanoma cell lines has also been established(14).

The PDGF-like factor produced by the human clonal gliomaline U-343 MGa CI 2 has been partially characterized (12, 13).The factor is released to serum-free culture medium as a dimerof M, 31,000 which competes with 125I-PDGFfor binding to its

cellular receptor, has growth promoting activity, and is immuno-logically related to authentic PDGF. Specific binding of 125I-PDGFto U-343 MGa CI 2 cells is essentially lacking, possibly due to ablocking and down regulation of the receptor by the endoge-nously produced factor. Analogous results have been obtainedwith other growth factor-producing cell lines (15). Interestingly,when 125I-PDGFbinding and PDGF production were assessed

on the parental cell line (U-343 MGa from which CI 2 was derived),

it was found that they displayed PDGF receptors but had abarely detectable production of the PDGF-like factor. This finding

suggested that the wild type population could be heterogeneouswith regard to production of glioma PDGF.

Clones with varying growth factor synthesis and PDGF bindingshould be valuable tools for further investigations on the role ofPDGF production in human tumor cells. In the present investigation we have therefore derived several clones from the U-343MGa lines, using cultures of low and high passage levels, as wellas formerly cloned cells, as starting material. Both 125I-PDGF

binding and production of PDGF receptor competing activityshow a considerable clonal variation and may be independentlyexpressed parameters in these tumor cells. The extent of gliomaPDGF production correlated with the growth rate in serum-freemedium. Moreover, a high degree of glioma PDGF productionwas found predominantly in clones derived from cells of highpassage levels. Our data thus indicate that the growth factorproduction is of selective advantage for the glioma cells.

3The abbreviations used are: PDGF, platelet-derived growth factor; GFAP, glialfibnllary acidic protein; Eagle's MEM, Eagle's minimum essential medium; NCS,

newborn calf serum; PBS, 137 miw NaCI, 2.7 mw KCI, 8 mu Na2HPO4, 1.5 mmKHzPCv EGF, epidermal growth factor; HEPES, 4-(2-hydroxyethyl)-1 -piperazineethanesulfonic acid.

CANCER RESEARCH VOL. 46 JANUARY 1986

332

Research. on August 29, 2018. © 1986 American Association for Cancercancerres.aacrjournals.org Downloaded from

CLONAL VARIATION IN THE PRODUCTION OF GLIOMA PDGF

MATERIALS AND METHODS

Human Glioma Cell Lines. The establishment of cell lines from humanmalignant gliomas has been described (16, 17). By dividing the biopsyof a glioblastoma multiforme from a 64-year-old man into two parts, we

established two cell lines with different phenotypic properties. Onefraction, giving rise to U-343 MG (17, 18) was originally plated intountreated culture dishes. The other fraction, giving rise to U-343 MGa

(19), was plated onto a feeder layer of normal human glial cells. Bypropagating these cells through many passages, the glioma cells totallyoutnumbered the glial cells. To date, both cell lines have been propagatedthrough a number of passages (about 400). The phenotypic characteristics of the two U-343 MG sublines have been described (17-19). Theline U-343 MG is composed of spindle, fibroblast-like cells which arefibronectin positive and GFAP negative. Low passage cultures of the U-

343 MGa line have a pleomorphic appearance, whereas high passagecultures are more homogeneous, with small polygonal or short fusiformcells. This cell line, irrespective of passage level, is fibronectin negativeand 100% GFAP positive. Data on two clonal derivatives of the line U-

343 MGa (CI 2 and CI 3) have been published previously (20). Theseclones are morphologically homogeneous and have the GFAP positive,fibronectin negative phenotype.

Culture Conditions. Cells were grown routinely in Nunc 50-mm plasticPetri dishes in Eagle's MEM supplemented with 10% NCS (Grand Island

Biological Company) and antibiotics (100 units of penicillin and 50 ¿¿gofstreptomycin per ml). Cultures were maintained at 37°Cin humidified air

containing 5% CO2 and were subcultivated at a split ratio 1:2-1:6 when

confluent (once or twice/week). A mixture of EDTA (0.2 mg/ml) andtrypsin (Difco) (2.5 mg/ml) in PBS was used as detaching agent.

Cloning Procedure. For cloning, U-343 MGa cells of low (<100) andhigh (>400) passage, as well as U-343 MGa CI 2 cells, were used. A

single cell suspension was prepared, and 1000 cells were seeded intoeach of 100-mm culture dishes, either untreated or containing a confluent

feeder layer of normal human glial cells (16). Cultures were incubated for2-5 weeks. Well demarcated cell colonies were isolated using glass

cloning cylinders glued to the bottom with silicone grease. Cells weredetached, suspended in culture medium, and replated in 35-mm culture

dishes and further propagated.Phase Contrast Microscopy. Cells were seeded sparsely on cover-

slips contained in 35-mm Petri dishes. After incubation for 3-4 dayscultures were fixed for 1 h at room temperature in 2% glutaralde-hyde:0.15 M sodium cacodylate:0.1 M sucrose and then kept at 4°C.

Coverslips were rinsed in PBS, mounted on glass slides, and viewed ina Zeiss photomicroscope with phase contrast equipment. Photomicrographs were taken on Kodak technical Pan film 2415 and used to dividethe glioma clones into 4 groups according to morphology and growthpattern. The assignment of morphological types to the clones was doneblindly using unmarked micrographs mixed together randomly.

Immunofluorescence Staining of Glial Fibrillary Acidic Protein. Fiftythousand cells, suspended in 2 ml Eagle's MEM 10% NCS, were seeded

on coverslips contained in 35-mm Petri dishes. After 3 days of incubationthe cells were rinsed three times in PBS and fixed in 2% paraformalde-

hyde in PBS for 20 min. They were postfixed in 96% ethanol for 5 min,stained for 20 min with rabbit anti-human GFAP (1:40) (kindly provided

by Dr. D. Dahl), and then, as a second layer, they were stained withfluorescein isothiocyanate-conjugated goat anti-rabbit IgG (1:20) for 20min. After washing, coverslips were mounted in glycerol-paraphenyl-

enediamine (Sigma). Micrographs were taken in a Leitz epifluorescencemicroscope with oil immersion objectives, using Kodak Tri X Pan professional film.

Growth Factors. Pure PDGF was prepared from human platelets asdescribed (21) and labeled with 125Iusing the chloramine T method to aspecific activity of 20,000-35,000 cpm/ng (22). Mouse EGF ("receptorgrade") was purchased from Collaborative Research (Waltham, MA) andwas also labeled with 125Iusing the chloramine T method to a specific

activity of 60,000-80,000 cpm/ng.

Assay for PDGF Receptor Competing Activity in Glioma Conditioned Medium. Serum-free conditioned media were harvested fromconfluent cultures incubated for 3-4 days in serum-free F-10 medium

with antibiotics. Medium volume was adjusted according to cell numbersin order to ensure an approximately constant volume/cell ratio. Theconditioned media were clarified by centrifugation and stored at -20°C

until tested. PDGF receptor competing activity was monitored usinghuman foreskin fibroblasts, line AG 1523, as test objects (AG 1523 cellswere obtained from the Human Mutant Cell Repository Institute forMedical Research, Camden, NJ). Cells were seeded in 12-well Linbroplates, grown to confluence, and washed once with ice-cold F-10 con

taining human serum albumin at 0.1 mg/ml and 20 mw HEPES, pH 7.4.Test samples, containing human serum albumin and HEPES, at concentrations as above, were added, and the plates were incubated at 4°C

for 2 h (the media from low passage U-343 MGa clones were tested

after concentration 5 times using disposable Amicon filters B 15, as wellas unconcentrated). Cells were washed three times with binding buffer(phosphate-buffered saline containing 1 mg human serum albumin per

ml, 0.01 mg CaCI2 2H2O per ml, and 0.01 mg MgSO4 7H2O per ml). Eachwell then received 0.5 ml binding buffer containing 5 ng (100,000-175,000 cpm) of 125I-PDGF. After incubation for 1 h at 4°C, cultures

were washed five times with binding buffer containing 1% NCS insteadof albumin. Cell associated radioactivity was extracted with 0.5 ml lysisbuffer [1% Triton X-100:20 mw HEPES, pH 7.4:10% (v/v) glycerol

supplemented with 0.1 mg human serum albumin per ml]. The Tritonlysate was sampled after 20 min, and the radioactivity was determinedin a gamma spectrometer. Nonspecific binding was determined as theamount of 125I-PDGF bound in the presence of a 50- to 100-fold molar

excess of unlabeled PDGF. PDGF receptor competing activity in sampleswas converted to equivalent concentration of PDGF, using a standardcurve constructed from results obtained with pure, unlabeled PDGF (5-

100 ng per ml).Assay of ml-PDGF and 125I-EGFBinding to Glioma Cells. Cells were

grown to confluence in 12-well Linbro plates. Cultures were washed

once with binding buffer (see above) at room temperature, and each wellreceived 0.5 ml binding buffer to which 5 ng 125I-PDGF(100,000-175,000

cpm) was added. Binding occurred for 1 h at room temperature and wasterminated by five washes with binding buffer containing 1% NCS(equilibrated at room temperature). Cells were lysed in 0.5 ml lysis bufferas above, and radioactivity and nonspecific binding were determined asabove. For each cell type the cell number per well was determined.Binding was estimated as fmol of tracer PDGF/106 cells. Scatchard

analysis was used to calculate the number of receptors per cell (23).Binding of 125I-EGFto cells was performed in an analogous manner

using 1 ng of 125I-EGF(60,000-80,000 cpm/ng). Nonspecific binding was

determined using a 500 molar excess of unlabeled EGF.Analysis of Serum Dependence of Glioma Cell Multiplication.

Cloned glioma cells were detached and suspended in Eagle's MEM, 1%

NCS with antibiotics. The cell number was determined, and cells wereseeded into 35-mm Petri dishes, with 1 x 10s cells in 2 ml medium per

dish. After 1 day of incubation, the number of cells per dish wasdetermined, and the dishes were divided into two groups, one receivingEagle's MEM 10% NCS and the other receiving serum-free MCDB 104.

Cultures were incubated for 2-3 weeks. Medium was changed every

second or third day. Cell number of duplicate dishes was determined atthe same intervals, and growth curves were constructed from the meanvalues. The population doubling time was deduced from the growthcurves.

Cell Counting. Trypsinized cells in single cell suspension were countedin an electronic cell counter (Celloscope).

RESULTS

Characteristics of Glioma Cell Lines: Production of GliomaPDGF and Binding of 125I-PDGF.Chart 1 shows the results of

CANCER RESEARCH VOL. 46 JANUARY 1986

333

Research. on August 29, 2018. © 1986 American Association for Cancercancerres.aacrjournals.org Downloaded from

CLONAL VARIATION IN THE PRODUCTION OF GLIOMA PDGF

i:0-l.l 8:0-2.7P:4-31 P.-0-5

Chart 1. Derivationof five glioma cell lines from one human glioblastoma multi-forme. The relative positions of the sublines approximately parallelsthe number ofin vitro passages at the time of use in the investigation. Binding (B) of 125I-PDGFto each cell line is depicted as fmol/106cells.The levelof bindingvaried on different

occasions with the ranges shown. The PDGF receptor competing activity inconditioned serum-free medium (P) is given as ng/ml PDGFequivalents.

the analysis of PDGF receptor competing activity in conditionedmedia and 125I-PDGFbinding of the various glioma sublines. A

PDGF receptor competing activity could be detected in mediaharvested from high passage U-343 MGa cells and from theclonal derivatives U-343 MGa CI 2 and U-343 MGa CI 3, whereas

no such activity could be demonstrated in the media of lowpassage U-343 MGa cells or in U-343 MG cells. On the otherhand, it could be shown that low passage U-343 MGa cells andU-343 MG cells displayed specific binding of 125I-PDGF,while cell

lines producing a PDGF-like factor were essentially devoid of125I-PDGFbinding capacity.

In the following series of experiments we focused our intereston the U-343 MGa line which, depending on passage level,expressed both 125I-PDGF binding and production of gliomaPDGF. First we analyzed the clonal variation in 125I-PDGFbinding

and production of PDGF receptor competing activity in this cellline.

Isolation of Clonal Derivatives of the U-343 MGa GliomaLine. When seeded sparsely, the cells from low passage U-343

MGa cells grew relatively slowly, and well defined colonies werenot formed on bare plastic, whereas cells grown on feeder cellsgave rise to well demarcated colonies. Twenty-three of 36 ran

domly selected colonies were successfully transferred and serially propagated. Cells from high passage U-343 MGa cultures

formed well defined colonies on bare plastic; 30 of 38 coloniescould be isolated and serially subcultivated. U-343 MGa CI 2

cells gave rise to well demarcated, fast growing colonies withtightly packed cells. From this cell line 30 of 35 selected coloniescould be subcultivated. Thus, 83 individual clonal sublines werederived from the U-343 MGa line and its clonal subline U-343

MGa CI 2. Based on their morphology, including occurrence ofcytoplasmic processes and growth pattern, the clonal sublineswere classified into 4 groups by the aid of phase contrastmicrographs (Fig. 1).

Thirty-five clones representing all of the 4 groups were ana

lyzed with regard to GFAP immunofluorescence. In all of these

clones, 100% of the cells contained a stainable fibrous network(examples are given in Fig. 2). This shows clearly that the cellsare of astrocytic origin.

Clonal Variation in 12SI-PDGF Binding and Production of

Glioma PDGF. Twenty-three clones from low passage and 30clones from high passage U-343 MGa and 16 clones from U-343MGa CI 2 were tested with regard to both binding of 125I-PDGF

and secretion of PDGF receptor competing activity. The result isshown in Chart 2. There was a striking variation in the twoparameters; glioma PDGF production varied from 0 to 87 ngPDGF equivalents/ml, and binding of 125I-PDGFranged from 0to 44 fmol/106 cells. High passage U-343 MGa cells gave rise to

a higher number of clones showing production of glioma PDGFthan did low passage cells. U-343 MGa CI 2 gave rise to a

number of high producing clonal sublines. A few clones displayedneither 125I-PDGF binding nor production of glioma PDGF,

whereas some clones expressed both parameters.The correlation between 125I-PDGFbinding and production of

glioma PDGF is graphically visualized in Chart 2. As can be seen,there appeared to be a negative con-elation between the two

parameters with a pronounced tendency of high producingclones to have a low binding capacity. The concentration dependence of 125I-PDGF binding was analyzed on two of the

receptor positive clones, both derived from low passage U-343MGa cells and belonging to the morphological group D. One ofthem produced PDGF, and another was practically nonproduc-ing. Scatchard analysis of the binding data is given in Chart 3. Anumber of 74,000 receptors per cell could be calculated on theclone with the highest binding, whereas the other one, i.e., theproducing clone, had 48,000 receptors per cell. The K, was 10"10

M in both cases. The number of receptors is somewhat lowerthan that reported for human cultured fibroblasts (22). Neverthe-

Chart2. Clonal variation in 1MI-PDGFbinding and glioma PDGF production.Three cell lines, all derived from the same human malignant glioma, were cloned:(a) U-343 MGa low passage (<100); (b) U-343 MGa high passage (>400); and (c)U-343 MGa CI 2. The clones were tested regarding bindingof 1AI-PDGF(fmol/106cells) and secretion of PDGF receptor competing activity to serum-free medium(ng/ml PDGFequivalents);each vertical line represents one clone.

CANCER RESEARCH VOL. 46 JANUARY 1986

334

Research. on August 29, 2018. © 1986 American Association for Cancercancerres.aacrjournals.org Downloaded from

CLONAL VARIATION IN THE PRODUCTION OF GLIOMA PDGF

0.30

0.20

0.10

1234Bound(ng POGF / 10°eels)

0.30

.0.20

0.10

1234Bound(ng PDGF/106cells)

Chart 3. Scatchard analysis of 125I-PDGFbinding. A, a clone from the morphological group D, that bound 28 fmol of tracer 125I-PDGF/106 cells and produced

only 0.5 ng/ml PDGF equivalents (detectable after concentration of the medium).The Scatchard plot gave an estimated amount of 74,000 PDGF receptors per cell.B, the single clone from the morphological group D2 with elongated cells (seen inFig. 1, D2) that bound 14 fmol of tracer 125I-PDGF/108cells and produced 1-5 ng/

ml PDGF equivalents. The Scatchard plot gave an estimated amount of 48,000PDGF receptors per cell. The K«was 10"'°M in both cases.

less the results show that saturable binding sites occur on cellsthat produce the PDGF-like factor.

The degree of glioma PDGF production and 125I-PDGFbinding

correlated to some extent with cellular morphology and growthpattern (Table 1 and Fig. 1). It is obvious that the high producing

cells belonged to the morphological group A, consisting of small,polyhedral, tightly growing cells. On the other hand, cells havingthe highest amount of PDGF receptors were large and star-

shaped, with many short cytoplasmic processes such as thoseseen in group D.

Clonal Variation of Serum-free Growth: Relation to the

Production of Glioma PDGF. Twelve clones derived from lowpassage and 5 from high passage U-343 MGa cultures, as wellas 5 from U-343 MGa CI 2 cultures, were tested for serum-freegrowth. Population doubling time in serum-containing mediumvaried from 1.2 to 8.7 days, and in serum-free medium it variedfrom 2.6 days to infinity. The growth rate in serum-free medium

correlated fairly well with the production of glioma PDGF, with acorrelation coefficient of 0.83 (Chart 4).

Binding of 125I-EGF. In order to learn whether the donaivariation in 125I-PDGFbinding was paralleled by a similar rangein number of available EGF receptors, the binding of 125I-EGF

was determined on all clones described above. Most of thesedisplayed high EGF binding, ranging from 20 to 200 fmol of 125I-EGF/106 cells. Two clones derived from low passage U-343 MGacultures had a lower binding capacity, 3 and 7 fmol/106 cells,respectively. There was no correlation to the degree of 125I-PDGF

binding (not shown).Since it has been demonstrated that human cells may release

an EGF-like growth factor which interferes with 125I-EGFbinding

(24), serum-free conditioned media of clones showing a relativelylow 125I-EGFbinding were analyzed for the presence of 125I-EGF

receptor competing activity. However, there was no negativecorrelation between 125I-EGF binding and production of EGF

receptor blocking activity of the glioma clones analyzed (data notshown). Though caution must be taken in the interpretation ofthese data due to the possibility of PDGF-mediated transmodu

lation of EGF receptors on the test cells (25), data suggest theexistence of a true clonal variation also in the expression of EGFreceptors.

DISCUSSION

We have presented evidence for a striking variation in theproduction of a PDGF-like growth factor in different cell lines

derived from the same human glioblastoma multiforme biopsy.Our data also indicate that by long-time culturing and cloning

one selects for sublines with a higher level of growth factor

Table 1125/-PDGFbinding and production of glioma PDGF in glioma clones with different morphology

For each of the 4 morphological groups the mean values of 125I-PDGFbinding, glioma PDGF production,

and population doubling time in serum-free medium were calculated. Population doubling times of >16 dayswere given a value of 16. The number of clones analyzed and their derivation is also shown.

No. ofclonesDerivationMorpho

logicalgroupA

BCDU-343

MGalowpassage1

1264U-343

MGahighpassage10

1910U-343

MGaCI230

000Total413174Glioma

PDGFproduction (ng/ml)Mean

value30

5.61.90.9No.

ofclonestested40

317

4125I-PDGF

binding (fmol/106

cells)Mean

value0.42

2.90.48

31No.

ofclonestested28

317

4Population

doubling time in

serum-free medium(days)Mean

value5.2

11>16

12.4No.

ofclonestested5

1133

CANCER RESEARCH VOL. 46 JANUARY 1986

335

Research. on August 29, 2018. © 1986 American Association for Cancercancerres.aacrjournals.org Downloaded from

CLONAL VARIATION IN THE PRODUCTION OF GLIOMA PDGF

$ 0.4

£ 0.3

I

I 02•

1"5

I < 0.06

O20 40 60 80

Production (ng/ml PDGF equivalents)

Chart 4. Growth of glioma clones in serum-free medium. Cells from 22 differentclones, originally derived from one human malignant glioma, were grown in serum-free medium as described in "Materials and Methods." Growth curves were

constructed for each clone. The number of cell generations per day was calculatedand compared to the amount of receptor competing activity in serum-free mediaharvested from confluent cultures of the same clones. Some clones grew exceptionally slowly in serum-free medium, and the exact number of cell generations perday was difficult to deduce. Values <0.06 were approximated (see also Table 1).The correlation coefficient was calculated to be 0.83 when all of the points wereconsidered. In the absence of the single outlying point, it was 0.59.

release. This finding implies that the production of a PDGF-like

growth factor is of selective growth advantage for glioma cellsin culture.

The variability in growth factor production of the various clonescorrelates fairly well to the cellular levels of c-s/s mRNA.4 The

genetic background to the selection for variants that express thec-s/s gene and produce PDGF is, however, as yet unknown.

Recent studies have shown that repeated cell division underselective pressure may lead to the amplification of specific genes(26). Amplification of certain oncogenes, such as c-myc (27, 28)and N-myc (29), has been demonstrated in human tumor celllines. However, preliminary analysis of the karyotype of the U-

343 MGa clones has given no gross structural evidence for anamplification (or gross rearrangement) of the c-s/s gene; thus no

homogeneous staining regions or double minutes have beenfound.5

The clonal variation in synthesis of glioma PDGF is consistentwith the notion that release of this growth factor occurs as a latephenomenon in tumor development and is related to tumorprogression. Previous studies on the expression of ras genes inhuman cancers have led to similar conclusions. Thus, an activated ras gene was present only in one out of five cell linesderived from individual métastases of a malignant melanoma(30). Moreover, an immunohistochemical study of the ras geneproduct in colon and mammary carcinoma revealed a considerable cellular heterogeneity; in the case of colon carcinoma, theexpression was related to the depth of invasion (31). In thepresent study the preponderance of high producing clones athigh passage levels suggests that progression with regard toPDGF production is a continual process in cultures; it may evenbe that both initiation of glioma PDGF production and subsequentprogressive changes in its expression are In vitro events. Comparative studies on c-s/s expression in primary tumors as well

* M. Nister, C-H. Heldin, and B. Westermark, unpublished observations.SJ. Mark, M. Nister, and B. Westermark, unpublished observations.

as in cell lines are thus warranted.A possibility, which has to be taken into account, is that the

extent of glioma PDGF production is related to the functional ordevelopmental state of the cell. This notion is suggested by thefinding of a morphological correlate to glioma PDGF production;most of the high producing clones are found among rather"immature" looking, tightly packed cells, whereas low- and non-

producing cells have a more astrocytic or glia-like morphology.

Recent studies have shown that certain normal cells such asarterial smooth muscle cells from young rats (32), human endo-thelial cells (33, 34) and placental cytotrophoblasts (35) expressthe c-s/s gene and produce PDGF. These findings suggest a rolefor PDGF in autocrine or paracrine regulation of normal cellgrowth and imply a wider role for PDGF in physiological growthmechanisms than is suggested by its presence in platelets. It isan interesting speculation that PDGF production is develop-mentally regulated and occurs at specific stages of normal glialdevelopment, here represented by the phenotypically differentglioma cell clones.

The present data also give evidence for a clonal variation inthe 125I-PDGFbinding capacity within the U-343 MGa cell line.

Scatchard analysis confirmed the presence of high affinity receptors with a KÖin the same order of magnitude as that found innormal fibroblasts (22, 36). To a considerable extent, the variation in binding capacity may be secondary to the range in gliomaPDGF production and thus reflect the degree of receptor blockingand down regulation, executed by the endogenous growth factor. However, the presence of non-producing clones devoid of125I-PDGFbinding suggests a true clonal variation in the expres

sion of PDGF receptors.The present results have bearing on the model of autocrine

growth stimulation of tumor cells (15). As indirect evidence forthis model we found a co-variation in the level of glioma PDGFproduction and the growth rate in serum-free medium. As foundpreviously in cultured U-2 OS osteosarcoma cells (37), we found

no growth inhibitory activity by adding PDGF antibodies to thehigh producing clones (data not shown). However, it has beensuggested that an endogenously produced growth factor mayinteract with its receptor in an intracellular compartment, inaccessible to exogenously added antibodies (37). It is thus possiblethat the lack of inhibitory effect of PDGF antibodies is, at leastin part, due to such a steric exclusion.

Autocrine growth stimulation presupposes a concomitant synthesis of growth factor and the corresponding receptor. Thepresence of clones within a population with varying expressionof PDGF and PDGF receptors suggests a more complex situation, involving both autocrine and paracrine growth stimulation.We therefore expect that the glioma clones described in thiscommunication will be of value for further studies on the role ofPDGF and the PDGF receptor in human glioma.

REFERENCES

1. Waterfield, M. D., Scrace, G. T., Whittle, N., Stroobant, P., Johnsson, A.,Wasteson, A., Westermark, B., Heldin, C-H., Huang, J. S., and Deuel, T. F.Platelet-derived growth factor is structurally related to the putative transforming protein p28** of simian sarcoma virus. Nature (Lond.), 304: 35-39,1983.

2. Doolittte, R. F., Hunkapiller, M. W., Hood, L. E., Devare, S. G., Robbins, K. C.,Aaronson, S. A., and Antoniades, H. N. Simian sarcoma virus one gene, /-s/s,is derived from the gene (or genes) encoding a platelet-derived growth factor.Science (Wash. DC), 22J: 275-277,1983.

CANCER RESEARCH VOL. 46 JANUARY 1986

336

Research. on August 29, 2018. © 1986 American Association for Cancercancerres.aacrjournals.org Downloaded from

CLONAL VARIATION IN THE PRODUCTION OF GLIOMA PDGF

3. Devare, S. G., Reddy, E. P., Law, J. D., Robbins, K. C., and Aaronson, S. A.Nucleotide sequence of the simian sarcoma virus genome: demonstration thatits acquired cellular sequences encode the transforming gene product p28**.Proc. Nati. Acad. Sci. USA, 80: 731-735, 1983.

4. Deuel, T. F., Huang, J. S., Huang, S. S., Strcobant, P., and WaterfiekJ, M. D.Expression of a platelet-derived growth factor-like protein in simian sarcomavirus transformed cells. Science (Wash. DC), 221: 1348-1350,1983.

5. Bowen-Pope, D. F., Vogel, A., and Ross, R. Production of platelet-derivedgrowth factor-like molecules and reduced expression of platelet-derived growthfactor receptors accompany transformation by a wide spectrum of agents.Proc. Nati. Acad. Sci. USA, 81: 2396-2400, 1984.

6. Johnsson, A., Betsholtz, C., von der Helm, K., Heldin, C-H., and Westermark,B. Platelet-derived growth factor agonist activity of a secreted form of the v-sis oncogene product. Proc. Nati. Acad. Sci. USA, 82: 1721-1725,1985.

7. Owen, A. J., Pantazis, P., and Antoniades, H. N. Simian sarcoma virus-transformed cells secrete a mitogen identical to platelet-derived growth factor.Science (Wash. DC), 225: 54-56,1984.

8. Johnsson, A., Heldin, C-H., Wasteson, A., Westermark, B., Deuel, T. F., Huang,J. S., Seeburg, P. H., Gray, A., Ullrich, A., Scrace, G., Strcobant, P., andWaterfield, M. D. The c-sis gene encodes a precursor of the B-chain of platelet-derived growth factor. EMBO J., 3: 921-928, 1984.

9. Westermark, B. and Wasteson, A. The response of cultured human normalglial cells to growth factors. Adv. Metab. Disord., 8: 85-100,1975.

10. Heldin, C-H., Westermark, B., and Wasteson, A. Chemical and biologicalproperties of a growth factor from human-cultured osteosarcoma cells: resemblance with platelet-derived growth factor. J. Cell Phystol., 105: 235-246,

1980.11. Nister, M., Heldin, C-H., Wasteson, À.,and Westermark, B. A platelet-derived

growth factor analog produced by a human donai glioma cell line. Ann. NYAcad. Sci., 397:25-33,1982.

12. Nister, M., Heldin, C-H., Wasteson, A., and Westermark, B. A glbma-derivedanalog to platelet-derived growth factor: demonstration of receptor competing

activity and immunological crossreactivity. Proc. Nati. Acad. Sci. USA, 8):926-930,1984.

13. Betsholtz, C., Heldin, C-H., Nistér,M., Ek, B., Wasteson, A., and Westermark,B. Synthesis of a PDGF-like growth factor in human glioma and sarcoma cellssuggests the expression of the cellular homologue to the transforming proteinof simian sarcoma virus. Biochem. Biophys. Res. Commun., 777: 176-182,

1983.14. Eva, A., Robbins, K. C., Andersen, P. R., Srinivasan, A., Tronick, S. R., Reddy,

E. P., Ellmore, N. W., Galen, A. T., Lautenberger, J. A., Papas, T. S., Westin,E. H., Wong-Staal, F., Gallo, R. C., and Aaronson, S. A. Cellular genes

analogous to retroviral one genes are transcribed in human tumour cells.Nature (Lend.), 295. 116-119,1982.

15. Sporn, M. B. and Todaro, G. J. Autocrine secretion and malignant transformation of cells. N. Engl. J. Med., 303: 878-880, 1980.

16. Ponten, J. and Macintyre, E. H. Long term culture of normal and neoplastichuman glia. Acta Pathpl. Microbio!. Scand. 74: 465-486,1968.

17. Westermark, B., Pontén,J., and Hugosson, R. Determinants for the establishment of permanent tissue culture lines from human gliomas. Acta Pathol.Microbio!. Scand. (Sect. A), 87: 791-805,1973.

18. Westermark, B. The deficient density-dependent growth control of humanmalignant glioma cells and virus-transformed glia-like cells in culture. Int. J.Cancer, 72:438-451,1973.

19. Ponten, J. and Westermark, B. Properties of human malignant glioma cells invitro. Med. Bid., 56:184-193,1978.

20. Westermark, B., Magnusson, A., and Heldin, C-H. Effect of epidermal growth

factor on membrane motility and cell locomotion in cultures of human clona!glioma cells. J. Neurosa. Res., 8: 491-507,1982.

21. Johnsson, A., Heldin, C-H., Westermark, B., and Wasteson, A. Platelet-derived

growth factor: identification of constituent polypeptide chains. Biochem. Biophys. Res. Commun., 704: 66-74, 1982.

22. Heldin, C-H., Westermark, B., and Wasteson, A. Specific receptors for platelet-derived growth factor on cells derived from connective tissue and glia. Proc.Nat!. Acad. Sci. USA, 78: 3664-3668,1981.

23. Scatchard, G. The attractions of proteins for small molecules and ions. Ann.NY Acad. Sci., 57: 660-672,1949.

24. Todaro, G. J. and De Larco, J. E. Growth factors produced by sarcoma virustransformed cells. Cancer Res., 38: 4147-4154,1978.

25. Wrann, M., Fox, C. F., and Ross, R. Modulation of epidermal growth factorreceptors on 3T3 cells by platelet-derived growth factor. Science (Wash. DC),270: 1363-1365, 1980.

26. Schimke, R. T. Gene amplification in cultured animal cells. Cell, 37: 705-713,1984.

27. Collins, S. and Groudine, M. Amplification of endogenous myc-related DNAsequences in a human myeloid leukaemia cell line. Nature (Lond.), 298: 679-681,1982.

28. Dalla Pavera. R., Wong-Staal, F.. and Gallo, R. C. One gene amplification inpromyelocytic leukaemia cell line HL-60 and primary leukaemia cells of thesame patient. Nature (Lond.), 299: 61-63,1982.

29. Schwab, M., Alitato, K., Ktempnauer, K-H., Varmus, H. E., Bishop, J. M.,Gilbert, F., Brodeur, G., Goldstein, M., and Trent, J. Amplified DNA with limitedhomology to myc cellular oncogene is shared by human neuroblastoma celllines and a neuroblastoma tumour. Nature (Lond.), 305: 345-348, 1983.

30. Albino, A. P., Le Strange, R., Oliff, A. I., Furth, M. E., and Old, L. J. Transformingras genes from human melanoma: a manifestation of tumour heterogeneity?Nature (Lond.), 308: 69-72, 1984.

31. Thor, A., Horan Hand, P., Wunderlich, D., Caruso, A., Muraro, R., and Schlom,J. Monoclonal antibodies define differential ras gene expression in malignantand benign colonie diseases. Nature (Lond.), 371: 562-565,1984.

32. Seifen, R. A., Schwartz, S. M., and Bowen-Pope, D. F. Devetopmentallyregulated production of platelet-derived growth factor-like molecules. Nature(Lond.), 377: 669-671,1984.

33. DiCorteto, P. E. and Bowen-Pope, D. F. Cultured endothelial cells produce aplatelet-derived growth factor-like protein. Proc. Nati. Acad. Sci. USA, 80:1919-1923, 1983.

34. Barrett, T. B., Gajdusek, C. M., Schwartz, S. M., McDougall, J. K., and Bendi«,E. P. Expression of the sis gene by endothelial cells in culture and in vivo.Proc. Nati. Acad. Sci. USA, 87: 6772-6774,1984.

35. Goustin, A. S., Betsholtz, C., Pfeifer-Ohlsson, S., Persson, H., Rydnert, J.,Bywater, M., Holmgren, G., Heldin, C-H., Westermark, B., and Ohlsson, R.Co-expression of the s/s and myc proto-oncogenes in developing humanplacenta suggests autocrine control of trophoWast growth. Cell, 41:301-312,

1985.36. Bowen-Pope, D. F., and Ross, R. Platelet-derived growth factor. II. Specific

binding to cultured cells. J. Bio!. Chem., 257: 5161-5171,1982.37. Betsholtz, C., Westermark, B., Ek, B., and Heldin, C-H. Coexpression of a

PDGF-like growth factor and PDGF receptors in a human osteosarcoma cellline: implications for autocrine receptor activation. Cell, 39: 447-457, 1984.

CANCER RESEARCH VOL. 46 JANUARY 1986

337

Research. on August 29, 2018. © 1986 American Association for Cancercancerres.aacrjournals.org Downloaded from

CLONAL VARIATION IN THE PRODUCTION OF GLIOMA PDGF

Fig. 1. Morphology and growth pattern of glioma clones. Eighty-three clonal sublines from one human malignant glioma cell line (U-343 MGa) were matched into 4different morphologicalgroups (A-D) with the help of phase contrast micrographs. One typical and one extreme cell clone are shown as examples from each group. A,,representative clone with small polyhedral, sometimes slightly spindle-shapedcells growing tightly and piling up. A¡,three clones within group A grew as exceptionallywell demarcated colonies.B,, representativeclone with varying proportions of large polyhedraland middle-sizedspindle-formedcells and only a few cells with processes.BÎ,one clone of special type with middle-sized polygonal cells almost devoid of processes. C,. flat polyhedral cells intermingled with a variable amount of cells withrounded cell bodies and long slender processes. C2,a subtype in which almost all cells developed long branched processes (resemblingfibrous astrocytes). Dt, largestar-shaped cells with many short processes.D2,a subtype with more elongated, but still star-shaped appearance.Bar, 30 t¡m.

CANCER RESEARCH VOL. 46 JANUARY 1986

338

Research. on August 29, 2018. © 1986 American Association for Cancercancerres.aacrjournals.org Downloaded from

CLONAL VARIATION IN THE PRODUCTION OF GLIOMA PDGF

WJij&ê^v&^t *..„.> ••L**%A¡?>4«rev>¡fï•**&»>*'t': ..v 4%O'v "^

ÉiwS^^è •^"*"*:"•>>!

/ ,Ksj

CANCER RESEARCH VOL. 46 JANUARY 1986

339

Research. on August 29, 2018. © 1986 American Association for Cancercancerres.aacrjournals.org Downloaded from

CLONAL VARIATION IN THE PRODUCTION OF GLIOMA PDGF

Fig. 2. GFAP immunofluorescent staining of human glioma cell clones. A, polyhedral PDGF producing cells from group A,. 8, nonproducing, receptorbearing, largestarshaped cell from group D,. C, the special clone representing group C¡,with cells remindingone of fibrous astrocytes. 0, polyhedral and round cells from group C1with long slender cell processes. Bar, 10 urn.

CANCER RESEARCH VOL. 46 JANUARY 1986

340

Research. on August 29, 2018. © 1986 American Association for Cancercancerres.aacrjournals.org Downloaded from

1986;46:332-340. Cancer Res   Monica Nistér, Carl-Henrik Heldin and Bengt Westermark  in a Human Malignant GliomaFactor-like Protein and Expression of Corresponding Receptors Clonal Variation in the Production of a Platelet-derived Growth

  Updated version

  http://cancerres.aacrjournals.org/content/46/1/332

Access the most recent version of this article at:

   

   

   

  E-mail alerts related to this article or journal.Sign up to receive free email-alerts

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

  Permissions

  Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/46/1/332To request permission to re-use all or part of this article, use this link

Research. on August 29, 2018. © 1986 American Association for Cancercancerres.aacrjournals.org Downloaded from