The phenotypic characteristics of simian sarcoma virus ... - NCBI

2 downloads 62 Views 2MB Size Report
2Present address: Ludwig Institute for Cancer Research (Uppsala Branch),. Box 595, BMC, 5-751 85 Uppsala, Sweden. Communicated by J.Schlessinger.
The EMBO Journal vol.5 no.7 pp.1535-1541, 1986

The phenotypic characteristics of simian sarcoma virus-transformed human fibroblasts suggest that the v-sis gene product acts solely as a PDGF receptor agonist in cell transformation

Ann Johnsson, Christer Betsholtzl, Carl-Henrik Heldin2 and Bengt Westermark1 Department of Medical and Physiological Chemistry, Box 575, BMC, S-751 23 Uppsala and 'Department of Pathology, University Hospital, S-751 85 Uppsala, Sweden 2Present address: Ludwig Institute for Cancer Research (Uppsala Branch), Box 595, BMC, 5-751 85 Uppsala, Sweden Communicated by J.Schlessinger

Previous studies have indicated that the oncogene v-sis of simian sarcoma virus (SSV) encodes a growth factor that is structurally and functionally similar to platelet-derived growth factor (PDGF). In the present investigation we have analysed the phenotypic characteristics of human foreskin fibroblasts transformed by SSV. It was found that the PDGF receptors were extensively down-regulated. This finding is consistent with a high, local, extracellular concentration of a PDGF-like factor, synthesized by the transformed cell. The receptors were up-regulated by suranin, a drug that is known to dissociate PDGF and the v-sis product from the PDGF receptors. A cell-associated v-sis product of mol. wt 24 000 was identified by immunoprecipitation with PDGF antibodies; release of this component was induced by a high concentration of exogenous PDGF, indicating that a fraction of the product is associated with the PDGF receptors. SSV was not found to be an immortalizing virus; when serially passaged, SSV-transformed cells had essentially the same life-span as their non-transformed counterparts. Moreover, SSV did not induce growth in soft agar beyond the level afforded by exogenously added PDGF. Thus, the present study favors the notion that SSV transformation is mediated by a growth factor that mimics PDGF but has no further cellular effects. Key words: v-sis oncogene/simian sarcoma virus/platelet-derived growth factor/transformation Introduction The finding of a structural homology between the transforming gene product of simian sarcoma virus (SSV) and human plateletderived growth factor (PDGF) (Devare et al., 1983; Waterfield et al., 1983; Doolittle et al., 1983; Johnsson et al., 1984) suggests that SSV transformation is mediated by a PDGF-like growth factor. This notion has recently been substantiated by a number of observations on the structural and functional properties of the oncogene of SSV (v-sis) and its translation product in relation to PDGF (Heldin and Westermark, 1984; Heldin et al., 1986). The nucleotide sequence of the open reading frame of v-sis predicts a peptide of mol. wt 28 000, denoted p28sis (Devare et al., 1983); the normal cellular homolog (c-sis) encodes a precursor of the B-chain of PDGF (Johnsson et al., 1984; Josephs et al., 1984; Chiu et al., 1984; Collins et al., 1985). Apart from lacking the most 5' part of c-sis, encoding, e.g. a signal peptide, the v-sis gene is homologous to the entire coding region of c-sis with a 94% homology in the predicted amino acid sequences. Translation of the v-sis transcript is apparently initiated IRL Press Limited, Oxford, England

within the helper virus-derived env sequence (King et al., 1985) the function of which, in addition to providing the initiation codon, is to encode a signal peptide that compensates for the loss of the cellular signal sequence. Deletion of the signal sequence abrogates the transforming activity of SSV (Hannink and Donoghue, 1984; King et al., 1985). Analysis of the metabolic processing of p28sis has provided evidence for an intimate structural relationship between the mature v-sis product and PDGF (Robbins et al., 1983). In SSVtransformed cells, p28sis is dimerized and proteolytically processed to an apparently stable end product of mol. wt. 24 000 that is structurally similar or identical to a PDGF B-chain homodimer (Robbins et al., 1983). Human PDGF consists of dimers of A-chains and B-chains (Johnsson et al., 1982); these are extensively homologous in their amino acid sequences and encoded by different genes (Johnsson et al., 1984; Betsholtz et al., in preparation). Thus, the v-sis translation product is structurally more similar to pig PDGF that seems to be a homodimer of polypeptide chains homologous to the human PDGF B-chain (Stroobant and Waterfield, 1984). If the v-sis gene product were to function as a PDGF agonist in cell transformation, it obviously has to be compartmentalized together with the PDGF receptor, which is an integral membrane protein with the ligand binding domain outside of the cell membrane (Heldin et al., 198 la, 1983; Glenn et al., 1982). The strict requirement for a signal peptide conforms with this prediction. In addition, there is compelling experimental evidence that the v-sis gene product is externalized and exerts its function via the PDGF receptor on the cell surface. Thus, addition of PDGF antibodies attenuates SSV-induced focus formation and inhibits growth in serum-free medium of SSV-transformed human foreskin fibroblasts infected with SSV (Johnsson et al., 1985b). The proposed mechanism of SSV transformation in vitro raises the question about the neoplastic properties of the SSVtransformed cells. Is the endogenous production of a PDGF-like growth factor alone sufficient to induce and maintain a truly malignant phenotype? In relation to this question we have considered it important to analyse further the phenotypic properties of SSV-transformed cells in vitro as compared with normal cells responding to exogenous PDGF. The present communication provides additional evidence that an externalized v-sis gene product interacts with the cellular PDGF receptor. Moreover, we have found that SSV-transformed human fibroblasts have a finite life span and grow in soft agar to about the same extent as nontransformed cells exposed to PDGF. We therefore conclude that the v-sis gene product in all respects may mimic PDGF in its cellular functions.

Results Immunoprecipitation of the v-sis gene product in lysates and conditioned medium of SSV-transformed cells PDGF antibodies can be used for the immunoprecipitation of the v-sis translational product (Robbins et al., 1983; Johnsson et al., 1985a). In the present investigation it was found that the anti1535

A.Johnsson et al.

Medium b a

Cell lysate a b

_soI 444 $

-

4mr ei

-24 kDa

Fig.

1. Identification of p24Sis produced by SSV-transformed human fibroblasts. Confluent cultures of SSV-transformed cells were incubated with 500 ng/ml partially purified PDGF overnight, washed twice and then

metabolically labelled with [35S]cysteine in medium containing 10 ng/ml PDGF as designed in Materials and methods. Conditioned medium and cell lysates were subjected to sequential incubations with control rabbit serum (a) and PDGF antiserum (b). Immunoglobulin-bound radioactivity were analysed by SDS-gel electrophoresis under non-reducing conditions followed by fluorography. bodies recognized a 24-kd product in lysates of SSV-transformed AG 1523 fibroblasts (Figure 1). This component was not found in the corresponding non-transformed cells (not shown). Upon reduction the 24-kd species was converted to 12 kd (not shown). This finding suggests that the factor is identical to the dimer of mol. wt. 24 000, previously identified as the end product of processed p28sis (Robbins et al., 1983). However, no immunoprecipitable material was detected in the conditioned medium of SSV-transformed cells (not shown). This finding is in agreement with the observation that the v-sis product remains associated with the cells, in particular the membrane compartment (Robbins et al., 1985). The molecular structure(s) anchoring the protein at the cell membrane has not been identified. An obvious 1536

possibility is that the factor associates with the PDGF receptor immediately after externalization. If this were the case, blocking and down-regulating the receptor with exogenous PDGF should inhibit this route of binding and allow the release of the factor to the medium. Indeed, the 24-kd product could be immunoprecipitated from medium conditioned by SSVtransformed cells pre-treated in 500 ng/ml PDGF and maintained in 10 ng/ml PDGF (Figure 1). We therefore draw the conclusion that the release of the factor to the bulk medium is hindered to some extent by its interaction with the PDGF receptor. Down-regulation of the PDGF receptor in SSV-transformed fibroblasts It follows from the results above that the minute concentration of the v-sis product detected in the bulk medium may not be a relevant measure of the local concentration to which the cells are really exposed. As can be seen in Figure 2a, the PDGF receptor on SSV-transformed cells is extensively down-regulated as would be the case if the cells were indeed exposed to a relatively high concentration of a PDGF receptor-binding factor; binding of a tracer amount of [1251]PDGF is reduced by some 90% as compared with non-transformed cells. Treating the cells with suramin at 0°C (200 /Ag/ml, 15 min), a drug known to dissociate receptor-bound PDGF (Williams et al., 1984), as well as the vsis product (Garrett et al., 1984), leads to an increase in specific binding of [1251]PDGF to a level corresponding to 21 % of the binding of the non-transformed control cells (Figure 2a). This increase probably reflects the steady-state level of PDGF-receptor occupancy on the cell surface. Upon culture in suramin-containing medium at 37°C (200 /tg/ml), the binding capacity increased dramatically (Figure 2b); within 4 h a new steady-state level was reached, essentially corresponding to the binding capacity of the non-transformed control cells. Life span of SSV-transformed human fibroblasts To determine the life-span of SSV-transformed human fibroblasts, sparse AG 1523 cells were infected with SSV (2500 f.f.u./ml) at passage level 8 and after 10 days, when all cells appeared to have converted to the SSV-transformed phenotype, the cumulative increase in cell number was determined by serial passage as described by Ponten (1970). As seen in Figure 3, the population doubling time of the SSV-transformed cells was only a little shorter than that of the control cells. Upon prolonged culture, the growth of both cell types decelerated until a plateau phase was reached after 8 weeks of serial passage. At this point the cultures could not be subcultivated with any gain in net cell number. At low passage levels, the fraction of [3H]thymidinelabeled cells after a 24 h labeling period was high, 54 % in SSVtransformed cells and 45% in the control cells; at high passage levels, only 4 % and 6%, respectively, incorporated [3H]thymidine during the 24 h labeling period (cf. Table I). Senescent cells are known to differ in morphology from cells of low passage level (Bowman and Daniel, 1975); they become large and flat with a considerable increase in nuclear/cytoplasmic ratio. Interestingly, the SSV-transformed cells at the end of their life span attained a similar morphology, in fact indistinguishable from that of the non-transformed cells of comparable age (cf. Figure 4). At low passage level, however, the two cell types differed markedly in morphology; the SSV-transformed cells were more fusiform and tended to grow in a criss-cross pattern (Johnsson et al., 1985b). The disappearance of the transformed characteristics was not related to a loss of the viral genome as high passage cells were found to release infectious virus particles, -

The v-sis gene product as a PDGF receptor agonist

a

C,)

b a.

10

10-

00

CO 0

v

0

D

E

.0

0

.2

0

Q 0 -o

=

QO°

5

enCc)

D

n)

I

0

5L

CO l0 ) U-

cm

I X',

co

>

CO Co

I

I

0 2 Suramin.

I ct

C


C/ u: CD +

Fig. 2. Effect of suramin on [125I]PDGF binding to SSV-transformed human fibroblasts. (a) The binding of [125I]PDGF to confluent cultures of SSVtransformed fibroblasts and their non-transformed fibroblasts was determined in the absence (open bars) or presence (filled bars) of an excess of unlabelled PDGF as described in Materials and methods. In addition SSV-transformed fibroblasts were analysed for [125I]PDGF binding after pre-treatment in suramin 200 ug/mJ, 15 min at 4°C. (b) SSV-transformed (0-0) and non-transformed (0-0) human fibroblasts were analyzed for [1251]PDGF binding after incubation in the presence of 200 mmol/ml of suramin at 37°C for various time periods.

Table I. Comparison of high and low passage human fibroblasts, either non-transformed or transformed by SSV, with regard to fraction of growing cells, release of infectious virus particles and binding of [125I]PDGF

15K

% Labelled nuclei Virus production [125I]PDGF binding (f.f.u./106 cells) (c.p.m./106 cells) (24 h [3H]Tdr pulse) SSV- 1523 Low passage 54.3 High passage 4.1 AG 1523 Low passage 44.8 High passage 6.1

.0

E 420 748 -

-

2230 2418 22 420 28 817

Cultures of low passage (SSV-1523 passage 19, AG 1523 passage 17) and high passage (SSV-1523 passage 36, AG 1523 passage 38) cells, were used. The fraction of labelled nuclei was determined by autoradiography using subconfluent cultures after a 24 h incubation with [3H]thymidine in Eagle's MEM supplemented with 1 % newborn calf serum. SSV-transformed cell cultures of high and low passages were analysed for production of SSV/SSAV by subjecting conditioned medium from these cells to nontransformed fibroblasts; the number of foci after 10 days was used as an estimate of the production of infectious virus particles. Binding of [125I]PDGF was performed as described in Materials and methods.

measured as focus forming units, to the same extent as low passage cells (Table I). Moreover, the down-regulation of the PDGF receptor was as pronounced in senescent SSV-transformed cells as in young cells (Table I). Thus, the SSV-transformed cells at high passage level revert phenotypically despite a continuous expression of the viral genome and production of a PDGF-like

activity. Growth in semi-solid medium To analyse the effect of SSV transformation of PDGF on growth

10 a) co E 0 0

cn CD 0)

5

81) 0

20

40

60

80

Days in culture Fig. 3. Accumulated growth of SSV-transformed and non-transformed human fibroblasts. AG 1523 cells (0-0) and their SSV-transformed counterparts (0-0) were grown on 4.5-cm Falcon dishes as described in Materials and methods. When the cultures reached confluency they were split 1:4 and the cell numbers were determined in duplicate parallel cultures. The accumulated cell number from cultures containing 50 000 cells as estimated day 0 (one day after plating), is illustrated. 1537

A.Johnsson et al.

p

.'

t.4.

.

,

a. .

.':.sti.

-

.

.'

ki

N Fig. 4. Micrographs of representative fields of low passage (a) and high passage (b) SSV-transformed fibroblasts.

in semi-solid medium, SSV-1523 cells as well as the nontransformed counterparts were implanted in agarose in medium containing 10% newborn calf serum. One group of non-transformed cells received PDGF (10 ng/ml). Colonies were formed in cultures of SSV-transformed cells and in non-transformed cells

1538

given PDGF, whereas no apparent growth occured in nontransformed cells in the absence of PDGF (Figure 5). The colonies of SSV-transformed cells and PDGF-stimulated cells had the same morphology and consisted of densely packed cells surrounded by a zone of cell debris.

The v-sis gene product as a PDGF receptor agonist

Fig. 5. Effect of SSV transformation and PDGF stimulation on growth in semi-solid medium of human fibroblasts. Growth in soft agar was assayed as described in Materials and methods. Micrographs are shown illustrating average sized colonies formed of SSV-transformed (a) and PDGF-stimulated (b) human fibroblasts. The same frequency of colony formation was seen in cultures of SSV-transformed and PDGF-stimulated cells. No colonies were detected in control cultures of non-PDGF-stimulated human fibroblasts (c).

Discussion The present investigation has provided further evidence in favour of the notion that SSV transformation of fibroblasts in culture is mediated by a growth factor that mimics PDGF in its biological activity but that has no further effects on the target cells. Immortality and anchorage independence are commonly recognized traits of transformed cells that are generally assumed to be of significance for their tumorigenic properties. The present investigation clearly demonstrates the inability of SSV in immortalizing human fibroblasts. Rather, at high pasage level, the SSV-transformed cells apparently become refractory to the v-sis product and revert phenotypically despite a sustained expression of the viral genome. In this respect the cells resemble normal fibroblasts at the end of their life span. Such cells retain their PDGF receptors (cf. Table I) but are non-responsive to PDGF with regard to its effects on cell morphology and DNA synthesis (Paulsson et al., in preparation). Most likely, SSVtransformed cells and their normal counterparts senesce by the same mechanism involving a block in the post-receptor pathway. Since anchorage-independent growth is an in vitro parameter

that has been claimed to correlate well with tumorigenicity (Shin et al., 1975), it was of particular interest to analyse the effect of SSV transformation versus PDGF stimulation on growth in soft agar. Again, we could not discriminate between SSV transformation and PDGF stimulation. The induction of anchorage independence by PDGF may seem surprising. However, it is becoming increasingly evident that anchorage dependence is not absolute; normal cells, generally regarded as anchorage dependent, can form colonies in semi-solid media when given the proper growth factors (Sporn and Todaro, 1980; Massague, 1985; Sporn and Roberts, 1985). PDGF may be one such factor (cf. Assoian et al., 1984; Kaplan and Ozanne, 1983), the requirement for which can be substituted for by SSV transformation. It may seem paradoxical that the v-sis product remains associated with the cell (Robbins et al., 1985), although it has the structural features of a secretory protein and to all appearances has to be externalized in order to elicit its biological functions. We would like to propose that a significant amount of the cellassociated factor represents receptor-bound material. The exten1539

A.Johnsson et al.

sive down-regulation of the PDGF receptor on SSV-transformed cells (Garrett et al., 1984, and present investigation) suggests that the cells are exposed to a rather high concentration of a PDGF receptor-binding ligand (Heldin et al., 1982) i.e. considerably higher than the small quantitites of receptor-blocking activity found in the conditioned medium. This notion is further supported by the finding that suramin, that dissociates the receptor-bound ligand and reverts SSV transformation (Betsholtz et al., in preparation), is efficient in up-regulating the receptor (Figure 2). Based on current knowledge we would like to propose the following model for the synthesis, release and internalization of the v-sis product: the primary translation product is synthesised on the endoplasmic reticulum and translocated to the secretory compartment, including the Golgi apparatus. After dimerization and trimming, the product (of mol. wt. 24 000) is externalized. The juxtacellular diffusion boundary layer (Stoker, 1973) hinders free exchange with the bulk medium and thus favors an interaction of the factor with cell surface PDGF receptors. After binding, the factor is internalized in the same manner as PDGF. Obviously, this model does not exclude the possibility that the v-sis product also, in a non-specific fashion, associates with other structures on the cell membrane. The notion that SSV transformation is mediated by a growth factor that remains associated with the producing cell or in its near vicinity is in accordance with the morphology of the SSV-transformed foci. These appear as sharply demarcated bundles or whirls of densely packed cells in parallel arrays (Johnsson et al., 1985b). In the case of a freely diffusing growth factor one would rather expect the foci to be surrounded by a zone of reactive cells, reflecting the concentration gradient of the diffusing factor. The rather simplistic model for SSV transformation is sufficient to explain the in vitro characteristics of the transformed cells but hardly explains the well established tumorigenic properties of the virus; SSV is known to induce fibrosarcomas and malignant gliomas in newborn marmosets, when administered intramuscularly or intracerebrally, respectively (Deinhardt, 1980). It is conceivable that SSV infection of a particular cell type in vivo leads to a polyclonal expansion driven by the endogenous growth factor. Such cells may then be at high risk of undergoing secondary changes leading to the emergence of a population having the malignant phenotype. Clearly, analyses of the genotypic and phenotypic properties of SSV-induced tumors in relation to in vitro transformed cells are highly warranted.

Materials and methods Cells and cell culture The human diploid foreskin fiberoblast cell line AG 1523 was obtained from the Human Genetic Mutant Cell Repository (Camden, NJ). Transformation by SSAV/SSV was carried out as described (Johnsson et al., 1985b) using 2500 f.f.u. per dish. Cells were routinely grown in Eagle's minimal essential medium (MEM) supplemented with 10% (v/v) newborn calf serum and 100 units of penicillin and 50 itg of streptomycin per ml. PDGF Pure PDGF was prepared as described by Johnsson et al. (1982), and radiolabelled using the chloramin-T method to a specific activity of -50 000 c.p.m./ng (Heldin et al., 1981a). Metabolic labelling Confluent cultures of SSV-transformed human fibroblasts, grown in roller bottles, were incubated with 500 ng of partially purified PDGF in 10 ml of MCDB 104 overnight. After washing twice with 10 ml of medium, cells were incubated in the presence of pure PDGF (10 ng/ml), with [35S]cysteine (0.25 mCi/ml; 600 Ci/mmol) for 3 h in 4 ml of cysteine-free MCDB 104, followed by a 2-h incubation in 3 ml of unlabelled cysteine-containing medium. Medium and cell lysates (obtained by solubilization in 1% Triton X-100, 0.5 M NaCl, 0.01 M phosphate pH 7.4) were sequentially immunoprecipitated by control rabbit serum

1540

and PDGF antiserum (Heldin et al., 1981b; Betsholtz et al., 1983). Immunoprecipitated radioactivity was analysed by SDS-gel electrophoresis (Blobel and Dobberstein, 1975), using 13-18% gradient gels and fluorography. Cell counting Cell cultures were harvested with trypsin-versene and dispersed by repipetting to single cell suspensions as monitored by phase-contrast microscopy. Cell numbers were counted by Coulter counter. Analysis of DNA synthesis [3H]Thymidine incorporation in confluent cultures of AG 1523 and SSV-transformed AG 1523 was determined by autoradiography as described (Betsholtz and

Westermark, 1984). Cell binding of [1251JPDGF Confluent cultures of SSV-transformed or of control human fibroblasts, grown in 12-well Linbro culture dishes, were washed once in washing buffer (phosphatebuffered saline supplemented with 0.9 mM CaC12, 0.8 mM MgSO4 and 1 mg/mi bovine serum albumin) and incubated with 1 ng [125I]PDGF in 0.5 ml of washing buffer for 2 h on ice. After washing five times in ice-cold washing buffer, cell bound radioactivity was solubilized for 20 min in 1 % Triton X-100, 10% glycerol, 20 mM Hepes, pH 7.4 and was determined in a gamma spectrometer. Non-specific binding was determined as the amount of [1251]PDGF bound in a 250-fold molar excess of unlabelled PDGF (30% pure). Soft agar assay Anchorage-independent growth was assayed as described (Roberts et al., 1980). Culture wells (35 mm tissue culture dishes) received 2 ml of Eagle's MEM containing 10% newborn calf serum and 0.7% (w/v) agar (Difco). SSV-transformed or normal fibroblasts (2.5 x 104 cells/dish) were added in 0.2 ml of Eagle's MEM supplemented with 10% serum and 0.3% agar; PDGF (10 ng/ml) was added as detailed in the legend to Figure 5. The cells were incubated at 37°C in a humidified 5% C02/95 % air atmosphere. The assay was read unfixed and unstained at 14 days.

Acknowledgements We thank the Finnish Red Cross, Helsinki, and the Department of Virology, University of Uppsala, for the supply of blood platelets. These studies were supported by grants from the Swedish Cancer Society (689, 786 and 1794), the Swedish MRC (4486), Konung Gustav V:s attioarsfond and the University of

Uppsala.

References Assoian,R.K., Grotendorst,G.R., Miller,D.M. and Sporn,M.B. (1984) Nature, 309, 804-806. Betsholtz,C. and Westermark,B. (1984) J. Cell. Physiol., 118, 203 -211. Betsholtz,C., Heldin,C.-H., Nister,M., Ek,B., Wasteson,A. and Westermark,B. (1983) Biochem. Biophys. Res. Commun., 117, 176-182. Blobel,G. and Dobberstein,B.J. (1975) J. Cell Biol., 67, 835-851. Bowman,P.D. and Daniel,C.W. (1975) Mech. Ageing Dev., 4, 147-158. Chiu,I.M., Reddy,E.P., Givol,D., Robbins,K.C., Tronick,S.R., Aaronson,S.A. (1984) Cell, 37, 123-129. Collins,T., Ginsburg,D., Boss,J.M., Orkin,S.H. and Pober,J.S. (1985) Nature, 316, 748-750. Deinhardt,F. (1980) In Klein,G. (ed.), Viral Oncology. Raven Press, NY, pp. 357-398. Devare,S.G., Reddy,E.P., Law,J.D., Robbins,K.C. and Aaronson,S.A. (1983) Proc. Natl. Acad. Sci. USA, 80, 731-735. Dootlittle,R.F., Hunkapiller,M.W., Hood,L.E., Devare,S.G., Robbins,K.C., Aaronson,S.A. and Antoniades,H.N. (1983) Science, 221, 275-277. Garrett,J.S., Coughlin,S.R., Niman,H.L., Tremble,P.M., Giels,G.M. and Williams,L.T. (1984) Proc. Natl. Acad. Sci. USA, 81, 7466-7470. Glenn,K., Bowen-Pope,D.F. and Ross,R. (1982) J. Biol. Chem., 257, 5172-5176. Hannink,M. and Donoghue,D.J. (1984) Science, 230, 1197-1199. Heldin,C.-H. and Westermark,B. (1984) Cell, 37, 9-20. Heldin,C.-H., Westermark,B. and Wasteson, A. (198 la) Proc. Natl. Acad. Sci. USA, 78, 3664-3668. Heldin,C.-H., Westermark,B. and Wasteson, A. (198lb) Exp. Cell Res., 135, 255-261. Heldin,C.-H., Wasteson,A. and Westermark,B. (1982) J. Biol. Chem., 257, 4216-4221. Heldin,C.-H., Ek,B. and Ronnstrand,L. (1983) J. Biol. Chem., 258, 10054-10061. Heldin,C.-H., Betsholtz,C., Johnsson,A. and Westermark,B. (1986) Cancer Rev., in press.

The v-sis gene product as a PDGF receptor agonist Johnsson,A., Heldin,C.-H., Westermark,B. and Wasteson, A. (1982) Biochem. Biophys. Res. Commun., 104, 66-74. Johnsson,A., Heldin,C.-H., Wasteson, A., Westermark,B., Deuel,T.F., Huang,J.S., Seeburg,P.H., Gray,A., Ullrich,A., Scrace,G., Stroobant,P. and Waterfield,M.D. (1984) EMBO J., 3, 921-928. Johnsson,A., Betsholtz,C., von der Helm,K., Heldin,C.-H. and Westermark,B. (1985a) Proc. Natl. Acad. Sci. USA, 82, 1721-1725. Johnsson,A., Betsholtz,C., Heldin,C.-H. and Westeermark,B. (1985b) Nature, 317, 438-440. Josephs,S.F., Guo,C., Ratner,L., Wong-Staal,F. (1984) Science, 223, 487-490. Kaplan,P.L. and Ozanne,B. (1983) Cell, 33, 931-938. King,C.R., Giese,N.A., Robbins,K.C. and Aaronson,S.A. (1985) Proc. Natl. Acad. Sci. USA, 82, 5295-5299. Massague,J. (1985) Trends Biol. Sci., 10, 237-240. Ponten,J. (1970) Int. J. Cancer, 6, 323-332. Robbins,K.C., Antoniades,H.N., Devare,S.G., Hunkapiller,M.W. and Aaronson,S.A. (1983) Nature, 305, 605-608. Robbins,K.C., Leal,F., Pierce,J.H. and Aaronson,S.A. (1985) EMBO J., 4, 1783-1792. Roberts,A.B., Lamb,L.C., Newton,D.L., Sporn,M.B., De Larco,J.E. and Todaro,G.J. (1980) Proc. Natl. Acad. Sci. USA, 77, 3494-3498. Shin,S.-I., Freedman,V.H., Risser,R. and Pollack,R. (1975) Proc. Natl. Acad. Sci. USA, 72, 4435-4438. Spom,M.B. and Todaro,G.J. (1980) N. Engl. J. Med., 303, 878-880. Sporn,M.B. and Roberts,A.B. (1985) Nature, 313, 745-747. Stoker,G.M.P. (1973) Nature, 246, 200-203. Stroobant,P. and Waterfield,M.D. (1984) EMBO J., 3, 2963-2967. 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. (1983) Nature, 304, 35-39. Williams,L.T., Tremble,P.M., Lavin,M.F. and Sunday,M.E. (1984) J. Biol. Chem., 259, 5287-5294. Received on I April 1986

1541