[CANCER RESEARCH 53. 5.155-5360. November 15. ITO)

Advances in Brief

Met Expression and Sarcoma Tümorigenicity1

Sing Rong, Michael Jeffers, James H. Resau, lian Tsarfaty, Marianne Oskarsson, and George F. Vande Woude2

ABL-Basic Research Program. National Cancer Institute, Frederick Cancer Research and Development Center, Frederick, Maryland 21702

Abstract

The mei protooncogene tyrosine kinase receptor (Met) and its ligand,hepatocyte growth factor/scatter factor (HGF/SF), ordinarily constitute aparacrine signaling system in which cells of mesenchymal origin producethe ligand, which binds to the receptor that is predominantly expressed incells of epithelial origin. However, mouse NIH/3T3 fibroblasts overex-

pressing Met induce tumor formation in nude mice via an autocrinemechanism (S. Kong a al.. Mol. Cell. Biol., /_': 5152-5158, 1992). In this

study, we report that human cell lines established from various sarcomas

express high levels of activated Met receptor. HGF/SF is also detected inthe human sarcoma cell lines but at a reduced level when compared toprimary fibroblasts. These properties, high Met expression and reducedligand levels, are indistinguishable from the properties of NIH/3T3 tumorexpiant cells overexpressing Met (S. Rong n al.. Mol. Cell. Biol., 12:5152-5158, 1992; S. Rong et al.. Cell Growth & Differ., 4: 563-569, 1993).Moreover, paraffin-embedded sections of primary tumors from human

osteosarcomas, chondrosarcomas, and leiomyosarcoma stain intensely forMet and/or HGF/SF and display extensive tumor cell heterogeneity withregard to both paracrine and autocrine stimulation. On the basis of thesefindings, we propose that Mel-HGF/SF autocrine signaling may contribute

to the tumorigenic process in human sarcomas.

Introduction

Expression of the met protooncogene receptor tyrosine kinase (Met)occurs in a majority of adult and embryonic tissues, predominantly inepithelial cells (l^t). The ligand for Met is HGF/SF,3 a fibroblast-

derived mitogen for hepatocvtes as well as other cell types (5-7).

HGF/SF promotes cell movement and induces epithelial morphogenesis (8-10). Therefore, HGF/SF and Met can constitute a paracrine

signaling system in which cells of mesenchymal origin produceHGF/SF and this ligand hinds to the receptor predominantly expressedon cells and tissues of epithelial origin (11, 12).

The mei protooncogene is amplified in most spontaneous transformants of N1H/3T3 cells (13, 14) and we have shown that met complementary DNA efficiently transforms NIH/3T3 cells (1). More recently,we have demonstrated that mc/-induced NIH/3T3 cell tumorigenicity

is due to an autocrinc mechanism (15, 16). The highly tumorigenicbehavior of /Tier-transformed NIH/3T3 cells led us to examine human

primary fibroblast cells and cell lines established from various humansarcomas for Met expression and for evidence of autocrine stimulation. We find that Met is expressed at low levels in fibroblast cellswhich, in the presence of endogenously expressed HGF/SF, lead to anautocrine interaction. Our studies also show that Met overexpressionoccurs frequently in human sarcoma cells and tumors and could playan important role in their tumorigenicity.

Received 10/5/93; accepted 10/11/93.The costs of publication of this article were defrayed in part hy the payment of page

charges. This article must therefore he hereby marked adn-riisemenl in accordance with

18 U.S.C. Section 1734 solely lo indicale this fact.1 Research sponsored by the National Cancer Institute. DHIIS. under contraci NO1-

CO-74IOI with ABL. The contents of this publication do not necessarily reflect Ihe viewsor policies of Ihe Department of Health and Human Services, nor does mention of tradenames, commercial products, or organizations imply endorsement by the U.S. Government.

2 To whom requests for reprints should be addressed.( The abbreviations used are: HGF/SF, hepalocyte growth factor or scatter factor; SDS.

sodium dodecyl sulfate; rh, recombinant human; anti-P-Tyr, anti-phosphotyrosinc; Met1"",mouse Mel; Mel1"1,human met; HGFh", human HGF.

Materials and Methods

Cell Lines and Antibodies. Cell lines used in this study are listed in Table1. Most of the cell lines used in this study were obtained from American TypeCulture Collection and grown as recommended. A fibrosarcoma cell line, 8387(grown in Dulhccco's modified Eagle's medium with 10% bovine fetal serum),

and a rhabdomyosarcoma cell line, RD-1 (grown in McCoy's 5A with 15%

bovine fetal serum), were obtained from Doug Halverson (National CancerInstitute. Frederick, MD). NIH/3T3 cells transfected with near and mi'th" were

described by Rong et al. (15).The lyS anti-Met monoclonal antibody was generated against a bacterially

expressed p50 form of Met1"1(15. 17). The Met-specific C28 anti-peptide

antibody was raised by immunization of rabbits with the 28 amino acid COOH-

terminal peptide of human Met (18). A3.1.2 is a monoclonal antibody againsthuman recombinant HGF (15). 23C2 is a monoclonal antibody against humanplacental SF (19). Anti-rhHGF-1 is a rabbit polyclonal antibody against humanrecombinant HGF (10). Anti-rhHGF-2 is purified goat IgG against human

recombinant HGF (R&D Systems). 4G10 is a monoclonal phosphotyrosineantibody (anti-P-Tyr) (20).

Immunoprecipitation Analysis. Immunoprccipitation analysis for bothMet and HGF/SF was carried out as described previously (15).

Confocal Laser Scan Microscopy and Immunofluorescence Analysis.Immunofluorescence assays were performed as described previously (4, 16).

Western Immunoblot Analysis. Western analysis was done essentially asdescribed previously (15) except that the cells were lysed in RIPA buffer [1%Triton X-IOO, 1% sodium deoxycholate, 0.1% SDS, 0.15 M NaCl, 0.02 M

NaPOj (pH 7.2)]. containing 1.25 ITIMphenylmcthylsulfonyl fluoride, 2 ITIMEDTA, 50 mM NaF, 30 mM sodium pyrophosphate, 10 fig/ml aprotinin, 10fig/ml leupeptin. 1 mM sodium ortho-vanadate.

Scatter Assay. Scatter assay using MDCK cell line was carried out asdescribed previously (12, lo).

Northern Analysis. Total cellular RNA was isolated using RNAzoI asdescribed by the supplier (C1NNA/BIOTECX). Twenty /xg of total RNA weredenatured, electrophoresed on 1% formaldehyde agarose gel, and transferredto nylon membranes (Schleicher and Schuell) as described (1). Hybridizations were carried out for 2 days with 10" cpm/ml of probe (specific activity,10" cpm/ftg) (Random Priming Labeling Kit: Boehringer-Mannheim). Filters

were washed twice in 2 X standard saline citrate-0.1% SDS at room temperature for 10 min and then 3 times in 0.2 x standard saline citrate-0.\% SDS at55°C.

Mitogenic Assay. Three x It)3 cells were seeded into 96-well microtilerplates (Costar). After an overnight incubation at 37°C,these cells were starved

in serum-tree medium for 2 days. Different concentrations of purified HGFhu(Id) were added in the presence or absence of HGF1'1' neutralizing antibody(anti-rhHGF-2; R&D Systems; 1:20 dilution) and incubated overnight. |'Hj-

Thymidine was added at I piCi/well for 4 h and cells were lysed with O.I mlof 0.02 M NaOH-0.1% SDS. Aliquots of the lysate were used for scintillation

counting.

Results

Met and HGF/SF Expression in Human Fibroblast Cell Cultures and Sarcoma Cell Lines. Total RNA extracted from nonim-

mortalized human fibroblast cell cultures and from human cell linesestablished from various human sarcomas was analyzed for met RNAexpression using a full length met cDNA probe. The major metmRNA, a 9-kilobase transcript (2), was present in all of the samples

tested (Fig. \A; Table 1). We also examined these cells and cell linesfor Met protein expression by immunoblot analysis. Met was precipi-

5355

on April 13, 2017. © 1993 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Mel EXPRESSION AND SARCOMA TUMORICÃŒLNIOTY

Table 1 Met und ÕK/F/SF expression in human fihroblast and sarcoma cell lines

•¿�•§•'t •¿�123456789 10

9 kb

Br „¿�^ # ¿_/K.«s-

rfJVS*///4«•¿�p140

—¿�-»—p140

123456789 10 11

///**&//&/•¿�•¿�>w•¿�i

1 2 3 4 5 6 7 8 9 10 11 12 13

Fig. I. Met and HCïF/SFexpression in human cells. In A, 20 ¿igof total RNA wereloaded/lane for Northern analysis and full-length /«(•/'"'cDNA was used as probe. Lane I,

HEL299; Lane 2, Hems; Lane.?, Hs68; Lane •¿�/,8387; Lane 5, HTIOXO; Lane ft, Hs913T;Lane 7, HOS; Lane 8, Saos-2; Lane 9, V-2 OS; Lane W. RD. In B, l mg of cell lysate wasimmunoprecipitated with anti-C'28 peplide antibody, followed by SDS-polyacrylamide gel

electrophoresis and immunohlotting with a Met monoclonal 19S antibody («)or anti-P-Tyr antibody (h). Lane 1. Hems; Lane 2, HT1080: Lane .1. HEL299; Lane 4. V-2 OS; LaneS, Saos-2; Lane ft, RD; Lane 7, SW872; Lane 8, Hs68; Lane 9. 8387; Lane 10, RD-1 ; Lane11, Hs913T. In C. cells were metabolic-ally labeled with [15S]melhionine and |'5S]cysteine(Translabel; ICN) for 6 h. One ml of supernatant was concentrated 10-fold in a C'entricon

apparatus (Amicon; IOK cutoff); the volumes were adjusted to 0.35 ml with RIPA bufferfor immunoprecipitation with HGF monoclonal antibody A3.1.2. Lane l, HEL299; Lane2, Hems; Lane .1. HsoK; Lane 4, RD; Lane 5. RD-1; ¿am-6, SW872; Lane 7. HOS: Lane8, U-2 OS; Lane 9, Saos-2; Lane IO. HTI080; Lane 11, 8387; Lane 12, Hs913T; Lane 13,

SW684.

ScatterMet" mei1' HGF/SF' activity''P-Tyr-Met''Human

diploidfibrohlastHEL299 (fetallung)Hems(fetalmuscle)Hs68(newbornfibroblast)Malme-3

(skinfibroblast)Fibrosarcoma8387HTI080Hs913TSW684Osteogcnic

SarcomaHOSSaos-2U-2

OSChondrosarcomaSW1353RhabdomyosarcomaR

DRD-1A204A673Hs729LeiomyosarcomaSK-LMS-ISK-UT-1BLiposarcomaSW872Mesoderma!

tumorSK-UT-1Synovial

sarcomaSW982MelanomaMalme-3MWM1I5WM266-4+

+ + + + + + + ++++ + + + + + + + ++++ ++ + + + +++

ND + + + ++++++

++++ + ++++ + + + + + ++++ + + ++ + + + ++++ + + ND+++

+ ++++ + + + + + ++++ ++++

NDND+

+ + +++++ + ++NDND+

ND++ + ND +++

+ + + ND +++ND++++

+ + +ND++++

ND+

NI) + +++

+ +ND+ND++ND --•f-NI)+

++++++

+ +++++++

+++NI)+

+__++

+++

+ +++-+

++++

+ ++-

" Met protein level was assessed by immunoprecipilation and Western analysis. -, not

detected.''mei gene expression was detected by Northern analysis. ND, not determined.' HGF/SF protein level was assessed by immunoprecipitation analysis.'' Scatter activity was assayed with MDCK cells.'' P-Tyr-Met was analyzed by Western analysis with anti-P-Tyr.

tated from cell lysates using C28 anti-peptide antibody directedagainst the COOH-terminus of the human Met (18). The immunopre-cipitates were subjected to SDS-polyacrylamide gel electrophoresis,

followed by immunoblot analysis using the 19S monoclonal antibody, directed against the intracellular Met domain (15, 17) (Fig.Iß).Low levels of Met protein, pl40MlM, and its precursor, pl7()McI,

were detected in several of the primary fibroblast cultures (Fig. Iß,a, Lanes 1 and 8; Table 1), but much higher levels were present inmany of the human sarcoma cell lines tested (Fig. Iß,a. Lanes 2,4-7, 9-/7; Table 1). These levels were similar to the high levels of

Met protein observed in N1H/3T3 cells transformed by met (Ref. 16;data not shown). These analyses also showed that there was no directcorrelation between the level of met mRNA detected (Fig. 1/4) andthe level of Met protein expressed (Fig. Iß,a). For example, the levels of met mRNA detected in 8387 and Saos-2 cells were higher than

the levels detected in HT1080 cells; however, the HT1080 cells express higher levels of Met. Likewise, equivalent amounts of metRNA were detected in Hems and HT1080, but the level of Met inthe Hems fibroblasts was barely detectable (Fig. Iß,a). These results showed significant variation in the regulation of Met protein

Neutralizing ér NT'o;Antibody: c> <•<•<•

kDa200 —¿�

92.5— —¿�

69 —¿�

46 —¿�

1 234 56Fig. 2. HGF neutralizing antibody increases Met protein abundance. Primary fibroblasl

HEL299 (Lanes 1—4)and Hems (Lanes 5-6) cells were incubated for 48 h with or withoutHGF neutrali/ing antibody (anti-rhHGF-1): no antibody (Lanes I and 5); I:4(HM)(Lanes2 and A), 1:1(MK)(Lane 3), and 1:250 (Lane 4) dilutions (antibody was added at 0 timeand at 24 h). Cells were lysed in RIPA buffer and KM)fig of protein were resolved by 7,5%SDS-polyacrylamide gel electrophoresis and immunohlotted with WS anti-Methu mono

clonal antibody.

5356

on April 13, 2017. © 1993 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

§O

Ee

toe

400-

350-

300

250

ZOO-

150-

100

SO

o

3T3 neo

Mel KXPRliSSION AND SARCOMA TUMORIGEN1C1TY

400-

350-

300

250

200

150-

100-

50-

0

ooo

B

EeBu

B 3T3 humet

TIO 0.25 1 5 20 40 80 120

HGF/SF (units/ml)

O 0.25 1 5 20 40 80 120

HGF/SF (units/ml)

»-. 150Ooo

ce

1 DO

SK-LMS-1 (-Ab)Oeo

so

T

SK-LHS-1 (»Ab)

20 40 80 120 120

HGF/SF (units/ml) HGF/SF (units/ml)Fig. 3. Mitogcnic responses nf NIH/3T3 fibroblasls and SK-LMS-1 human sarcoma cell line to HCjF/SF: stimulation. In A and H. NIH/3T3 cells transfected with imi' or with nm'

plus met1'" complementary DNA (15) were seeded into Wi-well microtiter plates (Costar). After 2 days of scrum starvation, the indicated concentrations of HOF/SP1" were added. After4 h with | 'H|thymidinc, the cells were lysed and aliquots measured by scintillation counting. Each sample was done in triplicate. Normali/cd values of [ (H]thymidine incorporationare shown. In (' and I), SK-LMS-1 human sarcoma cells were seeded and serum-starved as in A. The indicated concentrations of HGF/SF were added in the presence or absence ofIIGF neutralising antibody (anti-rhH(iF-2) and incubated overnight. ['H]thymidinc labeling was performed as in A.

expression, although we could not exclude the possibility that thedifference in steady-state Met levels also reflects ligand-mediateddown-modulation of the receptor. The high levels of Met in the sar

coma cell lines and its presence in fibroblast cells represented novelfindings, since Met expression was thought to be preferentially present in epithelial cells, while only the ligand was restricted to mes-

enchymal cells (11, 21).We determined the levels of immunoprecipitable HGF/SF using

growth medium harvested from metabolically labeled fibroblast andsarcoma cells (Fig. 1C). Abundant levels of the HGF/SF pM a-sub-unit (p69IK'"sl') were observed in all of the primary fibroblast cell

cultures (Fig. 1C, Lanes 1-3; Table 1), but only one sarcoma cell line

(Hs913T) secreted high levels (Fig. 1C, Lane 12). The level ofHGF/SF was also determined by scatter assays performed on growthmedium that was conditioned on confluent cells for 72 h. Comparableto the high levels of poV1"'1 sl detected, high levels of scatter activity

were detected in conditioned medium from HFX2W and Hems fibroblast cultures (Table 1). However, lower activity was observed in

the scatter assays performed with Hs68 and Malme-3 conditionedmedium that did not correlate with the high levels of p69IK'"SI' de

tected (Fig. 1C, Lane 3; Table 1). The Hs°13T fibrosarcoma cellsalso express high levels of p69"iil'VSF (Fig. 1C, Lane 12), and also

exhibit low scatter activity. However, with the exception of Hs913Tcells, the levels of HGF/SF and scatter activity were low in cells thatexpressed high levels of Met (Table 1). A similar marked reductionin the endogenous HGF/SF was observed in NIH/3T3 cells overex-pressing Met"'" that was presumably due to the depletion of the li

gand by the receptor (16). Evidence for the HGF/SF receptor activation was indicated by the high reactivity of Met""1 with anti-P-Tyr

antibody (15). We, therefore, performed anti-P-Tyr immunoblot

analyses on the Met expressed both in the human fibroblast cell cultures and in the sarcoma cell lines (Fig. \B,b). These analysesshowed that, in general, the overexpressed Met is highly reactivewith anti-P-Tyr antibody (Fig. \B,b, Lanes 2, 4-6, 9, 10; Table 1),similar to Met"1" in NIH/3T3 cells (15). We also found that Met1"1

was weakly reactive with anti-P-Tyr antibody in the primary Hems

5357

on April 13, 2017. © 1993 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

•¿�[•¡g.4. Confoca! analysis of human sarcoma (issues. Paraffin-embedded human lunnir sections of A, leiomyosarcoma; /?, chondrosarcoma; C and /), two different osteosarcomas.

In each panel (A-Ü), the insels are: /, Nomarski; 2, Met staining: and .?. HGF/SF staining. Each paraffin section was double-stained with anti-Mel C28 pcptide antibody (fluoresceinisolhiocynate) and monoclonal anti-HGF/SF antibody 23C2 (Rhodamine). The immunofluorescence was determined by confocal laser scan microscopy and color is generated byconfocal laser scan microscopy look-up tables. In the insets (2 and .?), yellow-red is used to characterize Met and HGF. respectively. High intensity is indicated by yellow, red is lower,and black is background. In the overlays, panels A-D. Met staining is green. HGF/SF is red. and yellow staining corresponds to cells that are stained and express both Met and HGF/SF.Magnification, X 1000.

fibroblasts (Fig. \B,b, Lane /), suggesting that the receptor was activated in an autocrine fashion by endogenous HGF/SF.

Autocrine Interaction of Met and HGF/SF in Primary Fibro-blast Cells. We tested the possibility that an autocrine Met-HGF/SF

stimulatory pathway exists in the primary fibroblast cell cultures.Hems and HEL299 cells expressed low levels of Met and high levelsof HGF/SF compared to other cell lines tested (Fig. Iß;data notshown). To test whether the level of Met was being down-modulatedby HGF/SF, we added anti-rhHGF-1, a neutralizing HGF/SF antibody

(10), to the growth medium of HEL299 and Hems cells. After 48 h,cell lysates were analyzed for levels of Met by immunoblot analysis(Fig. 2). These analyses indicate that there is a significant increase inthe amount of p!40Mcl in HEL299 cells (Fig. 2, Lanes l^t) and Hems

cells (Fig. 2, Lanes 5-6) in the presence of the antibody. These resultsindicate that Met is down-regulated via extracellular-autocrine acti

vation. The requirement for HGF/SF to be activated by extracellularproteolytic cleavage (22, 23) was consistent with these results.

Mitogenic Response of Sarcoma Cells to HGF/SF. Met-HGF/SF

signaling has been implicated in both mitogenic and motogenic activities for epithelial cells (21). As a control for HGF/SF mitogenicity,we measured ['Hjthymidine incorporation in NIH/3T3 cells overex-pressing Met1"' (15) in response to the addition of exogenous HGF/

SFhu (Fig. 3). A 9-fold stimulation of [3H]thymidine incorporation was

observed when 5 units/ml of exogenous HGF/SF was added to NIH/3T3 cells overexpressing Methu (Fig. 3ß),but not to control NIH/3T3cells (Fig. 3A). These analyses showed that HGF/SFhu was mitogenicfor NIH/3T3 fibroblast cells overexpressing Mcthu; but, at high levels,the ligand was inhibitory. HGF/SF also stimulated [3H]thymidine

incorporation in the human sarcoma cell line SK-LMS-1 (Fig. 3C) andthis stimulation was prevented by anti-rhHGF-2 neutralizing antibody(Fig. 3D). Curiously, much higher levels of HGF/SFhu (16X) are

required for stimulation of the SK-LMS-1 cells compared to NIH/3T3cells overexpressing Methu. Two other cell sarcoma cell lines, HOSand RD, also showed increased ['H]thymidine incorporation in re

sponse to 80-120 units/ml HGF/SF; while SK-UT-1 and U-205 cellsdid not respond (data not shown) even though they express Meth" andHGF/SF was not detected (Table 1). We concluded that HGF/SFhu can

elicit a mitogenic signal in sarcoma cells expressing the Met receptor.The low (or lack of) response to HGF/SF in the sarcoma cells suggested the presence of an inhibitor or mutation that interrupts Metsignaling.

Expression of Met and HGF/SF in Human Primary Tumors.The elevated expression of Met in sarcoma cell lines compared toprimary fibroblust cultures and our previous demonstration that NIH/

5358

on April 13, 2017. © 1993 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

MCI EXPRESSION AND SARCOMA TUMORIGENIC1TY

3T3 cells overexpressing Met are tumorigenic (15) suggested that thehigh expression of Met may have contributed to the formation ofsarcomas in vivo. We therefore examined paraffin-embedded human

sarcoma sections stained for Met and HGF/SF by confocal laser scanmicroscopy. Seven of eight tumors examined were positive for bothMet and HGF/SF staining: one leiomyosarcoma (Fig. 4A) and one oftwo chondrosarcoma examined (the positive tumor is shown in Fig.4ß)expressed both Met and HGF/SF, while three osteosarcomasshowed significant Met and HGF/SF staining (two tumors are shown;Fig. 4C and D). For each sample, the Nomarski images are presentedin panel 1, whereas panels 2 and 3 show the tumor sections stainedwith anti-Met and anti-HGF/SF antibody, respectively. In the double-stained overlays (Fig. 4A-D), green corresponds to Met staining, red

corresponds to HGF/SF, and yellow represents colocalization of Metand HGF/SF staining. In each tumor, we observed cells which arepositive either for both Met and HGF/SF or for Met or HGF/SF alone,suggesting that both autocrine and paracrine modes of stimulation canoccur. The differential pattern of Met and HGF/SF expression observed in the sarcoma cells suggests heterogeneity in the population oftumor cells which might reflect the state of cell differentiation and/ortumor progression. These analyses demonstrate that Met overexpres-

sion occurs in the primary human sarcomas as well as in sarcoma celllines (Table 1) and therefore, as in the NIH/3T3 model system (15),may contribute to the formation of these tumors.

Discussion

Cells that synthesize growth factor(s) and express the cognate re-ceptor(s) have the potential for autocrine-mediated growth (24). In an

external autocrine loop, receptor binding and signal transduction occurs when a growth factor is secreted and subsequently interacts with

cells. This combination of responses could be causally associated withthe development of sarcomas.

Similar to our finding that endogenous HGF/SF expression is dramatically lowered in the NIH/3T3 Metmu tumor cells (16), the level of

HGF/SF is low in most of the human sarcoma cell lines tested presumably due to the overabundance of Met receptor (16). However, wecannot exclude the fact that other factors are responsible for reducedHGF/SF expression in the sarcoma lines; e.g., expression of HGF/SFin the MRC-5 fibroblast cell line can be positively or negatively

affected by conditioned medium from several carcinoma cell lines (29,30) and HGF/SF expression in MRC-5 cells can be inhibited by

transforming growth factor ß,epidermal growth factor, and transforming growth factor a (29, 30). We have not tested the human sarcomacell lines for these factors.

Spontaneous transformants of NIH/3T3 fibroblasts were frequentlyfound to have met protooncogene amplified and overexpressed (13,14) and Met overexpression in NIH/3T3 fibroblasts induces fibrosar-

comas via an autocrine mechanism (1, 15, 16). The results presentedhere suggest that a Met-HGF/SF autocrine mechanism may also con

tribute to the tumorigenic process in human sarcomas.

Acknowledgments

We are grateful to Toshikazu Nakamura for A3.1.2 and anti-rhHGF-1 anti

bodies: Eliot Rosen for 23C2 antibody; Deborah Morrison for 4G10 antibody; and John Cottrell for providing human sarcoma paraffin sections. Wealso thank Monica Murakami, Janellc Conner, and Han-Mo Koo for their

critical reading of the manuscript. We thank Michelle Reed for typing themanuscript.

References

its receptor on the surface of the secreting or neighboring cells (25). 1. Iyer,A.,Kmiecik,T. E.,Park,M.,Dear,I., Blair,P.,Dunn,K.J., Sutrave,P.,Ihlc,J.Met and its ligand, HGF/SF, were first shown to interact in a paracrinefashion (21). The ligand is produced by cells of mesenchymal origin,such as fibroblasts and smooth muscle cells, and this ligand caninduce both mitogenic, motogenic, and morphogenic responses inepithelial cells that express the Met receptor (11, 12, 26). In additionto this paracrine model, we have recently demonstrated by overexpressing Met1"1and HGF/SFhu in NIH/3T3 cells that autocrine-medi

ated signaling leads to tumorigenesis (15, 16). Here, we provideevidence that endogenous Met-HGF/SF autocrine signaling mediatesa mitogenic response in cells of mesenchymal origin. First, nonim-

mortalized human primary fibroblast cultures (HEL299; Hems) whichproduce abundant HGF/SF express low levels of Met, and Met prepared from one of these cultures (Hems) reacts with anti-P-Tyr. More

over, Met is increased in HEL299 and Hems cells when HGF/SFneutralizing antibody is added to the medium. This suggests that theMet receptor may be down-modulated by the abundant HGF/SF pro

duced by these cells. Consistent with this interpretation, these cells donot respond mitogcnically when exogenous HGF/SF is added (datanot shown), while the SK-LMS-1, HOS, and RD human sarcoma celllines expressing high levels of Methu do. Thus, similar to platelet

derived growth factor, fibroblast growth factor, and insulin-likegrowth factor I (27), Met-HGF/SF autocrine interaction may be a

fundamental property of fibroblast cell mitogenesis in vitro. It has alsobeen suggested that the HGF/SF may play a role in the determinationof fibroblast morphology (8, 21). Thus, fibroblasts have extensivepscudopodial extensions and are spontaneously motile, resembling themorphology and motility of epithelial cells exposed to HGF/SF. Furthermore. Met-HGF/SF signaling induces motogenic responses inNIH/3T3 fibroblasts (28),4 suggesting that this autocrine loop con

tributes to both mitogenic and motogenic phenotypes of mesenchymal

4 S. Rong el al., manuscript in preparation.

N.. Bodescot, M., and Vandc Woude, G. F. Structure, tissue-specific expression, andtransforming activity of the mouse mei proto-oncogene. Cell Growth & Differ., /:87-95, 1990.

2. Park. M.. Dean. M., Cooper. C. S.. Schmidt, M., O'Brien, S. J., Blair, D. G., and

Vande Woude. Ci. F. Mechanism of mei oncogene activation. Cell. 45: 895-904.19X6.

3. DiRenzo, M. F., Narsimhan, R. P.. Olivero. M., Bretli. S.. Giordano, S., Medico. E..Gaglia. P.. Zara, P., and Comoglio. P. M. Expression of the Met/HGF receptor innormal and neoplastic human tissues. Oncogene, 6: 1997-2003. 1991.

4. Tsarfaty, I., Resau, J. H., Rulong. S.. Keydar, I.. Falcilo, D. L., and Vande Woude. G.F. The mîtproto-oncogene receptor and lumen formalion. Science (Washington DC).257: 1258-1261. 1992.

5. Bottaro. D. P., Ruhin. J. S.. Paletto, D., Chain. A. M-L, Kmiecik, T. E., Vande Woude.G. F., and Aaronson, S. A. The hepatocyte growth factor receptor is Ihe c-melproto-oncogene product. Science (Washington DC). 152: 802-804, 1991.

6. Naldini, L... Vigna. E.. Narsimhan. R. P., Gandino. G.. Zarncgar, R.. Michalopoulos,G. K., and Comoglio. P. M. Hepatocyte growth factor (HGF) stimulates the tyrosinekinase activity of the receptor encoded by the proto-oncogene c-met. Oncogene, ft:501-504, 1991.

7. Nakamura, T., Nawa, K.. and Ichihara A. Partial purification and characterisation ofhepatocyte growth factor from serum of hepatectomizcd rats. Biochem. Biophys. Res.Commun., 122: 1450-1459, 1984.

8. Stoker. M. J. Effect of scatter factor on molility of epithelial cells and fibroblasto. CellPhysiol., 139: 565-569. 1989.

9. Rosen, E. M., Goldberg, I. D., Kacinski. B. M., Buckhdz. T., and Vinler, D. W.Smooth muscle releases an epithelial scatter factor which binds to heparin. In VitroCell Dev. Biol.. 25: 163-173, 1989.

10. Montesano. R.. Matsumoto. K.. Nakamura, T. and Orci, L. Identification of a fibro-blast-dcrived epithelial morphogcn as hcpatocyte growth factor. Cell. 67: 901-908,1991.

11. Birchmeier. C., Sonnenberg. E., Weidner, K. M., and Walter. B. Tyrosine kinasereceptors in the control of epithelial growth and morphogenesis during developing.BioEssays, 15: 1-6, 1993.

12. Slokcr, M., Gherardi. M.. and Perryman. M. Scatter factor is a fibroblast-derivedmodulator of epithelial cell mobility. Nature (Lond.). 327: 239-242, 1987.

13. Cooper. C. S., Tempest. P. R.. Beckman. P. M.. Heldin, C-H., and Breakers, P.Amplification and overexpression of met gene in spontaneously transformed NHI/3T3mouse fihroblast. EMBO J.. 5: 2623-2628. 1986.

14. Hudziak. R., Lewis, G. D.. Holmes, W. E.. Ullrich. A., and Shcpard. H. M. Selectionfor transformation and met proto-oncogene amplification in NIH/3T3 fibroblastsusing tumor neurosis factor a. Cell Growth & Differ.. /: 129-134, 1990.

15. Rong. S., Bodescol, M.. Blair. D.. Nakamura. T. Mizuno, K., Park, M.. Chan. A.,Aaronson. S., and Vande Woude, G. F. Tumorigenicity of Ihe met proto-oncogene andthe gene for hepalocylc growth factor. Mol. Cell. Biol., 12: 5152-5158, 1992.

5359

on April 13, 2017. © 1993 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

MCI EXPRESSION AND SARCOMA TUMORIGENICITY

16. Rong. S., Oskarsson, M., Paletto, D. L.. Tsarfaty, I., Resau,J., Nakamura, T., Rosen,I:., Hopkins, R., and Vande Woude G. F. Tumorigenesis induced by coexpression ofhuman hepatocyte growth factor and the human met proto-oncogene leads to highlevels of expression of the ligand and receptor. Cell Growth & Differ., 4: 563—569,

1993.17. Paletto, D. L„Tsarfaty, !.. Kmiecik, T. E., Gonzatti. M., Suzuki, T., and VandeWoude,

G. F. Evidence for noncovalent clusters of the c-rnt't proto-oncogene product. Oncogene, 7: 1149-1157, 1991.

18. Gonzatti-Haccs, M.. Seth,A., Park, M., Copcland. T., Oroszlan, S.. and VandeWoude.G. F. Characterization of the tpr-mei oncogene p65 and the mei proto-oncogene p14(1protein-tyrosine kinase. Proc. Nati. Acad. Sci. USA, 85: 21-25, 1988.

19. Bhargava, M.. Joseph,A., Kncsel, J., Halaban, R.. Li. Y.. Pang, S., Goldberg, I.. Setter,E., Donovan. M. A.. Zarnegar, R.. Michalopoulos, G. A., Nakamura, T., Paletto. D..and Rosen, E. M. Scatter factor and hepatocyte growth factor: activities, properties.and mechanism. Cell Growth & Differ.,.?: 11-20, 1992.

2(1. Morrison, D. K., Kaplan, D. R., Escobedo, J. A.. Rapp. J. R.. Roberts. T. M., andWilliams. L. T. Direct activation of the scrine/threonine kinasc activity of Ruf-ithrough lyrosine phosphorylation by the PDGF-ßreceptor. Cell, SS: 649-657, 1989.

21. Gherardi, E.. and Stoker. M. Hepalocvlc growth factor-scatter factor: mitogen, mo-togen, and Mel. Cancer Cells, 3: 227-232. 1991.

22. Miyazawa, K., Shimomura. T., Kitamura, A., and Kondo. J. Molecular cloning andsequenceanalysis of Ihe cDNA for a human serine proteaseresponsible for activationof hepalocyte growth factor. Structural similarity of the protease precursor to blood

coagulation factor XII. J. Biol. Chcm., 26«;10024-1(X)28, 1993.

23. Naka. D.. Ishii, T., Yoshiyama, Y., Miyazawa, K. Activation of hepatocyte growthfactor by protcolytic conversion of a single chain form to a heterodimer. J. Biol.Chem.. 267: 20114-20119, 1992.

24. Sporn, M. B., and Roberts,A. B. Autocrine growth factors and cancer.Nature (Lond.).313: 745-747, 1985.

25. Browdcr, T. M.. Dunbar, C. E.. and Nienhuis. A. W. Private and public autocrinc loopsin ncoplastic cells. Cancer Cells, /: 9-17, 1989.

26. Vande Woude, G. F. Hepatocyte growth factor: Mitogen. motogen, and morphogen.Jpn. J. Cancer Res., «.?.•1992.

27. Baserga, R.. Porcu, P..and Sell. C. Oncogenes, growth factors and control of the cellcycle. Cancer Surv., 16: 201-213, 1993.

28. Giordano, S., Zhen, Z., Medico, E., and Caudino, G. Transfer of motogenic andinvasive response to scatter factor/hepatocyte growth factor by transfection of human MET prolooncogene. Proc. Nati. Acad. Sci. USA, 90: 649-653, 1993.

29. Seslar, S. P.. Nakamura. T.. and Byers, S. W. Regulation of fibrohlast hepatocytegrowth faclor/scatter factor expression by human breast carcinoma cell lines andpeptide growth factors. Cancer Res., 53: 1233-1238, 1993.

30. Kamalati. T.. Thirunavukarasu. B., Wallace, A., Holder. N-, Brooks, R.. Nakamura, T.,Stoker. M.. Gherardi, E.. and Bulumela. L. Down-regulation of scatter factor inMRC-5 fibroblasts by epithelial-derived cells. A model for scatter factor modulation.J. Cell. Sci., 101: 323-332, 1992.

5360

on April 13, 2017. © 1993 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

1993;53:5355-5360. Cancer Res   Sing Rong, Michael Jeffers, James H. Resau, et al.   Met Expression and Sarcoma Tumorigenicity

  Updated version

  http://cancerres.aacrjournals.org/content/53/22/5355

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

  .permissions@aacr.orgDepartment at

To request permission to re-use all or part of this article, contact the AACR Publications

on April 13, 2017. © 1993 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from