The MET oncogene was causally involved in the pathogenesis of a rare tumor, i.e., the papillary renal cell carcinoma, in which activating mutations, either germline or somatic, were identified. MET activating mutations are rarely found in other human tumors, whereas at higher frequencies, MET is amplified and/or overexpressed in sporadic tumors of specific histotypes, including osteosarcoma. In this work, we provide experimental evidence that overexpression of the MET oncogene causes and sustains the full-blown transformation of osteoblasts. Overexpression of MET, obtained by lentiviral vector–mediated gene transfer, resulted in the conversion of primary human osteoblasts into osteosarcoma cells, displaying the transformed phenotype in vitro and the distinguishing features of human osteosarcomas in vivo. These included atypical nuclei, aberrant mitoses, production of alkaline phosphatase, secretion of osteoid extracellular matrix, and striking neovascularization. Although with a lower tumorigenicity, this phenotype was superimposable to that observed after transfer of the MET gene activated by mutation. Both transformation and tumorigenesis were fully abrogated when MET expression was quenched by short-hairpin RNA or when signaling was impaired by a dominant-negative MET receptor. These data show that MET overexpression is oncogenic and that it is essential for the maintenance of the cancer phenotype. (Cancer Res 2006; 66(9): 4750-7)

The MET oncogene encodes the tyrosine kinase receptor for the hepatocyte growth factor (HGF), which in vitro and in vivo elicits a unique physiologic program leading to morphogenesis, known as “invasive growth” (1). If deregulated, the invasive growth program contributes to cell transformation and tumor progression (2). The link between the MET oncogene and cancer has been unequivocally established after the identification of mutations in families suffering from a rare familial cancer, i.e., the hereditary papillary renal cell carcinoma (3). Thus far, MET mutations were also found in <20% of sporadic papillary renal cell carcinoma (4) and, at even lower frequency, in other human primary tumors (59) and metastases (7). Conversely, at higher frequencies, the MET oncogene is amplified and/or overexpressed in sporadic human tumors of specific histotypes (2). MET overexpression is associated with gene amplification in renal (10, 11), gastric (12), and colorectal cancers (13), and in gliomas (14). In other cancers, MET overexpression is attained by mechanisms other than gene amplification. For instance, other oncogenes, such as activated RAS (15), can induce MET overexpression, and increased transcription might also be caused by hypoxia (16).

The receptor encoded by the MET oncogene is expressed mainly in epithelial cells, whereas its ligand HGF is normally produced and secreted by cells of mesenchymal origin, suggesting that this ligand/receptor couple represents a paracrine signaling system for the mesenchymal-epithelial interaction in physiologic and pathologic conditions. Although barely detectable in adult tissues derived from the mesenchyme, during development, the MET receptor is expressed and has a nonredundant role in the migration of myogenic precursors (17), and possibly contributes to growth and/or survival of blood cells (18).

It is noteworthy that the MET proto-oncogene was originally identified as a transforming oncogene in a human osteosarcoma cell line (MNNG-HOS), which had acquired tumorigenicity after treatment with a chemical carcinogen (19). We and others showed that MET is misexpressed in mesenchymal tumors and that the highest levels of the receptor are detected in human osteosarcomas (>80%) in which MET expression correlates with an aggressive phenotype and poor prognosis (2023). Therefore, we studied the contribution of MET receptor overexpression and activation in the transformation of human osteoblasts. In this work, we show that overexpression of the MET oncogene, at levels mimicking those found in human osteosarcomas, drives the conversion of human primary cultured osteoblasts into osteosarcoma cells, and is essential for the persistence of the transformed phenotype.

Human primary osteoblasts, tissue samples, and cell lines. Primary cultures of human osteoblast-like cells (HOB) were harvested from skull and appendicular bones of 6-month to 65-year-old patients treated for nonneoplastic diseases. Thirteen preparations were harvested from bone obtained from the pathologists, and consisted mostly of trabecular bone. As controls, two osteoblast preparations were purchased from PromoCell, Heidelberg, Germany. Fragments were repeatedly washed in PBS and collagenase at low concentrations to remove bone marrow, and then treated with 0.25% collagenase (type I, Sigma, St. Louis, MO) for 2 hours at 37°C. Outgrowing cells were cultured in DMEM with 10% fetal calf serum. Alkaline phosphatase content was determined on cell cytospins by a cytochemical method (86R, Sigma). To analyze the osteogenic potential of HOB cells, cultures were also separately maintained for 3 weeks in differentiating medium supplemented with β-glycerophosphate (10 mmol/L), ascorbic acid (50 μg/mL), and dexamethasone (10−8 mol/L). The formation of mineralized calcium phosphate deposits was estimated by histochemical assay using alizarin red staining. Samples of human normal and tumor tissues were harvested and snap-frozen at surgery by the pathologists and stored at −80°C until RNA and protein extraction. The GTL-16 gastric carcinoma cell line has been previously described (24). The other cell lines were purchased from the American Type Culture Collection (Manassas, VA). Anchorage-independent growth was determined in 0.3% agarose (SeaPlaque; FMC Bioproducts, Rockland, MA) after 21 days culture. Colonies stained with 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (0.5 mg/mL, Sigma) were scored by two observers from three plates.

Lentiviral vector production and transduction. Replication-defective lentiviral vectors (LV) were generated by transiently transfecting 293-T cells with three separate plasmids. We used self-regulated LV carrying the expression cassettes for the tet-transactivator and for the transgene of interest within a single backbone (25) and bicistronic vectors, with the internal ribosome entry site of the encephalomyocarditis virus. The plasmids used included, the second-generation packaging construct which encodes the HIV-1 Gag and Pol precursors, as well as the regulatory proteins, Tat and Rev (pCMVΔR8.74), the VSV-G expressing construct, pMD.G, and the appropriate transfer construct. Dominant-negative MET receptor (DN-MET) vectors were generated as previously described (26). MET-specific short-hairpin RNA (MET-shRNA) sequences are as reported (16). We used the lentiviral transfer vector containing the polypurine tract of the pol sequences that acts in cis and enhances nuclear import and reverse transcription (25) in one transcriptional unit. Packaging 293-T cells were seeded in 10-cm diameter dishes 24 hours prior to transfection in Iscove's modified Dulbecco culture medium (JRH Biosciences, Inc., Lenexa, KS) with 10% fetal bovine serum, penicillin (100 IU/mL), and streptomycin (100 μg/mL). Serial dilutions of freshly harvested vector stocks were used to infect 105 cells in a six-well plate in the presence of Polybrene (8 μg/mL). The viral p24 antigen concentration was determined by HIV-1 p24 Core profile ELISA (Perkin Elmer/NEN, Wellesley, MA) to determine the amount of infective particles before transduction and to show that transduced cells did not produce viral particles after transduction. Short-term cultures of HOB (from two to six population doublings) were transduced.

Southern and Northern blot analyses, end-point and real-time reverse transcription-PCR. Analyses were carried out as previously described (7). End-point reverse transcription-PCR for the detection of MET- and HGF-specific mRNA was carried out as described (7, 20). Quantitative “real-time” reverse transcription-PCR with TaqMan detection (ABI Prism 7700 Sequence Detection System, Perkin-Elmer Applied Biosystems, Foster City, CA) for the measurement of wild-type and mutant MET-specific transcripts were carried out as described (7).

Immunoprecipitation and Western blot analysis. Immunoprecipitation and Western blot analysis to detect MET protein and tyrosine phosphorylated proteins were carried out as described (27). Polyclonal MET antibody (C-12, 1:100; Santa Cruz Biotechnology, Santa Cruz, CA), used for Western blot analysis, was raised against a peptide corresponding to 12 COOH-terminal amino acids of the c-MET human sequence; monoclonal antibodies DO-24 and DL-21 were used for immunohistochemistry and immunoprecipitation, and Western blotting, respectively, are directed against the extracellular domain of the MET protein.

Invasion assays in Matrigel and in three-dimensional collagen matrix. Matrigel invasion assay was done as described (27). To perform three-dimensional collagen matrix invasion assays, spheroids were generated as follows: 600 to 800 cells were suspended in culture medium containing 20% FCS and 0.24% methylcellulose (M-0512, Sigma), and seeded in round-bottomed 96-well plates. After overnight incubation at 37°C in a humidified atmosphere, all suspended cells contribute to the formation of a single spheroid. Thirty-six spheroids were harvested in each test tube, centrifuged for 15 minutes at 300 × g and resuspended in M-199 medium containing 0.696 μg/μL Rat Tail Collagene, type I (BD Biosciences, San Jose, CA), 0.016% methylcellulose, and 20% FCS. The spheroid-containing gel was rapidly transferred into each well of a prewarmed 96-well plate and allowed to polymerize for 1 minute. Fresh M-199 medium was added on gel tops and gels were incubated at 37°C in 5% CO2 at 100% humidity and inspected for 48 hours. Each experiment was repeated at least thrice with identical results.

Tumorigenicity assay. Female 5- to 7-week-old severe combined immunodeficiency (SCID) mice (Charles River Italia, Como, Italy) were cared for in accordance with standards of the Italian law. Tumorigenicity was determined after s.c. injection in the ventral flank. Tumor growth was assessed twice weekly and mice were sacrificed 2 to 6 months after injection. Each tumor was divided into two: one-half was frozen in the presence of a cryopreservant or formalin-fixed and paraffin-embedded for routine histopathologic examination and immunohistochemistry. The other half was snap-frozen in liquid nitrogen and stored at −80°C until the time of protein, RNA, and DNA isolation. Western blot analysis, reverse transcription-PCR, and Southern blot analysis were carried out as above. Routine histopathology and immunohistochemistry for the detection of human MET and vimentin proteins were done according to standard protocols. A mouse anti-human vimentin monoclonal antibody that does not cross-react with mouse protein was used (V9, 1:100; Santa Cruz Biotechnology). A biotinylated secondary monoclonal antibody and Dab+ substrate chromogen system were used (Dako, Glostrup, Denmark). Sections were counterstained with Mayer's hematoxylin. Specimens in which the incubation with the primary antibody was omitted were used as negative controls.

MET oncogene overexpression transforms primary cultured human osteoblasts. HOB preparations were propagated from 15 donors as monolayer cultures, and then characterized. HOB preparations were used in experiments when >80% of cells displayed the osteoblastic phenotype (Supplementary Fig. S1). As expected, in these cultures, the MET receptor was undetectable, as its expression is likely confined to more immature precursors (28) or is induced by hormones (29).

To obtain stable and sustained overexpression of the MET oncogene, multiple MET transgene copies were integrated into the HOB genome by means of a bicistronic LV. In this vector, the cytomegalovirus promoter powered the concurrent expression of MET and of a reporter gene (GFP). HOB cultures were transduced with either wild-type MET (METwt), or a mutated MET (METY1253D), carrying the activating mutation Y1253D (7). After 40 to 60 days, stacks of brightly fluorescent cells had overgrown the dark monolayer of flat, large, and slowly growing cells (Fig. 1A). These stacks corresponded to the multilayered mounds of rapidly dividing cells that are usually described as “foci of transformation.”

Figure 1.

MET transgene expression and activation in HOB clones. HOB preparations were transduced with bicistronic LVs carrying the MET-internal ribosome entry site-GFP cassette with either the wild-type (METwt) or the mutated (METY1253D) MET cDNA. Results of the transduction of the HOB preparation no. 1704. A, after 40 to 60 days, a similar number of foci (35-45 each/10-cm diameter culture dish) developed in METwt (left) and in METY1253D (middle and right) HOB cultures. Foci of birefringent (top) and strongly fluorescent (bottom) cells overgrew a monolayer of flat and nonfluorescent cells, which were visible at higher magnification (arrows). B, Southern blot analysis of DNA from cells propagated from individual foci showed stable integration of the transgenes. Numbers on top indicate different clones. The blot was labeled with a probe specific for the GFP gene. DNA was digested with SpeI. As the SpeI site is unique in the vector, the labeling of discrete different bands in different foci showed the random integration of the transgene in HOB cells and the clonality of foci. Activation (C) and expression (D) of the MET receptor in clones. Proteins immunoprecipitated by anti-MET monoclonal antibodies were labeled in Western blot analysis with a monoclonal antibody against phosphotyrosine (C) and subsequently with MET monoclonal antibody against the human COOH terminal peptide of the receptor (D). Proteins were separated in the presence of a reducing agent, which takes apart the α- and β-chains of the MET receptor. Antibodies labeled the MET β-chain (p145) and the MET precursor (pr170). As positive controls proteins from 293-T cells, which express the MET receptor (see also Table 1), were loaded. The receptor was not found in parental HOB cultures.

Figure 1.

MET transgene expression and activation in HOB clones. HOB preparations were transduced with bicistronic LVs carrying the MET-internal ribosome entry site-GFP cassette with either the wild-type (METwt) or the mutated (METY1253D) MET cDNA. Results of the transduction of the HOB preparation no. 1704. A, after 40 to 60 days, a similar number of foci (35-45 each/10-cm diameter culture dish) developed in METwt (left) and in METY1253D (middle and right) HOB cultures. Foci of birefringent (top) and strongly fluorescent (bottom) cells overgrew a monolayer of flat and nonfluorescent cells, which were visible at higher magnification (arrows). B, Southern blot analysis of DNA from cells propagated from individual foci showed stable integration of the transgenes. Numbers on top indicate different clones. The blot was labeled with a probe specific for the GFP gene. DNA was digested with SpeI. As the SpeI site is unique in the vector, the labeling of discrete different bands in different foci showed the random integration of the transgene in HOB cells and the clonality of foci. Activation (C) and expression (D) of the MET receptor in clones. Proteins immunoprecipitated by anti-MET monoclonal antibodies were labeled in Western blot analysis with a monoclonal antibody against phosphotyrosine (C) and subsequently with MET monoclonal antibody against the human COOH terminal peptide of the receptor (D). Proteins were separated in the presence of a reducing agent, which takes apart the α- and β-chains of the MET receptor. Antibodies labeled the MET β-chain (p145) and the MET precursor (pr170). As positive controls proteins from 293-T cells, which express the MET receptor (see also Table 1), were loaded. The receptor was not found in parental HOB cultures.

Close modal

Individual foci were picked up and propagated as monolayer cultures, which showed unlimited growth in culture (see below). Cells of each focus-derived clone had an average of five copies of the bicistronic transgene integrated at distinct sites (Fig. 1B). Cells harvested from monolayers outside the foci of transformation were found to contain the MET transgene, randomly integrated (mean, 0.8 copies/cell; data not shown). Nonfluorescent foci were never found. Foci did not develop in control HOB cultures or in cultures transduced with control LVs (carrying either no transgene, GFP transgene alone, or bicistronic luciferase-GFP cassette; data not shown). MET receptors were only detected in foci-derived clones, as shown by Western blot analysis (Fig. 1D) or immunohistochemistry (Supplementary Fig. S2A). In individual clones, levels of MET protein overexpression paralleled the levels of MET mRNA (Table 1). The overexpressed MET receptors were in a constitutively activated state, as shown by autophosphorylation at tyrosine (Fig. 1C). The mutant METY1253D receptors showed a slightly lower level of phosphorylation than the wild-type ones, as expected, due to the substitution of one of the major MET phosphorylation sites (30) with an aspartic acid. This did not impair its kinase activity that was even increased by the negative charge mimicking the phosphorylated tyrosine; these data are not shown because they were reported elsewhere (7). We tested whether any of the MET-overexpressing HOB clones produce the MET ligand HGF, using both reverse transcription-PCR with specific primers and Western blot analysis (data not shown). Although in these cells, the MET receptor was in a constitutively activated state, in no instance was HGF detectable either as a protein or mRNA. This ruled out the idea that MET receptor activation was due to an autocrine loop.

Table 1.

MET expression level in HOB clones and human tumors

CellsTissueTransgene*METwt expression (CT)METY1253D expression (CT)
HOBs     
    Parental  None >40 >40 
    Clone 15 Osteoblasts METwt 20.1 >40 
    Clone 26 Osteoblasts METwt 22.0 >40 
    Clone 35 Osteoblasts METwt 22.4 >40 
    Clone 12 Osteoblasts METY1253D Not tested 21.0 
    Clone 18 Osteoblasts METY1253D Not tested 20.2 
    Clone 42 Osteoblasts METY1253D Not tested 19.0 
Cell lines     
    293-T Embryonal kidney None 28.2 >40 
    GTL-16 Gastric carcinoma None 19.0 >40 
    SK-OV-3 Ovarian carcinoma None 22.3 >40 
    NIHOVCAR3 Ovarian carcinoma None 21.0 >40 
    U2-OS Osteosarcoma None 25.2 >40 
    SAOS-2 Osteosarcoma None 23.9 >40 
    MG-63 Osteosarcoma None 24.7 >40 
Samples     
    OS 1 Osteosarcoma None 21.7 >40 
    OS 2 Osteosarcoma None 22.3 >40 
    OS 3 Osteosarcoma None 24.4 >40 
    NL 95 Normal lymph node None >40 >40 
    NL 106 Normal lymph node None >40 >40 
    HN mucosa 95 Squamous epithelium None 29.7 >40 
    HN mucosa 106 Squamous epithelium None 31.5 >40 
    HNSCC-T65 Squamous cell carcinoma None 27.0 33.5§ 
    HNSCC-M65 Squamous cell carcinoma metastasis None 22.5 24.4§ 
    HNSCC-T66 Squamous cell carcinoma None 27.7 31.6§ 
    HNSCC-M65 Squamous cell carcinoma metastasis None 23.7 23.0§ 
    OVCAR H01 Ovarian carcinoma None 21.0 >40 
    OVCAR E12 Ovarian carcinoma None 22.3 >40 
    OVCAR A8 Ovarian carcinoma None 21.9 >40 
CellsTissueTransgene*METwt expression (CT)METY1253D expression (CT)
HOBs     
    Parental  None >40 >40 
    Clone 15 Osteoblasts METwt 20.1 >40 
    Clone 26 Osteoblasts METwt 22.0 >40 
    Clone 35 Osteoblasts METwt 22.4 >40 
    Clone 12 Osteoblasts METY1253D Not tested 21.0 
    Clone 18 Osteoblasts METY1253D Not tested 20.2 
    Clone 42 Osteoblasts METY1253D Not tested 19.0 
Cell lines     
    293-T Embryonal kidney None 28.2 >40 
    GTL-16 Gastric carcinoma None 19.0 >40 
    SK-OV-3 Ovarian carcinoma None 22.3 >40 
    NIHOVCAR3 Ovarian carcinoma None 21.0 >40 
    U2-OS Osteosarcoma None 25.2 >40 
    SAOS-2 Osteosarcoma None 23.9 >40 
    MG-63 Osteosarcoma None 24.7 >40 
Samples     
    OS 1 Osteosarcoma None 21.7 >40 
    OS 2 Osteosarcoma None 22.3 >40 
    OS 3 Osteosarcoma None 24.4 >40 
    NL 95 Normal lymph node None >40 >40 
    NL 106 Normal lymph node None >40 >40 
    HN mucosa 95 Squamous epithelium None 29.7 >40 
    HN mucosa 106 Squamous epithelium None 31.5 >40 
    HNSCC-T65 Squamous cell carcinoma None 27.0 33.5§ 
    HNSCC-M65 Squamous cell carcinoma metastasis None 22.5 24.4§ 
    HNSCC-T66 Squamous cell carcinoma None 27.7 31.6§ 
    HNSCC-M65 Squamous cell carcinoma metastasis None 23.7 23.0§ 
    OVCAR H01 Ovarian carcinoma None 21.0 >40 
    OVCAR E12 Ovarian carcinoma None 22.3 >40 
    OVCAR A8 Ovarian carcinoma None 21.9 >40 
*

Genes transduced by LVs.

MET transgene expression was measured by quantitative real-time RT-PCR with TaqMan assay. CT (threshold cycle) indicates the number of cycles needed to reach a threshold amount of PCR products and depends directly on the initial concentration of target nucleic acid; >40, not detectable.

HNSCC, head and neck squamous cell carcinomas.

§

See ref. 7.

Transformed human osteoblasts express MET levels comparable to that of human osteosarcomas. Using quantitative reverse transcription-PCR, we compared the level of MET expression with that of human cancer cell lines and bioptic samples (Table 1). In HOB clones, the level of MET expression was in the range measured in gastric and ovarian carcinoma and osteosarcoma cell lines, already classified as MET-overexpressing cell lines. Similar amounts of MET-specific mRNA were found in spontaneously occurring osteosarcomas and other human tumors.

Although the level of MET expression attained in HOBs was comparable to that of human osteosarcomas, we hypothesize that only HOBs expressing MET over a critical threshold acquired the transformed and immortalized phenotype. To verify this hypothesis, we expressed the MET receptor in HOBs at levels comparable to those observed in physiologic conditions in human tissues (13, 31), using a tet-dependent self-regulating LV (25). In this case, MET-transduced HOB cultures were indistinguishable from those of parental HOBs. Foci did not develop in these cultures which, however, showed the ability to form colonies in semisolid medium at high efficiency, as long as they expressed the MET receptor (Fig. 2). These cultures underwent growth arrest like the parental HOBs after 40 to 70 days, suggesting that the neoexpression of the MET receptor at physiologic levels is able to confer clonogenic ability to human primary osteoblasts, but not immortalization.

Figure 2.

MET transgene regulated expression and biological outcome in HOB cultures. A, HOB cultures were transduced with either the wild-type (METwt) or mutated (METY1253D) MET transgenes using tet-self-regulated LVs. The bulk of the unselected populations of transduced HOBs were studied. When grown in the absence of doxycyclin (D−), HOBs expressed detectable MET receptors, which decreased on the addition of doxycyclin (D+). In parental HOB cultures (no. 2036 and no. 2028), the MET receptor was never found. Total proteins were extracted from the bulk unselected cell populations and analyzed in Western blot using MET monoclonal antibodies against the human COOH terminal peptide. Proteins were extracted to allow the separation of the α- and β-chains of the MET receptor. Antibodies labeled the MET β-chain (p145) and the MET precursor (pr170). B, HOB cultures expressing either METwt or METY1253D transgene formed visible colonies in semisolid agar medium in the absence (D−) but not in the presence (D+) of doxycyclin. C, colonies formed by different HOB cultures transduced with either the METwt or the METY1253D transgene were stained and counted under the microscope: #, different HOB preparations from different donors; &, colonies were counted 21 days after plating 1 × 103 and 4 × 103 cells in triplicate; the percentage of cells which formed colonies out of those plated are the mean of three plates in a representative experiment; nt, not tested.

Figure 2.

MET transgene regulated expression and biological outcome in HOB cultures. A, HOB cultures were transduced with either the wild-type (METwt) or mutated (METY1253D) MET transgenes using tet-self-regulated LVs. The bulk of the unselected populations of transduced HOBs were studied. When grown in the absence of doxycyclin (D−), HOBs expressed detectable MET receptors, which decreased on the addition of doxycyclin (D+). In parental HOB cultures (no. 2036 and no. 2028), the MET receptor was never found. Total proteins were extracted from the bulk unselected cell populations and analyzed in Western blot using MET monoclonal antibodies against the human COOH terminal peptide. Proteins were extracted to allow the separation of the α- and β-chains of the MET receptor. Antibodies labeled the MET β-chain (p145) and the MET precursor (pr170). B, HOB cultures expressing either METwt or METY1253D transgene formed visible colonies in semisolid agar medium in the absence (D−) but not in the presence (D+) of doxycyclin. C, colonies formed by different HOB cultures transduced with either the METwt or the METY1253D transgene were stained and counted under the microscope: #, different HOB preparations from different donors; &, colonies were counted 21 days after plating 1 × 103 and 4 × 103 cells in triplicate; the percentage of cells which formed colonies out of those plated are the mean of three plates in a representative experiment; nt, not tested.

Close modal

MET-overexpressing human osteoblasts show a fully transformed phenotype in vitro and tumorigenicity in vivo. The MET-overexpressing HOBs propagated from foci-derived clones showed a fully transformed phenotype in vitro and were tumorigenic in vivo, showing an osteosarcoma-like phenotype.

The MET-overexpressing HOBs showed anchorage-independent growth, whereas the parental HOB did not (Table 2). Nevertheless, clones maintained the osteogenic phenotype, e.g., production of alkaline phosphatase and mineralized matrix (Supplementary Fig. S2B and C; data not shown). Supplementary Fig. S2 and Table 2 show that cells propagated from foci of transformation, when grown as monolayer cultures, were spindle-shaped and birefringent, whereas the parental HOBs were cuboidal and flat, did not show contact inhibition, showed anchorage-independent growth, displayed high mitotic activity, and exhibited unlimited growth.

Table 2.

Proliferation of MET-overexpressing HOB clones

CellsTransgene*Cloning efficiency in soft agar (%)PD after 180 days
Parental HOBs None Senescent after 40-70 days (15-25 PD) 
Clone 15 METwt 18.0 140 
Clone 15 DN-MET METwt, DN-MET 3.6 120 
Clone 15 shRNA METwt, shRNA 3.3 120 
Clone 26 METwt 17.5 140 
Clone 35 METwt 9.5 Not tested 
Clone 12 METY1253D 16.5 150 
Clone 12 DN METY1253D, DN-MET 7.0 130 
Clone 12 shRNA METY1253D, shRNA 5.0 130 
Clone 18 METY1253D 115 
Clone 42 METY1253D 28.0 125 
Clone 42 DN METY1253D, DN-MET 7.2 120 
Clone 42 shRNA METY1253D, shRNA 3.3 120 
CellsTransgene*Cloning efficiency in soft agar (%)PD after 180 days
Parental HOBs None Senescent after 40-70 days (15-25 PD) 
Clone 15 METwt 18.0 140 
Clone 15 DN-MET METwt, DN-MET 3.6 120 
Clone 15 shRNA METwt, shRNA 3.3 120 
Clone 26 METwt 17.5 140 
Clone 35 METwt 9.5 Not tested 
Clone 12 METY1253D 16.5 150 
Clone 12 DN METY1253D, DN-MET 7.0 130 
Clone 12 shRNA METY1253D, shRNA 5.0 130 
Clone 18 METY1253D 115 
Clone 42 METY1253D 28.0 125 
Clone 42 DN METY1253D, DN-MET 7.2 120 
Clone 42 shRNA METY1253D, shRNA 3.3 120 
*

Genes transduced by LVs.

Cell colonies were counted 21 days after plating 1 × 103 and 4 × 103 cells in triplicate.

PD, population doublings: in clones, the stability of transgene integration and expression was monitored every 20 PD.

In vitro and in vivo, the MET receptor activated by its ligand HGF triggers a unique biological program leading to the so-called “invasive growth” (1). Deregulated activation of the MET-mediated “invasive-growth” program has long been thought to contribute to cell transformation and tumor progression (2). We found that the MET-overexpressing HOBs displayed biological properties that recapitulate the activities elicited by MET signaling in vitro. They acquired the ability to invade an artificial basement membrane made of collagen, laminin, and glycosaminoglycans (Matrigel), whereas parental osteoblasts did not (Supplementary Fig. S3A). This chemoinvasion assay is commonly used to evaluate cancer cell invasiveness. These cells also showed the ability to invade a three-dimensional collagen gel (Supplementary Fig. S3B). This assay, also known as the “branching morphogenesis” assay, highlights the potential of cells to invade a surrogate extracellular matrix, forming typical branched structures. This invasion process represents the summa of the invasive growth phenotype and results from the fine integration of all the pleiotropic effects induced by HGF, including cell proliferation, motility, differentiation, and survival (16).

We found that the MET-overexpressing HOB clones had complex cytogenetic and chromosomal aberrations. For example, in a wild-type MET-overexpressing clone, we found 84% of hyper-diploid cells (mean chromosome number, 52 ± 5) and 6% of hyper-tetraploid cells. In a METY1253D-overexpressing clone, we found 83% of hyper-diploid cells (mean chromosome number, 54 ± 1) and 7% of hyper-tetraploid cells. In addition, only a few chromosomes were recognizable, as many were rearranged. However, common aberrations were not detected (data not shown). It is noteworthy that both the parental osteoblasts and osteoblasts transduced with control vectors were diploid and devoid of aberrations as long as they were propagated in culture. The genetic alterations of MET-overexpressing HOB clones are reminiscent of the karyotypes of human osteosarcoma (32), which is characterized by complex chromosomal abnormalities, often with pronounced cell-to-cell variation or heterogeneity, and, in contrast with other human sarcomas, is not associated with any specifically recurrent translocation or any other specific chromosomal rearrangement (33).

The MET-overexpressing HOB clones were tumorigenic in vivo (Fig. 3). As expected, due to the known low growth efficiency of human osteosarcoma cells in SCID mice (34), a fairly high number of transplanted cells was necessary to obtain xenografts, and a low incidence was observed (Table 3). All tumors showed the distinguishing features of highly aggressive human osteosarcomas (35), including atypical nuclei and aberrant mitoses (Fig. 3), the production of some osteoid extracellular matrix and alkaline phosphatase, the expression of human vimentin, and striking neovascularization (data not shown). Figure 3 also shows that all tumors expressed the MET transgene mRNA and the human MET receptors detectable by immunohistochemistry.

Figure 3.

Tumorigenic properties of HOB clones overexpressing the METwt or the METY1253D receptor. A, tumorigenic ability was measured in immunocompromised SCID mice. B-F, phenotypic characterization of tumors grown after s.c. injection. Tumors had the histology of a highly aggressive osteosarcoma, with several atypical nuclei and aberrant mitoses (B). Although mostly undifferentiated, tumors showed areas of osteoid production (C) and alkaline phosphatase expression (D). Tumors carried integrated copies of the transgene (data not shown); all expressed the relevant transgene. E, expression of the transgene in the propagated clones no. 12 and 15 (top) and in the corresponding xenograft (bottom); +, positive (GTL-16 cells) controls; −, negative (HOB parental cultures) controls. F, immunohistochemical detection of the MET receptors in a tumor with antibody against the human COOH terminal peptide of the receptor.

Figure 3.

Tumorigenic properties of HOB clones overexpressing the METwt or the METY1253D receptor. A, tumorigenic ability was measured in immunocompromised SCID mice. B-F, phenotypic characterization of tumors grown after s.c. injection. Tumors had the histology of a highly aggressive osteosarcoma, with several atypical nuclei and aberrant mitoses (B). Although mostly undifferentiated, tumors showed areas of osteoid production (C) and alkaline phosphatase expression (D). Tumors carried integrated copies of the transgene (data not shown); all expressed the relevant transgene. E, expression of the transgene in the propagated clones no. 12 and 15 (top) and in the corresponding xenograft (bottom); +, positive (GTL-16 cells) controls; −, negative (HOB parental cultures) controls. F, immunohistochemical detection of the MET receptors in a tumor with antibody against the human COOH terminal peptide of the receptor.

Close modal
Table 3.

Tumorigenic properties of MET-overexpressing HOB clones

CellsTransgene*No. of mice with tumor / injected miceLatency (days)
Parental HOBs None 0 of 8 — 
Clone 15 METwt 3 of 8 100-135 (mean, 110) 
Clone 12 METY1253D 9 of 12 40-120 (mean, 55) 
Clone 12 DN METY1253D, DN-MET 0 of 4 — 
Clone 42 METY1253D 6 of 8 50-70 (mean, 60) 
Clone 42 DN METY1253D, DN-MET 0 of 4 — 
U2-OS None 4 of 4 50-55 (mean 53) 
U2-OS DN DN-MET 0 of 4 — 
IOR-OS9 None 4 of 4 37-70 (mean, 65) 
IOR-OS9 DN DN-MET 0 of 4 — 
CellsTransgene*No. of mice with tumor / injected miceLatency (days)
Parental HOBs None 0 of 8 — 
Clone 15 METwt 3 of 8 100-135 (mean, 110) 
Clone 12 METY1253D 9 of 12 40-120 (mean, 55) 
Clone 12 DN METY1253D, DN-MET 0 of 4 — 
Clone 42 METY1253D 6 of 8 50-70 (mean, 60) 
Clone 42 DN METY1253D, DN-MET 0 of 4 — 
U2-OS None 4 of 4 50-55 (mean 53) 
U2-OS DN DN-MET 0 of 4 — 
IOR-OS9 None 4 of 4 37-70 (mean, 65) 
IOR-OS9 DN DN-MET 0 of 4 — 
*

Genes transduced by LV.

Mice s.c. injected with 30 × 106 cells; clones overexpressing the METY1253D transgene were more tumorigenic than the METwt expressing clones. The mean percentage (±2 SE) of mice with tumors out of mice injected with either METwt or the METY1253D-expressing cells were 16.7 ± 12.5% and 75 ± 16.7%, respectively (t = 4.08; P < 0.05).

U2-OS and IOR-OS9 human osteosarcoma cell lines, as well as the METY1253D HOB clones, were no longer able to form tumors in nude mice when transduced with the DN-MET transgene.

As mentioned in the first paragraph, we also obtained HOBs overexpressing a MET oncogene carrying the Y1253D activating mutation (7). As shown in Supplementary Figs. S1 and S2, HOB clones overexpressing the mutated METY1253D were phenotypically indistinguishable from HOBs overexpressing the wild-type MET, the notable exception being the lower in vivo tumorigenicity (Table 3).

MET “addiction” of human osteoblasts transformed by MET. We next evaluated if the transformed and tumorigenic phenotype displayed by MET-overexpressing HOB clones relied on sustained MET receptor activation. Activation was almost abolished (Supplementary Fig. S4) after either decreasing MET receptors by stable expression of MET-shRNA or by blocking MET receptor dimerization by stable expression of a DN-MET. The latter lacks the intracellular domain and thus forms inactive dimers with the full size receptor (26). In both cases, the transformed phenotype, measured as the ability to undergo anchorage-independent growth, was consistently reduced (Table 2), and the invasion assay in vitro scored negative (Supplementary Fig. S3B). In vivo, the stable expression of the DN-MET receptor abolished the tumorigenic properties of MET-overexpressing clones (Table 3). The DN-MET was also expressed in established human osteosarcoma cell lines, where we previously showed MET overexpression and constitutive activation [ref. (20); Supplementary Fig. S5]. The DN-MET fully abolished MET constitutive activation (Supplementary Fig. S5) and, intriguingly, also impaired their growth in vivo as xenografts (Table 3).

In this work, we show that the overexpression of the MET oncogene initiates the transformation of human primary osteoblasts and sustains tumor progression towards the osteosarcoma.

We obtained overexpression by transferring the wild-type MET transgene by means of LVs, which ensures the random integration of the MET transgene in the majority of cultured cells, as LVs are also able to transduce nondividing cells. The presence in each MET-overexpressing clone of the transgene integrated at distinct sites showed that individual foci originated from random and different integration events, and thus makes the contribution to transformation of activation of proto-oncogenes or inactivation of tumor suppressor genes as an unlikely consequence of vector insertions. The former was per se improbable, as we used LV with a self-inactivating LTR (36). Random integration also allowed us to rule out the role of integration sites on expression variation. Therefore, transgene expression only depends on its intrinsic regulation, which relies on the internal promoter. In addition, a particular advantage of transgene expression operated by LV gene transfer is a constant linear relationship between the number of integrated copies and amount of mRNA (36). We also show that foci of transformation did not develop in control osteoblast cultures or in cultures transduced with control LVs. These data strengthen the established notion that spontaneous transformation of in vitro cultured human cells is an infrequent event (37), and thus, rule out the unlikely possibility that we selected transformed cells arising independently from MET expression.

In vitro, the HOB clones overexpressing the wild-type MET display the same properties of those transformed by the MET oncogene activated by a known mutation. This was not surprising, as we found that in vitro, the overexpressed wild-type receptors are constitutively active, even in the absence of the ligand HGF, likely due to self-oligomerization (38). In vivo, HOB clones overexpressing the mutated MET receptors were more tumorigenic. It is conceivable that the threshold to activate the in vivo growth might be higher and more easily overcome in the presence of some cross-reactive mouse HGF present in circulating blood. As it has been shown that some MET mutations sensitize the MET receptor to very low HGF concentrations (39), we can speculate that the METY1253D receptor made osteoblasts more sensitive to mouse HGF.

In the transformed osteoblast clones, the level of MET expression was in the range measured in known overexpressing cell lines (24, 27, 40) and in human cancers (7, 31, 41), including osteosarcomas (20). Thus, the high level of expression, obtained by mimicking MET gene amplification using LV-mediated gene transfer, did not exceed that observed in spontaneously occurring osteosarcomas and other human tumors. We can conclude that osteoblast transformation and tumorigenesis are not the result of massive oncogene overexpression.

Human osteoblast transformation in vitro is a rare event. A likely hypothesis that explains the rarity is the number of susceptible target cells being very small. A candidate for this minor population could be the still elusive osteoprogenitor cell that maintains the potential to proliferate and to differentiate (42). Progenitors are indubitably present in primary osteoblast cultures, which undergo several population doublings before senescence (43), and previous work has shown that LVs transfer genes in bona fide stem cells (44), including bone precursors (45). Although unidentified, the osteoprogenitor targeted by MET likely belong to the family of mesenchymal stem cells, which are also the precursors of rhabdomyosarcomas. Interestingly, we (46) and others (47) found MET overexpression in a high percentage of human rhabdomyosarcoma; furthermore, in a transgenic mouse model, deregulated MET signaling contributes to rhabdomyosarcomagenesis (48). As we show here, MET misexpression alone does not transform, but it is able to confer clonogenic capability to osteoblasts. Altogether, these data suggest that the expression of MET oncogene creates an expanded premalignant osteoblast population; in this population, overexpression of the oncogene might either lead or simply allow the occurrence of other transforming events in mesenchyme-derived osteoblast progenitors, e.g., cancer stem cells, thus uniquely contributing to osteosarcomagenesis. In both instances, the transformed and tumorigenic phenotype relies on the continuous presence of an activated MET receptor.

This article provides conclusive experimental evidence supporting the long-sought link between MET overexpression, found in a large repertoire of human cancers, and progression toward malignancy (2). Moreover, these findings are consistent with the notion that increased copy number of the MET gene and its expression above a threshold are also critical requirements for the development of tumors harboring germ line MET mutations (10, 49). Interestingly, MET overexpression is required to sustain the fully blown osteosarcoma phenotype; both transformed osteoblasts and MET-overexpressing osteosarcoma cell lines showed MET “addiction.” This information highlights MET as a therapeutic target in human cancers accumulating genetic alterations.

Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/).

Grant support: Italian Ministry of Research and Education, MIUR Cofin projects (M.F. Di Renzo, P. Comoglio, L. Naldini) and the Italian Association for Cancer Research (AIRC) funding to M.F. Di Renzo, P. Comoglio, and N. Baldini. N. Coltella was a FIRC fellowship recipient.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank C. Roca for giving advise and help in the generation of cell spheroids; C. Bardella and L. Ferrantino for help in performing in vivo experiments; and G. Ponzio and L. Casorzo for help and suggestions for chromosome analysis.

1
Trusolino L, Comoglio PM. Scatter-factor and semaphorin receptors: cell signalling for invasive growth.
Nat Rev Cancer
2002
;
2
:
289
–300.
2
Birchmeier C, Birchmeier W, Gherardi E, Vande Woude GF. Met, metastasis, motility and more.
Nat Rev Mol Cell Biol
2003
;
4
:
915
–25.
3
Schmidt L, Duh FM, Chen F, et al. Germline and somatic mutations in the tyrosine kinase domain of the MET proto-oncogene in papillary renal carcinomas.
Nat Genet
1997
;
16
:
68
–73.
4
Schmidt L, Junker K, Nakaigawa N, et al. Novel mutations of the MET proto-oncogene in papillary renal carcinomas.
Oncogene
1999
;
18
:
2343
–50.
5
Lee JH, Han SU, Cho H, et al. A novel germ line juxtamembrane met mutation in human gastric cancer.
Oncogene
2000
;
19
:
4947
–53.
6
Park WS, Dong SM, Kim SY, et al. Somatic mutations in the kinase domain of the Met/hepatocyte growth factor receptor gene in childhood hepatocellular carcinomas.
Cancer Res
1999
;
59
:
307
–10.
7
Di Renzo MF, Olivero M, Martone T, et al. Somatic mutations of the MET oncogene are selected during metastatic spread of human HNSC carcinomas.
Oncogene
2000
;
19
:
1547
–55.
8
Ma PC, Kijima T, Maulik G, et al. c-Met mutational analysis in small cell lung cancer: novel juxtamembrane domain mutations regulating cytoskeletal functions.
Cancer Res
2003
;
63
:
6272
–81.
9
Aebersold DM, Landt O, Berthou S, et al. Prevalence and clinical impact of Met Y1253D-activating point mutation in radiotherapy-treated squamous cell cancer of the oropharynx.
Oncogene
2003
;
22
:
8519
–23.
10
Zhuang Z, Park WS, Pack S, et al. Trisomy 7-harbouring non-random duplication of the mutant MET allele in hereditary papillary renal carcinomas.
Nat Genet
1998
;
20
:
66
–9.
11
Fischer J, Palmedo G, von Knobloch R, et al. Duplication and overexpression of the mutant allele of the MET proto-oncogene in multiple hereditary papillary renal cell tumours.
Oncogene
1998
;
17
:
733
–9.
12
Kuniyasu H, Yasui W, Kitadai Y, Yokozaki H, Ito H, Tahara E. Frequent amplification of the c-met gene in scirrhous type stomach cancer.
Biochem Biophys Res Commun
1992
;
189
:
227
–32.
13
Di Renzo MF, Olivero M, Giacomini A, et al. Overexpression and amplification of the Met/HGF receptor gene during the progression of colorectal cancer.
Clin Cancer Res
1995
;
1
:
147
–54.
14
Reis RM, Konu-Lebleblicioglu D, Lopes JM, Kleihues P, Ohgaki H. Genetic profile of gliosarcomas.
Am J Pathol
2000
;
156
:
425
–32.
15
Ivan M, Bond JA, Prat M, Comoglio PM, Wynford-Thomas D. Activated ras and ret oncogenes induce over-expression of c-met (hepatocyte growth factor receptor) in human thyroid epithelial cells.
Oncogene
1997
;
14
:
2417
–23.
16
Pennacchietti S, Michieli P, Galluzzo M, Mazzone M, Giordano S, Comoglio PM. Hypoxia promotes invasive growth by transcriptional activation of the met protooncogene.
Cancer Cell
2003
;
3
:
347
–61.
17
Maina F, Casagranda F, Audero E, et al. Uncoupling of Grb2 from the Met receptor in vivo reveals complex roles in muscle development.
Cell
1996
;
87
:
531
–42.
18
Fehlner-Gardiner CC, Cao H, Jackson-Boeters L, et al. Characterization of a functional relationship between hepatocyte growth factor and mouse bone marrow-derived mast cells.
Differentiation
1999
;
65
:
27
–42.
19
Cooper CS, Park M, Blair DG, et al. Molecular cloning of a new transforming gene from a chemically transformed human cell line.
Nature
1984
;
311
:
29
–33.
20
Ferracini R, Di Renzo MF, Scotlandi K, et al. The Met/HGF receptor is over-expressed in human osteosarcomas and is activated by either a paracrine or an autocrine circuit.
Oncogene
1995
;
10
:
739
–49.
21
Wallenius V, Hisaoka M, Helou K, et al. Overexpression of the hepatocyte growth factor (HGF) receptor (Met) and presence of a truncated and activated intracellular HGF receptor fragment in locally aggressive/malignant human musculoskeletal tumors.
Am J Pathol
2000
;
156
:
821
–9.
22
MacEwen EG, Kutzke J, Carew J, et al. c-Met tyrosine kinase receptor expression and function in human and canine osteosarcoma cells.
Clin Exp Metastasis
2003
;
20
:
421
–30.
23
Scotlandi K, Baldini N, Oliviero M, et al. Expression of Met/hepatocyte growth factor receptor gene and malignant behavior of musculoskeletal tumors.
Am J Pathol
1996
;
149
:
1209
–19.
24
Giordano S, Ponzetto C, Di Renzo MF, Cooper CS, Comoglio PM. Tyrosine kinase receptor indistinguishable from the c-met protein.
Nature
1989
;
339
:
155
–6.
25
Vigna E, Amendola M, Benedicenti F, Simmons AD, Follenzi A, Naldini L. Efficient Tet-dependent expression of human factor IX in vivo by a new self-regulating lentiviral vector.
Mol Ther
2005
;
11
:
763
–75.
26
Giordano S, Corso S, Conrotto P, et al. The semaphorin 4D receptor controls invasive growth by coupling with Met.
Nat Cell Biol
2002
;
4
:
720
–4.
27
Coltella N, Manara MC, Cerisano V, et al. Role of the MET/HGF receptor in proliferation and invasive behavior of osteosarcoma.
FASEB J
2003
;
17
:
1162
–4.
28
Grano M, Galimi F, Zambonin G, et al. Hepatocyte growth factor is a coupling factor for osteoclasts and osteoblasts in vitro.
Proc Natl Acad Sci U S A
1996
;
93
:
7644
–8.
29
Blanquaert F, Pereira RC, Canalis E. Cortisol inhibits hepatocyte growth factor/scatter factor expression and induces c-met transcripts in osteoblasts.
Am J Physiol Endocrinol Metab
2000
;
278
:
E509
–15.
30
Longati P, Bardelli A, Ponzetto C, Naldini L, Comoglio PM. Tyrosines1234–1235 are critical for activation of the tyrosine kinase encoded by the MET proto-oncogene (HGF receptor).
Oncogene
1994
;
9
:
49
–57.
31
Cortesina G, Martone T, Galeazzi E, et al. Staging of head and neck squamous cell carcinoma using the MET oncogene product as marker of tumor cells in lymph node metastases.
Int J Cancer
2000
;
89
:
286
–92.
32
Helman LJ, Meltzer P. Mechanisms of sarcoma development.
Nat Rev Cancer
2003
;
3
:
685
–94.
33
Sandberg AA, Bridge JA. Updates on the cytogenetics and molecular genetics of bone and soft tissue tumors: osteosarcoma and related tumors.
Cancer Genet Cytogenet
2003
;
145
:
1
–30.
34
Scotlandi K, Serra M, Nicoletti G, et al. Multidrug resistance and malignancy in human osteosarcoma.
Cancer Res
1996
;
56
:
2434
–9.
35
Dahlin DC, Unni KK. Osteosarcoma of bone and its important recognizable varieties.
Am J Surg Pathol
1977
;
1
:
61
–72.
36
De Palma M, Naldini L. Transduction of a gene expression cassette using advanced generation lentiviral vectors.
Methods Enzymol
2002
;
346
:
514
–29.
37
Rubio D, Garcia-Castro J, Martin MC, et al. Spontaneous human adult stem cell transformation.
Cancer Res
2005
;
65
:
3035
–9.
38
Schlessinger J. Cell signaling by receptor tyrosine kinases.
Cell
2000
;
103
:
211
–25.
39
Maritano D, Accornero P, Bonifaci N, Ponzetto C. Two mutations affecting conserved residues in the Met receptor operate via different mechanisms.
Oncogene
2000
;
19
:
1354
–61.
40
Di Renzo MF, Narsimhan RP, Olivero M, et al. Expression of the Met/HGF receptor in normal and neoplastic human tissues.
Oncogene
1991
;
6
:
1997
–2003.
41
Takeuchi H, Bilchik A, Saha S, et al. c-MET expression level in primary colon cancer: a predictor of tumor invasion and lymph node metastases.
Clin Cancer Res
2003
;
9
:
1480
–8.
42
Calvi LM, Sims NA, Hunzelman JL, et al. Activated parathyroid hormone/parathyroid hormone-related protein receptor in osteoblastic cells differentially affects cortical and trabecular bone.
J Clin Invest
2001
;
107
:
277
–86.
43
Stenderup K, Justesen J, Clausen C, Kassem M. Aging is associated with decreased maximal life span and accelerated senescence of bone marrow stromal cells.
Bone
2003
;
33
:
919
–26.
44
Wiznerowicz M, Trono D. Harnessing HIV for therapy, basic research and biotechnology.
Trends Biotechnol
2005
;
23
:
42
–7.
45
Lee CI, Kohn DB, Ekert JE, Tarantal AF. Morphological analysis and lentiviral transduction of fetal monkey bone marrow-derived mesenchymal stem cells.
Mol Ther
2004
;
9
:
112
–23.
46
Ferracini R, Olivero M, Di Renzo MF, et al. Retrogenic expression of the MET proto-oncogene correlates with the invasive phenotype of human rhabdomyosarcomas.
Oncogene
1996
;
12
:
1697
–705.
47
Jankowski K, Kucia M, Wysoczynski M, et al. Both hepatocyte growth factor (HGF) and stromal-derived factor-1 regulate the metastatic behavior of human rhabdomyosarcoma cells, but only HGF enhances their resistance to radiochemotherapy.
Cancer Res
2003
;
63
:
7926
–35.
48
Sharp R, Recio JA, Jhappan C, et al. Synergism between INK4a/ARF inactivation and aberrant HGF/SF signaling in rhabdomyosarcomagenesis.
Nat Med
2002
;
8
:
1276
–80.
49
Welm AL, Kim S, Welm BE, Bishop JM. MET and MYC cooperate in mammary tumorigenesis.
Proc Natl Acad Sci U S A
2005
;
102
:
4324
–9.