Normal cells are dependent on exogenous, receptor-mediated growth stimulation for cell cycle entry and progression, providing a critical homeostatic mechanism regulating cellular proliferation. In contrast, tumor cells acquire some degree of growth signal autonomy, often through their ability to produce growth factors as well as their receptors (autocrine signaling). Recently, data have begun to emerge implicating heterotypic signaling between diverse cell types within a tumor in the genesis and progression of cancer; however, current experimental approaches in vivo have not adequately addressed this critical relationship. Here we used transgenic mice overexpressing hepatocyte growth factor/scatter factor (HGF/SF), or its growth antagonist NK2, as genetically modified hosts for transplantation of tumor cells expressing their receptor, c-Met, to directly assess the contribution of heterotypic signaling to metastatic colonization. We demonstrate that metastatic potential under nonautocrine signaling conditions (i.e., where tumor cells expressing c-Met are transplanted into transgenic hosts producing HGF/SF) rivaled that observed under conditions of autocrine signaling (i.e., where tumor cells expressing both HGF/SF and c-Met are transplanted into wild-type hosts). HGF/SF and NK2 were not functionally equivalent in vivo. Attenuation of NK2-associated growth inhibition by the presence of an HGF/SF-Met autocrine loop uncovered a shift in metastatic site preference from lung to liver only in NK2-transgenic hosts, a qualitative behavioral alteration likely detectable only through the genetic approach used here. Our data demonstrate that growth factors not intrinsic to malignant cells can have profound effects on metastatic efficiency in vivo and provide experimental support of a role for heterotypic signaling in tumor progression.

Growth of normal cells is highly dependent on extracellular growth stimulation. A cell carrying a cell-surface receptor can respond to specific growth factors secreted by a neighboring cell (paracrine signaling) or a distant organ (endocrine signaling). This heterotypic signaling configuration provides an important homeostatic mechanism by which normal cellular proliferation is regulated. In contrast, tumor cells acquire some degree of growth factor independence, often through an ability to generate their own growth signals (1). The term “autocrine signaling” describes the situation in which a tumor cell gains the ability to express a growth factor as well as its corresponding receptor, generating a positive feedback loop (2). Human sarcomas and carcinomas, for example, have been shown to possess autocrine signaling loops involving a number of cell-surface receptor tyrosine kinases, including c-MET (reviewed in Refs. 3 and 4).

c-MET mediates a broad range of normal cellular activities, including growth, motility, and morphogenetic branching by binding its ligand, HGF/SF.2 HGF/SF induces autophosphorylation of c-MET, generating phosphotyrosine docking sites able to engage an array of SH2-containing signal transducers, which in turn activate appropriate signaling pathways (5). Under normal conditions, c-MET is expressed on epithelial cells and activated in a paracrine fashion by mesenchymally derived HGF/SF (6, 7, 8). HGF/SF-Met signaling is required for normal development of skeletal muscle, liver, and placenta (9, 10, 11). HGF/SF has 38% overall sequence similarity with plasminogen (12) and contains an α chain with an N-domain and four kringle domains and a β chain with an enzymatically inactive serine protease domain. HGF/SF RNA can undergo alternative splicing to create truncated isoforms able to bind c-Met and either mimic or antagonize HGF/SF, depending on the cell type and environmental context (13, 14). One natural variant possessing the N-domain and the first two kringle domains, NK2, has been reported to inhibit growth but stimulate motility, invasiveness, and tubulogenesis in cultured cells (15, 16, 17, 18, 19, 20, 21). The ability of NK2 to antagonize HGF/SF-induced phenotypes in vivo was demonstrated in studies with HGF/SF-NK2 bitransgenic mice (22). c-MET is also a proto-oncogene (23), involved in the development of human solid tumors, including melanoma (24, 25). Constitutive c-MET activation can stimulate angiogenesis, extracellular matrix dissolution, invasiveness and metastasis (18, 26) and promotes broad tumorigenesis in HGF/SF-overexpressing transgenic mice (27, 28).

In human cancer, aberrant c-MET stimulation can be achieved through activating c-MET mutations, c-MET amplification/overexpression, and acquisition of autonomous growth control through creation of autocrine signaling loops (5, 29). HGF/SF-Met autocrine loops have been detected in human primary and metastatic tumors, including melanoma, glioblastoma, osteosarcoma, and breast cancer (30, 31, 32, 33). Experimental recreation of autocrine loops through forced coexpression of c-Met and HGF/SF in immortalized cultured cells can confer a metastatic phenotype (34, 35). Recently, data have begun to emerge implicating heterotypic signaling pathways in tumor development and progression, originating from fibroblasts, endothelial cells, and inflammatory cells (1, 36). However, analysis of the contribution of organ microenvironment to metastasis has been limited to experimental manipulation of tumor cells in vitro and assessment of their behavior when transplanted into immunodeficient or syngeneic wild type mice (37). We recently reported that mice harboring transgenes encoding ligands for c-Met strongly support metastasis of transplanted melanoma cells characterized by a highly potent HGF/SF-Met autocrine loop. Here we directly assess the in vivo efficacy of nonautocrine c-Met signaling with respect to metastatic colonization using HGF/SF and NK2 transgenic mice as transplantation recipients.

Cell Culture, Transfection, and Reagents.

B16F1 murine melanoma cells were grown in DMEM supplemented with 10% FBS (Life Technologies, Inc.). The 37-7 cell line was derived from a neoplasm arising in HGF/SF transgenic mouse line MH37 (28) and maintained in DMEM plus 15% FBS, 5 ng/ml epidermal growth factor, and 5 μg/ml insulin (Upstate Biotechnology). B16F1 and 37-7 parent cell lines were transfected with the following expression vectors: TRE-HGF/SF plus CMV-tetracycline transactivator (ptTA3; Clontech); MT-NK2; CMV-c-Met; CMV-c-Met plus TRE-HGF/SF plus ptTA3; or MT-NK2 plus TRE-HGF/SF plus ptTA3. Although available, the tetracycline-inducible feature was not exploited in these studies. Different promoters were used to minimize competitive depletion of common transcription factors in cells carrying more than one expression vector. Stable clones expressing HGF/SF, NK2, c-Met, HGF/SF plus c-Met, and NK2 plus HGF/SF were established by selection for G418 (Life Technologies, Inc.) resistance. For each construct, two or three individual clones were subjected to in vitro and in vivo testing (described below) and shown to behave in a similar fashion. Phosphotyrosine monoclonal antibody (α-PY20) was from Transduction Laboratories, and HGF(N17) polyclonal antibody and anti-m-Met (SP260) polyclonal antibody were from Santa Cruz Biotechnology.

Transgenic Mice.

HGF/SF and NK2 transgenic mice were generated on an albino FVB background, using the expression constructs MT-HGF/SF and MT-NK2, respectively (22, 38). Host animals immunologically compatible with B16F1 melanoma cells were generated through a single cross of either MT-HGF/SF or MT-NK2 mice with C57BL/6 mice. Transgene expression profiles in both MT-promoter driven HGF/SF and NK2 transgenic mice were characterized by relatively strong expression in most tissues, including the liver and lung (Refs. 22 and 38; data not shown). Hepatic HGF/SF and NK2 transgene expression levels diverged by <3-fold based on measurement of RNA and protein (Ref. 22; data not shown). Positive HGF/SF and NK2 transgenic mice were identified by PCR using the following primer sets: for HGF/SF transgenic mice, MT-S (5′-ACTCGTCCAACGACTATA-3′) and HGF292 (5′-CTGAGGAATGTCACAGACTTCGTA-3′); and for NK2 transgenic mice, MT-S and YYL7 (5′-GCACTGGAGTGGCAACTTCC-3′). For histopathological analysis, mouse tissues were fixed in 10% buffered formalin, embedded in paraffin, sectioned at 5 μm, and stained with H&E. All mouse work was performed in accordance with guidelines established by the NIH.

Experimental Metastasis Analysis.

Either 1 × 105 or 106 cells in 0.2 ml of PBS were i.v. injected via the tail vein into 5–6-week-old male hosts: wild-type FVB, HGF/SF transgenic, or NK2 transgenic mice. Conclusions were essentially the same whether 105 or 106 melanoma cells were injected. Tumor numbers were obtained by visual inspection in mice euthanized 21 days after transplantation. Representative tissues were subjected to histopathological analysis. Nude mouse hosts confirmed metastasis patterns observed in wild-type mice (data not shown). Statistical analyses were performed using the paired Student t test.

Cell Growth Analysis.

Proliferation rates of cell lines were assessed by [3H]thymidine incorporation (39). Briefly, cells were plated in DMEM plus 10% FBS at 2 × 104 cells/well in 96-well plates in quadruplicate; after 24 h, cells were labeled for 4 h with [3H]thymidine (1 μCi/ml; NEN). Cells were then washed with ice-cold PBS and 5% trichloroacetic acid and solubilization in 0.25 n NaOH. After solubilization, lysates were neutralized with 3 n HCl and counted by liquid scintillation (Beckman).

Motility and Invasion Assays.

Motility and invasion assays were performed in 8-μm-pore transwells (6.5 mm; Costar) in quadruplicate (40). Cells were detached by trypsin, and 5 × 104 cells were plated in 200 μl of DMEM plus 10% FBS in the upper wells. Bottom wells contained 500 μl of the same medium. Cells were incubated at 37°C in humidified 5% CO2 for 24 h. Cells on the transwells were rinsed in PBS, fixed with 1% glutaraldehyde in PBS, and stained with 0.1% crystal violet in water for 30 min. After destaining in water, nonmigrating cells on top of the filter were removed with a cotton swab. Migrating cells on the bottom of the filter were solubilized in 500 μl 0.2% Triton X-100, and the absorbance measured at 590 nm. For invasion assays, transwell filters were layered with 100 μl of Matrigel (Collaborative Research) diluted 1:20 in PBS. After rinsing with PBS, cells were plated as above. Seventy-two h later, cells migrating to the bottom of the filter were evaluated, after removal of material from the upper side of the filter, by 0.1% crystal violet staining and measurement of solubilized dye at A590. Results using Matrigel and growth factor-reduced Matrigel were the same and were combined for statistical analysis.

Biochemical Analysis.

Immunoblots were performed on lysates obtained from cells grown in cell culture or tissues, solubilized in RIPA buffer [50 mm Tris (pH 7.4), 50 mm NaCl, 1% Triton X-100, 5 mm EDTA, plus protease inhibitor mixture (Boehringer Mannheim)]. Equivalent amounts of lysates were incubated with α-c-Met antibody (Santa Cruz Biotechnology) for 1 h, and then protein A/G agarose was added for 1 h and washed in RIPA buffer four times. Samples were fractionated by SDS 4–20% PAGE and transferred to nitrocellulose. Membranes were stripped using Re-Blot (Chemicon International) and reprobed as described.

HGF/SF-Met Signaling and Tumor Cell Behavior in Vitro.

C57BL/6-derived B16F1 melanoma cells, which contain very low levels of c-Met (40), were used as the parent cell line to generate stable transfectants that overexpress cDNAs encoding c-Met, HGF/SF, or both. Clonal cell lines of each type were selected that demonstrated readily detectable and comparable expression levels (Fig. 1,A and B). The same constructs were also used to make stable transfectants of the FVB-derived 37-7 tumor cell line (data not shown), which also makes very low levels of c-Met (28). Each selected cell line was characterized with respect to c-Met phosphorylation. c-Met phosphorylation was undetectable in either vector-transfected B16F1 control cells or in transfectants ectopically overexpressing c-Met alone (Fig. 1,B). In contrast, c-Met activity was detectable in cells expressing HGF/SF alone, and more pronounced in cells overexpressing both HGF/SF and ectopic c-Met (Fig. 1,B). These B16F1 cell lines were subjected to a series of in vitro assays for growth, motility, and invasiveness. Fig. 2 shows that the consequence of forced expression of c-Met and/or HGF/SF was similar for all three activities. In the absence of ligand, ectopic c-Met expression had a minor effect on growth, motility, or invasiveness. In contrast, forced expression of HGF/SF stimulated growth, motility, and invasiveness, an effect that was not appreciably enhanced by coexpression of ectopic c-Met. These data indicate that B16F1 cells are responsive to HGF/SF in vitro, and that within these cells c-Met is not limiting when stimulated through an autocrine mechanism.

Autocrine and Nonautocrine c-Met Signaling in Metastasis.

The stable cell lines described above were introduced into the tail vein of FVB wild-type, HGF/SF transgenic, or NK2 transgenic mice. Fig. 3 shows that endogenous levels of c-Met present in control B16F1 and 37-7 cells were sufficient to permit a modest but significant increase in lung metastasis in the ligand-rich environment provided by the HGF/SF transgenic, but not NK2 transgenic, host mice. Consistent with in vitro data (Fig. 2), genetic enhancement of c-Met expression alone did not significantly elevate the efficiency of pulmonary metastasis of B16F1 or 37-7 cells in FVB hosts (Fig. 3). In contrast, c-Met-overexpressing B16F1 and 37-7 cells demonstrated robust metastasis to the lung when introduced into HGF/SF transgenic or NK2 transgenic hosts (Fig. 3). These data show that the requirement for ligand exposure exhibited by tumor cells in vitro can be satisfied cell nonautonomously in vivo, and that nonautocrine stimulation of c-Met can promote metastasis.

We next directly compared the ability of autocrine and nonautocrine c-Met signaling to support metastasis of B16F1 cell lines. In these experiments, we used immunologically compatible transgenic hosts, the F1 product of cross-breeding between HGF/SF transgenic mice and C57BL/6 inbred mice. As before, ectopic c-Met expression alone did not affect metastatic efficiency to wild-type lungs, whereas forced HGF/SF expression significantly stimulated pulmonary metastasis, an affect that was marginally enhanced by ectopic c-Met coexpression (Fig. 4,A). These data indicate that ectopic cellular expression of HGF/SF, but not ectopic c-Met, is required for efficient B16F1 cell metastasis. To compare these results with nonautocrine stimulation, the same B16F1 lines were injected into HGF/SF transgenic mice. Fig. 4,A shows that the efficiency of pulmonary metastasis of B16F1 cells expressing low endogenous levels of c-Met was modestly but significantly (P < 0.01) heightened in HGF/SF transgenic mice relative to wild-type animals. However, host expression of the HGF/SF transgene had a dramatic effect on metastatic colonization of the c-Met-overexpressing B16F1 cell line (Fig. 4,A), a transfectant variant that did not metastasize efficiently in wild-type hosts. This differential, which could only be accounted for by a paracrine or endocrine mechanism, indicates that given ample receptor, constitutive c-Met activation through heterotypic stimulation can be as potent as autocrine signaling in promoting metastatic colonization. Fig. 4 shows representative sections of lung metastases induced through enhanced autocrine (Fig. 4,D) or nonautocrine (Fig. 4,E) signaling conditions. c-Met phosphorylation levels in tumor cells growing at metastatic sites in HGF/SF transgenic host mice rivaled those achieved through addition of HGF/SF in vitro (Fig. 1 C); in contrast, c-Met phosphorylation in tumor cell metastases from wild-type host mice was modest.

Qualitative Difference between HGF/SF- and NK2-mediated Signaling in Tumor Cell Behavior.

The i.v. introduction of various B16F1 lines into NK2 transgenic mice produced a pattern of pulmonary metastasis similar to HGF/SF transgenic hosts (Fig. 4,A). As in HGF/SF transgenic mice, B16F1 cells with ectopic c-Met expression showed a significant increase in metastatic potential to the lungs of NK2 hosts, and c-Met kinase activity was evident within metastases arising in this NK2-rich environment (Fig. 1,C). Notably, B16F1 cells expressing HGF/SF also demonstrated a dramatic shift in target organ preference to the liver, but only in NK2 transgenic hosts (Fig. 4, B and F). This shift in organotropism was most dramatic in cells ectopically expressing both HGF/SF and c-Met, creating a potent autocrine loop (Fig. 1,B). To elucidate the effects of NK2 on B16F1 behavior, in vitro assays were performed. Fig. 2 shows that although forced expression of NK2 alone in B16F1 cells stimulated motility and invasiveness, it strongly inhibited growth. When coexpressed with NK2, HGF/SF rescued the B16F1 cells from NK2-induced growth inhibition (Fig. 2,A). A similar pattern was observed when these B16F1 lines, or similarly constructed 37-7 lines, were i.v. injected into wild-type mice (Fig. 5). Whereas HGF/SF was stimulatory, expression of NK2 alone either failed to stimulate, or actually inhibited, pulmonary metastasis. However, coexpression of HGF/SF and NK2 produced a metastatic response that was significantly elevated relative to HGF/SF alone (Fig. 5).

It is evident that cell autonomous growth signals are not sufficient to explain the development or behavior of tumor cells. The tissue microenvironment at both primary and metastatic tumor sites can supply critical signaling factors for numerous functions, including cellular growth, motility, invasiveness, and survival, thought to be provided through heterotypic signaling (1, 36, 37). Recent data suggest that aspiring tumor cells “learn” to co-opt the aid of surrounding fibroblasts, endothelial cells, and immune cells (41, 42, 43, 44, 45). However, little experimental in vivo evidence exists directly demonstrating the efficacy of such nonautonomous signaling in tumor progression. Here we describe a genetically tractable experimental system capable of evaluating the potential contribution of heterotypic receptor tyrosine kinase signaling in vivo. Specifically, we show that pulmonary metastasis of c-Met-overexpressing tumor cells is stimulated when introduced into transgenic mice overexpressing either HGF/SF or its variant NK2, and that the metastatic potential of the resulting heterotypic c-Met signaling was equivalent to that of the HGF/SF-Met autocrine signaling loop.

c-MET has been associated with metastatic progression in melanoma and other tumors (reviewed in Refs. 4 and 46). B16F1 melanoma cells harbor very low levels of c-Met and as used here were weakly metastatic to the lungs of wild-type mice. This basal level of c-Met was nevertheless sufficient to support enhanced in vitro growth, motility, and invasiveness of B16F1 cells ectopically expressing HGF/SF. Similarly, B16F1 metastasis could be stimulated by exposure to high levels of HGF/SF, whether originating directly from the cell (autocrine stimulation) or from host tissue (paracrine and/or endocrine stimulation). However, greater c-Met levels were required to achieve efficient metastasis under conditions of nonautocrine stimulation, perhaps indicating that higher local concentrations of HGF/SF are reached through the autocrine mechanism. This observation is consistent with reports correlating tumor progression with c-MET overexpression/amplification (4, 46).

We are currently unable to discern the relative contributions of paracrine and endocrine c-Met stimulation to metastasis in our mouse models. HGF/SF and NK2 levels are elevated in both the target tissues and sera of HGF/SF transgenic and NK2 transgenic mice, respectively (22, 38). However, both paracrine and endocrine signaling represent highly relevant, and perhaps overlapping, mechanisms in human cancer. Elevated HGF/SF serum levels have been significantly associated with histological invasiveness, progressive metastasis, and prognosis in patients with breast, gastric, colorectal, and lung cancer (47, 48, 49, 50). Surgical resection of breast and gastric neoplasms significantly reduced serum HGF/SF levels (47, 51), suggesting that the source of the elevated HGF/SF was the tumor itself; however, the specific cellular origin within the tumor has not been clarified. Paracrine regulatory loops have been described in which human tumor cells coax neighboring stromal cells to secrete HGF/SF, in turn promoting expression of angiogenic factors by, and invasiveness of, the tumor cells (52, 53, 54, 55, 56). Furge et al.(57) have proposed that paracrine release of HGF/SF from host stromal tissue is essential for metastasis of Ras-transformed human carcinoma cells, which typically do not produce HGF/SF.

It has been suggested that c-Met overexpression and constitutive activation can favor the liver as a metastatic target of both mouse and human tumor cells (25, 40, 58). Here we show that HGF/SF transgene overexpression, either by tumor cells or transplantation hosts, did not by itself alter the preference of B16F1 or 37-7 tumor cells for metastasis to the lungs. These data are consistent with Lin et al.(59), who concluded that although c-Met activation could enhance the metastatic potency of B16 cells, it did not overtly influence organotropism. We were therefore surprised to discover that B16F1 cells possessing an HGF/SF-Met autocrine loop were highly metastatic to the liver in immunologically compatible NK2 transgenic hosts. These results were reminiscent of our earlier observation that host-derived NK2 facilitated hepatic metastasis of 37-32 melanoma cells, which also highly express both HGF/SF and c-Met (28), when transplanted into these same NK2 transgenic mice (22). Although in vitro growth and pulmonary metastasis of tumor cells were inhibited by expression of NK2 alone, consistent with previous reports of NK2 as an HGF/SF growth antagonist (15, 16, 22), this block was relieved by forced coexpression of HGF/SF (Fig. 2). One attractive hypothesis is that overcoming NK2 growth antagonism through HGF/SF overexpression uncovers NK2-initiated prometastatic pathways, such as phosphatidylinositol 3′-kinase/Akt activation. Although an explanation for the preference for hepatic metastasis in the NK2 transgenic host is currently obscure, our data clearly show that non-cell autonomous effects are critical in determining metastatic behavior.

In this report, we demonstrate the feasibility of using a genetically tractable transplantation system consisting of transgenic recipient mice and transfected tumor cells to directly assess the contribution of host factors to metastasis. As a test model, we compared the role of autocrine and nonautocrine c-Met signaling in experimental metastasis of B16F1 melanoma cells. We show that the metastatic potential of heterotypic signaling, initiated by growth factors originating from host tissue and not intrinsic to the tumor cell, can reach levels achieved through autocrine signaling. By using emerging techniques to achieve specific targeting and inducibility of transgene expression, we anticipate that this approach can be greatly refined to uniquely address questions about the relationship between malignant cells and their microenvironment in tumorigenesis.

Fig. 1.

Comparative expression and activity of c-Met and its ligands in cultured cells and metastatic tissue. A and B, Western blot analysis of lysates from cultured B16F1 transfectants harboring vectors expressing c-Met and/or HGF/SF (HGF). HGF/SF was visualized directly with an α-HGF/SF antibody (A), whereas c-Met was immunoprecipitated from lysates and visualized by probing with an α-c-Met antibody (B). The location of both p170 and p140 forms is shown. c-Met phosphorylation was determined by reprobing with antiphosphotyrosine antibody (α-pY20; B). C, quantification of immunoprecipitated c-Met and phosphorylated c-Met (as above) in lung metastases derived from transplantation of c-Met-transfected 37-7 and B16F1 cells. Cells were injected i.v. into immunologically compatible wild-type (WT), HGF/SF transgenic (TgHGF), or NK2 transgenic (TgNK2) mice. As a basis for comparison, lysates from cultured cells without and with HGF/SF addition were analyzed in identical fashion. D, Western blot analysis of lysates from B16F1 cells transfected with HGF/SF and/or NK2. Direct blotting with α-HGF antibody recognized both HGF and NK2.

Fig. 1.

Comparative expression and activity of c-Met and its ligands in cultured cells and metastatic tissue. A and B, Western blot analysis of lysates from cultured B16F1 transfectants harboring vectors expressing c-Met and/or HGF/SF (HGF). HGF/SF was visualized directly with an α-HGF/SF antibody (A), whereas c-Met was immunoprecipitated from lysates and visualized by probing with an α-c-Met antibody (B). The location of both p170 and p140 forms is shown. c-Met phosphorylation was determined by reprobing with antiphosphotyrosine antibody (α-pY20; B). C, quantification of immunoprecipitated c-Met and phosphorylated c-Met (as above) in lung metastases derived from transplantation of c-Met-transfected 37-7 and B16F1 cells. Cells were injected i.v. into immunologically compatible wild-type (WT), HGF/SF transgenic (TgHGF), or NK2 transgenic (TgNK2) mice. As a basis for comparison, lysates from cultured cells without and with HGF/SF addition were analyzed in identical fashion. D, Western blot analysis of lysates from B16F1 cells transfected with HGF/SF and/or NK2. Direct blotting with α-HGF antibody recognized both HGF and NK2.

Close modal
Fig. 2.

In vitro characterization of stable B16F1 transfectants expressing c-Met, HGF/SF, c-Met plus HGF/SF, NK2, or HGF/SF plus NK2. The proliferation rates (A) of cell lines were assessed by measuring incorporation of [3H]thymidine. Motility (B) and invasiveness (C) were determined using Transwell culture plates for 24 or 72 h, respectively. Control, transfection with vector alone. Bars, SD. All presented values have been normalized to vector-transfected control cells (set at 1.0). Relative to control, B16F1 cells transfected with c-Met, HGF/SF, c-Met plus HGF/SF, or NK2 plus HGF/SF expression vectors demonstrated significant increases in growth, motility, and invasiveness (from P < 0.05 to <0.0001). Cells expressing NK2 alone showed a significant decrease in growth (P < 0.0005) and increase in invasiveness (P < 0.05). Biochemical analysis of 37-7 cell transfectants yielded a similar pattern (data not shown).

Fig. 2.

In vitro characterization of stable B16F1 transfectants expressing c-Met, HGF/SF, c-Met plus HGF/SF, NK2, or HGF/SF plus NK2. The proliferation rates (A) of cell lines were assessed by measuring incorporation of [3H]thymidine. Motility (B) and invasiveness (C) were determined using Transwell culture plates for 24 or 72 h, respectively. Control, transfection with vector alone. Bars, SD. All presented values have been normalized to vector-transfected control cells (set at 1.0). Relative to control, B16F1 cells transfected with c-Met, HGF/SF, c-Met plus HGF/SF, or NK2 plus HGF/SF expression vectors demonstrated significant increases in growth, motility, and invasiveness (from P < 0.05 to <0.0001). Cells expressing NK2 alone showed a significant decrease in growth (P < 0.0005) and increase in invasiveness (P < 0.05). Biochemical analysis of 37-7 cell transfectants yielded a similar pattern (data not shown).

Close modal
Fig. 3.

Nonautocrine stimulation of c-Met signaling induces metastasis in vivo. Clones of B16F1 (105 cells) and 37-7 (106 cells), transfected with vector alone (Control) or CMV-c-Met (Met), were i.v. injected into 5–6-week-old male wild-type mice (FVB), HGF/SF transgenic mice (HGF), or NK2 transgenic mice (NK2). Between 5 and 12 recipient mice were used as transplantation hosts for each cell line tested. At 21 days after transplantation, all recipient mice were examined for the presence of tumors. Numbers (>100) over the columns, estimated minimal average tumor number; 80% of the mice from each group had >100 lung tumors, while the rest had at least 50. In addition to the displayed statistics, control B16F1 and 37-7 cells both showed significantly increased metastasis in HGF/SF hosts relative to wild-type FVB (P < 0.01 and <0.05, respectively). Bars, SD. NS, not significant.

Fig. 3.

Nonautocrine stimulation of c-Met signaling induces metastasis in vivo. Clones of B16F1 (105 cells) and 37-7 (106 cells), transfected with vector alone (Control) or CMV-c-Met (Met), were i.v. injected into 5–6-week-old male wild-type mice (FVB), HGF/SF transgenic mice (HGF), or NK2 transgenic mice (NK2). Between 5 and 12 recipient mice were used as transplantation hosts for each cell line tested. At 21 days after transplantation, all recipient mice were examined for the presence of tumors. Numbers (>100) over the columns, estimated minimal average tumor number; 80% of the mice from each group had >100 lung tumors, while the rest had at least 50. In addition to the displayed statistics, control B16F1 and 37-7 cells both showed significantly increased metastasis in HGF/SF hosts relative to wild-type FVB (P < 0.01 and <0.05, respectively). Bars, SD. NS, not significant.

Close modal
Fig. 4.

Comparison of autocrine and nonautocrine c-Met signaling in metastasis. A, autocrine and nonautocrine stimulation of c-Met signaling are both effective at inducing pulmonary metastasis in vivo. Clones of B16F1 transfectants expressing c-Met, HGF/SF, or c-Met plus HGF/SF were i.v. injected (see Fig. 3 legend) into immunologically compatible wild-type, HGF/SF transgenic or NK2 transgenic hosts, generated through a single cross with C57BL/6 mice. B, the ability to metastasize predominantly to liver emerges in melanoma cells possessing an HGF/SF-Met autocrine loop transplanted into transgenic host mice expressing the variant NK2. Between 4 and 5 recipient mice were used as transplantation hosts for each cell line tested. Control, transfection with an empty vector. >20 over the column, estimated minimal average tumor number. Bars, SD. NS, not significant. CF, representative histopathology of metastases of B16F1 transfectants; cells were transplanted into mice of the FVB × C57BL/6 genetic background. C, poor pleural metastatic potential demonstrated by c-Met-overexpressing B16F1 cells transplanted into a wild-type host. D, pulmonary metastasis in a wild-type host by B16F1 cells possessing an HGF/SF-Met autocrine loop. E, pleural metastasis driven by nonautocrine stimulation as a consequence of transplanting c-Met-overexpressing B16F1 cells into an HGF/SF transgenic host. F, B16F1 cells with an HGF/SF-Met autocrine loop demonstrate robust liver metastasis only in the NK2 transgenic host. Arrowheads, metastatic tissue. All sections were stained with H&E.

Fig. 4.

Comparison of autocrine and nonautocrine c-Met signaling in metastasis. A, autocrine and nonautocrine stimulation of c-Met signaling are both effective at inducing pulmonary metastasis in vivo. Clones of B16F1 transfectants expressing c-Met, HGF/SF, or c-Met plus HGF/SF were i.v. injected (see Fig. 3 legend) into immunologically compatible wild-type, HGF/SF transgenic or NK2 transgenic hosts, generated through a single cross with C57BL/6 mice. B, the ability to metastasize predominantly to liver emerges in melanoma cells possessing an HGF/SF-Met autocrine loop transplanted into transgenic host mice expressing the variant NK2. Between 4 and 5 recipient mice were used as transplantation hosts for each cell line tested. Control, transfection with an empty vector. >20 over the column, estimated minimal average tumor number. Bars, SD. NS, not significant. CF, representative histopathology of metastases of B16F1 transfectants; cells were transplanted into mice of the FVB × C57BL/6 genetic background. C, poor pleural metastatic potential demonstrated by c-Met-overexpressing B16F1 cells transplanted into a wild-type host. D, pulmonary metastasis in a wild-type host by B16F1 cells possessing an HGF/SF-Met autocrine loop. E, pleural metastasis driven by nonautocrine stimulation as a consequence of transplanting c-Met-overexpressing B16F1 cells into an HGF/SF transgenic host. F, B16F1 cells with an HGF/SF-Met autocrine loop demonstrate robust liver metastasis only in the NK2 transgenic host. Arrowheads, metastatic tissue. All sections were stained with H&E.

Close modal
Fig. 5.

HGF/SF expression can complement metastasis-promoting activities of NK2, overcoming growth-inhibitory effects. Transfected B16F1 and 37-7 cells were transplanted i.v. (see Fig. 3 legend) into wild-type mice. Between 5 and 14 recipient mice were used as hosts for each cell line transplanted. Numbers over columns, estimated minimal tumor number. Bars, SD.

Fig. 5.

HGF/SF expression can complement metastasis-promoting activities of NK2, overcoming growth-inhibitory effects. Transfected B16F1 and 37-7 cells were transplanted i.v. (see Fig. 3 legend) into wild-type mice. Between 5 and 14 recipient mice were used as hosts for each cell line transplanted. Numbers over columns, estimated minimal tumor number. Bars, SD.

Close modal

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.

2

The abbreviations used are: HGF/SF, hepatocyte growth factor/scatter factor; CMV, cytomegalovirus; FBS, fetal bovine serum; MT, metallothionein; TRE, tetracycline responsive element.

We thank Fred Vieira for assistance with tail vein injections; Drs. Bill LaRochelle, Andrew Chan, and Stuart Aaronson for the original HGF/SF and NK2 cDNAs; and Drs. George Vande Woude and Michael Jeffers for the mouse c-met cDNA. We are indebted to Richard Sharp and Drs. Hisashi Takayama and Toshi Otsuka for technical advice and to Drs. Lalage Wakefield, Jeff Rubin, and Chamelli Jhappan for stimulating discussions and critical reading of the manuscript.

1
Hanahan H., Weinberg R. A. The hallmarks of cancer.
Cell
,
100
:
57
-70,  
2000
.
2
Sporn M. B., Todaro G. J. Autocrine secretion and malignant transformation of cells.
N. Engl. J. Med.
,
303
:
878
-880,  
1980
.
3
Fedi P., Tronick S. R., Aaronson S. A. Growth factors Holland J. F. Bast R. C. Morton K. L. Frei E. Kufe D. W. Weichselbaum R. R. eds. .
Cancer Medicine
,
41
-64, Williams & Wilkins Baltimore  
1997
.
4
Jeffers M., Rong S., Vande Woude G. F. Hepatocyte growth factor/scatter factor-Met signaling in tumorigenicity and invasion/metastasis.
J. Mol. Med.
,
74
:
505
-513,  
1996
.
5
Furge K. A., Zhang Y. W., Vande Woude G. F. Met receptor tyrosine kinase: enhanced signaling through adapter proteins.
Oncogene
,
19
:
5582
-5589,  
2000
.
6
Iyer A., Kmiecik T. E., Park M., Daar I., Blair D., Dunn K. J., Sutrave P., Ihle J. N., Bodescot M., Vande Woude G. F. Structure, tissue-specific expression, and transforming activity of the mouse met protooncogene.
Cell Growth Differ
,
1
:
87
-95,  
1990
.
7
Sonnenberg E., Meyer D., Weidner K. M., Birchmeier C. Scatter factor/hepatocyte growth factor and its receptor, the c-met tyrosine kinase, can mediate a signal exchange between mesenchyme and epithelia during mouse development.
J. Cell. Biol.
,
123
:
223
-235,  
1993
.
8
Tsarfaty I., Rong S., Resau J. H., Rulong S., Pinto da Silva P., Vande Woude G. F. The Met proto-oncogene mesenchymal to epithelial cell conversion.
Science (Wash. DC)
,
263
:
98
-101,  
1994
.
9
Bladt F., Rlethmacher D., Isenmann S., Aguzzi A., Birchmeier C. Essential role for the c-met receptor in the migration of myogenic precursor cells into the limb bud.
Nature (Lond.)
,
376
:
768
-771,  
1995
.
10
Schmidt C., Bladt F., Goedecke S., Brinkman V., Zschiesche W., Sharpe M., Gherardi E., Birchmeier C. Scatter factor/hepatocyte growth factor is essential for liver development.
Nature (Lond.)
,
373
:
699
-702,  
1995
.
11
Uehara Y., Minowa O., Mori C., Shiota K., Kuno J., Noda T., Kitamura N. Placental defect and embryonic lethality in mice lacking hepatocyte growth factor/scatter factor.
Nature (Lond.)
,
373
:
702
-705,  
1995
.
12
Gherardi E., Stoker M. Hepatocyte growth factor-scatter factor: mitogen, motogen, and met.
Cancer Cells
,
3
:
227
-232,  
1991
.
13
Gherardi E., Hartmann G., Hepple J., Chirgadze D., Srinivasan N., Blundell T. Domain structure of hepatocyte growth factor/scatter factor (HGF/SF).
Ciba Found. Symp.
,
212
:
84
-93,  
1997
.
14
Jakubczak J., LaRochelle W. J., Merlino G. NK1, a natural splice variant of hepatocyte growth factor/scatter factor, is a partial agonist in vivo.
Mol. Cell. Biol.
,
18
:
1275
-1283,  
1998
.
15
Chan A. M.-L., Rubin J. S., Bottaro D. P., Hirschfield D. W., Chedid M., Aaronson S. A. Identification of a competitive HGF antagonist encoded by an alternative transcript.
Science (Wash. DC)
,
254
:
1382
-1385,  
1991
.
16
Miyazawa K., Kitamura A., Naka D., Kitamura N. An alternatively processed mRNA generated from human hepatocyte growth factor gene.
Eur. J. Biochem.
,
197
:
15
-22,  
1991
.
17
Hartmann G., Naldini L., Weidner K. M., Sachs M., Vigna E., Comoglio P. M., Birchmeier W. A functional domain in the heavy chain of scatter factor/hepatocyte growth factor binds the c-Met receptor and induces cell dissociation but not mitogenesis.
Proc. Natl. Acad. Sci. USA
,
89
:
11574
-11578,  
1992
.
18
Jeffers M., Rong S., Vande Woude G. F. Enhanced tumorigenicity and invasion-metastasis by hepatocyte growth factor/scatter factor-met signalling in human cells concomitant with induction of the urokinase proteolysis network.
Mol. Cell. Biol.
,
16
:
1115
-1125,  
1996
.
19
Stahl S. J., Wingfield P. T., Kaufman J. D., Pannell L. K., Cioce V., Sakata H., Taylor W. G., Rubin J. S., Bottaro D. P. Functional and biophysical characterization of recombinant human hepatocyte growth factor isoforms produced in Escherichia coli.
Biochem. J.
,
326
:
763
-772,  
1997
.
20
Castagnino P., Soriano J. V., Montesano R., Bottaro D. P. Induction of tissue inhibitor of metalloproteinases-3 is a delayed early cellular response to hepatocyte growth factor.
Oncogene
,
17
:
481
-492,  
1998
.
21
Montesano R., Soriano J. V., Malinda K. M., Ponce M. L., Bafico A., Kleinman H. K., Bottaro D. P., Aaronson S. A. Differential effects of hepatocyte growth factor isoforms on epithelial and endothelial tubulogenesis.
Cell Growth Differ.
,
9
:
355
-365,  
1998
.
22
Otsuka T., Jakubczak J., Vieira F., Bottaro D. P., Breckenridge D., LaRochelle W. J., Merlino G. Disassociation of met-mediated biological responses in vivo: the natural hepatocyte growth factor/scatter factor splice variant NK2 antagonizes growth but facilitates metastasis.
Mol. Cell. Biol.
,
20
:
2055
-2065,  
2000
.
23
Cooper C. S., Blair D. G., Oskarsson M. K., Tainsky M. A., Eader L. A., Vande Woude G. F. Characterization of human transforming genes from chemically transformed, teratocarcinoma, and pancreatic carcinoma cell lines.
Cancer Res.
,
44
:
1
-10,  
1984
.
24
Natali P. G., Nicotra M. R., Di Renzo M. R., Prat M., Bigotti A., Cavaliere R., Comoglio P. M. Expression of the c-Met/HGF receptor in human melanocytic neoplasms: demonstration of the relationship to malignant melanoma tumour progression.
Br. J. Cancer
,
68
:
746
-750,  
1993
.
25
Hendrix M. J., Seftor E. A., Seftor R. E., Kirschmann D. A., Gardner L. M., Boldt H. C., Meyer M., Pe’er J., Folberg R. Regulation of uveal melanoma interconverted phenotype by hepatocyte growth factor/scatter factor (HGF/SF).
Am. J. Pathol.
,
152
:
855
-863,  
1998
.
26
Rong S., Bodescot M., Blair D., Dunn J., Nakamura T., Mizuno K., Park M., Chan A., Aaronson S., Vande Woude G. F. Tumorigenicity of the met proto-oncogene and the gene for hepatocyte growth factor.
Mol. Cell. Biol.
,
12
:
5152
-5158,  
1992
.
27
Takayama H., LaRochelle W. J., Sharp R., Otsuka T., Kriebel P., Anver M., Aaronson S. A., Merlino G. Diverse tumorigenesis associated with aberrant development in mice overexpressing hepatocyte growth factor/scatter factor.
Proc. Natl. Acad. Sci. USA
,
94
:
701
-706,  
1997
.
28
Otsuka T., Takayama H., Sharp R., Celli G., LaRochelle W. J., Bottaro D., Ellmore N., Vieira W., Owens J. W., Anver M., Merlino G. c-Met autocrine activation induces development of malignant melanoma and acquisition of the metastatic phenotype.
Cancer Res.
,
58
:
5157
-5167,  
1998
.
29
Herynk M. H., Radinsky R. The coordinated functional expression of epidermal growth factor receptor and c-Met in colorectal carcinoma metastasis.
In Vivo
,
14
:
587
-596,  
2000
.
30
Ferracini R., Di Renzo M. F., Scotlandi K., Baldini N., Olivero M., Lollini P., Cremona O., Campanacci M., Comoglio P. M. The Met/HGF receptor is over-expressed in human osteosarcomas and is activated by either a paracrine or an autocrine circuit.
Oncogene
,
10
:
739
-749,  
1995
.
31
Tuck A. B., Park M., Sterns E. E., Boag A., Elliott B. E. Coexpression of hepatocyte growth factor and receptor (Met) in human breast carcinoma.
Am. J. Pathol.
,
148
:
225
-232,  
1996
.
32
Koochekpour S., Jeffers M., Rulong S., Taylor G., Klineberg E., Hudson E. A., Resau J. H., Vande Woude G. F. Met and hepatocyte growth factor/scatter factor expression in human gliomas.
Cancer Res.
,
57
:
5391
-5398,  
1997
.
33
Li G., Schaider H., Satyamoorthy K., Hanakawa Y., Hashimoto K., Herlyn M. Downregulation of E-cadherin and desmoglein 1 by autocrine hepatocyte growth factor during melanoma development.
Oncogene
,
20
:
8125
-8135,  
2001
.
34
Rong S., Segal S., Anver M., Resau J. H., Vande Woude G. F. Invasiveness and metastasis of NIH 3T3 cells induced by Met-hepatocyte growth factor/scatter factor autocrine stimulation.
Proc. Natl. Acad. Sci. USA
,
91
:
4731
-4735,  
1994
.
35
Jeffers M., Rong S., Anver M., Vande Woude G. F. Autocrine hepatocyte growth factor/scatter factor-Met signaling induces transformation and the invasive/metastatic phenotype in C127 cells.
Oncogene
,
13
:
853
-861,  
1996
.
36
Liotta L. A., Kohn E. C. The microenvironment of the tumour-host interface.
Nature (Lond.)
,
411
:
375
-379,  
2001
.
37
Fidler I. J. Seed and soil revisited: contribution of the organ microenvironment to cancer metastasis.
Surg. Oncol. Clin. North Am.
,
10
:
257
-269,  
2001
.
38
Takayama H., LaRochelle W. J., Anver M., Bockman D. E., Merlino G. Scatter factor/hepatocyte growth factor as a regulator of skeletal muscle and neural crest development.
Proc. Natl. Acad. Sci. USA
,
93
:
5866
-5871,  
1996
.
39
Shao R., Karunagaran D., Zhou B. P., Li K., Lo S. S., Deng J., Chiao P., Hung M. C. Inhibition of nuclear factor-κB activity is involved in E1A-mediated sensitization of radiation-induced apoptosis.
J. Biol. Chem.
,
272
:
32739
-32742,  
1997
.
40
Rusciano D., Lorenzoni P., Burger M. M. Expression of constitutively activated hepatocyte growth factor/scatter factor receptor (c-met) in B16 melanoma cells selected for enhanced liver colonization.
Oncogene
,
11
:
1979
-1987,  
1995
.
41
Kinzler K. W., Vogelstein B. Landscaping the cancer terrain.
Science (Wash. DC)
,
280
:
1036
-1037,  
1998
.
42
Skobe M., Fusenig N. E. Tumorigenic conversion of immortal human keratinocytes through stromal cell activation.
Proc. Natl. Acad. Sci. USA
,
95
:
1050
-1055,  
1998
.
43
Olumi A. F., Grossfeld G. D., Hayward S. W., Carroll P. R., Tlsty T. D., Cunha G. R. Carcinoma-associated fibroblasts direct tumor progression of initiated human prostatic epithelium.
Cancer Res.
,
59
:
5002
-5011,  
1999
.
44
Coussens L. M., Raymond W. W., Bergers G., Laig-Webster M., Behrendtsen O., Werb Z., Caughey G. H., Hanahan D. Inflammatory mast cells up-regulate angiogenesis during squamous epithelial carcinogenesis.
Genes Dev.
,
13
:
1382
-1397,  
1999
.
45
Holash J., Maisonpierre P. C., Compton D., Boland P., Alexander C. R., Zagzag D., Yancopoulos G. D., Wiegand S. J. Vessel cooption, regression, and growth in tumors mediated by angiopoietins and VEGF.
Science (Wash. DC)
,
284
:
1994
-1998,  
1999
.
46
Comoglio P. M., Boccaccio C. Scatter factors and invasive growth.
Semin. Cancer Biol.
,
11
:
153
-165,  
2001
.
47
Taniguchi T., Toi M., Inada K., Imazawa T., Yamamoto Y., Tominaga T. Serum concentrations of hepatocyte growth factor in breast cancer patients.
Clin. Cancer Res.
,
1
:
1031
-1034,  
1995
.
48
Taniguchi T., Kitamura M., Arai K., Iwasaki Y., Yamamoto Y., Igari A., Toi M. Increase in the circulating level of hepatocyte growth factor in gastric cancer patients.
Br. J. Cancer
,
75
:
673
-677,  
1997
.
49
Siegfried J. M., Weissfeld L. A., Singh-Kaw P., Weyant R. J., Testa J. R., Landreneau R. J. Association of immunoreactive hepatocyte growth factor with poor survival in resectable non-small cell lung cancer.
Cancer Res.
,
57
:
433
-439,  
1997
.
50
Fukuura T., Miki C., Inoue T., Matsumoto K., Suzuki H. Serum hepatocyte growth factor as an index of disease status of patients with colorectal carcinoma.
Br. J. Cancer
,
78
:
454
-459,  
1998
.
51
Beppu K., Uchiyama A., Morisaki T., Nakamura K., Noshiro H., Matsumoto K., Nakamura T., Tanaka M., Katano M. Elevation of serum hepatocyte growth factor concentration in patients with gastric cancer is mediated by production from tumor tissue.
Anticancer Res.
,
20
:
1263
-1267,  
2000
.
52
Rosen E. M., Goldberg I. D. Regulation of scatter factor (hepatocyte growth factor) production by tumor-stroma interaction Goldberg I. D. Rosen E. M. eds. .
Epithelial-Mesenchymal Interactions in Cancer
,
17
-31, Birkhauser Verlag Basel  
1995
.
53
Iwazawa T., Shiozaki H., Doki Y., Inoue M., Tamura S., Matsui S., Monden T., Matsumoto K., Nakamura T., Monden M. Primary human fibroblasts induce diverse tumor invasiveness: involvement of HGF as an important paracrine factor.
Jpn. J. Cancer Res.
,
87
:
1134
-1142,  
1996
.
54
Nakamura T., Matsumoto K., Kiritoshi A., Tano Y., Nakamura T. Induction of hepatocyte growth factor in fibroblasts by tumor-derived factors affects invasive growth of tumor cells: in vitro analysis of tumor-stromal interactions.
Cancer Res.
,
57
:
3305
-3313,  
1997
.
55
Hasina R., Matsumoto K., Matsumoto-Taniura N., Kato I., Sakuda M., Nakamura T. Autocrine and paracrine motility factors and their involvement in invasiveness in a human oral carcinoma cell line.
Br. J. Cancer
,
80
:
1708
-1717,  
1999
.
56
Dong G., Chen Z., Li Z-Y., Yeh N. T., Bancroft C. C., Van Waes C. Hepatocyte growth factor/scatter factor-induced activation of MEK and PI3K signal pathways contributes to expression of proangiogenic cytokines interleukin-8 and vascular endothelial growth factor in head and neck squamous cell carcinoma.
Cancer Res.
,
61
:
5911
-5918,  
2001
.
57
Furge K. A., Kiewlich D., Le P., Vo M. N., Faure M., Howlett A. R., Lipson K. E., Vande Woude G. F., Webb C. P. Suppression of Ras-mediated tumorigenicity and metastasis through inhibition of the Met receptor tyrosine kinase.
Proc. Natl. Acad. Sci. USA
,
98
:
10722
-10727,  
2001
.
58
Di Renzo M. F., Olivero M., Giacomini A., Porte H., Chastre E., Mirossay L., Nordlinger B., Bretti S., Bottardi S., Giordano S. Overexpression and amplification of the met/HGF receptor gene during the progression of colorectal cancer.
Clin. Cancer Res.
,
1
:
147
-154,  
1995
.
59
Lin S., Rusciano D., Lorenzoni P., Hartmann G., Birchmeier W., Giordano S., Comoglio P., Burger M. M. c-met activation is necessary but not sufficient for liver colonization by B16 murine melanoma cells.
Clin. Exp. Metastasis
,
16
:
253
-265,  
1998
.