The deregulation of Met/hepatocyte growth factor (HGF) receptor tyrosine kinase signaling constitutes a common event in colorectal cancers. However, the physiopathological functions of such a deregulation remain poorly understood. In the present study, we investigated the role of the deregulation of Met receptor in the neoplastic transformation of intestinal epithelial cells. To do so, the normal, well-established and characterized rat intestinal epithelial IEC-6 cells were transduced with a retrovirus carrying the oncogenic constitutive active form of Met receptor, Tpr-Met. Herein, we show that compared with control IEC-6 cells, Tpr-Met-IEC-6 cells exhibit enhanced proliferation, loss of growth-contact inhibition, cell morphological alterations, actin cytoskeletal reorganization, loss of E-cadherin expression and anchorage-independent growth. Moreover, Tpr-Met-IEC-6 cells are conferred the capacity to produce the proangiogenic factor VEGF and to reduce the potent antiangiogenic factor thrombospondin-1. Of significance, Tpr-Met-IEC-6 cells are endowed with the ability to elicit angiogenic responses and to form tumors and metastases in vivo. Hence, our study demonstrates for the first time that the sole oncogenic engagement of Met receptor in normal intestinal epithelial cells is sufficient to induce a wide array of cancerous biological processes that are fundamental to the initiation and malignant progression of colorectal cancers.
- intestinal epithelial transformation
- colorectal cancer
colorectal cancer (CRC) arises through a series of specific and well-characterized morphological stages known as the adenoma-carcinoma sequence (15). This sequence initiates with intestinal epithelial hyperplasia, which then becomes dysplastic, resulting in the formation of aberrant intestinal crypt foci and of benign tumors (termed adenomas). Invariably, a subset of those adenomas will progress toward malignant carcinomas unless removed surgically. The transition from benign adenomas to metastatic carcinomas involves the acquisition by the cancer cells of key biological properties, such as the ability to elicit angiogenic responses, anchorage-independent growth, and invasion (22). Around 15% of CRCs occur within the context of familial/hereditary predispositions, such as familial adenomatous polyposis and hereditary nonpolyposis colon cancer, whereas the remaining 85% arise sporadically (19). Previous studies have shown that the deregulation of components of the adenomatous polyposis coli/Wnt/catenin signaling pathway may contribute to the initiation of CRCs, whereas others such as K-Ras, p53, or DNA mismatch repair enzymes have been found to be mutated during the adenoma-carcinoma sequence (64). Despite the identification of such genetic alterations in CRCs, the precise molecular events responsible for neoplastic transformation and the molecular mechanisms leading to the transition of CRCs from a noninvasive to a metastatic malignant phenotype, remain poorly understood. To this effect, increasing evidence indicates that the deregulation of growth factor receptor tyrosine kinases (RTKs), including the Met/hepatocyte growth factor (HGF) receptor, may play a fundamental role in the etiology and progression of CRCs (reviewed in Refs. 9, 51).
The Met receptor was first identified as an oncoprotein, Tpr-Met, which is the product of a chromosomal rearrangement, fusing a leucine zipper dimerization domain (Tpr) to the nonmutated Met cytoplasmic tyrosine kinase region, thus resulting in a constitutively active cytosolic tyrosine kinase in the absence of ligand (11, 39, 46). Whereas the Met RTK is mainly expressed in epithelial and endothelial cells, its ligand HGF is a pleiotropic growth factor produced by cells of mesenchymal origin (3, 41). The Met/HGF signaling axis has been shown to play key roles in morphogenesis, organogenesis, and homeostasis, by mediating the regulation of several biological processes such as cell proliferation, migration, invasion, and survival, as well as angiogenesis (3, 41). Furthermore, deregulation of Met receptor signaling has been implicated in a variety of human malignancies, including CRC (3, 41, 51). Of significance, the Met receptor is not only overexpressed in a majority of CRCs at the earliest stages of the disease, but is furthermore overexpressed in virtually all invasive colorectal carcinomas (reviewed in Ref. 51). Incidentally, a recent study has shown that the coexpression of Met and HGF in primary colon cancers can predict tumor stage and clinical outcome (27). Overall, this suggests that aberrant activation of Met/HGF receptor axis may contribute to the etiology and progression of CRC by virtue of this receptor to mediate a variety of biological functions that, when deregulated, may cause intestinal epithelial cells to acquire neoplastic and malignant activities (22). However, the investigation of the potential contributions of Met receptor in CRC has been restricted so far to cancer-derived epithelial colorectal cell lines (51). Therefore, the physiopathological significance for activation of the Met receptor signaling in the early neoplastic transformation of the intestinal epithelium, as well as in the transition of CRCs from a noninvasive to a metastatic malignant phenotype remains poorly understood. This prompted us to investigate the roles of Met deregulation in intestinal epithelial cell tumorigenic transformation, using the well-characterized and normal-derived rat intestinal epithelial cell model, the IEC-6 cells.
In this present study, we show that the expression of the oncogenic Met receptor, Tpr-Met, in IEC-6 cells induces morphological cell alterations, enhanced proliferation, loss of growth-contact inhibition, anchorage-independent growth, along with the production of VEGF and the downregulation of thrombospondin-1 (TSP-1) protein. Of significance, we demonstrate that the oncogenic Met receptor confers to IEC-6 cells in vivo cancerous activities, such as eliciting angiogenic responses and the formation of subcutaneous tumors, as well as the capacity to form lung and liver metastases. Therefore, the oncogenic engagement of Met alone in normal intestinal epithelial cells is sufficient to promote a wide array of cancerous biological processes that are relevant to both the initiation and malignant progression of CRCs. This supports the concept that a deregulation of the Met receptor signaling pathways plays a fundamental role in the etiology and progression of CRCs.
MATERIALS AND METHODS
Antibodies and reagents.
The polyclonal Met 145 antibody that was raised in rabbit against a carboxy-terminal peptide of human Met was kindly provided by Dr. M. Park (McGill University, Montreal, QC, Canada). The p-Tyr100 phosphotyrosine antibody was from Cell Signaling Technology; Grb2 and E-cadherin specific antibodies (a gift of Dr. D. Ménard, Univ. de Sherbrooke) were from BD Transduction Laboratories (Lexington, KY). The VEGF, cyclin D, pan-Shc, and phosphotyrosine (Tyr239/240)-Shc specific antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA), whereas the α-tubulin and β-actin were from Sigma-Aldrich Canada (Oakville, ON, Canada), and the TSP-1 from Calbiochem-Novabiochem (San Diego, CA). Alexa Fluor-conjugated secondary antibody (555), Alexa Fluor-labeled phalloidin (488) and 4,6-diamidino-2-phenylindole (DAPI) were obtained from Molecular Probes (Invitrogen), whereas the anti-bromodeoxyuridine (BrdU) fluorescein-conjugated antibody was from Roche Applied Science (Mannheim, Germany).
The normal-derived rat intestinal epithelial crypt cell line, IEC-6, was provided by Dr. A. Quaroni (Cornell University, Ithaca, NY). The IEC-6 cells are well established and characterized to exhibit typical features of undifferentiated intestinal crypt cells, including an epithelioid morphology, sparse microvilli, and E-cadherin cell-cell interactions (43). Furthermore, IEC-6 cells exhibit characteristics of normal nontransformed epithelial cells, including growing as a monolayer with strong density inhibition of growth, lack of growth in soft-agar, low plating efficiency when seeded at low density, and inability to produce tumors when injected in syngenic animals (31, 43, 54, 63). The IEC-6 cells were maintained in DMEM containing 10% fetal bovine serum and 50 μg/ml gentamicin (Invitrogen). Retroviral infection with Tpr-Met pLXSN cDNA (kindly provided by Dr. M. Park, McGill University, Montreal, QC, Canada) was performed as described previously (50). Populations of pLXSN Tpr-Met expressing IEC-6 cells (Tpr-Met-IEC-6) were expanded from a pool of at least 50 neomycin-resistant colonies (400 μg/ml, Invitrogen). For each experiment, Tpr-Met-IEC-6 cells were paired with control IEC-6 cells that had comparable number of passages (±2) and a range between 11 to 25 passages.
Cells were seeded on glass coverslips (Bellco Glass, Vineland, NJ) at a density of 1–4 × 104 in 24-well plates (Nalgene, NUNC, Rochester, NY). After 2 days, serum-starved cells were washed twice in PBS (137 mM NaCl, 2.7 mM KCl, 10 mM Na2HPO4, 1.8 mM KH2PO4, pH 7.4) and fixed for 20 min at room temperature (RT) with 3% paraformaldehyde (PFA, Sigma-Aldrich Canada). After the removal of PFA by three 5-min washes with 100 mM glycine/PBS, the cells were permeabilized at RT for 5 min in PBS containing 0.2% Triton X-100 and were blocked for 30 min at RT in blocking buffer (5% BSA, 0.2% Triton X-100, 0.05% Tween 20, PBS). For detection of E-cadherin, cells were incubated with gentle rocking for 1 h at RT with primary E-cadherin antibody (1:1,000) diluted in blocking buffer, washed three times with PBS, and then incubated for 45 min in the presence of Alexa Fluor-labeled secondary antibody (555, 1:1,000) and Alexa Fluor-labeled phalloidin (488, 1:1,000) for visualization of polymerized actin. Nuclei were counterstained with DAPI, and coverslips were mounted onto slides in Immu-Mount medium (Thermo Scientific, Pittsburgh, PA) and images were photographed by using a Zeiss Axiovert 200 microscope (Carl Zeiss Canada, Toronto, ON, Canada).
Immunoprecipitation and immunoblots.
The methodology for the preparation of cell lysates including SDS-PAGE and Western blot analysis, in addition to the detection of VEGF in conditioned media of cells was previously described (35, 49, 50). All primary antibodies were used at a 1:1,000 dilution, with the exception of E-cadherin, cyclin D, α-tubulin, and β-actin that were diluted 1:5,000. Secondary antibodies were used at a dilution of 1:10,000 and proteins were visualized by enhanced chemiluminescence (GE Healthcare). Each experiment was performed with cells that were serum starved overnight and at least three times with independent preparation of cell lysates.
Proliferation, focus formation, and soft agar growth assays.
Proliferation assays were done with cells seeded at a density of 2.5 × 104 in six-well plates. The number of cells was counted daily and growth curve analyses were performed by use of Prism v3.0c (GraphPad software). Focus formation assays were performed with 200 of the tested cells seeded in six-well plates along with 5 × 105 of parental IEC-6 cells to form the monolayer. After ∼15 days, foci produced were photographed, fixed with 10% formalin buffer solution, and then stained with Giemsa for counting. Soft agar assays were performed with 5,000 cells embedded in Noble Agar in six-well plates as described previously (50). These assays were all performed at least three times in triplicate.
BrdU incorporation assay.
Cells were seeded on poly-l-lysine coated coverslips in 24-well plates at a density of 5,000 cells/well. After 3 days, the cells were incubated for 75 min with a 1:100 dilution of BrdU labeling reagent (Invitrogen) in complete medium and then were washed three times with cold PBS and were incubated for 45 min with fixing solution (3 volumes of 50 mM glycine pH 2 and 7 volumes 100% EtOH) and then rinsed two times with PBS. Cells were then denatured with 4 M HCl for 15 min, followed by 3 × 5-min wash and incubated for 10 min with incubation buffer (0.5% BSA, 0.1% Tween 20, PBS), and then 45 min with a 1:50 dilution of anti-BrdU in incubation buffer at 37°C. Coverslips were washed three times with PBS (with DAPI in the last wash) and then mounted onto slides in Immu-Mount medium. For each well, at least 15 random field views (×40) were photographed by using a Zeiss Axiovert 200 microscope (Carl Zeiss Canada, Toronto, ON, Canada), from which the number of BrdU-positive and DAPI-marked cells were enumerated.
In vivo tumorigenesis, angiogenesis, and experimental metastases assays.
The in vivo tumorigenesis, angiogenesis, and lung metastases assays were performed essentially as previously described (49, 50) with 4- to 5-wk-old female nude mice (CD1 nu/nu, Charles River) and under protocols approved by the Université de Sherbrooke Ethics Committee for Animal Care and Use. Briefly, for tumorigenesis assays, 106 cells/100 μl of DMEM were injected subcutaneously into mice and tumor volumes were measured periodically. The mice were euthanized prior to the tumor reaching 1 cm3 or at any signs of ulceration. For the in vivo angiogenesis assays, 106 cells resuspended in 250 μl of serum-depleted Matrigel (Becton Dickinson Labware, Bedford, MA) were injected subcutaneously into mice, hence allowing the maintenance of cells within the Matrigel. The resulting Matrigel plugs were photographed and collected after 10 days (6 Matrigel plugs/cell line). Experimental lung metastasis assays were performed by the injection of 106 cells/100 μl of DMEM into the tail vein of mice. For liver metastases assays via the intrasplenic/portal route, 106 cells were injected into the spleen of mice under anesthesia, which then later was removed. For these assays, animals were euthanized at any sign of respiratory distress or weight loss, or at a maximum of 35 days after cell injection. Livers and lungs, or any other tissues displaying metastases upon macroscopic examination were collected and photographed. The lungs were fixed in Bouin's staining solution to allow better visualization of pulmonary metastases.
Oncogenic Met induces transformation of intestinal epithelial cells.
As a first step to identify cancerous biological responses evoked by the oncogenic engagement of Met receptor in normal intestinal epithelial cells, the nontransformed IEC-6 cells were infected with a retrovirus encoding a known oncogenic constitutive active form of the Met receptor, Tpr-Met, or with a retrovirus carrying the empty expression vector (pLXSN). As shown in Fig. 1A, Tpr-Met-IEC-6 cells displayed drastic morphological alterations after 2 wk of selection. Such changes included an apparent breakdown of cellular junctions and cell dispersal, as well as the acquisition of a fibroblastic-like morphology, as evidenced by cells being highly refractile and spindle shaped (Fig. 1A). By contrast, the pLXSN control IEC-6 cells remained normal-behaving compared with sham-infected cells (not shown), as evidenced by still growing as colonies and their unaltered epithelioid morphology (Fig. 1A). The expression of the Tpr-Met oncoprotein was confirmed by immunoblot analysis (Fig. 1B).
We next verified that the oncogenic Met receptor could recruit and engage known Met-associated proximal signaling pathways when expressed in intestinal epithelial cells (Fig. 2A). Consistent with Tpr-Met being a constitutively activated tyrosine kinase, the latter was detected as a tyrosine phosphorylated protein when expressed in IEC-6 cells (Fig. 2B). As previously characterized for the Met receptor (16, 50), coimmunoprecipitation (co-IP) of Grb2 with Tpr-Met oncoprotein was evidenced in Tpr-Met-IEC-6 cells but not in control IEC-6 cells (Fig. 2B). Furthermore, specific activation of Shc-dependent signaling in Tpr-Met-IEC-6 cells was confirmed by detection of Grb2 in complex with Shc proteins (Fig. 2C; IP:Shc), as well as by detection of Shc proteins being phosphorylated on tyrosine residues by immunoblot analyses conducted with phospho-specific Shc antibody (Fig. 2C; TCL). Lastly, an enhanced tyrosine phosphorylation of the Gab1 adaptor protein, a major substrate for the Met receptor in epithelial cells (33, 65), was also evidenced in Tpr-Met-IEC-6 cells compared with control IEC-6 cells (Fig. 2D).
Oncogenic Met induces cytoskeleton reorganization and a loss of E-cadherin expression in intestinal epithelial cells.
The morphological alterations observed in Tpr-Met-IEC-6 cells suggested that the expression of the oncogenic Met in intestinal epithelial cells could induce cytoskeletal reorganization. To verify this, immunofluorescent staining for the detection of polymerized actin microfilaments was performed using Alexa Fluor 488-labeled phalloidin. As previously reported (31, 44) and expected from their normal epithelioid colony-growing behavior (Fig. 1A), control IEC-6 cells exhibited cuboidal morphologies with cortical actin architecture (Fig. 3A). However, in concordance with their observed cell dispersal and fibroblastic-like morphology (Fig. 1A), actin cytoskeletal rearrangements were readily apparent in Tpr-Met-IEC-6 cells, exhibiting a drastic loss of cortical actin at cell-cell junctions along with the formation of actin stress fibers mainly localized in lamellipodial-like structures (Fig. 3A).
The E-cadherin protein is a cell adhesion molecule involved in epithelial cell-cell contacts through the formation of adherens junctions, whose expression is often lost from cell-cell membrane contact points, or even shut down altogether, in transformed epithelial cancer cells (60). We therefore next examined the cellular distribution and the expression level of E-cadherin. Consistent with the normal, nontransformed features of IEC-6 control cells, E-cadherin was predominantly detected by indirect immunofluorescence at membrane cell-cell contact areas (Fig. 3A). On the contrary, E-cadherin was found localized in the cytosol of Tpr-Met-IEC-6 cells (Fig. 3A). Furthermore, the expression level of E-cadherin protein was severely reduced in Tpr-Met-IEC-6 cells, relative to control cells (Fig. 3B). These data overall indicate that the oncogenic engagement of Met is sufficient to transform intestinal epithelial cells, altering their morphological features and spreading behavior in culture, as well as their functional expression of E-cadherin.
Oncogenic Met induces proliferation, loss of growth-contact inhibition, and anchorage-independent growth in intestinal epithelial cells.
We next assessed which oncogenic activities were evoked in culture upon constitutive activation of the Met receptor in intestinal epithelial cells. As previously demonstrated (31, 42, 63), control IEC-6 cells were forming a monolayer composed of well-organized epithelial cells upon reaching confluence, indicative of them having normal growth inhibition at cell-cell contacts (Fig. 4A). In sharp contrast, Tpr-Met-IEC-6 cells continued to grow beyond confluence in a multilayered and highly disorganized fashion (Fig. 4A). Furthermore, Tpr-Met-IEC-6 cells were found to grow at considerably faster rates than control cells (doubling time in hours, Tpr-Met-IEC-6: 15.4 ± 0.3 and Control: 29.5 ± 1.9; Fig. 4B). Consistent with Tpr-Met promoting proliferation in the IEC-6 cells, BrdU incorporation (i.e., absolute numbers of BrdU-positive cells) and cyclin D1 protein levels, which are indicative of cell-cycle transition from G1 to S phase (34, 55), were much more elevated in subconfluent Tpr-Met-IEC-6 cells compared with control IEC-6 (Fig. 4, C and D). The loss of growth-contact inhibition exhibited by Tpr-Met-IEC-6 cells was further confirmed in focus-forming assays, whereby these cells rapidly formed numerous large foci of transformed cells, but not the control IEC-6 cells (Fig. 5, A and B). The ability of these cells to proliferate in absence of anchorage to the extracellular matrix was next analyzed in soft-agar assays. As anticipated (31, 42, 63), control IEC-6 cells were unable to grow in soft-agar (Fig. 5, C and D). On the other hand, Tpr-Met-IEC-6 cells grew efficiently in soft-agar, forming numerous large colonies (Fig. 5, C and D). Hence these data indicate that the oncogenic engagement of Met receptor signaling in normal intestinal epithelial cells is sufficient to confer enhanced proliferation, loss of growth-contact inhibition, and anchorage-independent growth.
Oncogenic Met induces in vivo angiogenic and tumorigenic activities in intestinal epithelial cells.
Our data demonstrated so far that constitutive engagement of Met receptor was sufficient to confer in intestinal epithelial cells a wide array of oncogenic properties in culture. We therefore next evaluated whether this could translate into oncogenic activities in vivo. We first tested the ability of Tpr-Met-IEC-6 vs. control IEC-6 cells to form tumors following their subcutaneous injection into nude mice. As previously reported (31, 42, 63), control IEC-6 cells failed to develop any tumors (Fig. 6A), even as far as 90 days after their injection. In contrast, Tpr-Met-IEC-6 cells were forming palpable tumors within a very short latency (∼5 days) that rapidly expanded (estimated doubling time of ∼6 days) into well-vascularized tumors (Fig. 6A and see below).
The prominent ability of Tpr-Met-IEC-6 cells to form fast growing and highly vascularized tumors raised the possibility that these cells had acquired the ability to induce angiogenesis, a host-tumor cell response critical for the expansion of neoplasms (2, 18). To validate this, in vivo angiogenesis Matrigel plug assays were performed, where Tpr-Met or control IEC-6 cells were mixed with a cold Matrigel solution depleted of growth factors for their subcutaneous injection into nude mice, hence allowing the confinement of cells within a solidified Matrigel plug. As shown in Fig. 6B, Matrigel plugs of Tpr-Met-IEC-6 cells were extensively infiltrated by blood vessels, whereas the Matrigel plugs containing control cells remained clear and poorly vascularized (Fig. 6B).
It is an overall switch in the balance of pro- and antiangiogenic factors that govern the formation of new blood vessels in tumors (2). Among the many angiogenic regulators, VEGF represents the most prevalent proangiogenic factor being upregulated in human cancers, whereas a reduction in TSP-1 expression, a natural potent inhibitor of angiogenesis, is frequent in malignancies (2). Incidentally, activation of the Met receptor in cells was shown to trigger such regulation in VEGF and TSP-1 expression (48, 67). We thus next evaluated the extent of VEGF protein produced in the culture media [conditioned media (CM)] of control or Tpr-Met-IEC-6 cells, and the TSP-1 protein level in lysates of these cells. As shown in Fig. 6C, VEGF protein was poorly detected in the CM of control IEC-6 cells, whereas drastically stronger levels of VEGF protein were found in the CM of Tpr-Met-IEC-6 cells. In addition, whereas TSP-1 protein was readily detected in control IEC-6 cells lysates, this antiangiogenic factor was barely detectable in Tpr-Met-IEC-6 cell lysates (Fig. 6D). Hence the proficiency of Tpr-Met-IEC-6 cells to induce angiogenic responses and to form fast growing tumors in nude mice correlates with their ability to concurrently enhance the VEGF production and to induce a drastic downmodulation in TSP-1 protein expression.
Oncogenic Met induces in vivo metastatic activities in intestinal epithelial cells.
Lastly, the in vivo metastatic activity of control and Tpr-Met-IEC-6 cells was evaluated by performing experimental lung and liver metastasis assays. Following their injection into the tail vein of nude mice, widespread lung metastasis was observed within 25 days in the case of Tpr-Met-IEC-6 cells, whereas the lungs of mice injected with control IEC-6 cells failed to show any signs of lung metastases (Fig. 7A). On the other hand, intrasplenic injection of Tpr-Met-IEC-6 cells into nude mice gave rise to the formation of multiple focal liver metastases (Table 1; Fig. 7B), as well as, remarkably, of lung and/or mesenteric lymph node metastases and hemorrhagic ascites, in some cases. In contrast, control IEC-6 cells failed to metastatically propagate when injected into mice via the splenic/portal route (Table 1; Fig. 7B).
Altogether, these data demonstrate that the oncogenic engagement of Met is sufficient to induce a variety of oncogenic behaviors in normal intestinal epithelial cells, conferring them strong in vivo tumorigenic, angiogenic, and metastatic activities.
In the present study, we demonstrate that constitutive engagement of the Met receptor signaling is sufficient to promote the oncogenic transformation of normal intestinal epithelial cells. Specifically, we show in a nontransformed rat IEC-6 cell model that the expression of an oncogenic active form of Met receptor, Tpr-Met, induces cell-cell contact breakdown, cell dispersal, loss of cortical actin architecture and of E-cadherin, and the acquisition of a fibroblastic-like morphology (Figs. 1 and 3). Moreover, we establish that Tpr-Met enhances the intestinal epithelial cell capacity to proliferate, in addition to escaping normal contact growth-inhibitory signals and acquiring anchorage-independent growth (Figs. 4 and 5).
Most of our current knowledge on the role of oncogenes in intestinal epithelial cell carcinogenesis is derived from studies conducted with cell cultures from human colon cancers and rodents. This is explained by the paucity of accessible normal intestinal epithelial cell culture models of human origin and by the fact that those existing are nonimmortalized cells, for which oncogenic-induced senescence represents a recurrent limitation for studying oncogene-mediated functions (10). Indeed, whereas the expression of a constitutively activated MEK was reported to transform the IEC-6 cells, the latter was shown to induce premature senescence in the nonimmortalized human intestinal epithelial crypt cell model (5, 31). Hence the reliance on a rat-derived intestinal epithelial cell line represents a limitation of our study, and to what extent the Met-induced oncogenic behaviors in rat IEC-6 cells are actually taking place in the human context remains to be ascertained. Nevertheless, our results support a study showing that the expression of a mutagenesis-generated constitutive active mutant of Met receptor in primary fetal colon epithelial cells promotes the downregulation of E-cadherin expression level, the formation of colonies in soft agar, and tumorigenic capability (4). Furthermore, we provide additional functional bases, showing notably for the first time that the sole oncogenic activation of the Met receptor signaling in normal intestinal epithelial cells concurrently evokes enhanced production of VEGF and downregulation of TSP-1 as well as promotes potent angiogenic, tumorigenic, and metastatic activities in vivo (Figs. 6 and 7).
The drastic cell morphological alterations observed in this study for Tpr-Met-expressing IEC-6 cells, including actin cytoskeletal reorganization and loss of E-cadherin expression (Fig. 3), are highly reminiscent of those occurring during epithelial-mesenchymal transition (EMT). This is a process viewed as being prerequisite for epithelial cell migration and invasion, taking place in normal developmental processes, in addition to playing a determining role in the metastatic dissemination of epithelial tumor cells (60). The EMT is commonly associated with actin cytoskeletal rearrangements and the breakdown of cell-cell junctions, therefore facilitating cell dissociation. Hallmarks of EMT include a redistribution of E-cadherin from cell-cell membrane contacts to the cytosol and/or a frequent loss of E-cadherin expression altogether (60). Incidentally, physical interaction between the Met receptor and E-cadherin, as well as their colocalization at cell-cell contacts in CRC, cells has been reported, and EMT-like changes, such as a loss of E-cadherin, are common in colorectal tumors (24, 45, 61). Similarly, several studies point to the importance of Met receptor signaling in E-cadherin regulation in gastric carcinogenesis. For instance, HGF promotes EMT-like morphological changes and invasion, along with a relocalization of E-cadherin at the membrane and a loss in its expression in gastric cancer cell lines, and a relationship between high levels of HGF and E-cadherin cytosolic expression in gastric tumors was reported (21, 59). Interestingly, activation of the Met/HGF signaling is also evidenced to play predominant roles in Helicobacter pylori infection-mediated gastric carcinogenesis. In this respect, H. pylori infection augments HGF expression in gastric mucosa and also promotes proliferation, EMT-like morphological changes, and loss of E-cadherin expression, thereby enhancing migratory and invasive phenotypes in gastric epithelial cells (8, 30, 52). Notably, the H. pylori virulent bacterial CagA protein, which is required for these H. pylori-mediated responses in gastric epithelial cells, has the ability to interact and mediate Met receptor activation (7, 57). Considering these observations, it is therefore not surprising for an oncogenic form of the Met receptor to evoke EMT-like changes in normal-derived colonic and intestinal epithelial cells as established previously and in this present study (Fig. 1⇑⇑–4). Further analyses will be required to determine whether such observed Met receptor-driven EMT-like changes in intestinal epithelial cells do constitute a genuine EMT process and to identify the specific signaling pathways engaged by the Met receptor that produce these changes in intestinal epithelial cells.
The metastatic dissemination of colorectal carcinoma cells to the liver or lungs represents the principal cause of death in patients afflicted with CRC. The metastatic cascade requires the cancer cells to first dissociate from primary sites and enter into the blood circulation, where they need to survive in the bloodstream, arrest in the vasculature of a distant organ, extravasate into the surrounding tissue, and finally reinitiate proliferation and sustain angiogenesis, all to expand in the newly colonized distant organ (6). Hence the conversion of adenoma into metastasizing carcinoma in CRCs implies the acquisition by cancer cells of a wide array of biological properties. Studies have documented that HGF stimulation or overexpression of the Met receptor can evoke metastatic properties in human and mouse CRC tumor-derived cell lines, including cell scattering, migration, and invasion in culture (13, 28, 29, 32, 56). Conversely, inhibition of Met receptor signaling in highly metastatic CRC cell lines through various approaches was shown to reduce their metastatic activities, namely blocking cell motility and invasiveness in vitro or reducing their capacity to form metastases in animal models (1, 23, 40, 56, 62, 66). These studies performed within the context of established CRC-derived cell lines infer for a role of the Met/HGF receptor-signaling axis in CRC metastases. Our finding that Tpr-Met-expressing IEC-6 cells form extensive lung and liver metastases in mice (Fig. 7) not only substantiates a role of the Met receptor signaling in CRC metastases but furthermore implies that deregulation of the Met receptor, by evoking a variety of oncogenic biological processes, is itself an instigating event in malignant progression of CRC. Consistent with this notion, increased expression of Met receptor is being recognized as a powerful indicator for early stages of invasion and metastasis of colorectal human tumors (12, 14, 27, 58). Our observed EMT-like changes in oncogenic Met-expressing IEC-6 cells, along with their concomitant loss of growth-contact inhibition and acquisition of anchorage-independent growth ability (Figs. 1⇑⇑⇑⇑–6), suggest that activation of Met receptor signaling may provide a strong propensity for intestinal epithelial cancer cells to leave the primary tumor site and enter the circulation, subsequently reaching target organs for metastasis. Notably, our experimental metastasis assays clearly demonstrate the strong proficiency of Tpr-Met-expressing IEC-6 cells to form lung and liver metastases (Fig. 7). Because these are assays recapitulating all the late stages of the metastatic cascade, this therefore provides evidence that deregulation of Met receptor is sufficient to endow noncancerous intestinal epithelial cells with capacities to survive in the bloodstream, to extravasate and to home in a new environment, and eventually to form tumor masses in the targeted organ.
Another key ability for tumor growth and secondary organ colonization is the capacity of cancer cells to induce angiogenesis (2). The induction of angiogenesis is suggested to constitute a critical event during the early premalignant phase of CRC development and metastatic progression (20). Enhanced microvessel density in human primary colonic adenomas and invasive carcinomas, relative to normal colonic epithelial mucosa is associated with metastatic dissemination and unfavorable prognosis (20). Highlighting the importance of VEGF and TSP-1 deregulation in CRC angiogenesis, elevated VEGF and lower TSP-1 expression in primary colorectal tumors were correlated with microvessel density, tumor progression, and patient survival (20, 26). Herein, we show that the expression of the Tpr-Met oncoprotein in normal-derived intestinal epithelial IEC-6 cells promotes VEGF production and reduces TSP-1 protein level, thus providing potential mechanistic basis for the ability of Tpr-Met-IEC-6 cells to evoke strong angiogenic properties in vivo (Fig. 6). Considering that the Met receptor is often expressed at higher levels in dysplastic crypt foci and in most colonic adenomas than in the normal adjacent mucosa (12, 25), our results support the notion that deregulation of Met receptor signaling in CRC might contribute to the early onset of the angiogenic switch, in part by regulating the expression of both VEGF and TSP-1.
The ability of the Met receptor to regulate a variety of biological functions is attributed to its capacity to recruit multiple signaling proteins and thereby to activate a wide array of downstream cellular signaling pathways (3, 41). Structure/function studies indicate similar and distinct involvement for Grb2, Shc, and Gab1 adaptor proteins in Met-driven oncogenesis. Whereas the recruitment of Grb2 or Shc by the Tpr-Met oncoprotein is sufficient and critical for promoting cell-cycle progression, oncogenic transformation, anchorage-independent growth, and metastasis, Shc plays a unique and critical role in Met-driven VEGF production and angiogenesis (17, 36, 49, 50). Alternatively, Gab1 is essential for Met-mediated invasive branching morphogenesis of MDCK epithelial cells, as well as proliferation and transformation of fibroblasts (35, 38, 65). The role of the Met receptor proximal signaling effectors in intestinal epithelial carcinogenesis remains poorly characterized. Nonetheless, it was recently shown that enhanced tumor progression induced by overexpression of the Met receptor in CRC cells harboring a K-Ras mutation is dependent on Gab1, but not on Grb2 or Shc (53). Herein, we have confirmed that the Tpr-Met oncoprotein associates with the adaptor protein Grb2 in intestinal epithelial IEC-6 cells and engages Shc- and Gab1-dependent signaling pathways (Fig. 2). These likely constitute part of the signaling basis of our observed biological processes driven by oncogenic Met receptor in nontransformed intestinal epithelial cells. Of particular relevance, Grb2, Shc, and Gab1 adaptor proteins share the capacity to activate the mitogenic RAS/MAPK and survival PI3K/Akt signaling cascades downstream of the Met receptor (51). Additional intracellular signaling pathways activated by Met receptor might as well be involved, including the c-Jun NH2-terminal kinase (JNK), c-Src kinase, and the signal transducer and activator of transcription 3 (STAT3) signaling pathways, for which studies point to their involvement in Met-induced tumorigenesis and metastasis, as well as in CRC progression (3, 37, 41, 47). Further studies will be necessary to define the specific biological relevance of Met receptor-proximal signaling proteins and downstream activated cellular signaling pathways in CRC progression. In this respect, our study puts emphasis on the oncogenic activation of Met receptor playing an integral part in the initial development of CRC by regulating discrete specific oncogenic biological activities. It also unveils that the intestinal epithelial crypt-derived IEC-6 cell model, in conjunction with the Tpr-Met oncoprotein, represent a powerful cell system to further elucidate oncogenic Met signaling in neoplastic transformation of normal intestinal epithelial cells.
In summary, we have demonstrated that a variety of cancerous biological processes, including angiogenic and metastatic activities, are engaged in normal intestinal epithelial cells upon the sole oncogenic activation of Met receptor. This strongly supports the concept that a deregulation of Met signaling can account for early events in neoplastic transformation of the intestinal epithelium, as well as in the transition of CRCs from a noninvasive to a metastatic malignant phenotype. A better understanding of the specific proximal signaling pathways involved in these processes would certainly support the rationale for targeting Met and/or its signaling in colorectal malignancies.
This work was supported by a CIHR grant (MOP 84382) awarded to C. Saucier, who is a scholar from the Fonds de la Recherche en Santé du Québec (FRSQ) and member of the FRSQ-funded Centre de Recherche Clinique Étienne Lebel.
No conflicts of interest, financial or otherwise, are declared by the author(s).
The authors are grateful to members of the Canadian Institutes of Health Research (CIHR) team on the digestive epithelium for providing equipments and reagents. We also thank members of the Saucier laboratory and Dr. Pierre H. Vachon for helpful comments upon critical reading of the manuscript. In memorial of Dr. Jeremy R. Jass, thank you so much for your inspiration.
- Copyright © 2010 the American Physiological Society