Abstract
Non–small cell lung cancer (NSCLC) is a difficult disease to treat. The c-Met receptor is an attractive potential target for novel therapeutic inhibition in human cancers. We provide strong evidence that c-Met is overexpressed, activated, and sometimes mutated in NSCLC cell lines and tumor tissues. Expression of c-Met was found in all (100%) of the NSCLC tumor tissues examined (n = 23) and most (89%) of the cell lines (n = 9). Sixty-one percent of tumor tissues strongly expressed total c-Met, especially adenocarcinoma (67%). Specific expression of phospho-Met (p-Met) [Y1003] and [Y1230/1234/1235] was seen by immunohistochemistry. p-Met expression was preferentially observed at the NSCLC tumor invasive fronts. c-Met alterations were identified within the semaphorin domain (E168D, L299F, S323G, and N375S) and the juxtamembrane domain (R988C, R988C + T1010I, S1058P, and alternative splice product skipping entire juxtamembrane domain) of a NSCLC cell line and adenocarcinoma tissues. We validated c-Met as potential therapeutic target using small interfering RNA down-regulation of the receptor expression by 50% to 60% in NSCLC cells. This led to inhibition of p-Met and phospho-AKT and up to 57.1 ± 7.2% cell viability inhibition at 72 hours. The selective small molecule inhibitor of c-Met SU11274 inhibited cell viability in c-Met-expressing NSCLC cells. SU11274 also abrogated hepatocyte growth factor–induced phosphorylation of c-Met and its downstream signaling. Here, we provide first direct evidence by small interfering RNA targeting and small molecule inhibitor that c-Met is important in NSCLC biology and biochemistry. These results indicate that c-Met inhibition will be an important therapeutic strategy against NSCLC to improve its clinical outcome.
Introduction
There are expected to be 173,770 new cases of lung cancer with a high estimated death rate of 160,440 cases in the United States for 2004 (1). Seventy-five percent to 85% will be composed of non–small cell lung cancer (NSCLC). NSCLC is further classified into adenocarcinoma (40%), including bronchoalveolar carcinoma, squamous cell carcinoma (30-35%), and large cell carcinoma (5-15%). Current therapy for NSCLC is surgery and adjuvant chemotherapy for early-stage disease and palliative chemotherapy (and/or radiation therapy) for advanced disease. For metastatic disease, doublet cytotoxic chemotherapy is the standard; however, the median survival time is still only 8 months (2). The epidermal growth factor receptor pathway has been shown to play a vital role in the pathogenesis of NSCLC, leading to the development of targeted therapeutic agents using small molecule inhibitors, such as gefitinib, erlotinib, and cetuximab (3–5). At best, response rates are 15%, with the population responding to epidermal growth factor receptor inhibition carrying a tyrosine kinase mutation (6, 7). Disappointingly, combinations of gefitinib or erlotinib with cytotoxic chemotherapies do not yield any superior advantage over chemotherapy alone (5, 8–10). There are several other molecules being used in NSCLC; however, no significant activities have been observed (3). Suboptimal results with epidermal growth factor receptor inhibition, however, have led to investigations of other receptor tyrosine kinases, such as c-Met, as potential therapeutic targets (3).
The c-Met gene is located on chromosome 7, band 7q31, and spans >120 kb long, consisting of 21 exons separated by 20 introns (11). In wild-type cells, the primary transcript produces a 150-kDa polypeptide, which gets partially glycosylated, and produces a 170-kDa precursor protein. This is further glycosylated and then cleaved to produce a 50-kDa α-chain and a 140-kDa β-chain, which are then linked by disulfide bonds. The ligand for c-Met has been identified as hepatocyte growth factor (HGF), also known as scatter factor (for review, see refs. 12, 13). Signaling through the c-Met/HGF pathway has been shown to trigger a variety of cellular responses that may vary based on the cellular context. In vivo, c-Met/HGF signaling plays key role in growth, motility, invasion, metastasis, angiogenesis, wound healing, and tissue regeneration. Higher levels of HGF have also been associated with more aggressive biology and a worse prognosis in NSCLC (14) and small cell lung cancer (SCLC; ref. 15). c-Met is normally expressed by epithelial cells and has been found to be overexpressed and amplified in a variety of human tumor tissues (16–20).
Many missense mutations of c-Met have been reported in a variety of cancers, with the majority of them identified in the cytoplasmic activation loop tyrosine kinase domain (13). Identification of activating germ-line mutations of c-Met in hereditary papillary renal carcinomas provided the first direct evidence linking c-Met directly to human oncogenesis (21). Juxtamembrane domains of receptor tyrosine kinases are thought to be key regulators of catalytic functions. In a recent study, we have examined the complete c-Met gene for potential mutations in SCLC (22). We identified unique activating mutations of c-Met in the juxtamembrane domain (R988C and T1010I) and semaphorin (Sema) domain (E168D) as well as insertions and alternative splice forms involving the juxtamembrane and extracellular domains of the proto-oncogene. Mutations of c-Met had not been reported previously in NSCLC.
With the normal wild-type c-Met transcript and altered mutations of c-Met, there have been several strategies to inhibit this receptor tyrosine kinase and its downstream pathway. These include antibody inhibition, small molecule inhibitors, and antisense techniques (13).
In this report, we have analyzed the functional expression and activation of c-Met in NSCLC—both cell lines and tumor tissues. We show that c-Met is functional and inhibitable by small interfering RNA (siRNA) and SU11274 in NSCLC. Finally, unique mutations and sequence variants of c-Met were also identified in NSCLC. With these findings, we believe that c-Met will be an attractive target for therapeutic inhibition in NSCLC clinical trials.
Materials and Methods
Cell Lines, Cell Culture, and Cell Viability. All NSCLC cell lines (A549, H1838, H2170, SW900, SW1573, H358, SK-LU-1, H1993, and H661) were obtained from the American Type Culture Collection (Rockville, MD) and maintained as described (http://www.atcc.org/). For stimulation studies with HGF, cells were deprived of growth factors by incubation in serum-free medium containing 0.5% bovine serum albumin for up to 72 hours.
For cell viability assay using SU11274, cells were plated in six-well culture plates and allowed to adhere overnight. SU11274 was then added the following day at the indicated concentrations, and DMSO was added to the control at a dilution of 1:1,000. The viability of the cells was determined 72 hours after drug addition by trypan blue exclusion assay at least in duplicates as described previously (22). Cell viability assay of A549 cells treated with or without siRNA was done using 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay (Sigma-Aldrich, Inc., St. Louis, MO) in triplicates as described previously (22).
Reagents. SU11274: [(3Z)-N-(3-chlorophenyl)-3-({3,5-dimethyl-4-[(4-methylpiperazin-1-yl)carbonyl]-1H-pyrrol-2-yl}methylene)-N-methyl-2-oxo-2,3-dihydro-1H-indole-5-sulfonamide] (Pfizer, Inc., San Diego, CA) was dissolved in DMSO and used at the concentrations described.
Tumor Samples and DNA Sequence Analysis. NSCLC tumor samples were collected with informed consent and in conformation with institutional guidelines as described previously (23). The cDNA from cell lines and tissue samples were prepared by standard procedures. The coding regions of the Sema domain, juxtamembrane domain, and tyrosine kinase domain in the c-Met cDNAs were sequenced using standard PCR and sequencing techniques as described previously (22). The adjacent “normal” lung tissues of the surgically resected tumor specimens (identified and separated histopathologically) where available were also sequenced using the genomic DNA isolated. The PCR primer sequences used for the c-Met mutational analyses (Sema domain, juxtamembrane domain, and tyrosine kinase domain) are available on request. The nucleotide positions numbering is relative to the first base of the translational initiation codon according to the full-length human c-Met cDNA (22, 24). This numbering corresponds to the full-length cDNA, including the “wild-type” full-length 154-bp exon 10 found as the minor alternative splice transcript. The predominant alternative splice transcript version is the one with the first 54 bp of the exon 10 skipped. Hence, the phosphospecific antibodies against the c-Met phosphoepitopes [pY1003] and [pY1230/1234/1235] (BioSource International, Inc., Camarillo, CA) described here, with sequence numbering by the manufacturer based on the major alternative splice transcript without the 54 bp in exon 10, would correspond to [pY1021] and [pY1248/1252/1253], respectively, as in the “wild-type” full-length cDNA sequence of the c-Met gene.
Signal Transduction via Immunoblotting. Cells were lysed in lysis buffer as described previously (20), with the cell lysates separated by 7.5% SDS-PAGE. Immunoblotting was done using standard procedures as described previously (22). A549 cells were prestarved in serum-free medium containing 0.5% bovine serum albumin for up to 72 hours before being stimulated with HGF (40 ng/mL) for the indicated time durations. The following antibodies were used in immunoblotting or immunohistochemistry staining: c-Met (C-12) from Santa Cruz Biotechnology (Santa Cruz, CA), phosphatidylinositol 3′-kinase (PI3K) from Upstate Cell Signaling (Waltham, MA), β-actin from Sigma-Aldrich, phospho-AKT (p-AKT) [S473] from Cell Signaling Technology (Beverly, MA), phospho-ERK1/2 [T202/Y204] from BioSource International, phospho-PDK-1 [S241] and phospho-mTOR [S2448] from Cell Signaling Technology, and phospho-S6K [T421/S424] from New England Biolabs (Beverly, MA). Phosphospecific c-Met antibodies against the phosphoepitopes [pY1003] and [pY1230/1234/1235] were obtained from BioSource International.
c-Met Gene Expression in Affymetrix Gene Chip Microarray Transcriptome Profile. The level of c-Met gene expression was analyzed using the gene expression data obtained on the cell lines as reported previously (25) using the Affymetrix HG-U133A gene chip. The c-Met gene expression for the H1993 cells was assessed using data from the Affymetrix HG-U95Av2 gene chip. The arrays were normalized for overall intensity for hybridization and scanning. Data for the SK-LU-1 and A549 cell lines, available on both U95 and U133 chips, were used as a basis for establishing the platform to platform comparison.
Immunofluorescence and Immunohistochemistry. A549 cells were grown on coated glass coverslips and prestarved in 0.5% bovine serum albumin containing medium. After treatment with or without HGF (40 ng/mL, 15 minutes), cells were fixed in ice-cold methanol/acetone (1:1) for 10 minutes and blocked in PBS containing 1% bovine serum albumin for 30 minutes. Primary antibody against phospho-Met (p-Met) [Y1230/1234/1235] was incubated at a 1:50 dilution in the blocking solution for 1 hour followed by three washes with PBS. Texas red–conjugated secondary antibody was incubated for an additional hour in blocking solution at 1:500 dilutions. After three washes with PBS, cells were mounted in fluorescent Vectashield mounting medium (Vector Laboratories, Inc., Burlingame, CA), and images were captured using a confocal microscope and saved as digital images.
For immunohistochemistry, we analyzed paraffin-embedded, formalin-fixed tissues from 32 patients with lung neoplasms diagnosed at the University of Chicago Medical Center (Chicago, IL) between 2001 and 2003 with an institutional review board–approved protocol. We examined the following subtypes of lung neoplasm: NSCLC (N = 23), including adenocarcinoma (n = 9), large cell carcinoma (n = 7), and squamous cell carcinoma (n = 7); lung carcinoid (N = 5); and small cell carcinoma, SCLC (N = 4).
Slides were deparaffinized in xylene and hydrated with alcohol before being placed in 3% H2O2/methanol blocking solution to quench endogenous peroxidase activity followed by subsequent antigen unmasking. To ascertain the specificity of the p-Met immunostain signals, comparison was also made with the immunohistochemical results using anti-phosphotyrosine antibody (4G10, Upstate Cell Signaling). Incubation with the primary antibodies was done with the following dilution: total c-Met 1:200 or phosphospecific c-Met antibodies 1:100. After TBS washing, the slides were incubated for 30 minutes at room temperature with goat anti-rabbit IgG conjugated to a horseradish peroxidase–labeled polymer (Envision+ System, DAKO, Carpinteria, CA) or to an alkaline phosphatase–labeled polymer (MACH3TM, Biocare Medical, Walnut Creek, CA). Reactions were developed with 3,3′-diaminobenzidine chromogen or Vulcan Red and counterstained with hematoxylin. Appropriate negative controls for the immunostaining were prepared by omitting the primary antibody step and substituting it with nonimmiune rabbit serum.
All of the slides were reviewed by two independent pathologists and immunoscoring was done. For each case, the most representative tumor section was examined at 100× magnification to identify the percentage of tumor cells with c-Met expression.
c-Met siRNA and Transfection. siRNAs targeting c-Met mRNA were obtained from Dharmacon, Inc. (Lafayette, CO) and used according to the manufacturer's instructions. Briefly, a pool of four c-Met-specific 21-nucleotide RNA oligonucleotides forming a 19-bp duplex core with 2-nucleotide 3′ overhangs were used in combination. siRNA duplexes were transiently transfected into the A549 cells using the Oligofectamine reagent according to the manufacturer's instructions (Invitrogen, Inc., Carlsbad, CA). Mock transfection was done in parallel using SignalSilence control siRNA (Cell Signaling Technology) as negative control.
Results
Functional Expression and Activation of c-Met in NSCLC Cell Lines. We examined the expression of c-Met receptor tyrosine kinase protein in nine NSCLC cell lines using standard immunoblotting with an anti-human polyclonal c-Met antibody (Fig. 1A,, top). There was expression of total c-Met in most of the cell lines, with the highest level seen in H1993 followed by H1838, SW900, H358, and SK-LU-1. Medium level of c-Met expression was identified in A549, H2170, and SW1573 cell lines. H661 did not have any significant expression of c-Met. Next, we examined the level of c-Met gene expression using the Affymetrix HG-U133A gene chip expression data obtained on the cell lines as reported previously (25). The c-Met gene expression for the H1993 cells was assessed using microarray data from the Affymetrix HG-U95Av2 gene chip. The arrays were normalized for overall intensity as described in Materials and Methods. A high degree of correlation was observed between gene expression assessed at the transcriptome level shown here (Fig. 1A,, bottom) and protein level assessed by immunoblot (Fig. 1A , top), with the notable exceptions of the SK-LU-1 and H661 cells.
c-Met is expressed and functional in NSCLC. A, expression of c-Met in NSCLC cell lines. Top, expression of the c-Met receptor protein was analyzed with immunoblotting. Whole cell lysates prepared from the nine NSCLC cell lines were immunoblotted using antihuman c-Met polyclonal antibody (C-12). 170-kDa protein band, precursor form of the glycosylated c-Met; 140-kDa protein band, biologically active transmembrane β-subunit of c-Met. The membrane was stripped and reprobed with anti-PI3K antibody as internal loading control. Bottom, expression of the c-Met mRNA was analyzed in gene chip microarray study. Affymetrix HG-U133A gene chip microarray was used for all the cell lines, except H1993. H1993 expression was assessed using the Affymetrix HG-U95Av2 gene chip and adjusted for comparison as described in Materials and Methods. A high degree of correlation is observed between gene expression assessed at the transcriptome level shown here and protein level assessed by immunoblot (top) with the notable exceptions of the SK-LU-1 and H661 cells. B, HGF stimulation kinetics of the lung adenocarcinoma cell line A549. Prestarved A549 cells were stimulated with HGF (40 ng/mL) at 37°C for the indicated time durations. Whole cell lysates were prepared at the various indicated time intervals, separated in 7.5% SDS-PAGE, and immunoblotted using anti-human phosphospecific antibodies against the p-Met at the phosphoepitopes [pY1230/1234/1235] (gel 3) and [pY1003] (gel 2). Immunoblots were also done using antibodies against phospho-ERK1/2 [T202/Y204] (gel 4), phospho-S6K [T421/S424] (gel 5), phospho-PDK-1 [S241] (gel 6), p-AKT [S473] (gel 7), and phospho-mTOR [S2448] (gel 8). Internal control was done with anti-total c-Met (C-12) antibody (gel 1) and anti-PI3K antibody (gel 9). C, HGF stimulated activated p-Met expression in NSCLC A549 cells. A549 cells were grown on coated glass coverslips. After treatment with (bottom) or without (top) HGF for 15 minutes, activation of p-Met was visualized using immunofluorescent study as described in Materials and Methods. Anti-p-Met [Y1230/1234/1235] antibody was used as the primary antibody followed by a Texas red–conjugated secondary antibody incubation. After three washes with PBS, cells were subsequently mounted in fluorescent mounting medium (Vectashield), and images were captured using a confocal microscope. Activated p-Met [Y1230/1234/1235] was dramatically induced by HGF in the A549 cells (+HGF, bottom right).
c-Met is expressed and functional in NSCLC. A, expression of c-Met in NSCLC cell lines. Top, expression of the c-Met receptor protein was analyzed with immunoblotting. Whole cell lysates prepared from the nine NSCLC cell lines were immunoblotted using antihuman c-Met polyclonal antibody (C-12). 170-kDa protein band, precursor form of the glycosylated c-Met; 140-kDa protein band, biologically active transmembrane β-subunit of c-Met. The membrane was stripped and reprobed with anti-PI3K antibody as internal loading control. Bottom, expression of the c-Met mRNA was analyzed in gene chip microarray study. Affymetrix HG-U133A gene chip microarray was used for all the cell lines, except H1993. H1993 expression was assessed using the Affymetrix HG-U95Av2 gene chip and adjusted for comparison as described in Materials and Methods. A high degree of correlation is observed between gene expression assessed at the transcriptome level shown here and protein level assessed by immunoblot (top) with the notable exceptions of the SK-LU-1 and H661 cells. B, HGF stimulation kinetics of the lung adenocarcinoma cell line A549. Prestarved A549 cells were stimulated with HGF (40 ng/mL) at 37°C for the indicated time durations. Whole cell lysates were prepared at the various indicated time intervals, separated in 7.5% SDS-PAGE, and immunoblotted using anti-human phosphospecific antibodies against the p-Met at the phosphoepitopes [pY1230/1234/1235] (gel 3) and [pY1003] (gel 2). Immunoblots were also done using antibodies against phospho-ERK1/2 [T202/Y204] (gel 4), phospho-S6K [T421/S424] (gel 5), phospho-PDK-1 [S241] (gel 6), p-AKT [S473] (gel 7), and phospho-mTOR [S2448] (gel 8). Internal control was done with anti-total c-Met (C-12) antibody (gel 1) and anti-PI3K antibody (gel 9). C, HGF stimulated activated p-Met expression in NSCLC A549 cells. A549 cells were grown on coated glass coverslips. After treatment with (bottom) or without (top) HGF for 15 minutes, activation of p-Met was visualized using immunofluorescent study as described in Materials and Methods. Anti-p-Met [Y1230/1234/1235] antibody was used as the primary antibody followed by a Texas red–conjugated secondary antibody incubation. After three washes with PBS, cells were subsequently mounted in fluorescent mounting medium (Vectashield), and images were captured using a confocal microscope. Activated p-Met [Y1230/1234/1235] was dramatically induced by HGF in the A549 cells (+HGF, bottom right).
The c-Met-expressing adenocarcinoma A549 cell line was used to study the signal transduction of the c-Met/HGF pathway. A549 cells were responsive to HGF (40 ng/mL) stimulation up to 60 minutes in a time-dependent fashion (Fig. 1B). Significant induction of tyrosine phosphorylation of the catalytic kinase domain autophosphorylation site [pY1230/1234/1235] as well as the regulatory juxtamembrane c-Cbl binding epitope [pY1003] was evident starting at 7.5 minutes of HGF stimulation, with substantial induction still evident up to 60 minutes. Moreover, phosphorylation of the signaling mediators of the PI3K/PDK-1/AKT/mTOR/S6K pathway was significantly induced after HGF stimulation (Fig. 1B). Specifically, HGF stimulation of c-Met induced phosphorylation of the [pY1003] and [pY1230/1234/1235] phosphoepitopes of Met itself. In addition, it also activated phospho-ERK1/2 [T202/Y204], phospho-S6K [T421/S424], phospho-PDK-1 [S241], p-AKT [S473], and phospho-mTOR [S2448] phosphoepitopes in a time-dependent manner. There was notably a slight decrease in the total c-Met level seen at 30 and 60 minutes after HGF induction, which may be a result of ligand-induced ubiquitination of c-Met mediated by c-Cbl binding to the c-Met juxtamembrane domain pY1003 (26, 27). HGF stimulation (40 ng/mL, 10 minutes) of A549 cells dramatically induced c-Met activation (as assessed by p-Met [Y1230/1234/1235]) visualized by immunofluorescence under fluorescence confocal microscopy (Fig. 1C).
Functional Expression of c-Met and Activated p-Met in NSCLC Tumor Tissues. We analyzed paraffin-embedded, formalin-fixed tissues from 32 patients with lung neoplasm. The subtypes of lung neoplasm examined were adenocarcinoma (n = 9), large cell carcinoma (n = 7), squamous carcinoma (n = 7), lung carcinoid tumor (n = 5), and small cell carcinoma (n = 4). Expression of both total c-Met and activated p-Met (i.e., p-Met [Y1003] and [Y1230/1234/1235]) was shown in paraffin-embedded, formalin-fixed lung neoplasm tumor tissues using standard immunohistochemical techniques (Figs. 2 and 3). There was expression of c-Met detected in all (100%) of the lung neoplasm tissues examined (n = 32), with representative digital images shown in Fig. 2A. We further classified immunohistochemical staining intensity and extent of c-Met using the three-scale scoring system: negative (0), weak (1+), and strong (2+) for further evaluation in the subtype analysis. Overall, 61% (14 of 23) of the NSCLC tumor tissues (n = 23) examined strongly expressed total c-Met, whereas 60% of lung carcinoid (n = 5) and 25% SCLC tumor tissues (n = 4) strongly expressed c-Met. Among NSCLC, strong c-Met expression (2+) was evident in 6 of 9 (67%) adenocarcinoma, 4 of 7 (57%) large cell, and 4 of 7 (57%) squamous cell carcinoma. We did not see any significant total c-Met staining in normal lung tissues (Fig. 2A).
Functional expression of c-Met and activated p-Met in lung neoplasm paraffin-embedded tumor tissues by immunohistochemistry. A, expression of total c-Met in lung neoplasm tumor tissues. Immunohistochemical staining for total c-Met using anti-human c-Met antibody was done as described in Materials and Methods on paraffin-embedded, formalin-fixed NSCLC tumor tissues, including adenocarcinoma, large cell, and squamous cell carcinoma. Lung carcinoid and SCLC tumor tissue staining was included for comparison (2+, strong immunostain intensity were shown). Digital pictures were taken at 20× magnification. Representative digital images of each of the tumor tissue types. Specific staining was evident with negative signal in the stromal tissues as well as in the normal lung control. B, analysis of the frequency of expression of activated p-Met expression by immunohistochemical staining in lung neoplasm tumor tissues. Expression of activated c-Met with tyrosine phosphorylation at the [pY1003] phosphoepitope and the [pY1230/1234/1235] catalytic autophosphorylation site was studied using the corresponding phosphospecific antibodies in immunohistochemical staining as described in Materials and Methods. Analysis of the frequency of expression of activated p-Met at both [pY1003] and [pY1230/1234/1235] sites seen among various lung cancer subtypes.
Functional expression of c-Met and activated p-Met in lung neoplasm paraffin-embedded tumor tissues by immunohistochemistry. A, expression of total c-Met in lung neoplasm tumor tissues. Immunohistochemical staining for total c-Met using anti-human c-Met antibody was done as described in Materials and Methods on paraffin-embedded, formalin-fixed NSCLC tumor tissues, including adenocarcinoma, large cell, and squamous cell carcinoma. Lung carcinoid and SCLC tumor tissue staining was included for comparison (2+, strong immunostain intensity were shown). Digital pictures were taken at 20× magnification. Representative digital images of each of the tumor tissue types. Specific staining was evident with negative signal in the stromal tissues as well as in the normal lung control. B, analysis of the frequency of expression of activated p-Met expression by immunohistochemical staining in lung neoplasm tumor tissues. Expression of activated c-Met with tyrosine phosphorylation at the [pY1003] phosphoepitope and the [pY1230/1234/1235] catalytic autophosphorylation site was studied using the corresponding phosphospecific antibodies in immunohistochemical staining as described in Materials and Methods. Analysis of the frequency of expression of activated p-Met at both [pY1003] and [pY1230/1234/1235] sites seen among various lung cancer subtypes.
Normalized level of expression of c-Met in genomics Affymetrix microarray in lung cancers. c-Met gene overexpression was examined using the genomics databases through the Oncomine bioinformatics resource as described in Materials and Methods. The c-Met mRNA was found to be expressed at significantly higher levels in lung adenocarcinoma (P = 0.0002; adjusted P = 0.0007), lung carcinoid (P = 0.0003; adjusted P = 0.0013), and SCLC (P = 0.0041; adjusted P = 0.0165) when compared with normal lung in the Bhattacharjee et al. microarray study using the human U95A oligonucleotide probe arrays (29). To assure a conservative estimate of statistical significance, adjusted Ps were calculated using the Bonferroni correction to correct for multiple hypothesis testing. *, adjusted P < 0.05, statistically significant.
Normalized level of expression of c-Met in genomics Affymetrix microarray in lung cancers. c-Met gene overexpression was examined using the genomics databases through the Oncomine bioinformatics resource as described in Materials and Methods. The c-Met mRNA was found to be expressed at significantly higher levels in lung adenocarcinoma (P = 0.0002; adjusted P = 0.0007), lung carcinoid (P = 0.0003; adjusted P = 0.0013), and SCLC (P = 0.0041; adjusted P = 0.0165) when compared with normal lung in the Bhattacharjee et al. microarray study using the human U95A oligonucleotide probe arrays (29). To assure a conservative estimate of statistical significance, adjusted Ps were calculated using the Bonferroni correction to correct for multiple hypothesis testing. *, adjusted P < 0.05, statistically significant.
Aside from total expression of c-Met in NSCLC tumors, we also analyzed its functional activity in terms of the expression of p-Met [Y1003] and [Y1230/1234/1235] by immunohistochemical staining using phosphospecific antibodies targeting the corresponding tyrosine phosphorylation sites of c-Met (Fig. 2B). Analysis of the frequency of detection by immunohistochemical staining as above showed that c-Met was activated in adenocarcinoma tissues sections at phosphoepitopes [pY1003] (44%) and [pY1230/1234/1235] (33%). Large cell lung cancer exhibited c-Met activation at phosphoepitopes [pY1003] (86%) and [pY1230/1234/1235] (57%). We found that 71% of squamous cell lung cancer tissues stained positively with p-Met [Y1003], with predominantly nuclear staining, whereas none was detected with p-Met [Y1230/1234/1235] antibodies. Forty percent of lung carcinoid tumors showed p-Met [Y1003] expression, also predominantly in the nucleus, but none stained positively with p-Met [Y1230/1234/1235]. For SCLC controls, 100% of the tumor sections examined expressed p-Met [Y1003] and 50% expressed p-Met [Y1230/1234/1235].
We also observed preferential expression of activated p-Met in the tumor cells in the invasive front of the NSCLC tumor tissues in both p-Met [Y1003] and [Y1230/1234/1235] (data not shown). It was also evident that the activated p-Met [Y1003] and [Y1230/1234/1235] were predominantly membranous in localization (data not shown).
The c-Met protein overexpression may correlate with the gene overexpression in NSCLC. We have determined the gene overexpression of c-Met through the Oncomine bioinformatics database resource, which provides an infrastructure of data mining tools to efficiently query genes and data sets of interest as well as meta-analyze groups of studies (28). The c-Met mRNA was found to be expressed at significantly higher levels in lung adenocarcinoma (P = 0.0002; adjusted P = 0.0007), lung carcinoid (P = 0.0003; adjusted P = 0.0013), and SCLC (P = 0.0041; adjusted P = 0.0165) when compared with normal lung in the microarray study using the human U95A oligonucleotide probe arrays in our previous data (ref. 29; Fig. 3). In squamous cell lung carcinoma, although there was higher relative expression of c-Met seen compared with normal lung, the difference was not statistically significant (P = 0.7926; adjusted P > 1; data not shown).
Novel c-Met Alterations in NSCLC. We evaluated 4 NSCLC cell lines and 127 lung adenocarcinoma tumor tissues (T1-T127) for the presence of c-Met mutations using standard DNA sequencing techniques as described previously (22). Based on our prior experience in the SCLC study (22), we chose to screen for c-Met mutations in lung adenocarcinoma tissues in the following functional domains: extracellular Sema domain, juxtamembrane domain, and cytoplasmic catalytic tyrosine kinase domain. Analysis of the adenocarcinoma tumor tissue c-Met cDNA sequencing results again showed sequence alterations clustering within the Sema and juxtamembrane domains but apparently absent from the tyrosine kinase domain (Table 1; Fig. 4). Within the Sema domain, we identified four novel missense changes, E168D (T74), L229F (T1), S323G (T100), and N375S (T63, T80, and T117). The N375S (c1124A>G) missense change was notably seen in a total of three different tumor tissue samples. Also identified were juxtamembrane domain mutations R988C in one NSCLC cell line (H1437) and S1058P in a tumor tissue (T123), and both a R988C and a T1010I mutation were detected in the tumor tissue (T9). Finally, there was an alternative splice variant with the 47-amino acid exon 14 (juxtamembrane domain) missing in-frame from the c-Met cDNA in the tumor specimen (T103). Most of the c-Met sequence changes that were identified in the lung adenocarcinoma tumor tissue cDNA were heterozygous. The adjacent, histologically normal lung counterparts from the surgically resected specimens where available were also sequenced for these mutations. A homozygous wild-type sequence was found in normal samples [N1 (with T1: L229F) and N103 (with T103: splice variant skipping entire exon 14)], suggesting somatic nature of the alterations identified in these tumors. In the remaining five c-Met cDNAs from adjacent, histologically normal tissues, the same variant nucleotide that had been detected in the tumor in the heterozygous state was similarly detected. Of note, the wild-type allele was the more predominant signal seen in the adjacent histologically normal tissues (N63 and N80). The juxtamembrane domain somatic mutations of c-Met R988C and T1010I were identified recently in SCLC by our group (22). R988C was found in H69 (SCLC) cell line as heterozygous and in H249 (SCLC) cell line as homozygous/hemizygous. T1010I-Met was identified in a SCLC tumor tissue to be heterozygous. None of the mutations we have identified are present in the National Center for Biotechnology Information Entrez single nucleotide polymorphism database search.
Mutations and sequence variants of c-Met identified in NSCLC
Mutations of c-Met in NSCLC . | . | . | . | . | . | |||||
---|---|---|---|---|---|---|---|---|---|---|
Tumor ID . | Nucleotide change . | Exon 2 (Sema domain) . | Exon 14/15 (juxtamembrane domain) . | Mutant genotype . | Adjacent “normal” (N) . | |||||
NSCLC tumor tissues (T1-T127) | ||||||||||
T1 | c687G>T (TTG>TTT) | L229F | Heterozygous | − | ||||||
T9 | c2692 C>T (CGC>TGC) + c3029C>T (ACT>ATT) | R988C + T10101 | Heterozygous | + Heterozygous | ||||||
T63 | c1124A>G (AAC>AGC) | N375S | Heterozygous | + Heterozygous* | ||||||
T74 | c504G>T (GAG>GAT) | E168D | Heterozygous | NA | ||||||
T80 | c1124A>G (AAC>AGC) | N375S | Heterozygous | + Heterozygous* | ||||||
T100 | c967A>G (AGC>GGC) | S323G | Heterozygous | + Heterozygous | ||||||
T103 | del 141-bp 2942-3082 | Splice variant (exon 14 skipped in-frame) | Homozygous | − | ||||||
T117 | c1124A>G (AAC>AGC) | N375S | Homozygous | NA | ||||||
T123 | c3172T>C (TCT>CCT) | S1058P | Heterozygous | + Heterozygous | ||||||
NSCLC cell lines | ||||||||||
A549 | Wild-type | |||||||||
H1395 | Wild-type | |||||||||
H1437 | c2692C>T (CGC>TGC) | R988C | Heterozygous | |||||||
H2087 | Wild-type |
Mutations of c-Met in NSCLC . | . | . | . | . | . | |||||
---|---|---|---|---|---|---|---|---|---|---|
Tumor ID . | Nucleotide change . | Exon 2 (Sema domain) . | Exon 14/15 (juxtamembrane domain) . | Mutant genotype . | Adjacent “normal” (N) . | |||||
NSCLC tumor tissues (T1-T127) | ||||||||||
T1 | c687G>T (TTG>TTT) | L229F | Heterozygous | − | ||||||
T9 | c2692 C>T (CGC>TGC) + c3029C>T (ACT>ATT) | R988C + T10101 | Heterozygous | + Heterozygous | ||||||
T63 | c1124A>G (AAC>AGC) | N375S | Heterozygous | + Heterozygous* | ||||||
T74 | c504G>T (GAG>GAT) | E168D | Heterozygous | NA | ||||||
T80 | c1124A>G (AAC>AGC) | N375S | Heterozygous | + Heterozygous* | ||||||
T100 | c967A>G (AGC>GGC) | S323G | Heterozygous | + Heterozygous | ||||||
T103 | del 141-bp 2942-3082 | Splice variant (exon 14 skipped in-frame) | Homozygous | − | ||||||
T117 | c1124A>G (AAC>AGC) | N375S | Homozygous | NA | ||||||
T123 | c3172T>C (TCT>CCT) | S1058P | Heterozygous | + Heterozygous | ||||||
NSCLC cell lines | ||||||||||
A549 | Wild-type | |||||||||
H1395 | Wild-type | |||||||||
H1437 | c2692C>T (CGC>TGC) | R988C | Heterozygous | |||||||
H2087 | Wild-type |
NOTE: Summary of the cDNA sequencing analysis of c-Met in 127 lung adenocarcinoma tumor tissues (T1-T127) and 4 NSCLC cell lines revealing novel sequence variants clustered within the Sema and juxtamembrane domains with missense mutations and alternative splice variant. No unambiguous mutations were identified within the tyrosine kinase domain. Nucleotide positions are shown in parentheses and are numbered relative to the first base of the translational initiation codon according to the full-length human c-Met cDNA (including the full-length exon 10; refs. 22, 24). Results of the sequence analysis of the adjacent “normal” lung tissues where available are also shown.
Abbreviations: −, wild-type sequence; +, mutant sequence variant; NA, not available; *, wild-type allele more predominant.
Unique mutations and sequence variants of c-Met identified in NSCLC. Unique mutations and sequence variants of c-Met in the NSCLC mutational analysis are illustrated schematically here in the context of the functional domains of c-Met. SEMA, Sema domain; PSI, plexins, semaphorins, and integrins domain; IPT Repeats 1-4, found in immunoglobulin-like regions, plexins, and transcriptional factors (also known as TIG domains); TM, transmembrane domain; Juxtamembrane, juxtamembrane domain encoded by exon 14/15; TK, catalytic tyrosine kinase domain. See text for details of the mutations and sequence variants.
Unique mutations and sequence variants of c-Met identified in NSCLC. Unique mutations and sequence variants of c-Met in the NSCLC mutational analysis are illustrated schematically here in the context of the functional domains of c-Met. SEMA, Sema domain; PSI, plexins, semaphorins, and integrins domain; IPT Repeats 1-4, found in immunoglobulin-like regions, plexins, and transcriptional factors (also known as TIG domains); TM, transmembrane domain; Juxtamembrane, juxtamembrane domain encoded by exon 14/15; TK, catalytic tyrosine kinase domain. See text for details of the mutations and sequence variants.
K-Ras mutational data are available in some of these adenocarcinoma tissues from previous analysis (23). In the available tumor tissues T63, T80, T100, and T103 where there are c-Met alterations identified, K-Ras was not mutated, whereas K-Ras was mutated in the tumor sample T1.
Inhibition of Cell Viability and c-Met/HGF Signaling in NSCLC by c-Met siRNA Gene Silencing. siRNA inhibitory strategy (30) was used to target against c-Met mRNA to test and validate the potential of therapeutic inhibition of c-Met in NSCLC. siRNA duplexes targeting against c-Met mRNA were transiently transfected into A549 cells, which express wild-type c-Met, to achieve c-Met gene silencing using Oligofectamine method. c-Met protein expression was successfully inhibited by 50% to 60% over 72 hours in the NSCLC A549 cells using c-Met-specific siRNA gene silencing technique (Fig. 5A). The HGF-induced activation of the p-Met [Y1003] and [Y1230/1234/1235] as well as p-AKT [S473] activation were all substantially inhibited by siRNA-Met. siRNA down-regulation of c-Met protein expression in A549 cells resulted in inhibition of the serum-stimulated cell growth and viability (16.3 ± 7.4% at 48 hours and 57.1 ± 7.2% at 72 hours, n = 3) as determined by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay.
Inhibition of c-Met in NSCLC. A, inhibition of c-Met protein expression via siRNA gene silencing targeting c-Met mRNA. NSCLC A549 cells with or without HGF stimulation were transfected with c-Met-specific siRNA duplexes or control siRNA (mock) as described in Materials and Methods and analyzed using immunoblots (IB) with the following antibodies: c-Met (gel 1), p-Met [Y1003] (gel 2), p-Met [Y1230/1234/1235] (gel 3), p-AKT [S473](gel 4), and β-actin as internal loading control (gel 5). Down-regulation of c-Met protein expression by c-Met-specific siRNA in NSCLC A549 cells is evident (gel 1). Note that there was also dramatic inhibition of HGF-induced p-Met [Y1003] [Y1230/1234/1235] and p-AKT expression by inhibiting c-Met using siRNA. Quantitation of the c-Met immunoblot (gel 1) was done using the Bio-Rad Chemidoc XRS gel imaging system (Bio-Rad Laboratories, Inc., Hercules, CA) for image analysis and quantitation (see text for details). HGF-stimulated mock-transfected A549 cells were standardized as the c-Met expression control (100%) and that of the c-Met-specific siRNA-transfected A549 cells was expressed as percentage of the mock after being normalized with the β-actin signal level. siRNA gene silencing of c-Met also resulted in inhibition of A549 cells viability by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay (see text for details). B, inhibition of c-Met with the specific small molecule c-Met inhibitor SU11274 in NSCLC cell lines. Cells from each of the nine NSCLC cell lines were plated on tissue culture plates and allowed to adhere overnight followed by pretreatment with the indicated concentrations of the drug SU11274. Cell viability was assayed at 72 hours after drug addition using trypan blue exclusion assay in at least duplicates experiments as described in Materials and Methods. C, SU11274 inhibited c-Met/HGF signaling as well as its downstream mTOR/S6K pathway. Prestarved A549 and H1993 cells were stimulated with or without HGF (40 ng/mL, 7.5 minutes) and in the absence or presence of SU11274 (5 μmol/L) as described in Materials and Methods. Whole cell lysates were prepared at the completion of the drug treatment and HGF stimulation, separated in 7.5% SDS-PAGE, and immunoblotted using anti-phosphospecific antibodies against the following: p-Met [Y1230/1234/1235] (gel 1), p-Met [Y1003] (gel 2), phospho-S6K [T421/S424] (gel 3), p-AKT [S473] (gel 4), and phospho-ERK1/2 [T202/Y204] (gel 5) phosphoepitopes. Internal loading control immunoblots using antibodies against total c-Met (gel 6) and β-actin (gel 7).
Inhibition of c-Met in NSCLC. A, inhibition of c-Met protein expression via siRNA gene silencing targeting c-Met mRNA. NSCLC A549 cells with or without HGF stimulation were transfected with c-Met-specific siRNA duplexes or control siRNA (mock) as described in Materials and Methods and analyzed using immunoblots (IB) with the following antibodies: c-Met (gel 1), p-Met [Y1003] (gel 2), p-Met [Y1230/1234/1235] (gel 3), p-AKT [S473](gel 4), and β-actin as internal loading control (gel 5). Down-regulation of c-Met protein expression by c-Met-specific siRNA in NSCLC A549 cells is evident (gel 1). Note that there was also dramatic inhibition of HGF-induced p-Met [Y1003] [Y1230/1234/1235] and p-AKT expression by inhibiting c-Met using siRNA. Quantitation of the c-Met immunoblot (gel 1) was done using the Bio-Rad Chemidoc XRS gel imaging system (Bio-Rad Laboratories, Inc., Hercules, CA) for image analysis and quantitation (see text for details). HGF-stimulated mock-transfected A549 cells were standardized as the c-Met expression control (100%) and that of the c-Met-specific siRNA-transfected A549 cells was expressed as percentage of the mock after being normalized with the β-actin signal level. siRNA gene silencing of c-Met also resulted in inhibition of A549 cells viability by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay (see text for details). B, inhibition of c-Met with the specific small molecule c-Met inhibitor SU11274 in NSCLC cell lines. Cells from each of the nine NSCLC cell lines were plated on tissue culture plates and allowed to adhere overnight followed by pretreatment with the indicated concentrations of the drug SU11274. Cell viability was assayed at 72 hours after drug addition using trypan blue exclusion assay in at least duplicates experiments as described in Materials and Methods. C, SU11274 inhibited c-Met/HGF signaling as well as its downstream mTOR/S6K pathway. Prestarved A549 and H1993 cells were stimulated with or without HGF (40 ng/mL, 7.5 minutes) and in the absence or presence of SU11274 (5 μmol/L) as described in Materials and Methods. Whole cell lysates were prepared at the completion of the drug treatment and HGF stimulation, separated in 7.5% SDS-PAGE, and immunoblotted using anti-phosphospecific antibodies against the following: p-Met [Y1230/1234/1235] (gel 1), p-Met [Y1003] (gel 2), phospho-S6K [T421/S424] (gel 3), p-AKT [S473] (gel 4), and phospho-ERK1/2 [T202/Y204] (gel 5) phosphoepitopes. Internal loading control immunoblots using antibodies against total c-Met (gel 6) and β-actin (gel 7).
SU11274 Inhibits Cell Viability and c-Met/HGF Signaling of NSCLC Cells. SU11274 is a highly specific inhibitor of c-Met with in vitro IC50 for the Met enzyme at 0.012 μmol/L (31). We used SU11274 to test for its inhibitory effects on various NSCLC cell lines (Fig. 5B). It was evident that the c-Met-selective inhibitor was effective against all but one of the nine cell lines tested, with IC50 ranging from ∼0.8 to 4.4 μmol/L. The H661 cell line, negative for c-Met expression, showed negligible inhibition of cell growth and viability by SU11274 up to 5 μmol/L.
Using SU11274, we have further defined its inhibitory effects on the c-Met/HGF signaling pathway in the two NSCLC adenocarcinoma cell lines (A549 and H1993) that overexpress c-Met (Fig. 5C). HGF-induced tyrosine phosphorylation of cellular proteins in both cell lines was inhibited by SU11274 (5 μmol/L; data not shown). The baseline and HGF-induced autophosphorylation of c-Met at the activation loop site phosphoepitope [pY1230/1234/1235] was abrogated by SU11274. Similarly, tyrosine phosphorylation at the phosphoepitope [pY1003] was also blocked by SU11274. In addition, specific HGF-induced phosphorylation of phospho-S6K [T421/S424], p-AKT [S473], and phospho-ERK1/2 [T202/Y204] was also inhibited by SU11274 in both NSCLC cell lines (Fig. 5C).
Discussion
In this study, we show that c-Met was functionally expressed in NSCLC cell lines and tumor tissues. We show that, in tumor tissues, phosphorylated activated forms of c-Met were detectable with unique preferential expression at the tumor invasive front. Unique c-Met alterations in the Sema domain and the juxtamembrane domain in NSCLC were identified. Finally, targeted inhibition of c-Met either via specific siRNA or by using the small molecule c-Met inhibitor SU11274 leads to decreased NSCLC cell growth and viability. SU11274 inhibition of c-Met also abrogates tyrosine phosphorylation of cellular proteins, including c-Met itself as well as its downstream signaling.
c-Met activation (as assessed by phosphorylation) can be a result of overexpression, gene amplification, or activating mutations (13). Expression of c-Met has been widely investigated in several solid tumors, including head and neck cancer (32, 33), esophageal cancer (34), breast cancer (35), renal cell cancer (16), and prostate cancer (36). In this study, we show that there was overexpression of c-Met in the majority of NSCLC cell lines examined, and expression of c-Met was also seen in 100% of NSCLC tumor tissues sections, with 61% of the NSCLC tumor tissues showing strong overexpression. Although c-Met overexpression has been seen in a variety of solid tumors, ours is the first study to show overexpression of phosphorylated forms of c-Met in tumor tissues. It would be useful now to compare the potential phosphospecific c-Met expression in frozen versus paraffin-embedded tumor tissues.
Recently, Pennacchietti et al. (37) showed that the levels of c-Met mRNA and protein were augmented by the hypoxia-inducible factor proteins sensing hypoxia within the tumors. This provides a mechanism for c-Met overexpression in tumors as a molecular response to intratumoral oxygen deprivation. This likely translates into a hypoxia-induced “invasive switch” accompanying an increase in an angiogenesis-independent malignant phenotype. In our c-Met immunohistochemical staining of the lung tumor tissues, we observed a pattern of preferential c-Met activation with induction of c-Met phosphoepitopes [pY1003] and [pY1230/1234/1235] along the invasive fronts of the NSCLC tumor tissues (also evident in SCLC; data not shown). Taken together, targeting c-Met/HGF signaling in combination with antivascular or antiangiogenic therapies, such as bevacizumab (38), might be more efficacious than bevacizumab alone in abrogating tumor invasion and metastasis. Interestingly, for squamous cell NSCLC and carcinoids, the p-Met expression was observed predominantly as nuclear staining. Although the precise mechanism of the nuclear localization of p-Met is not fully understood yet, this may implicate a different biology in these tumor types compared with other NSCLC.
We have identified several novel missense changes and an alternative splice product of c-Met among the lung adenocarcinoma tissues involving the Sema domain (E168D, L229F, S323G, and N375S) and the juxtamembrane domain (R988C + T1010I, S1058P, and an alternative splice variant skipping the entire 47-amino acid exon 14 in-frame). In addition, there was also a R988C juxtamembrane mutation identified in the NSCLC cell line H1437. Although several of the identified heterozygous mutations of c-Met in tumor were also reflected in the adjacent histologically normal counterpart lung tissues available for analysis, molecular contamination by the tumor cell genome itself in the specimen is possible because the “normal” was identified and separated only histopathologically from the surgically resected tumor tissues. Because c-Met is expressed at a very low level in normal lung tissue and because the sequence analysis was done on cDNA, it is possible that the c-Met cDNA produced from the normal tissue was actually from histologically undetected tumor cells, with elevated c-Met expression, which were present in the normal tissue. Alternative potential considerations include germ-line mutations or polymorphisms. It is also possible that the “normal” adjacent tissue is in fact already mutated. For example, it is possible that the adjacent “normal” lung parenchyma is altered in the context of “field defect” from cigarette smoking (such as chronic obstructive pulmonary disease), and the alteration of c-Met occurs early in the pathogenesis of lung cancer. In the future, analysis of paired normal/tumor tissues for mutations would be best done using laser capture microdissection to avoid molecular contamination by cancer cells. The interpretation that the nucleotide changes are tumor specific is supported by our observation of the wild-type allele being the more predominant signal seen in the adjacent “normal” tissues for both T63 and T80. The possibilities above can also potentially be resolved by sequencing the c-Met from genomic DNA obtained through peripheral blood lymphocytes for germ-line mutational determination. Recently, we have reported unique Sema missense mutation (E168D) and juxtamembrane mutations (R988C and T1010I) of c-Met in SCLC (22). The functional consequences of the two novel juxtamembrane missense mutations (R988C and T1010I) were also identified. They were found to be activating mutations regulating various SCLC cell biology, including cellular transformation, anchorage-dependent proliferation, cytoskeletal functions, cell motility, and migration. Moreover, the two juxtamembrane mutations also enhanced tyrosine phosphorylation of cellular proteins, including the key cytoskeletal focal adhesion protein paxillin at tyrosine residue [pY31] (the first CRKL binding site; refs. 22, 39). Here, we have identified other Sema domain and juxtamembrane domain alterations of c-Met in the NSCLC adenocarcinoma tumors. The biological effects of these novel c-Met mutations and sequence variants are currently being investigated.
The majority of the activating mutations of c-Met reported previously are located within the catalytic tyrosine kinase domain (21, 40); however, we did not find any unambiguous tyrosine kinase domain mutations in adenocarcinoma NSCLC or in SCLC (22). When compared with primary human sporadic cancers, activating somatic mutations of c-Met were more frequently identified in human carcinoma metastases, conferring a motile-invasive tumor cell phenotype (24). These studies support the hypothesis that the activated c-Met oncoprotein leads to tumor progression of primary carcinomas to metastasis. We present here the first immunohistochemical evidence that c-Met is activated preferentially at the NSCLC tumor invasive front, further reinforcing the key role of c-Met in tumor invasion and progression. Interestingly, our mutational analysis of c-Met suggests a novel emerging paradigm of the importance of the “hotspot clustering” of activating mutations within the HGF binding Sema domain and the regulatory juxtamembrane domain in lung cancer biology. The mutations were investigated in early-stage adenocarcinomas, and it would be useful to eventually compare the stage of disease with potential mutations.
Although the biological significance of mutations in the Sema domain is unclear at present and that of the juxtamembrane domain is just beginning to emerge, they represent plausible mechanisms of activating the c-Met signaling pathway promoting the invasive signaling program in lung cancer cells. Of particular interest, the extracellular Sema domain mutations are the first to be reported for c-Met. The three-dimensional structure of the c-Met Sema domain has recently been determined to adopt a seven-blade β-propeller structure as seen in a wide array of other biological molecules of little sequence homologies (41–43). In addition, it is now known that the Sema domain is necessary for the c-Met receptor binding to HGF, dimerization, and activation (44). Sema domain mutations may therefore potentially have altering effects on the receptor dimerization or on the ligand binding properties of c-Met to HGF. The study of the paradigm of structure-function interrelation in the context of various c-Met mutations is now feasible. It would also be useful for the future to determine the potential inhibition of the various c-Met mutations using both c-Met-specific antibodies and small molecule inhibitors, such as SU11274.
Down-regulation of c-Met via either siRNA or adenovirus expressing c-Met ribozyme represents powerful inhibition strategies to further validate the therapeutic role of c-Met in lung cancer (13, 45). However, the novel siRNA inhibitory strategy used against c-Met has not been reported in lung cancer. Here, we successfully showed that down-regulation of c-Met protein expression by siRNA gene silencing caused significant inhibition of NSCLC cell viability. This inhibitory strategy should be very applicable also to validate the role of c-Met as therapeutic target in other thoracic malignancies, such as SCLC, mesothelioma, and head and neck squamous cell carcinoma.
Small molecule inhibitors specifically targeting against c-Met represent an attractive novel targeted therapeutic approach. We have reported recently for the first time the effectiveness of a novel small molecule–specific inhibitor of c-Met (SU11274) in cells transformed by the oncogenic Tpr-Met as a model as well as in SCLC (31). Inhibition of the Met kinase activity by the drug SU11274 led to time- and dose-dependent reduced cell growth and induced G1 cell cycle arrest and apoptosis. Met kinase autophosphorylation was reduced on sites that have been shown previously to be important for activation of pathways involved in cell growth and survival, especially the PI3K and the Ras pathway. PI3K/AKT/mTOR is a crucial pathway downstream of c-Met that can regulate many of the biological phenomena, such as cell proliferation and survival, motility and migration, and tumor cell invasion into the stroma. Here, we further show that SU11274 inhibited cell viability, c-Met/HGF and its downstream cell proliferation, and survival pathway signaling in the c-Met-expressing NSCLC cells. It would now be useful to study c-Met inhibitors against NSCLC in mouse models, such as the orthotopic lung model (46).
In summary, we have identified unique functional expression and activation of c-Met in NSCLC cell lines and tumor tissues. The key role of activated p-Met in NSCLC tumor progression and invasion is supported by our immunohistochemical evidence of its preferential expression along the tumor invasive front. In addition, inhibition with siRNA and small molecule c-Met inhibitor was specific to NSCLC cell growth and viability as well as signal transduction. With the exciting finding of novel sequence alterations of c-Met in NSCLC, c-Met is further implicated to be an attractive molecular target for NSCLC inhibition. The underlying biology of these c-Met mutations and sequence variants must be better understood with further studies. It would be important to examine the efficacy of c-Met inhibitors in the context of the various c-Met mutations. Finally, with c-Met as a potential therapeutic target for NSCLC, it would be useful to design appropriate clinical trials to test compounds, such as the small molecule c-Met inhibitors.
Acknowledgments
Grant support: AACR-AstraZeneca-Cancer Research and Prevention Foundation Fellowship in Translational Lung Cancer Research, American Cancer Society Institutional Research Grant ACS IRG-58-004-44, and American Cancer Society (Illinois Division)-LUNGevity Foundation Lung Cancer Treatment Research Award (P.C. Ma) and American Cancer Society Research Scholar Grant RSG-02-244-02-CCE and Institutional Cancer Research Awards from the University of Chicago Cancer Center with the American Cancer Society and the V-Foundation (R. Salgia).
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.