Modulation of c-Met signaling and cellular sensitivity to radiation

Potential implications for therapy

Authors

  • Vikas Bhardwaj PhD,

    1. Department of Radiation Oncology, The University of Texas MD Anderson Cancer Center, Houston, Texas
    Search for more papers by this author
  • Tina Cascone MD,

    1. Department of Thoracic/Head and Neck Medical Oncology, The University of Texas MD Anderson Cancer Center, Houston, Texas
    Search for more papers by this author
  • Maria Angelica Cortez PhD,

    1. Department of Radiation Oncology, The University of Texas MD Anderson Cancer Center, Houston, Texas
    Search for more papers by this author
  • Arya Amini MD,

    1. Department of Radiation Oncology, The University of Texas MD Anderson Cancer Center, Houston, Texas
    2. University of California at Irvine Medical Center, Orange, California
    Search for more papers by this author
  • Jaden Evans BS,

    1. Department of Radiation Oncology, The University of Texas MD Anderson Cancer Center, Houston, Texas
    2. Texas Tech University Health Sciences Center, Lubbock, Texas
    Search for more papers by this author
  • Ritsuko U. Komaki MD,

    1. Department of Radiation Oncology, The University of Texas MD Anderson Cancer Center, Houston, Texas
    Search for more papers by this author
  • John V. Heymach MD,

    1. Department of Thoracic/Head and Neck Medical Oncology, The University of Texas MD Anderson Cancer Center, Houston, Texas
    Search for more papers by this author
  • James W. Welsh MD

    Corresponding author
    1. Department of Radiation Oncology, The University of Texas MD Anderson Cancer Center, Houston, Texas
    • Department of Radiation Oncology, Unit 97, The University of Texas MD Anderson Cancer Center, 1515 Holcombe Blvd, Houston, TX

    Search for more papers by this author

Abstract

The c-Met/hepatocyte growth factor receptor and its family members are known to promote cancer cell migration and invasion. Signaling within and beyond this pathway contributes to the systemic spread of metastases through induction of the epithelial-mesenchymal transition, a process also implicated in mediating resistance to current anticancer therapies, including radiation. Induction of c-Met has also been observed after irradiation, suggesting that c-Met participates in radiation-induced disease progression through the epithelial-mesenchymal transition. Therefore, c-Met inhibition is an attractive target for potentially mitigating radiation resistance. This article summarizes key findings regarding crosstalk between radiotherapy and c-Met and discusses studies performed to date in which c-Met inhibition was used as a strategy to increase cellular radiosensitivity. Cancer 2013. © 2013 American Cancer Society.

INTRODUCTION

c-Met, a receptor tyrosine kinase, has important roles in the malignant transformation of cancer cells. Also known as hepatocyte growth factor (HGF) receptor, c-Met is a 170-kilodalton (kD) transmembrane protein that is activated by the binding of its ligand, HGF, to its extracellular region. At the time of ligand binding, c-Met activates downstream signaling pathways that have been implicated in the invasion and migration of cancer cells. Several investigators have shown links between c-Met signaling and cellular response to radiotherapy.

Radiotherapy, alone or in combination with chemotherapy, is the foundation of treatment for various solid tumors. However, because of the proximity of critical normal tissues, tumoricidal radiation doses often cannot be used. Elevation of c-Met has been noted in various tumors and is known to contribute to treatment resistance. Indeed, targeting c-Met has shown promise for the treatment of various types of cancer and is currently being tested in phase 3 trials for non-small cell lung cancer (NSCLC). In the current study, we review the current literature regarding c-Met and the potential use of inhibitors of the HGF/c-Met axis to enhance the sensitivity of cancers to radiation.

c-Met Signaling

c-Met, a transmembrane receptor tyrosine kinase, consists of a 45-kD α and a 150-kD β subunit.1 HGF, the only known ligand, binds to the extracellular region, leading to receptor dimerization and phosphorylation of intracellular tyrosine kinase domains.2 The activation/phosphorylation of the intracellular domain prompts activation of downstream signaling, mainly through the phosphatidylinositide 3-kinase (PI3K)/Akt, mitogen-activated protein kinase (MAPK), and signal transducers and activators of transcription (STAT) pathways (Fig. 1).1, 3, 4 Activation of c-Met has been implicated in cell scattering, in which cells lose their contact with the surrounding cells and extracellular matrix, attain a mesenchymal phenotype, invade surrounding tissues, and proliferate. Cell scattering is crucial for wound healing, in which cells migrate to sites of injury for tissue repair and reconstruction. Migration and invasion are also important in cancer progression, in which epithelial cancer cells lose their cell-cell contacts and acquire a mesenchymal phenotype5, 6 that circumvents anoikis and allows the cells to invade through extracellular structures until they ultimately lodge at secondary sites.

Figure 1.

Key components of c-Met signaling are shown. Activation of c-Met signaling triggers downstream responses in the Ras-Raf-MAPK (mitogen-activated protein kinase), phosphatidylinositide 3-kinase (PI3K)-Akt, and signal transducers and activators of transcription (STAT) pathways, ultimately enhancing cellular motility and invasion as well as angiogenesis and protection against apoptosis. After irradiation, activation of c-Met is increased by both hepatocyte growth factor (HGF)-dependent and HGF-independent mechanisms. Irradiation also increases the transcription of both HGF and c-Met. HGFR, hepatocyte growth factor receptor; Gab1, GRB2-associated binding protein 1.

HGF/c-Met Expression in Cancer

HGF is produced by cancer cells7 and stromal cells, including cancer-associated fibroblasts.8 The c-Met receptor is expressed mainly by epithelial cells,9 but also by other cell types, including vascular and lymphatic endothelial cells,10 hepatocytes,11 and hematopoietic cells.12 The HGF/c-Met axis is dysregulated in many human neoplasms. Ectopic expression of the ligand, its receptor, or both in human and mouse cells reportedly leads to tumorigenesis and promotes metastasis.13, 14 Activation of c-Met can also occur in a ligand-dependent autocrine manner.15, 16 Other potential mechanisms for dysregulation of the HGF/c-Met axis include gene amplification, receptor overexpression, and activating mutations.17 The activation/upregulation of both HGF and its receptor is a negative prognostic indicator in various types of cancer.18-21 Raghav et al. reported high levels of c-Met and phosphorylated Met in several molecular subtypes of breast cancer.22 High levels of total and activated receptor have been correlated with poor prognosis in terms of recurrence-free and overall survival rates. In patients with gliomas, expression of HGF and c-Met correlates with tumor grade23, 24 and elevated coexpression is correlated with recurrence of meningioma.25 A comparison of unselected patients with colon carcinoma revealed significant differences in mRNA and protein levels of HGF and c-Met in tumor versus normal mucosa. Overexpression of HGF and c-Met mRNA were associated with lymph node metastasis and disease stage.21 c-Met overexpression also was found to be correlated with a shorter median progression-free survival (PFS) and overall survival.26 HGF/c-Met expression has been implicated in the resistance of colorectal cancer cell lines to the epidermal growth factor receptor (EGFR) inhibitor cetuximab.27

Overexpression of c-Met mRNA has also been found in both NSCLC and small cell lung cancer.28, 29 In patients with NSCLC, c-Met activation appears to be associated with increased tumor differentiation and an overall worse prognosis.30, 31 Cappuzzo et al. found that patients with NSCLC and amplified Met levels had shorter survival times after surgical resection than patients without Met amplification.32 Another group found significant correlations between high circulating levels of c-Met in patients with NSCLC and nodal disease status and early disease recurrence.28

The epithelial-mesenchymal transition (EMT) in relation to HGF/c-Met and radiation

HGF or scatter factor was first identified as a cytokine that can dissociate a colony of cells into individual cells.33 Early studies found that HGF also increases cellular migration and invasion33, 34 via the urokinase plasminogen activator system.34 HGF/c-Met activation also induces EMT and is therefore important in embryogenesis and organ regeneration. Expression of c-Met was found to be increased in the epithelial cells of the developing mouse, whereas the surrounding mesenchymal cells had high HGF expression.35, 36 EMT promotes cancer progression via upregulating cancer cell migration, invasion, and, ultimately, angiogenesis. Activation of the HGF/c-Met axis is known to promote invasive growth in both cell lines and transgenic animal models of various types of cancer.13, 37, 38 In patients with colorectal cancer, c-Met expression can be induced by activation of the Wnt-β-catenin pathway.39 Hypoxia also promotes the invasive growth of cancer cells40; increases in the expression of hypoxia-inducible factor (HIF)-1α (an oxygen sensor that is stabilized in hypoxic environments) have been associated with increased c-Met expression and HIF-1α was inhibited by small interfering RNA (siRNA) to c-Met.41 Because both Wnt signaling and hypoxia induce the invasive phenotype, these findings further implicate c-Met in promoting invasion. Jahn et al. recently demonstrated that acquisition of EMT characteristics in an in vivo model was correlated with upregulation of c-Met mRNA and increased responsiveness to HGF.42

Radiotherapy is an integral component of the treatment of many solid tumors, and improvements in treatment planning and delivery have led to improvements in local control and reductions in toxicity. However, systemic dissemination of disease continues to be a challenge in many types of tumors. As noted above, the EMT contributes to tumor progression and metastasis.43, 44 Cancer therapies such as radiotherapy have been shown to contribute to the elevation of tumor growth factor-β, a known inducer of EMT,45 which may lead to the development of treatment resistance. Breast cancer cells treated with ≥ 20 Gy begin to display changes consistent with the EMT.46 Additional studies have also confirmed that sublethal doses of radiation prompt the induction of EMT in various cancer cell lines.45 For example, irradiated colorectal cancer cells undergo changes characteristic of EMT,47 and patients with rectal cancer show increased levels of mesenchymal markers such as vimentin and fibronectin after chemoradiation therapy.47 Thus far, the clinical assumption that a radiation-induced shift to the mesenchymal phenotype would facilitate metastasis has not been borne out. Although both preoperative and postoperative radiation can suppress local recurrence, the appearance of metastases does not seem to differ between patients given preoperative versus postoperative radiation for rectal cancer48 or soft tissue sarcoma.49 Clearly additional research is needed to clarify how radiation-induced EMT affects the biological behavior of tumors in patients with cancer.

c-Met Signaling in Angiogenesis

Angiogenesis and lymphangiogenesis are critical processes in tumor development and metastasis. Activation of c-Met signaling stimulates several cellular processes, including morphogenesis, motility, tumor progression, proliferation, survival pathways, and angiogenesis.10, 50 Studies have shown that c-Met can promote tumor angiogenesis in cell lines and in preclinical models.51 The vascular endothelial growth factor/receptor (VEGF/VEGFR) pathway is a key mediator of tumor angiogenesis. HGF/c-Met signaling can increase the expression of angiogenic mediators, including VEGF/VEGFR family members, activating survival pathways, proliferation, and migration of vascular endothelial cells. HGF can upregulate proangiogenic factors (VEGF) and downregulate the expression of the natural antiangiogenic protein thrombospondin-1, thereby functioning as a regulator of the angiogenic switch.52 A vast body of evidence indicates that both the HGF and VEGF pathways cooperate in inducing angiogenesis in vitro and in vivo. c-Met and VEGFR can synergistically activate common signaling downstream molecules, including extracelluar signal-regulated kinase (ERK)/MAPK, AKT, and focal adhesion kinase (FAK).53 Similar to VEGF, the expression of both c-Met and HGF is induced by HIF-1α, suggesting a crucial contributory role for this axis in promoting angiogenesis in microenvironments possessing low oxygen tension, such as tumors.41

c-Met Signaling in DNA Damage and Radiation Response

A growing body of evidence has suggested that c-Met activation is also important in imparting cellular resistance to DNA-damaging agents, including ionizing radiation.54 Fan et al. demonstrated that pretreating breast cancer cells with HGF protected them from DNA fragmentation induced by DNA-damaging agents. They further found that this HGF-induced protection depended on both dose and time and could be reversed by the HGF antagonist neurokinin-1 (NK1), a truncated form of the HGF protein.55 That same group subsequently showed that PI3K-Akt signaling is important in the way HGF protects cells from DNA damage and suggested a signaling flow of HGF→c-Met→PI3K→Akt→DNA repair.55 The mechanism behind HGF-induced prevention of DNA damage was suggested to be upregulation of polycystic kidney disease-1 (PDK1, a survival-promoting component of cadherin-catenin complexes) and downregulation of 51C (an inositol polyphosphate 5-phosphatase), TOPBP1 (a topoisomerase IIB-binding protein), and doxorubicin-induced Gu protein (which participates in RNA synthesis and processing).56 Further support for the role of c-Met in DNA repair came from a study by Medova et al.,57 in which inhibition of c-Met with the small-molecule inhibitor PHA665752 and with siRNA in cell lines with abnormal c-Met signaling inhibited DNA repair by homologous recombination. Treatment with PHA665752 also was found to prevent the formation of the BRCA1-RAD51 complex involved in homologous recombination-mediated DNA repair, presumably by preventing radiation-induced accumulation of RAD51 in the nucleus.57 In addition, HGF has been shown to inhibit γ radiation-induced apoptosis in nontumor models such as human umbilical vein endothelial cell cultures.58

Clinically, c-Met expression was found to be an independent and significant predictor of impaired local failure-free survival among patients undergoing definitive radiation for squamous cell carcinoma of the oropharynx,59 and HGF expression was found to be inversely correlated with failure-free survival.59 Immunohistochemical analysis of nasopharyngeal carcinoma specimens from patients treated with radiotherapy found c-Met to be a poor prognostic marker, with 5-year overall survival rates of 84% to 48% for specimens with low versus high c-Met expression.60

Irradiation Upregulates c-Met Signaling

Similar to EGFR, c-Met levels are also upregulated in irradiated cancer cells.61 De Bacco et al. found that irradiation increased the levels of c-Met in biphasic fashion with respect to both radiation dose and time.62 Increases in c-Met promoter activity and the phosphorylation of c-Met with subsequent activation of the c-Met downstream signaling molecules GRB2-associated binding protein 1 (Gab1) and MAPK, without increases in HGF levels of the cells, after irradiation led the authors to suggest that activation of c-Met signaling at the time of irradiation is not ligand (HGF)-dependent.62 Qian et al.63 also demonstrated that irradiation increased c-Met levels in a panel of pancreatic cancer cells; however, the increase in c-Met was dose-dependent. Irradiation also was found to increase the activation of c-Met and increased the migration and invasion of pancreatic cancer cells in the presence of HGF.63

Although the studies noted above did not show stimulation of HGF secretion at the time of radiation, Chu et al. found that irradiation enhanced the secretion of HGF by glioblastoma cells.64 They found that the glioblastoma cells with the highest basal levels of HGF were the most radioresistant, leading them to suggest that the radioresistance may have been related to the higher basal HGF levels. Similarly, Schweigerer et al.65 found that irradiated neuroblastoma cells expressed higher levels of HGF mRNA than unirradiated cells. It is interesting to note that irradiation increased the invasiveness of neuroblastoma cells with high basal c-Met expression but not that of cells with low basal c-Met expression.65

Collectively, these findings implicate c-Met in radiation-induced invasion and suggest that c-Met-targeting agents might be able to overcome radiation-induced EMT and potentially sensitize cancer cells to radiation. The known involvement of c-Met in cancer initiation and progression has led to its being identified as a target for therapy. Several groups have tried to identify which cancer cell populations can be specifically targeted with anti-c-Met agents.61, 66, 67 The effect of c-Met inhibition on various aspects of cancer progression such as cellular survival, EMT, invasion, migration, and angiogenesis has been well studied.41, 68 The importance of c-Met in DNA repair has led to its being used as a target for sensitizing cancer cells to radiation as well. Inhibition of c-Met by using small-molecule inhibitors and siRNA can radiosensitize cancer cells both in vitro and in vivo. We found that glioma cells can be radiosensitized by MP470, a small-molecule inhibitor of c-Met, and that combining radiation with MP470 significantly increased the percentage of apoptotic cells. This combination affected the radiation-induced DNA damage response by increasing the γ-H2Ax foci and reducing Rad51 levels.69 In gastric carcinoma cells, radiosensitization by c-Met inhibition involved an increase in γ-H2Ax levels and phosphorylation of ataxia telangiectasia mutated (ATM). c-Met inhibition, alone and in combination with radiation, reduced the phosphorylation of ataxia-telangiectasia-related (ATR) and checkpoint kinase 1 (Chk1) and reduced radiation-induced S-phase arrest.70 Other studies in which HGF/c-met inhibition was used to radiosensitize cancer cells are mentioned in Table 1.71-75

Table 1. Studies Targeting the HGF/c-Met Axis in Combination With Radiation
StudyCancer TypeObservation
  1. Abbreviations: HGF, hepatocyte growth factor; NSCLC, non-small cell lung cancer.

Yu 201271ProstateSU 11274 enhances radiation response of Du145 cells
Li 201272NSCLCHigher c-Met-expressing cells are radiosensitized by c-Met inhibition (AMG 458)
Lin 201073Thyroid cancerAutophagy activation radiosensitizes thyroid cancer cells by Met dephosphorylation
Buchanan 201174GlioblastomaAMG 102 radiosensitizes glioblastoma in vitro and in vivo (tumor growth delay)
Lal 200575GlioblastomaU1/ribozyme-mediated knockdown of HGF and c-Met sensitizes glioblastoma

Clinical Targeting of HGF/c-Met Signaling

Several therapeutic agents that target HGF/c-Met signaling are currently available. These agents act by blocking either the c-Met receptor or the ligand HGF and include both small-molecule inhibitors and monoclonal antibodies and are being tested in a variety of solid tumors. A recent review by Peters and Adjei mentions several agents that are currently in clinical trials for their activity against the HGF/c-Met signaling axis.76 Some of the agents mentioned in the review include anti-HGF antibodies such as ficlatuzumab, rilotumumab, and TAK-701; anti-Met antibodies such as onartuzumab; and anti-Met tyrosine kinase inhibitors such as tivantinib, foretinib, cabozantinib, and others. For NSCLC, the only therapeutic agent targeting c-Met that has been approved by the US Food and Drug Administration to date is crizotinib, for ALK-positive NSCLC.76 The following paragraphs summarize the status of some other c-Met–targeted therapeutics currently being developed.

Tivantinib (ARQ197) is a selective, small-molecule c-Met blocker that is not competitive with adenosine triphosphate (ATP). Given the crosstalk between c-Met and EGFR, and the potential for c-Met inhibition to overcome resistance to EGFR inhibitors, this agent is being studied in combination with erlotinib. In one randomized phase 2 trial, patients with refractory, stage IV NSCLC were randomly assigned to receive erlotinib only or erlotinib plus tivantinib; the combination treatment resulted in longer progression-free survival (PFS) time (16.1 weeks vs 9.7 weeks for erlotinib alone; hazard ratio [HR], 0.81 [95% confidence interval (95% CI), 0.57-1.15]) (P = .23). A planned multivariable Cox regression model adjusting for prognostic factors (including histology and genotype) yielded an HR for PFS of 0.68 (95% CI, 0.47-0.98 [P < .05]).77, 78 The improvement in PFS was particularly prominent among patients with tumors of nonsquamous histology, wild-type EGFR, and mutated KRAS. These promising findings have led to a phase 3 trial, which is currently underway.

Another small-molecule inhibitor that had shown promise is cabozantinib (XL184), which targets both c-Met kinase and VEGF. Because the development of resistance to anti-VEGF therapeutic agents is known to be mediated at least in part through the HGF/c-Met axis, the strategy of targeting both pathways has gained considerable interest.79, 80 This therapy has demonstrated single-agent antitumor activity in several types of solid tumors, particularly androgen-resistant prostate cancer.81

Rilotumumab (AMG 102) is a fully humanized monoclonal antibody (immunoglobulin G2) that specifically targets HGF and is being developed by Amgen Inc (Thousand Oaks, Calif). It binds to amino acid residues at the NH2-terminus of the beta-chain of human HGF, with a preference for the mature, heterodimeric form, leading to the complete inhibition of c-Met autophosphorylation.82 Unfortunately, it demonstrated no sign of activity when tested in a phase 2 trial of patients with recurrent glioblastoma.83 Although the authors did not speculate on the underlying cause of AMG 102′s inactivity, they suggested that pretreatment with anticancer therapies including bevacizumab may have been responsible. Because multiple tyrosine kinase receptors are active in patients with glioblastoma, targeting only c-Met may not improve PFS. Furthermore, the authors also suggested that although unlikely, the blood-brain barrier may have limited the transport of the drug.

DN30, developed by Metheresis Translational Research SA (Lugano, Switzerland), is a monoclonal antibody targeting the extracellular domain of c-Met directly at a site distinct from the binding site of its HGF ligand.84 Although treatment with DN30 significantly downregulates c-MET,85 the process is complicated by its bivalent structure, which results in partial agonism of the c-Met receptor and thus has hindered clinical development of this agent.84

MetMAb (OA-5D5) from Genentech (South San Francisco, Calif) is perhaps the furthest along in clinical development at this time. MetMAb is a monovalent antibody specifically designed to prevent dimerization of the receptor. In phase 2 clinical testing, erlotinib was tested with or without MetMAb as second-line or third-line treatment of patients with stage IIIB or IV NSCLC. More than 120 patients were equally randomized to receive the combination of erlotinib and MetMAb versus erlotinib alone.86 Met-positive tumors were those in which > 50% of tumor cells stained with an intensity of 2 or 3 on a scale of 0 to 3. Patients with this type of tumor who were treated with MetMAb plus erlotinib had a trend toward improved overall survival (HR, 0.55; P = .11) and possibly PFS as well. It interesting to note that the reverse was observed for patients with Met-negative tumors in that MetMAb plus erlotinib produced worse overall survival compared with erlotinib plus placebo (HR, 3.02; P = .02). Based on these positive results, MetMAb currently is being tested in a phase 3 trial exclusively for patients with Met-positive tumors.

Given the encouraging findings from trials of targeted therapies performed to date, particularly for patients with high c-Met–expressing tumors, it is logical to expand these findings by combining targeted therapeutic agents with cytotoxic agents such as chemotherapy, radiotherapy, or both. Furthermore, because blocking EGFR has also been found to be effective in overcoming radiation resistance, blocking both EGFR and c-Met in combination with radiotherapy is an attractive therapeutic strategy to overcome resistance to both EGFR inhibitors and radiation.

Conclusions

With ever-increasing numbers of studies linking it with cancer progression, c-Met has become an attractive candidate for targeted anticancer therapy, often in combination with chemotherapy. With tivantinib, a c-Met inhibitor being tested with erlotinib in a phase 3 trial for the treatment of nonsquamous cell NSCLC (the MARQUEE trial [Phase 3 Met inhibitor ARQ 197 plus Erlotinib vs Erlotinib plus placebo in NSCLC]), and with the emerging significance of c-Met in cancer, EMT, and DNA damage repair pathways, the logical next step is to test c-Met–targeting agents in combination with radiotherapy in clinical settings.

FUNDING SUPPORT

This work was made possible through the generosity of the family of M. Adnan Hamed, the Orr Family Foundation, and the Wiegand Foundation. It was also supported in part by National Cancer Institute grants K12 11111246, P01CA06294, CA155196, CA160398, CA168484, and CA016672; Department of Defense awards W81XWH-06-1-0303 (the BATTLE trial) and W81XWH-07-1-03,06 (the PROSPECT trial); the University of Texas Southwestern Medical Center; and The University of Texas MD Anderson Cancer Center Lung SPORE grant P50 CA070907.

CONFLICT OF INTEREST DISCLOSURE

Dr. James Welsh is listed on the patent describing the use of MP470 for cancer.

Ancillary