Emerging discoveries of MET chromosomal fusions provide additional promise as therapeutic genomic targets in MET cancer therapy and would need to be tested vigorously.
Over the last several years of clinical MET targeting therapeutics development, mixed clinical trial studies results especially in the phase 3 study outcomes have resulted in disappointment and hurdles in the drug development. More recent advances in technological profiling platforms and their increasingly common implementation to inform therapeutics in cancer therapies have uncovered and validated novel MET alterations especiallyMETex14juxtamembrane skipping alternative splicing variant, and MET chromosomal fusion products. These findings alongwith some emerging clinical tumor response in MET therapy targetingMETex14in lung adenocarcinoma, soft tissue sarcoma and pulmonary sarcomatoid carcinoma have regenerated new enthusiasm in further exploring MET targeted therapy. It is also becoming clear in recent years that concurrent genomic, proteomics and immune surveillance milieu beyond the identifiable MET alterations could play important roles in determining the ultimate tumor response to MET targeted therapy. These bioinformatics data should be considered essential in modern omics era drug development and in the elucidation of predictive biomarkers. Emerging discoveries of MET chromosomal fusions provide additional promise as therapeutic genomic targets in MET cancer therapy and would need to be tested vigorously.
MET is a multifunctional receptor tyrosine kinase (RTK) that normally plays an important role in embryonic developmental signaling. In adulthood tissues and cellular environment, it is primarily activated during tissue injuries and wound healing, otherwise it remains mostly quiescent. Pathologically, MET signaling dysregulation leads to malignant cellular transformation, proliferation, survival, motility and migration, scattering, epithelial-mesenchymal transition (EMT), angiogenesis, invasion, and metastasis.1The MET protooncogene resides on the 7q31 locus of chromosome 7, consisting of 21 exons separated by 20 introns.2,3MET is composed of an a-chain (50 kDa) and a transmembrane b-chain (140 kDa) subunit linked by a covalent disulphide bond. The MET b-chain adopts a 7-blade b-propeller fold structural domain within the aminoterminal Semaphorin (or Sema) domain, which has homology with the plexin family of semaphorin receptors.
The Sema domain plays a key role in the ligand binding and receptor homodimerization of MET.4-6Following the Sema domain is the PSI domain (found in plexins, semaphorins, and integrins), then 4 IPT repeats (found in immunoglobulin, plexins, and transcription factors), and a single a-helix transmembrane domain. The cytoplasmic tyrosine kinase domain contains a number of key serine and tyrosine phosphorylation sites important in the recruitment of SRC-homology-2-domain (SH2) containing signaling transducers and intermediaries. The natural ligand for MET is hepatocyte growth factor (HGF), alternatively called scatter factor (SF), which can be found and is produced in the stromal and mesenchymal cells, acting in endocrine and/or paracrine fashion on the MET- expressing epithelial cells.7Besides, autocrine tumoral source of HGF can also mediate oncogenic signaling.8-12Upon ligand binding to HGF, MET is phosphorylated at multiple residues, with subsequent catalytic activation of cellular signaling transduction cascades.13Phosphorylation of the major autophosphorylation sites Y1230, Y1234, and Y1235, located within the activation loop of the tyrosine kinase domain, activates the intrinsic catalytic MET kinase activity. It leads then to further recruitment of intracellular adaptor molecules through the SH2 domains and other recognition motifs, such as GAB1 (key signaling adaptor coordinator of the cellular responses to MET). Downstream GRB2- mitogen-activated protein kinases (MAPK) signaling cascade, PI3K-mTOR pathway and the STAT signaling pathway are also eventually activated to effectuate various cellular functions.14,15
MET/HGF signaling has been designated as one of the new generation “Hallmarks of Cancer”, as a key oncogenic pathway for “activating invasion and metastasis.”16Extensive past preclinical studies have documented the role of MET/HGF signaling in various processes in the development of human malignancies. The first evidence of clinical relevance of MET kinase mutations was reported in the germline and somatic mutations of MET in the kinase domain, homologous to those in the KIT and Ret Proto-Oncogene (RET) proto-oncogenes, in hereditary and sporadic papillary renal cell carcinoma.17Besides, multitudes of MET mutations including those within the kinase domain and also extracellular sema domain, and more recently within the juxtamembrane domain, have been reported in various human cancers.18-20Oncogenic MET and/or HGF over- expression have also been demonstrated in various human cancers, including both hematologic and solid malignancies (http://www.vai.org/met).8,19.21Phosphorylated MET was shown to be highly expressed in lung cancer tissues among other common solid cancers.8,22The crosstalk between MET and epidermal growth factor receptor (EGFR) family receptors,23,24as well as Kirsten Rat Sarcoma Viral Oncogene Homolog (KRAS) signaling has been implicated in recent years.25-28MET/ HGF signaling can also crosstalk with other oncogenic pathways, such as EGFR, erb-B2 receptor tyrosine kinase 3 (ERBB3), AXL, macrophage stimulating 1 receptor (RON), CD44, a6b4-integrin and Hypoxia-inducible factor 1 (HIF)-1a.24,29-31Some of these “bypass” pathways have emerged as possible mechanisms contributing to acquired EGFR-tyrosine kinase inhibitor (TKI) resistance.32Basic and translational research studies on oncogenic MET/HGF signaling have ushered the therapeutic development of clinical MET targeted agents in recent years in cancer therapy.19-21,33
MET/HGF signaling can be activated through genomic amplification or transcriptional upregulation, protein overexpression of the ligand and/or receptor itself, oncogenic mutations, or alternative splicing. MET oncogenic signaling can be induced via both HGF-ligand dependent or independent fashion. MET gene mutations and copy number variations have been reported in a wide variety of human tumor tissues.8,19,34-39MET mutations can be found clustered in the non-tyrosine kinase domain, namely in the juxtamembrane and Sema domains, in addition to the kinase domain. MET kinase domain mutations have been found to be somatically selected in the metastatic tissues, compared with the primary solid cancers.40Previous studies characterizing the juxtamembrane (JM) mutations (R988C, T1010I, and alternative spliced JM-skipping variant) demonstrated that these are activating oncogenic variants (FIG. 1A), with enhanced oncogenic signaling, tumorigenicity, cell motility, and migration.34,35Somatic and germline variants of MET were first reported in sporadic and hereditary papillary renal cell carcinoma, and later in a variety of other human cancers.34,35,41,42The relative significance and relevance of many specific reported mutations in cancer biology and progression remain to be better defined. By using quantitative real time polymerase chain reaction (QPCR) assay for MET amplification, multiple studies have reported primary MET amplification to be 2-21% in nonsmall-cell lung cancer (NSCLC) lung adenocarcinomas.36-39Based on the QPCR method using formalin-fixed paraffin embedded (FFPE) tumor tissues, MET gene copy number >4 in lung adenocarcinoma tissue samples has been associated with better prognosis (median survival 39 vs 16 months,P= .06).37 Using fluorescence in-situ hybridization (FISH) assay, the Lung Cancer Mutation Consortium (LCMC) reported 4.1% of adenocarcinoma with MET amplification >2.2 (MET/CEP7).43A TCGA lung adenocarcinoma study identified MET amplification to be 2.2%.43 Cappuzzo et al44 reported in a study of 447 NSCLC patients, a high MET gene copy number (greater than or equal to 5 copies/cell) was negatively associated with survival (hazard ratio [HR], 0.66,P= .04).
Most recent attention has turned to the MET exon 14 alternative splicing variant (METex14) that generates an activating form of MET receptor skipping the juxtamembrane domain.18Genomic alterations involving exon 14 skipping alternative splicing of MET were first reported in lung cancer in 2003-200534,35. In the MET mutational analysis studies, 2METex14variants were identified in small cell lung cancer (SCLC) involving a 2 base-pair insertion in a splice acceptor site 5 of exon 14 and one in a non-small cell lung cancer (NSCLC) sample involving an in-frame skipping of exon 14.34,35MET JM domain activating mutations R988C and T1010I were also reported in lung cancer. Soon after, another series of somatic intronic mutations was reported in lung cancer cell lines and patient samples immediately flanking exon 14, and Y1003 residue that serves as the JM domain binding site for Cbl Proto-Oncogene (CBL), the E3-ubiquitin ligase to regulate MET receptor turnover.41
In addition, a recent TCGA lung adenocarcinoma study reported the presence of mRNA transcripts expressingMETex14skipping in 4.3% of samples as a result of somatic in-cis DNA exon 14 splice site mutation (ss mut), splice site deletion (ss del) or a Y1003* mutation45. The frequency ofMETex14skipping in lung adenocarcinoma was determined to be 4.3%. Another transcriptome meta-anal- ysis of lung cancer detectedMETex14skipping in 15 samples with 14 occurring in driver-unknown tumor samples (3.6%).46 Most recently, there have been several additional reports further validating the existence and, importantly, the clinical relevance ofMETex14in human cancers, including but not limiting to lung adenocarcinoma. Paik et al reported that mutations of RNA splice acceptor and donor sites involving MET exon 14 lead to exon skipping and in-frame deletion of the JM domain. Furthermore, clinical evidence of tumor response to MET targeted therapies using crizotinib or carbozantinib was reported.47In another article, Frampton et al identified through a Clinical Laboratory Improvement Amendments (CLIA)-certified clinical tumor genomic profiling platform from over 38,000 patient samples, recurrent and highly diverse genomic alterations in multiple tumor types, with overall frequency at 0.6% (221/38,028) ofMETex14. Among the positive cases, there were 126 distinct genomic sequence variants. Of note,METex14mutations were found most frequently in lung adenocarcinoima (3%), but also in other lung neoplasms (2.3%), brain glioma (0.4%), and tumors of unknown primary origin (0.4%).48 A small series of patients harboringMETex14was identified that demonstrated tumor response toward MET inhibitors crizotinib and INC280. Moreover,METex14variants have most recently been reported to occur very frequently in pulmonary sarcomatoid carcinoma (PSC) at 22%,49with an example of remarkable tumor response to crizotinib observed in a chemo-refractory patient with PSC.
Recently, chromosomal translocations have emerged as a class of oncogenic aberrancy more important than previously recognized in solid cancers especially in prostate and lung cancers. Interestingly, novel chromosomal translocations involving MET have also been uncovered in a variety of human cancers.50In particular, at least 2 fusion variants (ie, KIF5BMET in lung adeno-carcinoma and TFG–MET in thyroid papillary carcinoma) evidently possess the predicted chimeric protein confirmed with the classic fusion activation paradigm, joining the dimerization motifs to an intact kinase domain.50These findings provide support for chromosomal MET translocation products to be a candidate as abona fideoncoprotein. Moreover, protein tyrosine phos- phatase, receptor type, Z polypeptide 1(PTPRZ1)METfusion transcripts have been identified as a novel and recurrent alteration in grade III astrocytomas (7.7%) and in secondary glioblastomas (15%) through RNA-seq analysis.51 Recent studies also suggested thatPTPRZ1METenhanced expression and phosphorylation of the fusion-MET receptor.52
There are generally three main cancer therapeutic strategies in targeting the MET/HGF pathway: HGF antibodies, MET receptor antibodies, and MET receptor small molecule inhibitors (ATP-competitive and non-competitive). Many of them have undergone various phases of clinical trial studies in recent years, resulting in mixed study outcomes. To this date, there have not been an inhibitory agent against MET or HGF that have been FDA-approved primarily for targeting MET/ HGF in cancer therapy. The following section will review succinctly the rationale and updates for targeting various forms of MET genomic alterations in cancer therapy.
Up to date, there have been two mature phase III clinical trial studies targeting MET in cancer therapy, both followed successful positive phase II clinical study results. One involved an anti-MET, one-arm monoclonal antibody onartuzumab (MetMab) which was found to be efficacious in advanced NSCLC patients selected for high MET expression by protein IHC assay, leading to the development of the phase III METLung trial. The METLung study was a biomarker-selected clinical trial to investigate onartuzumab/erlotinib versus erlotinib/ placebo in previously treated stage IIIB-IV NSCLC with centrally confirmed MET-positive expression by IHC (MET-IHC 2+/3+ in ≥50% tumor cells). While the phase II study results strongly supported the notion that MET expression status by IHC may predict clinical benefit from onartuzumab/erlotinib combination; the phase III METLung study was terminated early due to lack of clinically meaningful efficacy. Another phase III study involved the use of tivantinib (ARQ197) which is a non- ATPcompetitive small molecule MET inhibitor. A global randomized phase II trial ARQ197-209, comparing erlotinib/tivantinib (ET) versus erlotinib/placebo (EP) in unselected advanced NSCLC, revealed progression- free-survival (PFS) as the primary endpoint to be prolonged in the ET group. Biomarker analysis highlighted that among nonsquamous tumors, 75% were MET-positive by IHC(2+/3+), compared with only 12% among squamous subtype. Exploratory analysis demonstrated significant delay in time-to-development of new me- tastases among patients treated with ET (HR 0.49,P<0.01), most notable in the nonsquamous population. This study led to the activation of a global randomized phase III clinical trial, MARQUEE, focusing on patients with nonsquamous NSCLC histology, and patients enriched for MET-high expression, using overall survival (OS) as the primary endpoint.53The MARQUEE trial was again discontinued early after a planned interim analysis revealed study futility. Nonetheless, final analysis showed both PFS and overall response rates (ORR) were improved. The tivantinib treatment group did show significant OS improvement in the subgroup with MET-high expression, essentially recapitulating the phase II onartuzumab study results.
Although crizotinib was initially developed as a putative MET inhibitor, its anaplastic lymphoma kinase (ALK) activities were validated as being clinically relevant in inhibiting ALK(2p23) translocation-positive NSCLC leading to its FDA approval in 2011. Thereafter, MET genomic amplification associated with various tumor types has been reported to correlate with crizotinib treatment clinical response.33Hence, it was believed that MET amplification represents a predictive biomarker for a subgroup of tumors potentially addicted to MET-driven oncogenic pathways. A recent TCGA Research Network report on lung adenocarcinoma confirmed a frequency of 2.2% MET amplification, and provided evidence for it as “driver” alteration.45A report of crizotinib treatment in MET-amplified NSCLC from the original phase I study included 14 patients accrued to the NSCLC cohort, mostly adenocarcinoma and with positive smoking status.54This study observed objective response rates to be: 0% in the low-MET (MET/CEP7 ratio ≥1.8-≤2.2), 17% in the intermediate-MET (ratio >2.2-<5.0), and 67% in the high- MET (ratio ≥5.0) group, hence supporting the notion of an improved efficacy correlating with an increasing MET amplification ratio.
Recently, clinical validation of precision therapy targeting the alternative splicing variants of MET receptors with the JM domain (encoded by exon 14) skipped, with the use of MET inhibitors, has been reported. The two recent reports by Paik et al47 and Frampton et al48 further enrich our understanding of MET as molecular target in precision cancer therapy. In the largest tumor genomic profiling cohort performed for MET alteration, Frampton et al reported 221 positive cases (0.6%) expressMETex14mutations out of 38,028 profiled tumors. Most interestingly,METex14alterations comprised a highly diverse 126 distinct sequence variants that were observed commonly but not exclusively in lung adenocarcinoma (3%).METex14alterations were identified also in other non-adenocarcinoma lung tumor types (2.3%), brain glioblastoma (0.4%), tumors of unknown primary (0.4%), and a number of other tumor types (0.06%) including renal cell carcinoma, soft tissue sarcoma, and hepatocellular carcinoma. Both reports highlightedMETex14conferring sensitivity towards MET-targeting inhibitors with clinical response either by tumor measurement or metabolic positron-emission tomography (PET) response, further raising the specter ofMETex14collectively as “actionable” genomic alterations and cancer predictive biomarkers. Since the JM domain is a key negative regulatory region for the intracellular kinase domain in the human kinome, its disruption through splicing skipping in MET likely can transition the closed kinase conformation to a more open and thus active conformation, akin to the effects of oncogenic fms-related tyrosine kinase 3(FLT3)-ITD(internal tandem repeat) in acute myeloid leukemia.METex14variants also stabilize the altered MET receptor through decreased CBL-mediated MET-ubiquitination.
Short interfering RNA silencing of MET and MET inhibition with crizotinib displayed marked effectiveness on cell viability and in downstream Protein kinase B (AKT) and MAPK survival-proliferative signaling inhibition in the gastric cancer cell line that harborsMETex14as driver alteration alone, Hs746T. Moreover,METex14variants have most recently been reported to occur very frequently in pulmonary sacomatoid carcinoma at 22%.49The study also highlighted a patient within the study who hadMETex14and concurrent MET amplification, and was chemotherapy-refractory, but found to demonstrate remarkable response to crizotinib. Nonetheless, it is of note that often con- current genomic mutations could modulate tumor response to precision therapies against a driver-alterations as in the case of H596 adenosquamous cell line which expresses not only theMETex14skipping variant but alsoPIK3CAmutation. H596 cells were found to be insensitive to MET inhibitor crizotinib alone but had synergistic sensitivity to combined MET/PI3K in- hibitors treatment in preclinical models.49
The identification of novel MET fusion transcripts of Kinesin Family Member 5B (KIF5B)MET, TRK-Fused Gene (TFG)–MET, andPTPRZ1MET50-52 in various cancers recently raises the question of whether chromosomal MET translocations would indeed represent predictive genomic biomarkers that can respond to MET targeting inhibitors/antibodies. This notion is rather attractive although definitive clinical evidence thus far is currently lacking. That said, recent clinical tumor response in genomics-matched targeted therapies against various novel fusion onco-kinases involving RET,55,56ROS Proto-Oncogene 1 (ROS1),57,58and also remarkably NTRK59in NSCLC, sarcoma, and other solid tumor types does offer promising notes to speculate perhaps similar clinical response might soon be observed in MET-fusion driven tumors.
Multitudes of MET mutations have been reported in various cancer tissues and also now more systemically compiled in TCGA database over the years [cBioPortal for Cancer Genomics database (http://www.cbioportal.org/index.do)]. While many of the specific mutations involving the kinase domain and the JM domain have been shown to play oncogenic roles and promote tumor cellular motility/migration, invasion, and metastasis, the specific oncogenic role of most of the identified somatic mutations nowadays remains less clear. Furthermore, specific mutations of MET kinase domain may constitute differing oncogenic signaling in a tissue-specific fashion.60,61Clinically, if MET mutations play a role as a critical signaling network to promote tumor invasion and metastasis, and not primarily in tumor cell proliferation or survival per se, targeting MET therapeutically might not result in remarkable tumor shrinkage response as seen in other classic oncogene-addicted tumor scenarios under genomics-matched precision inhibitor therapy.
Recent report on the novel role of MET as expressed on neutrophils within the tumor-associated stromal compartment and its requirement for neutrophil chemoattraction and cytotoxicity in response to HGF is provocative.62MET deletion in mouse neutrophils was found to upregulate tumor growth and metastasis, correlating with reduced tumor-infiltrating neutrophils both to primary and metastatic lesion sites. Most intriguingly, the authors showed that MET targeting inhibitor cancer in vivo treatment had confounding pro-tumoral effects stemming from the MET-blockade in neutrophils thereby counteracting the therapeutic anti-tumoral effects.62More work is needed to further understand if the model translates closely in the clinical setting of MET therapy in cancer patients. It would now be critically important to better comprehend the potential interplay between MET/HGF signaling pathway and the tumor-host immune surveillance network.
Over the last several years of clinical MET targeting therapeutics development, it has become clear that the invasive MET/HGF signaling axis likely is not a simple molecular target. The signaling pathway plays multifaceted roles in tumorigenesis and in the process of tumor progression and metastasis. The case for MET alterations to represent an oncogenic addictive event whereby clinically targeting the aberrant MET receptor with specific inhibitors could bring about remarkable tumor shrinkage response has been an attractive goal in clinical therapeutic development. Nonetheless, genuine tumor addiction to MET for proliferation and survival may not be an exceedingly common occurrence in human tumors. Tumor dependence on activated MET pathway to promote EMT, invasion and metastasis, is also a key consideration in clinical MET targeting but might be difficult to measure in inhibitory efficacy. Furthermore, the multitudes of activating venues of MET/HGF signaling in human cancers, including genomic amplification and transcriptional upregulation, protein overexpression, somatic and possibly germline mutations,METex14JM domain skipping splicing alterations, and chromosomal fusions raise an important question of whether each of these alterations may have differential impact over oncogenic signaling and tumor biology. The biologic and clinical relevance of the genomic target alterations is key to successful clinical molecular targeting.
In recent years, with technological advances in vari- ous genomic and molecular profiling methodologies and platforms, it has become more feasible and routine to uncover various MET/HGF alterations in tumor tissues. Nonetheless, it cannot be over-emphasized in this new era of cancer genomics that understanding the impact of concurrent genomic alterations beyond the MET/HGF aberrancies is essential in refining and defining the role of MET/HGF alterations as predictive biomarkers for targeted therapies. It is important to implement broad and comprehensive molecular-genomic tumor profiling in MET targeting clinical studies and in clinical therapy. For instance, the finding ofMETex14skipping alterations with concurrent MET amplification warrants further definitive investigations.45Similarly, the coincident genomic events of genomic amplification of MDM2 and CDK4 withMETex14also require further studies to address their underlying clinical implications.45Interestingly, a recent study also highlighted the occurrence of molecular heterogeneity and receptor co-amplification in driving tumor resistance against targeted therapy in MET-amplified esophagogastric carcinomas.63
Window-of-opportunities in MET targeting therapies now include alsoMETex14, MET-fusions, and MET amplification. More thoughtful clinical outcomes measurement would need to be developed to implement in clinical trial studies of MET/HGF therapeutics to increase the chance of success. To this end, novel emerging genomic biomarkers assays in genomics and proteomics as well as circulating tumor DNA (ctDNA) liquid biopsy platforms to detect and monitor genomic MET alterations undoubtedly would facilitate our abil- ity to better understand the impact of MET inhibition and to further define clinically relevant predictive biomarkers in MET cancer therapy. Lastly, the potential for other interacting and counteracting pathways in MET therapy, especially in tumor-host immune surveillance pathways would deserve in-depth investigations to help arrive at rational and optimal combinational MET targeting cancer therapeutic regimens.
Patrick C. Ma, MD, MSC, is supported by WVU Cancer Institute - Mary Babb Randolph Cancer Center, WVU Medicine, West Virginia University, and also West Virginia Clinical and Translational Science Institute (WVCTSI), by the National Institute Of General Medical Sciences of the National Institutes of Health under Award Number U54GM104942 (IDeA CTR). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. Xiaoliang Wu, MD,. is supported by a scholarship to study overseas by the Sun Yat-sen University Cancer Center, China.