Acquired MET Exon 14 Alteration Drives Secondary Resistance to Epidermal Growth Factor Receptor Tyrosine Kinase Inhibitor in EGFR-Mutated Lung Cancer

CORRESPONDING AUTHOR Marc Ladanyi, MD, Memorial Sloan Kettering Cancer Center, Molecular Diagnostics Service, 1275 York Ave, New York, NY 10065; ladanyim@mskcc.org. K.S., M.O., and A.J.S. contributed equally to this work as co-first authors. R.S. and M.L. contributed equally to this work as co-senior authors. AUTHOR CONTRIBUTIONS Conception and design: Ken Suzawa, Michael Offin, Adam J. Schoenfeld, Maria E. Arcila, Alexander Drilon, Gregory J. Riely, Romel Somwar, Marc Ladanyi Provision of study material or patients: Daniel Lu, Charles M. Rudin, Helena A. Yu, Gregory J. Riely Collection and assembly of data: Ken Suzawa, Michael Offin, Adam J. Schoenfeld, Andrew J. Plodkowski, Daniel Lu, William W. Lockwood, Maria E. Arcila, Charles M. Rudin, Alexander Drilon, Helena A. Yu, Romel Somwar, Marc Ladanyi Data analysis and interpretation: Ken Suzawa, Michael Offin, Adam J. Schoenfeld, Igor Odintsov, Daniel Lu, Alexander Drilon, Helena A. Yu, Romel Somwar, Marc Ladanyi Manuscript writing: All authors Final approval of manuscript: All authors


INTRODUCTION
Novel resistance mechanisms to epidermal growth factor receptor (EGFR) tyrosine kinase inhibitors (TKIs) in EGFR-mutant lung cancer continue to be defined. 1 With the recent approval of the thirdgeneration EGFR TKI osimertinib as first-line treatment, and expanded use of large-panel next-generation sequencing (NGS)-based testing, many novel resistance mechanisms are emerging. Resistance to osimertinib can result from on-target mechanisms, such as the acquisition of second-site EGFR mutations, or from the activation of off-target mechanisms or bypass pathways, including acquired oncogenic fusions of RET, ALK, BRAF, and FGFR 1,2 and amplification or mutation of HER2, BRAF, MEK, KRAS, PIK3CA, and MET. 3 MET amplification is reported in 5% to 22% of EGFR TKI resistance. 1,4 Preclinical studies have shown that acquired resistance to EGFR TKIs resulting from MET amplification can be reversed by combined therapy with EGFR and MET inhibitors. 4 MET exon 14 skipping alteration (METex14) is present in 3% to 4% of lung adenocarcinomas, 5 and the MET receptor lacking exon 14 shows decreased protein turnover because of loss of the ubiquitination site encoded by exon 14, resulting in aberrant MET activation and oncogenesis. 6 In preclinical and clinical studies, responses to MET inhibitors, such as crizotinib, have been reported in patients with lung cancer with METex14 as a primary driver. [6][7][8][9][10][11] However, it has not been previously implicated in acquired resistance to EGFR-TKIs. In this study, we used targeted NGS with Memorial Sloan Kettering Integrated Mutation Profiling of Actionable Cancer Targets (MSK-IMPACT), 12 immunohistochemistry, cell-free DNA testing, and fluorescence in situ hybridization to evaluate acquired resistance mediated by METex14. Furthermore, we used in vitro functional studies to establish METex14 as a novel mechanism of acquired resistance to EGFR TKIs.
We used MSK-IMPACT, a large-panel NGS assay, to detect mutations, copy-number alterations, and select gene fusions involving up to 468 cancer-associated genes. 12 METex14 was introduced into PC9 and H1975 cells as follows. Briefly, full-length METex14 was polymerase chain reaction amplified and subcloned into pLenti-CMV-blast lentiviral vector (plasmid 17451; Addgene, Cambridge, MA). The lentiviral plasmids were cotransfected with packaging plasmids into HEK 293 T cells using FuGENE HD (Promega, Madison, WI), and lentiviruses were generated. Cells were infected with lentivirus-expressing METex14 cDNA, followed by selection with blasticidin (20 mg/mL) for 8 days. The Data Supplement provides more detailed methods.

CASE REPORT
A 73-year-old woman who had never smoked presented with lung adenocarcinoma, which was diagnosed via bronchoscopy with biopsy of the left upper lobe, and underwent a left upper lobe lobectomy and lymph node dissection, which showed a stage IIB (pT2bN0M0) poorly differentiated adenocarcinoma. Sequenom mass spectrometry 13 revealed an EGFR L858R mutation, and the patient was administered adjuvant erlotinib (100 mg daily). 14 After 24.7 months of erlotinib, given no recurrence, adjuvant therapy was discontinued ( Fig 1A). The patient was observed for 20.5 months, when imaging revealed new bilateral pulmonary nodules, right-sided paratracheal lymphadenopathy, and a sclerotic T11 lesion. Right upper lobe biopsy confirmed recurrent disease, and MSK-IMPACT testing showed the presence of EGFR L858R without EGFR T790M mutation. The patient restarted erlotinib (100 mg daily) with clinical and radiologic response for 12.5 months, at which time computed tomography revealed an increase in the dominant right upper lobe mass. Fluorescence in situ hybridization of right upper lobe biopsy material revealed MET amplification, and cell-free DNA testing 15 Table A1). Therapy was changed to osimertinib with savolitinib daily (ClinicalTrials. gov identifier: NCT02143466) for 1.4 months, after which savolitinib was stopped because of toxicity and single-agent osimertinib 80 mg daily was continued. Progressive disease in the lung was noted after 2.4 months of osimertinib ( Fig 1B). Crizotinib 250 mg twice daily was then administered for 1.9 months, at which time further pulmonary progression of disease was noted ( Fig 1C). Treatment was changed to combination osimertinib (80 mg daily) with crizotinib (250 mg twice daily). The combination was tolerated without any report of toxicity. At follow-up 2.3,  Suzawa et al 4.6, and 7.7 months after starting combination therapy, she had ongoing clinical benefit and stable disease by RECIST (version 1.1; −12.2% response; Fig 1D). The patient continued to receive combination therapy with durable clinical and radiographic benefit for more than 9 months.
To define the role of METex14 in mediating resistance to EGFR TKIs, we generated two isogenic EGFR-mutant non-small-cell lung cancer (NSCLC) cell models using PC9 (exon 19 deletion) and H1975 cells (L858R and T790M) by transduction with lentiviral vectors driving expression of METex14 (Fig 2A). Western blot analysis showed that phosphorylation of EGFR and its downstream effectors AKT and ERK was inhibited by osimertinib in PC9 cells transduced with empty plasmids (PC9 empty), but phosphorylation of EGFR, METex14, and downstream effectors remained unaffected by osimertinib treatment in PC9 METex14 cells (Fig 2B). Notably, METex14 expression correlated with upregulation of phosphorylated EGFR in PC9 METex14 cells. In cell viability assays, the presence of METex14 reduced sensitivity to osimertinib by approximately 20-fold (half maximal inhibitory concentration: PC9
We next investigated whether METex14-mediated resistance to EGFR TKIs could be overcome by combination therapy with EGFR and MET inhibitors. As expected, crizotinib inhibited METex14 phosphorylation in PC9 METex14 cells; however, phosphorylation of EGFR, AKT, and ERK remained largely unchanged (Fig 3A), suggesting that EGFR is still signaling effectively in PC9 METex14 cells. Similarly, crizotinib was ineffective at modulating growth of EGFR-mutated cell lines, with or without METex14 expression ( Fig 3B). However, a combination of osimertinib and crizotinib inhibited activation of EGFR, MET, AKT, and ERK ( Fig 3C). Moreover, addition of crizotinib restored the growth inhibitory effects of osimertinib in PC9 METex14 cells (Fig 3C). Identical results were observed in the H1975 model (Figs 3C and 3D). In agreement with these results, dual inhibition of EGFR and MET caused significantly higher activation of caspase 3/7 (Fig 3E).

DISCUSSION
Our study highlights the importance of serial and diverse molecular analyses, including NGS, to evaluate acquired alterations in the post-TKI setting. Here, we show how acquired METex14 mediated resistance to osimertinib. Although the patient did not respond to MET-targeted therapy alone, the patient continued to have a durable clinical response to combination osimertinib and crizotinib, with stable disease by RECIST criteria and without notable toxicity. Our functional data were consistent with these clinical observations. We found that expression of METex14 in NSCLC cell lines with activating EGFR mutation resulted in resistance to osimertinib. Crizotinib restored sensitivity to EGFR TKIs; however, crizotinib alone was not enough to suppress growth.
Two previous reports have demonstrated co-occurrence of EGFR and METex14 mutations. In the first report, three (0.2%) of 1,590 patients with NSCLC harbored concomitant EGFR and METex14 mutations, one of whom received combination treatment with MET-(volitinib) and EGFRtargeted therapies (gefitinib), yielding a partial response. 16 The second report noted one patient case of sarcomatoid carcinoma with EGFR and METex14 alterations. 17 In our MSK-IMPACT testing experience of 866 patient cases of EGFR-mutant lung adenocarcinomas (as of October 15, 2018), only two patients showed this combination of alterations: the patient described here, with acquired resistance, and a patient in whom the alterations were present in separate primaries at diagnosis.
The clinical benefit of the combination of MET-and EGFRtargeted therapies in patients with NSCLC with acquired MET amplification-mediated resistance to EGFR TKIs has been explored in clinical trials, with varying tolerability dependent upon the agents being combined. 18 Our findings provide a rationale for future clinical evaluation of this combination approach, given its tolerability and efficacy in this case, for patients with EGFR and METex14 mutations. Furthermore, given the recent reports of secondary-site mutations in the MET kinase domain, such as D1228N/V and Y1230C, as mechanisms of acquired resistance to crizotinib in patients with METex14, 19 it is plausible that these secondary MET mutations will also emerge as mechanisms of resistance to the combination of osimertinib and crizotinib.
We found that expression of METex14 upregulated phosphorylation of EGFR itself, presumably via crossphosphorylation, because MET is known to interact with EGFR and drive the activity of EGFR, 20 which resulted in blunting of the inhibition of EGFR phosphorylation by EGFR TKIs. In addition, MET inhibition restored the antagonistic effect of osimertinib on EGFR signaling. Taken together, the results suggest a complex interaction between MET and EGFR in NSCLC in the presence of EGFR TKIs and provide unique insight into potential resistance mechanisms and management strategies in EGFR/METex14-altered lung cancers.
In summary, METex14 is a novel mechanism of acquired resistance to EGFR TKI therapy in EGFR-mutant lung cancer. We show in this case that this mechanism of resistance can be effectively treated with a combination of osimertinib and crizotinib.