Clinical trials ID number NCT00612703, at ClinicalTrials.gov
Amplification of the mesenchymal-epithelial transition factor (MET) gene can promote tumor resistance to epidermal growth factor receptor (EGFR) inhibition. Dual EGFR-MET inhibition may overcome this resistance. Tivantinib (ARQ 197) is a selective, oral, non–ATP-competitive, small-molecule inhibitor of the MET receptor tyrosine kinase. This phase 1 trial assessed the safety, pharmacokinetics, and preliminary antitumor activity of tivantinib combined with the EGFR inhibitor erlotinib.
Patients with advanced solid malignancies were administered oral tivantinib at escalating doses of 120, 240, 360, and 480 mg twice daily (BID) plus 150 mg erlotinib once daily (QD). Single or multiple intrapatient dose escalation was planned in the absence of dose-limiting toxicity in the first cycle of therapy (21 days).
Thirty-two patients received combination treatment. Tivantinib serum concentrations were not dose-proportional. The most common (≥20%) adverse events (AEs) regardless of causality included rash (n = 17), fatigue (n = 12), nausea (n = 10), abdominal pain (n = 10), diarrhea (n = 9), bradycardia (n = 9), and anemia (n = 7). AEs considered related to study treatment occurred in 28 patients (87.5%), and 5 patients (15.6%) had treatment-related serious AEs, including neutropenia, leukopenia, syncope, sinus bradycardia, and sick sinus syndrome. Fifteen of 32 patients (46.8%) had a partial response (n = 1) or stable disease (n = 14) as assessed by Response Evaluation Criteria in Solid Tumors. Six of 8 patients with nonsmall cell lung cancer achieved stable disease. The recommended phase 2 dose is tivantinib 360 mg BID plus erlotinib 150 mg QD.
Cancer therapy directed against the epidermal growth factor receptor (EGFR) pathway has improved outcomes for patients with cancers of the lung, colon, pancreas, and head and neck.1, 2 A low objective response rate to erlotinib for patients with nonsmall cell lung cancer (NSCLC) (8.9% in an unselected cohort) was demonstrated in the pivotal BR.21 trial.1 Concurrently, the pathophysiology underlying response to EGFR inhibition was elucidated in several studies, demonstrating that the presence of somatic mutations, primarily in exons 19 and 21 of the EGFR gene, led to enhanced tyrosine kinase activity and increased sensitivity to the EGFR tyrosine kinase inhibitors (TKIs) erlotinib and gefitinib.3, 4 Evidence of clinical benefit to TKIs in patients with wild-type, amplified EGFR has also been reported.5 However, the duration of even the most dramatic responses observed in these studies was often short (eg, median duration of response was 7.9 months in the BR.21 study), and ultimately all patients developed resistance.
The efficacy of EGFR TKI therapy is undermined by both primary and secondary resistance mechanisms. Primary resistance to EGFR TKI therapy has been identified in patients with mutated KRAS (v-Ki-ras2 Kirsten rat sarcoma viral oncogene homolog).6 In addition, mechanisms of acquired resistance (ie, secondary resistance) to EGFR TKI therapy have also been elucidated. A secondary mutation in exon 20 of EGFR, a substitution of methionine for threonine at position 790 (T790M), has been identified in approximately half of the patients with acquired resistance to EGFR TKIs.7, 8 More recently, 2 other second-site EGFR mutations have been demonstrated to confer resistance to EGFR TKIs: a T854A substitution in exon 21 and a D761Y substitution in exon 19.8, 9
EGFR TKI resistance can also be mediated through parallel activation of downstream signal transducers. Amplification of the mesenchymal-epithelial transition factor (MET) gene has been associated with activation of the phosphoinositide 3-kinase (PI3K)/Akt pathway, a pathway that is also activated during EGFR-mediated signaling. Indeed MET gene amplification has been associated with tumor resistance to EGFR TKI therapy in both in vitro and in vivo studies.9, 10 In support of this, MET amplification has been detected as a secondary event in approximately 22% of patients with EGFR-mutated NSCLC that have acquired erlotinib or gefitinib resistance.9, 10 Mechanistically, MET amplification may contribute to EGFR TKI resistance through increased MET heterodimerization with human EGFR-related protein 3 (HER3) and activation of the PI3K/Akt cell survival pathway (Figure 1).10 Consistent with this, EGFR inhibition can also be overcome by exposure to the MET ligand hepatocyte growth factor (HGF) in preclinical models.11
Given the interplay of the EGFR and MET signaling pathways, preclinical studies have evaluated the possible synergistic benefit of dual inhibitory strategies.10, 12 In lung cancer cell lines that developed resistance to gefitinib because of MET amplification, treatment with a MET inhibitor restored sensitivity to gefitinib.10 Likewise, treatment of erlotinib-resistant H1975 cells harboring the EGFR T790M mutation with a small-molecule MET inhibitor plus erlotinib enhanced inhibition of EGFR downstream signaling pathways and promoted tumor regression in the H1975 xenograft model.12 Furthermore, tivantinib (ARQ 197), a selective MET inhibitor, demonstrated synergistic activity when combined with erlotinib in xenograft studies with the NCI-H441 cell line (unpublished results from ArQule, Inc., Woburn, Massachusetts, and Kyowa Hakko Kirin Co., Ltd., Tokyo, Japan). These studies demonstrated that dual inhibition can overcome EGFR TKI resistance, whether due to the EGFR T790M mutation or MET amplification.
Tivantinib is a selective, oral, non–ATP-competitive small-molecule inhibitor of the MET receptor tyrosine kinase with a 50% inhibitory concentration (IC50) of 355 nmol/L in biochemical assays, and an IC50 of 300 nmol/L in studies of the MET-amplified human lung carcinoma cell line NCI-H441.13 In 2 phase 1 dose-escalation studies that evaluated the safety of orally administered tivantinib monotherapy, the recommended phase 2 dose was established as 360 mg twice daily (BID), and a tolerable toxicity profile was observed, with the most common adverse events (AEs) comprising fatigue, gastrointestinal disorders (nausea, vomiting, and diarrhea), and anemia.14, 15
On the basis of the strong molecular rationale and preclinical data, dual EGFR-MET inhibition has been proposed as a strategy for overcoming primary and secondary resistance to EGFR inhibition in the clinical setting. As such, patients were enrolled in a dose-escalation trial to define the safety, tolerability, pharmacokinetics (PK), and preliminary antitumor activity of tivantinib in combination with erlotinib.
MATERIALS AND METHODS
Eligible patients were ≥18 years of age and had a pathologically or cytologically confirmed solid tumor that was advanced, inoperable, or metastatic. Patients were required to have an Eastern Cooperative Oncology Group performance status of 0 to 1; adequate marrow, renal, and hepatic function; absence of pregnancy; absence of a gastrointestinal disorder that could interfere with oral absorption of the study medications; and no coexisting severe medical conditions. Patients were required to have measurable disease as defined by Response Evaluation Criteria in Solid Tumors (RECIST). Anticancer therapy including surgery, chemotherapy, radiotherapy, immunotherapy, or investigational agents had to be completed at least 4 weeks prior to the first treatment received in this trial, with the exception that patients previously taking erlotinib monotherapy could be started on the trial without a treatment break. Central nervous system metastases had to be treated, stable for 3 months, and asymptomatic before enrollment. The study was approved by a local institutional review board and in accordance with an assurance filed with and approved by the US Food and Drug Administration. Patients gave written informed consent according to federal and institutional guidelines before any study procedures were performed.
Drug Administration and Trial Design
Both tivantinib and erlotinib were administered orally 1 hour before or 2 hours after a meal in an open-label fashion on a continuous basis in cycles of 21 days. Erlotinib was administered at 150 mg once daily (QD), and tivantinib was administered at escalating doses in 3 sequential cohorts at 120, 240, and 360 mg BID. Single or multiple intrapatient dose escalation at 120 mg BID increments was planned in the absence of dose-limiting toxicity (DLT), defined as grade 4 neutropenia, grade 3 or 4 thrombocytopenia with bleeding, or any other toxicity of grade 3 or higher possibly or probably related to tivantinib in the first cycle of therapy, including diarrhea, nausea, and vomiting that could not be controlled by optimal medical management. This study design was employed to minimize the interpatient PK variability that was observed in the phase 1 monotherapy study and to better assess the PK of tivantinib at different dose levels. Eight patients were enrolled in the first cohort, and their dose was escalated as tolerated. An additional 3 to 6 patients were enrolled in each subsequent cohort. The maximum tolerated dose (MTD) was defined as the dose level at which no more than 1 patient of 6 in a cohort experienced a DLT in the first 21 days of treatment. Once the MTD was determined, an additional 6 to 15 patients could be treated at that dose. Patients received erlotinib and tivantinib combination therapy until disease progression (clinical or radiographic), unacceptable toxicity, or other discontinuation criteria were met.
Pretreatment and Follow-Up Studies
The pretreatment evaluation included a complete medical history; physical examination; a complete blood count with white cell differential; blood chemistry analysis; liver function tests; urinalysis; serum pregnancy test, if applicable; serum tumor markers, if applicable; and a blood sample for cytochrome P450 2C19 (CYP2C19) genotyping. Radiographic assessments by computed tomography or magnetic resonance imaging were performed at baseline (<4 weeks before the first dose) and at 8-week intervals throughout the study. Blood samples for complete blood count with white cell differential, chemistry, and liver function tests were collected on day 1 and day 15 of each cycle. An electrocardiogram was performed before treatment and repeated on day 1 of therapy at 0.5, 1, 2, 3, 4, 6, 10, 12, and 24 hours. Objective tumor response was evaluated by RECIST 1.0. Adverse events were graded according to National Cancer Institute Common Terminology Criteria for Adverse Events (NCI CTCAE) version 3.0 and were assessed throughout the study.
When available, tumor specimens from lung cancer patients were assessed for mutations in EGFR and KRAS by polymerase chain reaction testing for the most common mutations (performed by Clarient, Aliso Viejo, Calif, and Response Genetics, Los Angeles, Calif). These specimens were also tested for MET and EGFR amplification by fluorescent in situ hybridization with gene-specific probes and a centromere 7-specific probe (performed by Caris MPI, Phoenix, Ariz).
Blood samples were collected on days 1 and 15 (before dosing and 1, 2, 4, 6, 10, 12, and 24 hours following tivantinib and erlotinib concurrent administration) of cycle 1. If intrapatient dose escalation occurred, repeat PK samples were obtained on day 15 of that cycle. Pharmacokinetics parameters were analyzed by noncompartmental methods using the WinNonLin 4.0 program (Pharsight, St. Louis, Mo).
The study enrolled 32 patients with a mean age of 60 years (range, 24-86 years) (Table 1). The most common tumor types were NSCLC (8 patients [25%]), and colorectal, kidney, and head and neck tumors (3 patients each [9.4% each]). Patients had received a median of 3 (range, 0-8) prior lines of systemic therapy. Patients in cohort 1 (n = 8), cohort 2 (n = 4), and cohort 3 (n = 20) received tivantinib at a starting dose of 120 mg BID, 240 mg BID, or 360 mg BID, respectively, in combination with erlotinib 150 mg QD.
Table 1. Patient Demographic and Clinical Characteristics
Median days on tivantinib/erlotinib therapy, n (range)
The most common treatment-emergent AEs, regardless of causality, included rash, fatigue, nausea, abdominal pain, diarrhea, bradycardia, and anemia; these were mostly grade 1 or 2 in severity (Table 2). Hematologic toxicities included grade 3 or 4 neutropenia in 3 patients, grade 4 leukopenia in 1 patient, grade 2 and 3 thrombocytopenia in 2 patients, and grade 1 or 2 anemia in 7 patients. Six deaths occurred during study treatment due to disease progression; none were attributed to the study drugs.
Table 2. Treatment-Emergent Nonhematologic Adverse Events Reported in ≥20% of Patients and Hematologic Adverse Events by Tivantinib Dose at Occurrence of Event
A total of 28 patients (87.5%) reported AEs considered definitely, possibly, or probably related to study drug by the investigator. Treatment-related serious AEs were reported in 5 patients, including 1 patient in cohort 1 with grade 3 sinus bradycardia, 1 patient in cohort 2 with grade 3 sick sinus syndrome, and 3 patients in cohort 3 with grade 4 neutropenia (2 events), grade 4 leukopenia, and grade 3 syncope associated with dehydration. The grade 3 sinus bradycardia and sick sinus syndrome occurred on day 36 and day 812 of treatment, respectively, and both events resolved, although the grade 3 sinus bradycardia resolved with sequelae (patient did not require a pacemaker but was concurrently found to have hypothyroidism and was treated with levothyroxine).
Nine of 12 patients in cohorts 1 and 2 had dose escalation of tivantinib to 360 mg BID, and 5 patients escalated further to 480 mg BID, where dose escalation was halted. One patient in cohort 3 dose-escalated to 480 mg BID. Dose reductions occurred in 1 patient in cycle 1 (from 480 mg to 360 mg BID) and 1 patient in cycle 5 (from 360 mg to 240 mg BID), both due to fatigue. No DLTs were observed at the 120- and 240-mg dose levels. Two DLTs occurred at the 360-mg dose level and led to discontinuation of study treatment. One patient developed grade 4 neutropenia on day 15 that resolved after 8 days of drug holiday. A second patient developed grade 3 thrombocytopenia on day 14 that was not resolved when the patient died due to disease progression 32 days later. Neither patient restarted treatment. The MTD was not established; however, tivantinib 360 mg BID plus erlotinib 150 mg QD was considered the recommended phase 2 dose.
The PK profile of tivantinib (360 mg BID) and erlotinib are summarized in Table 3 and Figure 2. Exposure to tivantinib increased in a manner that was less than proportional to dose. There was no accumulation of tivantinib or erlotinib with successive treatment cycles. Although this study was not designed to test for drug–drug interactions, there did not appear to be a significant interaction between tivantinib and erlotinib in this study.
Table 3. Pharmacokinetic Parameters on Cycle 1, Day 1, and Day 15
AUC indicates area under concentration-time curve; Cl/F, apparent clearance; Cmax, maximum plasma concentration; CV, coefficient of variation; SD, standard deviation; t1/2 terminal half-life; Tmax, time to reach maximum plasma concentration.
Morning doses of tivantinib were coadministered with doses of erlotinib unless otherwise stated.
Data shown are arithmetic mean ± SD, geometric mean (geometric CV%) for AUC and Cmax, or median (minimum, maximum) for tmax
n = 4 tivantinib; n = 1 erlotinib
All doses of tivantinib were administered twice daily.
CYP2C19 genotype data were obtained for 31 of 32 patients (97%). Most patients had an extensive (n = 19) or intermediate (n = 11) metabolizer genotype, and only 1 patient had a poor metabolizer genotype. Because of the low number of poor metabolizers, rigorous PK comparisons between CYP2C19 genotypes could not be performed.
Twenty-two patients had at least 1 posttreatment tumor response assessment, and 15 of 32 patients (46.8%) had either stable disease (SD; n = 14) or partial response (PR, n = 1) by RECIST (Table 4). One patient with a chordoma involving the lumbosacral spine had SD for more than 30 months and continues to receive treatment with tivantinib 360 mg BID at the time of manuscript preparation. Two patients with head and neck cancer had disease control for 7 and 8 months, respectively, and 1 of these patients had a PR. Among the 8 patients with NSCLC enrolled in this study, 6 of whom had received prior erlotinib monotherapy, 6 had SD for 3 to 23 months, 1 had progressive disease, and 1 discontinued early due to an unrelated AE and received no posttreatment tumor assessments. Median progression-free survival (PFS) among all 32 patients was 4.1 months (95% confidence interval, 2.0-8.1 months).
Table 4. Treatment Duration and Best Overall Responses for Evaluable Patients (n = 22)
Squamous cell carcinoma of the skin (right axillary)
Microcystic adnexal carcinoma (upper lip/face)
Squamous cell carcinoma (scalp)
Exploratory Analysis of Biologic Profiles for Patients With NSCLC
Specimens were available for exploratory biomarker analysis from 6 of 8 patients with NSCLC (Table 5). One of 5 patients tested was positive for an EGFR mutation, and 3 of 6 patients tested were positive for MET amplification. Further analysis revealed that increased MET gene copy number (GCN), ranging from 3 to 4.5 copies per cell, was associated with a concomitant increase in chromosome 7 centromeres consistent with high polysomy rather than gene amplification. Patients with increased MET GCN remained on study for a mean of 14.7 months (range, 6.6-23.3 months) compared with 8.6 months (range, 1.9-15.8 months) for patients with normal MET GCN. One patient with a KRAS mutation, wild-type EGFR, and MET amplification remained on study for 14.3 months.
Table 5. Biomarker Data and Patient Outcomes for NSCLC Subset of Patients
Continuous therapy with the combination of erlotinib and tivantinib was well tolerated. The occurrence of rash and diarrhea was consistent with previous reports of erlotinib monotherapy.1 Two patients experienced serious AEs potentially related to arrhythmia (sinus bradycardia and sick sinus syndrome) that were considered by the investigator as possibly related to treatment; however, there was no consistency in the timing of these events (day 36 and day 812 of treatment, respectively). Although no formal MTD was identified, a regimen of erlotinib 150 mg QD plus tivantinib 360 mg BID demonstrated a favorable safety profile and preliminary clinical activity and was considered the recommended phase 2 dose. Activity was primarily evidenced by prolonged disease stabilization for at least 2 months in 13 patients (41%) and for more than 6 months (range, 6.5-29.5 months) in 8 patients (25%). The best responses were observed in patients with NSCLC (5 patients [15.6%]), and chordoma, ovarian cancer, microcystic adnexal carcinoma, and squamous cell carcinoma of the skin (1 patient each [3.1% each]).
MET is a cell-surface tyrosine kinase that is essential for embryonic development and wound healing, and it has been implicated in oncogenesis.16 It is expressed physiologically in the liver, gastrointestinal tract, thyroid, kidney, and brain, and its only known ligand is scatter factor, otherwise known as HGF. Activation of MET leads to its binding to and phosphorylation of adaptor proteins such as Gab-1, Grb-2, Shc, and c-Cbl, with subsequent activation of proliferative cascades, including the Ras, PI3K/Akt, signal transducer and activator of transcription, and beta-catenin/Wnt pathways.17 Overexpression of MET frequently occurs in carcinomas and, to a lesser extent, in lymphoma and leukemia. In addition, MET overexpression has been associated with the neoplastic phenotype, including dysfunctional cell survival, growth, angiogenesis, invasion, and metastasis.18-23 Furthermore, MET expression has also been demonstrated to characterize self-renewing cancer stem cells.24
MET activation may play a prominent role in the etiology of NSCLC, because MET axis activation, demonstrated by MET phosphorylation, is present in more than 70% of NSCLC tissue specimens.21 Furthermore, 2 retrospective studies demonstrated that an increased GCN of MET (ie, MET amplification) leading to high MET activity was associated with a poor prognosis in patients with NSCLC following surgical resection and in patients with advanced NSCLC.25, 26 Increased MET GCN is also highly associated with EGFR amplification but has not been observed de novo in EGFR mutant tumors.25, 26
MET amplification may arise from either specific MET gene duplication or through more global chromosome aberrations. True MET gene amplification has been reported in 3.9% to 4.1% of NSCLC specimens, whereas an increase in MET GCN due to high polysomy has been reported in 11.1% to 12.8% of cases.21, 22 Amplification of MET was not associated with sex, histology, smoking status, or cancer stage in the analysis by Go et al, whereas Cappuzzo et al found an association with advanced cancer stage.25, 26 However, the actual incidence of increased MET GCN is unclear, because results from other series have reported rates of MET amplification ranging from 7.3% to 21% of surgical specimens.27, 28 The large variation for the reported incidence of MET amplification may arise from differences in the assays. Furthermore, MET amplification may be underestimated, because de novoMET-amplified subclones may be present in many lung cancers, but not at a frequency high enough for detection until the proper treatment-associated selective pressures are applied.29
On the basis of the aforementioned scientific rationale for this combination of tivantinib and erlotinib, there was particular interest in its use for patients with NSCLC in this study. A cohort of 8 NSCLC patients remained on therapy for a mean of 36.9 weeks (range, 1.9-93 weeks). By way of cross-trial comparison, given all the limitations of such an analysis, the median PFS in the BR.21 study was 9.7 weeks.1 In the current trial, 4 of the 5 patients tested for EGFR status were wild type (the EGFR mutation status for 3 patients was unknown), and, therefore, would generally not be predicted to have such prolonged benefit from single-agent EGFR inhibition. A further review of biomarker data suggested that patients with MET amplification may be more likely to benefit from combination EGFR and MET blockade, and that KRAS mutations may not confer primary resistance to combination treatment.
In the present study, 5 of 6 NSCLC patients previously exposed to erlotinib monotherapy achieved SD when treated with erlotinib and tivantinib combination therapy. The safety and efficacy of erlotinib and tivantinib combination therapy have been further evaluated in a randomized phase 2 trial comparing erlotinib 150 mg QD plus tivantinib 360 mg BID with erlotinib plus placebo in previously treated patients with advanced NSCLC.30 In this study, PFS in the intent-to-treat population adjusted for sex, previous chemotherapy, best previous response, and EGFR mutation status was improved in the erlotinib plus tivantinib group (hazard ratio = 0.68; 95% confidence interval, 0.47-0.98; P < .05) whereas overall survival in the intent-to-treat population was not significantly different between treatment groups. A preplanned exploratory analysis in patients with nonsquamous histology showed a trend toward improved PFS and overall survival.30 Based on these results, a phase 3 trial comparing erlotinib plus tivantinib to erlotinib plus placebo in previously treated, EGFR-inhibitor–naive patients with nonsquamous histology NSCLC was initiated and is currently enrolling patients.31
The results of this phase 1 dose-escalation study demonstrate that tivantinib can be combined safely with erlotinib. Moreover, the combination of tivantinib plus erlotinib demonstrated activity in patients who had progressed on erlotinib monotherapy. This study serves as a proof-of-concept that inhibition of multiple pathways through combination therapy provides clinical benefit and offers promise as a strategy to overcome drug resistance.
We thank Bret A. Wing, PhD, Accuverus, a division of ProEd Communications, Inc., Beachwood, Ohio, and Robert Gillespie, PhD, for their medical editorial assistance with this manuscript. We thank Beverly Fellows, ArQule, Inc., for assistance with data verification.
This study was sponsored by ArQule, Inc., Woburn, Mass. Support for medical editorial assistance with this manuscript was provided by ArQule, Inc., Woburn, Mass, and Daiichi Sankyo, Inc., Parsippany, NJ.
CONFLICT OF INTEREST DISCLOSURE
Jonathan Goldman, Isett Laux, Richard Just, and Lee Rosen have no conflicts to disclose. Feng Chai, Ronald Savage, and Dora Ferrari are employed by ArQule, Inc. Edward Garmey was an employee and held stock in ArQule, Inc. at the time that work on this manuscript was performed.