Targeting HER2 aberrations as actionable drivers in lung cancers: phase II trial of the pan-HER tyrosine kinase inhibitor dacomitinib in patients with HER2-mutant or amplified tumors

introduction

Aberrations in human epidermal growth factor receptor 2 (HER2, ERBB2) have emerged as oncogenic drivers and therapeutic targets in lung cancers, with HER2 mutations occurring in ∼1%–4% [1–5], and HER2 amplifications occurring in 2%–5% [6–8] of lung adenocarcinomas. Preclinical studies have described the antitumor activity of HER2-targeted therapies, inhibiting both HER2-mutant and HER2-amplified lung cancer cell lines [9, 10], including mutations in the HER2 extracellular domain [11]. The growing use of multiplexed genomic assays that can detect both mutations and amplifications has led to the identification of an increasing number of individuals with HER2-driven lung cancers [12]. The availability of HER2-targeted agents makes HER2 an ‘actionable’ target, further spurring the implementation of assays to detect it. Several reports have demonstrated the activity of HER2-directed therapies in patients with HER2-mutant [13–16] and HER2-amplified [15] lung cancers.

Dacomitinib (PF-00299804; Pfizer, New York, NY) is a pan-HER inhibitor that irreversibly binds to HER2, HER1 (EGFR), and HER4 tyrosine kinases [17, 18]. We have reported the results of the phase II trial of dacomitinib in patients with lung adenocarcinomas clinically and/or molecularly selected for the presence of epidermal growth factor receptor (EGFR) mutations [19]. As part of this larger effort, we studied a separate molecularly selected cohort of individuals with tumors harboring HER2 mutations or amplification.

methods

results

We enrolled and treated 30 individuals with pathologically confirmed recurrent or de-novo lung cancers between January 2011 and February 2013. Patient characteristics are detailed in supplementary Table S1, available at Annals of Oncology online. Half were women, 60% were never smokers, and all but two had stage IV disease. Dacomitinib was the first systemic therapy for 17% and 83% had received at least one prior i.v. cytotoxic chemotherapy. Two patients (7%) were previously treated with trastuzumab. Twenty-eight patients were enrolled from Lung Cancer Mutation Consortium sites.

The specific aberrations detected in the 26 patients with HER2 mutations are detailed in Table 1. Thirteen tumors harbored identical 12-bp exon 20 insertions [A775_G776insYVMA, alternative nomenclature p.Y772_A775dup(c.2313_2324dup)]. There were four 9-bp insertions and three 3-bp insertions. Two patients had HER2 exon 20 indels and one a missense mutation. Mutation details were not specified in three cases with exon 20 insertions. Data on the four HER2-amplified cases is detailed in supplementary Table S2, available at Annals of Oncology online. Only one case had high-level amplification. Of the 10 patients with HER2 mutations that were also tested for HER2 amplification, none were amplified.

Confirmed partial responses were seen in three patients, two with tumors harboring a 9-bp exon 20 insertion, lasting 11 and 14 months and one with a tumor harboring an exon 20 indel lasting 3+ months. The two patients with the exon 20 insertions that had the longest responses had identical amino acid changes (p. P780_Y781insGSP, Table 1). The overall response for the patients with HER2-mutant disease was 12% (95% CI 2% to 30%) and 0% (95% CI 0% to 60%) for the four patients with HER2-amplified tumors. Figure 1 is a waterfall plot depicting the greatest change in target lesion size for each patient.

For the HER2-mutant cohort, the median progression-free survival was 3 months (95% CI 2–4 months). The four individuals with HER2-amplified tumors remained progression free for 1, 1, 5, and 5 months. The median overall survival was 9 months from the start of dacomitinib (95% CI 7–21 months) for patients with HER2-mutant cancers and their 1-year survival was 44% (95% CI 25–62). Overall survivals were 5, 7, 15, and 22 months for the individuals with HER2-amplified tumors. The patient with the highest degree of amplification (HER2/CEP17 ratio 17) had the longest survival. Kaplan–Meier survival curves from the start of dacomitinib for the HER-mutant and HER2-amplified groups are shown in supplementary Figure S1, available at Annals of Oncology online.

Treatment-related toxicities are listed in supplementary Table S3, available at Annals of Oncology online, reporting the maximum toxicity seen at any time while on study for all 30 patients. Some diarrhea was seen in 90% and skin rash in 73%. There was one case of grade 4 diarrhea and no grade 4 rash. Rash and diarrhea resolved with either supportive measures or dacomitinib cessation. Due to treatment-related toxicity, the dose of dacomitinib was reduced in 17% and in 13%, dacomitinib was stopped. There was one drug-related death. This individual developed hepatic failure likely due to an interaction of mirtazapine and dacomitinib. Mirtazapine is metabolized by CYP2D6 and dacomitinib is a strong inhibitor of the same cytochrome. Mirtazapine was started 6 days after the initiation of dacomitinib. Pharmacokinetic evaluations after development of liver failure revealed plasma mirtazapine levels increased 26-fold over expected. The patient died 13 days after starting mirtazapine.

discussion

This is the first phase II trial of a HER2-targeted agent in patients with lung cancers molecularly selected for study based on the presence of HER2 mutations or amplification in their tumors. Durable partial responses were documented in 12% of individuals with HER2 mutations using dacomitinib, an inhibitor of the HER2 tyrosine kinase.

Stephens et al. [1] and Shigematsu et al. [2] independently identified mutations in HER2 in lung cancers that were mutually exclusive with EGFR and KRAS. Reports exploring the association between HER2 amplification and HER2 mutation in lung cancers have yielded divergent results [3, 22]. Kinase domain mutations, mainly exon 20 insertions and point mutations [3] lead to constitutive HER2 kinase activation which is felt to underlie the sensitivity of cancers driven by these mutations to kinase inhibitors [9]. Subsequently, there have been several reports that HER2 kinase inhibitors, including afatinib [13, 14], lapatinib [23], neratinib [24], and neratinib plus temsirolimus [16, 24], can lead to regressions in patients with tumors harboring these mutations. Cappuzzo et al. [15] also reported a partial response in a patient with lung cancer harboring both a HER2 exon 20 (G776 > LC) mutation and HER2 amplification, following treatment with the combination of trastuzumab and paclitaxel. The clinical series by Mazieres et al. [13] reported 10 of 16 patients with tumors with YMVA HER2 exon 20 insertions experienced partial responses to HER2-targeted therapies, 9 of them with concurrent chemotherapy and 8 with trastuzumab. One partial response with trastuzumab alone was reported. Tomizawa reported a response with trastuzumab plus vinorelbine in a similar patient [25]. While a randomized phase II trial by Gatzemeier et al. [21] concluded that the addition of trastuzumab to chemotherapy produced no added activity in HER2 protein overexpressed lung cancers, the six patients with either 3+ HER2 overexpression or HER2 amplification had a response rate of 83%. A trial of ado-trastuzumab emtansine (TDM-1) in patients with lung cancers with HER2 protein overexpression is underway (ClinicalTrials.gov, NCT02289833). Reviews of this topic have recently been published [26, 27]. But despite these reviews and series, no prospective trials have explored single-agent HER2-targeted agents in patients molecularly selected by the presence of HER2 mutations or amplifications. The majority of the responses reported were using a HER2-targeted agent with chemotherapy.

The activity of dacomitinib in patients with HER2 mutations in this trial confirms the earlier reports of benefit with HER2 kinase inhibitors in similar patients. While the responses all provided clinical benefit, they were few in number and were seen only in two of four patients with 9-bp HER2 exon 20 insertions and one of two patients with indels in HER2 exon 20. Preclinical studies demonstrate that a number of HER2 mutations are oncogenic both in vitro and in vivo, similar to EGFR mutations [9, 28]. The sensitivity to dacomitinib among the HER2 mutations may vary, in part explaining the clinical findings. The most common HER2 mutation is located in the analogous position as EGFR exon 20 insertion mutations and some but not all of these are also dacomitinib sensitive. No partial responses were documented here in patients with the most common HER2 mutation, a 12-bp insertion in exon 20 (p. A775_G776insYVMA (c. 2324_2325ins12). In contrast, the responses reported with afatinib were described in patients with tumors with the YVMA exon 20 insertion found in 13/26 of the patients in this trial [13, 14]. These observations suggest that the benefits of an individual agent may be confined to specific aberrations and the term ‘HER2-positive lung cancers’ does not adequately define these illnesses. The specific type of HER2 mutation, presence and degree of HER2 amplification, and HER2 protein expression should be precisely defined for each patient in future studies of HER2-targeted agents.

Advances in technology have made it possible to routinely identify oncogenic drivers in patients with lung cancers using multiplex panels [29] which are today recommended for use in all patients with adenocarcinomas in treatment guidelines [30]. The Lung Cancer Mutation Consortium in the United States and the French Cooperative Thoracic Intergroup have detected mutually exclusive HER2 mutations in 3% (95% CI 2% to 4%) and 1% (95% CI 1% to 2%) of tumors from patients with lung cancers respectively [4, 5]. Next-generation sequencing (NGS) is quickly replacing current multiplex panels. With regard to HER2, in addition to identifying mutations in exon 20, NGS panels can assess extracellular domain mutations and the presence of amplifications [12]. These expanding capabilities are increasing the number of patients with HER2 aberrations that can treated by both approved and investigational agents.

The data reported here have limitations. This was an exploratory study with no predefined primary end point, study outcome, or sample size specific to this HER2 cohort. The end points chosen for this report were the same as for the EGFR cohort previously reported [19]. Because we were uncertain about the frequency of HER2 mutations and amplifications, no sample sizes were prespecified for either group. Central confirmation of molecular abnormalities was not mandated since HER2 amplification testing was a standard test for breast cancer specimens and multiplex testing for oncologic drivers carried out within CLIA standards were in place at all study sites. Tissue was not collected to determine HER2 protein expression.

These results with dacomitinib emphasize both the promise and pitfalls of testing drugs in patients with tumors harboring HER2 aberrations. While durable responses are documented, their frequency is a fraction of what we have observed with kinase inhibitors targeting EGFR and ALK in mutation-positive cases where response rates in excess of 50% are seen, including a 76% partial response rate and 18 months progression-free survival with dacomitinib in patients with sensitizing EGFR mutations [19]. Beyond the diversity of the specific molecular aberrations in HER2 in lung cancers, the variable effectiveness of HER2 kinase inhibitors may also reflect the inherent activity of the drugs themselves. Although we demonstrated partial responses here with dacomitinib in HER2-mutant lung cancers, no responses were reported with the HER2 kinase inhibitor neratinib in similar patients [24]. Care must be taken before concluding that the results for any one agent define the overall effectiveness of ‘HER2-targeted therapies’ in lung cancers. Future trials must assess adequate numbers of cases with precise molecular alterations for each agent under study. The relationships among HER2 amplification, HER2 protein expression, and HER2 mutation are largely unknown. The impact of these abnormalities on therapeutic efficacy of HER2-targeted agents either alone or in the context of HER2 mutations is equally unclear. Activation of other pathways, particularly PI3KCA/mTOR, may also play a role as suggested by the observation that the combination of the mTOR inhibitor temsirolimus plus neratinib (a HER2 kinase inhibitor) produced a higher response rate in patients with HER2-mutant lung cancers than with neratinib alone [24]. Further complicating the picture is the fact that in other cancers, HER2-targeted antibodies show greater clinical benefit than tyrosine kinase inhibitors in HER2-amplified cancers despite being far weaker inhibitors of oncogenic HER2 signaling [31]. These knowledge gaps for an important target in a rapidly growing patient group make further research and comprehensive assessment of HER2 mutation, HER2 copy number, and HER2 protein expression in all cases enrolled on trials of HER2-targeted agents.

This study also emphasizes the complexity of determining the existence of ‘actionable’ mutations in lung cancers. Precise molecular targets for ‘actionability’ require that they are of sufficient frequency to make their study in prospective clinical trials feasible. If the activating mutations of EGFR had represented only 15% of all EGFR mutations, rather than the 80+ percent that have been found, could clinical trials detect a ‘signal’ of sufficient clarity to justify drug development? The challenges and resources to identify a drug and prove that it has meaningful clinical activity are substantial. Tracking an activity ‘signal’ in a ‘segment of a segment’ of a molecularly defined patient population could prove to be challenging in today's drug development and testing environment. A commitment to sharing precise molecular profiles of tumors using a standardized nomenclature and a searchable database that bridges institutional and pharmaceutical company efforts could go a long way in unraveling these complexities.

This phase II trial of dacomitinib demonstrated that trials of agents targeting HER2 in patients with lung cancers molecularly selected for tumors with HER2 aberrations are possible and can lead to meaningful responses, but only for selected genotypes. We advocate that single-agent activity for HER2-targeted therapies be documented before proceeding to combination trials and that study of trastuzumab, pertuzumab, and TDM-1 be pursued in patients with HER2-amplified tumors. The results presented here can help guide future studies and further raise the hope that the significance of HER2 as a target in lung cancers will continue to grow, both in the numbers of individuals treated and the magnitude of the benefits achieved.

funding

This work was supported by Pfizer, Inc. No grant applied.

disclosure

MGK has received consultancy fees from AstraZeneca, Clovis Oncology, Daiichi Sankyo, and Genentech/Roche; BTL has received consultancy fees from Roche and Biosceptre International. JO, IT, HZ, and ZG are employees and stockholders of Pfizer. MEA has received consultancy fees from AstraZeneca. PAJ has received consultancy fees from Pfizer, AstraZeneca, Boehringer Ingelheim, Clovis Oncology, Chugai, and Genentech; DRC, GG, and TH declare no competing interests.

Supplementary Material