HER2 Mutations in Non–Small Cell Lung Cancer: A Herculean Effort to Hit the Target

The HER2/neu (encoded by ERBB2) gene is one of the most extensively studied oncogenes in solid tumors, and HER2 deregulation can occur through multiple mechanisms that include gene amplification, protein overexpression, and HER2 mutations. Over the last two decades, several authors have focused their research interest on HER2 aberrations in advanced non–small cell lung cancer (NSCLC), evaluating the potential clinical implications on sensitivity to anti-HER2 agents with conflicting results (1). Following the successful history of trastuzumab in HER2-overexpressing (IHC 3+) and/or amplified breast cancer, the development of HER2-targeting agents in lung cancer has focused on protein expression by IHC or HER2 amplification by FISH. The lower frequency in NSCLC (approximately 2% vs. 20% in breast cancer) prompted the evaluation of trastuzumab plus chemotherapy in patients with low-to-intermediate IHC expression, with generally modest results (1). More recently, the antibody–drug conjugate (ADC) ado-trastuzumab emtansine (T-DM1) was evaluated in HER2 IHC-positive NSCLC (IHC 2+ or 3+), showing only moderate activity in HER2-overexpressing tumors (2, 3). These data further confirm that IHC expression is not a definitive biomarker for anti-HER2 activity in NSCLC and suggest the need for additional strategies to refine the target population for these agents.

Over the last several years, HER2 mutations have emerged as a novel potential target for anti-HER2 agents in NSCLC, identifying a relatively small subgroup of patients (approximately 3%) with peculiar clinicopathologic characteristics that resemble those of other molecularly defined subgroups of patients with NSCLC (adenocarcinoma histology and never/light smoking history), although tissue testing for HER2 mutations should not be limited to these categories (1). Since the first report of Cappuzzo and colleagues demonstrating the actionability of HER2 mutations in lung cancer with a sustained response to trastuzumab-based chemotherapy in a heavily pretreated HER2-mutated patient (4), different therapeutic strategies have been tested in this subgroup of patients with variable results (Fig. 1). As a consequence, to date, no targeted agents are approved in this subgroup of patients.

However, the term “HER2 mutations” encompasses a large number of mutations that can occur in the extracellular, transmembrane/juxtamembrane or kinase domain, and the location of the mutation appears to be associated with the tumor histology (5). HER2 kinase domain mutations are the most commonly observed in NSCLC, either as insertions or missense mutations. The most common variant is the YVMA mutation that involves a 12–base pair in-frame insertion in exon 20 at codon 775. Less frequently, the mutations can be observed single amino-acid substitutions in exon 20 or in other exons (such as 8, 17, or 19). Some studies with pan-HER inhibitors have suggested a potential differential impact of HER2 mutation variants on therapeutic activity of anti-HER2 agents. This clinical evidence was recently confirmed in preclinical studies demonstrating that common HER2 mutants yielded differential sensitivities to multiple pan-HER tyrosine kinase inhibitors (TKI) in vitro, with a reduced drug-binding pocket volume associated with decreased affinity for larger TKIs. Across all HER2 mutants and TKIs tested, poziotinib has the lowest average IC₅₀ and is significantly more effective than afatinib, neratinib, or tarloxotinib for mutations in exons 19 and 20, with no significant difference in average IC₅₀ for exon 21 mutants (6). In contrast, the ADC T-DM1 seems to be associated with activity regardless of the HER2 mutation type (exon 20 insertions or transmembrane and extracellular domain point mutations), as recently reported in a phase II basket trial (7).

In this issue of Cancer Discovery, two studies further expand the knowledge on therapeutic exploitation of HER2 in lung cancer. Li and colleagues demonstrate that HER2 phosphorylation, ubiquitination, and internalization rate, rather than its overexpression, are key mechanisms underlying the endocytosis and consequent efficacy of the anti-HER2 ADCs T-DM1 and trastuzumab deruxtecan (T-DXd, also known as DS8201a) in lung cancer cell lines and patient-derived xenograft models (8). These preclinical data were confirmed clinically through the analysis of the molecular basis of T-DM1 efficacy in 49 patients with HER2-mutated or HER2-amplified NSCLC, in part enrolled into a phase II basket trial (7). The same extent of response in lung cancers was observed with either amplified or mutant HER2 [overall response rate (ORR) 55% for HER2-amplified patients, 50% for HER2-mutated patients, and 50% for concurrently HER2-amplified and -mutated patients; ref. 8]. Interestingly, the combination of T-DM1 plus a pan-HER inhibitor (neratinib) or switching to another ADC with a different chemotherapy payload (T-DXd) resulted in superior activity compared with T-DM1 monotherapy (8). Similarly, recent studies have reported that the novel pan-HER TKI poziotinib upregulates HER2 cell-surface expression and potentiates the activity of T-DM1 in HER2-mutated preclinical models (6).

In a second article accompanying this article, Tsurutani and colleagues report on the safety and efficacy of T-DXd in advanced non–breast cancer solid tumors (9). Results presented here represent a major step forward in HER2 targeting, especially in HER2-mutated NSCLC where no approved treatment options are available at the moment, and the use of the pan-HER inhibitors afatinib, dacomitinib, and neratinib showed only modest activity [ORR ranging from 0% to 19% and median progression-free survival (PFS) of 2.9–5.5 months; Fig. 1]. In contrast, the use of trastuzumab-based therapies (either with chemotherapy or in combination with pertuzumab) and the ADC T-DM1 have been associated with improved outcomes (ORR 14.8%–50.9%, median PFS 4.8–5.0 months; Fig. 1) and seem to be a more valuable therapeutic option compared with irreversible EGFR/HER2 TKIs. T-DXd represents a potential novel therapeutic option in this setting with ORR (approximately 70%) and PFS (approximately 11 months) comparable to those seen in other oncogene-addicted NSCLCs treated with targeted therapies, despite the high number of previous therapies received (≥5 prior anticancer regimens in approximately one third of patients; ref. 9). Clinical activity was also reported in T-DM1–resistant patients, and approximately one third of these patients had been treated with anti-HER2 agents before trial enrollment.

T-DXd is currently under evaluation in multiple phase I/II studies in NSCLC either as monotherapy (NCT03505710) or in combination with immune checkpoint inhibitors (NCT04042701). In addition, several novel TKIs are currently under evaluation in HER2-mutated and/or -amplified NSCLCs, such as pyrotinib (NCT04063462; NCT04144569), tarloxotinib (NCT03805841), TAK-788 (NCT02716116), and poziotinib (NCT03318939; NCT04044170), although toxicity might represent a major issue with some of these agents.

Despite promising activity, some open questions still remain unanswered. HER2-mutated NSCLCs are associated with central nervous system (CNS) involvement during the course of the disease in approximately half of the patients. The activity of T-DXd in patients with brain metastases should be further explored in this context, as T-DM1 CNS activity is well known in breast cancer. Another issue is the clinical challenge of the not-infrequent situation of concomitant mutations in HER2 mutants (2), which can impair the activity of anti-HER2 agents and affect prognosis. Furthermore, HER2 mutations and/or HER2 amplification have been reported as a mechanism of acquired resistance to EGFR TKIs in NSCLC. Data on concomitant genetic aberrations in the study by Tsurutani and colleagues, with the exception of KRAS mutations, were not reported, although some patients received prior EGFR or ALK inhibitors before study entry (9). The activity of T-DXd in these cases might provide the rationale for therapeutic exploitation and expansion of this agent to treat TKI-resistant patients with acquired HER2 bypass-signaling mechanisms of resistance.

One of the most clinically relevant findings of the work of Tsurutani and colleagues is the potential risk of interstitial lung disease. Whether this adverse event is associated with preexisting pulmonary conditions, previous radiotherapy-induced lung damage, or with ethnicity should be further evaluated in future studies.

Finally, the activity seen in HER2-expressing nonbreast/nongastric and/or HER2-mutant solid tumors treated with T-DXd is noteworthy (9) and might expand the use of anti-HER2 blockage beyond currently approved indications, including tumor types with limited therapeutic options, such as biliary tract and salivary gland cancers. Similarly, another novel ADC, trastuzumab duocarmazine (also known as SYD985), showed intriguing clinical activity in a phase I clinical trial in heavily pretreated patients with HER2-expressing metastatic breast cancer, including HER2-positive T-DM1–resistant and HER2-low patients, as well as other solid tumors (10).

As the use of next-generation sequencing (NGS) is constantly expanding, the identification of HER2 mutations is consequently destinated to grow, especially in tumor types where complete tumor genotyping is necessary for optimal treatment selection, as in NSCLC. However, the remarkable activity seen in the study by Tsurutani and colleagues suggests that HER2 mutations might expand the list of current tumor-agnostic biomarkers beyond microsatellite instability and neurotrophic tyrosine receptor kinase (NTRK) gene fusions, potentially affecting outcomes in a large number of patients for whom no targeted therapeutic strategies are approved to date. Unfortunately, all of these Herculean efforts, if not accompanied by increasing efforts to expand molecular testing in clinical practice through tissue NGS and/or plasma cell-free DNA analyses, will be applicable only for a small portion of patients, limiting access to a highly effective therapy. Having promising drugs for largely untested targets is like having a Ferrari without a driving license.

C. Rolfo has received speakers bureau honoraria from AstraZeneca and MSD, has an advisory board role at Inivata, ARCHER, MD Serono, and Mylan, has an advisory board role for Oncompass (non-financial), and has received honoraria from Elsevier. No potential conflicts of interest were disclosed by the other author.