You are here

HER2 Transmembrane Domain (TMD) Mutations (V659/G660) That Stabilize Homo- and Heterodimerization Are Rare Oncogenic Drivers in Lung Adenocarcinoma That Respond to Afatinib

Journal of Thoracic Oncology, Volume 12, Issue 3, March 2017, Pages 446 - 457



Erb-b2 receptor tyrosine kinase (HER2) transmembrane domain (TMD) mutations (HER2V659E, HER2G660D) have previously been identified in lung adenocarcinomas, but their frequency and clinical significance is unknown.


We prospectively analyzed 8551 consecutive lung adenocarcinomas using hybrid capture–based comprehensive genomic profiling (CGP) at the request of the individual treating physicians for the purpose of making therapy decisions.


We identified 15 cases (0.18%) of HER2 TMD mutations (HER2V659E/D, HER2G660D) through CGP of 8551 lung adenocarcinomas. HER2 TMD mutations were mutually exclusive from HER2 kinase domain mutations and other oncogenic drivers in lung adenocarcinoma. Only two cases with HER2 TMD mutations (13%) had concurrent Erb-b2 receptor tyrosine kinase 2 gene (HER2) amplification. Structural analysis of HER2 TMD association revealed that mutations at positions V659 and G660 to the highly polar residues glutamic acid, aspartic acid, or arginine should stabilize homodimerization and heterodimerization of HER2 in the active conformation. Treatment with afatinib, a pan-HER inhibitor, resulted in durable clinical response in three of four patients with lung adenocarcinoma, with two harboring HER2V659E and one with double HER2V659E/G660R mutations. HER2 TMD mutations (V659 and G660) are found in other non-NSCLC malignancies, and analogous TMD mutations are also found in EGFR, HER3, and HER4.


HER2 TMD mutations represent rare but distinct targetable driver mutations in lung adenocarcinoma. CGP capable of detecting diverse HER2 alterations, including HER2 TMD mutations, should be broadly adopted to identify all patients who may benefit from HER2-targeted therapies.

Keywords: NSCLC, HER2 V659, HER2 G660, Transmembrane mutation, Afatinib, Actionable driver mutation.


The erb-b2 receptor tyrosine kinase (HER2 [ERBB2]) receptor tyrosine kinase (RTK) belongs to the human epidermal growth factor receptor family (HER/ErbB) and is a known oncogenic driver in multiple malignancies. Small molecule inhibitors of EGFR receptor tyrosine kinase (HER1) and anti-HER2 monoclonal antibodies have been among the most successful examples of targeted cancer therapies to date.1 The four HER family members mediate a broad spectrum of cellular responses transducing biochemical signals through lateral homodimerization and heterodimerization at the plasma membrane, whereas inactive receptors can exist in both monomeric and dimeric forms. Normally, ligand-induced conformational rearrangements and specific interaction of extracellular domains (ECDs) in the course of dimer formation lead to allosteric activation of the cytoplasmic kinase domains (KDs) and downstream signaling.2 HER2 does not bind any ligands but instead activates signal transduction primarily through a heterodimerization with HER1 and erb-b2 receptor tyrosine kinase 3 (HER3) rather than homodimerization.1 and 2 Additionally, specific protein-protein and protein-lipid interactions of single-span HER transmembrane domains (TMDs) are important for proper receptor activation and mechanism(s) that reduce or enhance such interactions (e.g., by means of mutations) and can affect downstream activity independently of KD mutations.3

HER2 kinase domain mutations have been identified in 1% to 3% of lung adenocarcinomas,2, 3, and 4 and these KD mutations have been postulated to be targetable by pan-HER tyrosine kinase inhibitors (TKIs).5, 6, and 7 The transforming ability of HER2 was first shown from the rat HER2 homolog (neu receptor) owing to a single amino acid mutation (V664E) in the TMD region and not in the KD.8 Recently, sporadic mutation in the human HER2 TMD, V659E, which is analogous to the rat V664E mutation, has been identified as a potential driver mutation in one patient with lung adenocarcinoma with clinical response to lapatinib plus chemotherapy.9 Subsequently, a single HER2V659 mutation has been reported in two lung adenocarcinoma genomic data sets.10 and 11 Mutation of the adjacent amino acid residue (G660D) in the TMD has been identified as the germline mutation in a family with familial lung adenocarcinoma.10

Here we describe the largest series to date investigating the frequency and clinical features of HER2 TMD (V659 and G660) alterations across a cohort of more than 8000 lung adenocarcinoma samples. The structural basis of TMD alterations is analyzed by molecular modeling using nuclear magnetic resonance (NMR) studies, and their clinical impact is highlighted by response to HER2-directed therapy in three patients with NSCLC with HER2 TMD mutations.


Patients and CGP

To identify cases with HER2 TMD mutations, we interrogated sequencing data from 8551 lung adenocarcinoma clinical samples in the Foundation Medicine, Inc., database that were profiled between August 2012 and February 2016. DNA was extracted from 40-μm fresh frozen paraffin-embedded sections, and comprehensive genome profiling (CGP) was performed on hybridization-captured, adaptor ligation–based libraries to a mean coverage depth of more than 650× for at least 3769 exons of 236 cancer-related genes plus 47 introns from 19 genes that are frequently rearranged in cancer, as previously described.12 Patient characteristics were abstracted from the submitted pathology reports. A subspecialty board-certified thoracic pathologist reviewed all HER2 TMD–positive cases. Additional patient characteristics and treatment outcomes for patients with HER2 TMD mutations were provided by the treating physicians. Approval for this study, including a waiver of informed consent and a Health Insurance Portability and Accountability Act waiver of authorization, was obtained from the Western Institutional Review Board (protocol no. 20152817). This approval includes review of available pathology reports as well as a collection of clinical treatment and outcomes data from the treating physician. The relationship between patient age and HER2 TMD mutational status was examined by using Student’s t test.

Surface-Based Reconstruction of the TMD Association in the HER2/HER2 and EGFR/HER2 Homodimers and Heterodimers

The alternative dimerization states of helical HER2/HER2 and EGFR/HER2 TMD homodimers and heterodimers for identified mutation forms of HER2 (V659E, V659D, G660D, V659E/V660R, V659_660VE, V664F, V665M, and I675M) were predicted by the PREDDIMER algorithm.13 and 14 Starting from the TM sequences K642–R671 of EGFR and R647–R678 of HER2 (including helical TMD flanked by short polar N- and C-terminal regions), the PREDDIMER algorithm performs scanning of possible superposition of helical surfaces (dimer conformations), estimation of their complementarity, dimer structure reconstruction followed by geometry optimization, and ranking and filtering of the results. The sequence numbering corresponds to the Swiss-Prot annotation of the human receptors P00533 and P04626. Energy relaxation of the obtained dimeric TMD structures using the N- and C-terminal so-called GG4-like motifs, which is believed to contribute to HER dimerization in active and inactive receptor states,3 and 15 was performed with Gromacs 4.6 (amber99sb-ildn force field)16 in two steps: (1) energy minimization of the dimer structures by using a steepest descend algorithm for optimization of strong overlapping and (2) 50 ps of molecular dynamics relaxation with constrained position of the backbone atoms for optimization of side chains' orientation. Temperature was maintained at 310 K with a Nose-Hoover thermostat. The van der Waals and electrostatic interactions were truncated using the twin-range 1.0/1.4-nm spherical cutoff. Molecular hydrophobicity potential maps of the molecular surfaces of the helical TMD dimer subunits with the isolines encircling hydrophobic and hydrophilic (polar) regions were obtained with the PREDDIMER program. The dimeric TMD structures were analyzed and visualized with MOLMOL.17

Patient Studies

Review of the available treatment options and detailed risk-benefit discussion were undertaken, and informed consent was obtained from all patients before initiation of treatment. Patient care was conducted in accordance with the Declaration of Helsinki. Treatment was conducted during the course of clinical care, and medications were obtained through insurance and/or the drug manufacturer and monitored in accordance with the approved label.


HER2 Transmembrane Domain Mutations Are Rare Events in Lung Adenocarcinoma

From a series of 8551 lung adenocarcinomas profiled in the course of clinical care as previously described,12 15 cases (0.18%) were found to harbor HER2 TMD mutations at amino acid residues 659 or 660. These included 11 cases with V659 mutations (eight cases with HER2V659E and three cases with HER2V659D), two cases with HER2G660D, one case with HER2V659E/G660R, and one case with an HER2 V659_660VE insertion mutation (Table 1). In addition to the 15 cases, we identified four additional cases with non-V659/G660 HER2 TMD mutations (two cases with V664F, one case with V665M, and one case with I675M). Seventy-nine percent of the patients were female. Of the nine patients with a known smoking history, five were never-smokers and three were former light-smokers. Of the 26 nucleotide substitutions, 19 were transitions and six were transversions. The fact that transitions were more frequently observed supports the observation that HER2 TMD mutations are not likely smoking induced. Among the 15 cases with HER2 V659 or G660 TMD mutations, no concurrent alterations in EGFR, anaplastic lymphoma receptor tyrosine kinase gene (ALK), ROS1, ret proto-oncogene (RET), BRAF, hepatocyte growth factor receptor kinase (MET), neurotrophic receptor tyrosine kinase type 1 gene (NTRK1), neurotrophic receptor tyrosine kinase type 2 gene (NTRK2), neurotrophic receptor tyrosine kinase type 3 gene (NTRK3), or KRAS were observed. HER2 amplification was uncommon in TMD-mutant samples, occurring in two of 15 cases (13%) (with 14 copies and 30 copies, respectively). In comparison, HER2 KD mutations, most of which were exon 20 insertions, were identified in 273 of the 8551 samples (3.2%), which is consistent with prior reports.4, 18, and 19 The average tumor mutational burden of the 15 patients with HER2 TMD-mutations was 5.0 mutations per megabase of DNA, which is low compared with the average 9.2 mutations per megabase for all lung adenocarcinomas tested (G. Frampton, Foundation Medicine, unpublished data) (see Table 1). Among the 8551 patients with lung adenocarcinomas, the median age was 64, which significantly older than the median age of 56 for patients whose tumors harbored HER2 TMD mutations (p = 0.027). In the 14 cases available for review, acinar features were identified in five (35.7%), followed by a solid component in three cases (21.4%). Micropapillary, papillary, and lepidic features were each found in two of the 14 cases (14.3%). Importantly, five of the 14 cases with HER2 TMD mutations (35.7%) were classified as invasive mucinous adenocarcinomas.

Table 1

Clinicopathologic Features of HER2 TMD Mutations in NSCLC


Case Age Sex Stage Smoking Status TMD Protein Alteration Codon Change Nucleotide Change Concurrent HER2 Amplification MAF TMB (mutations/Mb) Response to HER2 TKIa
1 62 F 4 Never-smoker V659E GTT → GAA 1976_1977TT>AA N 0.42 7.19 First-line afatinib, PR, 5 mo
2 54 M 4 Never-smoker V659E
N 0.16 3.99 Second-line afatinib, PR, 18 mo, ongoing
3 73 M 4 Never-smoker V659E GTT → GAG 1976_1977TT>AG N 0.09 3.99 Third-line afatinib, 5 mo of symptomatic improvement and metabolic response
4 53 M 4 Positive history G660D GGC → GAC 1979G>A N 0.14 3.19 Second-line afatinib, PD, 10 weeks
5 52 F NR Light smoker V659E GTT → GAA 1976_1977TT>AA N 0.70 10.81 Second-line afatinib, not yet evaluable
6 59 F 4 Minimal remote smoking history V659D GTT → GAT 1976T>A N 0.13 6.60 NA
7 69 M 4 NR V659D GTT → GAT 1976T>A Y (30 copies) 0.94 6.60 NA
8 47 F 4 Never-smoker V659D GTT → GAT 1976T>A N 0.14 5.50 NA
9 51 F NR NR V659E GTT → GAG 1976_1977TT>AG Y (14 copies) 0.45 2.20 NA
10 59 F 2 NR V659E GTT → GAG 1976_1977TT>AG N 0.11 1.10 NA
11 74 F NR NR V659E GTT → GAG 1976_1977TT>AG N 0.27 7.19 NA
12 48 F NR NR V659E GTT → GAA 1976_1977TT>AA N 0.46 7.99 NA
13 47 F 4 Minimal remote smoking history V659E GTT → GAA 1976_1977TT>AA N 0.08 2.70 NA
14 33 F NR NR V659_660VE GTG → GGGTTGAAG 1976_1979TG>GGTTGAAG N 0.2 2.20 NA
15 66 F NR Never-smoker G660D
N 0.12 3.30 NA

a Follow-up cutoff was October 1, 2016.

Note: Bold text indicates nucleotide changes.

HER2, erb-b2 receptor tyrosine kinase 2 gene; TMD, transmembrane domain; MAF, mutant allele frequency; TMB, tumor mutational burden; TKI, tyrosine kinase inhibitor; F, female; N, no; PR, partial response; M, male; PD, progressive disease; NR, not reported; NA, not applicable; Y, yes.

HER2 TMD Mutations Favor a Kinase Active Conformation

To understand the structural implications of the identified TMD mutations (see Table 1) we reconstructed the association of mutant HER2 TMD in the HER2/HER2 and EGFR/HER2 homodimers and heterodimers by molecular modeling (Fig. 1B). On the basis of previous NMR studies of wild-type TMD of HER receptors, we distinguished homodimeric and heterodimeric HER2 TMD structures by using the so-called N- and C-terminal tetrad GG4-like motifs, T652S653xxS656A657xxG660 and G668xxxG672, which are believed to contribute to the receptor dimerization in active and inactive states, respectively (Fig. 1A, C, and F and Supplementary Figs. 1 and 2).3, 12, and 20 Small residues in these motifs create weakly polar cavities that complement the surface of an adjacent TMD helix and allow the helices to approach closely, constituting “hot spots” of TMD interaction in the hydrophobic environment of the membrane (Fig. 1A, F, and G). Although folding of helical membrane proteins can be driven by van der Waals interactions alone (resulting from the excellent geometric fit of the TMD surfaces), polar interactions (e.g., between CαH and carbonyl groups, which are afforded by a GG4-like motif) and hydrogen bonding, which become much stronger in the membrane, may be the primary factor in TMD helix-helix association. As can be seen from Figure 1E to G, in addition to the polar and hydrogen bonding interactions already existing in the wild type, extra hydrogen bonds and/or salt bridges occur in the N-terminal dimerization mode for each of the identified polar mutations V659E, V659D, G660D, V659E/G660R, and V659_660VE, which are situated in or on the boundary of the polar cavity formed by the N-terminal dimerization interface. The TMD-mutant driven additional interactions stabilize the N-terminal dimerization, as does enhancement of polarity of the cavity itself, caused by the extra polar residue introduced by the TMD mutation (Fig. 1G and Supplementary Figs. 1 and 2). The predicted effect is a preferential adoption of the kinase active state, with a net result of increased downstream signaling (Fig. 1D).


Figure 1

Erb-b2 receptor tyrosine kinase 2 (HER2) transmembrane domain (TMD) mutations stabilize homodimerization and heterodimerization. (A) HER/ErbB family TMD sequence homology (bold) and shared mutational hot spots. TMD helix packing interfaces are usually covered by weakly polar small side chain residues such as G, A, S, and T (blue). (B) Wild-type HER2 homodimerization and heterodimerization. (C, left) Before ligand binding EGFR/HER2 heterodimers exist in equilibrium with monomeric receptors.3 In this inactive setting the extracellular domain (ECD) of EGFR is in an unliganded “tethered” conformation, whereas the HER2 ECD remains in an “inherently extended” (similar to ligand-bound) conformation. TMD helices are dimerized through C-terminal motifs (yellow ovals), whereas more polar N-terminal motifs (green ovals) face the lipid bilayer locally perturbed by tethered ECD, making it favorable for cytoplasmic juxtamembrane-A (JM-A) regions to be buried near the lipid polar head plane. Consequently, kinase domains (KDs) have no access to phosphorylation sites (open orange circles). (C, right) Ligand binding to EGFR causes ECD rearrangement to the extended state, in which the ECD interaction with the lipid in the vicinity of the TMD is weakened, the lipid returns to its “original” unperturbed state, and TMD dimerization switches to polar N-terminal motifs whereas the C-terminal motifs get exposed to the unperturbed (more hydrophobic) lipid environment. All these changes trigger release from the membrane and subsequent antiparallel dimerization of JM-A helices, ultimately allowing formation of asymmetric KD dimer and phosphorylation of the target tyrosine residues. (D) HER2 TMD mutations promote strong dimerization of inherently active HER2. (E) Additional intermonomeric hydrogen bonding and salt bridge formation (yellow dashed lines with green stars) caused by polar mutations which stabilize the active N-terminal conformation. (F and G, left) Ribbon structure of the EGFR/HER2 heterodimer (Protein Data Bank identifier: 2ks1) demonstrates EGFR (blue) and HER2 (pink) TMDs associate in a right-handed α-helical bundle through the N-terminal GG4-like motifs forming a weakly polar cavity on the helix surface. (F, right) Hydrophobicity map of the wild-type HER2 TMD helix, colored according to the molecular hydrophobicity potential (MHP) values, presented in cylindrical coordinates associated with the helix. Green and yellow ovals encircle the N-terminal (more polar) and C-terminal (more hydrophobic) HER2 TMD dimerization interfaces, presumably corresponding to the active and inactive receptor states, respectively. (G, right) HER2 TMD polar amino acid substitutions perturb the N-terminal TMD dimerization interfaces (see also Supplementary Figs. 1 and 2), facilitating kinase activation.


Recurrent TMD Mutations Exist Across HER Family Members in Lung Adenocarcinoma

Given the known heterodimerization and sequence homology, we investigated the presence of TMD mutations in other HER family members from our lung adenocarcinoma data set. We retrospectively reviewed a database of more than 67,000 clinical tumor samples that were subjected to comprehensive genomic profiling as previously described12 and identified rare analogous mutations in EGFR in three cases of lung adenocarcinoma, as well as analogous mutations in EGFR, HER3, and erb-b2 receptor tyrosine kinase 1 (HER4) in other tumor types (Supplementary Table 1).

HER2 TMD Mutations in Other Malignancies

A survey of the Foundation Medicine database revealed only three cases of HER2V659E (pancreatobillary, colorectal carcinoma, and NSCLC [histologic subtype not otherwise specified]), and no HER2 V659 mutations were found in The Cancer Genome Atlas (TCGA) database,21 indicating that HER2 TMD V659 mutation may be somewhat unique to lung adenocarcinoma (Supplementary Table 2). On the other hand, nine cases of HER2G660D mutation in other tumor types were identified in the Foundation Medicine database and one case of HER2G660R was identified in the TCGA database. Additionally, HER2S310F co-occurred with HER2G660D in seven of 11 cases (64%), including one of two lung adenocarcinomas (see Table 1 and Supplementary Table 2).

HER2 TMD Mutations Respond to the pan-HER Inhibitor Afatinib

On the basis of genomic profiling results, five patients with lung adenocarcinoma with tumors harboring HER2 TMD mutations were treated with the irreversible pan-HER TKI afatinib, which was selected on the basis of its predicted activity and approval for use in EGFR-mutated and squamous cell lung cancer, allowing easier authorization for off-label use. Clinical outcomes were availbale in four cases. Three patients described in the following paragraphs had clinical benefit, and a fourth patient with a tumor harboring HER2G660D was treated with second-line afatinib for 10 weeks but had progressive disease (see Table 1, case 4).

The patient in case 1 is a 62-year-old Asian female never-smoker who initially presented with cough and in whom with stage IV poorly differentiated adenocarcinoma characterized by a right upper lobe mass with associated bulky right mediastinal adenopathy and bony metastases was diagnosed. CGP of the mediastinal lymph node biopsy specimen revealed HER2V659E. On the basis of the fact that a previous report that HER2V659E could be a driver mutation in lung adenocarcinoma,9 the patient began receiving first-line afatinib, 40 mg once daily, and her chest pain resolved within 3 days, followed by improved shortness of breath after 1 week of treatment. Computed tomography of the chest after 4 weeks of therapy revealed a 34% reduction in tumor measurement by the Response Evaluation Criteria in Solid Tumors, version 1.1 (Fig. 2A and B). The patient required a dose reduction to 30 mg because of facial rash and paronychia but had a confirmed partial response (PR) lasting for 5 months of afatinib treatment and near resolution of the metastatic lesion in the right upper lobe (Fig. 2C).


Figure 2

Response to afatinib as first-line treatment in a patient with advanced treatment-naive lung adenocarcinoma harboring V659E erb-b2 receptor tyrosine kinase 2 transmembrane domain mutation. Computed tomography before treatment (A) and after 35 days (B) and 90 days of afatinib (C), demonstrating partial response to therapy with a decrease of the right upper lobe mass and mediastinal lymphadenopathy (yellow circles).


The patient in case 2 is a 54-year-old white male never-smoker with stage IV adenocarcinoma of the lung that was negative for activating EGFR mutations or ALK rearrangement. He began receiving carboplatin/docetaxel/bevacizumab with a best response of stable disease in the primary tumor but evidence of lymphangitic spread and disease progression in the lung and mediastinal lymph nodes and symptomatic deterioration that required 4 to 6 liters of oxygen (Fig. 3A). CGP of a repeat bronchoscopic biopsy specimen at a tertiary comprehensive cancer center revealed both an HER2V659E/G660R. The patient began receiving single-agent afatinib, 40 mg once daily, and achieved symptomatic improvement and a significant tumor reduction after 6 weeks. The patient has maintained an ongoing confirmed PR lasting 18 months at the time of submission of this article (Fig. 3B).


Figure 3

Response to afatinib in a patient with NSCLC with V659E/G660R erb-b2 receptor tyrosine kinase 2 transmembrane domain mutations. Pretreatment imaging highlighting a right upper lobe mass and bilateral lymphangitic disease pattern (A), with near resolution after 7 months of afatinib monotherapy (B).


The patient in case 3 is 75-year-old white male never-smoker with stage IV adenocarcinoma (malignant pleural effusion) of the lung who progressed while receiving cisplatin/gemcitabine chemotherapy. On progression with peritoneal metastasis, repeat biopsy of pleural nodules revealed HER2V659E, and he was treated with vinorelbine/trastuzumab for three cycles with a metabolic response but no significant symptom improvement (nausea and abdominal pain) followed by trastuzumab for 5 months with a stable response (Supplementary Fig. 3A). On further progression, the patient began taking afatinib, 40 mg once daily orally, with significant symptom improvement and a metabolic response lasting for 5 months (Supplementary Fig. 3B and C).


Understanding the structural implications of RTK alterations is important in guiding therapeutic options. The transmembrane domain of HER2 (amino acids 650–675) is encoded entirely by exon 17. Recent solution NMR studies followed by molecular modeling indicate that the TMD V659E mutation can stabilize the active HER2 homodimer through intermonomeric hydrogen bonding of the glutamate side chains with the opposite hydroxyl, carboxyl, and carbonyl groups, resulting in uncontrolled receptor activation and subsequent cell transformation.20 Indeed, both HER2V659E and HER2V660D mutations have been shown to stabilize the HER2 protein, slowing degradation in the plasma membrane.10 Both V659 and G660 amino acids are conserved from zebrafish to human, demonstrating their importance.10 Additionally, both HER2V659E and HER2G660D mutations have been shown to increase autophosphorylation and phosphorylation of EGFR, AKT, and mitogen-activated protein kinase when compared with wild-type HER2, and this activation is inhibited by afatinib but not by gefitinib.22

According to a recently proposed mechanism of receptor activation, HER activation requires an asymmetric configuration of two kinase domains in the receptor dimer, which can be achieved when the C-termini of the TMDs are spaced apart while the N-termini are close to each other.2 and 3 Activation is accompanied by a transition of the TMDs to an alternative dimerization mode involving a more polar N-terminal dimerization motif, which is driven by rearrangement of both ECDs and presumably mediated by a change of local properties of the lipid membrane (Fig. 1B and C).3 In a broader context, any mutation increasing the hydrophilic surface of the N-terminal dimerization motif or enhancing its polarity (often promoting intramembrane hydrogen bonding) is potentially activating, as it allows the receptor dimer to transition to an active state independently of ligand binding and regardless of the state of the extracellular domains and the surrounding membrane (see Fig. 1D and Supplementary Figs. 1 and 2). Predictably, the identified oncogenic mutations (V659E, V659D, G660D, V659E/G660R, and V659_660VE) all fall into this category, introducing more polar residues into the motif or adding polar residues near its boundary (Fig. 1F and G). Similarly, the three identified nonpolar mutations (V664F, V665M, and I675M) do not affect the polarity of the N-terminal or the C-terminal dimerization interface and therefore should not be expected to have a notable impact on receptor activation or tumorigenesis (Fig. 1G). However, we did not have treatment and clinical outcome data on these patients with nonpolar mutations. It should be kept in mind that these effects apply equally to homodimerization of HER2 and heterodimerization with other species, such as EGFR (HER1) and HER3, for which HER2 acts as a ligand-independent superactivator (Fig. 1B and C).1, 2, and 22 In addition, the activating homodimerization and heterodimerization of the mutant HER2 TMD regardless of the ECD state would predict inadequate tumor suppression by targeting the extracellular HER domain, an observation supported by the limited responsiveness to trastuzumab.

Being involved in formation of the active and inactive state of HER, both N- and C-terminal GG4-like motifs evidently have to be regions functionally sensitive to mutations that are capable of either activating the receptor or restraining it in the inactive state. An inhibitory mutation (nonnative) of EGFR has indeed been found in the C-terminal motif area.23 At the same time, a number of pathogenic amino acid substitutions on the cysteine residues causing ligand-independent activation of receptors were found in the N-terminal TMD region of other RTK representatives (e.g. from the fibroblast growth factor receptor family).24 As can be seen in Figure 1A, all the identified polar oncogenic mutations are located in a narrow region on the boundary of the N-terminal motif (toward the interior of the membrane, where polar interactions are most effective). This region can be characterized as a hot spot for gain-of-function mutations. This finding justifies a prediction that similar gain-of-function mutations (inclusively E/D amino acid substitutions in positions V651 and G652 of EGFR, A650 and G651 of HER3, I658 and G659 of HER4) enhancing TMD dimerization in certain conformation and thus activating the receptor independently of ligand binding can be found for other HER family representatives and suggests that searching for them is a promising idea for future clinical studies. This is supported by the fact that HER4 I658E has indeed been found to cause receptor activation with appropriate cellular response.25 Combined, these data suggest an oncogenic class of polar TMD mutations and that our findings are likely generalizable to the entire HER family of RTKs. Indeed, some of the predicted orthologous mutations in other HER2 family members have been identified in both NSCLC and other tumor types (Supplementary Table 1).

In our series, HER2V659E occurs in 0.09% of lung adenocarcinomas (eight of 8551), whereas other case reports have the incidence of HER2V659E ranging from 0.07% to 0.4%.10 and 11 In total, mutation at V659 accounted for 0.15% of all lung adenocarcinomas (13 of 8551) and was found infrequently in other tumor types, indicating that HER2 V659 mutation may be unique to lung adenocarcinoma. HER2G660D mutation was identified as a germline mutation in a family with familial adenocarcinoma of the lung.10 However, none of the patients in our series with G660 mutations reported a family history of lung cancer. HER2G660D was found in nine cases of other solid malignancies in the Foundation Medicine database, and one case of HER2G660R was reported in a uterine adenocarcinoma in the TCGA database.25 Whether HER2V659E, and in particular HER2G660D, are oncogenic drivers in other solid malignancies remains to be determined.

This report, together with the in vitro work by Suzawa et al.,22 provides convincing evidence that HER2 V659 and G660 mutations are rare but targetable driver mutations in adenocarcinoma of the lung. A confirmed and durable PR was observed with single-agent afatinib as either first-line or second-line treatment in cases with HER2 TMD mutation (V659E andV659E/G660R), and ongoing metabolic response and clinical benefit were observed with single-agent afatinib as third-line treatment in a case with V659E. Antitumor response irrespective of therapy timing (first- versus second- versus third-line therapy) is consistent with the characteristics of a driver mutation, in which case a response should be obtained when the target is inhibited regardless of therapy timing. Whether an antibody that inhibits ligand-induced dimerization such as trastuzumab would be effective in preventing the gain-of-function activity of such HER2 TMD mutants remains to be determined. Case 3 in this series indicated that trastuzumab resulted in minor metabolic response but minimum symptomatic benefit, whereas subsequent single-agent afatinib led to significant metabolic response and clinical symptom improvement. Case 4 in this series harbored only HER2G660D and did not respond to afatinib after 10 weeks of treatment; therefore, additional clinical data are needed to determine whether mutations at both V659 and G660 are responsive to afatinib as well as what other factors may influence response in this setting. Furthermore, determining the mechanisms of resistance to afatinib in these patients with HER2 TMD mutation will provide further mechanistic insight into HER2 TMD–driven pathogenesis.26

Besides the well-studied activating mutations in the kinase domain of the HER family member EGFR, EGFR kinase domain duplications and oncogenic EGFR fusions have recently been identified as actionable oncogenic drivers in lung adenocarcinoma.26 and 27 This report adds TMD mutations to the list of actionable oncogenic alterations in NSCLC to be screened and underscores the need for broad adoption of testing methodologies capable of detecting multiple mechanisms of oncogenic RTK activation across tumor types.

The patient responses observed in our series warrant further prospective investigation of HER2-directed therapy, either with single-agent pan-HER TKIs (i.e., afatinib, dacomitinib, and neratinib) or in combination with trastuzumab/pertuzumab in patients harboring HER2 TMD V659 and G660 mutations. Efforts such as the ongoing European Thoracic Oncology Platform (ETOP-NICHE) with afatinib (NCT02369484) are likely to refine the response (and resistance) patterns of HER2 TMD alterations. As in the case of other rare driver alterations, it is difficult to collect large prospective data sets required for regulatory approval, and this may delay access for some patients whose tumors harbor HER2 TMD mutations. In the meantime, we hope that our report raises awareness of these extremely rare but targetable mutations and can provide the most robust literature-based approached to help manage these patients.


This work was partially (structural analysis of HER2 transmembrane domain–mutant forms) supported by the Russian Science Foundation, Project No. 14-50-00131 (to Dr. Bocharov). Dr. Ou was partly supported by the University of California Irvine National Institutes of Health P30 grant CA062203-19. The authors express their sincere thanks to Drs. P.E. Volynsky, K.V. Pavlov, and K.A. Beirit for helpful discussions.

Supplementary Data


Supplemental Figure 1

Structural analysis of influence of HER2 TMD mutations on hetero-dimerization. A, NMR structure of the EGFR/HER2 TMD hetero-dimer (PDB ID: 2ks1) presumably corresponding to the receptor active state (16). Ribbon structure of the EGFR/HER2 TMD hetero-dimer, hydrophilic/hydrophobic properties and hydrophobicity map of molecular the surface of the HER2 TMD helix are presented. B, Predicted structure of the EGFR/HER2 TMD hetero-dimer presumably corresponding to the receptor inactive state. C and D, Predicted hetero-dimeric TMD structures of mutant EGFR/HER2-V659E and EGFR/HER2-G660D presumably corresponding to the receptor active state.



Supplemental Figure 2

Structural analysis of influence of HER2 TMD mutations on homo-dimerization. A, NMR structure of the HER2 TMD homo-dimer (PDB ID: 2jwa) presumably corresponding to the receptor active state (15). Ribbon structure of the HER2 TMD homo-dimer, hydrophilic/hydrophobic properties and hydrophobicity map of the molecular surface of the HER2 TMD helix are presented. B, Predicted structure of mutant HER2-V664F TMD homo-dimer presumably corresponding to the receptor inactive state. C and D, Predicted homo-dimeric TMD structures of mutant HER2-V659E and HER2-G660D presumably corresponding to the receptor active state.



Supplemental Figure 3

Response to afatinib as third-line treatment in an advanced lung adenocarcinoma patient harboring V659E HER2 TMD mutation. Computed tomography (left, yellow arrows) and FDG-PET (right, red arrows) images of A, peritoneal lesions pre-treatment (left) and after 3 cycles of trastuzumab in combination with vinorelbine (right), and B, peritoneal lesions pre-afatinib treatment (left) and post-three months of afatinib treatment (right), and C, left lower lobe pulmonary metastasis pre-afatinib treatment (left) and post-three months of afatinib treatment (right)


Download file

Supplementary Tables 1 and 2


  • 1 R. Roskoski. ErbB/HER protein-tyrosine kinases: structures and small molecule inhibitors. Pharmacol Res. 2014;87:42-59 Crossref
  • 2 E. Kovacs, J.A. Zorn, Y. Huang, T. Barros, J. Kuriyan. A structural perspective on the regulation of the epidermal growth factor receptor. Annu Rev Biochem. 2015;84:739-764 Crossref
  • 3 E.V. Bocharov, D.M. Lesovoy, K.V. Pavlov, Y.E. Pustovalova, O.V. Bocharova, A.S. Arseniev. Alternative packing of EGFR transmembrane domain suggests that protein-lipid interactions underlie signal conduction across membrane. Biochim Biophys Acta. 2016;1858:1254-1261 Crossref
  • 4 M.G. Kris, B.E. Johnson, L.D. Berry, et al. Using multiplexed assays of oncogenic drivers in lung cancers to select targeted drugs. JAMA. 2014;311:1998-2006 Crossref
  • 5 M.G. Kris, D.R. Camidge, G. Giaccone, et al. 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. Ann Oncol. 2015;26:1421-1427
  • 6 J. Mazières, F. Barlesi, T. Filleron, et al. Lung cancer patients with HER2 mutations treated with chemotherapy and HER2-targeted drugs: results from the European EUHER2 cohort. Ann Oncol. 2016;27:281-286
  • 7 P. Stephens, C. Hunter, G. Bignell, et al. Lung cancer: intragenic ERBB2 kinase mutations in tumours. Nature. 2004;431:525-526 Crossref
  • 8 C.I. Bargmann, M.C. Hung, R.A. Weinberg. Multiple independent activations of the neu oncogene by a point mutation altering the transmembrane domain of p185. Cell. 1986;45:649-657 Crossref
  • 9 V. Serra, A. Vivancos, X.S. Puente, et al. Clinical response to a lapatinib-based therapy for a Li-Fraumeni syndrome patient with a novel HER2V659E mutation. Cancer Discov. 2013;3:1238-1244 Crossref
  • 10 H. Yamamoto, K. Higasa, M. Sakaguchi, et al. Novel germline mutation in the transmembrane domain of HER2 in familial lung adenocarcinomas. J Natl Cancer Inst. 2014;106:djt338 Crossref
  • 11 R. Wang, Y. Zhang, Y. Pan, et al. Comprehensive investigation of oncogenic driver mutations in Chinese non-small cell lung cancer patients. Oncotarget. 2015;6:34300-34308
  • 12 G.M. Frampton, A. Fichtenholtz, G.A. Otto, et al. Development and validation of a clinical cancer genomic profiling test based on massively parallel DNA sequencing. Nat Biotechnol. 2013;31:1023-1031 Crossref
  • 13 A.A. Polyansky, P.E. Volynsky, R.G. Efremov. Multistate organization of transmembrane helical protein dimers governed by the host membrane. J Am Chem Soc. 2012;134:14390-14400
  • 14 PREDDIMER Prediction Tool for an Ensemble of Transmembrane α-Helical Dimer Conformations. Predict a dimer structure. Accessed November 7, 2013.
  • 15 K.S. Mineev, E.V. Bocharov, Y.E. Pustovalova, O.V. Bocharova, V.V. Chupin, A.S. Arseniev. Spatial structure of the transmembrane domain heterodimer of ErbB1 and ErbB2 receptor tyrosine kinases. J Mol Biol. 2010;400:231-243 Crossref
  • 16 B. Hess, C. Kutzner, D. van der Spoel, E. Lindahl. GROMACS 4:  algorithms for highly efficient, load-balanced, and scalable molecular simulation. J Chem Theory Comput. 2008;4:435-437
  • 17 R. Koradi, M. Billeter, K. Wüthrich. MOLMOL: a program for display and analysis of macromolecular structures. J Mol Graph. 1996;14:51-55 29–32 Crossref
  • 18 F. Barlesi, J. Mazieres, J.-P. Merlio, et al. Routine molecular profiling of patients with advanced non-small-cell lung cancer: results of a 1-year nationwide programme of the French Cooperative Thoracic Intergroup (IFCT). Lancet. 2016;387:1415-1426 Crossref
  • 19 K. Tomizawa, K. Suda, R. Onozato, et al. Prognostic and predictive implications of HER2/ERBB2/neu gene mutations in lung cancers. Lung Cancer. 2011;74:139-144 Crossref
  • 20 E.V. Bocharov, K.S. Mineev, P.E. Volynsky, et al. Spatial structure of the dimeric transmembrane domain of the growth factor receptor ErbB2 presumably corresponding to the receptor active state. J Biol Chem. 2008;283:6950-6956 Crossref
  • 21 CBio Portal for Cancer Genomics. Visualize, analyze, discover. Accessed May 2, 2016.
  • 22 K. Suzawa, S. Toyooka, M. Sakaguchi, et al. Antitumor effect of afatinib, as a human epidermal growth factor receptor 2-targeted therapy, in lung cancers harboring HER2 oncogene alterations. Cancer Sci. 2016;107:45-52 Crossref
  • 23 N.F. Endres, R. Das, A.W. Smith, et al. Conformational coupling across the plasma membrane in activation of the EGF receptor. Cell. 2013;152:543-556 Crossref
  • 24 E.V. Bocharov, D.M. Lesovoy, S.A. Goncharuk, M.V. Goncharuk, K. Hristova, A.S. Arseniev. Structure of FGFR3 transmembrane domain dimer: implications for signaling and human pathologies. Structure. 2013;21:2087-2093 Crossref
  • 25 G.A. Vidal, D.E. Clark, L. Marrero, F.E. Jones. A constitutively active ERBB4/HER4 allele with enhanced transcriptional coactivation and cell-killing activities. Oncogene. 2007;26:462-466 Crossref
  • 26 J.N. Gallant, J.H. Sheehan, T.M. Shaver, et al. EGFR kinase domain duplication (EGFR-KDD)i a novel oncogenic driver in lung cancer that is clinically responsive to afatinib. Cancer Discov. 2015;5:1155-1163 Crossref
  • 27 K. Konduri, J.N. Gallant, Y.K. Chae, et al. EGFR fusions as novel therapeutic targets in lung cancer. Cancer Discov. April 21, 2016;6:601-611


a Chao Family Comprehensive Cancer Center, University of California Irvine School of Medicine, Orange, California

b Foundation Medicine, Inc., Cambridge, Massachusetts

c Department of Structural Biology, Shemyakin-Ovchinnikov Institute of Bioorganic Chemistry, Russian Academy of Sciences, Moscow, Russian Federation

d The Angeles Clinic and Research Institute, Los Angeles, California

e Hospital Sao Jose, Sao Paulo, Brazil

f Affiliated Oncologists, LLC, Oak Lawn, Illinois

g Sarah Cannon Research Institute, Tennessee Oncology, Nashville, Tennessee

h University of Southern California Norris Comprehensive Cancer Center, Los Angeles, California

i Hospital Israelita Albert Einstein, Sao Paulo, Brazil

j Albany Medical College, Albany, New York

k Cleveland Clinic, Cleveland Clinic Main Campus, Cleveland, Ohio

Corresponding author. Address for correspondence: Sai-Hong Ignatius Ou, MD, PhD, Department of Medicine, Division of Hematology-Oncology, Chao Family Comprehensive Cancer Center, University of California Irvine School of Medicine, 101 City Dr., Bldg. 56, RT 81, Rm. 241, Orange, CA 92868.

Readers of this article may receive CME credit. Further information can be found at

Drs. Ou, Schrock, and Bocharov equally contributed to this work.

Disclosure: Dr. Ou has received honoraria as an advisory board and speaker bureau member for Boehringer Ingelheim. Drs. Schrock, Chung, Campregher, Ross, Stephens, Miller, Suh, and Ali are employees of and have equity interest in Foundation Medicine, Inc. Dr. Klempner has received honoraria from Foundation Medicine. The remaining authors declare no conflict of interest.

Commentary by Tom Stinchcombe

With the development of widespread next generation sequencing and the development of large databases, it is now possible to detect rare mutations, and compile the efficacy results. In this study comprehensive genomic profiling (CPD) was performed 8,551 consecutive lung adenocarcinomas. HER2 transmembrane mutations at V659/G660 were detected 15 cases (0.18%). These mutations stabilize homodimerization and heterodimerization of HER2 in the active conformation leading constitutive activation. In a retrospective review four patients were treated were treated with afatinib and had response data available. One patient received afatinib for 10 weeks and had progressive disease, one patient had partial response lasting 5 months, one patient has a partial response lasting 18 months, and one patient had a metabolic response lasting 5 months. 

This study demonstrated the value of CPD in identifying rare mutations that are susceptible to targeted therapy. Afatinib has demonstrated some activity in this patient population, but hopefully other agents will greater activity.