If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Department of Medical Oncology, Sun Yat-sen University Cancer Center, Guangzhou, People’s Republic of ChinaState Key Laboratory of Oncology in South China, Collaborative Innovation Center for Cancer Medicine, Sun Yat-sen University Cancer Center, Guangzhou, People’s Republic of China
Department of Medical Oncology, Sun Yat-sen University Cancer Center, Guangzhou, People’s Republic of ChinaState Key Laboratory of Oncology in South China, Collaborative Innovation Center for Cancer Medicine, Sun Yat-sen University Cancer Center, Guangzhou, People’s Republic of China
Department of Medical Oncology, Sun Yat-sen University Cancer Center, Guangzhou, People’s Republic of ChinaState Key Laboratory of Oncology in South China, Collaborative Innovation Center for Cancer Medicine, Sun Yat-sen University Cancer Center, Guangzhou, People’s Republic of China
Department of Medical Oncology, Sun Yat-sen University Cancer Center, Guangzhou, People’s Republic of ChinaState Key Laboratory of Oncology in South China, Collaborative Innovation Center for Cancer Medicine, Sun Yat-sen University Cancer Center, Guangzhou, People’s Republic of China
Department of Medical Oncology, Sun Yat-sen University Cancer Center, Guangzhou, People’s Republic of ChinaState Key Laboratory of Oncology in South China, Collaborative Innovation Center for Cancer Medicine, Sun Yat-sen University Cancer Center, Guangzhou, People’s Republic of China
State Key Laboratory of Oncology in South China, Collaborative Innovation Center for Cancer Medicine, Sun Yat-sen University Cancer Center, Guangzhou, People’s Republic of China
Corresponding author. Address for correspondence: Li Zhang, MD, Department of Medical Oncology, Sun Yat-Sen University Cancer Center, 651 Dongfeng East Road, Guangzhou 510060, People’s Republic of China.
Affiliations
Department of Medical Oncology, Sun Yat-sen University Cancer Center, Guangzhou, People’s Republic of ChinaState Key Laboratory of Oncology in South China, Collaborative Innovation Center for Cancer Medicine, Sun Yat-sen University Cancer Center, Guangzhou, People’s Republic of China
HER2 exon 20 insertion (ex20ins) is one of the most intractable problems in lung cancer. Most ex20ins are resistant to available EGFR or pan-HER tyrosine kinase inhibitors (TKIs), with the exception of a few mutants. However, the mechanism for TKI response and resistance of HER2 ex20ins remains poorly understood.
Methods
Next-generation sequencing-based genomic profiling data of 4139 patients with lung cancer were interrogated for HER2 ex20ins. Structural modeling and molecular dynamics simulations of common HER2 ex20ins were carried out to provide insights into the mechanism of activation and response heterogeneity of ex20ins. Molecular docking was performed to predict affinity to TKIs. Therapeutic decisions for patients were made on the basis of the results of genomic profiling.
Results
From 155 HER2-mutant lung cancer cases, Y772_A775dup and G778_P780dup were identified in 74 (47.7%) and 18 (11.6%) cases, respectively. Molecular dynamics simulations revealed that HER2 ex20ins led to ligand-independent kinase activation by changing the conformational landscape of HER2 kinase and restricting kinase conformation in the active state. G778_P780dup had a three-amino acid extension in the αC-β4 loop and retained the HER2-characteristic G776 and G778. Compared with Y772_A775dup, it had less restriction on kinase conformational sampling and higher affinity to afatinib, dacomitinib, pyrotinib, and poziotinib. Treating lung adenocarcinomas carrying G778_P780dup with these inhibitors led to sustained tumor responses in six of the 10 patients.
Conclusions
The kinase conformational landscape dictated by the length of the αC-β4 loop and residues at HER2 776 and 778 position explains TKI sensitivity in ex20ins. This finding could guide therapeutic decisions with currently available therapies and future drug development strategies.
Exon 20 insertions (ex20ins) that occur predominantly in the αC-β4 loop of EGFR or HER2 kinase represent 10% to 15% of EGFR-mutant lung cancers and 90% of HER2-mutant lung cancers.
Recently, novel exon 20-selective inhibitors such as poziotinib and TAK-788 have revealed promising efficacy in EGFR ex20ins, but few responses were documented in HER2 insertions.
First report of safety, PK, and preliminary antitumor activity of the oral EGFR/HER2 exon 20 inhibitor TAK-788 (AP32788) in non–small cell lung cancer (NSCLC).
HER2 exon 20 insertion remains an intractable problem in the treatment of lung cancer.
Despite the overall poor response to kinase inhibitors, some TKI-sensitive ex20ins have been identified. EGFR A763_Y764insFQEA is erlotinib-responsive.
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.
We recently reported that HER2 G776delinsVC and G778_P780dup were afatinib-sensitive, whereas the most common exon 20 insertion in HER2, Y772_A775dup, was resistant to afatinib and dacomitinib.
Understanding the mechanism behind TKI response and resistance in ex20ins will guide therapeutic decisions for this population using currently available drugs and provide implications for future drug development strategies. An earlier study has proposed that G770 in EGFR and G778 in HER2 were features for identifying TKI-responsive insertions.
The structural analysis of EGFR A763_Y764FQEA suggests that its responsiveness is attributed to the insertion of a glutamic acid (E), which allows an N-terminal shift of the αC-helix and maintains the intactness of the αC-β4 loop.
However, no HER2 exon 20 insertion identified thus far carries a glutamic acid molecule, suggesting that TKI-responsive HER2 insertions may adopt a different mechanism.
In this study, to provide structural insights into the mechanism of activation and response heterogeneity of ex20ins, we carried out structural and molecular dynamics (MD) analysis of two common HER2 insertions, Y772_A775dup and G778_P780dup. Molecular docking was performed to assess their affinity to afatinib, dacomitinib, pyrotinib, and poziotinib. Findings in structural analysis are supported by clinical outcomes in patients carrying HER2 ex20ins treated with these inhibitors.
Materials and Methods
Patients and Tumor Genotyping
To identify HER2 ex20ins, we interrogated the sequencing data of 4139 samples of lung cancer from the OrigiMed Inc. database (n = 2035) and Sun Yat-Sen University Cancer Center (SYSUCC) (n = 2104). Samples at OrigiMed Inc. database were derived from patients with stage IV lung cancer who underwent a next-generation sequencing-based genomic testing (OrigiMed, Shanghai) during routine clinical care as previously described.
Samples at SYSUCC were derived from patients who were enrolled for genomic profiling under an institutional review board-approved, prospective protocol for the SYSUCC Personalized Therapy Program for advanced lung cancer (ChiCTR1900027582) between October 2016 and October 2019.
Tumor samples were collected by means of computed tomography (CT)-guided biopsy or bronchial biopsy. DNA was extracted from paraffin-embedded sections and the matched blood samples. Genomic profiling was performed using a hybrid capture-based next-generation sequencing-based platform at Beijing Genomics Institute (n = 659),
Clinical significance of circulating tumor cells in predicting disease progression and chemotherapy resistance in patients with gestational choriocarcinoma.
with a mean coverage depth of greater than 800× for 22 to 450 lung cancer related genes.
Structural Modeling and MD Simulations
To investigate the mechanism of TKI resistance and the response of HER2 ex20ins, structure models of HER2 Y772_A775dup and G778_P780dup were generated on the basis of the crystal structure of the human HER2 kinase (PDB ID: 3PP0). Homology modeling was performed using the Molecular Operating Environment software version 2019.01 (Chemical Computing Group, Montreal, Canada).
which calculated the root-mean-square deviation (rmsd) value for superimposition between the structure models and the template. The representative homology modeling result, which had the highest agreement with the template backbone, was used as the starting conformation in the following-MD simulations.
MD simulations were performed using AMBER16 (University of California, San Francisco).
AMBER, a package of computer programs for applying molecular mechanics, normal mode analysis, molecular dynamics and free energy calculations to simulate the structural and energetic properties of molecules.
Sodium or chloride ions were applied to neutralize the charge of the system. The simulations started with a 5000-step optimization of the solvent and side chains of the protein, followed by a 5000-step nonrestrained optimization. The Langevin thermostat was employed under the canonical ensemble (NVT ensemble) to heat up the system from 0 K to 300 K in 200 ps.
Afterward, another 200 pico seconds NPT ensemble equilibrium process was carried out under the Berendsen barostat to balance the total density of the system (to 1 kg/L). Finally, a long-time (>300 ns) MD simulation, started from the equilibrium structure under the canonical ensemble, was performed. During the simulation, the SHAKE6 algorithm was used to restrict all covalent bonds involving hydrogen atoms within a time step of 2fs.
The cutoff values for electrostatic and van der Waals interactions were set at 10 Å. The above simulation procedure was repeated three times for each protein system.
Molecular Docking
To predict the binding affinity of HER2 ex20ins to TKIs, molecular docking of four kinase inhibitors, afatinib, dacomitinib, pyrotinib, and poziotinib, into G778_P780dup and Y772_A775dup was performed. The protonation state of the protein and the orientation of the hydrogen atoms were optimized by LigX at a pH of 7 and temperature of 300 K. AMBER10:EHT force field and the implicit solvation model of reaction field were applied before docking.
The weight used for tethering side-chain atoms to their original positions was set at 10. The docked poses were ranked by their London ΔG scores followed by a force field refinement.
Treatment
On the basis of the genomic profiling results, patients who carry actionable mutations (EGFR, ALK, ROS1, BRAF, and MET) will be referred to clinical trials investigating the corresponding targeted therapies at SYSUCC for screening. Patients who test negative for potentially actionable mutations (EGFR, ALK, ROS1, BRAF, MET, RET, HER2, and NTRK) will be referred to clinical trials investigating chemotherapies or immunotherapies. Patients who refuse to participate in clinical trials or those who are ineligible will receive standard care.
There is currently no clinical trial for patients carrying EGFR or HER2 ex20ins at SYSUCC, and a platinum-based doublet chemotherapy is usually recommended as the first-line treatment. For patients who refuse chemotherapy, or are considered unsuited, or experience disease progression after the first-line treatment, targeted therapies will be considered. After a detailed discussion of all available treatment options, patients who provide informed consent will receive targeted therapies. Patients mentioned here have provided us their informed consent for use of their clinical data. This study was approved by the SYSUCC Institutional Review Board and was conducted in accordance with the Declaration of Helsinki.
Results
Y772_A775dup and G778_P780dup Are Common HER2 Ex20ins in Lung Cancer
From 4139 lung cancer samples, HER2 mutations were identified in 155 cases (155/4139, 3.7%). Y772_A775dup was identified in 74 cases (74/4139, 1.8%; 74/155, 47.7%), making it the most common HER2 mutation in lung cancer. G778_P780dup was identified in 18 cases (118/4139, 0.4%), accounting for 11.6% of HER2-mutant lung cancers. Y772_A775dup and G778_P780dup both are ex20ins in the αC-β4 loop. Y772_A775dup represents three ex20ins, E770_A771insAYVM, A771_Y772insYVMA, and A775_G776insYVMA. G778_P780dup represents two insertions, V777_G778insGSP and P780_Y781insGSP (Fig. 1A).
Figure 1HER2 exon 20 insertions. (A) Sequence alignment among insertion hotspots of EGFR, HER2, HER3, and HER4. Residues in the αC-β4 loop are shown in yellow boxes; (B) the Ramachandran plot and superimposition between structural model of HER2 G778_P780dup and the template (HER2 WT: 3PP0). Structural differences among the HER2 G778_P780dup and HER2 wild-type (HER2 WT) are highlighted in red square; (C) the Ramachandran plot and superimposition between structural model of HER2 Y772_A775dup and the template (HER2 WT: 3PP0). Structural differences between the HER2 G778_P780dup and HER2 WT are highlighted in red square.
All the patients carrying Y772_A775dup and G778_P780dup had a diagnosis of lung adenocarcinoma. A total of 64 patients were women (64/92, 69.6%). The median age was 58 years, which was not significantly different from that of the study population (median age = 61 y, p = 0.071). Of the 62 patients, 53 were never smokers and nine were light smokers. In terms of concurrent genomic alterations, five cases carried concurrent HER2 amplifications (four in Y772_A775dup, one in G778_P780dup). One case with Y772_A775dup carried concurrent EGFR L858R and EGFR A864P mutations. No concurrent alterations in EGFR, ALK, KRAS, ROS1, MET, RET, BRAF, or NTRK were detected in other cases. As of November 30, 2019; 39 of the 92 patients had died (G778_P780dup = 7 [7/8, 39%], Y772_A775dup = 32 [32/74, 43%]), 49 patients were still alive (G778_P780dup = 10 [10/18, 56%], Y772_A775dup = 39 (39/74, 53%)) and four patients were lost to follow-up (G778_P780dup = 1 [1/18, 6%], Y772_A775dup = 3 [3/74, 4%]).
HER2 Ex20ins Restrict Kinase Conformational Sampling to Varying Degrees
As shown in the Ramachandran plot, the backbones of the two structure models were generally in good agreement with the template. The rmsd value for superimposition between the structure model and the template was 0.16 Å for G778_P780dup (Fig. 1B) and 0.18 Å for Y772_A775dup (Fig. 1C), suggesting that the mechanism of kinase activation and TKI resistance of ex20ins were not attributable to marked structural changes in the adenosine triphosphate (ATP)-binding site or drug-binding pocket, but might arise from long-range allosteric effects on kinase conformation.
To investigate the impact of ex20ins on HER2 kinase conformation, we used conformational sampling to analyze the frequency and distribution of potential kinase conformations of HER2 wild-type (HER2 WT), HER2 G778_P780dup, and HER2 Y772_A775dup. The rmsd of the αC-helix and the distance between the nitrogen atom of K753 and the carbon atom of E770 (K753-E770 interaction) in MD simulations were used as variables to describe kinase conformation. The joint probability distribution of the two variables is depicted in Figure 2A. Each point in Figure 2A represents the log-transformed frequency of one potential kinase conformation in MD simulations, in which the kinase adopts the corresponding αC-helix rmsd and K753-E770 distance.
Figure 2Molecular dynamics (MD) simulations of HER2 wild-type, G778_P780dup and Y772_A775dup. (A) The joint probability distribution of the rmsd of the αC-helix (y axis) and the distance between the nitrogen atom of K753 and the carbon atom of E770 (K753-E770 interaction) (x axis). Each point represents the log-transformed frequency of the corresponding αC-helix rmsd and K753-E770 distance in a MD snapshot. (B) Structural illustrations for representative snapshots of the conformational states identified in MD simulations. Rmsd, root-mean-square deviation.
As shown in Figure 2A, four distinct states could be identified for HER2 WT, whereas two states were identified for G778_P780dup and one state for Y772_A775dup. State 1 is shared by HER2 WT and the two ex20ins. It corresponds to the “DFG (aspartic acid (D)-phenylalanine (F)-Glycine (G))-in, αC-in” active state in which the αC-helix is in an inward position and a salt bridge between K753 and E770 is formed (Fig. 2B). State 2 only exists in HER2 WT and G778_P780dup. It is a meta-active state in which weak interactions take place between K753 and E770 through a water molecule (Fig. 2B). This K753-E770 interaction in state 2 is weakened by charged interactions between K753-D863 and E770-R868. States 3 and 4 are only observed in HER2 WT. State 3 is a meta-inactive state in which the K753-E770 interaction is further weakened, and the E770-R868 interaction becomes stronger (Fig. 2B). State 4 corresponds to the “inactive state” where the K753-E770 interaction is completely lost and the E770-R868 salt bridge is formed. The αC-helix in state 4 is partially unfolded.
In addition, the range of K753-E770 distance could serve as an independent variable to describe structural flexibility of the HER2 kinase. On the basis of the MD simulations, the range of K753-E770 distances in HER2 WT, G778_P780dup, and Y772_A775dup was 16 Å, 9 Å, and 6 Å, respectively. Collectively, these results indicate that ex20ins in the αC-β4 loop change the conformational landscape of HER2 kinase and restrict kinase conformation in the active state. Kinase conformational landscapes are also different among different HER2 insertions. Compared with the more common Y772_A775dup, G778_P780dup has less restriction on kinase conformational sampling.
G778_P780dup Is More Accessible to Small Molecular Kinase Inhibitors
To evaluate the impact of ex20ins on TKI binding, three pan-HER inhibitors (afatinib [Fig. 3A], dacomitinib [Fig. 3B], and pyrotinib [Fig. 3C]), and one exon 20-selective inhibitor (poziotinib, Fig. 3D) were respectively docked into the G778_P780dup and Y772_A775dup system. The free energy of binding (London ΔG) was calculated to assess their affinity to TKIs.
Afatinib, dacomitinib, pyrotinib, and poziotinib are irreversible kinase inhibitors, whose function depends on the formation of a covalent bond between the cysteine residue within the kinase catalytic domain (EGFR C797, HER2 C805) and the electrophilic moiety of inhibitors. As illustrated in Figure 3, the G778_P780dup system allows a reorientation of kinase inhibitors toward C805. Meanwhile, the interactions among the basic residues (K753, R849) and the nucleophilic moiety of the inhibitors further stabilize the binding between G778_P780dup and kinase inhibitors. In the Y772_A775dup system, the strong K753-E770 interaction as indicated by the small range of K753-E770 distance (6 Å), tightens its drug-binding pocket and restricts its kinase conformation in the DFG-in, αC-in state. Therefore, kinase inhibitors could not completely fit into its pocket and leave more ligands exposed (Fig. 3). For the part of the inhibitor that fits in, a highly restricted conformation of the protein prevents its reorientation toward C805, thus abrogating the activity of kinase inhibitors. The predicted free energy of binding (London ΔG) with afatinib, dacomitinib, pyrotinib, and poziotinib is also lower in the G778_P780dup system compared with Y772_A775dup (Fig. 3).
Figure 3Interactions and space fill models of HER2 G778_P780dup and HRE2 Y772_A775dup with afatinib (A), dacomitinib (B), pyrotinib (C), and poziotinib (D) in-molecular docking. Predicted free energy of binding (London ΔG) is given in the bottom right table.
Treating Patients Carrying G778_P780dup With Kinase Inhibitors
Ten patients with lung adenocarcinoma carrying G778_P780dup were treated with kinase inhibitors during the course of clinical care. The clinicopathologic characters and treatment histories are listed in Table 1 and Supplementary Table 1.
Table 1Clinicopathological Characteristics of Patients With Lung Adenocarcinoma Carrying HER2 G778_P780dup Treated With Kinase Inhibitors
Seven patients received afatinib as monotherapy (n = 6) or in combination with chemotherapy (n = 1). Tumor responses were observed in three patients. Sustained clinical benefits (tumor response or stable disease for at least 6 mo) were observed in six patients. Patient 1 was a 44-year-old woman never smoker who was diagnosed with lung adenocarcinoma with brain metastasis. Genomic profiling using a 450-gene OrigiMed panel revealed a HER2 G778_P780dup without HER2 amplification or concurrent mutation. She received stereotactic body radiotherapy for the brain metastasis before systemic treatment. After one cycle of pemetrexed plus carboplatin, the patient refused to continue chemotherapy owing to severe nausea and vomiting. She was then treated with afatinib 30 mg every day. The chest CT scan after 4 weeks reveal a 51% reduction in tumor measurement on the basis of Response evaluation criteria in solid tumors 1.1 (Fig. 4A). She developed new brain metastasis after 8 months of afatinib and received a second SBRT for the metastasis. She then started poziotinib (16 mg every day) with a stable disease as the best response. She has maintained disease control on poziotinib for 9 months before disease progression (Fig. 4A).
Figure 4Computed tomography scans from patients with lung adenocarcinomas carrying HER2 G778_P780dup treated by kinase inhibitors. (A) Patient 1 who achieved a partial response to second-line treatment with afatinib and maintained disease control on third-line treatment with poziotinib for 9 months. (B) Patient 6 who achieved a partial response to third-line treatment with pyrotinib. (C) Patient 10 who received second-line treatment with dacomitinib and reported a marked tumor response.
Four patients received pyrotinib monotherapy as the second-line (n = 1) or third-line and beyond treatment (n = 3). At the time of article submission, these patients have maintained disease control on pyrotinib for 8 months (patient 6), 3 months (patient 7), 6 months (patient 8), and 11 months (patient 9), respectively. Tumor responses were documented in two patients (patient 6, patient 8). Patient 6, a 51-year-old woman, presented with pain in the left side of the chest and cough. She was diagnosed with lung adenocarcinoma with pleural effusion and subclavicular lymph node metastasis. She developed disease progression after 4 cycles of tegafur plus carboplatin and two cycles of tegafur. She then received second-line treatment with afatinib (30 mg every day) and maintained a stable disease for 6 months. After progression on afatinib, the patient was treated with pyrotinib (400 mg every day). Her symptoms and the pleural effusion were almost completely resolved after 1 month of treatment (Fig. 4B).
Patient 10 is a 65-year-old man who received second-line treatment with dacomitinib. The patient was diagnosed with stage IV lung adenocarcinoma with bone and pleural metastases. He received first-line pemetrexed plus carboplatin plus bevacizumab for six cycles, followed by pemetrexed plus bevacizumab. After four cycles of pemetrexed plus bevacizumab, the patient reported evidence of disease progression on CT scan and symptom deterioration including severe dyspnea, hemoptysis, and chest pain. He was considered unsuited for second-line chemotherapy and received dacomitinib 30 mg every day instead. The patient reported a complete resolution of symptoms after 2 weeks of dacomitinib. Dosage modification to 15 mg every day was made after 4 weeks owing to paronychia. Marked tumor response was observed after 3 months of treatment (Fig. 4C), and the patient stayed on dacomitinib at the time of the previous follow-up.
Discussion
Identifying molecular features affecting TKI sensitivity in ex20ins could guide therapeutic decisions with currently available drugs and provide implications for future drug research & development (R&D) strategies. Our MD simulations of two common HER2 ex20ins, Y772_A775dup, and G778_P780dup, reveal that activating ex20ins lead to constitutive, ligand-independent kinase activation by changing the conformational landscape of HER2 kinase and restricting kinase conformation in the active state. Kinase conformational landscapes are also different among different HER2 insertions. The distinctive kinase conformational landscape of one insertion, characterized by its degree of restriction on kinase conformational sampling, affects its inhibitor binding affinity and, therefore, TKI sensitivity.
Compared with Y772_A775dup, G778_P780dup that has less restriction on kinase conformational sampling shows higher affinity to afatinib, dacomitinib, pyrotinib, and poziotinib, which is supported by clinical responses to these inhibitors in six out of 10 patients carrying G778_P780dup. Our modeling of HER2 ex20ins and the crystallography of EGFR insertions (D770_N771insNPG) both reveal that ex20ins do not dramatically alter the ATP-binding site or drug-binding pocket of the kinase.
In this study, we focused on the local conformational dynamics of regions involved in kinase regulation. On the basis of the MD simulations results, it was noted that HER2 G778_P780dup and Y772_A775dup both alter the conformational landscape of HER2 kinase and restrict the kinase conformation in the active state. Different HER2 ex20ins restrict kinase conformational sampling to different degrees, as reflected by their distinctive conformational landscapes. Compared with Y772_A775dup, the conformational landscape of HER2 G778_P780dup allows inhibitor reorientation in its drug-binding pocket and effective TKI binding. These data suggest that the constitutive, ligand-independent kinase activation of activating ex20ins arise from its ability to restrict the kinase conformation in the active state. For an exon 20 insertion, its degree of restriction on kinase conformational sampling, as reflected by the conformational landscape of the insertion, affects its affinity and sensitivity to TKIs.
Next, we examined potential determinants of kinase conformational landscape of HER2 ex20ins. The length of EGFR or HER2 ex20ins range from one to four amino acids.
Sequence alignment of EGFR family shows that S768 and D770 are conserved in EGFR, HER3, and HER4 from mouse to human, whereas G776 and G778 are characteristic in HER2 (Fig. 1A). Compared with Y772_A775dup, TKI-responsive G778_P780dup has a three-amino acid extensions in the αC-β4 loop and retains both G776 and G778. Furthermore, HER2 M774delinsWLV, a dacomitinib-responsive exon 20 insertion,
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.
has a two-amino acid extensions in the αC-β4 loop and retains only G778. Recently, we reported an afatinib-sensitive insertion, HER2 G776delinVC, which had a one-amino acid extension in the αC-β4 loop and no retainment of G776 or G778.
Collectively, these data suggest that the kinase conformational landscape of a HER2 exon 20 insertion is probably a function of two factors, the length of the αC-β4 loop and residues at HER2 776 and 778 positions. Predictably, HER2 ex20ins with these two factors identical to G778_P780dup, such as HER2 S779_P780insVGS and G778_S779insCPG, will have a similar kinase conformational landscape and are also sensitive to TKIs. In addition, HER2 G776delinsAVGC that has a 2-amino acid extensions and retains G778 is predicted to be dacomitinib-responsive like M774delinsWLV. Future studies are warranted to expand on these findings.
Results of this study also provide implications for drug R&D strategies against Y772_A775dup, the predominant HER2 mutation in lung cancer. We previously reported that Y772_A775dup responded poorly to afatinib.
Here, we further dig into its structural basis. Currently available HER2 TKIs all target the ATP-binding site, which is also considered the drug-binding pocket of the kinase. According to the conformational landscape of HER2 ex20ins, the YVMA insertion facilitated the formation of K753-E770 salt bridge and restricted the kinase in the active conformation. This conformational rigidity renders the ATP-binding site of Y772_A775dup less accessible to small molecular TKIs. Simply reducing the size of TKIs seems futile in improving their efficacy. Pyrotinib, a pan-HER TKI that was developed on the basis of the structure of neratinib, has the largest size among the four TKIs mentioned above.
Discovery and development of pyrotinib: a novel irreversible EGFR/HER2 dual tyrosine kinase inhibitor with favorable safety profiles for the treatment of breast cancer.
However, it shows high binding affinity to HER2 ex20ins (London ΔG, G778_P780dup = -15.17; Y772_A775dup = -10.08) and delivers promising results in clinical trials.
Considering the highly restricted accessibility of the traditional drug-binding pocket (ATP-binding site) of Y772_A775dup, developing an allosteric inhibitor that could bind to the kinase away from the ATP-binding site may be a more rational strategy. The discovery of EAI045, a fourth-generation EGFR TKI that could overcome osimertinib-resistant EGFR C797S, is an example worth emulating.
In addition, some HER2 TKIs that are effective in vitro fail to deliver satisfactory efficacy in patients with HER2-mutant LUAD, such as poziotinib and neratinib.
This suggests that aside from TKI binding affinity, there may be other factors further complicating the suppression of HER2 signaling in vivo. Potential confounding factors include concomitant mutations, immune microenvironment and changes in the equilibrium between active and inactive kinase conformations in vivo.
Further investigation in this field is required to improve treatment outcomes of this population.
In summary, the present work identifies the kinase conformational landscape, dictated by the length of the αC-β4 loop and the residues at HER2 776 and 778 positions, as a molecular feature that affects TKI sensitivity in HER2 ex20ins. It provides a structural basis for understanding response heterogeneity to kinase inhibitors in ex20ins, which could guide a rational selection of currently available drugs and shed some light on future drug R&D strategies against ex20ins in lung cancer.
Acknowledgments
This study was funded by National Key Research & Development Program of People’s Republic of China (2016YFC0905500, 2016YFC0905503), Science and Technology Program of Guangdong (2017B020227001) and National Natural Science Funds of People’s Republic of China (81772476, 81602011). The funding organizations had no role in study design, data collection, data analysis, data interpretation, or writing of the report. The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication. We thank OrigiMed, Beijing Genomics Institute and Guangzhou Mygene Medical Institute for conducting genomic profiling.
First report of safety, PK, and preliminary antitumor activity of the oral EGFR/HER2 exon 20 inhibitor TAK-788 (AP32788) in non–small cell lung cancer (NSCLC).
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.
Clinical significance of circulating tumor cells in predicting disease progression and chemotherapy resistance in patients with gestational choriocarcinoma.
AMBER, a package of computer programs for applying molecular mechanics, normal mode analysis, molecular dynamics and free energy calculations to simulate the structural and energetic properties of molecules.
Discovery and development of pyrotinib: a novel irreversible EGFR/HER2 dual tyrosine kinase inhibitor with favorable safety profiles for the treatment of breast cancer.
Drs. Zhao, Fang, and Pan equally contributed to this work.
Disclosure: Dr. Wang reports that he is a shareholder of OrigiMed Inc. (Shanghai, People’s Republic of China). The remaining authors declare no conflict of interest.
30094-0&id=gr1.jpg){kind=link}
30094-0&id=gr2.jpg){kind=link}
30094-0&id=gr3.jpg){kind=link}
30094-0&id=gr4.jpg){kind=link}