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
A significant proportion of patients with lung cancer carry mutations in the EGFR kinase domain. The presence of a deletion mutation in exon 19 or L858R point mutation in the EGFR kinase domain has been shown to cause enhanced efficacy of inhibitor treatment in patients with NSCLC. Several less frequent (uncommon) mutations in the EGFR kinase domain with potential implications in treatment response have also been reported. The role of a limited number of uncommon mutations in drug sensitivity was experimentally verified. However, a huge number of these mutations remain uncharacterized for inhibitor sensitivity or resistance.
Methods
A large-scale computational analysis of clinically reported 298 point mutants of EGFR kinase domain has been performed, and drug sensitivity profiles for each mutant toward seven kinase inhibitors has been determined by molecular docking. In addition, the relative inhibitor binding affinity toward each drug as compared with that of adenosine triphosphate was calculated for each mutant.
Results
The inhibitor sensitivity profiles predicted in this study for a set of previously characterized mutants correlated well with the published clinical, experimental, and computational data. Both the single and compound mutations displayed differential inhibitor sensitivity toward first- and next-generation kinase inhibitors.
Conclusions
The present study provides predicted drug sensitivity profiles for a large panel of uncommon EGFR mutations toward multiple inhibitors, which may help clinicians in deciding mutant-specific treatment strategies.
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).
Of these, classical activating mutations such as deletion in the exon 19 and a point mutation L858R together account to approximately 85% of mutated NSCLC tumors. The classical activating mutations were shown to cause more sensitivity than that of the wild-type (WT) kinase toward EGFR inhibitor treatment.
Functional characterization and drug sensitivity profiling of mutations aids in stratification of patients with cancer for appropriate inhibitor treatment. Drug sensitivities of a broad panel of NSCLC-specific EGFR mutations toward first-generation EGFR kinase inhibitors were previously reported.
However, the EGFR inhibitor sensitivity for a significant number of EGFR mutations has remained elusive. In addition, clinical and preclinical studies regarding the effect of next-generation inhibitors on nonclassical EGFR mutants are limited. Thus, computational analysis has been performed to predict the drug sensitivity or resistance of clinically relevant EGFR kinase domain mutations.
Materials and Methods
A total of 298 cancer-specific EGFR mutations (in the Catalogue of Somatic Mutations in Cancer database) affecting 136 amino acid residues (Fig. 1) were analyzed for their sensitivity toward EGFR kinase inhibitors. Mutant EGFR kinase structures were generated by using RosettaBackrub software based on the WT crystal structure (PDB: 1XKK). Deletion and insertion mutant structures were constructed using SWISS-MODEL software based on WT structure as a template. Each of these mutant structures was docked with either adenosine triphosphate (ATP) or one of the eight EGFR kinase inhibitors gefitinib, erlotinib, lapatinib, dacomitinib, afatinib, vandetanib, neratinib, and osimertinib. The binding affinities of the docked ligands were calculated by using the Schrodinger software in extraprecision (xP) mode and the output for each receptor-ligand docking was measured as an xP score. Further, the relative inhibitor binding affinity (RIBA) of EGFR mutants for each kinase inhibitor was calculated as the ratio of xPinhibitor to xPATP.
Figure 1Schematic representation of EGFR mutations analyzed in this study. Cancer-specific point mutations that were reported in the EGFR kinase domain were collected from the Catalogue of Somatic Mutations in Cancer database (http://cancer.sanger.ac.uk/cosmic). A total of 298 substitutions affecting 136 amino acid residues that were analyzed in the present study are schematically shown.
In total, 2392 dockings were performed by using seven EGFR kinase inhibitors as well as ATP against WT kinase and 298 mutants. The RIBA, which is the ratio of xPinhibitor to xPATP, reflects the ability of a kinase inhibitor to competitively displace ATP from the active site; a ratio higher than 1 is considered good inhibitor efficacy and a ratio lower than 1 is considered inhibitor resistance. Initially, the validation of RIBA values for a panel of 12 mutants was performed by using published experimental cellular concentration that inhibits 50% values for gefitinib and erlotinib as a reference (Supplementary Table 1).
The increased drug-binding affinity of the most frequent point mutation (L858R) compared with that of the WT kinase is a remarkable feature that underlies the success of EGFR kinase inhibitors.
In line with earlier reports, in the present study EGFR L858R also displayed sensitivity (RIBA ≥1) whereas EGFR T790M displayed maximum resistance (see Supplementary Table 1).
Importantly, as reported earlier, other common mutations G719S and L861Q showed reduced binding affinity toward EGFR inhibitors compared with that of the reference L858R mutation (Supplementary Table 2).
Additionally, when compared with the L858R mutant, most of the EGFR mutants displayed lower RIBA values toward gefitinib and erlotinib, thus conforming with the higher experimental concentration that inhibits 50% values that were reported earlier (see Supplementary Table 1).
Among the mutations in the validation set, nine of 12 remained sensitive to at least one of the seven inhibitors (see Supplementary Table 2). The remaining three mutants, R784F, T790M, and N826S, showed cross-resistance to all the inhibitors (see Supplementary Table 2).
Unbiased listing of the top 20 drug-sensitive EGFR mutants on the basis of their RIBAs revealed differential drug sensitivity profiles (Fig. 2A). The most frequent (classical) inhibitor-sensitizing point mutation, L858R, featured in the top 20 lists for all EGFR inhibitors except that for neratinib (see Fig. 2). Importantly, about half of the top 20 EGFR-sensitive mutations identified in this study are located in one of the three regions of the kinase domain: P-loop (719–724), C-loop (836–842), and A-loop (855–876). In particular, the mutations G721A, M766V, L792P, N842H, K852N, and T854I displayed drug-sensitizing effects toward all the seven EGFR inhibitors, with a relative binding affinity greater than that of the WT kinase (Supplementary Table 3).
Figure 2Distinct inhibitor sensitivity/resistance profiles of EGFR mutants. (A) An unbiased listing of the top 20 inhibitor-sensitizing mutants (with the highestxPInhibitor/xPATP) was performed for each of the seven EGFR kinase inhibitors and is presented. xPInhibitor/xPATP for the wild-type (WT) kinase is shown as the reference. Mutants encompassing important regions of the EGFR kinase are depicted in the indicated colors. (B) The top 20 highly resistant EGFR mutants (those with the lowest xPInhibitor/xPATP) for each indicated kinase inhibitor are listed, and mutants covering important hotspot regions are color-coded. xPInhibitor/xPATP for the WT kinase is shown as the reference, and the gatekeeper mutant T790M is shown in pink. ATP, adenosine triphosphate.
An additional unbiased listing based on the RIBA values revealed that most of the top 20 drug-resistant mutations for each inhibitor mapped to either the conserved V742-A750 region (V742I, A743M, A743P, A743S, A743T, A743V, E746G, E746K, E746V, L747P, L747S, L747V, and R748I) or the activation loop (T854S, D855G, D855N, F856S, G857E, G857R, K860I, L861G, L861P, L861Q, L862R, G863S, A864E, E865K, and E866K) (Fig. 2B). Notably, the gatekeeper mutation T790M displayed the highest gefitinib resistance among the 298 mutations analyzed (see Fig. 2B). In addition, T790M is also featured in the top 20 resistant mutations for the inhibitors erlotinib, lapatinib, and afatinib (see Fig. 2B). T790M mutation also displayed cross-resistance toward other inhibitors such as dacomitinib, vandetanib, and neratinib, albeit to a lesser extent (Supplementary Table 4). This is in line with previous observations that T790M is resistant to gefitinib, erlotinib, lapatinib, and vandetanib but sensitive to higher concentrations of dacomitinib, afatinib, and neratinib.
In total, 31 mutations showed a high level of drug resistance toward at least one of the seven EGFR inhibitors (see Supplementary Table 4). Most of these 31 mutants displayed inhibitor cross-resistance (see Supplementary Table 4). It is important to note that many of the drug-resistant mutated residues identified in this study (E709, L718, V742, A743, P794, G796, T854, L861, and A864) are in conformity with the results of previous studies in displaying resistance toward one or more EGFR inhibitors.
A743M, 743S, 743V, and T790M mutations exhibited cross-resistance to multiple EGFR inhibitors (see Supplementary Table 4). A743S and A743V showed higher affinity for ATP in addition to the reduced binding affinity for EGFR inhibitors, thus resulting in significantly lower RIBAs than that of the WT kinase (Fig. 3A–C). Further analysis revealed that these three mutants caused a significant reduction in size of the ligand-binding pocket, leading to reduced binding affinity and consequent resistance to all the EGFR inhibitors (see Fig. 3A–C).
Figure 3Drug resistance caused by A743M/S/V and compound mutations. Molecular mechanism underlying inhibitor resistance was studied by superposing EGFR wild-type (WT) structure over the mutants A743M (A), A743S (B) and A743V (C). (D) A panel of cancer-specific compound mutations that were reported in the EGFR kinase domain were modeled and xPInhibitor/xPATP for each inhibitor based on molecular docking analysis is shown. The most frequent NSCLC-specific highly resistant L858R plus T790M mutant is indicated in gray.
The existence of a drug-resistant T790M mutation in combination with activating mutations has been reported in approximately 50% of responders to kinase inhibitor treatment at the time of relapse. In addition to this compound mutation, several less frequent compound mutations have been reported in lung cancers. However, the role of these compound mutations in drug sensitivity is unclear. Analysis of a panel of nine compound mutations revealed that the L858R-T790M mutant conferred significant resistance toward all seven inhibitors (Fig. 3D). Similar to the T790M single mutation, the L858R-T790M compound mutation also displayed an increase in the binding affinity for ATP, as well as a reduction in the binding affinity toward EGFR kinase inhibitors (Supplementary Table 4). In contrast, the mutation G719D-L861R displayed sensitivity toward all the inhibitors tested (see Fig. 3D). Although L858R-R776H showed sensitivity toward five inhibitors, the remaining compound mutations displayed intermediate sensitivity toward most of the inhibitors (see Fig. 3D). Thus, differential kinase inhibitor sensitivities were observed among compound mutations that were analyzed in this study.
Multiple substitutions were frequently observed at key residues in the EGFR kinase. However, how these multiple mutations respond to kinase inhibitor treatment is unclear. Comparative analysis of relative binding affinities indicated that all substitutions at residues E746, L747, and R748 caused very high resistance toward all the docked EGFR inhibitors (Supplementary Table 5). Although E709V, E709K, E709A, and E709H are resistant to most EGFR inhibitors, E709G remained sensitive to gefitinib, erlotinib, dacomitinib, afatinib, and vandetanib (see Supplementary Table 5). G721A and G721D remained sensitive to most EGFR inhibitors, whereas G721S was sensitive to dacomitinib and vandetanib (see Supplementary Table 5). L861Q, L861R, L861F, L861E, L861P, and L861G displayed cross-resistance to multiple inhibitors, whereas L861V remained sensitive to all the inhibitors tested (see Supplementary Table 5). Thus, multiple substitutions at critical residues displayed differential EGFR inhibitor sensitivity.
After the recent approval, we extended the docking analysis to osimertinib, which demonstrated a drug resistance profile distinct from that of the seven first- and second-generation inhibitors (Supplementary Fig. 1 and Supplementary Table 6). Further docking analysis of most common exon19 deletion and exon20 insertion mutants revealed that the observed inhibitor sensitivities are in line with preclinical and clinical findings (Supplementary Fig. 2), indicating the applicability of this approach for indel mutations also.
In conclusion, the present study predicted drug sensitivity profiles of 298 EGFR point mutants toward eight first-, second-, and third-generation EGFR kinase inhibitors (see Supplementary Table 6). The relative binding affinities as measured by xP score ratios of inhibitor to that of the ATP largely correlated with the previous experimental data. Specifically, the drug sensitivity profiles of both frequent and uncommon mutations correlated well with our experimental findings obtained by using cell culture models (see Supplementary Table 1). Further validation revealed that the RIBA scores of uncommon mutations are in conformity with cell culture–based inhibitor sensitivity profiles generated by various groups (Supplementary Table 7). Notably, the computational data on rare EGFR mutations also correlated well with the clinical data (see Supplementary Table 6).
Synchronous occurrence of squamous-cell carcinoma “transformation” and EGFR exon 20 S768I mutation as a novel mechanism of resistance in EGFR-mutated lung adenocarcinoma.
Furthermore, the data also correlated well with previous computational studies that analyzed the drug sensitivity or resistance of a limited number of EGFR mutants.
A systematic profile of clinical inhibitors responsive to EGFR somatic amino acid mutations in lung cancer: implication for the molecular mechanism of drug resistance and sensitivity.
Only those mutants that displayed a higher RIBA toward a particular inhibitor than that of the WT kinase were considered sensitive in this study. However, there is still a possibility of drug efficacy despite a lower RIBA for certain mutations owing to a chance of achieving higher serum concentrations of the drug. It is difficult to estimate the precise efficacy of irreversible inhibitors (dacomitinib, afatinib, and neratinib), thus adding complexity to predict inhibitor potency.
Taken as a whole, the predicted drug sensitivity profiles generated in this study provide hints to physicians in selecting potential strategy to treat patients with lung cancer with rare single or compound EGFR mutations.
Acknowledgments
Dr. Kancha acknowledges funding from the Department of Science and Technology-Science and Engineering Research Board (extramural grant No. EMR/2014/000904) and University Grants Commission-Faculty Recharge Programme (salary grant No. F.4-5(136-FRP)/2014(BSR)). Dr. Vudem acknowledges funding from the UGC-UPE (grant No. F-14-5/2012(NS/PE)).
Supplementary Tables 1–7 and Supplementary Figures 1 and 2
References
Barlesi F.
Mazieres J.
Merlio J.P.
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).
Synchronous occurrence of squamous-cell carcinoma “transformation” and EGFR exon 20 S768I mutation as a novel mechanism of resistance in EGFR-mutated lung adenocarcinoma.
A systematic profile of clinical inhibitors responsive to EGFR somatic amino acid mutations in lung cancer: implication for the molecular mechanism of drug resistance and sensitivity.
30004-2&id=gr1.jpg){kind=link}
30004-2&id=gr2.jpg){kind=link}
30004-2&id=gr3.jpg){kind=link}