The CLIP1-LTK fusion was recently discovered as a novel oncogenic driver in non-small cell lung cancer (NSCLC). Lorlatinib, a third-generation ALK inhibitor, exhibited a dramatic clinical response in a NSCLC patient harboring CLIP1-LTK fusion. However, it is expected that acquired resistance will inevitably develop, particularly by LTK mutations, as observed in NSCLC induced by oncogenic tyrosine kinases treated with corresponding tyrosine kinase inhibitors (TKIs). In this study, we evaluate eight LTK mutations corresponding to ALK mutations that lead to on-target resistance to lorlatinib. All LTK mutations show resistance to lorlatinib with the L650F mutation being the highest. In vitro and in vivo analyses demonstrate that gilteritinib can overcome the L650F-mediated resistance to lorlatinib. In silico analysis suggests that introduction of the L650F mutation may attenuate lorlatinib-LTK binding. Our study provides preclinical evaluations of potential on-target resistance mutations to lorlatinib, and a novel strategy to overcome the resistance.
A novel strategy using gilteritinib can overcome resistance to lorlatinib in non-small cell lung cancer with CLIP1-LTK-L650F mutation.
These authors contributed equally: Shunta Mori, Hiroki Izumi.
The discovery of several actionable oncogenic drivers in non-small cell lung cancer (NSCLC) and the development of corresponding targeted therapies have changed the treatment strategy, leading to great improvement in patient outcome[
In this study, we demonstrated that LTK mutations, especially the L650F mutation, potentially confer resistance to lorlatinib treatment, and that L650F-mediated resistance to lorlatinib can be overcome by gilteritinib.
LTK and ALK belong to the insulin receptor subfamily of receptor tyrosine kinases, which consist of an extracellular region, transmembrane region, and intracellular region. The kinase domain of LTK and ALK contains 268 amino acids (Fig. 1a). Intriguingly, LTK and ALK exhibit 79% amino acid homology in their respective kinase domains[
Graph: Fig. 1Lorlatinib is predicted to bind to LTK and ALK.a Schematic representation of LTK and ALK protein. LTK and ALK are formed with three regions: extracellular, transmembrane, and intracellular region. They consist of 864 amino acids (aa) and 1620 aa respectively. MAM, meprin, A-5 protein, and receptor protein-tyrosine phosphatase μ domain; LDLa, low-density lipoprotein class A motif; Gly_rich, glycine rich region; TM, transmembrane region. b The lorlatinib-binding mode for the LTK kinase domain. The protein is depicted by a surface model (I565, blue; L590, red; L592, orange; G596, cyan; D597, yellow; L650, magenta; others, gray) and lorlatinib is depicted by sticks (C, green; N, blue; and O, red). An enlarged view of the ATP-binding pocket is shown in the left panel. In the right panel, hydrogen bonds between LTK residues (E591 and M593) and lorlatinib are shown as dashed yellow lines.
We first evaluated the binding affinity of lorlatinib to LTK proteins using in silico approaches to further support the efficacy of lorlatinib in CLIP1-LTK fusion-expressing cells[
The most common resistance mechanism to genotype-matched therapy is caused by acquired genetic alterations in the on-target gene, including gatekeeper or solvent-front mutations[
Graph: Fig. 2 LTK mutations are analogous to ALK resistant mutations.a Amino acid sequences between full LTK and ALK kinase domain. Asterisks represent all the conserved amino acid residues across LTK and ALK. Their homological mutations are surrounded with red boxes. b LTK and ALK corresponding mutations.
We then established Ba/F3 and NIH3T3 cells expressing CLIP1-LTK fusion proteins with the aforementioned LTK mutations to clarify the impact of these mutations on sensitivity to lorlatinib, as well as other targeted agents. Cell viability assays using Ba/F3 cells expressing WT CLIP1-LTK or each CLIP1-LTK mutation revealed that Ba/F3 cells expressing mutant CLIP1-LTK were less sensitive to lorlatinib compared with those expressing WT CLIP1-LTK (Fig. 3a). The western blotting assay also showed that the effect of lorlatinib on LTK tyrosine phosphorylation was attenuated in Ba/F3 cells expressing mutant CLIP1-LTK compared with Ba/F3 cells expressing WT CLIP1-LTK. Ten nM of lorlatinib or higher inhibited the LTK tyrosine phosphorylation of Ba/F3 cells expressing WT CLIP1-LTK, whereas that of CLIP1-LTK with kinase mutations was not inhibited by 10 nM of lorlatinib (Fig. 3b). In particular, the L650F mutation was the most resistant to lorlatinib in terms of inhibition of cell proliferation with IC
Graph: Fig. 3 LTK mutations are associated with resistance to lorlatinib.a Cell viability curves for Ba/F3 cells expressing WT CLIP1-LTK and mutant CLIP1-LTK treated with lorlatinib at the indicated concentrations for 48 h. Cell viability was evaluated using Cell Counting Kit-8. Error bars are indicated as mean ± SD from three independent experiments. b The attenuation of LTK phosphorylation in Ba/F3 cells expressing CLIP1-LTK with eight mutations treated with lorlatinib at the increasing concentrations for 24 h. Cell extracts were analyzed by western blotting assay using the indicated antibodies. p-LTK, phospho-LTK. c The percentage of apoptosis in Ba/F3 cells expressing CLIP1-LTK with eight mutations treated with 10 nM lorlatinib for 24 h. The cells were stained with AnnexinV and propidium iodide. Apoptotic cells were then measured by flow cytometry. Error bars are indicated as mean ± SD from three independent experiments. *p < 0.001 (Dunnett's test).
Furthermore, cell apoptosis was evaluated in Ba/F3 cells expressing WT or mutant CLIP1-LTK treated with lorlatinib. Ba/F3 cells expressing CLIP1-LTK-L592F were more susceptible to lorlatinib than those expressing other CLIP1-LTK mutations possibly due to differences in cell growth rates. However, apoptosis was significantly suppressed in all CLIP1-LTK mutant Ba/F3 cells compared to those expressing WT CLIP1-LTK (Fig. 3c). These results suggested that these LTK mutations are resistant to lorlatinib-induced LTK kinase inhibition and cell apoptosis.
Next, we explored potential compounds that could overcome lorlatinib resistance mediated by these LTK mutations and evaluated the sensitivity of the following compounds: lorlatinib, crizotinib, alectinib, ceritinib, brigatinib, entrectinib, repotrectinib, and gilteritinib. To compare the sensitivity to each compound in Ba/F3 cells expressing WT or mutant CLIP1-LTK, the IC
Graph: Fig. 4IC 50 values of eight compounds in Ba/F3 cells expressing indicated LTK mutations.Parental Ba/F3 cells and Ba/F3 cells expressing WT CLIP1-LTK and mutant CLIP1-LTK were treated with the eight indicated inhibitors at several concentrations for 48 h. Cell viability was evaluated using the Cell Counting Kit-8. The mean IC 50 values are shown.
Moreover, gilteritinib can also inhibit Ba/F3 expressing WT CLIP1-LTK, with IC
We further focused on CLIP1-LTK-L650F, which was the most resistant strain to lorlatinib in this study. Among the tested compounds, gilteritinib was the most potent in Ba/F3 cells expressing CLIP1-LTK-L650F (Fig. 5a). Therefore, we investigated whether gilteritinib could overcome resistance to lorlatinib induced by CLIP1-LTK-L650F. The western blotting assay showed that gilteritinib successfully inhibited LTK phosphorylation in Ba/F3 cells expressing CLIP1-LTK-L650F. Indeed, at 100 nM, gilteritinib strongly attenuated AKT and ERK phosphorylation, whereas lorlatinib did not. In addition, gilteritinib increased the levels of the stabilized form of BIM and cleaved caspase-3, the hallmark of apoptosis (Fig. 5b). Fluorescence-activated cell sorting (FACS) analysis using annexin V/propidium iodide (PI) staining also confirmed that gilteritinib induced apoptosis in Ba/F3 cells carrying CLIP1-LTK-L650F (Fig. 5c). An increase in caspase activity by gilteritinib, but not lorlatinib, also supported the successful induction of apoptosis by gilteritinib (Supplementary Fig. 3).
Graph: Fig. 5Gilteritinib is potent in overcoming lorlatinib resistance by CLIP1-LTK -L650F.a Cell viability curves for Ba/F3 cells expressing CLIP1-LTK -L650F treated with the indicated compounds at the increasing concentrations for 48 h. Cell viability was evaluated using Cell Counting Kit-8. Error bars are indicated as mean ± SD from three independent experiments. b Western blotting showing LTK and its downstream signaling molecules in Ba/F3 cells expressing CLIP1-LTK -L650F. The cells were treated with lorlatinib and gilteritinib at the indicated concentrations for 16 h. Cell extracts were analyzed by western blotting assay using the indicated antibodies. p-LTK, phospho-LTK; p-AKT, phospho-AKT; p-ERK, phospho-ERK. c The percentage of apoptosis in Ba/F3 cells expressing CLIP1-LTK -L650F treated with 0.1% DMSO, lorlatinib (1 μM) and gilteritinib (1 μM) for 24 h. The cells were stained with AnnexinV and propidium iodide. Apoptotic cells were then measured by flow cytometry. Error bars are indicated as mean ± SD from three independent experiments. * p < 0.001; n.s, not significant (Turkey's test). d The diameters of NIH3T3 cells carrying CLIP1-LTK -L650F treated with 0.1% DMSO, lorlatinib (1 μM) and gilteritinib (1 μM) for 14 days. Error bars are indicated as mean ± SD from three independent experiments. Scalebars, 100 μm. * p < 0.001; n.s, not significant (Turkey's test). e Inhibition of lorlatinib and gilteritinib against mouse tumors bearing NIH3T3 cells expressing CLIP1-LTK -L650F. Mice were treated with either lorlatinib (10 mg/kg once daily), gilteritinib (30 mg/kg once daily) or vehicle control. Error bars are indicated mean ± SD (n = 6 for each group) * p < 0.05; ** p < 0.01; n.s, not significant (Turkey's test) f Body weight changes in mice indicated in e. Error bars are indicated as mean body weight± SD (n = 3 for each group) and statistically analyzed by Turkey's test. n.s, not significant.
We subsequently investigated the inhibitory effect of gilteritinib in another cell model, NIH3T3 cells carrying CLIP1-LTK-L650F, using a soft agar colony formation assay. The diameter of colonies treated with gilteritinib was significantly smaller than that treated with lorlatinib or dimethylsulfoxide (DMSO), whereas lorlatinib did not inhibit colony formation compared with DMSO (Fig. 5d).
Finally, we tested the activity of gilteritinib against CLIP1-LTK-L650F cells using a xenograft model. Consistent with the results of the in vitro experiments, there was no significant difference in tumor size between the lorlatinib and vehicle control groups, suggesting the robust resistance of CLIP1-LTK-L650F to lorlatinib. In contrast, gilteritinib significantly inhibited tumor growth compared with the vehicle control or lorlatinib (Fig. 5e). Notably, no significant difference in body weight was detected among these three groups, suggesting that gilteritinib showed less toxicity (Fig. 5f). Collectively, gilteritinib potentially overcame the L650F-mediated resistance to lorlatinib in tumors expressing CLIP1-LTK-L650F.
We further explored how these LTK mutations affect sensitivity to lorlatinib and gilteritinib. As lorlatinib failed to inhibit LTK phosphorylation in cells with LTK mutations, we speculated that these mutations affected LTK-lorlatinib binding. We estimated the binding affinity of lorlatinib against WT CLIP1-LTK and its mutants using the Massively Parallel Computation of Absolute binding Free Energy with well-equilibrated states (MP-CAFEE) method[
Graph: Fig. 6Gilteritinib has potent to overcome lorlatinib resistance by CLIP1-LTK-L650F.The plots of the binding free energy (ΔG) of lorlatinib (a) or gilteritinib (b) to WT CLIP1-LTK or each mutant CLIP1-LTK against experimental IC 50 values of the lorlatinib or gilteritinib in Ba/F3 cells expressing the corresponding CLIP1-LTK as shown in Fig. 4. These ΔG values were calculated by MP-CAFEE. MD-relaxed structures of c lorlatinib or d gilteritinib-bound WT CLIP1-LTK (green) and the L650F mutant (magenta). The energetically-stable structure for each LTK–drug complex was extracted from five independent 50 ns MD simulations. The protein backbone is represented by a ribbon diagram, and L/F650 and lorlatinib/gilteritinib are depicted by sticks (C, green/magenta; N, blue; O, red). ΔG values and electrostatic (Coulomb) and van der Waals (vdW) contributions to them are also indicated.
This study was the first to explore potential LTK resistance alternations against lorlatinib in tumors expressing the CLIP1-LTK fusion protein. We found that all eight LTK tested mutations were responsible for lorlatinib resistance, among which the L650F mutation showed the most robust resistance to lorlatinib. We also demonstrated that gilteritinib was an exquisite and potent inhibitor of CLIP1-LTK-L650F in in vivo and in vitro experiments.
LTK fusion is a rare but actionable oncogenic driver in NSCLC[
In this study, we showed that gilteritinib can inhibit kinase activity of CLIP1-LTK with LTK mutations including L650F, as well as WT CLIP1-LTK. Gilteritinib may be an alternative option for LTK fusion-positive NSCLC as either a first TKI treatment or second TKI after lorlatinib treatment.
Considering similar pattern of drug sensitivity between ALK and LTK, and the rarity of patients harboring specific resistance mechanisms, basket-type trials of targeted therapy for patients with specific resistance mutations might be useful, for example for L1256F/L650F mutated ALK/LTK fusion-positive NSCLC to efficiently develop targeted therapy for rare fusion-positive NSCLC resistant to prior targeted therapies.
In summary, LTK mutations analogous to ALK mutations were resistant to lorlatinib, with the L650F mutation being the most potent. Our preclinical models demonstrate that gilteritinib may be a promising strategy to overcome L650F-mediated resistance.
NIH3T3 cells were purchased from American Type Culture Collection (ATCC). Ba/F3, WEHI, and BOSC23 cells were kindly provided by Dr. Daniel G. Tenen (Harvard Medical School). Crizotinib, ceritinib, alectinib, brigatinib, lorlatinib, entrectinib, gilteritinib, and repotrectinib were purchased from Selleck. NIH3T3 cells were maintained in Dulbecco's Modified Eagle Medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 100 units/ml penicillin, and 100 mg/ml streptomycin (P/S). Parental Ba/F3 cells were maintained in RPMI1640 supplemented with 5% WEHI (as a source of IL-3), 10% FBS, and P/S. Ba/F3 cells expressing CLIP1-LTK mutants were maintained in RPMI1640 supplemented with 10% FBS and P/S. All cell lines were routinely tested for mycoplasma infection and negative for mycoplasma infection.
The MIGR1 retroviral vector harboring CLIP1-LTK fusion protein was constructed as previously described[
Ba/F3 and NIH3T3 cells expressing WT CLIP1-LTK fusion or various mutant CLIP1-LTK fusions were generated by retroviral transduction as previously described[
Cells were lysed in sodium dodecyl sulfate (SDS) sample buffer and boiled for 5 min. Lysates were subjected to SDS polyacrylamide gel electrophoresis and blotted onto poly (vinylidene fluoride) (PVDF) membranes (Millipore). The antibodies and dilutions used are listed in Supplementary Table 2. Images were captured using ImageQuant LAS 4000 (GE Healthcare) and analyzed using the ImageJ software (ver. 1.53). All images were assembled, and figures were generated using the Affinity Designer (ver. 1.10.5), and Microsoft PowerPoint 2016 (ver. 2108).
Ba/F3 cells (
NIH3T3 cells expressing CLIP1-LTK-L650F (
Ba/F3 cells (
The Caspase-Glo3/7 Assay System (Promega) was used to evaluate cell apoptosis. Ba/F3 cells (5000 cells/well) were seeded in 96-well plates and treated with the indicated drugs for 12 h. Data were captured using the Spectra Max Paradigm (Molecular Devices) with SoftMax Pro software (ver.7.10). Absorbance was measured at 490 nm.
Molecular docking of alectinib, gilteritinib, and lorlatinib with the LTK-tyrosine kinase domain was performed using GOLD 5.5. Standard default settings for the genetic algorithm were used. The structure of the LTK kinase domain was predicted using AlphaFold2[
Each of I565N, F568C, L590M, L592F, G596R, D597N, L650F, and G663A mutations were introduced into the structural model of WT LTK using the MODELER program[
The LTK-drug ΔG was calculated using MP-CAFEE, which is one of the chemical free energy perturbation methods[
The Institutional Animal Care and Use Committee of the National Cancer Center (K20-009) approved all the animal experiments. We have complied with all relevant ethical regulations for animal use. To establish tumor xenografts, NIH3T3 cells transduced with CLIP1-LTK-L650F were transplanted into the flanks of athymic nude mice (female, 8-weeks old BALB/cAJcl-Foxn1
The group size was based on previous experience. Unless otherwise noted, each experiment was repeated three or more times with similar results. One-way ANOVA and post-hoc analysis, including Dunnett's test and Tukey's test, were used to determine statistical significance among more than three groups. All statistical analyses were conducted on data from three or more biologically independent experimental replicates using the GraphPad Prism software (ver. 9.3.1). Statistical significance was set at p < 0.05.
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
We thank Dr. Takashi Kohno at National Cancer Research Institute for helpful discussion. We also thank Ms. Yuri Murata and PREMIA for administrative assistance with managing clinical samples, molecular screening and the clinico-genomic database in LC-SCRUM-Asia and the members of the Division of Translational Genomics, Exploratory Oncology Research and Clinical Trial Center, and National Cancer Center for helpful discussion for valuable comments on the manuscript. This study was supported by MEXT/JSPS KAKENHI (JP20K17215 to H.I., JP21K06510 to M.A., 16K21746 to S.S.K., and 22H03084 to S.S.K.), JSPS Bilateral Joint Research Projects grant number 120207408 (S.S.K.), Princess Takamatsu Cancer Research Fund 18-250 (S.S.K.), the National Cancer Center Research and Development Fund 31-A-6 (S.S.K) and National Institute of Health 1R01CA240257 (S.S.K.). This study was also supported by MEXT as "Program for Promoting Researches on the Supercomputer Fugaku (Application of Molecular Dynamics Simulation to Precision Medicine Using Big Data Integration System for Drug Discovery)" (Y.O.), and FOCUS Establishing Supercomputing Center of Excellence (Y.O.). This research used computational resources of the supercomputer Fugaku provided by the RIKEN Center for Computational Science through the HPCI System Research Project (Project ID: hp210172 and hp220164).
S.Mo., H.I., S.Ma., and S.S.K conceived the study and designed the experiments. M.A. and Y.O. designed and supervised the simulation. Y.I. modeled the LTK structure, B.M. and Y.Sas. performed molecular docking, and Y.Sas. performed molecular dynamics simulation (MP-CAFEE). S.Mo., H.I., and J.L. performed cloning and mutagenesis of the expression constructs for in vitro analysis. S.M, J.L., Y.K. and S.K. generated stable cell lines. S.Mo., H.I., J.L., Y.K., Y.Sak., K.T., S.Y., Y.T. H.U. and S.S.K. performed biochemical analysis. K.Y., K.G. and S.S.K. supervised this project. S.Mo., H.I., S.Ma., Y.O. and S.S.K wrote the manuscript with input from all the authors.
Communications Biology thanks the anonymous reviewers for their contribution to the peer review of this work. Primary Handling Editors: Toril Holien and Christina Karlsson Rosenthal.
The sequence data used in this study are publicly available in the National Center for Biotechnology Information (https://
The authors declare the following competing interests. S.Mo. reports no conflicts of interest in this study. Y.T. reports personal fees (honoraria) from Chugai, Eli Lilly, AstraZeneca, Taiho. H.I. reports research support from Amgen, Ono, Takeda, Eisai and personal fees (honoraria) from Ono, Chugai, AstraZeneca, Merck. Y.S. reports research support from Ono, MSD, and personal fees (honoraria) from Ono, Chugai, AstraZeneca, Eli Lilly, Bristol-Myers Squibb, Pfizer. H.U. reports research support from Takeda, Boehringer Ingelheim, Taiho and personal fees (honoraria) from Taiho. S.Ma. reports research support from Chugai, Novartis, Eli Lilly, Merck, MSD, and personal fees (honoraria) from AstraZeneca, Chugai, Novartis, Pfizer and Eli Lilly. K.Y. reports research support from AstraZeneca, Eli Lilly, Phizer, Diichi sankyo, Abbvie, Taiho, MSD, Takeda, Chugai, and personal fees (honoraria) from Chugai, AstraZeneca, Bristol-Myers Squibb, Daiichi sankyo, Janssen, Eli Lilly, Taiho, Novaritis, Kyowa kirin, Boehringer Ingelheim. G.K. reports research support from Amgen, Amgen Astellas BioPharma, AstraZeneca, Bayer, Boehringer Ingelheim Japan, Bristol-Myers Squibb, Blueprint Medicines, Chugai, Daiichi sankyo, Eisai, Eli Lilly, Haihe Biopharma, Ignyta, Janssen, KISSEI, Kyowa Kirin, Life Technologies, Loxo Oncology., Medical & Biological Laboratories, Merck, Merus, MSD, NEC Corporation, Novartis, Ono, Pfizer, Sumitomo Dainippon, Spectrum Pharmaceuticals, Sysmex Corporation, Taiho, Takeda, Turning Point Therapeutics, and personal fees (honoraria) from Amgen, Amoy Diagnosties, Amgen Astellas BioPharma, AstraZeneca, Bayer, Boehringer Ingelheim, Bristol-Myers Squibb, Chugai, Daiichi sankyo, Eisai, Eli Lilly Japan, Guardant Health, Janssen, Thermo Fisher Scientifi, Medpace, Merck, MSD, Novartis Pharma, Ono, Otsuka, Taiho, and Takeda. SSK reports grants from Boehringer Ingelheim, MiRXES, Johnson&Johnson, and Taiho Therapeutics, as well as personal fees from AstraZeneca, Boehringer Ingelheim, Bristol Meyers Squibb, Chugai Pharmaceutical, and Takeda Pharmaceuticals plus royalties from Life Technologies. Other authors declare no conflicts of interest.
Graph: Supplementary Information
Graph: Description of Additional Supplementary Files
Graph: Supplementary Data 1
Graph: Reporting Summary
The online version contains supplementary material available at https://doi.org/10.1038/s42003-024-06116-6.
Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
By Shunta Mori; Hiroki Izumi; Mitsugu Araki; Jie Liu; Yu Tanaka; Yosuke Kagawa; Yukari Sagae; Biao Ma; Yuta Isaka; Yoko Sasakura; Shogo Kumagai; Yuta Sakae; Kosuke Tanaka; Yuji Shibata; Hibiki Udagawa; Shingo Matsumoto; Kiyotaka Yoh; Yasushi Okuno; Koichi Goto and Susumu S. Kobayashi
Reported by Author; Author; Author; Author; Author; Author; Author; Author; Author; Author; Author; Author; Author; Author; Author; Author; Author; Author; Author; Author