Drug lead compounds with ALK G1202R / L1196M double mutation inhibitory activity, virtual screening methods and applications

Abstract

This application relates to a drug lead compound with ALK G1202R / L1196M double mutation inhibitory activity, a virtual screening method, and its application. The method includes: obtaining resolved single-mutant complex structures from a database; constructing screening models based on the complex structures and performing standardized preprocessing on the crystal structures in each model; optimizing the structure and preparing the conformation of the compound library to be screened; determining the docking method, screening the optimal screening model, clustering the optimal screening model, docking it with the compound library to be screened, and obtaining drug lead compounds with ALK G1202R / L1196M double mutation inhibitory activity through drug-like property filtering, ligand structure clustering, and ADMET property prediction. The technical solution of this application can accurately screen ALK G1202R / L1196M double mutation inhibitory lead compounds, shortening the research and development cycle, reducing costs, and has clinical translational value.

Description

Technical Field

[0001] This application relates to the field of biomedical technology, and in particular to a drug lead compound with ALK G1202R / L1196M double mutation inhibitory activity, a virtual screening method, and its application. Background Technology

[0002] Anaplastic lymphoma kinase (ALK) fusion mutations are one of the key driver genes in advanced non-small cell lung cancer (NSCLC), accounting for approximately 5%-7% of all NSCLC patients. Tyrosine kinase inhibitors (TKIs) targeting ALK fusion mutations have become a core clinical treatment approach, significantly prolonging patient survival. However, acquired resistance commonly develops after long-term use, with ALK kinase domain mutations being the primary resistance mechanism. In particular, the double mutation formed by G1202R and L1196M significantly alters the ALK protein binding pocket conformation, leading to strong resistance to most existing TKIs and posing a major challenge to clinical treatment. Developing specific inhibitors targeting this double mutation and efficient virtual screening methods has become an urgent need in the field of ALK targeted therapy, and is of great significance for overcoming the clinical resistance dilemma and improving patients' quality of life.

[0003] In current technologies, ALK inhibitors have evolved to three generations. The first-generation drug crizotinib pioneered ALK-targeted therapy. While second- and third-generation drugs such as alectinib and lolatinib are effective for some single-mutation resistant patients, their inhibitory activity against the G1202R / L1196M double mutation is significantly insufficient, failing to meet clinical treatment needs. In terms of virtual screening, conventional methods often rely on molecular docking with a single static protein structure, failing to fully consider the dynamic conformational changes of proteins under physiological conditions. Furthermore, the design of drug-like property filtering and structural diversity assurance in the screening process is imperfect, resulting in low screening efficiency and poor accuracy. A large number of compounds without drug potential enter the subsequent validation stage, significantly prolonging the development cycle and increasing development costs. At the same time, existing drug lead compounds with ALK G1202R / L1196M double mutation inhibitory activity often have problems such as poor drug-like properties and insufficient specificity for the double mutation target, making it difficult to achieve clinical translation and failing to cover the needs of ALK double-mutation resistant NSCLC patients with different pathological subtypes such as squamous cell carcinoma and adenocarcinoma.

[0004] It should be noted that the information disclosed in the background section above is only used to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0005] To address the aforementioned issues, this application provides a drug lead compound with ALK G1202R / L1196M double mutation inhibitory activity, a virtual screening method, and its application. This method can accurately screen ALK G1202R / L1196M double mutation inhibitory lead compounds, solve the drug resistance problem, shorten the research and development cycle, reduce costs, and has clinical translational value.

[0006] To achieve the objectives of this application, the following technical solution is provided: In a first aspect, this application provides a virtual screening method for drug lead compounds possessing ALK G1202R / L1196M double mutation inhibitory activity, comprising: The resolved structures of the ALK L1196M single mutant protein-small molecule complex were obtained from the RCSB PDB database. The structures of the ALK L1196M single mutant protein-small molecule complex include PDBID: 4CLJ, resolution: 1.66 Å; PDBID: 2YFX, resolution: 1.7 Å; and PDBID: 4CD0, resolution: 2.23 Å. Screening models were constructed based on the structure of the ALK L1196M single mutant protein-small molecule complex, and the crystal structures of the protein-inhibitor complexes in each model were standardized and preprocessed. The screening models included: 2YFX-Crizotinib, 4CLJ-Lorlatinib, 4CLJ-TPX0005, 4CLJ-TPX0131 and 4CD0-8e. Structural optimization and conformation preparation were performed on the library of compounds to be screened. The docking method was determined, and the optimal screening model was selected from the standardized preprocessed screening models. After molecular dynamics trajectory clustering, the optimal screening model was docked with the library of compounds to be screened. Drug lead compounds with ALK G1202R / L1196M double mutation inhibitory activity were screened by filtering for drug-like properties, clustering of ligand structures and prediction of ADMET properties.

[0007] In one possible implementation, in the step of obtaining the resolved ALK L1196M single mutant protein-small molecule complex structure from the RCSB PDB database, the residue G1202 in the ALK L1196M single mutant protein structure is mutated to R1202 using the Mutate Residue module of the Schrodinger software. The structure of the mutated double mutant protein is then optimized by molecular dynamics (MD) simulation to adjust the spatial conformation of the newly added side chain to a stable state.

[0008] In one possible implementation, the standardized preprocessing involves using the Protein PreparationWizard module to remove water of crystallization molecules and impurity ligands, repair the loop region conformation, adjust the bond order, determine the protonation state of residues based on the hydrogen bonding mode of the original crystal structure, and perform energy optimization on all heavy atoms under the OPLS3 force field to achieve a root mean square deviation ≤ 0.3 Å.

[0009] In one possible implementation, in the step of structure optimization and conformation preparation of the compound library to be screened, SDF format molecules from the ChemDiv database are selected as the screening database, and the structure is preprocessed using the LigPrep module to remove duplicate molecules and unreasonable chemical bonds, correct tautomers and chiral centers, and retain the 10 lowest energy dominant conformations during conformation optimization for each small molecule.

[0010] In one possible implementation, the steps of determining the docking method, selecting the optimal screening model from the standardized preprocessed screening models, performing molecular dynamics trajectory clustering on the optimal screening model and docking it with the library of compounds to be screened, and screening for drug lead compounds with ALKG1202R / L1196M double mutation inhibitory activity through drug-like property filtering, ligand structure clustering and ADMET property prediction include: The standard precision (SP) and precision (XP) modes of the Glide module were used for re-docking evaluation, and the optimal docking method was determined by using the rationality of ligand binding conformation and docking score as indicators. For each screening model, a 100ns dynamic simulation was performed. The k-means algorithm was used to cluster the molecular dynamic trajectories into 5 categories. Based on the clustered conformations, the optimal screening model was selected by docking with the library of compounds to be screened. Dual filtering is performed based on the Lipinski five-rule and the Jorgensen three-rule; the Lipinski five-rule includes mol_MW<500, QPlogPo / w<5, donorHB≤5, accptHB≤10, and the Jorgensen three-rule includes QPlogS>-5.7, QPPCaco>22, and Metab<7. The three-dimensional coordinates of the top 500 docking scores of the ligands were imported into Canvas software, the cluster count was set to 50, and structural clustering was performed based on the Tanimoto coefficient to ensure the diversity of the drug lead compounds with ALK G1202R / L1196M double mutation inhibitory activity; wherein, the Tanimoto coefficient is ≥0.8. The QikProp module was used to calculate the absorption, distribution, metabolism, excretion, and toxicity parameters of the compounds, and to screen for drug lead compounds with ALK G1202R / L1196M double mutation inhibitory activity.

[0011] In one possible implementation, the method further includes: verifying the in vitro bioactivity of the drug lead compound with ALK G1202R / L1196M double mutation inhibitory activity using wet experiments; evaluating the binding stability of the drug lead compound with ALK G1202R / L1196M double mutation inhibitory activity to the double mutant protein using molecular dynamics simulations; and verifying the binding affinity and stability of the drug lead compound with ALK G1202R / L1196M double mutation inhibitory activity using energy calculations, thereby determining the inhibitory effect and binding process of the drug lead compound with ALK G1202R / L1196M double mutation inhibitory activity on the double mutant protein.

[0012] In one possible implementation, the wet assay involves using a kinase activity assay kit to determine the inhibitory activity of the drug lead compound with ALK G1202R / L1196M double mutation inhibitory activity against the ALK G1202R / L1196M double mutant protein. The system was constructed through a three-stage minimization method using molecular dynamics simulations. The system energy was optimized by steepest descent and conjugate gradient, and the system temperature was gradually heated from 0 K to 310 K within 200 ps. The steepest descent took 30,000 steps, and the conjugate gradient took 70,000 steps. The energy calculation was performed using the MM / GBSA method to calculate the binding free energy between the drug lead compound with ALK G1202R / L1196M double mutation inhibitory activity and the double mutant protein.

[0013] Secondly, this application provides a drug lead compound with ALK G1202R / L1196M double mutation inhibitory activity, which is obtained by screening using the virtual screening method for drug lead compounds with ALK G1202R / L1196M double mutation inhibitory activity described above.

[0014] In one possible implementation, the chemical structure of the drug lead compound having ALK G1202R / L1196M double mutation inhibitory activity is shown in Formula (I): .

[0015] Thirdly, this application provides the use of a drug lead compound with ALK G1202R / L1196M double mutation inhibitory activity in the preparation of an anti-non-small cell lung cancer (NSCLC) drug, wherein the non-small cell lung cancer includes squamous cell carcinoma, adenocarcinoma and large cell carcinoma.

[0016] The technical solution provided in this application may include the following beneficial effects: This application provides a drug lead compound with ALK G1202R / L1196M double mutation inhibitory activity, a virtual screening method, and its application. By constructing a multi-conformation hierarchical virtual screening system—integrating docking mode optimization, dynamic conformation clustering, dual drug-likeness rule filtering, ligand structure clustering, and ADMET property prediction—combined with OPLS3 force field and MM / GBSA energy calculation techniques, it achieves efficient enrichment of high-potential molecules from a million-level compound library, significantly improving the accuracy and efficiency of lead compound screening and avoiding the blindness of traditional screening methods. Furthermore, the screened lead compound of formula (I)... The compound can specifically bind to key sites such as Arg1202 and Met1199 of the double-mutant protein, forming stable hydrogen bonds and hydrophobic interactions. This addresses the clinical pain point of existing ALK inhibitors failing to inhibit this drug-resistant mutation. Furthermore, in vitro kinase activity and cell proliferation experiments have verified its significant inhibitory activity. At the same time, it significantly shortens the drug development cycle and reduces the cost of in vitro validation. Its application can cover most ALK double-mutant resistant non-small cell lung cancer subtypes, such as squamous cell carcinoma and adenocarcinoma, providing new treatment options for clinically resistant patients. It has significant clinical translational value and market prospects.

[0017] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Attached Figure Description

[0018] The accompanying drawings are provided to further illustrate the present application and form part of the specification. They are used together with the embodiments of the present application to explain the application and do not constitute a limitation thereof. Obviously, the drawings described below are merely some embodiments of this disclosure, and those skilled in the art can obtain other drawings based on these drawings without any creative effort.

[0019] Figure 1 A schematic flowchart of a virtual screening method for drug lead compounds with ALK G1202R / L1196M double mutation inhibitory activity provided for embodiments of this application; Figure 2 A hierarchical flowchart based on multiple conformations is provided for a virtual screening method of drug lead compounds with ALK G1202R / L1196M double mutation inhibitory activity, which is provided for the embodiments of this application. Figure 3 A schematic diagram illustrating the inhibition of wild-type and double-mutant cell proliferation by 38 compounds in a virtual screening method for drug lead compounds with ALK G1202R / L1196M double mutation inhibitory activity, provided for embodiments of this application. Figure 4 The effect of a drug lead compound with ALK G1202R / L1196M double mutation inhibitory activity on the viability of Ba / F3-EML4-ALK-L1196M / G1202R and Ba / F3-EML4-ALK-WT cells, as determined by MTT assay in this application embodiment, is related to the IC50 of the lead compound. 50 A diagram illustrating the values; Figure 5 This application provides a virtual screening method for drug lead compounds with ALK G1202R / L1196M double mutation inhibitory activity, which simulates the dynamic interaction mode between the lead compound and ALK G1202R / L1196M during the MD simulation process. Detailed Implementation

[0020] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, they are provided so that this disclosure will be more comprehensive and complete, and will fully convey the concept of the exemplary embodiments to those skilled in the art. The described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.

[0021] This example embodiment first provides a virtual screening method for drug lead compounds with ALK G1202R / L1196M double mutation inhibitory activity, including: Step S100: Obtain the resolved ALK L1196M single mutant protein-small molecule complex structure from the RCSB PDB database; wherein, the ALK L1196M single mutant protein-small molecule complex structure includes PDBID: 4CLJ, resolution: 1.66 Å, PDBID: 2YFX, resolution: 1.7 Å, PDBID: 4CD0, resolution: 2.23 Å.

[0022] Step S200: Construct screening models based on the structure of the ALK L1196M single mutant protein-small molecule complex, and perform normalization preprocessing on the crystal structure of the protein-inhibitor complex in each model; the screening models include: 2YFX-Crizotinib, 4CLJ-Lorlatinib, 4CLJ-TPX0005, 4CLJ-TPX0131 and 4CD0-8e.

[0023] Step S300: Optimize the structure and prepare the conformation of the compound library to be screened.

[0024] Step S400: Determine the docking method, select the optimal screening model from the standardized preprocessed screening models, perform molecular dynamics trajectory clustering on the optimal screening model and dock it with the library of compounds to be screened, and screen drug lead compounds with ALK G1202R / L1196M double mutation inhibitory activity through drug-like property filtering, ligand structure clustering and ADMET property prediction.

[0025] In one embodiment, in the step of obtaining the resolved ALK L1196M single mutant protein-small molecule complex structure from the RCSB PDB database, the Mutate Residue module of the Schrodinger software is used to mutate residue G1202 in the ALKL1196M single mutant protein structure to R1202, and the molecular dynamics simulation optimization is performed on the mutated double mutant protein structure to adjust the spatial conformation of the newly added side chain to a stable state.

[0026] It should be noted that the original crystal structure of the ALK G1202R / L1196M double mutant protein contained only the L1196M single mutation. The addition of the side chain conformation was optimized through MD simulation, which ensures that the double mutant protein structure conforms to the stable conformation under physiological conditions, providing a reliable target basis for subsequent virtual screening.

[0027] In one embodiment, the standardized preprocessing involves using the Protein Preparation Wizard module to remove water of crystallization molecules and impurity ligands, repair the loop region conformation, adjust the bond order, determine the protonation state of residues based on the hydrogen bonding mode of the original crystal structure, and optimize the energy of all heavy atoms under the OPLS3 force field to a root mean square deviation ≤ 0.3 Å.

[0028] It should be noted that the core purpose of standardized pretreatment is to eliminate redundant information (such as water of crystallization molecules and impurity ligands) and conformational defects (such as unstable conformations in the loop region) in the crystal structure. Optimization to RMSD≤0.3Å through OPLS3 force field is a recognized standard for protein structure stability in the field. It can ensure that the binding pocket of the screening model is clear and the atomic coordinates are accurate, providing high-quality targets for molecular docking.

[0029] In one embodiment, in the step of structure optimization and conformation preparation of the compound library to be screened, SDF format molecules from the ChemDiv database are selected as the screening database, and the structure is preprocessed by the LigPrep module to remove duplicate molecules and unreasonable chemical bonds, correct tautomers and chiral centers, and retain the 10 lowest energy dominant conformations when optimizing the conformation of each small molecule.

[0030] It should be noted that the ChemDiv database was chosen because it contains more than 1.7 million structurally diverse small molecules, which is suitable for large-scale virtual screening needs; the LigPrep module can unify the standard of compound structure, correct chirality and tautomerism, and retain the 10 lowest-energy dominant conformations, which can cover the main conformation types when small molecules bind to proteins, avoiding docking errors caused by conformational uniformity.

[0031] In one embodiment, step S400 includes: In step S410, the standard precision and accurate modes of the Glide module are used to evaluate the re-docking, and the optimal docking method is determined by using the rationality of the ligand binding conformation and the docking score as indicators.

[0032] It should be noted that the standard precision and exact modes of the Glide module are used for re-docking evaluation because they are the mainstream modes for molecular docking in the industry: the SP mode has high computational efficiency and is suitable for large-scale docking scenarios, while the XP mode has higher precision and can capture more refined protein-ligand interactions. Using the rationality of ligand binding conformation as an indicator means comparing the spatial fit between the ligand and the protein binding pocket, such as whether hydrogen bonds or hydrophobic interactions are formed with key residues, to determine whether the binding mode conforms to the known action rules of ALK inhibitors. Using docking score as an indicator is to quantify the binding potential of ligand and protein, and finally determine the optimal docking method. This can balance computational efficiency and result accuracy in subsequent large-scale database docking, and avoid screening bias caused by a single mode.

[0033] In step S420, a 100ns dynamic simulation is performed on each screening model. The k-means algorithm is used to cluster the molecular dynamic trajectories into 5 categories. Based on the clustered conformations, the optimal screening model is selected by docking with the library of compounds to be screened.

[0034] It should be noted that a 100ns dynamic simulation was performed on each screening model because the 100ns duration is sufficient to capture the dynamic conformational changes of proteins under physiological conditions, such as the flexible fluctuations of the binding pocket, avoiding omissions of binding modes due to reliance on static crystal structures. The k-means algorithm was used to cluster the trajectories into 5 classes, which is based on the industry standard for clustering and can cover the main stable conformational types of proteins. This avoids both conformational homogeneity caused by too few clusters and computational burden caused by too many clusters. The docking of the clustered conformations with the library of compounds to be screened simulates the dynamic binding characteristics of proteins under physiological conditions, rather than relying solely on static structures. Finally, the optimal model was selected to ensure that the target conformations of subsequent virtual screening are closer to the real state in vivo, improving the reliability of the screening results.

[0035] In step S430, dual filtering is performed based on the Lipinski five-rule and the Jorgensen three-rule; the Lipinski five-rule includes mol_MW<500, QPlogPo / w<5, donorHB≤5, accptHB≤10, and the Jorgensen three-rule includes QPlogS>-5.7, QPPCaco>22, and Metab<7.

[0036] It's important to note that the core functions of each parameter in Lipinski's five rules are as follows: mol_MW < 500 limits the molecular weight of the compound, preventing excessively large molecules from hindering intestinal absorption; QPlogPo / w < 5 controls the lipid-water partition coefficient, ensuring transmembrane transport capacity; donorHB < 5 and acptHB ≤ 10 limit the number of hydrogen bond donors / acceptors, reducing the risk of excessively rapid metabolic clearance; in Jorgensen's three rules, QPlogS > -5.7 ensures the compound's water solubility, preventing excessively low in vivo dissolution; QPPCaco > 22 ensures intestinal permeability, guaranteeing oral bioavailability; and Metab > 7 improves metabolic stability and prolongs the duration of action in vivo. Employing a dual-filter approach rather than a single rule simultaneously covers both the basic characteristics of drug-like compounds and core pharmacokinetic indicators, significantly reducing the number of compounds without drug potential and lowering resource consumption for subsequent experimental validation.

[0037] In step S440, the three-dimensional coordinates of the top 500 docking scores of the ligands are imported into the Canvas software, the cluster count is set to 50, and structural clustering is performed based on the Tanimoto coefficient to ensure the diversity of the drug lead compounds with ALK G1202R / L1196M double mutation inhibitory activity; wherein, the Tanimoto coefficient is ≥0.8.

[0038] It should be noted that Canvas software is a commonly used compound structure clustering tool in the biopharmaceutical field, which can accurately calculate structural similarity based on molecular fingerprints. Setting the cluster count to 50 is a balance between ensuring structural diversity and controlling the total number of candidates. 50 classes cover compounds with different chemical cores while avoiding excessive clustering that would lead to a surge in subsequent validation workload. Limiting the Tanimoto coefficient to ≥0.8 is a common threshold in the industry for defining highly similar compound structures. Grouping ligands with a similarity of ≥0.8 into one class can avoid the presence of compounds with repetitive structures and homogeneous mechanisms of action in the screening results. Clustering the top 500 high-potential ligands with docking scores balances binding potential and structural diversity, ensuring that the final drug lead compounds with ALKG1202R / L1196M double mutation inhibitory activity not only have the ability to bind to the double mutant protein but also cover a wider chemical space, increasing the probability of screening highly efficient and low-toxicity lead compounds.

[0039] In step S450, the QikProp module is used to calculate the absorption, distribution, metabolism, excretion and toxicity parameters of the compound, and to screen the drug lead compound with ALK G1202R / L1196M double mutation inhibitory activity.

[0040] It should be noted that the QikProp module is a commonly used ADMET property prediction tool in the biopharmaceutical field. Its calculated parameters such as absorption and distribution are trained based on experimental data from a large number of marketed drugs, and the prediction results have industry recognition. Screening drug lead compounds with ALK G1202R / L1196M double mutation inhibitory activity through this module allows for early assessment of the pharmacokinetic behavior and safety of compounds in vivo, such as whether there is hepatotoxicity and whether they can effectively cross the blood-brain barrier. This avoids the situation where resources are invested in subsequent wet experiments only to find that the compound cannot be developed into a drug due to ADMET property defects, which significantly shortens the drug development cycle and reduces costs.

[0041] In one embodiment, the virtual screening method for the drug lead compound with ALK G1202R / L1196M double mutation inhibitory activity further includes: verifying the in vitro bioactivity of the drug lead compound with ALK G1202R / L1196M double mutation inhibitory activity through wet experiments; evaluating the binding stability of the drug lead compound with ALK G1202R / L1196M double mutation inhibitory activity to the double mutant protein through molecular dynamics simulations; and verifying the binding affinity and stability of the drug lead compound with ALK G1202R / L1196M double mutation inhibitory activity through energy calculations, thereby determining the inhibitory effect and binding process of the drug lead compound with ALK G1202R / L1196M double mutation inhibitory activity on the double mutant protein.

[0042] The wet assay involves using a kinase activity assay kit to determine the inhibitory activity of the drug lead compound with ALK G1202R / L1196M double mutation inhibitory activity on the ALK G1202R / L1196M double mutant protein.

[0043] It should be noted that the use of a kinase activity assay kit to determine the inhibitory activity of a drug lead compound with ALK G1202R / L1196M double mutation inhibitory activity against the ALK G1202R / L1196M double mutant protein is because this kit can specifically target the catalytic domain of ALK kinase, directly reflecting the compound's inhibitory ability on the function of the double mutant protein kinase. This is the core mechanism by which ALK inhibitors exert their therapeutic effects. Compared with non-specific cell activity assays, this method can eliminate interference from other targets, accurately verify its inhibitory specificity against the core target of this invention, and avoid screening out active compounds that do not target the double mutant protein.

[0044] The system is constructed through a three-stage minimization method using molecular dynamics simulations. The system energy is optimized by steepest descent and conjugate gradient, and the system temperature is gradually heated from 0K to 310K within 200ps. The steepest descent takes 30,000 steps, and the conjugate gradient takes 70,000 steps.

[0045] It should be noted that the three-stage minimization construction method is used to gradually eliminate the steric hindrance and energy conflict of the protein-compound complex: the steepest descent of 30,000 steps can quickly alleviate the obvious problems of atomic overlap and chemical bond tension in the system, while the conjugation gradient of 70,000 steps can more finely optimize the system energy and ensure the stability of the initial structure in subsequent simulations; the system temperature is gradually heated from 0K to 310K within 200ps, which is the physiological temperature of the human body, to simulate the gradual process of temperature change in the body, avoid sudden rises that would cause drastic fluctuations in the system conformation, and ensure that the subsequent 100ns kinetic simulation can realistically reflect the binding behavior of the compound and the double mutant protein under physiological conditions.

[0046] The energy calculation employs the MM / GBSA method to calculate the binding free energy between the drug lead compound with ALK G1202R / L1196M double mutation inhibitory activity and the double mutant protein.

[0047] It should be noted that the MM / GBSA method is a classic algorithm for quantifying protein-ligand binding free energy. Its binding free energy result integrates core energy terms such as molecular mechanical energy and solvation energy, and can reflect the tightness of the binding between the compound and the double mutant protein from an energy perspective. The calculation of this binding free energy is to corroborate the binding stability results of molecular dynamics simulation. The more negative the binding free energy, the stronger the binding interaction and the higher the stability, forming a dual verification of dynamic conformational stability and energy stability. This eliminates false positive compounds with occasional in vitro activity and improves the reliability of lead compound development.

[0048] Furthermore, this example embodiment also provides a drug lead compound with ALK G1202R / L1196M double mutation inhibitory activity, which is obtained by screening using the virtual screening method for drug lead compounds with ALK G1202R / L1196M double mutation inhibitory activity described above.

[0049] The chemical structural formula of the drug lead compound with ALK G1202R / L1196M double mutation inhibitory activity is shown in Formula (I): .

[0050] It should be noted that the drug lead compound shown in formula (I) was obtained through a virtual screening method using the aforementioned drug lead compounds with inhibitory activity against the ALK G1202R / L1196M double mutation. The functional groups in its structure, such as the methoxybenzene ring and the N,N-dimethylamino side chain, can form stable hydrogen bonds and hydrophobic interactions with key residues such as Arg1202 and Met1199 of the ALK G1202R / L1196M double mutant protein. Compared with existing ALK inhibitors, this compound has specific inhibitory activity against the G1202R / L1196M double mutation, which can solve the problem of existing drugs being ineffective against this drug-resistant mutation.

[0051] Furthermore, this example embodiment also provides the application of a drug lead compound having ALK G1202R / L1196M double mutation inhibitory activity in the preparation of an anti-non-small cell lung cancer drug, wherein the non-small cell lung cancer includes squamous cell carcinoma, adenocarcinoma and large cell carcinoma.

[0052] It should be noted that ALK gene fusion mutations are one of the key driver mutations in advanced non-small cell lung cancer (NSCLC), with approximately 5%-7% of NSCLC patients carrying this mutation. The G1202R / L1196M double mutation is a common resistance mutation type in NSCLC patients after treatment with first- or second-generation ALK inhibitors, with a resistance rate of approximately 30%. Existing drugs show significantly reduced inhibitory activity against this double mutation. The drug lead compound of this invention exhibits specific inhibitory activity against this resistance double mutation and can therefore be used to prepare drugs against this type of drug-resistant NSCLC. Squamous cell carcinoma, adenocarcinoma, and large cell carcinoma are the three major pathological subtypes of NSCLC, covering approximately 90% of clinical NSCLC cases; therefore, the application of this compound can cover the majority of NSCLC patients with ALK double mutation resistance.

[0053] The drug lead compound (AMI007) with ALK G1202R / L1196M double mutation inhibitory activity described in this application is explained in detail below through specific examples: Example 1: See Figure 2 A virtual screening method for inhibitors targeting the ALK G1202R / L1196M double mutation includes: (1) Screening of crystal structures: The resolved structures of the ALK L1196M single mutant protein-small molecule complex were obtained from the RCSB PDB database (PDBID: 4CLJ, resolution: 1.66 Å; PDBID: 2YFX, resolution: 1.7 Å; PDBID: 4CD0, resolution: 2.23 Å). (2) Construction of screening models: Double mutant structures were obtained by manual mutation based on the resolved ALK L1196M single mutant protein structure (PDBID: 4CLJ; PDBID: 2YFX; PDBID: 4CD0). According to the literature review, a total of five screening models were constructed: 2YFX-Crizotinib, 4CLJ-TPX0005, 4CLJ-Lorlatinib, 4CLJ-TPX0131, and 4CD0-8e. The ligand structures Crizotinib (CID: 11626560), Lorlatinib (CID: 71731823), 8e (CID: 72710568), TPX-0005 (CID: 135565923), and TPX-0131 (CID: 156024486) used to construct the complex models were obtained from the PubChem database in 3D.

[0054] (3) Protein structure preparation: The amino acid residues of the protein in each model were renumbered, and redundant water molecules, buffer molecules and impurity ligands in the model were removed using the ProteinPreparation Wizard module; hydrogen atoms were added; the conformation of the amino acid side chain was optimized; the OPLS3 force field was used for energy minimization (convergence criterion: RMSD≤0.3 Å) to obtain a stable protein structure model.

[0055] (4) Download the library of compounds to be screened and perform structure optimization and conformation preparation: The ChemDiv database contains 175,6882 molecules. After removing duplicate molecules and molecules containing unreasonable chemical bonds, 173,5332 molecules remain. The Ligprep module was used to add hydrogen atoms and charges to all ligands, correct tautomers and chiral centers, and retain 10 3D dominant conformations for each small molecule during conformation optimization. The force field was configured as OPLS3, and the protonation state was determined at pH = 7.0 ± 2.0.

[0056] (5) Hierarchical virtual screening workflow based on multiple conformations: By integrating multiple molecular simulation techniques, compounds with potential biological activity are screened from a large set of molecules, providing high-quality compounds for subsequent experimental verification.

[0057] First, different docking simulations were used to ensure that small molecules and proteins could bind correctly and form reasonable binding conformations, thereby determining the optimal docking mode. Molecular docking calculations were performed using the Glide module of Schrödinger software. Docking boxes were generated using the Receptor Grid Generation module, with the inhibitor in the binding pocket as the center, and the van der Waals radius scale factor set to 1.0 and the maximum local charge of atoms set to 0.25. Co-crystallized ligand re-docking experiments were then conducted, re-docking the ligands to the active site of the double-mutant ALK protein, retaining the highest-scoring conformation for each ligand. For each model, the five conformations obtained after MD simulation trajectory clustering were re-docking using both SP and XP modes to screen for compounds most likely to form beneficial interactions with the double mutant. Based on the analysis of docking performance and screening efficiency, the 4CLJ-TPX-0131, 4CLJ-TPX0005, and 4CLJ-Lorlatinib models showed better docking performance in SP mode; the scores of the 4CD0-8e and 2YFX-Crizotinib models showed no significant difference between SP and XP modes; and the SP mode was significantly faster than the XP mode, making it more suitable for screening large-scale compound libraries (see Table 1). Based on these results, the SP docking mode was chosen for virtual screening of the five complex models.

[0058] Table 1

[0059] Secondly, the molecular MD trajectories from different models were clustered, and models with satisfactory ligand structures were obtained by docking with a small molecule database based on the different conformations obtained, which was then used for further research. Since there are differences in molecular structures, molecular weights, and other indicators in the dataset, the initial structure of the protein is crucial to the accuracy of molecular docking. To maximize the structural diversity of the selected proteins, after performing MD simulations on the five constructed models for 100 ns, the k-means algorithm was used to cluster the MD trajectories into 5 classes; this step helps to enhance the diversity of the screened protoproteins. As shown in Table 2, satisfactory ligand structures were obtained from both the 4CLJ-TPX-0131 and 4CLJ-Lorlatinib models, and 2055 ligand structures with scores < -7.0 kcal / mol were selected for subsequent work.

[0060] Table 2

[0061] Then, the structures of 2055 small molecule compounds in the optimal screening model were filtered using the Lipinski five-rule (mol_MW<500, QPlogPo / w<5, donorHB≤5, accptHB≤10) and the Jorgensen three-rule (QPlogS>-5.7, QPPCaco>22, Metab<7). The results showed that only 1509 ligands in the 4CLJ-Lorlatinib model met the criteria.

[0062] Subsequently, ligands with high docking scores were clustered based on their structures. Maintaining diversity in the candidate molecule set is crucial for covering a broader chemical space and avoiding homogenization of screening results in virtual screening. Using the Similarity and Clustering module of Canvas software, a similarity matrix was constructed based on molecular structure and physical properties. Hierarchical clustering analysis was then performed on the top 500 compounds with the highest docking scores out of 1509 ligands. Molecules with similar structures were grouped together, with a cluster count set of 50, to obtain a candidate molecule set exhibiting structural diversity.

[0063] Finally, from the clustered candidate molecules, 38 commercially available compounds were selected for ADMET (Absorption, Distribution, Metabolism, Excretion, and Toxicity) prediction. The QikProp module was used to calculate ADMET-related parameters, and the results were compared with reference values ​​in the software user manual. Compounds with excellent drug-like characteristics were prioritized. The results showed that all 38 compounds possessed good ADMET properties, meeting the requirements for subsequent experimental validation.

[0064] (6) Wet experiment verification: The 38 compounds obtained from screening (i.e., drug lead compounds with ALK G1202R / L1196M double mutation inhibitory activity) were subjected to in vitro bioactivity verification; (7) Molecular dynamics simulations were used to evaluate the binding stability of the active compound to the double mutant protein; (8) Energy calculations were used to verify the binding affinity and stability of the active compound.

[0065] Example 2: Wet experiment verification: Thirty-eight small molecule compounds from Example 1 were selected for in vitro antiproliferative activity evaluation. Ba / F3 cells were cultured in RPMI Medium 1640 (product number U21-279b, YOBIBIO) containing 10% fetal bovine serum (F8318, Sigma-Aldrich) and 10 ng / ml interleukin-3 (90143ES10, Yeasen); while NIH-3T3 cells and 293T cells were grown in DMEM medium (U21-265B, YOBIBIO) supplemented with 10% fetal bovine serum. Furthermore, all culture media used for cell growth were supplemented with 1% penicillin-streptomycin-glutamine (10378016, Gibco), and cells were cultured in flasks at 37°C and 5% CO2.

[0066] 1×10 4 Ba / F3 cells were seeded in 96-well plates, RPMI-1640 medium was added, and the cells were treated with gradient concentrations of the test compound for 48 hours. Then, 10 μL of 5 mg / mL MTT solution was added to each well, and incubation continued for 4 hours. Next, 100 μL of a triplet solution (10% SDS-0.1% HCl-PBS) was added to dissolve the formazan deposited at the bottom of the wells. Finally, the plates were incubated overnight. The absorbance at 570 nm was measured using a Synergy H1 microplate reader (BioTek) with 650 nm as the reference wavelength.

[0067] Figure 3 The study indicated that at a concentration of 10 μM, nine compounds exhibited inhibition rates exceeding 50% against EML4-ALK-WT, while six compounds achieved similar levels of inhibition against EML4-ALK-L1196M / G1202R. Notably, compound AMI007 achieved 100% inhibition in both cell lines. At concentrations reduced to 2 μM, the results were as follows... Figure 4 As shown, AMI007 inhibited cell proliferation by approximately 50% in both WT and mutant cell lines. Further studies on AMI007 revealed that its half-maximal inhibitory concentration (IC50) against Ba / F3-EML4-ALK-L1196M / G1202R was... 50 The effective concentration (1.83 μM) of AMI007 was superior to that of Ba / F3-EML4-ALK-WT (2.45 μM). The potent activity of AMI007 confirms the practicality and effectiveness of the targeted mutant screening model in real-world applications.

[0068] Example 3: Kinematic simulations were used to evaluate the binding stability of the active compound. Throughout the virtual screening process, all kinetic simulations employed a three-stage minimization approach to eliminate steric hindrance in the complex, allowing the system to better adapt to the simulation environment. The first stage fixed heavy atoms on all residue side chains in the system and applied an 8 kcal / mol·Å pressure. 2 The first stage constrains only hydrogen atoms, allowing them to move freely; the second stage restricts the protein backbone and ligand molecules in the system; the third stage minimizes all system atoms without constraints. System energy minimization includes 30,000 steps of steepest descent optimization and 70,000 steps of conjugate gradient optimization. Temperature is gradually increased: the system is first heated from 0 K to 100 K in 100 ps, ​​and then further heated to 310 K in another 100 ps. At pressure equilibrium, the constraint on all non-hydrogen atoms is set to 2.0 kcal / mol·Å. 2 Equilibrium was reached at 100 ps. An application of 8 kcal / mol·Å was applied to the main chain atoms of the protein. 2 Constraints were applied to maintain the system's temperature (T) and pressure (P) at 310 K and 1 atm for 100 ps; subsequently, 100 ps of equilibrium simulations were performed without any constraints. Finally, 100 ns simulations were conducted in both unconstrained and NPT systems with a time step of 2.0 fs. All covalent bonds were constrained by the SHAKE algorithm, the Langevin method was used to control the temperature, PME techniques were used to handle long-range electrostatic interactions, and the cutoff radius for non-bonded interactions was 8 Å.

[0069] Figure 5 The results show that the mutation of Gly1202 to Arg1202 leads to increased steric hindrance; residues Leu1122, Hie1124, and Gly1123 also extend outward, increasing the binding pocket space and facilitating the binding of AMI007. AMI007 can form stable hydrogen bonds with residues Hie1124, Gly1123, and Leu1122 in the G-loop region of the target protein, and with residues Met1199 and Arg1202 in the hinge region. These hydrogen bonds enhance the interaction between AMI007 and the mutant binding pocket, improving the stability of AMI007 and contributing to a prolonged duration of its pharmacological action.

[0070] Example 4: The binding energy differences of AMI007 with different systems were calculated to assess the binding stability. Table 3 shows the binding free energy results calculated from 400 snapshots extracted from the trajectory in the last 40 ns of the MD simulation. The predicted binding free energies of AMI007 in WT and G1202R / L1196M are -24.453 kcal / mol and -31.238 kcal / mol, respectively. In the wild-type system, the experimental IC50 of AMI007 is... 50The IC of the ALK G1202R / L1196M system is 2.45μM. 50 The concentration was 1.83 μM, and the predicted binding free energy was 6.785 kcal / mol higher than that of the WT system. Free energy calculations also demonstrated that AMI007 has a stronger binding affinity to the mutant ALK kinase domain.

[0071] Table 3

[0072] in, a ΔG pol =ΔE ele +ΔG gb、 b ΔG hydro =ΔE vdW +ΔG surf .

[0073] In summary, the virtual screening method for inhibitors targeting the ALK G1202R / L1196M double mutation constructed in this application, through the integration of multiple screening technologies and the implementation of a multi-level screening strategy, has for the first time discovered that the active compound AMI007 has highly efficient targeting of this double mutation. AMI007 exhibits a highly stable binding mode and excellent binding affinity in the ALK active pocket, providing key support for solving the drug resistance problem of NSCLC caused by this double mutation and opening up new pathways for the development of targeted drugs.

[0074] In the description of this specification, references to terms such as "one possible implementation," "further," "exemplary," "specific example," or "optional," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. The illustrative expressions of the above terms in this specification do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. In addition, those skilled in the art can combine and integrate the different embodiments or examples described in this specification.

[0075] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the invention filed herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not claimed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of this application are indicated by the appended claims.

Claims

1. A virtual screening method for drug lead compounds with ALK G1202R / L1196M double mutation inhibitory activity, characterized in that, include: The resolved structures of the ALK L1196M single mutant protein-small molecule complex were obtained from the RCSB PDB database. The structures of the ALK L1196M single mutant protein-small molecule complex include PDBID: 4CLJ, resolution: 1.66 Å; PDBID: 2YFX, resolution: 1.7 Å; and PDBID: 4CD0, resolution: 2.23 Å. Screening models were constructed based on the structure of the ALK L1196M single mutant protein-small molecule complex, and the crystal structures of the protein-inhibitor complexes in each model were standardized and preprocessed. The screening models included: 2YFX-Crizotinib, 4CLJ-Lorlatinib, 4CLJ-TPX0005, 4CLJ-TPX0131 and 4CD0-8e. Structural optimization and conformation preparation were performed on the library of compounds to be screened. The docking method was determined, and the optimal screening model was selected from the standardized preprocessed screening models. After molecular dynamics trajectory clustering, the optimal screening model was docked with the library of compounds to be screened. Drug lead compounds with ALK G1202R / L1196M double mutation inhibitory activity were screened by filtering for drug-like properties, clustering of ligand structures and prediction of ADMET properties.

2. The virtual screening method for drug lead compounds with ALK G1202R / L1196M double mutation inhibitory activity according to claim 1, characterized in that, In the step of obtaining the resolved ALK L1196M single mutant protein-small molecule complex structure from the RCSB PDB database, the Mutate Residue module of the Schrodinger software is used to mutate residue G1202 in the ALKL1196M single mutant protein structure to R1202. Molecular dynamics simulation is performed to optimize the structure of the mutated double mutant protein and adjust the spatial conformation of the newly added side chain to a stable state.

3. The virtual screening method for drug lead compounds with ALK G1202R / L1196M double mutation inhibitory activity according to claim 1, characterized in that, The standardized preprocessing involves using the Protein Preparation Wizard module to remove water of crystallization molecules and impurity ligands, repair the loop region conformation, adjust the bond order, determine the protonation state of residues based on the hydrogen bonding mode of the original crystal structure, and optimize the energy of all heavy atoms under the OPLS3 force field to a root mean square deviation ≤ 0.3 Å.

4. The virtual screening method for drug lead compounds with ALK G1202R / L1196M double mutation inhibitory activity according to claim 1, characterized in that, In the steps of structure optimization and conformation preparation of the compound library to be screened, SDF format molecules from the ChemDiv database are selected as the screening database. The structure is preprocessed using the LigPrep module to remove duplicate molecules and unreasonable chemical bonds, correct tautomers and chiral centers, and retain the 10 lowest energy dominant conformations during conformation optimization of each small molecule.

5. The virtual screening method for drug lead compounds with ALK G1202R / L1196M double mutation inhibitory activity according to claim 1, characterized in that, The steps of determining the docking method, selecting the optimal screening model from the standardized preprocessed screening models, docking the optimal screening model with the library of compounds to be screened after performing molecular dynamics trajectory clustering, and screening for drug lead compounds with ALK G1202R / L1196M double mutation inhibitory activity through drug-like property filtering, ligand structure clustering, and ADMET property prediction include: The standard and precision modes of the Glide module were used to evaluate re-docking, and the optimal docking method was determined by using the rationality of ligand binding conformation and docking score as indicators. For each screening model, a 100ns dynamic simulation was performed. The k-means algorithm was used to cluster the molecular dynamic trajectories into 5 categories. Based on the clustered conformations, the optimal screening model was selected by docking with the library of compounds to be screened. Dual filtering is performed based on the Lipinski five-rule and the Jorgensen three-rule; the Lipinski five-rule includes mol_MW<500, QPlogPo / w<5, donorHB≤5, accptHB≤10, and the Jorgensen three-rule includes QPlogS>-5.7, QPPCaco>22, and Metab<7. The three-dimensional coordinates of the top 500 docking scores of the ligands were imported into Canvas software, the cluster count was set to 50, and structural clustering was performed based on the Tanimoto coefficient to ensure the diversity of the drug lead compounds with ALK G1202R / L1196M double mutation inhibitory activity; wherein, the Tanimoto coefficient is ≥0.

8. The QikProp module was used to calculate the absorption, distribution, metabolism, excretion, and toxicity parameters of the compounds, and to screen for drug lead compounds with ALKG1202R / L1196M double mutation inhibitory activity.

6. The virtual screening method for drug lead compounds with ALK G1202R / L1196M double mutation inhibitory activity according to claim 1, characterized in that, Also includes: The in vitro bioactivity of the drug lead compound with ALK G1202R / L1196M double mutation inhibitory activity was verified by wet experiments. The binding stability of the drug lead compound with ALK G1202R / L1196M double mutation inhibitory activity to the double mutant protein was evaluated by molecular dynamics simulations. The binding affinity and stability of the drug lead compound with ALK G1202R / L1196M double mutation inhibitory activity were verified by energy calculations. Thus, the inhibitory effect and binding process of the drug lead compound with ALK G1202R / L1196M double mutation inhibitory activity on the double mutant protein were determined.

7. The virtual screening method for drug lead compounds with ALK G1202R / L1196M double mutation inhibitory activity according to claim 6, characterized in that, The wet assay was performed using a kinase activity assay kit to determine the inhibitory activity of the drug lead compound with ALKG1202R / L1196M double mutation inhibitory activity on the ALKG1202R / L1196M double mutant protein. The system was constructed through a three-stage minimization method using molecular dynamics simulations. The system energy was optimized by steepest descent and conjugate gradient, and the system temperature was gradually heated from 0 K to 310 K within 200 ps. The steepest descent took 30,000 steps, and the conjugate gradient took 70,000 steps. The energy calculation was performed using the MM / GBSA method to calculate the binding free energy between the drug lead compound with ALK G1202R / L1196M double mutation inhibitory activity and the double mutant protein.

8. A drug lead compound possessing ALK G1202R / L1196M double mutation inhibitory activity, characterized in that, The drug lead compounds with ALK G1202R / L1196M double mutation inhibitory activity as described in any one of claims 1-7 were obtained through a virtual screening method.

9. The drug lead compound with ALK G1202R / L1196M double mutation inhibitory activity according to claim 8, characterized in that, The chemical structure of the drug lead compound with ALK G1202R / L1196M double mutation inhibitory activity is shown in Formula (I): 。 10. The use of the drug lead compound with ALK G1202R / L1196M double mutation inhibitory activity according to claim 8 in the preparation of an anti-non-small cell lung cancer drug, wherein the non-small cell lung cancer includes squamous cell carcinoma, adenocarcinoma and large cell carcinoma.

Images