Discovery of pyridazine-3 (2H)-ketone derivative and application of pyridazine-3 (2H)-ketone derivative as HDAC6 inhibitor

Abstract

The invention belongs to the technical field of biological medicines, and particularly relates to discovery of pyridazine-3 (2H)-ketone derivatives and application of the pyridazine-3 (2H)-ketone derivatives as HDAC6 inhibitors. By means of artificial intelligence and a virtual screening technology, a small molecule compound with an HDAC6 selective targeting inhibition effect is obtained; the small-molecule inhibitor has strong HDAC6 enzyme inhibition activity and can be used for preparing an HDAC6 inhibitor; the compound has the potential of treating leukemia, colorectal cancer, melanoma, non-small cell lung cancer, breast cancer and other cancers, the inhibition effect of the compound is superior to that of a tool drug Tubastatin A (TubA), and the compound has very wide application prospects.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence and biomedicine technology, and specifically to the discovery of pyridazine-3(2H)-one derivatives and their application as HDAC6 inhibitors. Technical Background

[0002] Cancer, as one of the diseases that pose the greatest threat to human health and life, has attracted extensive attention and research. Studies have pointed out that abnormal epigenetic modifications are closely related to the occurrence and development of cancer. Among them, histone acetylation catalyzed by histone acetyltransferases (HATs) and deacetylation catalyzed by histone deacetylases (HDACs) are dynamic and reversible processes that directly affect the epigenetic regulatory mechanism, change the morphology of chromatin and interfere with gene expression. Abnormal expression and changes in biological functions of HDACs will lead to changes in the expression of related genes in cells, thereby affecting cell proliferation, differentiation and apoptosis, and then inducing diseases such as melanoma, colorectal cancer and breast cancer. So far, 18 subtypes of HDACs have been reported, which are divided into 4 categories: Class I (HDAC1, 2, 3, 8), Class II (IIa: HDAC4, 5, 7, 9; IIb: 6, 10), Class III (SIRT 1-7), and Class IV (HDAC11). Among them, Class Ⅰ, Ⅱ and Ⅳ belong to Zn 2+ Class III is NAD + Dependent enzyme.

[0003] Among the 18 subtypes of enzymes mentioned above, HDAC6 has attracted widespread interest among researchers due to its unique structure, distribution, and biological function. HDAC6 consists of two catalytic domains (CD1 and CD2) and a ubiquitination site, and is mainly distributed in the cytoplasm. HDAC6 can target many non-histone substrates such as α-tubulin, cortactin, and HSP90, thereby participating in a variety of biological processes such as cell migration, endocytosis, autophagy, apoptosis, and protein degradation. Abnormal expression of HDAC6 is closely related to processes such as tumorigenesis, invasion, and metastasis. Studies have shown that, unlike the phenotypic distortion and embryonic lethality caused by knockout of Class I (HDAC1-3), mice knocked out of the HDAC6 gene develop normally. The results show that drugs that selectively target HDAC6 have good safety and great application prospects and research value. Currently, a total of four selective HDAC6 inhibitors (ACY-1215, ACY-241, KA2507 and CS3003) have entered the clinical trial research stage for the development of anti-tumor drugs, which has further stimulated researchers' interest in designing small molecule drugs targeting HDAC6.

[0004] The pharmacophore model of selective HDAC6 inhibitors is mainly composed of three parts: Cap, Linker and ZBG. Among them, Cap can interact with the surface recognition area of ​​HDAC6 binding pocket; ZBG can chelate Zn at the bottom of the pocket. 2+ ; Linker connects Cap and ZBG and binds them in the hydrophobic cavity. At present, the design of selective HDAC6 inhibitors mainly focuses on modifying Cap and Linker. However, while only improving the activity and selectivity of the inhibitor by enlarging the Cap fragment and replacing the Linker with an aromatic fragment, it also greatly increases the complexity of the compound synthesis process, resulting in reduced ligand efficacy. Therefore, it is still of great significance to find HDAC6 inhibitors with novel structures.

[0005] In recent years, virtual screening (VS) technology based on molecular docking and molecular dynamics simulation has become an important strategy for the discovery of novel small molecule drugs. This technology can fully consider the interaction between ligands and residues in local active pockets and the binding mode. As an important branch of virtual screening technology, the compound-protein interaction (CPI) prediction model based on artificial intelligence technology does not rely on crystal information. It can evaluate the interaction mode between proteins and small molecules through the learned sequence knowledge, realize multi-target simultaneous screening, and provide a new screening perspective with strong versatility and higher efficiency for the discovery of selective molecules.

[0006] In the present invention, we disclose a novel drug screening technology, which combines the CPI prediction model based on voting strategy, cascade molecular docking and molecular simulation to establish an efficient and accurate screening pipeline. Through the evaluation of HDAC enzyme subtype inhibitory activity and anti-tumor cell activity, 2-(2-phenoxyethyl)pyridazine-3(2H)-one derivatives with selective HDAC6 inhibitory activity and anti-tumor cell activity were obtained. Summary of the invention

[0007] In response to the above technical problems, the present invention uses artificial intelligence and virtual screening technology to obtain a small molecule compound with selective targeted inhibition of HDAC6; the small molecule inhibitor has potent HDAC6 enzyme inhibition activity and can be used to prepare HDAC6 inhibitors; and has the potential to treat leukemia, colorectal cancer, melanoma, non-small cell lung cancer, breast cancer and other cancers.

[0008] Specifically include the following:

[0009] In a first aspect, the present invention provides a method for screening an HDAC6 inhibitor, the method comprising the following steps:

[0010] (1) Preparation of molecular library: Integrate the commercial molecular library, and then use Lipinski rule, Veber rule and PAINS rule to screen and remove compounds that do not meet drug-like properties and false positives;

[0011] (2) Rapid virtual screening based on voting strategy: Voting strategies were established based on two CPI models, the HDACban model and the TransformerCPI2.0 model, to complete the rapid screening of the molecular library;

[0012] (3) Structure-based virtual screening: Prepare receptor crystals and ligands, generate docking grids for receptor molecules based on the spatial coordinates of the ligands in the original crystals, use Glide-HTVS, Glide-SP / XP, and Prime-MMGBSA to screen small molecules based on the screening accuracy, and refer to the binding affinity of TubA in HDAC6 to select molecules that meet the requirements for subsequent molecular dynamics simulations;

[0013] (4) Target affinity evaluation: The GPU-accelerated PMEMD program in AMBER20 software was used to perform molecular dynamics simulation studies on the established research system.

[0014] Preferably, the step (2) comprises:

[0015] Construction of HDAC fine-tuning dataset: The relevant information of compounds in ChEMBL34 database was collected and processed as follows: (i) The biological activity indicators were set as ‘Ki’, ‘Kd’, ‘IC 50 ','EC 50 ', and the unit is uniformly converted to 'nM'; (ii) biological activity threshold (including pKi, pKd, pIC 50 and pEC 50 ) is set to 6.5. When the biological activity threshold is greater than 6.5, it is considered that the compound interacts with the target and is assigned a value of 1; when it is less than 6.5, it indicates that there is no interaction and is assigned a value of 0; the relevant information includes protein type, molecule type and activity;

[0016] Model building: The HDAC-specific CPI screening models HDACban and TransformerCPI2.0 were obtained by retraining and fine-tuning using large-scale CPI data and the constructed HDAC fine-tuning dataset.

[0017] Preferably, the step (3) comprises:

[0018] Receptor preparation: Download the required receptor crystal PDB ID: 5EDU from the Protein Data Bank (http: / / www.rcsb.org), import it into Maestro, and optimize it through the Protein Preparation Wizard, including the sorting of distribution bonds, creation of disulfide bonds, dehydration, hydrogenation, protonation state setting and energy optimization. When the RMSD between heavy atoms reaches When , the energy minimization is terminated;

[0019] Ligand preparation: The LigPrep module in Maestro was used to prepare the molecular library under OPLS-2005, including setting the ionization state, generating tautomers and stereoisomers to ensure the diversity of the ligand at pH = 7.0 ± 2.0;

[0020] Cascade molecular docking: The docking grid of the receptor molecule was generated according to the spatial coordinates of the ligand in the original crystal. Glide-HTVS, Glide-SP / XP and Prime-MMGBSA were used in turn for small molecule screening according to the screening accuracy. The binding affinity of TubA in HDAC6 was referred to and the molecules that met the requirements were selected for subsequent molecular dynamics simulation.

[0021] Preferably, the step (4) comprises:

[0022] Simulation system preparation: Use Gaussian09 and AMBER20 research systems to prepare, generate coordinate files of complexes, receptors, ligands, and topology files and coordinate files of solvated complexes in the SPCE water box;

[0023] Molecular dynamics simulation: The simulation system was optimized using the steepest descent method and the conjugate gradient method to eliminate high-energy collisions between molecules, and then heated up, and the PMEMD program was used to perform a 1μs simulation under the NPT ensemble;

[0024] Post-processing analysis based on simulation trajectory: Evaluate the equilibrium stability of the simulation system and extract representative conformations, calculate binding free energy and residue energy decomposition based on representative conformations;

[0025] ΔG MM / GBSA =ΔE vdw +ΔE ele +ΔG pol +ΔG nonpol ;

[0026] Where, ΔE vdw and ΔE ele Refers to van der Waals force and electrostatic force respectively; ΔG pol and ΔG nonpol Refers to polar interaction and nonpolar interaction respectively.

[0027] In a second aspect, the present invention provides the use of the screening method described in the first aspect in screening HDAC6 inhibitors or anti-tumor drugs.

[0028] In a third aspect, the present invention provides a compound as shown in the following formula (I), or a pharmaceutically acceptable salt, hydrate, deuterated substance, or prodrug thereof;

[0029]

[0030] Among them, A 1 , A 2 , A 3 are selected from CR' or N, R' is a substituent of C;

[0031] R' and R 1 Selected from H, C 1 -C 6 Alkoxy, C 1 -C 6 Alkylamino, C 1 -C 6 Alkoxycarbonyl, substituted or unsubstituted C 1 -C 6 Acylamino, substituted or unsubstituted C 1 -C 10 Alkyl, substituted or unsubstituted C 6 -C 12 aryl or heteroaryl, substituted or unsubstituted 3-12 membered alicyclic or heterocyclic group, substituted or unsubstituted benzyl;

[0032] X is selected from (CH 2 ) n , n is any integer greater than or equal to 0;

[0033] Y is selected from C, N or O;

[0034] Z is selected from substituted or unsubstituted C 1 -C 10 Alkyl, C 1 -C 6 Alkoxy, C 1 -C 6 Alkylamino, C 1 -C 6 Alkoxycarbonyl, substituted or unsubstituted C 6 -C 12 Aryl or heteroaryl, substituted or unsubstituted 3-12 membered alicyclic or heterocyclic group;

[0035] W is selected from (CH 2 ) n , substituted or unsubstituted C 6 -C14 Aryl, C 5 -C 14 Heteroaryl, C 7 -C 12 Aralkyl, C 6 -C 12 Arylheteroalkyl; n is any integer greater than or equal to 0;

[0036] The substitution refers to that one or more hydrogen atoms on the group are replaced by a substituent selected from any of the following:

[0037] Halogen atoms, carbonyl, carboxyl, hydroxyl, amino, nitro, cyano, C 1 -C 10 Alkyl, C 1 -C 6 Alkoxy, C 1 -C 6 Alkylamino, C 1 -C 6 Alkoxycarbonyl, C 1 -C 6 Acylamino, C 6 -C 10 Aryl, five-membered heteroaryl, six-membered heteroaryl;

[0038] or C substituted by the above substituents 1 -C 10 Alkyl, C 6 -C 10 Aryl, five-membered heteroaryl, six-membered heteroaryl; preferably C 1 -C 6 Alkyl, C 1 -C 6 Alkoxy or C 1 -C 6 Alkylamino.

[0039] Preferably, the substituted C 6 -C 10 Aryl is substituted phenyl.

[0040] Preferably, the substituted phenyl group is selected from the group consisting of an alkyl substituted phenyl group, an alkoxy substituted phenyl group, an alkylthio substituted phenyl group, a halogen substituted phenyl group, and a haloalkyl substituted phenyl group.

[0041] Preferably, the substitution is selected from halogen atoms, C 1 -C 6 Alkoxycarbonyl, C 1 -C 6 Alkyl, halogenated C 1 -C 6 Alkyl, C 1 -C 6 Alkoxy or phenyl.

[0042] Preferably, the structural formula of the compound is as shown in formula (II):

[0043]

[0044] In a fourth aspect, the present invention provides use of the compound described in the third aspect, or a pharmaceutically acceptable salt, hydrate, deuterated product, or prodrug thereof in the preparation of an HDAC6 inhibitor.

[0045] In a fifth aspect, the present invention provides use of the compound described in the third aspect, or a pharmaceutically acceptable salt, hydrate, deuterated substance, or prodrug thereof in the following pharmaceutical applications:

[0046] (1) Application in the preparation of anti-tumor drugs;

[0047] (2) Application in the preparation of drugs for treating autoimmune diseases;

[0048] (3) Application in the preparation of anti-inflammatory drugs;

[0049] (4) Application in the preparation of drugs for treating Alzheimer's disease.

[0050] Preferably, the tumor is selected from leukemia, colorectal cancer, melanoma, lung cancer, breast cancer, and cervical cancer.

[0051] Preferably, the tumor is colorectal cancer.

[0052] In a sixth aspect, the present invention provides any pharmaceutically acceptable dosage form prepared by adding a pharmaceutically acceptable salt and an auxiliary material to the compound described in the third aspect.

[0053] Preferably, the dosage forms include tablets, injections, granules, and suspensions.

[0054] The beneficial effects of the present invention are as follows: first, the present invention provides a technical solution for screening small molecule inhibitors targeting HDAC6 by means of artificial intelligence and virtual screening technology; secondly, the present invention uses the screening solution to obtain a compound 2-(2-phenoxyethyl) pyridazine-3(2H)-one hydroxamic acid derivative (Cmpd.18) with HDAC6 inhibitory activity, and the Cmpd.18 has high selectivity and strong inhibitory activity against HDAC6. At the same time, Cmpd.18 can significantly inhibit the proliferation of HCT116 and HL60 cells and exhibit anti-tumor activity. Further studies have found that Cmpd.18 exerts the proliferation effect of HCT116 cells by promoting cell apoptosis and blocking the cell cycle; the results of pharmacodynamic experiments show that the compounds involved in the present invention can be used as potential therapeutic drugs for diseases such as tumors, autoimmune diseases, inflammation, and Alzheimer's disease. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 A screening pipeline based on artificial intelligence and molecular dynamics simulation technology;

[0056] Figure 2 Molecular dynamics simulation of the HDAC6-Cmpd.18 docking system; (A) shows the fluctuations of the protein, binding pocket and ligand in the HDAC6-Cmpd.18 research system during the dynamics simulation, and (B) shows the changes in the representative conformation of Cmpd.18;

[0057] Figure 3 Representative binding conformation analysis of Cmpd.18 in HDAC6 during the 900-1000 ns trajectory;

[0058] Figure 4 Analysis of the decomposition free energy of the 900-1000ns trajectory of the HDAC6-Cmpd.18 research system;

[0059] Figure 5 Effects of the compounds Cmpd.18 and TubA of the present invention on the acetylation levels of HDAC6-related substrates in human colorectal cancer cells HCT116; wherein, (A) is a Western blot analysis of the effects of the compounds Cmpd.18 and TubA on the acetylation levels of α-tubulin and histone H3; (B) is a statistical data analysis of the effects of Cmpd.18 and TubA on the acetylation levels of α-tubulin and histone H3, compared with the control group *P<0.05, **P<0.01, ***P<0.001;

[0060] Figure 6 Effect of the compound Cmpd.18 of the present invention on apoptosis of human colorectal cancer cell HCT116;

[0061] Figure 7 The compound Cmpd.18 described in the present invention affects the cell cycle of human colorectal cancer cell HCT116. DETAILED DESCRIPTION

[0062] The technical solution described in the present application is described below in conjunction with specific embodiments, but the present application is not limited to the following.

[0063] The experimental methods in the following examples are conventional methods unless otherwise specified; the experimental materials used in the following examples are purchased from conventional biochemical reagent companies unless otherwise specified.

[0064] Example 1 Screening method for HDAC6 inhibitors

[0065] This embodiment provides a technical solution for screening small molecule inhibitors targeting HDAC6 by using artificial intelligence and virtual screening technology, specifically including:

[0066] (1) Preparation of molecular libraries: Commercial molecular libraries such as LifeChemicals, ZINC, ChemDiv and Enamine were integrated, and then the Lipinski rule, Veber rule and PAINS rule were used to further screen and remove compounds that did not meet the drug-like properties and false positives.

[0067] (2) Rapid virtual screening based on voting strategy: The present invention uses two CPI models (HDACban model and TransformerCPI2.0 model) to establish a voting strategy (such as Figure 1 As shown) , rapid screening of the molecular library is completed.

[0068] Dataset construction: First, collect the relevant information of compounds in the ChEMBL34 database, including protein type (single protein); molecule type (small molecule) and activity (binding). The collected information is processed as follows: (i) biological activity indicators are set as 'Ki', 'Kd', 'IC 50 ','EC 50 ', and the unit is uniformly converted to 'nM'; (ii) biological activity threshold (including pKi, pKd, pIC 50 and pEC 50 ) was set to 6.5. When the bioactivity threshold was greater than 6.5, the compound was considered to have an interaction with the target (assigned a value of 1), and less than 6.5 indicated no interaction (assigned a value of 0).

[0069] All protein sequence information was obtained from UniProt. The training set and validation set for retraining DrugBAN were obtained by dividing CPI data other than HDAC CPI data in a 9:1 ratio. Because the training dataset of TransformerCPI2.0 is derived from ChEMBL23, the data collection time of the HDAC CPI independent test set is limited to after 2017, and the HDAC CPI data before 2017 are also divided in a 9:1 ratio to obtain the training set and validation set of the DrugBAN fine-tuning dataset. Finally, a total of 848,027 CPI data were obtained, including: 763,219 DrugBAN training sets and 84,808 validation sets; 7,929 fine-tuning training sets and 882 validation sets; 2,879 independent test sets of each subtype of HDAC. The data were classified and converted into standard input format for retraining and optimization of the CPI model.

[0070] Model establishment and training: The HDACban model is a DrugBAN model developed based on the bilinear attention network framework, and then retrained and fine-tuned using large-scale CPI data and the constructed HDAC fine-tuning dataset to obtain the HDAC-specific CPI screening model, which is implemented in Python 3.8 and PyTorch 1.7.1. The TransformerCPI2.0 model is a pre-trained model, and its effectiveness and generalization ability have been fully verified by experiments.

[0071] (3) Structure-based virtual screening:

[0072] Receptor preparation: Download the required receptor crystal (PDBID: 5EDU) from the Protein Data Bank (http: / / www.rcsb.org), import it into Maestro, and optimize it through the Protein Preparation Wizard, which mainly includes the sorting of distribution bonds, creation of disulfide bonds, dehydration, hydrogenation, protonation state setting and energy optimization. When the RMSD between heavy atoms reaches When , the energy minimization is terminated;

[0073] Ligand preparation: The LigPrep module in Maestro was used to prepare the molecular library under OPLS-2005, including setting the ionization state, generating tautomers and stereoisomers to ensure the diversity of the ligands at the specified pH value (pH = 7.0 ± 2.0);

[0074] Cascade molecular docking: The docking grid of the receptor molecule is generated according to the spatial coordinates of the ligand in the original crystal. Then, Glide-HTVS, Glide-SP / XP and Prime-MMGBSA are used in turn for small molecule screening according to the screening accuracy. With reference to the binding affinity of TubA in HDAC6, molecules that meet the requirements are selected for subsequent molecular dynamics simulation.

[0075] (4) Target affinity evaluation

[0076] The present invention uses the GPU-accelerated PMEMD program in AMBER20 software to perform molecular dynamics simulation research on the established research system.

[0077] Simulation system preparation: Use Gaussian09 and AMBER20 research systems to prepare, generate coordinate files of complexes, receptors, ligands, and topology files and coordinate files of solvated complexes in the SPCE water box;

[0078] Molecular dynamics simulation: The simulation system was optimized using the steepest descent method and the conjugate gradient method to eliminate high-energy collisions between molecules, and then heated up, and the PMEMD program was used to perform a 1μs simulation under the NPT ensemble;

[0079] Post-processing analysis based on simulation trajectory: equilibrium stability evaluation and representative conformation extraction of the simulation system, binding free energy calculation and residue energy decomposition calculation based on the representative conformations.

[0080] ΔG MM / GBSA =ΔE vdw +ΔE ele +ΔG pol +ΔG nonpol

[0081] where ΔE vdw and ΔE ele Refers to van der Waals force and electrostatic force respectively; ΔG pol and ΔG nonpol Refers to polar interaction and nonpolar interaction respectively.

[0082] After drugability and PAINS rule screening, about 3.5 million molecules were obtained, in order to further improve the prediction accuracy of the CPI prediction model. The present invention first conducts a comprehensive scientific evaluation of the performance of the model on each target HDAC subtype target. On all selected subtypes, the AUPRC and PRE of TransformerCPI2.0 and HDACban are mostly more than 0.7 (as shown in Tables 1 and 2), proving the superior predictive ability of the two models in selecting HDAC subtypes. The PRE performance of the voting strategy using HDACban and TransformerCPI2.0 on the subtype test set was also evaluated. The experimental results showed that not only did the PRE performance on all subtype test sets surpass the single model, all exceeding 0.8, but also on the key HDAC6 subtype test set, PRE reached a high level of 0.832. In summary, the voting strategy combining HDACban and TransformerCPI2.0 models shows significant advantages in predicting and screening selective active molecules for HDAC6 subtypes, which can provide strong support for the subsequent development of HDAC6 selective inhibitors. Finally, approximately 60,000 predicted active molecules of HDAC6 (labeled as 1) and other HDAC enzyme isoforms (labeled as 0) were retained for subsequent evaluation through a voting strategy.

[0083] Table 1 PRE performance of DrugBAN, HDACban, TransformerCPI2.0 and voting strategy

[0084] HDAC1 HDAC2 HDAC6 HDAC8 HDAC11 DrugBAN 0.654 0.728 0.771 0.535 0.714 HDACban 0.778 0.753 0.824 0.678 0.727 TransformerCPI2.0 0.800 0.780 0.740 0.711 0.714 Voting strategy 0.845 0.830 0.834 0.800 0.800

[0085] Table 2 AUPRC performance of DrugBAN, HDACban and TransformerCPI2.0 on the selected HDAC isoform test set

[0086] HDAC1 HDAC2 HDAC6 HDAC8 HDAC11 DrugBAN 0.663 0.747 0.836 0.616 0.735 HDACban 0.794 0.807 0.898 0.697 0.757 TransformerCPI2.0 0.839 0.733 0.832 0.590 0.722

[0087] Example 2 HDAC enzyme inhibitory activity determination of compounds

[0088] After the receptor-based screening strategy, the top 5000 molecules with higher docking scores and binding free energies than TubA were retained and further clustered into 50 categories using the Butina algorithm. The top two molecules were subjected to 50ns molecular dynamics simulations, and TubA was used as a reference (MM-GBSA = -45.07 kcal / mol). 53 molecules were retained for preliminary HDAC6 enzymatic activity determination. The specific method is as follows:

[0089] Recombinant HDAC1, HDAC2, HDAC6 and HDAC8 enzyme kits were purchased from BPS Bioscience. First, the target compound and positive control drug (TubA) were dissolved in DMSO to prepare a series of dilutions. HDAC1, HDAC2, HDAC6 and HDAC8 human recombinant proteins were diluted with HDAC buffer to the specified concentrations of 1.4 ng / μL, 1 ng / μL, 7 ng / μL, and 4 ng / μL, respectively. Then, 5 μL of fluorescent HDAC substrate 3, 5 μL of BSA (1 mg / mL), and 30 μL of HDAC buffer were added to the 96-well plate in sequence, and then 5 μL of solutions containing different concentrations of target compounds or TubA were added to each well, and finally 5 μL of the corresponding enzyme solution was added to each well to start the reaction. The reaction system was incubated at 37°C for 30 minutes, and then 50 μL of HDAC developer (2×) containing Trichostatin A was added to each well and incubated at room temperature for 15 minutes. Finally, the fluorescence intensity was detected using a Varioskan LUX microplate reader at an excitation wavelength of 360 nm and an emission wavelength of 460 nm. Three independent experiments were performed in each group, and the enzyme inhibition activity of the target compound was calculated using GraphPadPrism 8 software.

[0090] The experiment found that there were significant differences in the inhibitory activities of these compounds (ranging from 4.59% to 95.86%). Among them, Cmpd.1, Cmpd.4, Cmpd.8, Cmpd.9, Cmpd.11, Cmpd.12, Cmpd.18 and Cmpd.25 had an inhibition rate of more than 50% on HDAC6 at 500nM, which deserves further study (as shown in Table 3). It is worth noting that the inhibition rates of Cmpd.12 (inhibition rate of 94.65%) and Cmpd.18 (inhibition rate of 95.86%) were comparable to those of TubA (94.82%). In addition, we used the Tanimoto algorithm of RDKit to evaluate the novelty and diversity of the screened molecular skeletons, indicating that the screened compounds were structurally specific (Tanimoto coefficient <50%). Molecular docking revealed that these eight molecules can occupy the active pocket of HDAC6 and interact with key residues in the active pocket. Hydroxylamine, amide, urea, aminopropanol, ester and hydroxyamide as ZBG fragments can interact with Zn at the bottom of the hydrophobic pocket. 2+ Chelation.

[0091] Based on the HDAC6-Cmpd.18 docking system, a long-term (1 μs) molecular dynamics simulation was performed. Figure 2 As shown, the amino acids on the receptor The amino acids that make up the active pocket (average ) and ligand The fluctuations were small throughout the simulation, and dynamic equilibrium was reached quickly. It is worth noting that the representative conformation had a certain degree of deflection compared with the initial conformation, especially in the Cap region. As the simulation time increased, the proportion of representative conformations also increased, especially in the last 200ns of the dynamic trajectory, where the proportion of representative conformations reached 55.8%, indicating that Cmpd.18 has a stable binding mode in HDAC6.

[0092] The representative binding conformations of the 900-1000 ns trajectory show that the fragment with alkoxyphenyl substituted pyridazin-3(2H)-one as Cap can form hydrogen bonds with H499 and form hydrophobic interactions with P748, P501 and L749. Linker and ZBG can penetrate into the hydrophobic cavity, where the phenyl on the linker forms a π-π stacking effect with H651, F680 and F620, and ZBG forms hydrogen bonds with G619, H611, C621 and Y782 (e.g. Figure 3 The carbonyl oxygen in the hydroxamic acid can be chelated with the zinc ion in a single reaction, and the chelation distance is stable at 1. To explore the receptor-ligand interaction in more detail, we performed amino acid residue decomposition free energy analysis and identified 24 residues that contributed to Cmpd.18 binding (|energy contribution| ≥ 0.1 kcal / mol), of which 11 key amino acids had significant energy contributions (|energy contribution| ≥ 0.5 kcal / mol) ( Figure 4 ). In summary, Cmpd.18 is able to interact with key residues and bind to the active pocket in an effective conformation, which is also an important prerequisite for its enzyme inhibitory activity.

[0093] The evaluation results of the enzymatic and anti-tumor cell proliferation activities of the compounds of the present invention are shown in Table 3.

[0094] Table 3 Computational simulation results and preliminary enzyme activity evaluation of the screened compounds

[0095] Cmpd. Docking Score XP GScore Prime MM-GBSA MM-GBSA Inhibition rate(%) Cmpd.1 –10.54 –10.54 –57.93 –53.94 71.85±3.78 Cmpd.4 –9.48 –9.48 –58.93 –48.65 58.47±2.58 Cmpd.8 –9.89 –9.89 –59.60 –52.29 75.24±3.13 Cmpd.9 –9.61 –9.61 –56.82 –49.02 54.92±1.89 Cmpd.11 –9.41 –9.41 –60.33 –50.11 70.45±5.74 Cmpd.12 –9.52 –9.53 –58.56 –49.43 94.65±2.35 Cmpd.18 –10.66 –10.67 –63.17 –52.90 95.86±1.31 Cmpd.25 –9.49 –9.55 –56.10 –47.51 51.89±2.52 TubA –9.35 –9.35 –54.05 –45.07 94.82±1.58

[0096] Example 3 Determination of the anti-tumor activity of the compounds of the present invention

[0097] 1. Tumor Cell Proliferation Assay

[0098] HL60, HCT116, A375, A549, MCF-7, MDA-MB-231 and HeLa cell lines were used to evaluate the inhibitory effects of the compounds on tumor cell proliferation. All cell lines were cultured in complete medium (containing 10% fetal bovine serum and 1% penicillin / streptomycin) at 37°C and 5% CO 2 The cells in the logarithmic growth phase were digested, counted, and inoculated in a 96-well plate (1×10 4 Each well was incubated for 24 h and treated with different concentrations of the compound and the positive control drug TubA for 48 h. CCK-8 reagent was then added to each well and incubated at 37 °C for 1-3 h. The absorbance was measured at 450 nm using a Varioskan LUX microplate reader. Three independent experiments were performed and IC was calculated using GraphPad Prism 8. 50 value.

[0099] As can be seen from Table 4, the compounds screened by the present invention have general antiproliferative activity against HL60 and HCT116 (IC 50 <50μM), among which Cmpd.1 (IC 50 =10.16μM), Cmpd.8 (IC 50 =8.52μM), Cmpd.12 (IC 50 =2.78 μM) and Cmpd.18 (IC 50=2.59 μM) had better antiproliferative activity against HCT116 than the positive control (TubA:IC 50 =11.97 μM).

[0100] Table 4 Inhibitory activity of compounds against tumor cell proliferation (IC 50 , μM)

[0101] Cmpd. HL60 HCT116 HeLa A549 MCF-7 MDA-MB-231 A375 Cmpd.1 16.54±1.33 10.16±1.01 38.92±3.12 21.89±1.67 43.87±4.12 32.76±2.62 32.57±1.55 Cmpd.4 22.95±1.54 31.34±1.69 >50 >50 >50 >50 29.41±2.46 Cmpd.8 9.42±0.56 8.52±0.67 23.45±1.43 37.97±2.58 23.79±2.24 25.58±1.35 25.95±2.21 Cmpd.9 31.62±1.65 22.75±1.24 30.68±2.37 >50 >50 >50 >50 Cmpd.11 25.25±1.37 17.73±1.45 22.39±1.97 >50 49.42±4.21 >50 35.31±1.79 Cmpd.12 3.53±0.16 2.78±0.21 11.39±0.49 12.78±1.15 4.75±0.24 9.45±0.48 3.46±0.18 Cmpd.18 3.91±0.32 2.59±0.28 4.97±0.22 9.35±0.62 7.17±0.23 11.75±1.32 4.54±0.14 Cmpd.25 46.98±2.76 48.32±3.28 >50 >50 >50 >50 49.52±4.85 TubA 8.87±0.43 11.97±1.03 17.64±1.23 27.81±2.31 27.4±1.13 24.67±1.95 14.76±1.26

[0102] The inhibitory activity of the above compounds Cmpd.1, Cmpd.8, Cmpd.12 and Cmpd.18 with excellent HCT116 proliferation inhibition activity against HDAC (representative subtypes HDAC1, HDAC2, HDAC6 and HDAC8) was further evaluated. As shown in Table 5, compared with the positive control (TubA:IC 50 =15.11 nM) compared to Cmpd.12 (IC 50 =12.79 nM) and Cmpd.18 (IC 50 =5.41 nM) showed stronger inhibitory activity against HDAC6, among which Cmpd.18 showed extremely strong selective inhibition against HDAC6.

[0103] Table 5 Inhibitory effects of compounds on HDAC1, HDAC2, HDAC6 and HDAC8 enzymes (IC 50 , nM)

[0104] Cmpd. HDAC1 HDAC2 HDAC6 HDAC8 Cmpd.1 798.28±24.16 1005.69±29.45 219.52±18.49 670.15±26.29 Cmpd.8 435.62±9.21 425.78±11.47 162.47±9.12 867.51±23.78 Cmpd.12 132.53±6.54 543.56±15.97 12.79±0.85 678.45±31.69 Cmpd.18 634.19±29.37 863.01±18.31 5.41±0.31 1367.37±47.82 TubA 2356.49±69.52 >5000 15.11±1.21 957.29±35.63

[0105] In summary, Cmpd.18 can interact with key residues and bind to the active pocket in an effective conformation, which is also an important prerequisite for its enzyme inhibitory activity. The structure of Cmpd.18 is shown in formula (II):

[0106]

[0107] 2. Western blot experiment

[0108] HCT116 cells were seeded in 6-well plates (1×10 6 / well) and cultured in an incubator for 24 h, and then treated with different concentrations (0.5 μM, 1.0 μM and 2.0 μM) of Cmpd.18 and TubA (2.0 μM) for 48 h. After that, the cells were collected and lysed with RIPA lysis buffer (SW104-02, SEVEN) containing 1% PMSF (SW106-01, SEVEN). Subsequently, the concentration of each group of protein samples was determined using a BCA protein kit (SW101-02, SEVEN), and electrophoretically separated in a 12% SDS-PAGE gel system, and then transferred to a PVDF membrane (ISEQ00010, Immobilon) and blocked with skimmed milk powder (SW128-03, SEVEN) for 2-4 h. The PVDF membrane was incubated with the primary antibodies Ac-α-tubulin (3971S, Cell Signaling), Ac-H3 (9649S, Cell Signaling) and GAPDH (ET1601-4, HUABIO) at 4°C overnight. The PVDF membrane was then washed and incubated with the corresponding secondary antibodies at room temperature in the dark for 2 h. Finally, the target bands were scanned using the Odyssey infrared imaging system, and quantitative analysis was performed using ImageJ software.

[0109] In this example, the HDAC6 selective inhibitor TubA was used as a positive control, and the Western blot method was used to explore the effect of the compound Cmpd.18 of the present invention on the acetylation level of HDAC6 related substrates in HCT116 cells. The experimental results are shown in Figure 5 As shown: Compared with the control group, 0.5μM Cmpd.18 can significantly upregulate the expression of Ac-α-tubulin, but does not affect the expression of Ac-H3 protein; Cmpd.18 can significantly affect the expression level of Ac-α-tubulin in a dose-dependent manner. In addition, at the same administration concentration (2.0μM), the effect of Cmpd.18 on the expression of Ac-α-tubulin exceeds that of TubA, which is consistent with the results of the enzyme activity inhibition experiment.

[0110] 3. Cell apoptosis and cell cycle experiments

[0111] HCT116 cells in the logarithmic growth phase were digested with trypsin (25200072, Gibco) and seeded in 6-well plates (1×10 6 / well) and cultured in an incubator for 24h. Subsequently, the cells were treated with different concentrations of Cmpd.18 (0.1μM, 0.5μM and 2.5μM) for 48h. The cell cycle experiment mainly includes four steps: (i) Cell collection: digest and collect cells with trypsin, centrifuge at 1000rpm for 5min to remove the supernatant, and then wash the cells twice with pre-cooled PBS. (ii) Cell fixation: add 1mL of pre-cooled 70% ethanol to each group of cells mentioned above, fix them at 4℃ for 24h, then centrifuge at 1000rpm for 5min to remove the supernatant, and wash the cells twice with pre-cooled PBS. (iii) Cell staining: add 0.5mL PI / RNase staining buffer (550825, BD Pharmingen) to each centrifuge tube to resuspend the cells and incubate them at room temperature in the dark for 30min. (iv) Detection and analysis: use flow cytometer to detect samples, and use GraphPad Prism 8 for data analysis.

[0112] The cell apoptosis experiment mainly includes three steps: (i) Cell collection: Digest and collect cells with EDTA-free trypsin, centrifuge at 1000rpm for 5 minutes to remove the supernatant, and then wash the cells twice with pre-cooled PBS. (ii) Cell staining: First, dilute the 10×Binding buffer in the apoptosis kit (556547, BD Pharmingen) with triple distilled water to 1×Binding buffer, then add 100μL 1×Binding buffer to each system to resuspend the cells, then add 5μL FITCAnnexin V and 5μL Propidium Iodide and mix gently, protect from light, react at room temperature (25℃) for 15 minutes, and then add 400μL 1×Binding buffer to each system. (iii) Detection and analysis: Detect the samples with a flow cytometer within 1 hour, and analyze the data with GraphPad Prism 8 software. The cell cycle and apoptosis experiments were repeated three times.

[0113] Considering that Cmpd.18 exhibited ideal selective inhibitory properties among representative HDAC isoforms and showed a strong inhibitory effect on HCT116 cells, the pro-apoptotic mechanism of Cmpd.18 was further characterized by Annexin V-FITC / PI and cell cycle distribution analysis.

[0114] The results are as follows Figure 6As shown: Cmpd.18 can induce apoptosis of HCT116 cells at a lower concentration (0.5μM) (Q2: 11.97%, Q4: 4.23%); and with the increase of concentration, the ability to induce apoptosis significantly increased (from 1.17% to 41.66%), especially the proportion of late apoptotic cells increased compared with the control group (from 5.83% to 36.38%), showing an obvious dose-dependency.

[0115] Flow cytometry was then used to detect the effect of Cmpd.18 on the cell cycle distribution of HCT116 cells. Figure 7 As shown, Cmpd.18 can significantly increase the proportion of G2 / M phase (from 17.50% to 44.64%), indicating that the compound can arrest mitosis in G2 / M phase. The results show that Cmpd.18 has a good inhibitory effect on HCT116 cell proliferation by inducing cell apoptosis and arresting cell cycle.

[0116] In summary, the HDAC6 selective inhibitor 2-(2-phenoxyethyl)pyridazine-3(2H)-one hydroxamic acid derivative (Cmpd.18) prepared by the present invention has good application prospects in the field of malignant tumor treatment. Other compounds in the present invention have similar anti-proliferative effects as the listed compounds, but it should not be limited to understand that the compounds of the present invention only have the above beneficial effects.

[0117] It should be noted that the specific embodiments in the present invention are for illustration and should not be construed as limiting the scope of the present invention. In addition, it should be understood that after reading the contents of the present invention, those skilled in the art may make various changes or modifications to the present invention, and such equivalent forms also fall within the scope defined by the claims attached to this application.

Claims

1. A method for screening HDAC6 inhibitors, characterized in that: The method comprises the following steps: (1) Preparation of molecular library: Integrate the commercial molecular library, and then use Lipinski rule, Veber rule and PAINS rule to screen and remove compounds that do not meet drug-like properties and false positives; (2) Rapid virtual screening based on voting strategy: Voting strategies were established based on two CPI models, the HDACban model and the TransformerCPI2.0 model, to complete the rapid screening of the molecular library; (3) Structure-based virtual screening: Prepare receptor crystals and ligands, generate docking grids for receptor molecules based on the spatial coordinates of the ligands in the original crystals, use Glide-HTVS, Glide-SP / XP, and Prime-MMGBSA to screen small molecules based on the screening accuracy, and select molecules that meet the requirements for subsequent molecular dynamics simulations with reference to the binding affinity of TubA in HDAC6; (4) Target affinity evaluation: The GPU-accelerated PMEMD program in AMBER20 software was used to perform molecular dynamics simulation studies on the established research system.

2. The screening method according to claim 1, characterized in that The step (2) comprises: Construction of HDAC fine-tuning dataset: The relevant information of compounds in ChEMBL34 database was collected and processed as follows: (i) The biological activity indicators were set as ‘Ki’, ‘Kd’, ‘IC 50 ','EC 50 ', and the unit is uniformly converted to 'nM'; (ii) biological activity threshold (including pKi, pKd, pIC 50 and pEC 50 ) is set to 6.

5. When the biological activity threshold is greater than 6.5, it is considered that the compound interacts with the target and is assigned a value of 1; when it is less than 6.5, it indicates that there is no interaction and is assigned a value of 0; the relevant information includes protein type, molecule type and activity; Model building: The HDAC-specific CPI screening models HDACban and TransformerCPI2.0 were obtained by retraining and fine-tuning using large-scale CPI data and the constructed HDAC fine-tuning dataset.

3. The screening method according to claim 2, characterized in that The step (3) comprises: Receptor preparation: Download the required receptor crystal PDB ID: 5EDU from the Protein Data Bank (http: / / www.rcsb.org), import it into Maestro, and optimize it through the Protein Preparation Wizard, including the sorting of distribution bonds, creation of disulfide bonds, dehydration, hydrogenation, protonation state setting and energy optimization. When the RMSD between heavy atoms reaches When , the energy minimization is terminated; Ligand preparation: The LigPrep module in Maestro was used to prepare the molecular library under OPLS-2005, including setting the ionization state, generating tautomers and stereoisomers to ensure the diversity of the ligand at pH = 7.0 ± 2.0; Cascade molecular docking: The docking grid of the receptor molecule was generated according to the spatial coordinates of the ligand in the original crystal. Glide-HTVS, Glide-SP / XP and Prime-MMGBSA were used in turn for small molecule screening according to the screening accuracy. The binding affinity of TubA in HDAC6 was referred to and the molecules that met the requirements were selected for subsequent molecular dynamics simulation.

4. The screening method according to claim 3, characterized in that The step (4) comprises: Simulation system preparation: Use Gaussian09 and AMBER20 research systems to prepare, generate coordinate files of complexes, receptors, ligands, and topology files and coordinate files of solvated complexes in the SPCE water box; Molecular dynamics simulation: The simulation system was optimized using the steepest descent method and the conjugate gradient method to eliminate high-energy collisions between molecules, and then heated up, and the PMEMD program was used to perform a 1μs simulation under the NPT ensemble; Post-processing analysis based on simulation trajectory: Evaluate the equilibrium stability of the simulation system and extract representative conformations, calculate binding free energy and residue energy decomposition based on representative conformations; ΔG MM / GBSA =ΔE vdw +ΔE ele +ΔG pol +ΔG nonpol ; Where, ΔE vdw and ΔE ele Refers to van der Waals force and electrostatic force respectively; ΔG pol and ΔG nonpol Refers to polar interaction and nonpolar interaction respectively.

5. Use of the screening method according to any one of claims 1 to 4 in screening HDAC6 inhibitors or anti-tumor drugs.

6. A compound represented by the following formula (I), or a pharmaceutically acceptable salt, hydrate, deuterated substance, or prodrug thereof; in, A1, A2, A3 are independently selected from CR' or N, R' is a substituent of C; R' and R1 are independently selected from H, C1-C6 alkoxy, C1-C6 alkylamino, C1-C6 alkoxycarbonyl, substituted or unsubstituted C1-C6 acylamino, substituted or unsubstituted C1-C 10 Alkyl, substituted or unsubstituted C6-C 12 aryl or heteroaryl, substituted or unsubstituted 3-12 membered alicyclic or heterocyclic group, substituted or unsubstituted benzyl; X is selected from (CH2) n , n is any integer greater than or equal to 0; Y is selected from C, N or O; Z is selected from substituted or unsubstituted C1-C 10 Alkyl, C1-C6 alkoxy, C1-C6 alkylamino, C1-C6 alkoxycarbonyl, substituted or unsubstituted C6-C 12 Aryl or heteroaryl, substituted or unsubstituted 3-12 membered alicyclic or heterocyclic group; W is selected from (CH2) n , substituted or unsubstituted C6-C 14 Aryl, C5-C 14 Heteroaryl, C7-C 12 Arylalkyl, C6-C 12 Arylheteroalkyl; n is any integer greater than or equal to 0; The substitution refers to that one or more hydrogen atoms on the group are replaced by a substituent selected from any of the following: Halogen atoms, carbonyl, carboxyl, hydroxyl, amino, nitro, cyano, C1-C 10 Alkyl, C1-C6 alkoxy, C1-C6 alkylamino, C1-C6 alkoxycarbonyl, C1-C6 acylamino, C1-C6 10 Aryl, five-membered heteroaryl, six-membered heteroaryl; or C1-C 10 Alkyl, C6-C 10 Aryl, five-membered heteroaryl, six-membered heteroaryl; preferably C1-C6 alkyl, C1-C6 alkoxy or C1-C6 alkylamino.

7. The compound according to claim 6, or a pharmaceutically acceptable salt, hydrate, deuterated substance, or prodrug thereof, characterized in that: The structural formula of the compound is shown in formula (II):

8. Use of the compound according to claim 6 or 7, or a pharmaceutically acceptable salt, hydrate, deuterated substance or prodrug thereof in the preparation of an HDAC6 inhibitor.

9. Use of the compound according to claim 6 or 7, or a pharmaceutically acceptable salt, hydrate, deuterated substance or prodrug thereof in the following pharmaceutical applications: (1) Application in the preparation of anti-tumor drugs; (2) Application in the preparation of drugs for treating autoimmune diseases; (3) Application in the preparation of anti-inflammatory drugs; (4) Application in the preparation of drugs for treating Alzheimer's disease.

10. The use according to claim 9, characterized in that The tumor is selected from leukemia, colorectal cancer, melanoma, lung cancer, breast cancer, and cervical cancer.

Images