Screening of CYP7A1 inhibitors and their application in the preparation of anti-hepatocellular carcinoma drugs

Abstract

This application provides a method for screening inhibitors targeting CYP7A1 and their use in the preparation of anti-liver cancer drugs, relating to the field of tumor targeted drug design technology. The method for screening inhibitors targeting CYP7A1 protein includes: using the crystal structure of human cholesterol 7α-hydroxylase as a receptor model, precise docking screening is performed using software to screen drug-like molecules with molecular weights of 250-500 Da and lipid-water partition coefficients of -1 to 5 from a compound database; a graph neural network active learning model is used to predict and rank the protein-small molecule binding affinity of compounds in the database; the ΔG between candidate molecules and CYP7A1 is calculated using molecular mechanics / generalized Born surface area methods, with molecules having ΔG ≤ -40 kcal / mol considered potential inhibitors. The inhibitors of this application do not rely on the enzymatic activity of CYP7A1 to achieve anti-tumor effects and do not significantly affect cholesterol and bile acid metabolism levels.

Description

Technical Field

[0001] This application relates to the field of tumor-targeted drug design technology, and in particular to the screening of an inhibitor targeting CYP7A1 and its use in the preparation of anti-liver cancer drugs. Background Technology

[0002] Liver cancer is one of the malignant tumors with high incidence and mortality rates worldwide. Current systemic treatment of liver cancer, especially for advanced-stage patients, still faces significant challenges. While commonly used first-line multi-kinase inhibitors such as sorafenib and lenvatinib can prolong patient survival to some extent, they generally suffer from strong drug resistance, short median progression-free survival, and side effects due to multi-target effects.

[0003] Intervening in tumor growth from the perspective of metabolic reprogramming is an important strategy in cancer treatment. However, traditional drugs targeting metabolic enzymes, such as IDH1 / 2 mutation inhibitors, have been found to potentially cause systemic metabolic disorders in clinical applications, such as a high incidence of hyperbilirubinemia. CYP7A1 is the rate-limiting enzyme in the classical pathway of bile acid synthesis, and research has long focused on its 7α-hydroxylase function. Reported CYP7A1 inhibitors (such as 7-ketocholesterol) mainly regulate bile acid metabolism by competitively binding to cholesterol and inhibiting enzyme activity, but this may lead to metabolic side effects such as enterohepatic circulation disorders, and it is difficult to distinguish their effects on enzyme catalytic function from other potential biological functions.

[0004] In terms of drug screening technology, traditional virtual screening methods have limitations when dealing with ultra-large-scale compound libraries. For example, screening libraries with tens of millions of compounds using tools such as Autodock Vina typically takes months and has a high false positive rate. Single computational screening strategies have limited capabilities in exploring new chemical spaces and may result in insufficient structural diversity of lead compounds. Summary of the Invention

[0005] The purpose of this application is to overcome at least some of the deficiencies of the prior art and provide a method for screening inhibitors based on novel functions of CYP7A1, the inhibitor itself, and its application in anti-liver cancer drugs. This inhibitor achieves its anti-tumor effect independently of the enzymatic activity of the CYP7A1 protein, and can block CYP7A1-mediated autophagosome-lysosome fusion and / or increase CD8+. + T-cell infiltration. At effective antitumor concentrations, the inhibitor showed an inhibition rate of less than 20% on cholesterol 7α-hydroxylase activity in cell or animal models, and / or had no statistically significant effect on serum total bile acid levels (p>0.05). The specific technical solution is as follows:

[0006] The first aspect of this application provides a method for screening inhibitors targeting the CYP7A1 protein, comprising the following steps:

[0007] Step S1: Using the human cholesterol 7α-hydroxylase crystal structure PDB ID: 3V8D as the receptor model, precise docking screening was performed using Schrödinger Maestro software to screen drug-like molecules with molecular weights of 250 Da-500 Da and lipid-water partition coefficients of -1 to 5 from the Zinc15 compound database.

[0008] Step S2: Using an active learning model of graph neural network, predict the protein-small molecule binding affinity of compounds in the Enamine RealSpace database, and sort the prediction results.

[0009] Step S3: Calculate the binding free energy of the candidate molecules obtained in Step S1 and / or Step S2 with the CYP7A1 protein using the molecular mechanics / generalized Born surface area method; wherein, molecules with a binding free energy ΔG ≤ -40 kcal / mol are selected as potential inhibitors targeting the CYP7A1 protein.

[0010] In some embodiments of this application, step S1 includes: defining a ligand-binding pocket composed of His101, Trp284, Val281, Ile125, Phe129, Gly485, Ile114, Leu486, Leu104, Leu361, Ser105, and Phe102 residues; identifying the carbonyl oxygen of the Leu361 residue and the carbonyl oxygen of the Gly485 residue as key hydrogen bond interaction sites; docking compounds that conform to the pharmacophore model constructed based on the PDB 3SN5 and 3V8D cocrystal ligands to the active site of CYP7A1, and sorting them from high to low docking scores.

[0011] In some embodiments of this application, step S2 includes: using Trp284, Gly485, and Ser360 as key binding sites to describe the interaction between CYP7A1 and ligands; and using protein-small molecule complex data from the PDBbind v2020 database to train and validate the active learning model.

[0012] The second aspect of this application provides the use of the screening method described in the first aspect of this application in the preparation of anti-hepatocellular carcinoma drugs.

[0013] A third aspect of this application provides the use of an inhibitor targeting the CYP7A1 protein in the preparation of an anti-liver cancer drug, wherein the inhibitor is used to block CYP7A1-mediated autophagosome-lysosome fusion and / or increase CD8+. + T-cell infiltration.

[0014] In some embodiments of this application, the inhibitor is obtained by screening using the screening method described in the first aspect of this application.

[0015] In some embodiments of this application, the inhibitor targeting CYP7A1 is selected from compounds represented by any of the following general formulas I to V, their stereoisomers, or pharmaceutically acceptable salts thereof:

[0016] , , , , ;

[0017] in,

[0018] In Formula I, Ra is selected from or Q1 is selected from hydrogen, -C(CH3)2, -C(O)CH3 or -C(O)O-CH2CH3; Rb, Rc and Rd are each independently selected from N or CH;

[0019] In Formula II, R1 and R4 are each independently selected from hydrogen or -O-CH3, R2, R3, R5, and R7 are each independently selected from hydrogen or methyl, and R6 is selected from methyl.

[0020] In Formula III, X1 is selected from N or CH, X2 is selected from hydrogen or Cl, X3 is selected from -CH2- or -C(O)-, and X4 is selected from... , or ;

[0021] In Formula IV, Y1 is selected from hydrogen or Y2 is selected from hydrogen or hydroxyl; Indicates a double bond or a single bond;

[0022] In formula V, Z1 is selected from hydrogen or hydroxyl, and Z2 is selected from... , or .

[0023] In some embodiments of this application, the inhibitors targeting CYP7A1 are selected from the following compounds, their stereoisomers, or pharmaceutically acceptable salts thereof:

[0024] .

[0025] In some embodiments of this application, the inhibitors targeting CYP7A1 are selected from the following compounds, their stereoisomers, or pharmaceutically acceptable salts thereof:

[0026] .

[0027] In some embodiments of this application, the inhibitors targeting CYP7A1 are selected from the following compounds, their stereoisomers, or pharmaceutically acceptable salts thereof:

[0028] .

[0029] The beneficial effects of this application are as follows: Through virtual screening, molecular dynamics simulation, and in vitro and in vivo experimental verification, this application reveals for the first time that the cholesterol-metabolizing enzyme CYP7A1 promotes liver cancer progression through a mechanism independent of its classical enzyme activity, specifically by promoting autophagosome-lysosome fusion and reducing CD8+ T cell infiltration. Small molecule inhibitors obtained based on this novel mechanism can effectively block the above process, thereby inhibiting tumor growth. Most importantly, the lead compounds (such as CA4), at effective antitumor concentrations, do not significantly affect the activity of cellular cholesterol 7α-hydroxylase or the overall cholesterol / bile acid metabolism level, providing a new pathway for developing highly selective and low-toxicity anti-liver cancer drugs.

[0030] Of course, implementing any product or method of this application does not necessarily require achieving all of the advantages described above at the same time. Attached Figure Description

[0031] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other embodiments can be obtained based on these accompanying drawings.

[0032] Figure 1 This document outlines the research process for non-enzymatic activity-dependent oncogenic mechanisms of CYP7A1 as a novel therapeutic target for liver cancer, and provides an overall experimental roadmap from target screening and mechanism validation to lead compound discovery and functional validation.

[0033] Figure 2A This is a representative immunohistochemical image of CYP7A1 in liver cancer tissue (T) and adjacent normal tissue (N) in liver cancer patient tissue sample 121 (case 121) in Example 1 of this application;

[0034] Figure 2B This is a representative immunohistochemical image of CYP7A1 in liver cancer tissue (T) and adjacent normal tissue (N) in liver cancer patient tissue sample 145 (case 145) in Example 1 of this application;

[0035] Figure 2CThis is a heatmap comparing the IHC staining scores of CYP7A1 in liver cancer tissue (T) and adjacent normal tissue (N) in Example 1 of this application.

[0036] Figure 3 This is a graph showing the relationship between CYP7A1 expression levels and patient survival in Example 1 of this application.

[0037] Figure 4 This is a protein band image obtained by Western blot analysis of protein levels in Example 2 of this application;

[0038] Figure 5 This is a diagram showing the results of the CYP7A1 expression and cell proliferation experiment in Example 2 of this application;

[0039] Figure 6 This is a diagram showing the Transwell migration experiment results in Example 2 of this application;

[0040] Figure 7A This is a schematic diagram of the intervention in the orthotopic liver cancer model in Example 2 of this application;

[0041] Figure 7B The tumor size of different groups of mouse liver orthotopic transplantation models in Example 2 of this application is shown;

[0042] Figure 7C This demonstrates the effect of CYP7A1 expression on tumor growth in the orthotopic hepatocellular carcinoma model of Example 2 of this application;

[0043] Figure 8A This is a schematic diagram of the patient-derived xenotransplantation (PDX) model intervention in Example 2 of this application;

[0044] Figure 8B The tumor volume of different groups of PDX models in Embodiment 2 of this application is shown;

[0045] Figure 8C This is a graph showing the tumor volume change of different groups of PDX models in Example 2 of this application;

[0046] Figure 9 This shows the location of the mutation site of the CYP7A1 enzyme inactivation mutant in Example 3 of this application within the full length of CYP7A1;

[0047] Figure 10 This is a schematic diagram illustrating the changes in the catalytic substrate in Example 3 of this application;

[0048] Figure 11 This is a graph showing the verification results of the catalytic activity of the CYP7A1 enzyme inactivation mutant in Example 3 of this application;

[0049] Figure 12This is a diagram showing the experimental results of CYP7A1 promoting liver cancer cell proliferation independently of enzyme activity in Example 3 of this application;

[0050] Figure 13A The results of differential gene function enrichment in Example 4 of this application are shown;

[0051] Figure 13B This is a heatmap drawn based on the expression levels of differentially expressed genes among samples in Example 4 of this application;

[0052] Figure 14 This is a graph showing the results of Western blot analysis of the expression levels of autophagy marker proteins LC3 and p62 after knocking down CYP7A1 in Example 4 of this application.

[0053] Figure 15A This is a confocal microscope image of the changes in autophagic flux detected by double-labeled adenovirus after knocking down CYP7A1 in Example 4 of this application;

[0054] Figure 15B The number of autophagosomes-lysosomes showing changes in autophagic flux detected by double-labeled adenovirus after knocking down CYP7A1 in Example 4 of this application is shown.

[0055] Figure 16 This is an electron micrograph of autolysosomes formed after CYP7A1 knockdown in Example 4 of this application;

[0056] Figure 17 This is a confocal microscope image of the lysosomal fluorescence intensity of the CYP7A1 knockdown group in Example 4 of this application;

[0057] Figure 18 This is a cell viability graph from the complementation experiment of CYP7A1 regulating tumor progression by promoting autophagy in Example 4 of this application;

[0058] Figure 19 This is a graph showing the results of the CYP7A1 and p62 co-precipitation immunoprecipitation experiment in Example 5 of this application;

[0059] Figure 20 This is a graph showing the GST pull-down experimental results of CYP7A1 and p62 in Example 5 of this application;

[0060] Figure 21A This is an immunohistochemical image of CYP7A1 expression and CD8+ T cell infiltration in Example 6 of this application;

[0061] Figure 21B This is a graph showing the statistical results of the CYP7A1 expression and CD8+ T cell infiltration experiment in Example 6 of this application;

[0062] Figure 22 This is a graph showing the contribution of key forces in Embodiment 7 of this application.

[0063] Figure 23 This illustrates the major interaction between the ligand (purple) and the receptor in the CYP7A1 crystal configuration (PDB ID: 3V8D) of Example 7 of this application;

[0064] Figure 24 The structure of the 3V8D and 3SN5 ligands after superposition in Example 7 of this application is shown as a pharmacophore;

[0065] Figure 25 IC of the lead compound CA4 in Example 9 of this application 50 Measurement curve;

[0066] Figure 26 This is a graph showing the effect of the lead compound CA4 in Example 9 of this application on primary hepatocytes.

[0067] Figure 27 This is a diagram showing the effect of the lead compound CA4 in Example 9 of this application on the colony formation of liver cancer cells;

[0068] Figure 28 This is a diagram showing the cell scratch test results in Example 9 of this application;

[0069] Figure 29A This is a schematic diagram illustrating the intervention of the lead compound CA4 in a liver cancer animal model in Example 9 of this application;

[0070] Figure 29B This is a comparison image of the size of subcutaneous tumors in nude mice after treatment in Example 9 of this application;

[0071] Figure 29C This is a growth curve of subcutaneous tumors in nude mice after treatment in Example 9 of this application;

[0072] Figure 30A The image shows the HE staining results of the heart, liver, spleen, lungs, and kidneys of nude mice in Example 9 of this application.

[0073] Figure 30B This is a graph showing the serum biochemical indicators of nude mice in Example 9 of this application;

[0074] Figure 31A This is a graph showing the results of a thermal migration assay (CETSA) of the lead compound CA4 and CYP7A1 in Example 10 of this application.

[0075] Figure 31B This is a CETSA melting curve of the lead compound CA4 and CYP7A1 in Example 10 of this application;

[0076] Figure 32 This is a SPR binding sensing image of the lead compound CA4 and CYP7A1 in Example 10 of this application;

[0077] Figure 33 This is a Western blot diagram from Example 10 of this application, validating the inhibition of liver cancer progression by CA4 directly targeting CYP7A1 to regulate autophagy.

[0078] Figure 34A This is a confocal microscope image from Example 10 of this application, verifying the direct targeting of CYP7A1 by CA4 to regulate autophagolysosome formation using adenovirus double-labeling.

[0079] Figure 34B This is a diagram illustrating the inhibition of liver cancer progression by CA4 directly targeting CYP7A1 to regulate autophagy and lysosome formation using a dual-labeled adenovirus in Example 10 of this application. Detailed Implementation

[0080] The technical solutions of this application will be clearly and completely described below with reference to the embodiments and accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. All other embodiments obtained by those skilled in the art based on this application are within the scope of protection of this application.

[0081] The first aspect of this application provides a method for screening inhibitors targeting the CYP7A1 protein (i.e., a CYP7A1 inhibitor screening method based on a dual AI synergistic strategy), which includes the following steps:

[0082] Step S1: Using the human cholesterol 7α-hydroxylase crystal structure (PDB ID: 3V8D) as the receptor model, precise docking screening was performed using Schrödinger Maestro software to screen drug-like molecules with molecular weights of 250 Da-500 Da and lipid-water partition coefficients of -1 to 5 from the Zinc15 compound database.

[0083] Step S2: Using an active learning model of graph neural network (GNN), predict the protein-small molecule binding affinity of compounds in the Enamine RealSpace database, and sort the prediction results;

[0084] Step S3: Calculate the binding free energy of the candidate molecules obtained in Step S1 and / or Step S2 with the CYP7A1 protein using the molecular mechanics / generalized Born surface area (MM / GBSA) method; wherein, molecules with a binding free energy ΔG ≤ -40 kcal / mol are selected as potential inhibitors targeting the CYP7A1 protein.

[0085] In some embodiments of this application, a ligand-binding pocket consisting of His101, Trp284, Val281, Ile125, Phe129, Gly485, Ile114, Leu486, Leu104, Leu361, Ser105, and Phe102 residues is defined; the carbonyl oxygen of the Leu361 residue and the carbonyl oxygen of the Gly485 residue are identified as key hydrogen bond interaction sites; compounds conforming to the pharmacophore model constructed based on the PDB 3SN5 and 3V8D co-crystal ligands are docked to the active site of CYP7A1 and sorted according to docking scores from high to low.

[0086] In some embodiments of this application, step S2 includes: using Trp284, Gly485, and Ser360 as key binding sites to describe the interaction between CYP7A1 and ligands; and using protein-small molecule complex data from the PDBbind v2020 database to train and validate the active learning model.

[0087] In some embodiments of this application, the method for screening inhibitors targeting the CYP7A1 protein includes the following synergistic steps: (1) performing precise docking with Glide-XP using Schrödinger Maestro software to screen the ZINC15 database:

[0088] Using the crystal structure of CYP7A1 PDB ID:3V8D (protein name: human cholesterol 7α-hydroxylase) as the receptor model, precise Glide-XP docking was performed using Schrödinger Maestro software to screen drug-like molecules with molecular weights of 250-500 Da and clog P 1-5 from the ZINC15 database. Specifically, this involved: a) defining ligand-binding pockets including His101, Trp284, Val281, Ile125, Phe129, Gly485, Ile114, Leu486, Leu104, Leu361, Ser105, and Phe102; b) identifying two sites that form stable hydrogen bond interactions with known ligands: the carbonyl oxygen of Leu361 residue and the carbonyl oxygen of Gly485 residue; and c) finally docking the ligand set that conforms to the pharmacophore model to the active site of CYP7A1 and sorting them according to the highest (most negative) docking score.

[0089] (2) The Enamine RealSpace database was filtered using the deep active learning algorithm of MolProphet™ software:

[0090] The Enamine RealSpace database was iteratively screened using the MolProphet™ platform, including: a) using Trp284, Gly485, and Ser360 as key binding sites to describe the interaction between CYP7A1 and ligands; b) using the PubBindv2020 public database to select 19,443 protein-small molecule compound data for active learning to obtain predicted protein-small molecule compound affinity values ​​and then ranking them.

[0091] This application provides a dual-AI collaborative virtual screening method for therapeutic target development, comprising: (a) using the Schrödinger Maestro platform combined with the Glide-XP precise docking strategy to perform pharmacophore-based initial screening of 9,204,306 drug-like molecules in the ZINC15 database; (b) using the active learning algorithm of the MolProphet™ platform to explore the substructure space of 3.9 billion compounds in the Enamine Real Space database through multi-round iterative optimization models; and (c) predicting and optimizing candidate molecules by combining free energy calculations (MM / GBSA ΔG≤-40 kcal / mol). The dual-AI collaborative virtual screening method developed in this application integrates Schrödinger Glide-XP precise docking with the MolProphet™ active learning algorithm, breaking through the efficiency bottleneck of traditional screening in ultra-large-scale compound libraries. This method ensures the identification of key interactions through structure-guided precise docking screening and improves efficiency by utilizing active learning-driven exploration of a vast chemical space. Ultimately, it successfully obtained the world's first class of highly efficient small molecule inhibitors targeting the non-enzymatic function of CYP7A1, providing a new pathway for precision treatment of liver cancer. It also provides an innovative compound library and efficient screening methodology for developing novel anti-liver cancer drugs targeting the non-enzymatic function of CYP7A1.

[0092] The second aspect of this application provides the use of the screening method described in the first aspect of this application in the preparation of anti-hepatocellular carcinoma drugs.

[0093] A third aspect of this application provides the use of an inhibitor targeting the CYP7A1 protein in the preparation of an anti-liver cancer drug, wherein the inhibitor is used to block CYP7A1-mediated autophagosome-lysosome fusion and / or increase CD8+. + T-cell infiltration.

[0094] In some embodiments of this application, the inhibitor is obtained by screening using the screening method described in the first aspect of this application.

[0095] In some embodiments of this application, the inhibitor is selected from compounds represented by any of the following general formulas I to V, their stereoisomers, or pharmaceutically acceptable salts thereof:

[0096] , , , , ;

[0097] in,

[0098] In Formula I, Ra is selected from or Q1 is selected from hydrogen, -C(CH3)2, -C(O)CH3 or -C(O)O-CH2CH3; Rb, Rc and Rd are each independently selected from N or CH;

[0099] In Formula II, R1 and R4 are each independently selected from hydrogen or -O-CH3, R2, R3, R5, and R7 are each independently selected from hydrogen or methyl, and R6 is selected from methyl.

[0100] In Formula III, X1 is selected from N or CH, X2 is selected from hydrogen or Cl, X3 is selected from -CH2- or -C(O)-, and X4 is selected from... , or ;

[0101] In Formula IV, Y1 is selected from hydrogen or Y2 is selected from hydrogen or hydroxyl; Indicates a double bond or a single bond;

[0102] In formula V, Z1 is selected from hydrogen or hydroxyl, and Z2 is selected from... , or .

[0103] In some embodiments of this application, the inhibitors targeting CYP7A1 are selected from the following compounds, their stereoisomers, or pharmaceutically acceptable salts thereof:

[0104] .

[0105] In this application, compounds 58, 90, 93 and 100 all include both cis and trans structures.

[0106] In this application, compounds falling within the scope of general formula I include: compound 30 (Terconazole, IC50). 50 =24.55 μM), compound 34 (Elubiol, neoconazole, IC50) 50=10.48 μM), compound 56 (Deacylketoconazole, IC50 = 10.48 μM), 50 =9.2 μM), compound 74 (ketoconazole, IC50) 50 =24.5 μM), compound 75 (Ketoconazole impurity E, IC50). 50 =11.85μM).

[0107] Compounds falling within the general formula II range include: compound 60 (Nepetin, zeylan flavonoid, IC50). 50 =10.36 μM), compound 81 (6-Methoxytricin, 6-methoxypiperidine, IC50) 50 =19.13 μM), compound 87 (Dabcyl acid, IC50) 50 =1.45μM).

[0108] Compounds falling within the scope of general formula III include: compound 66 (IC 50 =72.95μM), compound 71 (IC) 50 =21.53 μM), compound 72 (IC) 50 =24.59μM).

[0109] Compounds falling within the range of formula IV include: compound 64 (IC). 50 =18.73μM), compound 65 (IC) 50 =10.88μM).

[0110] Compounds located within the range of general formula V include: compound 58 (combretastatin A4, IC50). 50 =0.00648 μM), compound 90 (Combretastatin A1, IC50) 50 =0.3178μM), compound 93 (IC50) 50 =9.78 μM), compound 100 (IC50) 50 =0.6158μM).

[0111] In some embodiments of this application, the inhibitors targeting CYP7A1 are selected from the following compounds, their stereoisomers, or pharmaceutically acceptable salts thereof:

[0112] .

[0113] In this application, compound 2 (IC)50 =24.38μM), compound 8 (IC) 50 =51.14 μM), compound 11 (IC500) 50 =53.47 μM), compound 16 (Ezetimibe ketone, IC50) 50 =70.16μM), compound 22 (IC 50 =13.59μM), compound 35 (IC) 50 =86.72μM), compound 59 (IC50) 50 =22.07 μM), compound 61 (Bucindolol, IC50, 22.07 μM), 50 =28.39μM), compound 98 (IC) 50 =48μM).

[0114] This application provides 26 compounds (IC) that inhibit CYP7A1. 50 (<100 μM), and can be used to treat liver cancer. All 26 compounds mentioned above are commercially available or can be synthesized using conventional methods in the art.

[0115] In some embodiments of this application, the inhibitor targeting CYP7A1 is selected from the following compounds (cis structure of compound 58), their stereoisomers, or pharmaceutically acceptable salts thereof:

[0116] .

[0117] The cis structure of compound 58 provided in this application (i.e., lead compound CA4, Combretastatin A4) has an effect on the IC50 of Huh7 cells. 50 =6.48 nM, with the best activity compared to CYP7A1 KD=1.1 μM. In a xenograft (PDX) model derived from 3 patients with hepatocellular carcinoma, CA4 administration reduced tumor volume by 85%, with efficacy comparable to the clinical drug lenvatinib, and showed no significant toxicity to primary hepatocytes (selectivity index SI>100).

[0118] The fourth aspect of this application provides the use of CYP7A1 protein as a target in screening and / or preparing anti-hepatocellular carcinoma drugs; wherein, the inhibitor targeting the CYP7A1 protein is used to block CYP7A1-mediated autophagosome-lysosome fusion and / or increase CD8+. + T-cell infiltration.

[0119] In some embodiments of this application, the inhibitor is obtained by screening using the screening method described in the first aspect of this application.

[0120] In some embodiments of this application, the inhibitor is the compound described in the third aspect of this application, its stereoisomer, or a pharmaceutically acceptable salt thereof.

[0121] This application reveals for the first time that the cholesterol-metabolizing enzyme CYP7A1 promotes hepatocellular carcinoma (HCC) progression through a mechanism independent of its classical enzyme activity. It discovers that CYP7A1 drives tumor progression in HCC by promoting autophagy-lysosome formation (independent of its cholesterol 7α-hydroxylase activity) and reducing CD8+ T cell infiltration. The inhibitors targeting CYP7A1 described in this application achieve antitumor effects independent of the enzymatic activity of CYP7A1, and can block CYP7A1-mediated autophagosome-lysosome fusion and / or increase CD8+ T cell infiltration. + T cell infiltration was observed, but cholesterol and bile acid metabolism levels were not significantly affected. Therefore, CYP7A1 protein could serve as a novel biological target.

[0122] This application Figure 1 This document outlines the research process for non-enzyme-dependent oncogenic mechanisms of CYP7A1 as a novel therapeutic target for liver cancer, and provides a comprehensive experimental roadmap from target screening and mechanism validation to lead compound discovery and functional validation.

[0123] The following examples illustrate the implementation of this application in more detail. Various tests and evaluations were conducted according to the methods described below. Unless otherwise specified, the percentage content used in this application has the meaning commonly used in the art; otherwise, for solid-liquid mixtures and solid-phase-solid mixtures, it refers to mass percentage, and for liquid-phase-liquid mixtures, it refers to volume percentage.

[0124] Example 1: Validation of the clinical correlation between CYP7A1 and liver cancer

[0125] 1. Analysis of CYP7A1 protein expression levels in clinical samples

[0126] Experimental materials: Paraffin-embedded blocks of surgically removed liver cancer tissue and paired adjacent normal tissue were collected from 183 patients with liver cancer. All procedures strictly followed the Declaration of Helsinki and were approved by the Ethics Committee of the First Affiliated Hospital of Zhejiang University School of Medicine (IIT20240651B-R1).

[0127] Experimental Procedure: Paraffin-embedded blocks of liver cancer tissue and adjacent non-cancerous tissue were cut to obtain paraffin sections. After antigen retrieval, the paraffin sections were incubated overnight at 4°C with CYP7A1 protein primary antibody (purchased from Affinity Biotech, catalog number DF2612), developed using DAB staining, and scored using H-score (score range 0-300).

[0128] Results analysis: CYP7A1 expression results in tissue samples from 183 liver cancer patients are as follows: Figure 2A , Figure 2B and Figure 2C As shown, where, Figure 2A Immunohistochemical representations of CYP7A1 in liver cancer tissue (T) and adjacent normal tissue (N) from liver cancer patient tissue sample 121 (case 121); Figure 2B Immunohistochemical representations of CYP7A1 in liver cancer tissue (T) and adjacent normal tissue (N) from tissue sample 145 of a liver cancer patient (case 145); Figure 2C A heatmap comparing IHC staining scores of CYP7A1 in hepatocellular carcinoma (T) tissues and adjacent normal tissues (N). The results show that immunohistochemical (IHC) analysis of tissue samples from 183 hepatocellular carcinoma patients revealed a positive rate of 62.30% for high expression of CYP7A1 in hepatocellular carcinoma tissues, significantly higher than that in adjacent normal tissues (p<0.001).

[0129] 2. Analysis of CYP7A1 protein expression levels and prognostic survival

[0130] After adjusting for factors such as age, sex, and tumor stage using a Cox proportional hazards model and multivariate analysis, patients were grouped according to their relative CYP7A1 IHC score expression between the cancer cell and adjacent tissue. A score higher in the cancer cell than in the adjacent tissue defined as a high CYP7A1 expression group, and a score lower in the cancer cell than in the adjacent tissue defined as a low CYP7A1 expression group. Prognostic analysis was performed using Kaplan-Meier survival curves, and the Log-rank test was used to assess the difference in overall survival between the two groups.

[0131] The results showed that the correlation between CYP7A1 high and low expression and overall survival was analyzed using Kaplan-Meier curves. Figure 3 As shown, the overall survival of the CYP7A1 high expression group (red) was significantly lower than that of the CYP7A1 low expression group (blue), with the dashed line representing the 95% confidence interval. The median overall survival (OS) of the CYP7A1 high expression group was 18.4 months, while that of the CYP7A1 low expression group was 36.2 months (p=0.001). Cox regression analysis showed that CYP7A1 high expression was an independent prognostic factor (HR=2.38, p=0.0046).

[0132] Example 2: CYP7A1 Functional Verification

[0133] 1. Cell model construction

[0134] Experimental materials: human hepatocellular carcinoma cell lines Hep3B, HepG2, Huh7, HCC-LM3, SMMC-7721 and THLE2 normal hepatocytes (identified by STR), protease phosphatase inhibitor (Biosharp, catalog number: BL1439A), cell lysis buffer (Beyotime, catalog number: P0013B), BCA kit (Beyotime, catalog number: P0011), CYP7A1 antibody.

[0135] Experimental procedure: All cells were cultured in DMEM medium containing 10 vol% FBS at 37°C and 5 vol% CO2.

[0136] Western blot analysis of protein levels: All cells were collected and lysed separately using lysis buffer containing a protease phosphatase inhibitor. After centrifugation, the supernatant was collected, and protein concentration was determined using a BCA protein assay kit with consistent loading amounts. Equal amounts of protein from each group were separated by 10% SDS-PAGE (sodium dodecyl sulfate polyacrylamide gel) and transferred to polyvinylidene fluoride (PVDF) membranes. After blocking with 5% skim milk powder / TBST for 1 hour at room temperature, the corresponding CYP7A1 protein primary antibody was added and incubated overnight at 4°C. The next day, after washing with TBST, HRP-labeled secondary antibody was added and incubated at room temperature for 1 hour. After washing again with TBST, enhanced chemiluminescence (ECL) substrate development solution was added, and protein band images were captured. The grayscale values ​​of the protein bands were measured using ImageJ software, normalized to internal controls, and the relative expression level of CYP7A1 protein was calculated and statistically analyzed. Results analysis: Protein band images are shown below. Figure 4 As shown in the figure (where actin represents the internal reference), it can be seen that the expression level of CYP7A1 protein in cell lines HepG2 and HCC-LM3 is relatively high.

[0137] 2. Validation of gene knockdown effect and in vitro / in vivo experiments.

[0138] CYP7A1 gene knockdown group (shCYP7A1): The CYP7A1-shRNA plasmid (sequence 5'-ACGTGACACGTTCGGAGAATC-3', SEQ ID NO.1) was constructed using the U6 vector. A transfection complex was then prepared using the standard packaging system of psPAX2 plasmid, pMD2.G plasmid, and transfection reagent (Polyplus, catalog number: PT-114-15). This complex was transfected into 293T cells, and viral supernatant was collected 48 hours later. The viral supernatant was used to further infect HepG2 and HCC-LM3 cells, respectively, and stable transfected HepG2 and HCC-LM3 cell lines were selected using puromycin (2 μg / mL).

[0139] (1) To evaluate the effect of CYP7A1 knockdown on the proliferation of hepatocellular carcinoma cells, HCC-LM3 cells were first subjected to in vitro proliferation detection using the 5-ethynyl-2′-deoxyuridine (EdU) method. Cells from the control group (shNC group) and the CYP7A1 knockdown group (shCYP7A1 group) were cultured to the required experimental conditions. EdU incorporation was performed according to the instructions of the commercial EdU-488 detection kit (Beyotime, catalog number: C0071L), and Hoechst nuclear staining was performed according to the instructions of Hoechst nuclear staining solution (Beyotime, catalog number: C1017). The number of EdU-positive cells and the total number of cells in each field of view were recorded using microscopy, and the EdU positivity rate (%) was calculated. EdU positivity rate (%) = Number of EdU-positive cells / Total number of cells × 100%. Experimental replicates included at least three independent biological replicates.

[0140] The results of EdU positivity rates in each group of cells are as follows: Figure 5 As shown, the proliferation rate of cells in the shCYP7A1 group was significantly lower than that in the shNC group 72 h after infection (Edu positivity rate: shNC group 48.83% VS shCYP7A1 group 25.23%, p=0.0064). The above results indicate that the in vitro proliferation rate of human hepatocellular carcinoma cell line HCC-LM3 decreased after CYP7A1 gene knockdown treatment, that is, knockdown of CYP7A1 (shRNA efficiency >80%) can significantly inhibit the proliferation of HCC LM3 cells.

[0141] (2) To evaluate the effect of CYP7A1 gene knockdown on the motility of the human hepatocellular carcinoma cell line HepG2, the Transwell migration assay was used to compare the stable transgenic cells in the shNC group and the shCYP7A1 group.

[0142] The experiment was conducted using a commercial Transwell chamber (Corning, catalog number: 3422) and methods known in the art. The specific steps included: cell preparation, seeding cells in the upper chamber of the Transwell, setting up culture conditions in the lower chamber to induce migration, fixing with 4% paraformaldehyde and staining with crystal violet after 48 hours of culture, and using ImageJ software to analyze the migration area to assess migration and invasion capabilities.

[0143] Transwell migration experiment results are as follows: Figure 6 As shown, compared with the control group (shNC group), the number of Transwell cells in the CYP7A1 knockdown group (shCYP7A1 group) was reduced (control group 127.66 vs knockdown group 45.33, p<0.001); the above results indicate that the motility of the human hepatocellular carcinoma cell line HepG2 decreased after CYP7A1 gene knockdown treatment.

[0144] 3. Animal model validation

[0145] Establishing a DEN+CCl4 model: Male C57BL / 6 mice (6 weeks old) were intraperitoneally injected with diethylnitrosamine (DEN) (25 mg / kg) to initiate liver cancer. One week later, they were intraperitoneally injected weekly with 10% CCl4 solution (5 mL / kg) for 16 weeks. After ultrasound verification of tumor formation, the control group (shNC group) and the CYP7A1 knockdown group (shCYP7A1 group) were respectively injected via tail vein into C57BL / 6 mice with AAV2 / 9-hU6-shRNA(NC) (1×10⁻⁶). 11 vg / mouse), AAV2 / 9-hU6-shRNA(CYP7A1) (1×10 11 (vg / mouse), of which AAV was synthesized and identified by Shanghai Quanyang Biotechnology. A schematic diagram of the model intervention is shown below. Figure 7A As shown in the figure. Mice were sacrificed after 8 weeks to verify the effect of AAV, and the results are as follows. Figure 7B and Figure 7C As shown in the figure, the number of liver nodules was reduced in the CYP7A1 knockdown group (23.50 in the control group vs. 11.00 in the CYP7A1 knockdown group, p<0.001).

[0146] Patient-derived xenograft subcutaneous tumor model: Tissue from liver cancer patients is cut into 1-3mm pieces. 3 Small tissue fragments were extracted and transplanted into the subcutaneous tissue of 6-week-old male nude mice. When the tumor grew to 200 mm... 3 At the tumor volume level, an in vivo liver cancer inhibition experiment was conducted by intratumoral injection of in vivo siRNA (synthesized and identified by Shanghai Quanyang Biotechnology) to knock down CYP7A1. Mice were injected with siRNA (5 nmol / mouse) every 3 days for a total of 7 times. A schematic diagram of the model intervention is shown below. Figure 8A As shown. The results are as follows. Figure 8B and Figure 8C As shown in the figure, after a total of 21 days of CYP7A1 siRNA silencing, the subcutaneous tumors in the CYP7A1 knockdown group (siCYP7A1 group) PDX model were significantly reduced by 68.3% compared with the control group (siNC group), indicating that the tumor volume inhibition rate reached 68.3% after 21 days of intratumoral injection of siCYP7A1.

[0147] Example 3: Functional verification of the CYP7A1 R260L mutant

[0148] 1. Enzyme activity detection

[0149] According to reports by Qayyum, Tempel, and others, arginine 260 (R260) is a key residue in the heme-binding domain of CYP7A1, and its mutation to leucine (L) can cause it to lose its ability to bind to substrates and catalyze. The mutant plasmids pTSBX-CMV-MCS-EF1-CYP7A1-WT (Shanghai Quanyang Biotechnology Co., Ltd., catalog number: TSB304132-1) and pTSBX-CMV-MCS-EF1-CYP7A1-R260L (synthesized by Shanghai Quanyang Biotechnology Co., Ltd., catalog number: TSB304132-2) were constructed, and the control group (con) was transfected with the empty control plasmid. Figure 9 This shows the location of the mutation within the entire length of CYP7A1. Figure 10 This is a schematic diagram illustrating the changes in the catalytic substrate. The above plasmids were transfected into HepG2-shCYP7A1 cells (cells obtained in Example 2) using the Polyplus jetPRIME transfection kit (Polyplus, catalog number: PT-114-15). Cells were collected 48 hours after transfection, and cholesterol levels were detected using a cholesterol assay kit (Polyplus Gene, catalog number: E1015). The results are as follows... Figure 11 As shown, the cholesterol level in the CYP7A1-WT group was 9.56 μmol / L per 1×10⁻⁶ μmol / L. 6 Cells, CYP7A1-R260L group, 136.1 μmol / L per 1×10 6 The results showed that the mutant R260L significantly affected cholesterol metabolism (p<0.001) in a few cells.

[0150] 2. Proliferation capacity detection

[0151] Cells from Example 2 (2. Verification of Gene Knockdown Effect and In Vivo / In Vitro Experiments) were seeded in 96-well plates. When the cell density reached 60%-70%, transfection was performed using the Polyplus jetPRIME transfection kit according to its instructions. The experimental groups were transfected with either the WT plasmid or the R260L plasmid, while the control group (con) was transfected with the empty control plasmid. An equal amount of plasmid (final concentration 0.1 μg / well) and corresponding transfection reagent were added to each well. After 0 h (Day 0), 24 h (Day 1), 48 h (Day 2), and 72 h (Day 3) of transfection, 10 μL of CCK-8 reagent (MCE, catalog number: HY-K0301, a rapid and highly sensitive kit based on WST-8 widely used for cell viability and cytotoxicity detection) was added to each well. The plates were incubated at 37°C for 2 h, and the absorbance (OD) of each well was measured at 450 nm using a microplate reader. 450Each experiment was performed in triplicate, with three replicates per replicate. The proliferation rate was calculated using the formula in the instructions: Proliferation Rate (%) = (OD value of experimental group - OD value of blank well) / (OD value of Day 0 in experimental group - OD value of blank well in Day 0) × 100%. Results are as follows: Figure 12 As shown, the cell proliferation rates of the WT group and the R260L group after 48 hours were 45.3% and 43.7% respectively (p=0.32). There was no significant difference in cell proliferation rate between the two groups, but both were higher than the control group (21.6%, p<0.001), indicating that the cancer-promoting effect of CYP7A1 is independent of its enzyme activity.

[0152] Example 4: Elucidation of the autophagy-lysosome mechanism

[0153] 1. Detection of autophagy proteins LC3 and p62, observation of autophagy flux, and assessment of autophagy status using electron microscopy.

[0154] During autophagy, cytoplasmic LC3B-I undergoes lipidation to form membrane-bound LC3-II, which is localized on the autophagosome membrane. Increased LC3-II levels typically reflect an increase in the number of autophagosomes. On the other hand, p62 (also known as SQSTM1) is an autophagy substrate receptor protein that binds to ubiquitinated substrates and is degraded via the autophagy pathway through interaction with LC3. When the autophagy pathway is active, p62 is continuously degraded, and its level decreases; if autophagy flux is blocked or autophagy is overactivated, p62 accumulates. Therefore, by simultaneously detecting changes in the LC3-II / I ratio and p62 levels using Western blot, a comprehensive assessment of autophagy activity and the patency of autophagy flux can be made. Figure 13A The results of differential gene function enrichment are displayed, where the vertical axis represents the enriched pathway or functional category, the horizontal axis represents the significance level (or enrichment score), the bubble size represents the number of genes involved in the pathway, and the bubble color represents the significance of enrichment. Figure 13B A heatmap showing differentially expressed gene levels across samples is presented, with colors ranging from blue to red representing increasing expression levels. The multi-omics analysis revealed that CYP7A1 regulates the activation and increase of autophagy-related pathways, indicating that CYP7A1 positively regulates autophagy flux.

[0155] Cells were collected from stable cell lines in the control group (shNC group) and the CYP7A1 knockdown group (shCYP7A1 group), respectively. Expression levels were detected using LC3 (Abmart, catalog number: T55992) and p62 (Abmart, catalog number: T55546) antibodies, following the method described in Example 2 (1. Cell Model Construction). The results are as follows: Figure 14As shown, compared to the control group, the LC3-II / I ratio increased 2.8-fold (p<0.01) and the autophagy substrate p62 increased 3.1-fold (p=0.003) in the CYP7A1 knockdown group. However, after administering 50 μM chloroquine (shCYP7A1+CQ group) as a late autophagy inhibitor on top of CYP7A1 knockdown, subsequent testing 12 hours later revealed no further increase in LC3-II and p62, indicating that CYP7A1 knockdown may block late-stage autophagy. Autophagic flux was then measured to further assess the autophagic state.

[0156] The mRFP-GFP-LC3 dual-label adenovirus fluorescence system (Hanheng Biotechnology, catalog number: HB-AP2100001) allows adenovirus infection of target cells to express LC3, which exhibits both red and green fluorescence. mRFP is used to label and track LC3. Attenuation of GFP indicates the fusion of lysosomes and autophagosomes to form autolysosomes. When autophagosomes fuse with lysosomes, GFP fluorescence is quenched, and only red fluorescence is detected. After microscopic imaging, the yellow spots appearing after merging the red and green fluorescence are simply autophagosomes, while the red spots indicate autolysosomes. Counting the different colored spots clearly shows the intensity of the autophagic flux.

[0157] Stable shNC and shCYP7A1 cells were infected with adenovirus overexpressing mRFP-GFP-LC3. The medium was changed 4-6 hours after infection, and autophagosomes and autolysosomes were observed using confocal microscopy 48 hours later. After acquiring images using confocal microscopy, channels, thresholds, and counts were performed in ImageJ. The autophagosome-lysosome fusion rate per cell was defined as the number of individual red LC3 particles (GFP). - RFP + ) as a percentage of total LC3 particles (GFP) + RFP + +GFP - RFP + The proportion of ). The results are as follows Figure 15A and Figure 15B As shown, the mRFP-GFP-LC3 dual fluorescence system showed a significant decrease in autophagosome-lysosome fusion rate (control group: 61.67±7.57% vs knockdown group: 29.33±4.51%, p=0.003).

[0158] Electron microscopy observation: After culturing shNC and shCYP7A1 cells stably transformed, the culture medium was discarded, and trypsin was added for digestion. After digestion, culture medium was added to stop the digestion, and the cells were gently pipetted until they floated. The cells were then aspirated into centrifuge tubes and centrifuged at 3000 rpm for 3 minutes, resulting in cell clumps the size of mung beans. The supernatant was discarded, and electron microscopy fixative (2.5% room temperature glutaraldehyde fixative) was slowly added. The cells were then subjected to room temperature dehydration, infiltration embedding, polymerization, sectioning, and staining, followed by observation under a transmission electron microscope (HITACHI, catalog number: HT7800). The results are as follows: Figure 16 As shown in the figure, red arrows represent autolysosomes, and yellow arrows represent autophagosomes. The figure shows that compared to the shNC group, the shCYP7A1 group had a reduced number of autolysosomes (<20% vs. control >60%), and the knockdown group exhibited lysosomal structural rupture.

[0159] 2. Lysosomal acidification

[0160] A small amount of Lyso-Tracker Red (Beyotime, catalog number: C1046) was added to the cell culture medium at a ratio of 1:5000-1:20000. The cell culture medium was removed, and the prepared Lyso-Tracker Red staining working solution (pre-warmed to 37°C) was added. The cells were then incubated at 37°C for 60 minutes. The Lyso-Tracker Red staining working solution was removed, and fresh cell culture medium was added. The cells were then observed using a laser confocal microscope (Zeiss LSM 900 Airyscan2) at an excitation wavelength of 594 nm and an emission wavelength of 617 nm. The results are as follows: Figure 17 As shown, the fluorescence intensity of lysosomes in the CYP7A1 knockdown group was significantly reduced.

[0161] 3. To verify whether the autophagy inhibitor chloroquine (CQ) could inhibit the in vitro proliferation of CYP7A1-overexpressing cells, the CCK-8 assay was used to quantitatively measure changes in cell viability.

[0162] Experimental materials: CYP7A1-FLAG plasmid and empty vector plasmid that does not express FLAG (synthesized and identified by Shanghai Quanyang Biotechnology Co., Ltd., catalog number: TSB4006145-1), chloroquine (MCE, catalog number: HHY-17589) were prepared into a 10mM stock solution using PBS.

[0163] Experimental Procedure: Hep3B cells were cultured to the logarithmic growth phase, digested, counted, and seeded into 96-well plates at a density of 3000 cells per well, with 100 μL of complete culture medium added to each well. After cell adhesion, the cells were transfected using the transfection method described in Example 3 (1. Enzyme Activity Assay) with the above-mentioned plasmids (CYP7A1 overexpression group, CYP7A1 overexpression group combined with chloroquine (50 μM) (CYP7A1+CQ group), and the control group with an equal volume of empty vector and isosolvable vector). Each group contained 3 biological replicates (independent culture / transfection). Chloroquine was added within 36 h after transfection and incubation continued. 12 h later, 10 μL of CCK-8 solution was added to each well. The 96-well plates were incubated at 37°C for 2 h, and the absorbance (OD) of each well was measured at 450 nm using a microplate reader. 450 Meanwhile, blank wells (containing only culture medium and CCK-8, without cells) were set up to remove the background.

[0164] Cell proliferation rate calculation (%) = (OD value of experimental group - OD value of blank well) / (OD value of control group - OD value of blank well) × 100%.

[0165] The results are as follows Figure 18 As shown in the CCK-8 assay, CYP7A1 overexpression significantly promoted cell proliferation. However, treatment with 50 μM chloroquine significantly reduced the proliferation rate of CYP7A1-overexpressing cells, with no statistically significant difference compared to the vector control group under the same conditions. This suggests that the autophagy inhibitor chloroquine can reverse the CYP7A1-mediated cell proliferation promotion effect. These results indicate that the autophagy inhibitor chloroquine (50 μM) can inhibit the proliferation of CYP7A1-overexpressing cells, further confirming that autophagy promotion is a key downstream event in its cancer-promoting effects.

[0166] Example 5: Verification of CYP7A1-p62 protein interaction

[0167] 1. Use immunoprecipitation to verify whether there is a direct or indirect interaction between CYP7A1 and p62 proteins in cells.

[0168] Experimental materials: CYP7A1-FLAG plasmid without FLAG expression (Example 4, 3. Verification of whether the autophagy inhibitor chloroquine can inhibit the in vitro proliferation of CYP7A1 overexpressing cells, quantitative determination of cell viability using CCK-8 assay), HepG2 hepatocytes, transfection reagent (Example 3, 1. Enzyme activity detection), IP lysis buffer (Beyotime, catalog number: P0013), Flag-beads (MCE, catalog number: HY-K0207), p62 antibody (Abmart, catalog number: T55546).

[0169] Experimental Procedure: HepG2 cells were cultured in serum-containing medium to 70%-80% confluence. Cells were transfected with the plasmid described in Example 3 (1. Enzyme Activity Assay) using the transfection method described above (control group: empty vector without FLAG tag; experimental group: CYP7A1-FLAG plasmid). After 48 hours, cell lysis and protein extraction were performed. Cells were collected, washed with pre-chilled PBS, and lysed with IP lysis buffer (containing protease / phosphatase inhibitors) for 60 minutes on ice. The cells were centrifuged (12000 rpm, 4°C, 15 minutes) and the supernatant (total protein extract) was collected. An appropriate amount of total protein was added to 10 μL of Flag beads and gently incubated at 4°C for 2-4 hours. The beads were washed three times with lysis buffer to remove non-specific binding. 1×SDS loading buffer was added, and the mixture was heated at 95°C for 10 minutes to elute the bound proteins. The eluted sample was separated by 10% SDS-PAGE electrophoresis. Transfer the sample to a PVDF membrane and block with 5% skim milk at room temperature for 1 hour. Incubate overnight at 4°C with anti-p62 antibody while verifying the input volume (Input group). Incubate with secondary antibody (HRP labeling) at room temperature for 1 hour, then develop ECL.

[0170] The results are as follows Figure 19 As shown in the figure, p62 protein was detected in the IP:FLAG experimental group while no band was observed in the control group (con), confirming that the two have direct or indirect intracellular interactions.

[0171] 3. Verify whether there is a direct binding interaction between the CYP7A1 protein and the autophagy receptor protein p62.

[0172] Using GST pull-down technology, the p62 fusion protein labeled with GST was immobilized on glutathione agarose beads to capture exogenously added recombinant CYP7A1 protein.

[0173] Experimental materials: GST-p62 plasmid (purchased from Shanghai Quanyang Biotechnology Co., Ltd., catalog number: TSB5001138-1), Escherichia coli BL21 (DE3) competent cells, recombinant CYP7A1 protein (MCE, catalog number: HY-P702798), GST-tagged protein purification kit (Beyotime, catalog number: P2260S), IPTG (Sangon Biotech, catalog number: A100487-0005).

[0174] Experimental Procedure: The GST-p62 plasmid was transformed into BL21(DE3), and expression was induced by IPTG. The bacterial cells were lysed according to the GST protein purification reagent instructions, and the immobilized GST-p62 protein was purified using GSH agarose beads. The immobilized GST-p62 beads and recombinant CYP7A1 protein (10 μg) were incubated in binding buffer at 4°C with slow rotation for 3 h. An empty GST bead + CYP7A1 protein was used as a negative control.

[0175] After incubation, the sample was washed five times with binding buffer to remove non-specific bindings. Bound proteins were eluted with SDS loading buffer. The incubation system (Input) and elution buffer (Eluate) were separated by SDS-PAGE and detected by Western blot. Results are as follows: Figure 20 As shown in the figure, the CYP7A1 band can be detected in the GST-p62 bead eluent, while no obvious signal is found in the empty GST control; this confirms the direct interaction between the two.

[0176] The above co-immunoprecipitation (Co-IP) and in vitro pull-down experiments confirmed that CYP7A1 directly interacts with the autophagy adaptor protein p62.

[0177] Example 6: Mechanism analysis of CYP7A1 and immune regulation

[0178] To clarify the correlation between the expression level of cholesterol 7α-hydroxylase (CYP7A1) in hepatocellular carcinoma tissue and the degree of CD8+ T cell infiltration, to reveal the potential molecular mechanism by which CYP7A1 participates in the immune regulation of hepatocellular carcinoma, and to provide experimental evidence for subsequent immunotherapy strategies targeting CYP7A1 in hepatocellular carcinoma, the following experiments were conducted.

[0179] Experimental Procedure: Paraffin-embedded blocks of DEN+CCl4 model liver cancer tissue from Example 2 were used. The paraffin-embedded blocks were cut to obtain paraffin sections. After antigen retrieval, the paraffin sections were incubated overnight at 4°C with CD8 protein primary antibody (purchased from CST, catalog number 98941, 1:200 dilution), and observed under a microscope after DAB staining.

[0180] Results Analysis: Three high-power fields (×400) were selected from each section, and the number of CD8+ T cell-positive cells was counted (positive cells were determined by the appearance of brownish-yellow staining on the cell membrane or cytoplasm). The results of CYP7A1 expression and CD8+ T cell infiltration are as follows: Figure 21A and Figure 21BAs shown in the results above, the number of CD8+ T cells infiltrating in the AAV-shCYP7A1 group was (24.40±3.78) cells / high-power field, significantly higher than that in the control group (14.66±2.68) cells / high-power field, with a highly statistically significant difference (P<0.001). These results indicate that high expression of CYP7A1 in liver cancer tissue is significantly correlated with reduced CD8+ T cell infiltration, suggesting that CYP7A1 may participate in the immunosuppressive regulation of liver cancer by inhibiting CD8+ T cell infiltration into the tumor microenvironment.

[0181] Example 7: Optimization of Virtual Screening Method Based on Schrödinger Maestro Platform

[0182] 1. Energy analysis of receptor-ligand interactions

[0183] The molecular covalent bonding score (MCCS) algorithm was used to perform energy decomposition on the PDB 3V8D cocrystal complex, quantifying the contribution of key forces (e.g., Figure 22As shown in the diagram: First, His101 exhibits the largest energy contribution (-1.27 kcal / mol), indicating that it may participate in stabilizing the ligand conformation through its aromatic side chain or π-π stacking during ligand binding. Following closely is Trp284 (-1.22 kcal / mol), an aromatic residue with a strong π-electron cloud, which can form a stable π-π stack with the aromatic ring system of the ligand, effectively enhancing the binding stability and specificity of the ligand. Furthermore, Val281 (-1.13 kcal / mol), as a hydrophobic residue, embeds the hydrophobic group of the ligand through hydrophobic interactions, reducing solvent accessibility within the binding cavity and providing an important hydrophobic driving force for the system's binding. Ile125 (-0.74 kcal / mol) and Ile114 (-0.55 kcal / mol), two hydrophobic branched amino acids, further enhance the ligand's embedding depth and binding stability by forming a tight contact with the ligand through van der Waals forces. Aromatic residues Phe129 (-0.58 kcal / mol) and Phe102 (-0.36 kcal / mol) provide additional aromatic stacking facets within the binding pocket, potentially forming stable stacking and steric repulsion with the aromatic rings or polar segments of the ligand, further assisting the ligand in positioning to the optimal binding configuration. Several leucine residues, including Leu486 (-0.49 kcal / mol), Leu104 (-0.46 kcal / mol), and Leu361 (-0.45 kcal / mol), also enhance the overall stability of the ligand-receptor complex through their hydrophobic side chains. Furthermore, Ser105 (-0.45 kcal / mol) and Gly485 (-0.59 kcal / mol) each play a role in polar interactions. Overall, the ligand binding pocket of CYP7A1 exhibits typical hydrophobic-dominant binding characteristics, while aromatic stacking and polar hydrogen bonds form a complex and stable network, providing a favorable binding environment and a basis for high binding affinity for the ligands. Figure 23 The main interaction between the ligand (purple) and the receptor is shown in the CYP7A1 crystal configuration (PDB ID: 3V8D).

[0184] 2. Protein structure preprocessing and compound library construction

[0185] In protein preparation, the Protein PreparationWizard module in the Schrödinger software was used. First, the Prime module was used to automatically reconstruct and optimize the missing side chains in the protein structure, primarily involving the residues Glu276 to Ala289. The Prime module utilizes its built-in template database and optimization algorithm to rationally fill in the missing regions based on spatial conformation, ensuring the rationality of the side chain conformation and its spatial adaptation with surrounding residues. After side chain completion, an OPLS3 force field was applied to the entire protein to eliminate unreasonable conformations and stresses that might have been introduced during model reconstruction and hydrogen addition. During energy minimization, an RMSD convergence threshold of 0.25 Å was set to ensure that structure optimization was completed within a reasonable range while avoiding large-scale main chain drift. The entire energy minimization process was performed under default aqueous medium conditions, employing a position-restricted optimization strategy, optimizing only the side chains and hydrogen atoms to maintain the stability of the crystal main chain conformation.

[0186] Prior to virtual screening, 9,204,306 drug-like small molecules were first screened from the ZINC15 database. The screening criteria were a molecular weight of 250 Da-500 Da and a lipid-water partition coefficient (logP) between -1 and 5. Subsequently, the QikProp module in Schrödinger software was used to perform pharmacokinetic and toxicity prediction analyses on the molecular library obtained from the initial screening. Compounds with an hERG inhibition tendency score greater than 0.5 were further excluded to reduce the potential risk of cardiotoxicity. The final retained molecular library was used for subsequent pharmacophore screening and molecular docking analysis.

[0187] 3. Intelligent filtration with pharmacophores

[0188] Based on the co-crystal ligand structures of 3SN5 and 3V8D in the PDB database, five representative key pharmacophore features were identified and extracted through superposition analysis (RMSD = 0.39 Å after ligand alignment) for subsequent activity screening and virtual screening model construction. Figure 24 The pharmacophores are shown as the structures formed by the superposition of 3V8D and 3SN5 ligands, with purple representing hydrogendon / Acc (hydrogen bond donor and acceptor) and the other four green ones representing Hyd / Arm (hydrophobic or cyclic). The first round of filtration retained compounds that met at least three pharmacophore characteristics, reducing the library capacity from 9,204,306 to 800,000.

[0189] Step 4: Molecular docking and ligand selection

[0190] In this molecular docking screening, the candidate ligands obtained all exhibited good binding adaptability at the CYP7A1 active site, with Glide scores exceeding -7.5 kcal / mol, indicating that the ligands possess advantages in conformational matching and initial binding energy. Most ligands formed at least one hydrogen bond within the binding pocket, primarily interacting with polar residues such as Leu361 and Gly485, playing a supporting stabilizing role. Simultaneously, the hydrophobic functional groups of the ligands were fully embedded in the hydrophobic cavity composed of residues such as Val281, Ile125, His101, Phe129, and Ile114, forming a stable hydrophobic embedding effect, providing the main driving force for overall binding. Overall, the candidate ligand binding mode is dominated by hydrophobic interactions, with hydrogen bonds enhancing stability, consistent with the target pocket characteristics, providing a sound molecular basis for subsequent optimization and experimental verification.

[0191] Example 8: Optimization of Virtual Screening Method Based on MolProphet™ Platform

[0192] 1. Definition of binding site

[0193] The CYP7A1 PDB ID: 3V8D eutectic structure was used as a virtual structural screening model. First, the binding pockets were automatically predicted using the AI ​​algorithm of MolProphet software. Then, the key binding sites of the CYP7A1 eutectic structure were described according to the literature (Tempel, W., Grabovec, I., MacKenzie, F., Dichenko, YV, Usanov, SA, Gilep, AA, Park, HW, & Strushkevich, N. (2014). Structural characterization of human cholesterol 7α-hydroxylase. Journal of lipid research, 55(9), 1925–1932. https: / / doi.org / 10.1194 / jlr.M050765). The software defines identical binding sites and interaction modes (defining the interaction mode of amino acid residue Trp284 with small molecule compounds as hydrophobic interaction, the interaction mode of amino acid residue Gly485 with small molecule compounds as hydrogen bonding, and the interaction mode of amino acid residue Ser360 with small molecule compounds as hydrogen bonding). When selecting molecules from the virtual screening results, not only the AI ​​activity prediction score of the virtual screening is considered, but also the binding conformation of the ligand and pocket. The results are automatically sorted after the interaction mode is set.

[0194] 2. Construction of Active Learning Models

[0195] To efficiently and accurately score compounds, an active learning algorithm was employed in the software. The model used the PDBbindv2020 public dataset as both the training and testing set, containing 19,443 protein-small molecule compound pairs. The entire protein was divided into segments with a radius of 20 Å. Simultaneously, the 2D small molecule structure was used to generate an initial 3D crystal structure (3D Conformer) via the RDKit tool. The protein segments and small molecule compounds were processed together using a graph neural network (GNN) to obtain the protein segment graph mapping features hp and the small molecule compound graph mapping features hc. The protein segment-small molecule compound interaction matrix z was obtained from the protein segment mapping features and the small molecule mapping features. Node updates were performed using a geometry-based message passing update function, and the parameters of the L-layered model were updated using a contrast function. The final protein-small molecule compound affinity prediction was obtained by linearly weighting the output of the last layer of the L-layered model after passing through a linear layer. Because the entire protein is divided into multiple protein blocks, the predicted binding affinity of the ligand to the entire protein is equal to the predicted binding affinity of the protein block with which the ligand binds most strongly. Thanks to the contrastive affinity loss function, this model is able to predict the binding affinity between proteins and small molecule compounds.

[0196] 3. Large-scale library screening

[0197] Based on the construction of the above active learning model, the Enamine real space database is selected in the Structure Based VirtualScreening function of the MolProphet AI platform. The software automatically completes AI docking of compounds and active pockets in the Enamine Real Space database, and selects reliable molecules by comparing affinity and various physicochemical property scores.

[0198] Based on the above analysis methods, more molecular structures with skeletons were selected through AI molecular virtual screening, 2D and 3D structural similarity algorithm screening, etc. Based on the physicochemical property parameters of the compounds, including LogP, TPSA, NRot, HBD and HBA, a total of 36 candidate molecules were selected after final observation and analysis.

[0199] In the first round of screening (Schrödinger Maestro software identified 36 compounds + MolProphet active learning identified 36 compounds), CA4 was identified as the lead compound after in vitro activity verification. Using CA4 as the parent compound, 28 similar derivatives were further obtained from the Enamine library using 2D / 3D joint similarity search (Tanimoto coefficient ≥ 0.6).

[0200] Example 9: Validation of the in vitro anti-hepatocellular carcinoma activity of candidate compounds

[0201] 1. Cell viability assay

[0202] By adding different concentrations of drugs to cultured cells, CCK8 assays were used to detect changes in cell activity, screen for drugs with potential therapeutic effects, and simultaneously assess the safety and toxicity of the drugs.

[0203] Experimental materials: CCK8 reagent, HepG2 cells, Huh7 cells, and DMEM medium containing 10% FBS.

[0204] Experimental content: IC50 of all compounds 50 The following standard procedures were used for all measurements: (1) Take a cell culture dish that has been cultured for 2-3 days in the exponential growth phase, add an appropriate amount of Trypsin-EDTA solution to digest the cells, digest at room temperature to allow the adherent cells to detach, add an appropriate amount of DMEM culture medium containing 10% fetal bovine serum to stop the digestion, centrifuge at 600 rpm at room temperature for 3 min, discard the supernatant and resuspend the cells in culture medium. (2) Stain with trypan blue and count the cells on a cell counter. (3) Dilute the cell suspension with cell culture medium to prepare a solution containing 10,000 cells per 100 mL. (4) Take a 96-well plate and add 100 μL of cell suspension to each well. Place the plate in a 37℃ CO2 (5%) incubator for 24 hours. (5) Serially dilute the compound (0-100 μM) and treat the cells with a final DMSO concentration <0.1%. (6) Incubate the plate in a 37℃ CO2 (5%) incubator for 48 hours. (7) Prepare 10% CCK8 solution using serum-free DMEM culture medium, add 100 μL to each well, and incubate at 37°C for 2 hours. (8) Measure the absorbance of each well at 450 nm using a microplate reader. (9) Use GraphPad Prism 9.0 to perform four-parameter fitting, plot the cell viability curve, and calculate the IC50. 50 value.

[0205] Experimental results: IC50 of compound 2 50 =24.38 μM, IC50 of compound 8 50 =51.14 μM, IC50 of compound 11 50 =53.47 μM, IC50 of compound 16 50=70.16 μM, IC50 of compound 22 50 =13.59 μM, IC50 of compound 30 50 =24.55 μM, IC50 of compound 34 50 =10.48 μM, IC50 of compound 35 50 =86.72 μM, IC50 of compound 54 50 =169.1 μM IC50 of compound 56 50 =9.2 μM, IC50 of compound 58 50 =0.00648 μM, IC50 of compound 59 50 =22.07 μM, IC50 of compound 60 50 =10.36 μM, IC50 of compound 61 50 =28.39 μM, IC50 of compound 64 50 =18.73 μM, IC50 of compound 65 50 =10.88 μM, IC50 of compound 66 50 =72.95 μM, IC50 of compound 71 50 =21.53 μM, IC50 of compound 72 50 =24.59 μM, IC50 of compound 74 50 =24.5 μM, IC50 of compound 75 50 =11.85 μM, IC50 of compound 81 50 =19.13 μM, IC50 of compound 87 50 =1.45 μM, IC50 of compound 90 50 =0.3178 μM, IC50 of compound 93 50 =9.78 μM, IC50 of compound 98 50 =48 μM, IC50 of compound 100 50 =0.6158 μM. Specific information for each compound is shown in Table 1.

[0206] Table 1

[0207]

[0208] Among them, the cis-gutter of compound 58 (i.e., lead compound CA4) showed the strongest activity against Huh7 cells (IC50). 50 =6.48nM (e.g.) Figure 25 As shown), and with low toxicity to primary hepatocytes (IC50). 50 =10.73μM), selectivity index (SI) >100 (e.g. Figure 26 (As shown).

[0209] 2. Cell cloning experiments

[0210] The proliferative capacity of cells after treatment can be indicated by observing the formation of clones on cell culture plates after treating cells with small molecule compounds.

[0211] Experimental materials: CA4 (Taoshu Biotechnology, catalog number: T6212), DMEM medium containing 10% FBS, trypsin, PBS, penicillin-streptomycin solution (100×), paraformaldehyde fixation, crystal violet staining solution.

[0212] Experimental content: (1) After trypsin digestion of HepG2 cells in the logarithmic growth phase, resuspend them in complete culture medium (basal culture medium + 10% fetal bovine serum) to form a cell suspension and count them. (2) Cell seeding: 700 cells / well were seeded in each experimental group in a 6-well plate. (3) Continue to culture for 14 days, changing the medium every 3 days and observing the cell status. (4) After cloning, take pictures of the cells under a microscope, then wash once with PBS, add 1 mL of 4% paraformaldehyde to each well for fixation for 60 min, and wash once with PBS. (5) Add 1 mL of crystal violet staining solution to each well and stain the cells for 10 min. (6) Wash the cells 3 times with PBS, air dry, and take pictures with a digital camera (take pictures of the entire 6-well plate and each well separately). (7) Observe positive clones under a microscope, i.e., each clone >50 cells, take pictures with a camera, observe directly with the naked eye, and count the number of clones one by one. (The size is about 0.3-1.0 mm).

[0213] Experimental results: such as Figure 27 As shown in the figure, the number of cell clones was significantly reduced after CA4 treatment (control group DMSO 110.25±17.63 VS CA4 group 44.75±23.41, p=0.005).

[0214] 3. Cell scratch assay

[0215] By measuring the scratch spacing at different time points and calculating the difference, the effect of small molecule compounds on cell migration ability can be determined.

[0216] Experimental content: (1) Prepare a 6-well plate with back markings in advance. Use a ruler to mark lines evenly on the back, with each line spaced 0.5cm-1cm apart. Each well should have at least 3-5 lines. (2) Digest Hep3B cells in the logarithmic growth phase into a single-cell suspension using trypsin and inoculate them into a culture plate overnight to allow them to adhere (5×10⁻⁶ cells). 5(3) Cell culture: Cultured in a 37℃, 5% CO2 incubator for 24h. (4) Scratching: Using a pipette tip as a guide, make a horizontal scratch with the pipette tip perpendicular to the back of the cell. Wash the scratched cells 3 times with 1×PBS, 1mL / well, and then add serum-free medium. (5) Photographing: Cultured in an incubator for 48h and then the cells were taken out and photographed. (6) Result analysis: The results were analyzed using ImageJ software.

[0217] Experimental results: such as Figure 28 As shown in the figure, the cell healing rate decreased after CA4 treatment (34.13% in the control group vs. 3.3% in the CA4 group, p=0.003).

[0218] Colony formation and scratch assays further confirmed that CA4 significantly inhibits the proliferation and migration of liver cancer cells.

[0219] 4. Validation of the in vivo antitumor activity and safety of compound CA4

[0220] Experimental model: Nude mouse subcutaneous tumor model: Liver cancer patient tissue was cut into 1-3mm pieces. 3 Small tissue fragments were extracted and transplanted into the subcutaneous tissue of 6-week-old male nude mice. When the tumor grew to 200 mm... 3 When the volume is reached, the drug is administered.

[0221] Experimental materials: 4-week-old male BALB / c nude mice, CA4P (Fosbretabulin, MCE, catalog number HY-13226, CA4 phosphorylation prodrug), lenvatinib (Selleck, catalog number: S1164), alanine aminotransferase (ALT) assay kit (Prepulse Biotech, catalog number: ALT01), aspartate aminotransferase (AST) assay kit (Prepulse Biotech, catalog number: AST01), and serum creatinine (CR) assay kit (Prepulse Biotech, catalog number: G034).

[0222] Experimental groups: blank control group (solvent PBS), low-dose CA4P group (20 mg / kg), high-dose CA4P group (40 mg / kg), and lenvatinib group (4 mg / kg).

[0223] Experimental procedure: The experimental group received intravenous injection of the drug once daily via tail vein; the control group received an equal volume of blank solvent; and the lenvatinib group received oral gavage once daily. Figure 29ADuring the drug administration period, the mental state, activity level, weight, diet, and fur luster of nude mice were observed daily, and any abnormalities were recorded. Every two days, the long axis (a) and short axis (b) of the tumor in the nude mice were measured using calipers, and their weight was assessed to evaluate their survival status. The tumor volume was calculated using the formula V=1 / 2×a×b², and a tumor growth curve was plotted to compare the tumor growth inhibition of each group. After 28 days of drug administration, the nude mice were sacrificed, and the subcutaneous tumor volume was measured and the tumor inhibition rate was calculated. Blood was collected, centrifuged at 4000 rpm for 10 min, and serum was separated. 15 μL of serum biochemical indicators were tested. Heart, liver, spleen, lung, kidney, and brain organs were collected, fixed with paraformaldehyde, embedded in paraffin, sectioned, and stained with hematoxylin and eosin (HE) to observe tissue morphological changes and assess the toxicity of the drug to the organs.

[0224] Experimental results: Figure 29B and Figure 29C The growth curves of subcutaneous tumors in nude mice after treatment are shown. At sacrifice, the tumor size in the high-dose CA4P group was 20.21 ± 33.41 mm. 3 The tumor size was significantly smaller than that of the control group (703.00±211.07 mm). 3 (p < 0.0001), there was no significant difference in body weight between nude mice and the control group during the treatment period. There was no significant difference in the inhibition of subcutaneous tumor growth in nude mice by tail vein injection of CA4P and gavage with lenvatinib. These results indicate that the inhibitory effect of CA4P is comparable to that of lenvatinib. There were no significant changes in body weight among the groups; HE staining results of the heart, liver, spleen, lungs, and kidneys of nude mice (…). Figure 30A The results showed that the mice's major organs had almost no physiological morphological abnormalities or histopathological damage. Figure 30B The graph shows the results of serum biochemical indicators. It can be seen that there were no significant abnormalities in serum ALT, AST, and creatinine levels.

[0225] Example 10: In vitro and in vivo validation of CA4 targeting CYP7A1

[0226] I. Interaction between compound CA4 and recombinant CYP7A1 protein

[0227] The interaction between compound CA4 and recombinant CYP7A1 protein was investigated using cell thermal migration assays and surface plasmon resonance (SPR) technology.

[0228] Experimental instrument: Biacore T200 instrument (GE Healthcare, USA).

[0229] 1. Determine the direct binding of compound CA4 to the target protein CYP7A1 at the cellular level.

[0230] HepG2 cells (2×10) 5Cells were seeded (number per well) into 6-well plates. After cell adhesion, CA4 (1 μM) was added and co-cultured for 3 h. The control group received an equal volume of DMSO. After trypsin digestion, cells were centrifuged, the supernatant was discarded, and the cells were resuspended in PBS containing 1% protease inhibitor. After mixing, the cells were evenly distributed into PCR tubes. The PCR instrument was set to six different temperatures (37℃-65℃) for 3 min each, followed by incubation at room temperature for 3 min. After centrifugation at 3000 rpm, the supernatant was discarded, and an equal volume of RIPA lysis buffer was added to each tube for sonication lysis. Western blot (WB) samples were prepared and analyzed. Finally, image J was used for grayscale analysis. The experimental results are as follows: Figure 31A and Figure 31B As shown in the figure, CA4 at a low concentration (1 μM) can inhibit the temperature-induced stability disruption of CYP7A1 protein, indicating that CA4 has a direct binding effect on CYP7A1 at the cellular level.

[0231] 2. Determine the affinity between compound CA4 and the target protein CYP7A1.

[0232] (1) Protein fixation: 15 μg / mL of purified recombinant CYP7A1 protein was dissolved in PBST buffer (pH 7.2) and fixed on the surface of the CM-5 sensor chip at a flow rate of 10 μL / s for 600 seconds at 25°C.

[0233] (2) Compound injection: Different concentrations of CA4 (0.25 μM-4 μM) were injected into the chip surface at a flow rate of 20 μL / s, with PBST as the buffer. Signal recording: The response signal (RU) generated by the interaction between CA4 and CYP7A1 was monitored in real time.

[0234] (3) Data analysis: Analyze the binding and dissociation stages; calculate the equilibrium dissociation constant (KD) according to the Langmuir binding model, with the formula KD=Koff / Kon; the binding affinity (Ka) is obtained by the reciprocal of KD (Ka=1 / KD) to quantify the binding strength between CA4 and CYP7A1.

[0235] Experimental results are as follows Figure 32 As shown (different colors represent the binding-dissociation fitting curves when CA4 of different concentrations flows through), the measured KD=1.0924μM indicates the specific binding characteristics of CA4 with CYP7A1, providing direct evidence for verifying targeting.

[0236] II. Verify whether compound CA4 inhibits the proliferation and progression of liver cancer cells by directly affecting autolysosome formation through its action on CYP7A1 protein.

[0237] Experimental materials: HepG2 cells, CA4.

[0238] Experimental Procedure: a) Western blot analysis of the LC3-II / I ratio: Hepatocellular carcinoma cells were divided into a control group (treated with DMSO solvent) and a CA4 treatment group (treated with a final concentration of 3 nM for 24 h). Cells were collected and total protein was extracted using RIPA lysis buffer. After quantification using the BCA method, proteins were separated by SDS-PAGE electrophoresis and transferred to a PVDF membrane. Cells were incubated with LC3 and actin antibodies, respectively, and ECL was used for color development. The LC3-II / I ratio was calculated using image analysis software. b) The mRFP-GFP-LC3 plasmid was transfected into hepatocellular carcinoma cells for 24 h. After transfection, CA4 or an equal volume of solvent was added as a control for 24 h. Images were observed and acquired using a confocal microscope, and the number of red and yellow fluorescent spots in each cell was counted. Yellow fluorescence represents autophagosomes that have not fused with lysosomes, while red fluorescence represents autophagolysosomes that have fused with lysosomes.

[0239] Experimental results: Western blot analysis showed that the LC3B-II / I ratio of liver cancer cells in the CA4-treated group was significantly higher than that in the control group. Figure 33 The mRFP-GFP-LC3 dual fluorescence detection results showed that the number of autolysosomes in the CA4-treated group was significantly reduced, and the number of yellow fluorescent spots was increased (p<0.001). Figure 34A and Figure 34B This suggests that the conversion of autophagosomes to autolysosomes is inhibited.

[0240] The above results indicate that CA4 can directly target the CYP7A1 protein and inhibit the formation of autolysosomes, thereby inhibiting the biological function of liver cancer cells at the molecular level, providing experimental evidence for its potential as a candidate drug for liver cancer treatment.

[0241] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for screening inhibitors targeting CYP7A1 protein, characterized in that, Includes the following steps: Step S1: Using the human cholesterol 7α-hydroxylase crystal structure PDB ID: 3V8D as the receptor model, precise docking screening was performed using SchrödingerMaestro software to screen drug-like molecules with molecular weights of 250 Da-500 Da and lipid-water partition coefficients of -1 to 5 from the Zinc15 compound database. Step S2: Using an active learning model of graph neural network, predict the protein-small molecule binding affinity of compounds in the Enamine RealSpace database, and sort the prediction results. Step S3: Calculate the binding free energy of the candidate molecules obtained in Step S1 and / or Step S2 with the CYP7A1 protein using the molecular mechanics / generalized Born surface area method; wherein, molecules with a binding free energy ΔG ≤ -40 kcal / mol are selected as potential inhibitors targeting the CYP7A1 protein.

2. The screening method according to claim 1, characterized in that, Step S1 includes: Define a ligand-binding pocket consisting of His101, Trp284, Val281, Ile125, Phe129, Gly485, Ile114, Leu486, Leu104, Leu361, Ser105, and Phe102 residues; The carbonyl oxygen of Leu361 residue and the carbonyl oxygen of Gly485 residue were identified as key hydrogen bond interaction sites. Compounds conforming to the pharmacophore model constructed based on the PDB 3SN5 and 3V8D co-crystal ligands were docked to the active site of CYP7A1 and sorted from high to low docking scores.

3. The screening method according to claim 1 or 2, characterized in that, Step S2 includes: Trp284, Gly485, and Ser360 were selected as key binding sites for describing the interaction between CYP7A1 and its ligands. The active learning model was trained and validated using protein-small molecule complex data from the PDBbind v2020 database.

4. Use of the screening method according to any one of claims 1-3 in the preparation of anti-hepatocellular carcinoma drugs.

5. The use of an inhibitor targeting CYP7A1 protein in the preparation of an anti-liver cancer drug, characterized in that, The inhibitor is used to block CYP7A1-mediated autophagosome-lysosome fusion and / or increase CD8+. + T cell infiltration, and the inhibitor is obtained by screening according to any one of claims 1-3.

6. The use according to claim 5, characterized in that, The inhibitor is selected from compounds represented by any of the following general formulas I to V, their stereoisomers, or pharmaceutically acceptable salts thereof: 、 、 、 、 ； in, In Formula I, Ra is selected from or Q1 is selected from hydrogen, -C(CH3)2, -C(O)CH3 or -C(O)O-CH2CH3; Rb, Rc and Rd are each independently selected from N or CH; In Formula II, R1 and R4 are each independently selected from hydrogen or -O-CH3, R2, R3, R5, and R7 are each independently selected from hydrogen or methyl, and R6 is selected from methyl. In Formula III, X1 is selected from N or CH, X2 is selected from hydrogen or Cl, X3 is selected from -CH2- or -C(O)-, and X4 is selected from... , or ; In Formula IV, Y1 is selected from hydrogen or Y2 is selected from hydrogen or hydroxyl; Indicates a double bond or a single bond; In formula V, Z1 is selected from hydrogen or hydroxyl, and Z2 is selected from... , or .

7. The use according to claim 6, characterized in that, The inhibitor is selected from the following compounds, their stereoisomers, or pharmaceutically acceptable salts thereof: 。 8. The use according to claim 5, characterized in that, The inhibitor is selected from the following compounds, their stereoisomers, or pharmaceutically acceptable salts thereof: 。 9. The use according to claim 5, characterized in that, The inhibitor is selected from the following compounds, their stereoisomers, or pharmaceutically acceptable salts thereof: 。

Images