ERRalpha agonist screening method based on virtual screening and application thereof

Abstract

The invention discloses an ERRalpha agonist screening method based on virtual screening and application of the ERRalpha agonist screening method. The method comprises the following steps: firstly, constructing a consensus model for preliminary screening from a compound library, and further constructing a method for screening the ERRalpha agonist for multiple rounds by taking the consensus model as a round of screening process; according to the method disclosed by the invention, a candidate ERR alpha agonist N-(3-ethyl phenyl)-6H, 7H, 8H, 9H-pyrido [2, 1-h] purine-4-amine is finally and successfully screened and obtained. Cell experiments prove that the compound has high ERRalpha agonist activity, so that the method can be used for screening ERRalpha agonists, and help can be provided for designing and developing novel diabetes treatment drugs.

Description

Technical Field

[0001] This invention relates to the field of pharmaceutical technology, particularly to the fields of drug screening and diabetes prevention and treatment, and specifically to a drug screening method based on multi-round virtual screening and its application. Background Technology

[0002] Diabetes mellitus is a chronic, progressive disease characterized by a core pathological mechanism of relative or absolute insulin deficiency and decreased insulin sensitivity in target tissues, accompanied by persistent hyperglycemia. Against the backdrop of a global aging population, the proportion of diabetic patients over 60 years of age continues to rise, making diabetes a major public health issue and a significant global chronic disease burden. Based on pancreatic function, diabetes is mainly classified into type 1 diabetes mellitus (T1DM) and type 2 diabetes mellitus (T2DM), with the latter accounting for approximately 90%–95% of all cases. Clinical management of type 2 diabetes currently relies primarily on dietary adjustments, exercise therapy, and oral or injectable hypoglycemic agents. Commonly used drugs include metformin, thiazolidinediones, sulfonylureas, GLP-1 receptor agonists, and DPP-IV inhibitors. However, these drugs often experience a range of adverse reactions, such as gastrointestinal symptoms, weight gain, and hypoglycemic events due to long-term use, indicating room for improvement in efficacy and safety. Therefore, exploring new treatment approaches—including discovering new targets and developing novel drugs—has become an important research direction in this field.

[0003] Studies have shown that ERRα interacts with cofactors such as PGC-1α, directly participating in the regulation of energy metabolism processes such as glycolysis, fatty acid β-oxidation, and oxidative phosphorylation of OXPHOS. In patients with type 2 diabetes, OXPHOS gene expression is generally suppressed, leading to glucose and lipid metabolism disorders, which in turn cause lipid deposition in skeletal muscle and induce insulin resistance. Therefore, restoring OXPHOS function is considered an important pathway to improve insulin sensitivity and glucose homeostasis. At the molecular level, small molecule ERRα agonists can initiate a positive feedback regulatory loop between GABPA and ERRα, thereby enhancing mitochondrial biosynthesis and the transcriptional activity of related metabolic pathways. Based on the core regulatory role of ERRα in energy metabolism, this receptor has become an important potential target for the treatment of metabolic diseases such as type 2 diabetes and obesity. Developing novel small molecule ERRα agonists with high activity and selectivity has therefore become a cutting-edge direction in current drug development.

[0004] Therefore, there is an urgent need in this field for a novel drug screening method to identify new drugs that have preventive and / or therapeutic effects on type 2 diabetes. Summary of the Invention

[0005] The purpose of this invention is to provide a drug screening method, particularly a method for screening novel drugs that have preventive and / or therapeutic effects on type 2 diabetes.

[0006] Another object of the present invention is to provide a compound obtained by screening using the drug screening method of the present invention and its use in the preparation of a drug for the prevention and / or treatment of type 2 diabetes.

[0007] A first aspect of the present invention provides a method for constructing an ERRα agonist prediction model, the method comprising the steps of:

[0008] (A) Provides a compound dataset, the compound dataset including ERRα agonist data and non-ERRα agonist data;

[0009] (B) Each compound in the compound dataset is converted into a molecular fingerprint and a molecular descriptor to obtain the characteristics of each compound;

[0010] (C) The features are used to train multiple machine learning models to obtain multiple base models and the prediction results of each base model;

[0011] (D) The prediction results of each base model are merged to form an integrated feature. The logistic regression algorithm is used to train on the integrated feature, thereby stacking and integrating the multiple base models to finally obtain a model for screening ERRα agonists, namely the consensus model.

[0012] In another preferred embodiment, the compound database includes LOPAC and Tox21 10K.

[0013] In another preferred embodiment, the compound dataset is preprocessed, the preprocessing including:

[0014] (i) Perform SMILES standardization on each compound in the compound dataset, and perform deduplication and conflict data removal based on the standardized SMILES;

[0015] (ii) compounds that have been removed from their salt form and are neutralized by charge; and

[0016] (iii) Remove inorganic matter / metal-organic matter / mixtures.

[0017] In another preferred embodiment, the molecular fingerprint is selected from the group consisting of AP2D, ECFP4, Estate, FCFP4, MACCS, or combinations thereof.

[0018] In another preferred embodiment, the molecular descriptor is an RDKit molecular descriptor.

[0019] In another preferred embodiment, the features include single features and combined features.

[0020] In another preferred embodiment, the single feature includes AP2D, ECFP4, Estate, FCFP4, MACCS, and RDKit molecular descriptors.

[0021] In another preferred embodiment, the composition features include AP2D+RDKit molecular descriptors, ECFP4+RDKit molecular descriptors, Estate+RDKit molecular descriptors, FCFP4+RDKit molecular descriptors, and MACCS+RDKit molecular descriptors.

[0022] In another preferred embodiment, the machine learning model is selected from two or more of the following: Support Vector Machine (SVM), Random Forest (RF), Extreme Gradient Boosting (XGBoost), or Light Gradient Boosting (LightGBM).

[0023] In another preferred embodiment, step (C) further includes the step of: screening the base model based on cross-validation to obtain a preferred base model.

[0024] In another preferred embodiment, the method further includes the step of: (E) validating the consensus model on an external compound dataset.

[0025] A second aspect of the present invention provides a prediction system for ERRα agonists, the prediction system comprising:

[0026] An input module, configured to input data, including compound data;

[0027] A prediction module, configured as a prediction model, performs predictions on the compound data to obtain prediction results; the prediction model is constructed using the method described in the first aspect of the present invention.

[0028] An output module is configured to output the prediction results of the prediction module.

[0029] A third aspect of the present invention provides a method for multi-stage screening of ERRα agonists, the method comprising the steps of:

[0030] (1) Provide several candidate compounds;

[0031] (2) The candidate compounds are predicted using a consensus model, and compounds that reach a predetermined threshold are selected as the first round of screening compounds based on the prediction results; the consensus model is constructed using the method described in the first aspect of the present invention;

[0032] (3) The first-round screening compounds are evaluated using a pre-constructed multi-ligand pharmacophore model, and the first-round screening compounds that reach the predetermined level are selected as second-round screening compounds based on the evaluation results.

[0033] (4) The second-round screening compounds are evaluated using a combination of pre-constructed molecular docking precision and conformation filtering rules, and the second-round screening compounds that reach the predetermined level are selected as third-round screening compounds based on the evaluation results.

[0034] (5) The three-round screening compounds are comprehensively evaluated based on their properties, and the three-round screening compounds that reach the predetermined level are selected as ERRα agonists based on the evaluation results.

[0035] In another preferred embodiment, the pre-constructed multiligand pharmacophore model is constructed based on the known binding modes of ERRα agonists to ERRα receptors.

[0036] In another preferred embodiment, the bonding mode includes:

[0037] (a1) The compound forms hydrogen bonds or halogen bonds with the only conserved water molecule in the binding pocket and may undergo π–π stacking with Phe328, Phe382, Phe495, and Phe510.

[0038] (a2) The compound forms hydrogen or halogen bonds with Asp329, His517, or Pro505, and may be accompanied by π–π stacking with Phe328, Phe382, Phe495, or Phe510; and

[0039] (a3) The compound does not form hydrogen or halogen bonds with any residues, but only with π–π stacking with Phe328, Phe382, Phe495, and Phe510.

[0040] In another preferred embodiment, the combination of the pre-constructed molecular docking precision and conformational filtering rules is constructed based on the known binding modes of ERRα agonists and ERRα receptors.

[0041] In another preferred embodiment, the conformational filtering rule includes:

[0042] (b1) The compound must form a hydrogen bond or a halogen bond with the water molecule at position 701;

[0043] (b2) The compound must form a hydrogen bond or halogen bond with residues at Asp329, Pro505, or His517; and

[0044] (b3) Compounds that do not satisfy (b1) or (b2) must form a π–π stack with at least one of Phe328, Phe382, Phe495 or Phe510.

[0045] In another preferred embodiment, the binding mode of the known ERRα agonist to the ERRα receptor is obtained by molecular docking using the known ERRα agonist and the ERRα protein crystal structure, and by analysis and induction based on secondary bond interactions.

[0046] In another preferred embodiment, the PDB number of the ERRα protein crystal structure is 7E2E.

[0047] In another preferred embodiment, in step (2), "reaching the predetermined threshold" means that the predicted probability is greater than 0.5.

[0048] In another preferred embodiment, step (3) specifically includes: evaluating the first-round screening compounds using multiple pre-constructed multi-ligand pharmacophore models, wherein the multi-ligand pharmacophore models extract key functional features from the first-round screening compounds, and obtaining a good fit by performing three-dimensional conformation matching between the key functional features and the feature composition in the multi-ligand pharmacophore models; selecting the first-round screening compounds that reach a predetermined level based on the good fit in each multi-ligand pharmacophore model; and merging the selection results obtained from each multi-ligand pharmacophore model as second-round screening compounds.

[0049] In another preferred embodiment, the pre-constructed multi-ligand pharmacophore model consists of three elements.

[0050] In another preferred embodiment, the multi-ligand pharmacophore model and its characteristic components include:

[0051] (z1)AADRR: contains 2 hydrogen bond acceptors, 1 hydrogen bond donor and 2 aromatic rings;

[0052] (z2)ADRRR: Contains one hydrogen bond acceptor, one hydrogen bond donor, and three aromatic rings; and

[0053] (z3)AHHRR: contains one hydrogen bond acceptor, two hydrophobic groups and two aromatic rings.

[0054] In another preferred embodiment, in step (3), "reaching the predetermined level" means that the evaluation result ranking exceeds 50%, such as 40%, 30%, 20%, 10%, 5%, and more preferably exceeds 10%.

[0055] In another preferred embodiment, step (4) specifically includes: evaluating the second-round screening compound using a combination of multiple pre-constructed molecular docking precision and conformational filtering rules, wherein the second-round screening compound is evaluated according to the molecular docking precision and screened according to the conformational filtering rules; the combination of the multiple molecular docking precision and conformational filtering rules is applied to the selection results obtained in each of the multi-ligand pharmacophore models in step (3); and the selection results obtained in each combination are combined as the third-round screening compound.

[0056] In another preferred embodiment, the combination of the pre-constructed molecular docking precision and conformation filtering rules is three.

[0057] In another preferred embodiment, the combination of molecular docking precision and conformational filtering rules includes:

[0058] (y1) Combination 1: Consists of SP docking accuracy and conformation filtering rules shown in (b1);

[0059] (y2) Combination 2: Consists of XP docking accuracy and the conformation filtering rules shown in (b2); and

[0060] (y3) Combination 3: Composed of XP docking accuracy and conformation filtering rules shown in (b3).

[0061] In another preferred embodiment, in step (4), "reaching the predetermined level" means that the evaluation results are ranked in the top 50, such as 40, 30, 20, 10, and more preferably in the top 30.

[0062] In another preferred embodiment, the properties of the compound include: water solubility, binding free energy, molecular weight, lipid-water partition coefficient, number of hydrogen bond acceptors, number of hydrogen bond donors, and compound availability.

[0063] A fourth aspect of the present invention provides a system for multi-stage screening of ERRα agonists, the system comprising:

[0064] An input module, configured to input data, including compound data;

[0065] The filtering module includes:

[0066] (a1) A screening unit, the screening unit being configured as a prediction model, the prediction model predicting the compound data and selecting compounds that reach a predetermined threshold as screening compounds in the first round based on the prediction results; the prediction model is constructed using the method of claim 1;

[0067] (a2) A second-round screening unit, the second-round screening unit being configured to perform the following operations: evaluate the first-round screening compound using a pre-constructed multi-ligand pharmacophore model, and select the first-round screening compound that reaches a predetermined level as the second-round screening compound based on the evaluation results;

[0068] (a3) A three-round screening unit, the three-round screening unit being configured to perform the following operations: evaluate the two-round screening compounds using a combination of pre-constructed molecular docking precision and conformational filtering rules, and select the two-round screening compounds that reach a predetermined level as the three-round screening compounds based on the evaluation results;

[0069] (a4) A four-round screening unit, the four-round screening unit being configured to perform the following operations: comprehensively evaluate the three-round screened compounds based on the properties of the compounds, and select the three-round screened compounds that reach a predetermined level as ERRα agonists based on the evaluation results, thereby obtaining screening results;

[0070] An output module is configured to output the filtering results of the filtering module.

[0071] A fifth aspect of the present invention provides a compound obtained by the method described in the third aspect of the present invention, the compound having a structure as shown in formula (I):

[0072]

[0073] A sixth aspect of the present invention provides a medicament or pharmaceutical composition comprising the compound described in the fifth aspect of the present invention, and a pharmaceutically acceptable carrier, diluent, or excipient.

[0074] In another preferred embodiment, the drug or drug composition comprises an ERRα agonist.

[0075] In another preferred embodiment, the dosage form of the drug or drug composition includes tablets, granules, capsules, pills, solutions, emulsions, and suspensions.

[0076] A seventh aspect of the invention provides the use of a compound obtained by the method described in the third aspect of the invention for the preparation of medicaments, pharmaceutical compositions and / or kits for treating or preventing diabetes.

[0077] In another preferred embodiment, the compound comprises N-(3-ethylphenyl)-6H,7H,8H,9H-pyrido[2,1-h]purine-4-amine.

[0078] In another preferred embodiment, the N-(3-ethylphenyl)-6H,7H,8H,9H-pyrido[2,1-h]purine-4-amine has a structure as shown in formula (I):

[0079]

[0080] In another preferred embodiment, the diabetes is type 2 diabetes.

[0081] It should be understood that, within the scope of this invention, the above-described technical features of this invention and the technical features specifically described below (such as in the embodiments) can be combined with each other to form new or preferred technical solutions. Due to space limitations, they will not be described in detail here. Attached Figure Description

[0082] Figure 1 A schematic diagram of the screening process for compounds with ERRα agonist activity is shown.

[0083] Figure 2 The diagram shows an S-shaped curve of the concentration of the novel estrogen-related receptor α agonist of the present invention and its corresponding ERRα agonist rate. Detailed Implementation

[0084] Through extensive and in-depth research, the inventors have for the first time constructed a consensus model for the initial screening of ERRα agonists. Using this consensus model as a first-round screening process, a multi-round screening method for ERRα agonists was further developed. This method sequentially screens candidate ERRα agonists three more times based on multi-ligand pharmacophore models, molecular docking precision and conformation, and properties such as water solubility, ultimately successfully obtaining the desired compounds. Furthermore, cell experiments verified that N-(3-ethylphenyl)-6H,7H,8H,9H-pyrido[2,1-h]purine-4-amine among the candidate compounds possesses strong ERRα agonist activity, thus serving as a lead compound for the treatment of diabetes. Therefore, the method of this invention and the compounds obtained through screening can provide assistance for the future design and development of novel diabetes treatments. Based on this, this invention was completed.

[0085] It should be understood that the specific methods and experimental conditions of the invention described below in varying degrees of detail are intended to provide a substantive understanding of the invention. Definitions of certain terms used in this specification are provided below. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0086] the term

[0087] As used herein, the terms “containing” or “including (comprise)” can be open-ended, semi-closed, or closed-ended. In other words, the terms also include “consistently made of” or “made of”.

[0088] As used herein, the term “and / or” refers to and covers any and all possible combinations of one or more of the related listed items.

[0089] As used in this article, the term "significant" means that, in a hypothesis test, the observed effect (such as the difference between the experimental and control groups) is unlikely to be caused solely by random error. A hypothesis test includes: the null hypothesis (H0), which assumes that the observed effect does not exist (such as no difference between the experimental and control groups); the p-value, which is the probability of observing the current or more extreme effect when H0 is true; and the significance threshold (α). The significance threshold is typically used to determine whether a hypothesis test is significant. Generally, the significance threshold is 0.05. If the p-value ≤ α, then H0 is rejected, meaning the observed effect exists, and the result is called "significant."

[0090] As used in this article, the term "Ribinski's Five Rules" refers to five basic principles for screening drug molecules, including molecular weight, lipid-water partition coefficient, number of hydrogen bond acceptors, and number of hydrogen bond donors. Orally active drugs typically meet the following criteria: molecular weight less than 500, lipid-water partition coefficient not exceeding 5, number of hydrogen bond acceptors not exceeding 10, and number of hydrogen bond donors not exceeding 5.

[0091] As used herein, the terms “method described in this invention”, “screening method of this invention”, “screening method” and “multi-stage screening method” are used interchangeably to refer to the method of screening ERRα agonists of this invention.

[0092] Estrogen-related receptor α (ERRα)

[0093] In 1988, Giguère et al. successfully identified the orphan nuclear receptor ERRα for the first time using low-rigidity hybridization probe cloning technology. As a member of the nuclear receptor superfamily, ERRα is highly expressed in metabolically active tissues such as brown adipose tissue, skeletal muscle, and liver. Studies have shown that ERRα interacts with cofactors such as PGC-1α, directly participating in the regulation of energy metabolism processes such as glycolysis, fatty acid β-oxidation, and oxidative phosphorylation of OXPHOS. Of particular note is that ERRα is one of the direct downstream targets of PGC-1α, a key coactivator of mitochondrial biosynthesis and metabolic regulation. ERRα, along with PGC-1α and GA-binding protein A (GABPA), constitutes a transcriptional regulatory complex that activates the transcription of multiple OXPHOS-related genes.

[0094] This invention constructs a combination of multi-ligand pharmacophore models, molecular docking precision, and conformational filtering rules based on known ERRα agonists and the structure of ERRα, for screening candidate ERRα agonists from a compound library. Through screening, this invention discovered that N-(3-ethylphenyl)-6H,7H,8H,9H-pyrido[2,1-h]purine-4-amine can serve as an ERRα agonist, and thus can be used in drug development for type 2 diabetes.

[0095] N-(3-Ethylphenyl)-6H,7H,8H,9H-pyrido[2,1-h]purine-4-amine

[0096] Using the aforementioned multi-round screening method, this invention also provides a novel estrogen-associated receptor α agonist. This estrogen-associated receptor α agonist has a novel structure and strong estrogen-associated receptor α agonist activity, thus it can serve as a lead compound for the treatment of diabetes, providing assistance for the future design and development of novel diabetes treatments. This novel estrogen-associated receptor α agonist is N-(3-ethylphenyl)-6H,7H,8H,9H-pyrido[2,1-h]purin-4-amine, having the following structural formula:

[0097]

[0098] Molecular fingerprint

[0099] As used in this article, the term "molecular fingerprint" is a method for describing molecular structure that reduces computational costs and represents the chemical properties of a compound, including structural features and physicochemical properties.

[0100] The term "AP2D" refers to a two-dimensional fingerprint based on atomic pairs, which characterizes molecular structure by the distances between all atomic pairs and the types of atoms. Atom types typically include element type, hybridization state, bonding configuration, etc., while distances are represented by the number of bonds along the shortest path between atoms.

[0101] The term "EState" is a fingerprint based on atomic-level electrotopological state indices, which characterizes molecules by calculating the electrotopological state values ​​of each atom. These values ​​combine the electronic properties of the atom (such as charge and polarization) with the topological environment (such as connectivity).

[0102] The term "ECFP4" is an extended-connectivity fingerprint that represents molecular structure using circular atomic neighborhoods. Here, "4" refers to the computational diameter of an atomic neighborhood of 4; the computational diameter can also be 6 or 8, etc.

[0103] The term "FCFP4" is a variant of extended connectivity fingerprint (Functional-Class Fingerprints4), which represents molecular structures based on functional class atomic neighborhoods. Functional classes include chemical functional groups such as hydrogen bond donors, hydrogen bond acceptors, hydrophobic atoms, and charge centers.

[0104] The term "MACCS" is a substructure-based fingerprint that characterizes molecules using a series of predefined substructure patterns, represented by binary codes indicating the presence of these substructures. These substructures typically include common chemical groups and patterns, such as hydroxyl groups, amino groups, and conjugated double bonds.

[0105] The ERRα agonist screening method of the present invention

[0106] This invention provides a method for screening ERRα agonists, which is derived from a computer-simulated drug design technique to shorten the discovery time and process. This technique is called virtual screening. The compound library used for virtual screening, including the compounds in this invention, is derived from the ChemDiv compound library (https: / / www.chemdiv.com / ). Specifically, the method includes the following steps:

[0107] (1) Provide several candidate compounds;

[0108] (2) The candidate compounds are predicted using a consensus model, and compounds that reach a predetermined threshold are selected as the first round of screening compounds based on the prediction results; the consensus model is constructed using the method described in the first aspect of the present invention;

[0109] (3) The first-round screening compounds are evaluated using a pre-constructed multi-ligand pharmacophore model, and the first-round screening compounds that reach the predetermined level are selected as second-round screening compounds based on the evaluation results.

[0110] (4) The second-round screening compounds are evaluated using a combination of pre-constructed molecular docking precision and conformation filtering rules, and the second-round screening compounds that reach the predetermined level are selected as third-round screening compounds based on the evaluation results.

[0111] (5) The three-round screening compounds are comprehensively evaluated based on their properties, and the three-round screening compounds that reach the predetermined level are selected as ERRα agonists based on the evaluation results.

[0112] More specifically, the method includes the following steps: (1) First round of screening: The ChemDiv compound library is initially screened using an artificial intelligence consensus model built based on combined features and stacked integration methods. The screening results are sorted according to the model score, and compounds with a prediction probability greater than 0.5 (i.e., the prediction result is an ERRα agonist) are retained.

[0113] (2) Second round of screening: The compound molecules obtained in the first round of screening were further screened using three multi-ligand pharmacophore models based on three binding modes. The screening results were sorted according to the model scores, and the top 10% of the compounds in the three screenings were retained respectively.

[0114] (3) Third round of screening: The compounds obtained from the second round of screening were scored and filtered using three molecular docking methods and conformation filtering rules based on three binding modes. The top 30 compounds were retained, for a total of 90 compounds.

[0115] (4) Fourth round of screening: The water solubility of the compounds obtained in the third round of screening and the binding free energy score are considered, and candidate compounds are selected by combining Ribinsky's five rules and the availability of compounds.

[0116] Drugs or drug compositions and methods of administration

[0117] This invention provides a drug or pharmaceutical composition comprising N-(3-ethylphenyl)-6H,7H,8H,9H-pyrido[2,1-h]purine-4-amine within a safe and effective amount range, and a pharmaceutically acceptable carrier, diluent, and excipient. "Safe and effective amount" refers to an amount of the compound sufficient to significantly improve the condition without causing serious side effects. Typically, the pharmaceutical composition contains 1-2000 mg of the compound of this invention per dose, more preferably, 10-1000 mg of the compound of this invention per dose. Preferably, "one dose" is a capsule, tablet, or solution.

[0118] A typical formulation is prepared by mixing the traditional Chinese medicine monomers or combinations thereof described in the first aspect of this invention with a carrier, diluent, or excipient. Suitable carriers, diluents, or excipients are well known to those skilled in the art and include substances such as carbohydrates, waxes, water-soluble and / or expandable polymers, hydrophilic or hydrophobic substances, gelatin, oils, solvents, and water.

[0119] The specific carrier, diluent, or excipient used will depend on the method of use and purpose of the ERRα agonist according to the invention (such as N-(3-ethylphenyl)-6H,7H,8H,9H-pyrido[2,1-h]purine-4-amine). Solvents are generally selected based on those deemed safe and effective for administration to mammals by those skilled in the art. Generally, safe solvents are non-toxic aqueous solvents, such as water, and other non-toxic solvents that are soluble in or miscible with water. Suitable aqueous solvents include one or more of water, ethanol, propylene glycol, polyethylene glycol (such as PEG400, PEG300), etc. The formulation may also include one or more sustained-release agents, stabilizers, surfactants, lubricants, emulsifiers, suspending agents, preservatives, antioxidants, opacifiers, flow aids, processing aids, colorants, sweeteners, flavorings, or other known additives to manufacture or use the drug in an acceptable form.

[0120] The term "pharmaceutically acceptable carrier" refers to one or more compatible solid or liquid fillers or gel substances suitable for human use, and must have sufficient purity and sufficiently low toxicity. "Compatibility" refers to the ability of each component in the composition to combine with or mix with the herbal monomers of the present invention without significantly reducing the efficacy of the compound. Pharmaceutically acceptable carriers include: cellulose and its derivatives (such as sodium carboxymethyl cellulose, sodium ethyl cellulose, cellulose acetate, etc.), gelatin, talc, solid lubricants (such as stearic acid, magnesium stearate), calcium sulfate, vegetable oils (such as soybean oil, sesame oil, peanut oil, olive oil, etc.), polyols (such as propylene glycol, glycerin, mannitol, sorbitol, etc.), emulsifiers (such as... Wetting agents (such as sodium dodecyl sulfate), colorants, flavoring agents, stabilizers, antioxidants, preservatives, pyrogen-free water, etc.

[0121] When the ERRα agonist (such as N-(3-ethylphenyl)-6H,7H,8H,9H-pyrido[2,1-h]purine-4-amine) of the present invention is used in combination with one or more other drugs, the two or more drugs may be used separately or in combination, preferably administered in the form of a pharmaceutical composition. The ERRα agonist (such as N-(3-ethylphenyl)-6H,7H,8H,9H-pyrido[2,1-h]purine-4-amine) of the present invention can be administered orally, by inhalation, parenteral administration, oral cavity, sublingually, rectally, vaginally, by patch, by pump or transdermal administration, and accordingly formulated into pharmaceutical compositions. Parenteral administration includes, for example, intravenous, intra-arterial, subcutaneous, nasal, intrapulmonary, rectal, and topical application methods. These pharmaceutical compositions may also contain one or more sustained-release agents, stabilizers, surfactants, wetting agents, lubricants, emulsifiers, suspending agents, preservatives, antioxidants, opacifiers, flow aids, processing aids, colorants, sweeteners, flavorings, or other known additives to make or use the pharmaceutical composition in an acceptable form.

[0122] The administration method of the drug or pharmaceutical composition of the present invention is not particularly limited, and representative administration methods include (but are not limited to): oral, intratumoral, rectal, parenteral (intravenous, intramuscular, or subcutaneous), and local administration. The specific administration method depends on the dosage form of the drug or pharmaceutical composition and whether it is convenient for the subject to ingestion so as to exert its effect in the subject. Preferably, the dosage form of the drug or pharmaceutical composition includes tablets, granules, capsules, pills, solutions, emulsions, and suspensions.

[0123] The medicaments or pharmaceutical compositions of the present invention can be administered alone or in combination with other medicaments for treating diabetes (such as type 2 diabetes). When administering the medicaments or pharmaceutical compositions, a therapeutically and / or preventively effective amount of the medicaments or pharmaceutical compositions of the present invention is given to a subject (e.g., a human) requiring treatment and / or prevention. For a person weighing 60 kg, the daily dose is typically 1-2000 mg, preferably 50-1000 mg. The actual dosage, rate of administration, and timing of administration depend on the nature and severity of the disease / condition and the specific medicament or pharmaceutical composition being administered. Prescribing treatments, such as determining the dosage, falls within the skill scope of a skilled physician and typically takes into account the disease / condition to be treated, the individual subject's condition, the site of administration, the method of administration, and other factors.

[0124] The drug of the present invention is preferably administered orally. Solid dosage forms for oral administration include capsules, tablets, pills, powders, solutions, and granules. In the solid dosage form, the active compound is mixed with at least one conventional inert excipient (or carrier), such as sodium citrate or dicalcium phosphate, or with the following components: (a) fillers or compatibilizers, such as starch, lactose, sucrose, glucose, mannitol, and silica; (b) binders, such as hydroxymethyl cellulose, alginate, gelatin, polyvinylpyrrolidone, sucrose, and gum arabic; (c) humectants, such as glycerin; (d) disintegrants, such as agar, calcium carbonate, potato starch or cassava starch, alginate, certain complex silicates, and sodium carbonate; (e) slowing agents, such as paraffin wax; (f) absorption accelerators, such as quaternary ammonium compounds; (g) wetting agents, such as cetyl alcohol and glyceryl monostearate; (h) adsorbents, such as kaolin; and (i) lubricants, such as talc, calcium stearate, magnesium stearate, solid polyethylene glycol, sodium dodecyl sulfate, or mixtures thereof. In capsules, tablets, and pills, dosage forms may also contain buffers.

[0125] Solid dosage forms such as tablets, sugar pills, capsules, pellets, and granules can be prepared using coatings and shells, such as casings and other materials known in the art. They may contain opacifying agents, and the release of the active compound or compound in such compositions can be delayed in a portion of the digestive tract. Examples of encapsulating components that can be used are polymeric substances and waxes. If necessary, the active compound may also form microcapsules with one or more of the excipients described above. In the case of capsules and tablets, the dosage form may also include buffers. Similar types of solid compositions can also be used as fillers in soft and hard filled gelatin capsules, using lactose and polyethylene glycol as excipients. Other materials known in the art include liposomes. Liposomes are composed of various phospholipids, such as cholesterol, stearamide, or phosphatidylcholine.

[0126] Liquid dosage forms for oral administration include pharmaceutically acceptable emulsions, solutions, suspensions, syrups, or tinctures. In addition to the active compound, liquid dosage forms may contain inert diluents conventionally used in the art, such as water or other solvents, solubilizers and emulsifiers, e.g., ethanol, isopropanol, ethyl carbonate, ethyl acetate, propylene glycol, 1,3-butanediol, dimethylformamide, and oils, particularly cottonseed oil, peanut oil, corn germ oil, olive oil, castor oil, and sesame oil, or mixtures of these substances.

[0127] In addition to the inert diluents mentioned above, the composition may also contain auxiliaries such as wetting agents, emulsifiers and suspending agents, sweeteners, flavoring agents and fragrances.

[0128] In addition to the active compound, the suspension may contain suspending agents such as ethoxylated isooctadecyl alcohol, polyoxyethylene sorbitol and dehydrated sorbitol esters, microcrystalline cellulose, aluminum hydroxide, bentonite, agar, astragalus gum, aluminum methoxide and agar or mixtures of these substances.

[0129] Compositions for parenteral injection may comprise physiologically acceptable sterile aqueous or anhydrous solutions, dispersions, suspensions, or emulsions, and sterile powders for reconstitution into sterile injectable solutions or dispersions. Suitable aqueous and non-aqueous carriers, diluents, solvents, or excipients include water, ethanol, polyols, and suitable mixtures thereof.

[0130] Compounds obtained by the ERRα agonist screening method of this invention and their uses

[0131] This invention provides the use of a compound obtained by the method described herein for the preparation of medicaments, pharmaceutical compositions and / or kits for the treatment or prevention of diabetes.

[0132] The screening method of this invention includes fitting candidate compounds to the known ERRα agonist-ERRα protein interaction structures, selecting compounds with high fitting accuracy; evaluating the properties of these compounds, selecting compounds with favorable properties; thereby obtaining newly discovered ERRα agonists. These newly discovered ERRα agonists are predicted to bind to the ERRα protein and are predicted to have good biological properties. Therefore, it can be understood that, provided a sufficient number of candidate compounds are provided, the method of this invention can screen for new ERRα agonists. Furthermore, it can be understood that new ERRα agonists obtained by screening existing compounds outside the compound library used in this study using the method of this invention can also be used to prepare medicaments, pharmaceutical compositions, and / or kits for treating or preventing diabetes; or new ERRα agonists obtained by screening newly synthesized or newly discovered unknown compounds using the method of this invention can also be used to prepare medicaments, pharmaceutical compositions, and / or kits for treating or preventing diabetes.

[0133] The medicine box typically includes a medicine or pharmaceutical composition containing the ERRα agonist discovered in this invention, as well as other medicines that can treat diabetes, such as other ERRα agonists. Preferably, the diabetes is type 2 diabetes.

[0134] The main advantages of this invention include:

[0135] (1) The present invention first constructs a consensus model, which uses machine learning algorithms to screen compounds based on their molecular fingerprint data, and achieves good accuracy in both the test set and the validation set.

[0136] (2) This invention proposes a multi-round drug screening method. The method uses a consensus model for the first round of screening, scores the candidates based on the multi-ligand pharmacophore model for the second round of screening, then scores the candidates based on molecular docking accuracy and conformation for the third round of screening, and finally scores the candidates based on a combination of properties including water solubility and binding free energy for further screening, thereby obtaining the desired candidate compounds.

[0137] (3) Using the drug screening method of the present invention, a compound N-(3-ethylphenyl)-6H,7H,8H,9H-pyrido[2,1-h]purine-4-amine was first discovered to be an ERRα agonist and could be used for the development of novel diabetes drugs.

[0138] The present invention will be further illustrated below with reference to specific embodiments. It should be understood that these embodiments are for illustrative purposes only and are not intended to limit the scope of the invention. Experimental methods in the following embodiments, unless otherwise specified, are generally performed under conventional conditions, such as those described in Sambrook et al., Molecular Cloning: A Laboratory Manual (New York: Cold Spring Harbor Laboratory Press, 1989), or as recommended by the manufacturer. Unless otherwise stated, percentages and parts are weight percentages and parts by weight.

[0139] Example 1: Multi-level virtual screening based on combined pattern classification

[0140] This embodiment relates to a method for constructing a multi-level virtual screening of a drug, the process of which can be found in [link to documentation]. Figure 1 Specifically, a multi-level virtual screening process was used to screen the ChemDiv compound database using an AI consensus model incorporating binding mode information, a multi-ligand pharmacophore model, molecular docking conformation filtering, and comprehensive analysis, ultimately yielding the compounds described in this invention. The specific steps are as follows:

[0141] (1) First round of screening:

[0142] An artificial intelligence consensus model based on combinatorial features and stacked integration methods was used to perform preliminary screening of the ChemDiv compound library.

[0143] ERRα agonist data were obtained from the LOPAC and Tox21 10K compound libraries (corresponding to PubChem BioAssay AID1224820 and AID1259404). The raw data were preprocessed using RDKit (2023.9.2) and MolVS (0.1.1), specifically including: SMILES normalization, salt removal, charge neutralization, removal of inorganic / organometallic / mixture compounds, and deduplication and conflict removal based on the normalized SMILES. The data was then randomly divided into training and test sets in a 9:1 ratio. To alleviate the imbalance between positive and negative samples, a random undersampling method was used on the training set to balance the positive and negative sample ratio. This random partitioning and undersampling process was repeated 10 times to enhance the stability of the model evaluation results. In addition, supplementary data were collected from the ChEMBL database, relevant literature, and patents to construct an external validation set for evaluating the model's generalization ability.

[0144] This study employs five molecular fingerprints (AP2D, ECFP4, EState, FCFP4, and MACCS) and a set of RDKit two-dimensional molecular descriptors to characterize compound structures. Furthermore, each molecular fingerprint is combined with a descriptor to construct five combined features. Finally, the six single features and the five combined features are used together as model input.

[0145] Base models were constructed based on the aforementioned 11 features and 4 machine learning algorithms (RF, SVM, XGBoost, and LightGBM). The hyperparameters of all models were optimized using grid search combined with AUC values ​​from 5-fold cross-validation. Furthermore, based on the 5 base models with the best cross-validation performance, a stacked ensemble strategy was adopted to construct a consensus model: the first layer integrates the prediction results of each base model to form a new feature matrix; the second layer is trained using logistic regression to generate more robust ensemble predictions. Model performance was primarily evaluated using the area under the ROC curve (AUC) and balanced accuracy (BA). The constructed consensus model performed well on both the test set (AUC = 0.876, BA = 0.816) and the external validation set (AUC = 0.867, BA = 0.777), demonstrating reliable predictive ability.

[0146] The ChemDiv compound database was screened using a constructed consensus model. The screening results were sorted according to the model scores, and compounds with a prediction probability greater than 0.5 (i.e., predicted to be ERRα agonists) were retained. A total of 49,108 compounds were selected for the next round of screening.

[0147] (2) Second round of screening:

[0148] Based on the Glide module of Maestro 13.4, XP-precision molecular docking was performed on 72 known ERRα agonists and ERRα crystal structures (PDB: 7E2E) collected from literature and patents. Three different binding modes were identified through analysis of secondary bond interactions. Binding Mode 1: The compound forms a hydrogen bond or halogen bond with the only conserved water molecule in the binding pocket, and may undergo π–π stacking with Phe328, Phe382, Phe495, and Phe510. Binding Mode 2: The compound forms a hydrogen bond or halogen bond with Asp329, His517, or Pro505, and may also undergo π–π stacking with the four phenylalanine residues mentioned above. Binding Mode 3: The compound does not form hydrogen bonds or halogen bonds with any residues, relying solely on π–π stacking with the four phenylalanine residues mentioned above.

[0149] Based on the above binding modes, multi-ligand pharmacophore models were further constructed using the Phase module of Maestro 13.4. Model performance was evaluated using AUC and ROC enrichment rate (ROCE), where AUC reflects the overall discriminative ability of the model, and ROCE@5% indicates that the model's efficiency in enriching active compounds is folded relative to that of random screening when screening the top 5% of compounds. The three pharmacophore models constructed are characterized by the following features and performance: Pharmacophore model 1 is AADRR (AUC = 0.770, ROCE 5% = 7.89), containing 2 hydrogen bond acceptors, 1 hydrogen bond donor, and 2 aromatic rings; pharmacophore model 2 is ADRRR (AUC = 0.964, ROCE 5% = 13.56), containing 1 hydrogen bond acceptor, 1 hydrogen bond donor, and 3 aromatic rings; pharmacophore model 3 is AHHRR (AUC = 0.830, ROCE 5% = 10.00), containing 1 hydrogen bond acceptor, 2 hydrophobic groups, and 2 aromatic rings. These models can extract the key functional features required for ERRα binding from molecules with the same binding mode. Pharmacophore screening uses three-dimensional conformation matching to rank compounds based on fit, thereby enabling rapid identification and experimental verification of potential active molecules in a large-scale compound library.

[0150] The three constructed multi-ligand pharmacophore models were applied to the compounds obtained in the first round of screening. The compounds were ranked according to the model scores, and the top 10% of each model were retained.

[0151] (3) Third round of screening:

[0152] Based on the three binding modes mentioned above, the Glide module in Maestro 13.4 was used to construct corresponding conformational filtering rules, and the optimal docking precision (HTVS, SP, or XP) for each was determined. The specific conformational filtering rules are set as follows: Rule 1: The compound must form a hydrogen bond or halogen bond with the water molecule at position 701; Rule 2: The compound must form a hydrogen bond or halogen bond with residues at positions Asp329, Pro505, or His517; Rule 3: The compound does not satisfy the interactions described in Rules 1 and 2, but must form a π–π stack with at least one of Phe328, Phe382, Phe495, or Phe510.

[0153] The screening performance of various docking precision and conformational filtering combinations was evaluated using AUC and ROCE@5%. The results are as follows: Binding Mode 1: SP docking + Rule 1 (AUC = 0.885, ROCE@5% = 10.53); Binding Mode 2: XP docking + Rule 2 (AUC = 0.831, ROCE@5% = 8.34); Binding Mode 3: XP docking + Rule 3 (AUC = 0.792, ROCE@5% = 5.00). This combined screening strategy first ranks the compounds based on their docking scores (glide_gscore), and then excludes molecules that do not conform to the corresponding binding modes through conformational filtering, thereby effectively improving the enrichment efficiency of active compounds.

[0154] The above three combined methods were applied to the compounds obtained in the second round of three screenings. The top 30 compounds in docking scores were retained in each screening, and a total of 90 compounds were finally selected.

[0155] (4) Fourth round of screening: A comprehensive evaluation was conducted on the 90 compounds selected in the third round, including water solubility and binding free energy scores. Potentially active compounds were selected based on Ribinsky's five rules and compound availability. The binding free energy score was calculated using the MM-GBSA module in Prime, allowing for adjustments to ligand molecules and... Protein atoms within the range can move flexibly, while other atoms in the protein remain fixed.

[0156] Ultimately, through a multi-level virtual screening strategy involving an artificial intelligence consensus model, a multi-ligand pharmacophore model, molecular docking conformation filtering, and comprehensive analysis, a total of 43 compounds were obtained for bioactivity testing, including the novel estrogen-related receptor α agonist of this invention.

[0157] Example 2: Bioactivity testing of candidate compounds

[0158] This embodiment relates to verifying the bioactivity of candidate compounds through cell experiments.

[0159] Experimental materials: Firefly luciferase reporter gene assay kit purchased from Yisheng Biotechnology; Ad-ERRE-SV40-luc purchased from Hanheng Biotechnology; Ad-ERRα purchased from Hanheng Biotechnology; DMEM high glucose medium purchased from Shanghai Fuheng Biotechnology; fetal bovine serum purchased from Shanghai Fuheng Biotechnology; white 96-well fluorescent plates purchased from LABSELECT; Synergy 2 microplate reader from Biotek; HeLa cells from Shanghai Fuheng Cell Bank; DMSO purchased from Yuanye Biotechnology; MTT reagent purchased from Beyotime Biotechnology.

[0160] Experimental instrument: Biotek Synergy 2 microplate reader

[0161] The specific steps include:

[0162] 1. Cell model construction: First, Ad-ERRα with an infection multiplicity (MOI) of 100 and Ad-ERRE-SV40-luc with an infection multiplicity (MOI) of 300 were transfected into cells to construct an ERRα agonist screening model.

[0163] 2. ERRα agonist screening: The activity assay method used was the luciferase reporter gene assay. Assay principle: During the assay, firefly luciferin was added as a substrate, and luciferase reacted in Mg... 2+ With the participation of O2, luciferin is catalyzed to be oxidized to oxyluciferin. During the oxidation process, luciferin emits bioluminescence. The activity of luciferase can be determined by detecting the intensity of the fluorescence. The better the activating effect, the stronger the fluorescence.

[0164] Experimental steps:

[0165] (1) Place Ad-ERRE-SV40-luc and Ad-ERRa on ice for later use.

[0166] (2) The compound was prepared into a 10 mM stock solution using DMSO for later use.

[0167] (3) Prepare a cell suspension from HeLa cells that have grown to 80% concentration, seed them in a 96-well plate, and incubate overnight in a culture incubator.

[0168] (4) Infect cells with Ad-ERRα with a multiplicity of infection (MOI) of 100 and Ad-ERRE-SV40-luc with a MOI of 300, and replace with fresh complete culture medium after 8 hours.

[0169] (5) 36 hours after virus transfection, the compound stock solution was diluted to the required concentration with DMEM high sugar medium and added to the cells for incubation. The control group was given medium containing 0.1% DMSO.

[0170] (6) After culturing for 24 hours, the activity of luciferase was detected by microplate reader according to the instructions of the firefly luciferase reporter gene assay kit.

[0171] (7) The activation rate of the compound was calculated. The calculation formula is as follows:

[0172] Activation rate % = [luc(compound) - luc(control)] / 1uc(control) × 100%

[0173] Where luc represents fluorescence intensity.

[0174] (8) To eliminate the influence of high-concentration drug cytotoxicity on the experimental results, the MTT assay was used to detect the drug's toxic effects on cells. After co-incubating the drug with transfected HeLa cells for 24 h, MTT reagent (final MTT concentration 0.5 mg / mL) was added. After 4 h, the old culture medium was discarded, and DMSO was added to dissolve formazan crystals. The absorbance was then measured at 490 nm using a Synergy2 microplate reader, and cell viability was calculated. When fitting the drug activation curve, concentrations that showed significant differences in cell viability compared to the control group were removed to eliminate the influence of cytotoxic drug concentrations on the experimental results.

[0175] In the experiment, a culture medium containing 0.1% DMSO was used as a blank control group. At least seven test compounds were used, and their concentrations in the reaction system and their corresponding activation rates were plotted as sigmoid curves to determine the EC50. 50 Values ​​(curve fitting and EC were performed using the data processing software GraphPad 10.1.2) 50 calculate).

[0176] like Figure 2 As shown, the results indicate that the compound exhibits agonistic activity towards ERRα, and its EC50... 50 With a value of 0.2529 μM, it can be used as a drug lead compound for further structural modification to develop novel ERRα agonists.

[0177] In summary, the novel estrogen-associated receptor α agonist of this invention has a novel structure and strong estrogen-associated receptor α agonist activity, thus it can serve as a lead compound for the treatment of diabetes and provide assistance for the future design and development of novel diabetes treatment drugs.

[0178] All documents mentioned in this invention are incorporated herein by reference as if each document were individually incorporated by reference. Furthermore, it should be understood that after reading the foregoing teachings of this invention, those skilled in the art can make various alterations or modifications to this invention, and these equivalent forms also fall within the scope defined by the appended claims.

Claims

1. A method of constructing an ERRa agonist prediction model, characterized by, The method comprises the steps of: (A) providing a compound data set comprising ERRa agonist data and non-ERRa agonist data; (B) converting each compound in the compound data set into a molecular fingerprint and a molecular descriptor, thereby obtaining a feature of each compound; (C) using the features for training of a plurality of machine learning models, thereby obtaining a plurality of base models and prediction results of the respective base models; (D) combining the prediction results of the respective base models to form an ensemble feature, training a logistic regression algorithm on the ensemble feature, thereby performing stacked ensemble of the plurality of base models, and finally obtaining a model for screening ERRa agonists, i.e. a consensus model.

2. A system for the prediction of ERRa agonists, characterized in that, The prediction system comprises: an input module configured to input data, the data comprising compound data; a prediction module configured as a prediction model, the prediction model being configured to perform prediction on the compound data, thereby obtaining prediction results; the prediction model being constructed by the method of claim 1; an output module configured to output the prediction results of the prediction module.

3. A method of multi-stage screening for ERRa agonists, characterized in that, The method comprises the steps of: (1) providing a plurality of candidate compounds; (2) using the consensus model to perform prediction on the candidate compounds, and selecting, according to the prediction results, compounds reaching a predetermined threshold as a round of screening compounds; the consensus model being constructed by the method of claim 1; (3) using a pre-constructed multi-ligand pharmacophore model to evaluate the round of screening compounds, and selecting, according to the evaluation results, the round of screening compounds reaching a predetermined level as a second round of screening compounds; (4) using a combination of pre-constructed molecular docking accuracy and conformation filtering rules to evaluate the second round of screening compounds, and selecting, according to the evaluation results, the second round of screening compounds reaching a predetermined level as a third round of screening compounds; (5) comprehensively evaluating the third round of screening compounds according to compound properties, and selecting, according to the evaluation results, the third round of screening compounds reaching a predetermined level as ERRa agonists.

4. The method of claim 3, wherein, The pre-constructed multi-ligand pharmacophore model is constructed according to a known binding mode of ERRa agonists and ERRa receptors; the combination of pre-constructed molecular docking accuracy and conformation filtering rules is constructed according to a known binding mode of ERRa agonists and ERRa receptors; and the known binding mode of ERRa agonists and ERRa receptors is obtained by molecular docking using known ERRa agonists and ERRa protein crystal structure, and is induced according to secondary bond interaction analysis.

5. A system for multi-stage screening of ERRa agonists, characterized in that, The system comprises: an input module configured to input data, the data comprising compound data; a screening module comprising: (a1) a first round of screening unit configured as a prediction model, the prediction model being configured to perform prediction on the compound data, and to select, according to the prediction results, compounds reaching a predetermined threshold as a first round of screening compounds; the prediction model being constructed by the method of claim 1; (a2) a second round screening unit configured to perform the following operation: evaluating the first round screening compounds by using a pre-constructed multi-ligand pharmacophore model, and selecting the first round screening compounds reaching a predetermined level as second round screening compounds according to the evaluation results; (a3) a third round screening unit configured to perform the following operation: evaluating the second round screening compounds by using a pre-constructed combination of molecular docking accuracy and conformational filtering rules, and selecting the second round screening compounds reaching a predetermined level as third round screening compounds according to the evaluation results; (a4) a fourth round screening unit configured to perform the following operation: comprehensively evaluating the third round screening compounds according to compound properties, and selecting the third round screening compounds reaching a predetermined level as ERRa agonists according to the evaluation results, thereby obtaining a screening result; an output module configured to output the screening result of the screening module.

6. A compound obtained by the method of claim 3, characterized in that, The compound has a structure as shown in formula (I):

7. A medicament or pharmaceutical composition, characterized by, The pharmaceutical or pharmaceutical composition comprises the compound of claim 6, and a pharmaceutically acceptable carrier, diluent or excipient.

8. The medicament or pharmaceutical composition of claim 7, wherein The pharmaceutical or pharmaceutical composition comprises an ERRa agonist.

9. Use of a compound obtained by the method of claim 3, characterized in that, A medicament, a pharmaceutical composition and / or a kit for preparing a medicament for treating or preventing diabetes.

10. Use according to claim 9, characterized in that, The compound comprises N-(3-ethylphenyl)-6H,7H,8H,9H-pyrido[2,1-h]purin-4-amine; the N-(3-ethylphenyl)-6H,7H,8H,9H-pyrido[2,1-h]purin-4-amine has a structure as shown in formula (I): The diabetes is type 2 diabetes.

Images