Identification of Novel Antagonists Targeting Cannabinoid Receptor 2 Using a Multi-Step Virtual Screening Strategy

The endocannabinoid system plays an essential role in the regulation of analgesia and human immunity, and Cannabinoid Receptor 2 (CB2) has been proved to be an ideal target for the treatment of liver diseases and some cancers. In this study, we identified CB2 antagonists using a three-step “deep learning–pharmacophore–molecular docking” virtual screening approach. From the ChemDiv database (1,178,506 compounds), 15 hits were selected and tested by radioligand binding assays and cAMP functional assays. A total of 7 out of the 15 hits were found to exhibit binding affinities in the radioligand binding assays against CB2 receptor, with a pK_i of 5.15–6.66, among which five compounds showed antagonistic activities with pIC₅₀ of 5.25–6.93 in the cAMP functional assays. Among these hits, Compound 8 with the 4H-pyrido[1,2-a]pyrimidin-4-one scaffold showed the best binding affinity and antagonistic activity with a pK_i of 6.66 and pIC₅₀ of 6.93, respectively. The new scaffold could serve as a lead for further development of CB2 drugs. Additionally, we hope that the model in this study could be further utilized to identify more novel CB2 receptor antagonists, and the developed approach could also be used to design potent ligands for other therapeutic targets.     

Structure-Based Discovery of ABCG2 Inhibitors: A Homology Protein-Based Pharmacophore Modeling and Molecular Docking Approach

ABCG2 is an ABC membrane protein reverse transport pump, which removes toxic substances such as medicines out of cells. As a result, drug bioavailability is an unexpected change and negatively influences the ADMET (absorption, distribution, metabolism, excretion, and toxicity), leading to multi-drug resistance (MDR). Currently, in spite of promising studies, screening for ABCG2 inhibitors showed modest results. The aim of this study was to search for small molecules that could inhibit the ABCG2 pump. We first used the WISS MODEL automatic server to build up ABCG2 homology protein from 655 amino acids. Pharmacophore models, which were con-structed based on strong ABCG2 inhibitors (IC₅₀ < 1 μM), consist of two hydrophobic (Hyd) groups, two hydrogen bonding acceptors (Acc2), and an aromatic or conjugated ring (Aro|PiR). Using molecular docking method, 714 substances from the DrugBank and 837 substances from the TCM with potential to inhibit the ABCG2 were obtained. These chemicals maybe favor synthesized or extracted and bioactivity testing.     

Structure-based model profiles affinity constant of drugs with hPEPT1 for rapid virtual screening of hPEPT1’s substrate

The human proton-coupled peptide transporter (hPEPT1) with broad substrates is an important route for improving the pharmacokinetic performance of drugs. Thus, it is essential to predict the affinity constant between drug molecule and hPEPT1 for rapid virtual screening of hPEPT1’s substrate during lead optimization, candidate selection and hPEPT1 prodrug design. Here, a structure-based in silico model for 114 compounds was constructed based on eight structural parameters. This model was built by the multiple linear regression method and satisfied all the prerequisites of the regression models. For the entire data set, the r² and adjusted r² values were 0.74 and 0.72, respectively. Then, this model was used to perform substrate/non-substrate classification. For 29 drugs from DrugBank database, all were correctly classified as substrates of hPEPT1. This model was also used to perform substrate/non-substrate classification for 18 drugs and their prodrugs; this QSAR model also can distinguish between the substrate and non-substrate. In conclusion, the QSAR model in this paper was validated by a large external data set, and all results indicated that the developed model was robust, stable, and can be used for rapid virtual screening of hPEPT1’s substrate in the early stage of drug discovery.     

Machine learning-, rule- and pharmacophore-based classification on the inhibition of P-glycoprotein and NorA

The efflux pumps P-glycoprotein (P-gp) in humans and NorA in Staphylococcus aureus are of great interest for medicinal chemists because of their important roles in multidrug resistance (MDR). The high polyspecificity as well as the unavailability of high-resolution X-ray crystal structures of these transmembrane proteins lead us to combining ligand-based approaches, which in the case of this study were machine learning, perceptual mapping and pharmacophore modelling. For P-gp inhibitory activity, individual models were developed using different machine learning algorithms and subsequently combined into an ensemble model which showed a good discrimination between inhibitors and noninhibitors (acc_{train-diverse} = 84%; acc_{internal-test} = 92% and acc_{external-test} = 100%). For ligand promiscuity between P-gp and NorA, perceptual maps and pharmacophore models were generated for the detection of rules and features. Based on these in silico tools, hit compounds for reversing MDR were discovered from the in-house and DrugBank databases through virtual screening in an attempt to restore drug sensitivity in cancer cells and bacteria.     

Drug Repurposing for Influenza Virus Polymerase Acidic (PA) Endonuclease Inhibitor

Drug repurposing can quickly and effectively identify novel drug repurposing opportunities. The PA endonuclease catalytic site has recently become regarded as an attractive target for the screening of anti-influenza drugs. PA N-terminal (PA_N) inhibitor can inhibit the entire PA endonuclease activity. In this study, we screened the effectivity of PA_N inhibitors from the FDA database through in silico methods and in vitro experiments. PA_N and mutant PA_N-I38T were chosen as virtual screening targets for overcoming drug resistance. Gel-based PA endonuclease analysis determined that the drug lifitegrast can effectively inhibit PA_N and PA_N-I38T, when the IC₅₀ is 32.82 ± 1.34 μM and 26.81 ± 1.2 μM, respectively. Molecular docking calculation showed that lifitegrast interacted with the residues around PA or PA-I38 T’s active site, occupying the catalytic site pocket. Both PA_N/PA_N-I38T and lifitegrast can acquire good equilibrium in 100 ns molecular dynamic simulation. Because of these properties, lifitegrast, which can effectively inhibit PA endonuclease activity, was screened through in silico and in vitro research. This new research will be of significance in developing more effective and selective drugs for anti-influenza therapy.     

A Knowledge-Based Approach for Identification of Drugs Against Vivapain-2 Protein of Plasmodium vivax Through Pharmacophore-Based Virtual Screening with Comparative Modelling

Malaria is one of the most infectious diseases in the world. Plasmodium vivax, the pathogen causing endemic malaria in humans worldwide, is responsible for extensive disease morbidity. Due to the emergence of resistance to common anti-malarial drugs, there is a continuous need to develop a new class of drugs for this pathogen. P. vivax cysteine protease, also known as vivapain-2, plays an important role in haemoglobin hydrolysis and is considered essential for the survival of the parasite. The three-dimensional (3D) structure of vivapain-2 is not predicted experimentally, so its structure is modelled by using comparative modelling approach and further validated by Qualitative Model Energy Analysis (QMEAN) and RAMPAGE tools. The potential binding site of selected vivapain-2 structure has been detected by grid-based function prediction method. Drug targets and their respective drugs similar to vivapain-2 have been identified using three publicly available databases: STITCH 3.1, DrugBank and Therapeutic Target Database (TTD). The second approach of this work focuses on docking study of selected drug E-64 against vivapain-2 protein. Docking reveals crucial information about key residues (Asn281, Cys283, Val396 and Asp398) that are responsible for holding the ligand in the active site. The similarity-search criterion is used for the preparation of our in-house database of drugs, obtained from filtering the drugs from the DrugBank database. A five-point 3D pharmacophore model is generated for the docked complex of vivapain-2 with E-64. This study of 3D pharmacophore-based virtual screening results in identifying three new drugs, amongst which one is approved and the other two are experimentally proved. The ADMET properties of these drugs are found to be in the desired range. These drugs with novel scaffolds may act as potent drugs for treating malaria caused by P. vivax.     

Evaluation of Neurotropic Activity and Molecular Docking Study of New Derivatives of pyrano[4″,3″:4′,5′]pyrido[3′,2′:4,5]thieno[3,2-d]pyrimidines on the Basis of pyrano[3,4-c]pyridines

Background: Heterocyclic compounds and their fused analogs, which contain pharmacophore fragments such as pyridine, thiophene and pyrimidine rings, are of great interest due to their broad spectrum of biological activity. Chemical compounds containing two or more pharmacophore groups due to additional interactions with active receptor centers usually enhance biological activity and can even lead to a new type of activity. The search for new effective neurotropic drugs in the series of derivatives of heterocycles containing pharmacophore groups in organic, bioorganic and medical chemistry is a serious problem. Methods: Modern methodology of drugs involves synthesis, physicochemical study, molecular modeling and selection of active compounds through virtual screening and experimental evaluation of the biological activity of new chimeric compounds with pharmacophore fragments. For the synthesis of new compounds, classical organic methods were used and developed. For the evaluation of neurotropic activity of new synthesized compounds, some biological methods were used according to indicators characterizing anticonvulsant, sedative and antianxiety activity as well as side effects. For docking analysis, various soft ware packages and methods were used. Results: As a result of multistep reactions, 11 new, tri- and tetracyclic heterocyclic systems were obtained. The studied compounds exhibit protection against pentylenetetrazole (PTZ) seizures as well as some psychotropic effects. The biological assays evidenced that nine of the eleven studied compounds showed a high anticonvulsant activity by antagonism with pentylenetetrazole. The toxicity of the compounds is low, and they do not induce muscle relaxation in the studied doses. According to the study of psychotropic activity, it was found that the selected compounds have an activating behavior and anxiolytic effects on the “open field” and “elevated plus maze” (EPM) models. The data obtained indicate the anxiolytic (antianxiety) activity of the derivatives of tricyclic thieno[2,3-b]pyridines and tetracyclic pyridothieno[3,2-d]pyrimidin-8-ones, especially pronounced in compounds 3b–f and 4e. The studied compounds increase the latent time of first immobilization on the “forced swimming” (FS) model and exhibit antidepressant effects; compounds 3e and 3f especially exhibit these effects, similarly to diazepam. Docking studies revealed that compounds 3c and 4b bound tightly in the active site of γ-aminobutyric acid type A (GABA_A) receptors with a value of the scoring function that estimates free energy of binding (∆G) at −10.0 ± 5 kcal/mol. Compound 4e showed the best affinity ((∆G) at −11.0 ± 0.54 kcal/mol) and seems to be an inhibitor of serotonin (SERT) transporter. Compounds 3c–f and 4e practically bound with the groove of T4L of 5HT_1A and blocked it completely, while the best affinity observed was in compound 3f ((∆G) at −9.3 ± 0.46 kcal/mol). Conclusions: The selected compounds have an anticonvulsant, activating behavior and anxiolytic effects and at the same time exhibit antidepressant effects.     

From Combinations to Single-Molecule Polypharmacology—Cromolyn-Ibuprofen Conjugates for Alzheimer’s Disease

Despite Alzheimer’s disease (AD) incidence being projected to increase worldwide, the drugs currently on the market can only mitigate symptoms. Considering the failures of the classical paradigm “one target-one drug-one disease” in delivering effective medications for AD, polypharmacology appears to be a most viable therapeutic strategy. Polypharmacology can involve combinations of multiple drugs and/or single chemical entities modulating multiple targets. Taking inspiration from an ongoing clinical trial, this work aims to convert a promising cromolyn–ibuprofen drug combination into single-molecule “codrugs.” Such codrugs should be able to similarly modulate neuroinflammatory and amyloid pathways, while showing peculiar pros and cons. By exploiting a linking strategy, we designed and synthesized a small set of cromolyn–ibuprofen conjugates (4–6). Preliminary plasma stability and neurotoxicity assays allowed us to select diamide 5 and ethanolamide 6 as promising compounds for further studies. We investigated their immunomodulatory profile in immortalized microglia cells, in vitro anti-aggregating activity towards Aβ₄₂-amyloid self-aggregation, and their cellular neuroprotective effect against Aβ₄₂-induced neurotoxicity. The fact that 6 effectively reduced Aβ-induced neuronal death, prompted its investigation into an in vivo model. Notably, 6 was demonstrated to significantly increase the longevity of Aβ₄₂-expressing Drosophila and to improve fly locomotor performance.     

Deep Learning and Structure-Based Virtual Screening for Drug Discovery against NEK7: A Novel Target for the Treatment of Cancer

NIMA-related kinase7 (NEK7) plays a multifunctional role in cell division and NLRP3 inflammasone activation. A typical expression or any mutation in the genetic makeup of NEK7 leads to the development of cancer malignancies and fatal inflammatory disease, i.e., breast cancer, non-small cell lung cancer, gout, rheumatoid arthritis, and liver cirrhosis. Therefore, NEK7 is a promising target for drug development against various cancer malignancies. The combination of drug repurposing and structure-based virtual screening of large libraries of compounds has dramatically improved the development of anticancer drugs. The current study focused on the virtual screening of 1200 benzene sulphonamide derivatives retrieved from the PubChem database by selecting and docking validation of the crystal structure of NEK7 protein (PDB ID: 2WQN). The compounds library was subjected to virtual screening using Auto Dock Vina. The binding energies of screened compounds were compared to standard Dabrafenib. In particular, compound 762 exhibited excellent binding energy of −42.67 kJ/mol, better than Dabrafenib (−33.89 kJ/mol). Selected drug candidates showed a reactive profile that was comparable to standard Dabrafenib. To characterize the stability of protein–ligand complexes, molecular dynamic simulations were performed, providing insight into the molecular interactions. The NEK7–Dabrafenib complex showed stability throughout the simulated trajectory. In addition, binding affinities, pIC50, and ADMET profiles of drug candidates were predicted using deep learning models. Deep learning models predicted the binding affinity of compound 762 best among all derivatives, which supports the findings of virtual screening. These findings suggest that top hits can serve as potential inhibitors of NEK7. Moreover, it is recommended to explore the inhibitory potential of identified hits compounds through in-vitro and in-vivo approaches.     

RNA-Targeting Splicing Modifiers: Drug Development and Screening Assays

RNA splicing is an essential step in producing mature messenger RNA (mRNA) and other RNA species. Harnessing RNA splicing modifiers as a new pharmacological modality is promising for the treatment of diseases caused by aberrant splicing. This drug modality can be used for infectious diseases by disrupting the splicing of essential pathogenic genes. Several antisense oligonucleotide splicing modifiers were approved by the U.S. Food and Drug Administration (FDA) for the treatment of spinal muscular atrophy (SMA) and Duchenne muscular dystrophy (DMD). Recently, a small-molecule splicing modifier, risdiplam, was also approved for the treatment of SMA, highlighting small molecules as important warheads in the arsenal for regulating RNA splicing. The cellular targets of these approved drugs are all mRNA precursors (pre-mRNAs) in human cells. The development of novel RNA-targeting splicing modifiers can not only expand the scope of drug targets to include many previously considered “undruggable” genes but also enrich the chemical-genetic toolbox for basic biomedical research. In this review, we summarized known splicing modifiers, screening methods for novel splicing modifiers, and the chemical space occupied by the small-molecule splicing modifiers.     

Deciphering the Interactions of Bioactive Compounds in Selected Traditional Medicinal Plants against Alzheimer’s Diseases via Pharmacophore Modeling,Auto-QSAR,and Molecular Docking Approaches

Neurodegenerative diseases, for example Alzheimer’s, are perceived as driven by hereditary, cellular, and multifaceted biochemical actions. Numerous plant products, for example flavonoids, are documented in studies for having the ability to pass the blood-brain barrier and moderate the development of such illnesses. Computer-aided drug design (CADD) has achieved importance in the drug discovery world; innovative developments in the aspects of structure identification and characterization, bio-computational science, and molecular biology have added to the preparation of new medications towards these ailments. In this study we evaluated nine flavonoid compounds identified from three medicinal plants, namely T. diversifolia, B. sapida, and I. gabonensis for their inhibitory role on acetylcholinesterase (AChE), butyrylcholinesterase (BChE) and monoamine oxidase (MAO) activity, using pharmacophore modeling, auto-QSAR prediction, and molecular studies, in comparison with standard drugs. The results indicated that the pharmacophore models produced from structures of AChE, BChE and MAO could identify the active compounds, with a recuperation rate of the actives found near 100% in the complete ranked decoy database. Moreso, the robustness of the virtual screening method was accessed by well-established methods including enrichment factor (EF), receiver operating characteristic curve (ROC), Boltzmann-enhanced discrimination of receiver operating characteristic (BEDROC), and area under accumulation curve (AUAC). Most notably, the compounds’ pIC₅₀ values were predicted by a machine learning-based model generated by the AutoQSAR algorithm. The generated model was validated to affirm its predictive model. The best models achieved for AChE, BChE and MAO were models kpls_radial_17 (R² = 0.86 and Q² = 0.73), pls_38 (R² = 0.77 and Q² = 0.72), kpls_desc_44 (R² = 0.81 and Q² = 0.81) and these externally validated models were utilized to predict the bioactivities of the lead compounds. The binding affinity results of the ligands against the three selected targets revealed that luteolin displayed the highest affinity score of −9.60 kcal/mol, closely followed by apigenin and ellagic acid with docking scores of −9.60 and −9.53 kcal/mol, respectively. The least binding affinity was attained by gallic acid (−6.30 kcal/mol). The docking scores of our standards were −10.40 and −7.93 kcal/mol for donepezil and galanthamine, respectively. The toxicity prediction revealed that none of the flavonoids presented toxicity and they all had good absorption parameters for the analyzed targets. Hence, these compounds can be considered as likely leads for drug improvement against the same.     

An Insight into the Medicinal Chemistry Perspective of Macrocyclic Derivatives with Antitumor Activity: A Systematic Review

The profound pharmacological properties of macrocyclic compounds have led to their development as drugs. In conformationally pre-organized ring structures, the multiple functions and stereochemical complexity provided by the macrocycle result in high affinity and selectivity of protein targets while maintaining sufficient bioavailability to reach intracellular locations. Therefore, the construction of macrocycles is an ideal choice to solve the problem of “undruggable” targets. Inspection of 68 macrocyclic drugs on the market showed that 10 of them were used to treat cancer, but this structural class still has been poorly explored within drug discovery. This perspective considers the macrocyclic compounds used for anti-tumor with different targets, their advantages and disadvantages, and the various synthetic methods of them.     

Synthesis,Structural Characterization,and In Vitro and In Silico Antifungal Evaluation of Azo-Azomethine Pyrazoles (PhN2(PhOH)CHN(C3N2(CH3)3)PhR,R = H or NO2)

The azo-azomethine imines, R¹-N=N-R²-CH=N-R³, are a class of active pharmacological ligands that have been prominent antifungal, antibacterial, and antitumor agents. In this study, four new azo-azomethines, R¹ = Ph, R² = phenol, and R³ = pyrazol-Ph-R’ (R = H or NO₂), have been synthesized, structurally characterized using X-ray, IR, NMR and UV–Vis techniques, and their antifungal activity evaluated against certified strains of Candida albicans and Cryptococcus neoformans. The antifungal tests revealed a high to moderate inhibitory activity towards both strains, which is regulated as a function of both the presence and the location of the nitro group in the aromatic ring of the series. These biological assays were further complemented with molecular docking studies against three different molecular targets from each fungus strain. Molecular dynamics simulations and binding free energy calculations were performed on the two best molecular docking results for each fungus strain. Better affinity for active sites for nitro compounds at the “meta” and “para” positions was found, making them promising building blocks for the development of new Schiff bases with high antifungal activity.     

Natural Products Extracted from Fungal Species as New Potential Anti-Cancer Drugs: A Structure-Based Drug Repurposing Approach Targeting HDAC7

Mushrooms can be considered a valuable source of natural bioactive compounds with potential polypharmacological effects due to their proven antimicrobial, antiviral, antitumor, and antioxidant activities. In order to identify new potential anticancer compounds, an in-house chemical database of molecules extracted from both edible and non-edible fungal species was employed in a virtual screening against the isoform 7 of the Histone deacetylase (HDAC). This target is known to be implicated in different cancer processes, and in particular in both breast and ovarian tumors. In this work, we proposed the ibotenic acid as lead compound for the development of novel HDAC7 inhibitors, due to its antiproliferative activity in human breast cancer cells (MCF-7). These promising results represent the starting point for the discovery and the optimization of new HDAC7 inhibitors and highlight the interesting opportunity to apply the “drug repositioning” paradigm also to natural compounds deriving from mushrooms.     

Determination of Anthelmintic and Antiprotozoal Drug Residues in Fish Using Liquid Chromatography-Tandem Mass Spectrometry

The objective of this study is to develop a comprehensive and simple method for the simultaneous determination of anthelmintic and antiprotozoal drug residues in fish. For sample preparation, we used the “quick, easy, cheap, effective, rugged, and safe” (QuEChERS) method with a simple modification. The sample was extracted with water and 1% formic acid in acetonitrile/methanol (MeCN/MeOH) (95:5, v/v), followed by phase separation (salting out) with MgSO₄ and NaCl (4:1, w/w). After centrifugation, an aliquot of the extract was purified by dispersive solid-phase extraction (d-SPE) prior to liquid chromatography-tandem mass spectrometry (LC-MS/MS) analysis. The method was validated at three concentration levels for all matrices, in accordance with the Codex guidelines (CAC/GL-71). Quantitative analysis was performed using the method of matrix-matched calibration. The recoveries were between 60.6% and 119.9%, with coefficients of variation (CV) <30% for all matrices. The limit of quantitation (LOQ) of the method ranged from 0.02 μg kg⁻¹ to 4.8 μg kg⁻¹ for all matrices. This comprehensive method can be used for the investigation of both anthelmintic and antiprotozoal drugs belonging to different chemical families in fishery products.     

New Potential Pharmacological Targets of Plant-Derived Hydroxyanthraquinones from Rubia spp.

The increased use of polyphenols nowadays poses the need for identification of their new pharmacological targets. Recently, structure similarity-based virtual screening of DrugBank outlined pseudopurpurin, a hydroxyanthraquinone from Rubia cordifolia spp., as similar to gatifloxacin, a synthetic antibacterial agent. This suggested the bacterial DNA gyrase and DNA topoisomerase IV as potential pharmacological targets of pseudopurpurin. In this study, estimation of structural similarity to referent antibacterial agents and molecular docking in the DNA gyrase and DNA topoisomerase IV complexes were performed for a homologous series of four hydroxyanthraquinones. Estimation of shape- and chemical feature-based similarity with (S)-gatifloxacin, a DNA gyrase inhibitor, and (S)-levofloxacin, a DNA topoisomerase IV inhibitor, outlined pseudopurpurin and munjistin as the most similar structures. The docking simulations supported the hypothesis for a plausible antibacterial activity of hydroxyanthraquinones. The predicted docking poses were grouped into 13 binding modes based on spatial similarities in the active site. The simultaneous presence of 1-OH and 3-COOH substituents in the anthraquinone scaffold were emphasized as relevant features for the binding modes’ variability and ability of the compounds to strongly bind in the DNA-enzyme complexes. The results reveal new potential pharmacological targets of the studied polyphenols and help in their prioritization as drug candidates and dietary supplements.     

Metal oxidation states in biological water splitting

A central question in biological water splitting concerns the oxidation states of the manganese ions that comprise the oxygen-evolving complex of photosystem II. Understanding the nature and order of oxidation events that occur during the catalytic cycle of five S_i states (i = 0–4) is of fundamental importance both for the natural system and for artificial water oxidation catalysts. Despite the widespread adoption of the so-called “high-valent scheme”—where, for example, the Mn oxidation states in the S₂ state are assigned as III, IV, IV, IV—the competing “low-valent scheme” that differs by a total of two metal unpaired electrons (i.e. III, III, III, IV in the S₂ state) is favored by several recent studies for the biological catalyst. The question of the correct oxidation state assignment is addressed here by a detailed computational comparison of the two schemes using a common structural platform and theoretical approach. Models based on crystallographic constraints were constructed for all conceivable oxidation state assignments in the four (semi)stable S states of the oxygen evolving complex, sampling various protonation levels and patterns to ensure comprehensive coverage. The models are evaluated with respect to their geometric, energetic, electronic, and spectroscopic properties against available experimental EXAFS, XFEL-XRD, EPR, ENDOR and Mn K pre-edge XANES data. New 2.5 K ⁵⁵Mn ENDOR data of the S₂ state are also reported. Our results conclusively show that the entire S state phenomenology can only be accommodated within the high-valent scheme by adopting a single motif and protonation pattern that progresses smoothly from S₀ (III, III, III, IV) to S₃ (IV, IV, IV, IV), satisfying all experimental constraints and reproducing all observables. By contrast, it was impossible to construct a consistent cycle based on the low-valent scheme for all S states. Instead, the low-valent models developed here may provide new insight into the over-reduced S states and the states involved in the assembly of the catalytically active water oxidizing cluster.     

Virtual screening of dipeptidyl peptidase-4 inhibitors using quantitative structure–activity relationship-based artificial intelligence and molecular docking of hit compounds

Dipeptidyl peptidase-4 (DPP-4) inhibitors are becoming an essential drug in the treatment of type 2 diabetes mellitus; however, some classes of these drugs exert side effects, including joint pain and pancreatitis. Studies suggest that these side effects might be related to secondary inhibition of DPP-8 and DPP-9. In this study, we identified DPP-4-inhibitor hit compounds selective against DPP-8 and DPP-9. We built a virtual screening workflow using a quantitative structure–activity relationship (QSAR) strategy based on artificial intelligence to allow faster screening of millions of molecules for the DPP-4 target relative to other screening methods. Five regression machine learning algorithms and four classification machine learning algorithms were applied to build virtual screening workflows, with the QSAR model applied using support vector regression (R²_pred 0.78) and the classification QSAR model using the random forest algorithm with 92.2% accuracy. Virtual screening results of > 10 million molecules obtained 2 716 hits compounds with a pIC₅₀ value of > 7.5. Additionally, molecular docking results of several potential hit compounds for DPP-4, DPP-8, and DPP-9 identified CH0002 as showing high inhibitory potential against DPP-4 and low inhibitory potential for DPP-8 and DPP-9 enzymes. These results demonstrated the effectiveness of this technique for identifying DPP-4-inhibitor hit compounds selective for DPP-4 and against DPP-8 and DPP-9 and suggest its potential efficacy for applications to discover hit compounds of other targets.