Docking Score ML: Target-Specific Machine Learning Models Improving Docking-Based Virtual Screening in 155 Targets.

In drug discovery, molecular docking methods face challenges in accurately predicting energy. Scoring functions used in molecular docking often fail to simulate complex protein-ligand interactions fully and accurately leading to biases and inaccuracies in virtual screening and target predictions. We introduce the "Docking Score ML", developed from an analysis of over 200,000 docked complexes from 155 known targets for cancer treatments. The scoring functions used are founded on bioactivity data sourced from ChEMBL and have been fine-tuned using both supervised machine learning and deep learning techniques. We validated our approach extensively using multiple data sets such as validation of selectivity mechanism, the DUDE, DUD-AD, and LIT-PCBA data sets, and performed a multitarget analysis on drugs like sunitinib. To enhance prediction accuracy, feature fusion techniques were explored. By merging the capabilities of the Graph Convolutional Network (GCN) with multiple docking functions, our results indicated a clear superiority of our methodologies over conventional approaches. These advantages demonstrate that Docking Score ML is an efficient and accurate tool for virtual screening and reverse docking.

Design, synthesis, pharmacological evaluation, and computational study of benzo[d] isothiazol-based small molecule inhibitors targeting PD-1/PD-l1 interaction.

Blockade of the programmed cell death-1 (PD-1)/programmed cell death ligand 1 (PD-L1) pathway is an attractive strategy for immunotherapy, but the clinical application of small molecule PD-1/PD-L1 inhibitors remains unclear. In this work, based on BMS-202 and our previous work YLW-106, a series of compounds with benzo[d]isothiazol structure as scaffold were designed and synthesized. Their inhibitory activity against PD-1/PD-L1 interaction was evaluated by a homogeneous time-resolved fluorescence (HTRF) assay. Among them, LLW-018 (27c) exhibited the most potent inhibitory activity with an IC50 value of 2.61 nM. The cellular level assays demonstrated that LLW-018 exhibited low cytotoxicity against Jurkat T and MDA-MB-231. Further cell-based PD-1/PD-L1 blockade bioassays based on PD-1 NFAT-Luc Jurkat cells and PD-L1 TCR Activator CHO cells indicated that LLW-018 could interrupt PD-1/PD-L1 interaction with an IC50 value of 0.88 µM. Multi-computational methods, including molecular docking, molecular dynamics, MM/GBSA, MM/PBSA, Metadynamics, and QM/MM MD were utilized on PD-L1 dimer complexes, which revealed the binding modes and dissociation process of LLW-018 and C2-symmetric small molecule inhibitor LCH1307. These results suggested that LLW-018 exhibited promising potency as a PD-1/PD-L1 inhibitor for further investigation.

Application of molecular dynamics-based pharmacophore and machine learning approaches to identify novel Mcl1 inhibitors through drug repurposing and mechanics research.

Myeloid cell leukemia 1 (Mcl1), a critical protein that regulates apoptosis, has been considered as a promising target for antitumor drugs. The conventional pharmacophore screening approach has limitations in conformation sampling and data mining. Here, we offered an innovative solution to identify Mcl1 inhibitors with molecular dynamics-refined pharmacophore and machine learning methods. Considering the safety and druggability of FDA-approved drugs, virtual screening of the database was performed to discover Mcl1 inhibitors, and the hit was subsequently validated via TR-FRET, cytotoxicity, and flow cytometry assays. To reveal the binding characteristics shared by the hit and a typical Mcl1 selective inhibitor, we employed quantum mechanics and molecular mechanics (QM/MM) calculations, umbrella sampling, and metadynamics in this work. The combined studies suggested that fluvastatin had promising cell inhibitory potency and was suitable for further investigation. We believe that this research will shed light on the discovery of novel Mcl1 inhibitors that can be used as a supplemental treatment against leukemia and provide a possible method to improve the accuracy of drug repurposing with limited computational resources while balancing the costs of experimentation well.

Structural requirement of RARγ agonism through computational aspects.

CONTEXT: RARγ is a therapeutic target for many skin diseases and has potential in cancer treatment. In the current study, we put forward a comprehensive structure-activity relationship study of third and fourth generations of RARγ agonists, addressing multiple crystal structures of RARγ complexes and approved drugs. Adapalene and Trifarotene, through hybrid strategies including protein contacts Atlas analysis, molecular docking, dynamics simulations, MM-GBSA, ASM, and pharmacophore modeling. Our result revealed crucial amino acids Arg267, Ser278, Phe288, Phe230, Met272, Leu271, and Leu268 within the RARγ pocket, as well as pharmacophore features such as two hydrophobic groups, two aromatic rings, and negative ionic features, which are essential for the binding of RARγ agonists. Based on this study, the binding mechanism of RARγ agonists was elucidated, which will be helpful for the rational design of new RARγ agonists for skin diseases and cancer treatment. METHODS: In this study, Schrödinger suite 2021-2 with OPLS_4 force field, Discovery Studio program 3.0, LigandScout 4.3, and PyMOL are utilized in the investigation.

Selectivity mechanism of BCL-XL/2 inhibition through in silico investigation.

The BCL-XL protein is among the most important members of the anti-apoptotic subfamily of the BCL-2 protein family, and is currently a promising new target for anti-tumor drug research. However, the BCL-XL/2 proteins have similar structures and functions, which could lead to undesirable side effects because of inhibitors that can bind to both BCL-XL and BCL-2. Therefore, it is crucial to expound on the structural basis of the selective mechanism towards BCL-XL/2 inhibition. In the current study, we employed hybrid computational methods including molecular docking and dynamics simulation, MM/GBSA energy calculation, alanine scanning mutagenesis and Hirshfeld surface analysis to comprehensively reveal the selectivity mechanism towards BCL-XL/2 from multiple perspectives, revealing the significant effects of the BCL-XL residues SER106 and LEU108 as well as the BCL-2 residue ASP103 on the inhibitory selectivity. Overall, our findings provide useful references for the rational design of BCL-XL/2 selective inhibitors with better affinity.

Discovery of a novel and selective cathepsin L inhibitor with anti-metastatic ability in vitro and in vivo against breast cancer cells.

Asperphenamate is a natural product that has attracted considerable interest from researchers worldwide. In the last decade, aiming to increase the biological activity and improve druggability, modifications of the A-ring moiety in asperphenamate have been performed. Our laboratory has also recently reported functional derivatizations of the A ring and studied its effect on the inhibition of cysteine cathepsin L. However, the functional significance of the B-ring fragment toward cathepsin L has not been evaluated thus far. In this paper, forty-four derivatives of the B-ring substituted with different N-phenylsulfonyl groups were designed and synthesized. Among them, the paratrifluromethyl analog B-2a and the 2, 4-difluoro-5-chloro derivative B-11b showed more potent inhibitory activity against cathepsin L than the control compound, ABR, which displayed the strongest inhibitory effect on cathepsin L and S among all reported asperphenamate derivatives. In particular, compound B-2a showed more pronounced selectivity against cathepsin L than the other derivatives. Molecular docking revealed that the N-phenylsulfonylamide moiety was vital for the interactions between B-2a and cathepsin L. Moreover, B-2a displayed no toxicity against normal cells. Therefore, compound B-2a was selected for further studies. Wound-healing assays, Transwell chamber assays and breast cancer lung metastasis mouse models demonstrated that B-2a exhibited antimetastatic ability in vitro and in vivo.

Diagnosis Accuracy and Prognostic Significance of the Dickkopf-1 Protein in Gastrointestinal Carcinomas: Systematic Review and Network Meta-analysis.

Objective: To evaluate the diagnosis accuracy and prognostic significance of bio-marker dickkopf-1(DKK-1) protein in GIC, and also sub-type of hepatocellular carcinoma (HCC), pancreas carcinomas (PC), oesophageal carcinoma (EPC) and Adenocarcinoma of esophago-gastric junction (AEGJ), etc. Methods: Electronic databases were searched from inception to May 2020. Patients were diagnosed with gastrointestinal carcinomas, and provided data on the correlation between high and low DKK-1 expression and diagnosis or prognosis. Results: Forty-three publications involving 9318 participants were included in the network meta-analysis, with 31 of them providing data for diagnosis value and 18 records were eligible for providing prognosis value of DKK-1. DKK-1 has a moderate diagnostic value for overall GIC, HCC and PC. In addition, for the combined diagnosis value of DKK-1 +AFP, high diagnostic accuracy value could be determined in HCC and early HCC group, respectively. Whereas, diagnosis efficiency of DKK-1+CA19-9 was also better than that of DKK-1 alone with AUC value is above 0.95. For the prognosis meta-analysis of histopathological stratification, we found that EPC and AEGJ ranked the best for the histopathological stratification of prognosis from network meta-analysis. This systematic review protocol was registered with the PROSPERO registry (No.CRD42020167910). Conclusion: DKK-1 has good diagnostic accuracy, especially combination of DKK-1+AFP in HCC and DKK-1+CA19-9 in PC, whereas modest prognostic significant in GIC. Future head-to-head researches are warranted for DKK-1 expression in HCC and PC tissue.