GcForest-based compound-protein interaction prediction model and its application in discovering small-molecule drugs targeting CD47.

Identifying compound-protein interaction plays a vital role in drug discovery. Artificial intelligence (AI), especially machine learning (ML) and deep learning (DL) algorithms, are playing increasingly important roles in compound-protein interaction (CPI) prediction. However, ML relies on learning from large sample data. And the CPI for specific target often has a small amount of data available. To overcome the dilemma, we propose a virtual screening model, in which word2vec is used as an embedding tool to generate low-dimensional vectors of SMILES of compounds and amino acid sequences of proteins, and the modified multi-grained cascade forest based gcForest is used as the classifier. This proposed method is capable of constructing a model from raw data, adjusting model complexity according to the scale of datasets, especially for small scale datasets, and is robust with few hyper-parameters and without over-fitting. We found that the proposed model is superior to other CPI prediction models and performs well on the constructed challenging dataset. We finally predicted 2 new inhibitors for clusters of differentiation 47(CD47) which has few known inhibitors. The IC50s of enzyme activities of these 2 new small molecular inhibitors targeting CD47-SIRPα interaction are 3.57 and 4.79 µM respectively. These results fully demonstrate the competence of this concise but efficient tool for CPI prediction.

Combined inhibition of pyruvate dehydrogenase kinase 1 and lactate dehydrogenase a induces metabolic and signaling reprogramming and enhances lung adenocarcinoma cell killing.

Lung adenocarcinoma (LUAD) is one of the most prevalent and aggressive types of lung cancer. Metabolic reprogramming plays a critical role in the development and progression of LUAD. Pyruvate dehydrogenase kinase 1 (PDK1) and lactate dehydrogenase A (LDHA) are two key enzymes involved in glucose metabolism, whilst their aberrant expressions are often associated with tumorigenesis. Herein, we investigated the anticancer effects of combined inhibition of PDK1 and LDHA in LUAD in vitro and in vivo and its underlying mechanisms of action. The combination of a PDK1 inhibitor, 64, and a LDHA inhibitor, NHI-Glc-2, led to a synergistic growth inhibition in 3 different LUAD cell lines and more than additively suppressed tumor growth in the LUAD xenograft H1975 model. This combination also inhibited cellular migration and colony formation, while it induced a metabolic shift from glycolysis to oxidative phosphorylation (OXPHOS) resulting in mitochondrial depolarization and apoptosis in LUAD cells. These effects were related to modulation of multiple cell signaling pathways, including AMPK, RAS/ERK, and AKT/mTOR. Our findings demonstrate that simultaneous inhibition of multiple glycolytic enzymes (PDK1 and LDHA) is a promising novel therapeutic approach for LUAD.

Epigallocatechin gallate circumvents drug-induced resistance in non-small-cell lung cancer by modulating glucose metabolism and AMPK/AKT/MAPK axis.

Upon prolonged use of epidermal growth factor receptor (EGFR) tyrosine kinase inhibitors (TKIs) in non-small-cell lung cancer (NSCLC), acquired drug resistance inevitably occurs. This study investigates the combined use of EGFR-TKIs (gefitinib or osimertinib) with epigallocatechin gallate (EGCG) to overcome acquired drug resistance in NSCLC models. The in vitro antiproliferative effects of EGFR-TKIs and EGCG combination in EGFR-mutant parental and resistant cell lines were evaluated. The in vivo efficacy of the combination was assessed in xenograft mouse models derived from EGFR-TKI-resistant NSCLC cells. We found that the combined use of EGFR-TKIs and EGCG significantly reversed the Warburg effect by suppressing glycolysis while boosting mitochondrial respiration, which was accompanied by increased cellular ROS and decreased lactate secretion. The combination effectively activated the AMPK pathway while inhibited both ERK/MAPK and AKT/mTOR pathways, leading to cell cycle arrest and apoptosis, particularly in drug-resistant NSCLC cells. The in vivo results obtained from mouse tumor xenograft model confirmed that EGCG effectively overcame osimertinib resistance. This study revealed that EGCG suppressed cancer bypass survival signaling and altered cancer metabolic profiles, which is a promising anticancer adjuvant of EGFR-TKIs to overcome acquired drug resistance in NSCLC.

Repurposing the Hedgehog pathway inhibitor, BMS-833923, as a phosphatidylglycerol-selective membrane-disruptive colistin adjuvant against ESKAPE pathogens.

The rapid emergence and spread of multi-drug- or pan-drug-resistant bacterial pathogens, such as ESKAPE, pose a serious threat to global health. However, the development of novel antibiotics is hindered by difficulties in identifying new antibiotic targets and the rapid development of drug resistance. Drug repurposing is an effective alternative strategy for combating antibiotic resistance that both saves resources and extends the life of existing antibiotics in combination treatment regimens. Screening of a chemical compound library identified BMS-833923 (BMS), a smoothened antagonist that kills Gram-positive bacteria directly, and potentiates colistin to destroy various Gram-negative bacteria. BMS did not induce detectable antibiotic resistance in vitro, and showed effective activity against drug-resistant bacteria in vivo. Mechanistic studies revealed that BMS caused membrane disruption by targeting the membrane phospholipids phosphatidylglycerol and cardiolipin, promoting membrane dysfunction, metabolic disturbance, leakage of cellular components, and, ultimately, cell death. This study describes a potential strategy to enhance the efficacy of colistin and combat multi-drug-resistant ESKAPE pathogens.

Dichloroacetophenone biphenylsulfone ethers as anticancer pyruvate dehydrogenase kinase inhibitors in non-small cell lung cancer models.

Pyruvate dehydrogenase kinase 1 (PDK1) is an important metabolic enzyme which is often overexpressed in many types of cancers, including non-small-cell lung cancers (NSCLC). Targeting PDK1 appears to be an attractive anticancer strategy. Based on a previously reported moderate potent anticancer PDK1 inhibitor, 64, we developed three dichloroacetophenone biphenylsulfone ethers, 30, 31 and 32, which showed strong PDK1 inhibitions of 74%, 83% and 72% at 10 µM, respectively. Then we investigated the anticancer effects of 31 in two NSCLC cell lines, namely, NCI-H1299 and NCI-H1975. It was found that 31 exhibited sub-micromolar cancer cell IC50s, suppressed colony formation, induced mitochondrial membrane potential depolarization, triggered apoptosis, altered cellular glucose metabolism, with concomitant reductions in extracellular lactate levels and enhanced the generation of reactive oxygen species in NSCLC cells. Moreover, 31 significantly suppressed the tumor growth in an NCI-H1975 mouse xenograft model, outperforming the anticancer effects of 64. Taken together our results suggested that inhibition of PDK1 via dichloroacetophenone biphenylsulfone ethers may provide a novel direction leading to an alternative treatment option in NSCLC therapy.

The development of small-molecule inhibitors targeting hexokinase 2.

As one of the well-known hallmarks of cancer malignancy, most proliferating cancer cells exhibit enhanced rates of glycolysis. Hexokinase 2 (HK2) is the rate-limiting enzyme catalyzing the first step of glycolysis, and is often overexpressed in most cancer cells. Thus, targeting HK2 appears to be a promising anticancer therapy. However, selective inhibition of HK2 and the polar nature of the target site remain challenges to the development of small-molecule inhibitors, which could be addressed by targeting unique domains of HK2, such as its N-terminal domain. Here, we review different target-inhibitor binding modes and the associated pharmacological effects, which would be informative for rational molecular design. We also highlight further perspectives and strategies to develop novel HK2 inhibitors for cancer therapy.

Convolutional Neural Network-based Virtual Screening.

Virtual screening is an important means for lead compound discovery. The scoring function is the key to selecting hit compounds. Many scoring functions are currently available; however, there are no all-purpose scoring functions because different scoring functions tend to have conflicting results. Recently, neural networks, especially convolutional neural networks, have constantly been penetrating drug design and most CNN-based virtual screening methods are superior to traditional docking methods, such as Dock and AutoDock. CNNbased virtual screening is expected to improve the previous model of overreliance on computational chemical screening. Utilizing the powerful learning ability of neural networks provides us with a new method for evaluating compounds. We review the latest progress of CNN-based virtual screening and propose prospects.