Search

Machine learning is a well-known approach for virtual screening. Recently, deep learning, a machine learning algorithm in artificial neural networks, has been applied to the advancement of precision medicine and drug discovery. In this study, we performed comparative studies between deep neural networks (DNN) and other ligand-based virtual screening (LBVS) methods to demonstrate that DNN and random forest (RF) were superior in hit prediction efficiency. By using DNN, several triple-negative breast cancer (TNBC) inhibitors were identified as potent hits from a screening of an in-house database of 165,000 compounds. In broadening the application of this method, we harnessed the predictive properties of trained model in the discovery of G protein-coupled receptor (GPCR) agonist, by which computational structure-based design of molecules could be greatly hindered by lack of structural information. Notably, a potent (~ 500 nM) mu-opioid receptor (MOR) agonist was identified as a hit from a small-size training set of 63 compounds. Our results show that DNN could be an efficient module in hit prediction and provide experimental evidence that machine learning could identify potent hits in silico from a limited training set.

We report that compound 13 , a novel phosphatidylserine-targeting zinc(II) dipicolylamine drug conjugate, readily triggers a positive feedback therapeutic loop through the in situ generation of phosphatidylserine in the tumor microenvironment. Linker modifications, pharmacokinetics profiling, in vivo antitumor studies, and micro-Western array of treated-tumor tissues were employed to show that this class of conjugates induced regeneration of apoptotic signals, which facilitated subsequent recruitment of the circulating conjugates through the zinc(II) dipicolylamine-phosphatidylserine association and resulted in compounding antitumor efficacy. Compared to the marketed compound 17 , compound 13 not only induced regressions in colorectal and pancreatic tumor models, it also exhibited at least 5-fold enhancement in antitumor efficacy with only 40% of the drug employed during treatment, culminating in a >12.5-fold increase in therapeutic potential. Our study discloses a chemically distinct apoptosis-targeting theranostic, with built-in complementary functional moieties between the targeting module and the drug mechanism to expand the arsenal of antitumor therapy.

A series of zinc(II) dipicolylamine (ZnDPA)-based drug conjugates have been synthesized to probe the potential of phosphatidylserine (PS) as a new antigen for small molecule drug conjugate (SMDC) development. Using in vitro cytotoxicity and plasma stability studies, PS-binding assay, in vivo pharmacokinetic studies, and maximum tolerated dose profiles, we provided a roadmap and the key parameters required for the development of the ZnDPA based drug conjugate. In particular, conjugate 24 induced tumor regression in the COLO 205 xenograft model and exhibited a more potent antitumor effect with a 70% reduction of cytotoxic payload compared to that of the marketed irinotecan when dosed at the same regimen. In addition to the validation of PS as an effective pharmacodelivery target for SMDC, our work also provided the foundation that, if applicable, a variety of therapeutic agents could be conjugated in the same manner to treat other PS-associated diseases.