OpenChem: A Deep Learning Toolkit for Computational Chemistry and Drug Design.

Deep learning models have demonstrated outstanding results in many data-rich areas of research, such as computer vision and natural language processing. Currently, there is a rise of deep learning in computational chemistry and materials informatics, where deep learning could be effectively applied in modeling the relationship between chemical structures and their properties. With the immense growth of chemical and materials data, deep learning models can begin to outperform conventional machine learning techniques such as random forest, support vector machines, and nearest neighbor. Herein, we introduce OpenChem, a PyTorch-based deep learning toolkit for computational chemistry and drug design. OpenChem offers easy and fast model development, modular software design, and several data preprocessing modules. It is freely available via the GitHub repository.

Generative and reinforcement learning approaches for the automated de novo design of bioactive compounds.

Deep generative neural networks have been used increasingly in computational chemistry for de novo design of molecules with desired properties. Many deep learning approaches employ reinforcement learning for optimizing the target properties of the generated molecules. However, the success of this approach is often hampered by the problem of sparse rewards as the majority of the generated molecules are expectedly predicted as inactives. We propose several technical innovations to address this problem and improve the balance between exploration and exploitation modes in reinforcement learning. In a proof-of-concept study, we demonstrate the application of the deep generative recurrent neural network architecture enhanced by several proposed technical tricks to design inhibitors of the epidermal growth factor (EGFR) and further experimentally validate their potency. The proposed technical solutions are expected to substantially improve the success rate of finding novel bioactive compounds for specific biological targets using generative and reinforcement learning approaches.

Combined MEK and VEGFR inhibition in orthotopic human lung cancer models results in enhanced inhibition of tumor angiogenesis, growth, and metastasis.

PURPOSE: Ras/Raf/mitogen-activated protein-extracellular signal-regulated kinase (ERK) kinase (MEK)/ERK signaling is critical for tumor cell proliferation and survival. Selumetinib is a potent, selective, and orally available MEK1/2 inhibitor. In this study, we evaluated the therapeutic efficacy of selumetinib alone or with cediranib, an orally available potent inhibitor of all three VEGF receptor (VEGFR) tyrosine kinases, in murine orthotopic non-small cell lung carcinoma (NSCLC) models. EXPERIMENTAL DESIGN: NCI-H441 or NCI-H460 KRAS-mutant human NSCLC cells were injected into the lungs of mice. Mice were randomly assigned to treatment with selumetinib, cediranib, paclitaxel, selumetinib plus cediranib, or control. When controls became moribund, all animals were sacrificed and assessed for lung tumor burden and locoregional metastasis. Lung tumors and adjacent normal tissues were subjected to immunohistochemical analyses. RESULTS: Selumetinib inhibited lung tumor growth and, particularly at higher dose, reduced locoregional metastasis, as did cediranib. Combining selumetinib and cediranib markedly enhanced their antitumor effects, with near complete suppression of metastasis. Immunohistochemistry of tumor tissues revealed that selumetinib alone or with cediranib reduced ERK phosphorylation, angiogenesis, and tumor cell proliferation and increased apoptosis. The antiangiogenic and apoptotic effects were substantially enhanced when the agents were combined. Selumetinib also inhibited lung tumor VEGF production and VEGFR signaling. CONCLUSIONS: In this study, we evaluated therapy directed against MEK combined with antiangiogenic therapy in distinct orthotopic NSCLC models. MEK inhibition resulted in potent antiangiogenic effects with decreased VEGF expression and signaling. Combining selumetinib with cediranib enhanced their antitumor and antiangiogenic effects. We conclude that combining selumetinib and cediranib represents a promising strategy for the treatment of NSCLC.

Treatment with HIF-1alpha antagonist PX-478 inhibits progression and spread of orthotopic human small cell lung cancer and lung adenocarcinoma in mice.

INTRODUCTION: PX-478 is a potent small-molecule inhibitor of hypoxia-inducible factor 1alpha (HIF-1alpha). In prior preclinical studies, it had antitumor activity against various solid tumors in subcutaneous xenografts but had no measurable activity against a non-small cell lung cancer (NSCLC) xenograft. To determine the effectiveness of PX-478 against lung tumors, we investigated HIF-1alpha expression in several lung cancer cell lines, both in vitro and in vivo, and treated orthotopic mouse models of human lung cancer with PX-478. METHODS: Cells from two human lung adenocarcinoma cell models (PC14-PE6 and NCI-H441) or two human small cell lung cancer (SCLC) models (NCI-H187 and NCI-N417) were injected into the left lungs of nude mice and were randomized 16 to 18 days after injection with daily oral treatment with PX-478 or vehicle for 5 days. RESULTS: In the PC14-PE6 NSCLC model, treatment with 20 mg/kg PX-478 significantly reduced the median primary lung tumor volume by 87% (p = 0.005) compared with the vehicle-treated group. PX-478 treatment also markedly reduced mediastinal metastasis and prolonged survival. Similar results were obtained in a second NSCLC model. In SCLC models, PX-478 was even more effective. In the NCI-H187 model, the median primary lung tumor volume was reduced by 99% (p = 0.0001). The median survival duration was increased by 132%. In the NCI-N417 model, the median primary lung tumor volume was reduced by 97% (p = 0.008). CONCLUSIONS: We demonstrated that the PX-478, HIF-1alpha inhibitor, had significant antitumor activity against two orthotopic models of lung adenocarcinomas and two models of SCLC. These results suggest the inclusion of lung cancer patients in phase I clinical trials of PX-478.