Discovery of small-molecule PD-1/PD-L1 antagonists through combined virtual screening and experimental validation.

Inhibition of the interaction between the PD-1 protein on activated lymphocytes and the PD-L1 protein on tumors represents a novel therapeutic approach for selective activation of the innate immune response against a variety of cancers. Therefore, the present study utilized a combined virtual and experimental screening approach to screen databases of both lead-like and larger molecules for identification of novel inhibitors of PD-1/PD-L1 interaction. First, high-throughput virtual screening of ∼3.7 million lead-like molecules using a rigid-receptor docking approach against both human PD-1 and PD-L1 proteins revealed possible small-molecule tractability of PD-1, but not PD-L1, binding interface. The subsequent work, therefore, involved screening of the National Cancer Institute (NCI) compound database against the PD-1 pocket. Several NCI compounds were identified with potential to bind to the PD-1 pocket and in turn inhibit the PD-1/PD-L1 interaction. The dynamic binding behavior of these molecules was further investigated using long 100 ns molecular dynamics (MD) stimulation revealing NSC631535 to be a potentially stable binder at PD-1 interface pocket. In support of these MD data, the experimental testing of NSC631535 exhibited 50% inhibition at ∼15 µM test concentration. The observed activity of this compound is promising as despite its relatively low molecular weight (415.5 g/mol) it is still capable of inhibiting the PD-1/PD-L1 interaction having a large interface area (∼1970 Å2). In summary, our integrated computational and experimental screening led to identification of a novel PD-1 antagonist that may serve as a starting point for further optimization into more potent small-molecule PD-1/PD-L1 inhibitors for cancer immunotherapy.

Identification of antipsychotic drug fluspirilene as a potential p53-MDM2 inhibitor: a combined computational and experimental study.

The inhibition of tumor suppressor p53 protein due to its direct interaction with oncogenic murine double minute 2 (MDM2) protein, plays a central role in almost 50 % of all human tumor cells. Therefore, pharmacological inhibition of the p53-binding pocket on MDM2, leading to p53 activation, presents an important therapeutic target against these cancers expressing wild-type p53. In this context, the present study utilized an integrated virtual and experimental screening approach to screen a database of approved drugs for potential p53-MDM2 interaction inhibitors. Specifically, using an ensemble rigid-receptor docking approach with four MDM2 protein crystal structures, six drug molecules were identified as possible p53-MDM2 inhibitors. These drug molecules were then subjected to further molecular modeling investigation through flexible-receptor docking followed by Prime/MM-GBSA binding energy analysis. These studies identified fluspirilene, an approved antipsychotic drug, as a top hit with MDM2 binding mode and energy similar to that of a native MDM2 crystal ligand. The molecular dynamics simulations suggested stable binding of fluspirilene to the p53-binding pocket on MDM2 protein. The experimental testing of fluspirilene showed significant growth inhibition of human colon tumor cells in a p53-dependent manner. Fluspirilene also inhibited growth of several other human tumor cell lines in the NCI60 cell line panel. Taken together, these computational and experimental data suggest a potentially novel role of fluspirilene in inhibiting the p53-MDM2 interaction. It is noteworthy here that fluspirilene has a long history of safe human use, thus presenting immediate clinical potential as a cancer therapeutic. Furthermore, fluspirilene could also serve as a structurally-novel lead molecule for the development of more potent, small-molecule p53-MDM2 inhibitors against several types of cancer. Importantly, the combined computational and experimental screening protocol presented in this study may also prove useful for screening other commercially-available compound databases for identification of novel, small molecule p53-MDM2 inhibitors.