IDENTIFICATION AND VALIDATION OF DRUG-TARGET INTERACTIONS USING CRISPR/CAS-MEDIATED GENOME EDITING

Drug discovery and development are cornerstones of the pharmaceutical industry. However, these processes are difficult, time-consuming and costly. Many new drugs fail during clinical development due to insufficient demonstration of their clinical benefit and efficacy. One of the major reasons for this failing is the lack of robust understanding of a drug’s mechanism of action. In addition, phenotypic drug discovery is experiencing a revival due to disappointing results obtained through target-based drug discovery. However, phenotypic drug discovery requires challenging target deconvolution approaches to identify the cellular target of a drug. Thus, the requirement of unraveling drug-target interactions to support the drug development process is greater than ever. Many methods are available to identify drug-target interactions, although none provides a one-size-fits-all solution. Genetic approaches in particular provide strong evidence for target confirmation as they examine drug-target interactions in a living cell. Although genetic approaches for target confirmation are widely applied to simple organisms, their application to mammalian cells has been hindered for a long time. Fortunately, chemical-genetic approaches in a mammalian context have started to develop with the advent of next generation sequencing technologies, RNA interference and more recently, CRISPR/Cas-mediated genome editing. However, current chemical-genetic approaches for higher eukaryotes are often plagued by technical issues, are not well suited to probe essential genes or require costly whole-exome sequencing combined with complex bio-informatics. Therefore, the field would benefit from new approaches that can simplify drug mechanism of action studies in a mammalian context. In this thesis we explore whether CRISPR/Cas genome editing can provide such an alternative for identification of drug-target interactions in a human context.

In the first part we validated the cellular drug-target interaction of selective inhibitors of nuclear export (SINE). These small molecule inhibitors show potent anticancer activity in a variety of in vitro and in vivo models of cancer and are currently being tested in clinical trials against cancer. They have been shown to modulate the activity of the Exportin-1 (XPO1) protein through covalent interaction with XPO1’s cysteine528 residue. XPO1 is an essential protein of the karyopherin-β family of nuclear transport receptors and is required for the active nuclear export of many proteins from the nucleus to the cytoplasm. Inhibition of XPO1 function is considered a potential anticancer strategy. However, the direct causality between the observed anticancer activity of the SINE and the binding of the SINE to XPO1 has not been demonstrated. To validate this drug-target interaction we applied CRISPR/Cas9-induced homology directed repair to generate XPO1 mutant cancer cell lines that carry a serine substitution of the cysteine528 residue. These mutant cell lines were highly resistant to treatment of the SINE on various parameters and our results validate the anticancer activity of these inhibitors is caused by selective binding to the cysteine528 residue of XPO1.

In the second part we developed a new genetic target deconvolution method based on CRISPR/Cas-mediated genome editing. We reasoned that targeted CRISPR/Cas-induced DNA double strand breaks can be utilized for rational mutagenesis to derive gain-of-function drug resistance mutations, even in essential genes. We validated this concept using three anticancer drugs for which the drug-target interactions are well-known. We then applied this concept on a large scale to scan multiple genes simultaneously and identified nicotinamide phosphoribosyltransferase (NAMPT) as the cellular target protein of an investigational anticancer compound. We further validated this drug-target interaction by X-ray crystallography and CRISPR/Cas validation experiments. Finally, we show that the CRISPR mutagenesis approach is compatible with the class 2, type V CRISPR AsCpf1 endonuclease, increasing the resolution of the methodology. Taken together, our results highlight CRISPR/Cas-mediated genome editing provides a powerful tool for the validation and identification of drug target interactions in a human cellular context.