Phenotypic drug discovery
To achieve the promised benefits of cancer precision medicine, we need better targeted and more effective therapies. Oncology drug discovery has traditionally relied on target validation followed by high throughput screens to identify small molecule inhibitors. In contrast, phenotypic drug discovery offers the potential to unlock unexpected mechanisms as paths to radically new cancer treatments. We previously employed a cellular barcoding method to evaluate 4,518 drugs for their ability to kill 578 diverse cancer cell lines. We found that four dozen non-cancer drugs show unexpected selective cancer killing that is predictable from tumor genomic features. In the past, such discoveries would stagnate due to the difficulty of moving from a promising phenotype to biological mechanism. Using modern genomic and proteomic methods, we find that drugs with unexpected anti-cancer activities 1) often act through new molecular targets, 2) exploit non-inhibitory mechanisms such as stimulating protein-protein interactions, and 3) offer potential for future clinical translation. Discoveries resulting from these studies include compounds with selective activity against drug resistant cancer (Nature Cancer 2020), a new molecular glue mechanism (Nature 2020), and compounds with copper-dependent cell death (Science 2022). We are evaluating compound new mechanisms and their future translational potential.
Existing drugs have serendipitously been found to be effective for new uses in the past, but it has been challenging to identify these opportunities systematically. A major barrier is the lack of easy laboratory access to pharmaceutical collections. We founded the Drug Repurposing Hub to enable scientists to perform laboratory experiments using a high-quality and open source drug collection (Nature Medicine 2017). The collection (housed at the Broad Institute) now contains more than 6,000 existing drugs and other molecules developed by pharmaceutical companies and is accompanied by a freely available information resource. Compounds have been profiled across multiple information-rich cellular assays including perturbational gene expression, cancer cell viability, and cell morphology. We are performing focused screens of the collection using isogenic cancer models and other strategies.
Targeted therapy for gastrointestinal cancers
Patients with advanced gastrointestinal cancers, including gastroesophageal, colorectal, and liver cancers, have limited treatment options. Using results from large-scale cancer dependency screens, we identified the serine-threonine kinase casein kinase 1 as a selective vulnerability (AACR oral presentation 2022). Through medicinal chemistry collaboration, we developed potent and specific small molecule inhibitors with anti-tumor efficacy in vivo. We are working to select the target patient population for future development, identify mediators of compound response, and determine the downstream mechanism of cell death.