Super Method Seeks Cancer's Kryptonite

Decades of toil have led to significant breakthroughs in cancer diagnosis and treatment (1) – and many genetic loci have been pinpointed as key players in tumor development. But the search for more personalized therapy continues. Here, the combination of genome-wide CRISPR and computational biology may prove to be a powerful tool.

Fiona Behan is a Senior Scientist at the Wellcome Sanger Institute in Cambridge. With significant experience in genome-wide CRISPR, she was keen to apply the technology to oncology. “Collaborating with Kosuke Yusa, who was one of the leaders in developing genome-scale CRISPR technology, and Mathew Garnett, who had expertise in high-scale cancer screening, provided us with a great starting point for our research (2),” says Behan.

The goal? To determine new druggable targets. Although Behan acknowledges this is hardly a novel endeavor, she notes that previous approaches have largely been unsuccessful. “Many of these genome-scale projects in the past have established very long lists of potential targets,” she says. “Yet we’ve failed to be able to translate these long lists of targets into the clinic or, ultimately, into new therapies.”

Disrupting almost 20,000 genes in over 300 cancer models, Behan’s study is one of the largest of its kind. Working with such an enormous bank of expression, sequencing, and clinical data, the team had its work cut out. “We realized how huge the computational challenge was,” says Behan. “But our priority has always been to develop a data-driven system capable of selecting targets from our list.” By integrating cell fitness data, genomic biomarkers, and target tractability information into a single computational entity, they were able to systematically prioritize new targets by tissue and genotype.

The overall result? An online tool that integrates the work of multiple laboratories at the Sanger called the Cancer Dependency Map (3), which includes over 600 genes that the team believe could play a significant role in tumor development. “By identifying all these dependencies, we hope to develop a tool that can be brought into the clinic to inform decision making regarding precision medicine or targeted therapies,” says Behan. “Ideally – perhaps in 10 or 15 years – a patient could go into the clinic, have their tumor sequenced for as many known cancer genes or mutations, and – based on research that the dependency map has driven – access for more tailored treatment,” she says.

Beyond that specific goal, the work could aid research more broadly. “We hope it provides a starting point for other scientists, many of whom may lack the resources we have to carry out these large genome-scale CRISPR experiments. Instead, they can apply our data to their particular interest – whether it be a particular tumor type or cell.”

In fact, the tool is already bearing fruit. One marker – the WRN protein – is already being explored by researchers at the Institute. “The WRN protein has emerged in recent studies, including ours, as a dependency in microsatellite unstable cancers (4),” says Behan. “Gabrielle Picco, a colleague in our lab, is beginning a deep dive into WRN to see if we can determine the underlying mechanisms behind why these cancers are completely dependent on WRN.”

Progress is being made, but major hurdles remain; tumor heterogeneity and emergent tumor resistance (to name but two) could throw a spanner in the works. But Behan is resolute: “We hope to further refine our prioritization system, this time at a medium-throughput level, rather than the high-throughput approach that underpinned the Cancer Dependency Map.” To do so, Behan is working with Open Target to develop a Validation Lab that will allow researchers to pursue particular targets, confirm target suitability, and then provide preliminary data to drive further research.

As for uncovering cancer’s kryptonite? Genome-scale CRISPR experiments and powerful computational approaches appear to be super combination.