The Genomics of Drug Sensitivity in Cancer project recently added 4 years of additional data to a large-scale database designed to identify predictive biomarkers for cancer therapies. Freely available online, these data are driving research on the relationship between genomic features and drug responses to better tailor treatment to individual patients.

The Genomics of Drug Sensitivity in Cancer (GDSC) project has added 4 years of additional data to a database of cancer drug screens designed to identify predictive biomarkers for cancer therapies. Available at cancerrxgene.org, these data are driving research on the relationship between genomic features and drug responses to better tailor treatment to individual patients.

The GDSC project started 10 years ago and is part of the Cancer Dependency Map, a joint effort between the Wellcome Sanger Institute in Cambridge, UK, and the Broad Institute in Cambridge, MA, to identify vulnerabilities in cancer cells that could become drug targets. “We've pulled together one of the largest collections of cancer cell models in the world,” says Hayley Francies, PhD, of the Sanger—one that, according to a 2016 study, accurately captures the genomic landscape seen in tumor cells from patients (Cell 2016;166:740–54). The goal, she adds, is to fully characterize the genomic features of the models and then use the models for research.

In the GDSC project, researchers are performing high-throughput screens and cataloguing drug responses “to try to integrate the genomic information and the drug response information, and look for biomarkers of drug response,” Francies says. To date, the project has released data on 453 cancer drugs tested in 989 cell lines, which include 386,293 assessments of how cell lines respond to different doses and 494,973 associations between drug responses and genomic features such as mutations, gene amplifications and deletions, gene fusions, and changes in gene expression. The project mainly tests approved drugs and those in clinical trials.

graphic

Ewing sarcoma cells with the EWS–FLI1 translocation are sensitive to the PARP inhibitor olaparib, shown above binding to PARP.

“I think the key selling point of our project is that we've got a huge range of diversity of cancer models that we're using,” says Elizabeth Coker, PhD, of the Sanger. “You need a relatively small number of models to find the really big, really obvious hits, but you need statistical power to find the rarer biomarkers and the rarer variants, and that comes with an increased scale.” For example, Sanger researchers used the data to establish that Ewing sarcoma cells with the EWS–FLI1 translocation are sensitive to the PARP inhibitor olaparib (Lynparza; AstraZeneca; Nature 2012;483:570–5). Clinical trials are now testing olaparib and other PARP inhibitors against the disease.

Other institutions are querying the database, too: It is accessed by more than 350 people per day. “Because we do have the ability to hold these large numbers of cell models and to conduct these high-throughput drug screens, we are able to be a starting point for other labs to investigate drug response,” Francies says. One research group is developing a method for predicting patient responses to chemotherapy based on gene-expression levels and in vitro drug sensitivity, whereas another is investigating why MEK inhibitors show activity in BRAF-mutant but not KRAS-mutant melanoma, and a third is exploring RANBP2 as a therapeutic target in colon cancers that resemble BRAF-mutant disease but lack a BRAF alteration (Genome Biology 2014;15:R47; Cancer Cell 2014;25:697–710; Cell 2016;165:317–30).

The researchers are now adding more models and drugs to the project. They are generating organoids to represent more cancer subtypes, and they also plan to begin testing drug combinations. “Right now, the majority of patients don't receive precision cancer medicine, and that is simply because we don't understand what targeted drugs we should be giving to what patients,” Coker says.

Patricia Jaaks, PhD, also of the Sanger, adds, “We hope that finding new biomarkers will eventually help to position drugs in the right disease or cancer context.” –Catherine Caruso

For more news on cancer research, visit Cancer Discovery online at http://cancerdiscovery.aacrjournals.org/CDNews.