NCI Develops Large Dataset of Genomic Variants Across Nine Common Cancer Types

Yves Pommier, MD, PhD

Researchers at the National Cancer Institute (NCI) have developed a comprehensive analysis of coding variants in nine of the most frequently studied human tumor cell lines in cancer research, according to a study published online in the journal Cancer Research.

The NCI has decided to make this resource available for other investigators as they study these variants and associated drug response. The NCI-60 human tumor cell line panel is frequently used to discover novel anticancer therapies, and over the years, researchers have discovered a wide range of cancer-specific genetic variations in numerous types of cancer.

“Opening this extensive dataset to researchers will expand our knowledge and understanding of tumorigenesis, as more and more cancer-related gene aberrations are discovered,” said Yves Pommier, MD, PhD, chief of the Laboratory of Molecular Pharmacology at the NCI in Bethesda, Maryland, in a statement. “This comes at a great time, because genomic medicine is becoming a reality, and I am very hopeful this valuable information will change the way we use drugs for precision medicine.”

In this study, investigators at the NCI extracted DNA from the cells and used whole exome sequencing in order to reveal aberrant patterns of gene expression in tumor cells and novel cancer variants. The researchers then distinguished variants into two types: type I variants corresponded to genetic variants found in the normal population, whereas type II variants are variants that are specific to cancer.

Using a novel algorithm, the researchers were able to predict the sensitivity of tumor cells with type II variants to 310 drugs, including 103 FDA-approved anticancer drugs and 207 investigational agents.

The researchers found numerous correlations between specific variants in genes and their susceptibility to anticancer agents. For example, mutations of TP53, the most frequently mutated gene in this panel, were strongly correlated to drug response. The investigational MDM2 inhibitor nutlin-3 showed the highest statistically significant score of activity in p53 wild-type cell lines.

Additionally, specific variants for BRAF, KRAS, NRAS, PI3KCA, PTEN, and ERBBs showed influence on the response to clinically relevant targeted agents, including vemurafenib, selumetinib, hypothemycin, rapamycin, wortmannin, perifosine, erlotinib, afatnib, lapatinib, and neratinib.

“To date, this is the largest database worldwide, containing six billion data points that connect drugs with genomic variants for the whole human genome across cell lines from nine tissues of origin, including breast, ovary, prostate, colon, lung, kidney, brain, blood, and skin,” said Pommier. “We are making this dataset public for the greater community to use and analyze.”

The newly released data are now part of the NCI-60 cell line collection that can be accessed through multiple sources, including the CellMiner and Ingenuity Systems portals, whereby investigators can choose those cell line models most relevant to their research.

Abaan OD, Polley EC, Davis SR, et al. The exomes of the NCI-60 panel: a genomic resource for cancer biology and systems pharmacology. Cancer Res. 2013;73(14):4372-82.