Skip to main content
Premium Trial:

Request an Annual Quote

TCGA Study Provides 'Roadmap' for Targeted Therapy, Genomic Subtyping in Ovarian Cancer


By Monica Heger

Churning out its largest data set yet, the Cancer Genome Atlas project has sequenced the exomes of 316 ovarian tumors and their matched normal samples, identifying several mutations for which drugs currently exist, and revealing four molecular subtypes of high-grade serous ovarian cancer.

The team performed whole-exome sequencing, mRNA and microRNA expression profiling, methylation analysis, pathway analysis, and genotyping of the tumor samples, identifying potential treatment strategies and subtypes, and also providing a comprehensive picture of high-grade serous ovarian cancer tumor.

The results were published online today in Nature.

Paul Spellman, a research scientist at Lawrence Berkeley National Laboratory and project leader for this arm of TCGA, said that he was somewhat surprised to find that ovarian cancer does not appear to be defined by one, or even several, major driving cancer genes, but rather a "long tail of mutations that are contributing to the pathogenesis of the disease," and recurrent in only a small percentage of cases.

Additionally, he said, the study showed that structural variation was likely a driver of ovarian cancer — "nearly the entire genome is aberrant," he said.

Sequencing, copy number analyses, expression profiling, and methylation analyses were done at LBNL; Baylor College of Medicine; the Broad Institute; the Washington University Genome Institute; Harvard Medical School; Memorial Sloan-Kettering Cancer Center; the HudsonAlpha Institute for Biotechnology; the University of North Carolina, Chapel Hill; and the University of Southern California. Eighty tumor samples were sequenced on Life Technologies' SOLiD and 236 tumors were sequenced on the Illumina GA.

Nearly all of the tumors — 96 percent — harbored mutations to the tumor suppressor gene TP53, however there was not a set of driving mutations that were broadly recurrent across the population. Instead, nine genes were found to be statistically significant for mutations, which, aside from TP53, each occurred in just a small subset of the population.

Germline mutations to the BRCA1 and BRCA2 genes were found in 9 percent and 8 percent of cases, respectively, with somatic mutations to those genes found in an additional 3 percent of cases. The researchers identified six other statistically recurrently mutated genes: RB1, NF1, FAT3, CSMD3, GABRA6, CDK12.

Comparing the mutations to publicly available databases identified hundreds more genes that were mutated, including BRAF, KRAS, PIK3CA, and NRAS. The authors concluded that because of the "transforming activity" of mutations in those genes, that they were likely rare but important drivers in high-grade serous ovarian cancer.

Therapeutic Implications

A number of the findings have therapeutic implications. The team mined several databases, including Ingenuity Systems' pathway database,, and DrugBank, to search for inhibitors of genes that were amplified. They identified 22 targets that were over-expressed in at least 10 percent of the samples, including MECOM, MAPK1, KRAS, and CCNE1.

BRCA1/2 mutations were found in about 20 percent of the samples, and have been found to be responsive to PARP inhibitors.

In a pathway analysis, the team identified alterations to additional genes involved in homologous recombination, which the PARP inhibitors are designed to target, nearly doubling the number of patients who would be candidates for receiving PARP inhibitors, to 50 percent.

"It's a scale which may not be appreciated if you just looked at patients who had BRCA 1 or BRCA 2 lesions," Spellman said.

[ pagebreak ]

Further pathway analyses found that the RB signaling pathway, involved in cell cycle progression; the PI3K/RAS pathway, involved in proliferation and survival; and the NOTCH signaling pathway were altered in 67 percent, 45 percent, and 22 percent of cases, respectively. These pathways offer potential opportunities for therapy, as drugs are currently in development that target genes in those pathways.

Additionally, the BRAF V600E mutation was found in a small number of cases, about one out of 300. Drugs to target that mutation are currently in clinical trials for malignant melanoma, and the same mutation was recently found to be present in HCL (CSN 6/15/2011).

The BRAF finding adds support to the "long tail distribution" theory, said Spellman. "A lot of these genes we find in other cancer types that are more prevalent may show up in other cancers too, but less prevalent."

Spellman added that this suggests that screening cancer patients with a sequencing-based gene panel test could be a good way to identify potential therapeutic options that clinicians may not otherwise consider.

David Ting, a researcher at Massachusetts General Hospital's Cancer Center, said that the sheer amount of data that the team was able to generate on such a "complex and heterogeneous" disease was impressive. However, he said the next major challenge would be to figure out which "rabbit hole" to pursue.

Because the sequencing and analyses generated so many potential strategies to explore, it will be important to prioritize, he said, noting that the most promising first avenues would be to start testing drugs that have shown some benefit in similar mutations in other cancers — for instance, PIK3CA inhibitors, KRAS inhibitors, and inhibitors targeting the RAS pathway.

PARP inhibitors are already being tested in patients with ovarian cancer, he said, so this study further validates that approach.

He said the study could be used as a "roadmap" by other researchers working in ovarian cancer to help guide and prioritize translational medicine studies.

Stratifying Patients

The team also used arrays to do an expression analysis of 11,864 genes, finding four distinct expression clusters, which they dubbed: immunoreactive, differentiated, proliferated, and mesenchymal. The same analytical approach applied to a separate set of public data yielded the same four clusters, suggesting they are real subtypes.

The four subtypes "appear to have different biology that drives their tumors," said Spellman. "One would hope that now, like in breast cancer where subtypes are treated differently, that we could use these tumors to develop novel therapies that go after the specific subtypes," he said.

While prognosis did not differ significantly between the four subtypes, and was poor across the board, the different subtypes were characterized by different mutational spectrums. Spellman added that much more work will be needed to understand the differences between the subtypes and said that a subsequent study would look at the biology of each of the subtypes in more detail.

Aside from defining these four subgroups, the study further illustrated that high-grade serous ovarian cancer is distinct from other subtypes of ovarian cancer. For example, while nearly all cases had mutations in the TP53 gene, clear-cell ovarian cancer tumors have very few TP53 mutations, but recurrent ARID1A and PIK3CA mutations, while endometrioid ovarian cancer tumors also have lower rates of TP53 mutations, but frequent CTNNB1, ARID1A, and PIK3CA mutations.

Ting added that a larger, prospective study would also need to be done to look at the four subtypes, to determine how each of the pathways involved are important for pathogenesis and outcomes, and to figure out the relevant, targetable mutations.

Overall though, the TCGA study is "definitely something that has never been able to be accomplished before," added Ting. "There will be more of these, and we'll learn more about these cancers," he said. "It's not the solution, but it is the roadmap."

Have topics you'd like to see covered by Clinical Sequencing News? Contact the editor at mheger [at] genomeweb [.] com.

The Scan

Study Finds Sorghum Genetic Loci Influencing Composition, Function of Human Gut Microbes

Focusing on microbes found in the human gut microbiome, researchers in Nature Communications identified 10 sorghum loci that appear to influence the microbial taxa or microbial metabolite features.

Treatment Costs May Not Coincide With R&D Investment, Study Suggests

Researchers in JAMA Network Open did not find an association between ultimate treatment costs and investments in a drug when they analyzed available data on 60 approved drugs.

Sleep-Related Variants Show Low Penetrance in Large Population Analysis

A limited number of variants had documented sleep effects in an investigation in PLOS Genetics of 10 genes with reported sleep ties in nearly 192,000 participants in four population studies.

Researchers Develop Polygenic Risk Scores for Dozens of Disease-Related Exposures

With genetic data from two large population cohorts and summary statistics from prior genome-wide association studies, researchers came up with 27 exposure polygenic risk scores in the American Journal of Human Genetics.