Data from the Cancer Genome Atlas shows basal-like breast cancers share similar genetic origins and features with serous ovarian tumors. The large collaborative study, conducted by the National Human Genome Research Institute and the National Cancer Institute, also confirmed the previously identified four main breast cancer subtypes—luminal A, luminal B, HER2, and basal-like, each of which are molecularly distinct. The findings are published in Nature.
“This kind of comprehensive and systems approach to the problem is exactly the point of the Cancer Genome Atlas,” said Lisa A. Carey, MD, associate professor of medicine at the University of North Carolina, Chapel Hill, North Carolina.
The study is unique, analyzing a variety of different breast cancer types and using multiple technology platforms to evaluate the DNA, RNA, and protein of these tumors. The comprehensive study analyzed primary tumor samples from 825 patients using different, complementary technologies—genomic DNA copy number arrays, exome sequencing, DNA methylation, messenger RNA arrays, microRNA sequencing, and reverse-phase protein arrays. Each tumor sample was analyzed using at least one analytic tool, 463 patient tumor samples were analyzed using five of the six analytical techniques, and 348 patient samples with all six technologies.
Similarities Between Ovarian and Basal-Like Breast Cancers
In a copy number and genomic mutation analysis of all four breast cancer subtypes and ovarian cancer tumor samples, basal-like breast cancer and ovarian cancer were found be most similar. The authors highlight that the similarities between basal-like breast cancer and serous ovarian cancers indicate that these two difficult-to-treat cancer types may be able to be treated with the same therapies—chemotherapy drugs, anti-angiogenesis agents, and others in development.
Basal-like breast cancers account for approximately 10% of all breast cancers. The patient population for this subtype is younger and includes more African American than other cancer subtypes. The subtype is also associated with a higher incidence of germline BRCA1 mutations. This breast cancer type currently only has chemotherapy options. This breast cancer subtype includes triple-negative breast cancers, characterized by the aberrations they lack—HER2 overexpression, and estrogen- and progesterone(Drug information on progesterone)-receptor expression. However, about three quarters of triple-negative breast cancers are basal-like, while the other quarter had similarity to all of the other subtypes. The overall mutation spectrum of these two cancers was similar, based on previous results of ovarian cancer genome analyses. For both ovarian and basal-like breast cancer, TP53 was found mutated in 80% of tumors.
Basal-like breast cancers and ovarian cancers had similar rates of mutation and a similar spectrum of mutations. The basal-like tumors had a high frequency of mutations in the ATM, BRCA1, BRCA2, as well as RB1 loss and cyclin E1 amplification—the same mutations identified for ovarian cancers.
About 20% of basal-like breast tumors had a germline or somatic mutation in either BRCA1 or BRCA2. The authors suggest that these BRCA-mutated patients could potentially respond to Poly (ADP-ribose) polymerase (PARP) inhibitors. The analysis also showed various copy number amplifications and deletions that may be therapeutic targets.
“The central role of BRCA1 and BRCA2 in the [response to DNA damage] may have implications for chemotherapy choices or PARP inhibition,” said Dr. Carey, pointing out that both types of therapies have been pursued for both ovarian and basal-like breast cancer. “So while the genomic data are new, the therapeutic implications are being tested already.”
Breast Cancer Subtype Comparisons
Basal-like and HER2-overexpressing tumors had higher mutation rates compared to the other two subtypes. The luminal A subtype had the lowest mutation rate. These two types also had the least number of mutated genes. A fifth subtype was observed, but with only eight patient samples with this type, the authors could not provide definitive characterization of this “normal-like” breast tumor type.
Three genes—TP53, PI3KCA, and GATA3—were common to all of the four subtypes, occurring in more than 10% of the tumor samples. The spectrum of mutations identified by exome sequencing found novel and previously identified mutations. The luminal A subtype had the most “significantly mutated genes,” according to the authors, with 45% of the tumors having a PIK3CA mutation, followed by high frequencies of mutations in MAP3K1, GATA3, TP53, CDH1, and MAP2K4. Luminal B breast cancers had a 29% frequency of TP53 and PIK3CA mutations—the most frequently mutated genes in this breast cancer type. Basal-like cancers had a TP53 mutation in 80% of the tumors sample, and 9% had a PIK3CA mutation. These were the only two gene mutations found in common between the basal-like and luminal subtypes. HER2 was amplified in 80% of HER2-breast cancers. TP53 mutations were found in 72% of these cancers, and 39% had a PI3KCA mutation. In general mutations in TP53 have been linked to a poorer outcome for patients.
Analyzing the DNA of normal tissue samples, approximately 10% of the patients were found to harbor a germline mutation in nine different genes that may confer a predisposition to breast cancer, including BRCA1, BRCA2, PTEN, and RAD51C, among others. This suggests that some sporadic breast cancer patients actually may have a genetic component.
How feasible are these massive whole-genome and proteome tumor analyses? “They are less massive now than they used to be, and it is likely that this sort of comprehensive evaluation will be key to understanding individual tumor biology,” said Carey. To make this analysis really comprehensive, according to Carey, the tumor microenvironment should also be assessed, as it is likely to have effects on the evolution of a cancer.
