• National Cancer Institute
  • National Human Genome Research Institute

Last Updated: July 1, 2013

Comprehensive Characterization of Breast Cancer Enhances Understanding of the Disease

Sara Bloom Leeds

Breast cancer is one of the most common cancers among women — about one in eight will develop some form of it within their lifetimes. The Cancer Genome Atlas (TCGA) Research Network published its first comprehensive characterization of breast cancer in the October 4, 2012 issue of Nature, which revealed that breast cancer is even more complex than previously thought. The resulting data have provided researchers with a greater understanding of this disease while offering insights into the myriad of genomic changes that are associated with it. These findings may contribute to possible treatments in the future.

Clusters of Clusters of Findings

To characterize the tumors, researchers analyzed samples from more than 800 patients and collected data across multiple platforms: exome sequencing, DNA methylation, copy number, mRNA, miRNA and protein expression.

The combined results supported the notion that there are indeed four main molecular subtypes of breast cancer based on variations in mRNA expression, among other less common subtypes. The four principal ones are Luminal A, Luminal B, HER2-enriched (HER2E) and Basal-like. The results also showed that the Basal-like subtype was the most distinct of the four on a molecular level.  

Reverse Phase Protein Array (RPPA) analysis was used to investigate cancer-related protein expression in more than 400 cases. Based on these data, TCGA researchers believe they have found two potentially new protein-defined subgroups: Reactive I and Reactive II. The first mainly consisted of a subset of Luminal A tumors, and the second was composed of a mixture of mRNA-subtypes.

In addition to examining the tumor itself, TCGA also looks at the surrounding cellular environment, since some of these data are likely to capture the importance of this microenvironment in relation to tumor growth and survival. In this case, the researchers dubbed these subgroups “reactive” because they believe that the characteristic proteins are likely produced by the surrounding microenvironment, which regulates many aspects of tumorigenesis. Although these groups did not differ greatly in regards to microRNA expression, DNA methylation, mutation, DNA copy number or the percent of tumor cell content, there were indeed many differences in protein and mRNA expression.  

Overall, there were more than 30,000 genomic mutations found in the biospecimens studied. Using a statistical analyzing software package called MuSiC (Mutation Significance in Cancer), the scientists identified 35 significantly mutated genes (SMGs). These were mutated more frequently than expected, and may be contributing factors to tumor development and progression. Three of these genes, PIK3CA, TP53 and GATA3 had somatic mutations that occurred at an incidence greater than 10 percent across all breast cancer subtypes.

TCGA analysis also helped identify a number of SMGs that had not previously been associated with breast cancer. Researchers found that different mutations in some of these same genes are known to cause different conditions. Further study of these changes might show a correlation with cancer development, and if so, the data could prove valuable in the quest for a cure. For example, one of these SMGs called TBX3, is also mutated in patients with the developmental disorder Ulnar-Mammary Syndrome, where abnormalities of the limbs, mammary glands, teeth and genitals may be present.

Similarities Found between Basal-Like Breast Cancer and Serous Ovarian Cancer

Perhaps the most interesting finding from the study came from comparing the breast cancer subtypes to other common cancers (colorectal, lung, kidney, ovarian and prostate) for which data were available from the TGEN expO dataset within the Gene Expression Omnibus, and TCGA’s ovarian cancer data. The results showed that of all the breast cancer subtypes, Basal-like is more similar to high-grade serous ovarian cancer on a genomic level than to other subtypes of breast cancer. In fact, both contained particularly similar types and frequencies of genomic mutations which may suggest that the two cancers may have arisen from a similar evolutionary molecular origin.

Additional similarities between Basal-like breast cancer and serous ovarian cancer included not only the high frequency of TP53 tumor suppressor gene mutations, but high expression of the AKT3 gene, which normally codes for an enzyme that regulates certain cellular processes. Both cancers also showed inactivation of BRCA1, as well as amplification and high expression of cMYC, which codes for a transcription factor that helps regulate cell growth and proliferation. Finally, the two cancers display a loss of another tumor suppressor gene, RB1, as well as amplification of Cyclin E1, which can lead to chromosome instability.

Prospective Therapies

Due to the similarities between Basal-like breast cancer and serous ovarian cancer, researchers have already begun exploring the idea that the two cancers could potentially be susceptible to similar treatments.

For example, ovarian cancer is often treated with PARP inhibitors or platinum-based chemotherapy. PARP proteins have many functions, and one of these roles is assisting with DNA repair and programmed cell death in healthy cells. Cancerous cells, however, use PARP to repair themselves to survive and grow, so inhibiting PARP in these cells may prevent this from occurring.  However, drugs that inhibit PARP activity are rarely used in breast cancer patients. TCGA’s new findings suggest that patients with Basal-like breast cancer might benefit from these drugs and perhaps ought to avoid more common chemotherapies that come with other significant side effects. To try to prevent these particular side effects, some oncologists have begun trying to use the platinum-based therapies on Basal-like breast cancer patients.1 However, clinical trials are necessary to determine if this method is as effective as the other.  

Another potential idea for treatment stems from a pathway called HIF1-α/ARNT, which TCGA’s research found to be one of the key regulatory features of Basal-like breast cancer. Transcription factors in this pathway respond to a decrease in oxygen, and thus might be susceptible to “bioreductive” drugs. These drugs undergo a chemical change under low-oxygen conditions, such as within a tumor, and then begin to produce toxic metabolites to kill surrounding cells.

In addition, the study found a high rate of mutations of PIK3CA in Luminal tumors. Researchers are interested in investigating if inhibitors of this activated kinase or the PI3K pathway would be effective in treating these hormone-responsive tumors.

Together with other large-scale genomic data on breast cancer, TCGA’s comprehensive work is an invaluable resource to understanding the complexity of this type of cancer. It is TCGA’s hope that the data published in this study will help researchers focus on specific areas within the genome that might one day be effectively targeted by various drugs and perhaps even lead to alternative solutions to treating this disease.

The Cancer Genome Atlas Research Network. (4 Oct 2012) Comprehensive molecular portraits of human breast tumors. Nature. 490 (7418): 61-70. Read the full article

Selected Reference:

1From TCGA to Treatment. (2 Dec 2012). Cancer Discov. American Association for Cancer Research. 2 (12): 1068. Read the full article