We learned from this study how to start thinking about the complex analysis of not just seeing that a mutation exists in a tumor but learning to recognize whether a given mutation is likely to be important.
—Todd R. Golub, MD
Advances in next-generation DNA sequencing technologies are allowing scientists to decipher the whole genome or whole exome (ie, the coding region of the genome) of cancer specimens more quickly and inexpensively than ever before. And the results are revealing genes that had not previously been associated with cancer as well as multiple genetic mutations that disrupt common pathways and trigger cancer-related changes in a cell.
Findings in Multiple Myeloma
Todd R. Golub, MD, Director of the Cancer Program at the Broad Institute, Charles A. Dana Investigator in Human Cancer Genetics at Dana-Farber Cancer Institute, is a coauthor of the first large-scale study to compare the entire genomes of malignant cells and normal cells from 38 patients with multiple myeloma. Dr. Golub and co-investigators at the Broad Institute and Dana-Farber Cancer Institute, Boston, found mutations in the nuclear factor (NF)-kappaB pathway, an important transcriptional regulator in multiple myeloma. Although researchers had suspected that this pathway was involved in the development of the disease, they did not understand the chain of events that turned the pathway on. The study, published last year in the journal Nature,1 found that 11 different genes involved in the NF-kappaB pathway were altered in at least one tumor sample.
In addition, mutations were found in genes that are involved in two cellular processes in normal cell function: RNA processing and protein folding. About half of the multiple myeloma samples sequenced had defects in one or more of these genes, including FAM46C, which had never before been linked to cancer. Another surprising discovery was that a small number of multiple myeloma patients (just 4%) had BRAF V600E gene mutations—the same mutation found in some types of melanoma and colon cancer, which had not previously been associated with the development of multiple myeloma.
Driver vs Passenger Mutations
“We learned from this study how to start thinking about the complex analysis of not just seeing that a mutation exists in a tumor but learning to recognize whether a given mutation is likely to be important … because we know that probably the majority of the mutations observed in any individual tumor—the so-called passenger mutations—are not going to be diagnostically important, they are not going to point to good treatment strategies. And so there is a need for approaches to sort out the difference between the driving events—the mistakes that drive cells toward becoming cancerous—and the passenger mutations—genetic alterations that are just along for the ride,” said Dr. Golub.
One way to accomplish that goal, said Dr. Golub, is to develop sophisticated algorithms to establish the statistical significance of mutation frequency in a specific gene, a complicated process because the mutation rate is not the same across the various tumor types. Additional analysis of the myeloma samples is also showing a difference in the mutation rates among the subtypes of the cancer. “And we are now discovering that the mutation rate even within a single sample may be different across the genome. We did not appreciate that in our initial study,” said Dr. Golub.
Finding Therapeutic Targets
To distinguish the so-called driving events that propel cells toward becoming cancerous from passenger mutations, thousands of myeloma tumors will be needed for genomic analysis, said Dr. Golub. It may be possible to analyze such a vast number of tumors over the next 2 years, as the cost of genomic sequencing continues to come down. Currently, Dr. Golub is completing the genomic sequencing of 250 additional multiple myeloma tumors.
The next year, said Dr. Golub, will be spent analyzing how the new genes discovered in multiple myeloma contribute to the development of the disease and whether they represent new therapeutic targets for the cancer. “Simply identifying the existence of a mutation does not make for a treatment of the disease. It’s going to take time for the scientific community to understand what these mutations mean and what they do to the cell, and figure out if there is a therapeutic opportunity. But at least now people know what to work on,” said Dr. Golub.
Rare Mutations in Breast Cancer
A large-scale genomics investigation of breast cancer tumors is also revealing genes never before associated with the disease. Using DNA samples taken from 50 patients with luminal (estrogen receptor– and/or progesterone receptor–positive), HER2-negative breast cancer enrolled in a clinical trial and comparing them to the patients’ healthy cells, researchers from the Genome Institute at Washington University in St. Louis found over 1,700 mutations. Most of these mutations were unique to the individual patient and involved single-nucleotide variations, frame shifts, translocations, and deletions.
Similar to what has been published in earlier studies, the researchers found two relatively common mutations: PIK3CA, which was present in 50% of the tumors, and TP53, found in about 20% of the tumors. However, mutations in the tumor-suppressor gene MAP3KI, found in about 15% of ER-positive breast cancers, represented a significant new discovery.
“The MAP3KI suppressor gene was hitherto unknown in breast cancer,” said Matthew J. Ellis, MD, PhD, Professor of Medicine and Chief of Breast Oncology at Washington University, St. Louis, and lead investigator of the study. “MAP3KI is a stress kinase. It becomes activated when cells are responding to stress, including chemotherapeutic stress, and when activated, triggers cell death. The interesting clinical association is that MAP3KI-mutant tumors are associated with lower-grade histology. It is the opposite of what you see with mutant P53. The significance of this mutation was not appreciated before, because a lot of sequencing focused on cell lines or in higher-grade breast cancers, so mutations associated with low-grade breast cancers were missed.”
Because breast cancer is so common, discovering rare but drug-sensitive gene mutations that occur in as few as even 1% or 2% of all breast cancers is crucial, because those small percentages still translate to between 2,000 and 4,000 patients a year in the United States.
Dr. Ellis’ research was presented at last year’s annual meeting of the American Association for Cancer Research.2 This study has been expanded by the Washington University team to 77 patients and has documented 3,355 mutations. The complete findings of the genomic analysis of Dr. Ellis’ new study were presented at the ASCO Annual Meeting.3
Making Personalized Medicine a Reality
On a national level, The Cancer Genome Atlas (TCGA), a joint effort of the National Human Genome Research Institute and the NCI, is now sequencing the genomes of 22 cancers with the goal of completing its analysis in 2 years.
“Although we may not finish evaluating every one of the 500 samples of each of the 22 tumor types by 2014, we will have a large data set that will be available to researchers so they can begin their own investigations and make their own discoveries combined with their own data sets,” said Bradley A. Ozenberger, PhD, Program Director, The Cancer Genome Atlas.
The next phase of total genomic sequencing, said Dr. Ozenberger, is application in the clinic with analyses that match a patient’s specific tumor biomarkers with targeted therapies to inhibit cancer growth. And genomic analysis of malignant tumor samples is already starting to have an impact in the clinic. Some large cancer institutions are offering genomic testing for certain malignancies, including lung cancer and multiple myeloma, although more needs to be understood before clinical-grade genomic sequencing becomes standard of care for these patients, said Dr. Golub.
Much to Learn
“There is still much to learn about how to interpret the data and learn what aspects of it are actually clinically useful vs just a curiosity,” Dr. Golub said. “That all still needs to be worked out, but I think it is no longer a matter of when genomic sequencing will enter the clinic, because the when is now. It is going to be a matter of the pace at which it becomes used routinely in clinical research and in routine clinical care,” he added.
“Many of the mutations that we are uncovering from genomic sequencing are mysterious, but I think that the first use for some of them will be in predictive and prognostic medicine with respect to standard-of-care drugs” Dr. Ellis commented. “What we may find is that some of the mutations are quite predictive of lack of response to standard therapy. I think that some of these mutations we are uncovering in luminal breast cancer will be resistance markers, probably not only to endocrine therapy but also to chemotherapy.”
Because whole-genome sequencing is still a complex process, achieving true personalized medicine for individual patients with cancer is still years away from becoming a reality. In the meantime, “progress will be made in subsets of patients, one mutation-defined cancer subset at a time,” said Dr. Ellis. ■
Disclosure: Dr. Golub is an advisor to Foundation Medicine, Inc., which is developing sequencing-based diagnostic tests for cancer. Drs. Ellis and Ozenberger reported no potential conflicts of interest.
1. Chapman MA, Lawrence MS, Keats JJ, et al: Initial genome sequencing and analysis of multiple myeloma. Nature 471:467-472, 2011.
2. Ellis MJ, Ding L, Shen D, et al: Analysis of luminal-type breast cancer by massively parallel sequencing. American Association for Cancer Research 102nd Annual Meeting. Abstract LB-87. Presented April 3, 2011.
3. Ellis, MJ Ding L, Shen D, et al: Whole genome sequencing to characterize luminal-type breast cancer, 2012 ASCO Annual Meeting. Abstract 503. Presented June 3, 2012.