This article is more than 5 years old. Information may no longer be current.
Whole-genome sequencing yields knowledge, hope for future of cancer careWhich approach, whole-genome sequencing or targeted genomic sequencing, holds more promise for cancer research?
Click Here to Manage Email Alerts
Compared with other disciplines of scientific discovery, cancer genomics is in its infancy.
However, emerging data from comprehensive cancer genome analysis suggest that the burgeoning field will forever alter the future of cancer diagnosis and treatment.
The largest whole-genome sequencing program in the United States, The Cancer Genome Atlas (TCGA), debuted as a 3-year pilot program in 2006. Today, TCGA — an initiative of NCI and National Human Genome Research Institute (NHGRI) — includes more than 150 researchers at dozens of institutions across the nation.
Investigations conducted during the past 6 years have produced valuable insights into the mechanisms of breast, colorectal and lung cancers, and breakthroughs in other malignancies appear within reach.
“The data generated by the TCGA program comprise a vast resource that investigators will be analyzing for years to come,” said Eric D. Green, MD, PhD, director of the NHGRI. “The resource of information … will undoubtedly fuel myriad discoveries by the cancer research community.”
The success comes with a price.
Early opponents balked at the $275 million initial investment required for TCGA. Questions also remain about whether the molecular diagnostics that will be required to keep pace with the discovery of new pathways — as well as efforts to develop new therapeutic agents — will be cost effective.
David S. Strayer
Yet, proponents such as David S. Strayer, MD, PhD, medical director of the molecular analysis lab at the Kimmel Cancer Center of Thomas Jefferson University in Philadelphia, said the return on investment could be significant.
“I don’t know how you put a dollar sign on how valuable these efforts are,” Strayer, who is not directly involved with TCGA, told HemOnc Today. “Developing and testing targeted drugs is a very costly process. The information we get from these studies allows us to target our developmental therapeutics more specifically and accurately. That could make the delivery of care much more efficient and eventually decrease the cost of care. In the long run, the investment in these projects may be money well spent.”
Data integration
When TCGA launched, NCI officials established a goal to collect up to 500 specimens of each tumor type, including breast, colorectal, head and neck, lung, gynecologic, hematologic, skin, thoracic and urologic.
Investigators quickly accrued the target number of samples for common malignancies, but analysis on other less common tumor types began after the collection of 200 to 250 samples.
Still, the sample sizes are large, and normal tissue samples have been collected to allow determination of which genetic changes are only present in tumor cells.
During genomic research, the same samples are analyzed for several different types of genetic information.
Marc Ladanyi, MD, a molecular pathologist at Memorial Sloan-Kettering Cancer Center, said efforts are under way to develop highly efficient technical platforms that can genotype the few hundred genes that underlie the vast majority of human cancers.
Source: Photo courtesy of Marc Ladanyi, MD, reprinted with permission.
This in-depth investigation allows potentially important discoveries, according to Marc Ladanyi, MD, a molecular pathologist at Memorial Sloan-Kettering Cancer Center and co-director of the center’s TCGA group.
“There are many different levels of data integration,” Ladanyi said. “This data integration is helpful in identifying which amplified genes are also overexpressed, or which genes are the likely targets of overlapping genetic deletions, for instance.”
Researchers also are evaluating how the presence of mutations in one gene relate to mutations in other genes, he said. Some mutually exclusive mutations may be in the same pathway, and this points to biological relationships and critical oncogenic pathways.
Also, integrating new data with known pathways may show that a given pathway is almost always deregulated in a given cancer.
“All of this integration can highlight relationships that might not be obvious from looking at only one data type,” Ladanyi said.
‘Treasure trove’ of genetic information
One of the most recent discoveries, related to breast cancer, was uncovered by researchers who used genomic DNA copy number arrays, DNA methylation, exome sequencing, messenger RNA arrays, microRNA sequencing and reverse-phase protein arrays.
Data from five platforms and 825 patients suggested that four primary classes of breast cancer — HER-2–enriched, luminal A, luminal B and basal-like — demonstrate significant heterogeneity on the molecular level.
Investigators also determined that there are several molecular commonalities between basal-like breast cancer tumors and high-grade serous ovarian cancer tumors. Both ovarian and basal-like cancers demonstrated comparably high pathway activity of the HIF1-a/ARNT, MYC and FOXM1 regulatory hubs.
Computational analyses showed that basal-like breast cancer and serous ovarian cancer may be susceptible to compounds that target DNA repair, as well as to agents that inhibit blood vessel growth, thereby cutting off blood supply to the tumor.
Both diseases also demonstrated similar mRNA expression, widespread genomic instability and common gains of 1q, 3q, 8q and 12p, and loss of 4q, 5q and 8p.
“The molecular similarity of one of the principal subtypes of breast cancer to that found in ovarian cancer gives us additional leverage to compare treatments and outcomes across these two cancers,” Harold Varmus, MD, director of NCI, said in a press release. “This treasure trove of genetic information will need to be examined in great detail to identify how we can use it functionally and clinically.”
Targeted therapy for colorectal cancer
A genomic analysis of 276 colon and rectal tumors identified new markers for aggressive tumors, as well as potential therapeutic targets.
The researchers performed DNA copy number, promoter methylation, and messenger RNA and microRNA expression. They observed hypermutation in 16% of colorectal carcinomas. They also observed expectedly high microsatellite instability — usually with hypermethylation and MLH1 silencing — in three-quarters of these carcinomas, whereas one-quarter demonstrated somatic mismatch-repair gene and POLE mutations.
When the hypermutated cancers were excluded, the researchers said colon and rectum cancers had considerably similar patterns of genomic alteration.
Significant mutations occurred in 24 genes beyond the expected mutations in APC, TP53, SMAD4, PIK3CA and KRAS, including in RID1A, SOX9 and FAM123B. Potential new drug targets included ERBB2 and amplification of IGF2, and recurrent chromosomal translocations included the fusion of NAV2 and Wnt pathway member TCF7L1.
“We have a much better understanding of this serious and widespread disease,” said researcher Raju Kucherlapati, PhD, the Paul C. Cabot Professor of Genetics at Harvard Medical School. “The fact that we were able to find so many of these genetic changes that are targetable by a drug is exciting. The most commonly used single therapy for these patients is chemotherapy, and it’s only moderately effective. These findings open a new range of targeted therapies.”
Mutations found in lung cancer
Researchers also have identified potential therapeutic targets in squamous cell carcinoma of the lung, the second most common form of lung cancer.
The analysis included tissue samples from 178 patients. Results indicated that tumor type was characterized by a mean of 360 exonic mutations, 165 genomic rearrangements and 323 segments of copy number alteration per tumor. The researchers observed statistically recurrent mutations in 11 genes. These mutations included mutation of TP53 in nearly all samples.
Researchers identified previously unreported loss-of-function mutations in the HLA-A class I major histocompatibility gene. They also said the NFE2L2 and KEAP1 pathways were significantly altered in 34% of tumors, squamous differentiation gene pathways were altered in 44%, PI3K pathway genes were altered in 47%, and CDKN2A and RB1 were altered in 72% of tumors.
“These TCGA findings should stimulate a wide variety of new clinical trials for patients with squamous cell lung cancer and specific genotypic alterations,” project co-leader Matthew Meyerson, MD, PhD, a professor in the department of pathology at Harvard Medical School and associate professor of pathology at Dana-Farber Cancer Institute, said in a press release. “These will include clinical trials of PI3 kinase inhibitors and other tyrosine kinase inhibitors, as well as ways to use genomics to select patients for trials of lung cancer treatments that dial down the immune response.”
Cost vs. benefit
TCGA researchers also have identified novel genes that may drive a rare, aggressive form of uterine cancer, and data related to other types of cancer are expected soon.
However, the information comes at a price.
Cost concerns can be broken into three primary components.
The first is whether financially and clinically efficient diagnostic and therapeutic strategies will counter-balance the initial investment in TCGA and other genomic research efforts.
The second is the challenge of creating molecular diagnostics that keep pace with the discovery of new pathways. The third is the always costly process of developing therapeutic agents.
With a budget of $270 million per year, TCGA was controversial at the outset, Carl Morrison, MD, DVM, executive director of the Center for Personalized Medicine at Roswell Park Cancer Institute, told HemOnc Today.
“Some people argued that the program involved centralized research that was not distributing the wealth to various cancers,” Morrison said. “The defense of that was that no one could do a study this large in a setting by itself.”
It remains unclear whether the advancements made by genome sequencing projects ultimately will be cost effective, Strayer said.
“What I can say is, without these kinds of investments, pharmaceutical companies would definitely spin their wheels a lot more,” he said.
Charis Eng, MD, PhD, chair of the Genomic Medicine Institute at the Cleveland Clinic, said genomic research “will serve as an infrastructural resource for all clinicians around the world.”
“It will let smart people look at the data and solve problems in a very directed way,” Eng said in an interview. “TCGA will identify patterns of mutations, which will serve as an invaluable compendium for other researchers who can not only replicate and validate the results, but also ask clinically relevant research questions that would not have been possible without these data.”
Morrison invoked the basic tenet that time is money.
“Whole-genome sequencing has sped up our understanding by 10 or 20 years compared to what could have been accomplished by individual investigator-type studies,” he said. “Because of that, people are starting to realize the value of this. Now we have to figure out how to best maximize the usefulness.”
Genomic analysis efforts also have driven widespread technology development, Ladanyi said. That progress has increased the number of genetic markers that can be identified in one test.
“We want to be doing multi-plexed tests where we are looking at a single test that will pick up, for instance, 10 different mutations that are each in 5% of patients, rather than running multiple single-gene tests on every case, each positive in only 5% of cases,” Ladanyi said. “We are moving away from single biomarker tests.”
That will have tremendous benefit in the therapeutic realm, Eng said.
“We are able to see the mutations in much more detail,” Eng said. “We are able to see, right up front, the proportion of patients who will not respond to chemotherapy. This will save millions in trial-and-error treatments.”
Strayer agreed.
“Treating a patient with a therapy that costs $20,000 a month may extend their life a few months,” he said. “If that kind of short-term benefit can be replaced with something that provides longer benefit, it will eventually decrease cost of health care.”
Genomic analysis also may provide benefits unrelated to cancer therapy, Strayer said.
“These projects have led to discoveries in diagnostics that may be applied to other fields or specialties,” he said. “I think of it as an analogy to technologies that came out of the space program. These discoveries may end up reducing the burden on the health care system in ways that we can’t even see yet.”
Emerging data
Most clinicians who spoke with HemOnc Today suggest that the amount of data available now is only a fraction of what will emerge.
Comparative studies are just beginning to appear in peer-reviewed journals.
In a study published in 2012 in the Journal of Clinical Oncology, Sawyer and colleagues evaluated 1,143 women with breast cancer who had completed BRCA1 and BRCA2 mutation screening because they were at high risk for hereditary breast cancer.
The researchers genotyped 22 breast cancer–associated variants and calculated a polygenic risk score. Results from the patient population were compared with 892 controls from the Australian Ovarian Cancer Society.
Women in the hereditary breast cancer cohort had a highly significant excess of risk alleles compared with controls (P=1.0 × 10−16), according to the results. Women who tested negative for BRCA1 or BRCA2 mutations had higher risk scores than mutation carriers (P=2.3 × 10−6).
Women in the BRCA1/BRCA2 negative cohort who were in the top quartile of risk distribution were at an increased risk for breast cancer when aged younger than 30 years (OR=3.37; P=.03) vs. women with low polygenic risk. This same group of BRCA1/2 negative women had higher second breast cancer risk (OR=1.96; P=.02) than women in the low polygenic risk group.
“Genetic testing for common risk variants in women undergoing assessment for familial breast cancer may identify a distinct group of high-risk women in whom the role of risk-reducing interventions should be explored,” the researchers wrote.
In an accompanying editorial, Jennifer K. Litton, MD, and Ana Maria Gonzalez-Angulo, MD, MSc, both from The University of Texas MD Anderson Cancer Center, said the study took several important variables — including prophylactic mastectomy and salpingo-oophorectomy — into account.
“Interestingly, the [polygenic risk score] did not show any significance in women with a history of ovarian cancer as well, and an elevation in [polygenic risk score] was not identified as significant in families with a history of ovarian cancer,” they wrote. “Likely this is a result of the higher rates of BRCA mutations found in this population.”
The study offers a convincing argument that multiple single nucleotide polymorphisms can be used to assess breast cancer risk more effectively, Litton and Gonzalez-Angulo said. This can be done by evaluating links in this high-risk population.
“[The study is] compelling and well done, but to be developed into a meaningful clinical application, their approach will need to address a specific unmet need,” they wrote. “The clinical utility without validation in independent cohorts including other populations is uncertain.”
However, Litton and Gonzalez-Angulo said when the findings by Sawyer and colleagues are validated prospectively and in other populations, they may help predict who may benefit from prophylactic mastectomies.
‘Too much of a good thing’
Despite the vast amount of valuable data that whole-genome sequencing has the potential to produce, it may be “too much of a good thing, at lease in terms of clinical molecular diagnostics,” Ladanyi said.
“On the other hand, there are so many genetic alterations that are useful to screen that it’s becoming more and more difficult for clinical labs to screen for one genetic alteration at a time,” he said. “A current trend is to develop highly efficient technical platforms that can genotype the few hundred genes that underlie the vast majority of human cancers.”
Arul Chinnaiyan
Arul Chinnaiyan, MD, PhD, director of the Michigan Center for Translational Pathology at the University of Michigan Health System, said that personalized therapy may be more appropriately termed “precision therapy,” adding that it is unlikely that the industry will be able to develop completely personalized cocktails.
“It is more likely that we will see a set of mutations and know which combination works in a rational fashion,” Chinnaiyan said.
The pharmaceutical industry will be required to develop therapies that are conditional, that work in some cells but not others, Strayer said.
Obstacles also exist within the pharmaceutical industry, Chinnaiyan said.
“The first issue is that there are not enough drugs out there,” Chinnaiyan said. “Also, some pharmaceutical companies do not allow their drugs to be used in combination with those from a different manufacturer. They are protecting their competitive edge. This will likely change moving into the future, but for now it is still an issue.”
Efforts to establish clinical trials may be problematic, Chinnaiyan said.
“There are challenges in testing hypotheses for targeted therapies for personalized patients,” he said. “How do you pick patients? How do you predict for side effects? We have had challenges assembling the right number of drugs from different manufacturers, and those challenges will likely increase as the number of targets increases.”
Eng framed the issue in more practical terms.
“Prognostic biomarkers will come faster than treatment,” she said, “because the FDA will take longer to approve treatments.”
Strayer spoke to a broader issue of a culture or paradigm change that may be necessary.
“We may be hitting our limit on the ability of cytoreductive therapies,” he said. “If we are going to expand the target range for tumors, we need to develop different kinds of therapeutics. These sequencing projects have given us the information to develop TKIs and other types of sophisticated therapeutics that affect the tumor and also allow us to understand how certain aspects of normal processes are evaluated.”
Despite these advances, even the most advanced therapies may never catch up with the complexity of tumor biology, Strayer said.
“We may need to shift away from the idea of curing people toward an idea that reflects a goal of not eliminating the tumor, but keeping it in check so a patient can survive happily,” he said. “We may need to think in terms of prolonging life comfortably, which may be easier and more realistic than dealing with the complexity of tumor biology.”
Next generation of research
Clinicians and researchers await subsequent data sets from TCGA and other genomic analysis projects. However, the work is far from complete, Eng said.
“We need to be stratifying patient populations for race,” she said. “I would like to see a data set including all Asian patients. I am sure there are studies going on in Japan or in other places.”
Recent efforts have helped to advance the next generation of sequencing technology, in addition to advancing the understanding of tumors, Chinnaiyan said.
“We have pushed this field forward,” he said.
The next generation of research should focus more specifically on whether mutations and mutation pathways are the driver or the passenger of the tumor, Chinnaiyan said.
“Moving into the future, we’ll have more comprehensive mutational landscapes of a tumor,” he said. “This is an exciting time.”
The heterogeneity of malignancies will continue to be an obstacle.
“We are learning, for example, that it is possible that heterogeneity within tumors is in every cell at the beginning, which is different from the idea that it develops over the course of the disease,” Morrison said. “The more we sequence these tumors, we can identify these mutations and give patients first- and second-line therapies right off the bat. The process of figuring out which drugs work has become more sophisticated.”
In the end, strategies must be based on science, Strayer said.
“It is important to develop therapeutics based on knowledge rather than hope,” he said. “The unique thing about these studies is that they are giving us both knowledge and hope.” – by Rob Volansky
References:
Sawyer S. J Clin Oncol. 2012;30:4330-4336.
The Cancer Genome Atlas Network. Nature. 2012;487:330-337.
The Cancer Genome Atlas Network. Nature. 2012;489:519-525.
The Cancer Genome Atlas Network. Nature. 2012;490:61-70.
For more information:
Arul Chinnaiyan, MD, PhD, can be reached at Michigan Center for Translational Pathology, University of Michigan Medical School, 5316 Comprehensive Cancer Center, 1500 E. Medical Center Drive, Ann Arbor, MI 48109-5940; email: arul@umich.edu.
Charis Eng, MD, PhD, can be reached at Cleveland Clinic, Mail Code NE50, 9500 Euclid Ave., Cleveland, OH 44195; email: engc@ccf.org.
Marc Ladanyi, MD, can be reached at Memorial Sloan-Kettering Cancer Center, 1275 York Ave., New York, NY 10065; email: ladanyim@mskcc.org.
Carl Morrison, MD, DVM, can be reached at Roswell Park Cancer Institute, Elm and Carlton streets, Buffalo, NY 14263.
David S. Strayer, MD, PhD, can be reached at 471 Jefferson Alumni Hall, 1020 Locust St., Philadelphia, PA 19107; email: david.strayer@jefferson.edu.
Disclosure: Chinnaiyan, Eng, Ladanyi, Morrison and Strayer report no relevant financial disclosures.
Which approach, whole-genome sequencing or targeted genomic sequencing, holds more promise for cancer research?
Sameek Roychowdhury
The Cancer Genome Atlas and International Cancer Genome Consortium are essential programs in a mission to define the molecular taxonomy of cancer.
Sameek Roychowdhury
Already, they provide a groundwork to identify relevant pathways for novel drugs being developed through precision cancer trials enrolling patients with the putative predictive target.
The TCGA has produced an enormous amount of data that still are being integrated and analyzed, and they continue to support characterization and validation of new pathways for drug development and research. Some, but not all, of these genomic findings are ready for prime time translation into clinical oncology by affecting treatment decisions or enrollment for trials.
For clinical application for patients with advanced cancer, a genomic testing strategy must meet stringent standards (clinical grade), be scalable for thousands of patients, have a streamlined and locked-down analysis and interpretation, and be available for clinical oncologists in less than 2 weeks so that it can employed in a clinically meaningful timeframe.
For academic medical centers and clinical trials cooperative groups deliberating how to incorporate next-generation sequencing strategies, cost and resources are an essential aspect when considering evaluation of thousands of patients. Although the cost of next-generation sequencing continues to decrease, teams must consider resources for sample handling, processing, sequencing and analysis.
Strategies that focus on a targeted gene panel can meet many of these challenges. Research-oriented goals related to discovery could be pursued in parallel with broader approaches such as whole-genome and transcriptome sequencing. Therefore, as we translate genomics with a clinical mission in mind, we should focus on targeted gene panels rather than whole-genome sequencing.
Thus, precision cancer medicine will benefit from genomic sequencing through identifying the right patients for trials, accelerating the drug development process, and contributing to new drug development and rationale for combination therapies.
Sameek Roychowdhury, MD, PhD, is an assistant professor in the department of internal medicine, division of medical oncology, and the department of pharmacology at The Ohio State University Comprehensive Cancer Center — Arthur G. James Cancer Hospital and Richard J. Solove Research Institute. He can be reached at Comprehensive Cancer Center, The Ohio State University, Biomedical Research Tower, Room 508, 460 W. 12th Ave., Columbus OH 43210; email: sameek.roychowdhury@osumc.edu. Disclosure: Roychowdhury reports no relevant financial disclosures.
Sandeep Dave
While targeted sequencing has provided a powerful means to identify the role of mutations, especially in protein-coding genes, these methods are necessarily incomplete.
Sandeep Dave
Structural genetic alterations including translocations and changes in copy number cannot be assayed using these methods. For instance, our data show that PI3 kinase pathway alterations in diffuse large cell lymphoma can occur both as a result of allelic deletion in PTEN or as a result of mutations in PIK3CD (Zhang J. Proc Natl Acad Sci USA. 2013;published online ahead of print Jan. 4).
Whole-genome sequencing provides a powerful method for assaying both structural alterations as well as sequence alterations in the same tumor.
Further, there is a growing awareness of the importance of the 99% of the genome that does not encode proteins. For instance, recent work has found that more than 80% of the genome is actively transcribed as RNA (Dunham I. Nature. 2012;489:57-74).
These transcripts include non-coding RNAs such as microRNAs that regulate the expression of other genes. These non-coding regions and gene regulatory regions are usually not assayed by targeted sequencing, and their contribution to oncogenesis will remain obscure until we systematically apply whole-genome sequencing in tumors.
There remains a significant difference in the cost of whole-genome sequencing and targeted sequencing. However, that difference is growing smaller as the costs of sequencing continue to decline rapidly. It is conceivable that in a few years, whole-genome sequencing will completely supplant targeted sequencing and provide a more complete molecular portrait of cancer.
Sandeep Dave, MD, MS, is an associate professor and director of the molecular genetics and genomics program at Duke Cancer Institute. He can be reached at CIEIMAS, 2177C, 101 Science Drive, Box 33821, Durham, NC 27708; email: sandeep.dave@duke.edu. Disclosure: Dave reports no relevant financial disclosures.