“Genome” has been the biggest word in cancer research in the last decade. Thanks largely to the high throughput and relatively low cost of “next generation” DNA-sequencing technologies, researchers have screened thousands of tumors for gene mutations that could explain their malignant properties and reveal possible treatment targets.
Sequencing of adult tumors has revealed a broad spectrum of cancer-causing gene mutations. Childhood tumors, by contrast, have turned out to be relatively simple from a genomic point of view. By and large, they harbor few mutations in genes that code for relatively “druggable” targets with discrete effects, like kinases.
“Pediatric tumors are very ‘pure,’ with very low mutation rates,” says Carlos Rodriguez-Galindo, MD, director of the Solid Tumor Center at Dana-Farber/Boston Children’s Cancer and Blood Disorders Center. “To really understand the nature of pediatric cancer, we need to turn to epigenetics and gene regulation.”
Rodriguez-Galindo is not alone in this view. There is a trend afoot in pediatric cancer research: the study of gene regulation and epigenetics is beginning to overshadow classic tumor genetics and genomics.
A deafening genetic silence
Why are pediatric tumors so genomically quiet? The prevailing theory is that children’s cells haven’t accumulated the host of gene mutations that adults acquire from the ravages of time and environmental exposures.
Rather, pediatric tumors arise largely from perturbed regulation of normal tissue development. The mutations that are present tend to arise in genes with wholesale effects on how cellular functions are regulated. Think chromatin regulators, transcription factors, and factors that interact with non-coding elements like enhancers and super-enhancers, all of which impact gene activation, cellular maturity and cell identity.
“History has taught us that the genes mutated in early-onset pediatric cancers are often of fundamental importance to cancer more broadly.”
“The seeds of pediatric tumors are sown during embryonic development,” says Rani George, MD, PhD, a Dana-Farber/Boston Children’s neuroblastoma researcher. “They arise because normal developmental processes have gone awry, and these processes are governed by epigenetic or regulatory mechanisms.”
Peering into the fundamental
Rodriguez-Galindo notes that pediatric tumors’ genomic simplicity can help reveal fundamental concepts of cancer development and the roles regulatory elements play. He cites retinoblastoma, a rare pediatric eye tumor that harbors a single notable genetic alteration, in a gene called RB. RB encodes a regulatory factor influencing the cell cycle and apoptosis. “RB‘s discovery 30 years ago launched the concept of tumor suppression,” he says.
“History has taught us that the genes mutated in early-onset pediatric cancers are often of fundamental importance to cancer more broadly,” says Dana-Farber/Boston Children’s solid tumor researcher Charles Roberts, MD, PhD, who studies the SWI/SNF complex, the first chromatin regulatory factor to be linked to cancer. “Therefore, studying a rare pediatric cancer can be a model for transforming the field.
“Adult cancer genomes’ high complexity makes it challenging to identify contributions from altered chromatin structure and transcriptional regulation,” he adds. “These effects can be identified and studied more easily in pediatric cancer.”
Case-in-point: mutations that affect SWI/SNF’s chromatin-regulating functions first came to light in malignant rhabdoid tumor, an aggressive and rare childhood cancer. But as Dana-Farber/Boston Children’s researcher Cigall Kadoch, PhD, discovered during a postdoctoral fellowship at Stanford, SWI/SNF is broken in at least 20 percent of all human cancers.
What’s also becoming apparent is that in developing tissues, a specific mutation doesn’t mean as much as the context in which it occurs—cell type, maturation level, active gene programs, etc.
“It was calculated 20 years ago that a minimum of five genetic mutations may be needed to give rise to cancer,” Roberts says. “But our newer data raises the possibility that less may be needed. And that’s because not every cell interprets the same mutation in the same way. Many mutations only promote cancer formation when they occur in very specialized cell types.”
Windows of treatment opportunity
Even though the principles of gene regulation aren’t fully understood, academic labs and pharmaceutical companies are already trying to leverage epigenetics and gene regulation to treat intractable childhood tumors.
“One of the exciting things we’re finding is that understanding the interplay between abnormal transcription factors and epigenetic regulators could give you inroads to a new therapeutic,” says Dana-Farber/Boston Children’s cancer genomics researcher Kimberly Stegmaier, MD.
- This past spring, Dana-Farber/Boston Children’s hematologist/oncologist Birgit Knoechel, MD, PhD, published data in Nature Genetics showing that epigenetic mechanisms can influence drug resistance in T-cell acute lymphoblastic leukemia (T-ALL). In her study, drug-resistant T-ALL cells proved vulnerable to compounds that target BRD4, a protein that affects epigenetic structure.
- Another form of leukemia, called mixed lineage leukemia (MLL), may prove vulnerable to drugs that block an epigenetic enzyme called Dot1L. “We now know that these leukemias fully rely on this enzyme and the methylation pattern it generates in order to persist and grow,” said Scott Armstrong, MD, PhD, a former Dana-Farber/Boston Children’s oncologist now at Memorial Sloan Kettering Cancer Center, in an interview three years ago. The enzyme he refers to add tags to histones— proteins that package DNA, thereby altering gene regulation. “While methylation tags on histones are very difficult to manipulate directly, Dot1l is much easier to target therapeutically.”
- MYCN is a transcription factor that’s amplified—but not mutated—in aggressive neuroblastomas, and was one of the first oncogenes to be discovered. It works with other regulators through what’s called a super-enhancer: a stretch of non-coding regulatory DNA that supercharges gene transcription. While MYCN itself is considered undruggable, in a recent Cell paper George and her colleagues showed they could selectively quench its ramped-up activity and force neuroblastoma cells to die off by blocking one of its super-enhancer partners, a protein called CDK7.
Childhood cancers’ genomic purity could also help reveal hitherto unrecognized vulnerabilities in adult tumor cells. Healthy cells often cannot tolerate cancer-causing genetic alterations. To survive, nascent tumor cells have to adapt in ways that allow them to tolerate those cancer-causing events.
“Those adaptations could create new vulnerabilities within the cancer cell that are not present in normal cells,” says Stegmaier, who recently published a genomic survey of Ewing sarcoma (EWS). Her vision is to systematically screen genomically simple cancers for such adaptations, and then determine whether they exist in adult cancers with similar mutations.
“By studying a pediatric cancer such as EWS, where there is the EWS/FLI gene rearrangement and few other mutations, we expect to more easily find the adaptations and vulnerabilities associated with this genetic abnormality,” she says. “We could then address whether genetically complex adult cancers, such as prostate cancer, that have a genetic event similar to EWS/FLI, possess the same vulnerability.”
Time to change focus?
Given that exome sequencing reveals so few actionable mutations, should we abandon the exome and focus our efforts solely on the epigenome, regulatory DNA and functional studies?
Roberts says no. “The exome should be essentially identical in all cells. We can sequence tumor DNA, compare this against the person’s normal DNA, and rapidly identify abnormal mutations. While the large majority of these mutations cannot be targeted with a drug, a small fraction can be, and that list is growing.”
“There’s a lot of evidence to say that we haven’t even scratched the surface of how gene expression is controlled.”
George favors interrogating the tumor cell’s whole genome and epigenome, with an eye toward genes and genetic elements whose expression contribute to cancer initiation and maintenance. However, she notes, “These approaches aren’t trivial. Whole genome sequencing involves a lot of expense and manpower, and ChIP-seq, one of the major tools for studying regulation of gene expression, requires a lot of tumor tissue.”
Additionally, Roberts says, “It’s much harder to measure the epigenome because there is no standard truth. The epigenome of a heart cell looks different than that of a kidney cell.”
Stegmaier says that from a research and discovery perspective, broad exome and genome sequencing can identify relatively infrequent cancer-promoting events. “But should we be doing exomes clinically? It depends on the disease,” she says. “Our current data from EWS suggest you are unlikely to match a genetic mutation with a drug targeting it. In other diseases, like the acute leukemias and osteosarcoma, however, there is a higher incidence of druggable lesions.”
The exome could, though, still have value when looking at relapsed pediatric cancers. “In solid tumors, at least, the therapies can themselves cause mutations,” says George. “So the tumors of patients who have relapsed could be expected to have as many gene mutations as are seen in adults,” some of which might be actionable.
In a sense, epigenetic and regulatory research in pediatric tumors today is where genome sequencing was ten years ago, with teams just starting to plumb tumors’ regulatory DNA and report out new mechanisms and discoveries. Dana-Farber/Boston Children’s researchers are at the forefront of efforts to create this new picture of how pediatric tumors work, and turn that knowledge into potential treatments. It’s an exciting picture, but a daunting task because of how little we know right now.
“We are really at the dawn of understanding how altered chromatin structure and regulatory elements contribute to cancer,” Roberts says. “As time goes on, the hope is that better mechanistic understanding will lead to better therapies. And in a few cases it has already.”
George agrees, adding, “There’s a lot of evidence to say that we haven’t even scratched the surface of how gene expression is controlled. We have so much more to study.”