Collaboration aims to ‘truly accelerate better treatments’ for pediatric cancers

Advances in pediatric cancer research and treatment over the past 50 years have led to dramatic improvements in survival outcomes for this vulnerable population.

There is always room for further progress, however, particularly in the development of more targeted, less toxic therapies.

“We now cure four out of five children with cancer, but losing one out of five kids to cancer is just unacceptable,” John M. Maris, MD, pediatric oncologist, co-head of the Pediatric Cancer Dream Team and Giulio D’Angio chair in neuroblastoma research at Children’s Hospital of Philadelphia, told Healio | HemOnc Today. “Equally important, the way we’ve gotten to that four out of five is with very toxic chemotherapies and radiation therapies. So, of those four kids who survive, at least two or three will have significant long-term side effects of their chemotherapy and radiation therapy.”

Addressing this and other challenges in the fight against pediatric cancer is at the heart of a new academic collaboration involving Broad Institute of MIT and Harvard, Dana-Farber Cancer Institute and St. Jude Children’s Research Hospital.

The Pediatric Cancer Dependencies Accelerator project aims to identify and exploit genetic vulnerabilities in pediatric cancers in order to develop better treatments.

Kimberly Stegmaier

“With our patients now living into adulthood, we are really seeing that although we’re curing these cancers, we’re doing so at a cost,” project co-leader Kimberly Stegmaier, MD, professor of pediatrics at Harvard Medical School, attending physician in pediatric oncology at Boston Children’s Hospital and Ted Williams investigator and vice chair of pediatric oncology research at Dana-Farber Cancer Institute, told Healio | HemOnc Today. “So, our hopes are twofold: to help those patients whom we’re not yet treating effectively with our drugs, and to have better, less toxic treatments for all children with cancer.”

Expanding the map

The three participating institutions are supporting the collaboration with a joint investment of more than $60 million over 5 years. The investment will fund infrastructure development and scientific work of a team that currently includes more than 80 investigators, data scientists and research staff from the institutions.

The project expands upon research initiatives at each institution, including the St. Jude-Washington University Pediatric Cancer Genome Project. This initiative and other large-scale sequencing efforts have uncovered valuable information about the genomic landscape of pediatric cancers.

Among the co-leaders of the collaboration is Francisca Vazquez, PhD, director of the Broad Institute Cancer Dependency Map (DepMap) project. DepMap uses whole-genome CRISPR screens to identify vulnerabilities in hundreds of cancers.

“We’re trying to map vulnerabilities across human cancers,” Vazquez told Healio | HemOnc Today.

The success of DepMap in adult cancers led to the 2015 launch of the Pediatric Cancer Dependency Map project, which served as proof-of-concept for extending the DepMap approach to pediatric cancers.

The first-generation of the map looked at 13 pediatric solid and brain tumor types. Researchers conducted genome-scale CRISPR-Cas9 loss-of-function screening on 82 pediatric cancer cell lines to reveal genes necessary for cell survival.

“Our very first paper in Nature Genetics was the bird’s eye perspective on that first build of the data,” Stegmaier said. “It was incredible. We were finding new candidate drug targets. We’re using that data to better understand the mechanisms of the various diseases that we studied, and we’re building a community of researchers who are very invested in the data.”

The pilot project proved valuable to the research community and inspired the researchers to take the initiative further.

Charles W.M. Roberts

“Of course, 100 cell lines isn’t sufficient to cover the diversity of pediatric cancer,” Vazquez said. “So, when Charles W.M. Roberts, MD, PhD, moved to St. Jude a number of years ago, he talked to the leadership there about the opportunity to expand this dependency map.”

The proposal for the Pediatric Cancer Dependencies Accelerator project came from a computational biologist at St. Jude, according to Roberts, director of St. Jude Comprehensive Cancer Center.

“The Pediatric Cancer Dependency Map had an incredible impact on him, and he was looking for ways to get involved and help accelerate and expand that,” Roberts told Healio | HemOnc Today. “So, 2 and a half years ago, we sent a senior leadership group from St. Jude, including our CEO, to the Broad Institute. We spent the day there, and everyone got excited.”

From there, Roberts and his collaborators began “thinking deeply about the science” in small groups, later initiating larger discussions among the institutions.

“We realized that this is how we can truly accelerate better treatments for childhood cancers,” he said. “All three institutions are very enthusiastic about this, right up to the level of the CEOs of all three places.”

‘Ultra-orphan diseases’

The genetic complexity of pediatric cancers compared with adult malignancies makes collaboration among researchers crucial, according to Maris.

John M. Maris

“Childhood cancer is not a single entity — it’s actually very, very complicated because it’s a collection of ultra-orphan diseases,” said Maris, who is not involved in the accelerator project. “Although the current dependency map includes some childhood cancers, the majority are not there. That is one of the things that makes this so important. It’s a collaborative effort to source childhood cancer models from investigators worldwide to get them onto the screen.”

The beauty of the map is in its simplicity, Maris said.

“The goal is to simply remove every gene in our genome, one by one, to see which ones the cancer cells don’t tolerate,” he said. “The Broad Institute has had this program for a long time, focused mainly but not exclusively on adult cancers.”

Although pediatric tumors generally have few, if any, mutations, there are exceptions, Maris said.

“Twenty-five percent of neuroblastomas have mutations in the ALK gene … but for the vast majority of childhood cancers, it is so-called epigenetic developmental programs that are deregulated,” he said.

Identifying these deregulated expressions and determining their importance to the cancer is part of what the dependency map will do. Based on what he has seen in neuroblastoma, Maris believes the project will have a significant impact in advancing the field of pediatric oncology.

“The cancer I study most frequently, but not exclusively, is neuroblastoma, and investigators in neuroblastoma have been quite successful over the years in establishing cell lines,” he said. “Neuroblastoma was one of the very early childhood cancers studied at the Broad Institute. So, there has been dependency data on this disease for a few years now, and we use it all the time in my lab. If a student in my lab says they have discovered this protein or that protein that they think is important, my first question is often, ‘Is it a dependency?’”

Research objectives

The collaboration is designed to pool the academic leadership, technical expertise and institutional resources of the three institutions to address knowledge gaps in the biology of pediatric cancers and identify more effective approaches to treatment.

The project aims to expedite progress in the development of treatments for aggressive pediatric malignancies through the following research goals:

• Develop and use genome-editing techniques to determine hidden vulnerabilities or dependencies in a range of high-risk childhood brain, solid and blood cancers.
• Utilize new technologies to assess and describe the genetic and epigenetic landscape of pediatric cancers.
• Develop model systems where they are currently lacking for high-risk pediatric cancers with poor outcomes.
• Find effective combination treatments, identify mechanisms of drug resistance and shorten the timeline for developing new therapies.
• Develop computational systems for mining and assimilating data and create software tools for data sharing.

Understanding the mutations in childhood cancers and how they can be effectively targeted with drugs is one of the main challenges the Pediatric Cancer Dependencies Accelerator project aims to surmount, Stegmaier said.

“The mutations we find [in pediatric cancer] tend to encode for proteins that are difficult drug targets,” she said. “For example, a very common theme in childhood cancers is the presence of fusion oncoproteins that tend to involve transcription factors that are difficult to make drugs against.”

The project’s early efforts improved understanding of pediatric cancers with good cell-line models, but the expanded project will tackle some of the more challenging diseases, according to Stegmaier.

“There were a number of high-risk diseases we were not able to represent in the first build of the data,” she said. “That is a big focus of the collaboration now across our institutions — to say, ‘OK, we evaluated some of the low-hanging fruit, but now we have to do some of the harder work to tackle diseases such as pediatric brain tumors.’”

Researchers may utilize CRISPR screens of cell lines in some of these diseases, but in others, different types of screens may be needed.

“Some of the pediatric high-risk leukemia subsets were also not represented,” Stegmaier said. “So, our goal now is to pool forces across the St. Jude, Dana-Farber and Broad Institute teams to really drill down on those high-risk leukemia types.”

Identifying and understanding biomarkers of response is another key goal of the project.

“One of the other efforts is going to be to do a deeper characterization of the epigenetic state of these childhood cancers,” Stegmaier said. “We think that knowing this will help us find better biomarkers as to why cells may be dependent upon one gene or another. We also think this will be important for understanding the heterogeneity of childhood cancers and some of their basic mechanistic underpinnings.”

Addressing limitations

Developing targeted therapy and/or immunotherapy for the pediatric cancer population is an arduous endeavor, Roberts said.

“New therapies take decades when you work gene by gene,” he said. “We know that one drug is never enough, so identifying effective combinations takes even longer. The cancer genome projects at St. Jude and elsewhere have been great at defining the landscape of cancer, but a large number of these mutations are not druggable.”

Understanding and overcoming drug resistance is another goal of the Pediatric Cancer Dependencies Accelerator project, Roberts said.

“We’ve all had the experience of family members or friends who start treatment for a cancer, whether it’s an adult or pediatric cancer, and it looks good — the tumor is shrinking,” he said, “then 1 month or 6 months later it comes racing back. So, the resistance is there.

“Cancer genome projects at St. Jude and elsewhere have been great at defining the landscape of cancer, but a large number of those mutations are not druggable,” he noted. “Many of them are tumor suppressors that are gone, so how do you drug that? It’s not even there anymore. In addition, the oncogene mutations often turn out to be a real challenge for the chemists.”

Other limitations include the lack of laboratory models for many childhood cancers and difficulty sharing increasingly complex data on pediatric cancer dependencies, he said.

“So, we asked whether there might be an ability to leapfrog each of these challenges and build upon the pediatric dependency project,” Roberts said. “The goal of it is to go from the lab to the clinic faster.”

Roberts said the pediatric dependency map started by conducting functional, genome-wide CRISPR screening on hundreds of cell lines.

“Cancer sequencing projects often ask the question, ‘What genes, when mutant, cause cancer?’ The functional genomic screening in the pediatric cancer dependency accelerator asks the question, ‘What genes, when turned off, kill cancer?’ It’s designed to reveal the gene targets that make cancer cells vulnerable,” he said.

This approach has proved effective and yielded a large number of publications in high-impact journals, according to Roberts.

In addition to searching for new vulnerabilities, the researchers are also working to identify more effective drug combinations.

“It turns out that if we can take a drug that we know works but is not enough, and we perform CRISPR screening to systematically turn off every gene, we can identify cooperative targets and also reveal mechanisms of resistance,” he said. “Collectively, we’re so excited about Pediatric Cancer Dependencies Accelerator because we think it addresses each of these limitations.”

‘Cutting-edge’ data sharing

Another key resource being leveraged for the accelerator project is St. Jude Cloud, a cloud-based, data-sharing genomic platform for pediatric cancer.

St. Jude Cloud provides gene expression data collected from analysis of tumor sample RNA sequencing data for more than 2,000 pediatric tumor samples. This data is available to all users and represents a wide range of pediatric tumor subtypes.

“Federated data-sharing can be a real problem in adult cancer settings, as well, as far as how to share data and integrate it across disease types,” he said. “Because we have the Broad Institute, because we have St. Jude with St. Jude Cloud, we have incredible expertise,” Roberts said. “We’re very excited about the federated science and the tools we’re developing and freely sharing. The NCI has funded us to export St. Jude tools to NCI because of how useful they are. We really think this is an opportunity to be cutting edge in data-sharing that will go beyond pediatrics.”

Stegmaier agreed that the collaborative nature of the project, as well as the unique resources each individual institution brings to bear, will make this initiative very powerful in advancing the fight against pediatric cancers.

“The idea of building a community of people working collaboratively is one of the main goals of the project,” Stegmaier said. “Our goal is to cure children with cancer with better drugs, and that takes a very large community.”

References:

For more information:

John M. Maris, MD, can be reached at Children’s Hospital of Philadelphia, 3401 Civic Center Blvd., Philadelphia, PA 19104; email: [email protected].

Charles W.M. Roberts, MD, PhD, can be reached at St. Jude Children’s Research Hospital, 262 Danny Thomas Place, Memphis, TN 38105-3678.

Kimberly Stegmaier, MD, can be reached at Stegmaier Labs, Dana-Farber Cancer Center, 450 Brookline Ave., Boston, MA 02215; email: [email protected].

Francisca Vazquez, PhD, can be reached at Broad Institute of MIT and Harvard, Merkin Building, 415 Main St., Cambridge MA 02142; email: [email protected].