St. Jude-Washington University Pediatric Cancer Genome Project yields ‘important new findings’ in childhood malignancies

In collaboration with Washington University School of Medicine in St. Louis, St. Jude Children’s Research Hospital created the Pediatric Cancer Genome Project, which has led to sequencing the complete normal and cancer genomes of more than 800 children and adolescents with 23 different childhood cancers.

James R. Downing

The project has led to new areas of research involving high-risk leukemia, as well as brain and solid tumors in pediatric patients.

HemOnc Today spoke more in-depth about the Pediatric Cancer Genome Project with James R. Downing, MD, president and CEO of St. Jude Children’s Research Hospital.

Question: How did the Pediatric Cancer Genome Project come about?

Answer: St. Jude has been doing a fair amount of work in the genomics of pediatric cancer, using the contemporary technology available. My own group has done work on pediatric acute lymphoblastic leukemia using gene expression arrays and molecular-based approaches. Other groups have conducted work on brain tumors, solid tumors and pharmacogenomics. For more than a decade, we also had been doing target gene-sequencing and collaborating with projects such as the TARGET Initiative through the NCI. We saw the promise in gene sequencing, and the need to have a project focused solely on pediatric cancer. In 2009, we gathered the faculty of St. Jude together at an offsite retreat where we laid out why St. Jude should conduct a pediatric cancer genome project. In the end, there was overwhelming support from the faculty, and everyone agreed we should do it. At that point, we had only an idea but lacked true practical experience in the methodologies and data analysis approaches that would be needed to accomplish such a project. We laid out a rather detailed approach, exploring all possible options for tumor selection, data acquisition and data interpretation. In the end, we estimated that the project would cost approximately $60 million and that it would be best pursued in collaboration with one of the major DNA sequencing centers that participated in the Human Genome Project. We decided to partner with the Washington University Genome Center and initiated the project in February 2010 with the goal of conducting whole-genome sequencing on 600 pediatric cancer types and matched germ lines (normal tissue) samples and to complete the effort in a three-year period.

Q: Can you describe the need for this initiative, as well as its mission?

A: When we began the project, only one human cancer — an adult acute myeloid leukemia — had been analyzed by whole-genome sequencing. That effort was performed by Washington University. Based on that accomplishment and on the information that had emerged from genomic analysis of human tumors using lower-resolution approaches, we believed that defining the landscape of somatic and germ line mutation that underlie pediatric cancers would transform our understanding of this disease and would alter the way we diagnose, risk stratify and ultimately treat children with cancer. We, therefore, mapped out a project that was extremely bold in scope: to perform whole-genome sequencing on 600 pediatric tumor and matched normal tissue samples (1,200 genomes) to map the landscape of mutations and to focus our efforts on those cancers where our current therapies were insufficient. We began by recruiting people, buying the equipment needed, establishing the operational structure for the collaboration and identifying the initial tumors to be sequenced. We initiated the project with a goal of performing 50 whole genomes during the first year and then accelerated the rate of sequencing to 250 whole genomes the second year and 300 the third year. In the end, we sequenced over 700 tumors and accomplished this under budget and within the established timeframe. The whole-genome sequencing was done at Washington University, and almost all of the validation work was conducted at St. Jude. One of the key decisions we made up front was that St. Jude and Washington University were both going to independently analyze the data. We would compare results and re-sequence everything to validate which lesions were real, which lesions were not real and then modify algorithms as we went forward. This approach was key to the project and resulted in our efforts, leading to the development of some of the most accurate algorithms for detecting single nucleotide variations, structural variation and DNA copy number variations.

Q: What makes this effort different from other research efforts that target childhood cancers?

A: This project is large in scope. It involved 350 people across the two institutions. In every cancer we examined, we made fundamental insights into the landscape of mutations that underlie the individual cancer subtypes. In addition, for some cancers, we discovered new clinically relevant cancer subtypes, prognostic markers and/or therapeutic targets. The effort has had a major impact on the kind of research that is now being pursued not only at St. Jude but really around the globe. Our DNA sequence database has been accessed by over 100 institutions. Some of the findings generated through the project have led to new laboratory research programs at our institution. The Pediatric Cancer Genome Project was an amazing effort that brought together people from across many different disciplines to focus on pediatric cancer. At St. Jude, would routinely have clinicians, molecular pathologists, pharmacists, chemical biologists, basic scientists and computational biologists discussing the findings and their impact on clinical practice and fundamental biological studies. It was the exact kind of exchange that drives innovation and discovery that will move medicine forward.

The data generated during the initial project continues to yield important new findings, with publications continuing to be published. We initiated a second phase of the project in 2013 that cost $30 million and was focused on defining the epigenetic landscape of pediatric cancer and to move next-generation sequencing into our clinical operations. The data from this second phase will start to emerge in publications later this year.

Q: Which malignancies are the primary areas of focus, and why were they chosen?

A: About one-third of cancer seen in children are hematopoietic malignancies; one-third are tumors of the central nervous system (CNS); and one-third are non-CNS solid tumors. We decided on several criteria across the various tumor types. First, we wanted to work on tumors that had a poor outcome with contemporary therapies, or “high-risk” tumors were our current approaches are inadequate. Secondly, we wanted to focus on tumors where we had access to experimental systems relevant to the tumor that could be used to directly assess the functional significance of the tumor. St. Jude has spent considerable efforts to develop murine model systems for many of the cancers we see in children and we felt these established resources if appropriately used would add significant value to the interpretation of the functional significance of the identified somatic and germ line mutations. Lastly, we focused on tumors where we had a sufficient number of primary tumor samples and associated outcome data on the individual patients. Early on, we had discussions on how many tumors are needed to identify the presence of mutation that would be predicted to occur at a specified frequency. For many of the high-risk tumor types, we did not have the calculated number of tumors. This did not discourage us. We decided to begin with a limited number of tumors of specific cancer subtypes and see what emerged. We began with some very rare tumor types and what emerged was amazing — clear novel insights that significantly advanced our understanding of the biology of the specific cancer subtype.

Q: Can you talk about some of the key findings so far?

A: One of the most amazing findings was in a rare but lethal type of pediatric tumor called diffuse intrinsic pontine glioma. This brain tumor typically is seen in children around the age of 5 years and has an OS of less than 5%. Because of its location within a critical region of the brainstem, the simple act of biopsy the tumor can kill a child. One of our investigators set up a Web-based autopsy protocol where parents of children with this disease could consent for an autopsy at the time of the child’s death and then have the tissue sent to St. Jude. Parents from around the globe enrolled in this study. The samples collected through this protocol along with other samples were then sequenced as part of the Pediatric Cancer Genome Project. Remarkably, we discovered that 80% of these tumors had the exact same mutation, a change that resulted in a lysine to methionine change in Histone H3, one of the key proteins that controls the packing of chromosomes and thus which genes are expressed and which are shut off. This mutation had never been seen in any other tumor in either pediatric or adult patients. This was an unbelievable finding that has spread rapidly around the world and today a large number of groups have used this information to advance our understanding of gene regulation and how alteration through this mutation and through other mechanisms contribute to the development of cancer. Our efforts also have led to the development of computational approaches to improve the ability to analyze DNA sequencing data. The algorithms we have developed are now widely used across the globe. In addition, the primary data that was generated through the effort had served as an invaluable resource for the entire scientific community with it being accessed daily by investigators from around the globe. People have used our data to make significant advance on a variety of other human diseases including degenerative neurological conditions.

Q: How do you see the project evolving/expanding in the near future?

A: We opened a new clinical protocol a little over a month ago, Genomes for Kids, and have enrolled six patients to date. Our goal for this project is to perform whole-genome, whole-exome and RNA sequencing on every child with a diagnosis of cancer who enters St. Jude. We will perform an integrated analysis and the place the result into the patient’s medical record. The data will be used for discovery but ultimately will directly influence clinical decisions. In parallel with the Genomes for Kids protocol, we have established a hereditary cancer predisposition clinic that is staffed with pediatric hematologists, oncologists, clinical genetic counselors and ethicists. This clinic will work closely with the genomics team and the pediatric hematologist/oncologist and will follow any children who are found to have a germ line mutation that predisposes the child to cancer. For these children, the entire family unit will be tested and any members found to carry the mutation, we will follow them with comprehensive surveillance testing. Our efforts are beyond what is currently the standard of practice but it is our hope that these research efforts will help to define how this technology should be incorporated into the diagnostic work-up and treatment of every child with cancer. This is a very exciting integrated effort that is unparalleled in scope. – by Jennifer Southall

For more information:

James R. Downing, MD can be reached at St. Jude Children’s Research Hospital, 262 Danny Thomas Place, Memphis, TN 38105; email: media@stjude.org.

Disclosure: Downing reports no relevant financial disclosures.