New Insights into Genomics of Pediatric Cancers

Historically, much less has been known about the genetic basis of childhood cancers than that of adult cancers. Recently, two studies analyzed the genomes of more than 2,600 pediatric leukemias and solid tumors, revealing that many of the mutations that drive childhood cancers are not found in adult cancers but can potentially be targeted with existing drugs.

“Both papers together are landmarks,” says Josh Stuart, PhD, of the University of California, Santa Cruz, who was not involved in either study and is co-leader of The Cancer Genome Atlas Pan-Cancer Analysis Working Group. “This represents our first comprehensive view of childhood cancer at the molecular level.”

In the first study, researchers analyzed 961 tumor samples from children with 24 distinct cancer types, focusing largely on brain and other solid tumors. Stefan Pfister, MD, of the German Cancer Research Center and Hopp Children's Cancer Center in Heidelberg, senior author of the study, explains that his team intentionally took a broad approach to identify trends across pediatric cancers.

Pfister's group found that pediatric tumors had a mutation rate 14 times lower than that of adult tumors. However, pediatric patients had a high incidence of germline mutations (7.6%), often in DNA repair genes. Only 30% of the driver mutations identified by researchers had previously been reported in adult cancers. However, 50% of the pediatric tumors contained a genetic change predicted to encode a drug target.

“We now have a much better idea how pediatric tumors are different from adult tumors, how important hereditary predisposition is in pediatric oncology, and how common putative drug targets are across pediatric malignancies,” says Pfister, emphasizing the importance of the latter point in planning future clinical trials.

In the second study, researchers analyzed 1,699 samples from six types of pediatric cancer. Jinghui Zhang, PhD, of St. Jude Children's Research Hospital in Memphis, TN, senior author of the study, explains that her team's goal was to gain an in-depth understanding of leukemias and several solid tumors, including neuroblastoma, Wilms tumor, and osteosarcoma.

Like Pfister's team, they found that many of the driver genes identified in pediatric cancers (45%) had not been previously reported in adult cancers. In addition, they found that 62% of the genetic events in pediatric cancers consisted of structural alterations and copy-number variations, not point mutations. These results show that “clinical genomic testing for pediatric cancer cannot use the same gene panels as those designed for adult cancer,” says Zhang.

Both studies also identified several novel mutational signatures in pediatric cancers by analyzing the pattern of genetic changes present, though Zhang says some of these patterns could simply be sequencing artifacts. If the unique signatures hold up to scrutiny, it raises the question of what is causing them, says Stuart.

Zhang and colleagues detected a known but surprising mutational signature as well: A pattern consistent with UV exposure was present in eight cases of B-cell acute lymphoblastic leukemia. “It's intriguing that UV may cause cancers outside of skin and eye cancer,” observes Stuart.

Both groups emphasize that their data are freely available to the public to facilitate future research, on the Pediatric Pan-Cancer website maintained by Pfister's group and the St. Jude's ProteinPaint Portal maintained by Zhang's group. In particular, how the genetic changes specific to pediatric patients lead to cancer will need to be studied.

Pfister also says that multiple public–private partnerships are devoted to systematically testing which drugs can target the genetic changes identified thus far, including the NCI's Pediatric Preclinical Testing Consortium and the European Union's Innovative Therapies for Children with Cancer Pediatric Preclinical Proof-of-Concept Platform. –Kristin Harper