NEW YORK (GenomeWeb) – Members of the Pediatric Cancer Genome Project have started characterizing the prevalence of germline alterations in known cancer predisposition genes in individuals who develop cancer before age 20.
As they reported online last night in the New England Journal of Medicine, the researchers did genome sequencing and/or whole-exome sequencing on germline samples from 1,120 pediatric cancer cases. Starting from analyses on almost 600 suspicious genes, they found that 8.5 percent of the individuals carried risk mutations in genes previously implicated in autosomal dominant cancer predisposition syndromes.
Germline mutations in the same genes were far less frequent in germline samples from individuals enrolled in the 1000 Genomes Project or in a study of autism spectrum disorder (ASD), the team reported. In those groups, seemingly risky mutations in the genes showed up 1.1 percent and 0.6 percent of the time, respectively.
Tumor genome sequencing — such as that being done for the broader Pediatric Cancer Genome Project — is still expected to offer more clues about mutations with potential relevance for targeted treatment, explained first author Jinghui Zhang, chair of computational biology at St. Jude Children's Research Hospital.
Nevertheless, the new analysis suggests risky germline mutations may be more common than previously appreciated. And that may have implications for understanding disease and for finding those who could benefit from enhanced screening or risk reduction surgery, Zhang told GenomeWeb.
Researchers have searched for germline risk factors within several individual types of pediatric cancer in the past, the researchers noted. But the current effort is believed to be the most comprehensive, covering the main pediatric cancer subtypes.
The team did whole-genome sequencing on germline samples from 595 pediatric cancer patients. The protein-coding portions of the germline genome were assessed in 456 pediatric cancer cases, while 69 patients had both their germline genomes and exomes sequenced.
Within these sequences, the researchers assessed 595 genes linked to cancer, focusing in particular on 60 autosomal dominant hereditary cancer predisposition syndrome-associated genes.
The team found that 95 of the patients had mutations classified as pathogenic or probably pathogenic in 21 of these genes.
The researchers had access to family histories for just a few dozen of the pediatric cases. And a familial predisposition to cancer was apparent in just 40 percent of those individuals, with even fewer cases fitting anticipated cancer predisposition syndrome patterns.
When the team tested 966 unselected individuals from the 1000 Genomes Project, it uncovered pathogenic or probably pathogenic changes to the genes in 1.1 percent of individuals.
In a group of individuals from another cohort brought together for an ASD study, which included 515 individuals with ASD and 208 individuals without, the group saw apparently risky germline changes to the same gene set in 0.6 percent of individuals.
Although germline mutations in TP53 are thought to be relatively rare in the germline of adults who develop cancer, Zhang noted, such glitches were common in the childhood and adolescent cases, accounting for more than half of the predisposition mutations detected.
In a study published in Nature Genetics in 2013, Zhang and her colleagues found that TP53 mutations were also over-represented in the germlines of individuals who developed a form of leukemia called hypodiploid acute lymphoblastic leukemia (ALL).
As such, the over-representation of TP53 mutations found in the analysis may partly reflect the fact that the pediatric cancer cohort tested included proportionally more hypodiploid ALL and adrenocortical carcinoma tumors than typically occur in pediatric cancer cases in the US — a bias stemming from the team's focus on cancer types with poor outcomes.
After TP53, the most common germline mutations in the study fell in the APC, NF1, PMS2, RB1, and RUNX1 genes, though germline ALK mutations were over-represented in individuals who developed neuroblastoma.
The team also detected germline glitches in some genes not usually linked to pediatric cancer risk such as EGFR, PALB2, and the BRCA1/2 genes.
"This was the first time we've done a systematic evaluation [of these risk genes] in pediatric cancer," Zhang said. "The significance of these will require more study."
She noted that the germline mutation frequency described in the current pediatric cancer analysis prove to be an underestimate, since the full suite of risky mutations has not yet been accounted for.
But because genome or exome sequences are now available for the pediatric patients included in the study, the researchers plan to re-assess the data down the road as new germline alterations with potential ties to cancer risk are detected in other studies.
"We like to take the whole-genome approach so that you don't need to draw blood from patients over and over again," she noted. "When new variations in genes are discovered, we can go back."
Germline mutation screening will also make up one aspect of the Genomes for Kids program, a prospective study that will involve genome or exome sequencing on samples from new pediatric cancer cases at St. Jude.
Zhang noted that she and her colleagues are interested in exploring germline mutations related to pediatric patient outcomes and survivorship. They also hope to translate the approach used in the cancer study to other, non-malignant childhood conditions.
In an accompanying editorial in NEJM, Children's Hospital of Philadelphia's John Maris cautioned that the current study's focus on children with high-risk cancer types might skew the findings somewhat, making autosomal dominant predisposition mutations appear more common in the germline than they might be in a broader population of childhood cancer cases.
Still, Maris did not discount the potential risk posed by germline mutations in general. He noted that childhood cancers "remain fertile ground for making the discoveries that will be essential for furthering our understanding of susceptibility to cancer in general," since these cancer cases occur in a context of more limited exposure to potential risk factors in the environment.
"[T]hese data provocatively suggest that germline sequencing should be routinely incorporated into clinical care," he wrote, "given that many pediatric patients with cancer have a genetic predisposition."
Maris also emphasized the need for characterizing germline mutations further, considering their interactions with the rest of the genome, and exploring the role of mosaic germline risk mutations, which were found in a handful of the Pediatric Cancer Genome cases.