NEW YORK (GenomeWeb) – A pair of new studies by independent research teams is offering a look at the molecular alterations found in thousands of childhood tumors from a broad swath of cancer subtypes.
An international team that includes members of the International Cancer Genome Consortium (ICGC) brought together single nucleotide variant, small insertion and deletion, copy number, and structural variant profiles for 961 tumor and matched normal samples from 914 children, adolescents, and young adults under the age of 25 who were diagnosed with cancer from 24 types for one of the studies, published online today in Nature.
That set spanned 879 primary tumors and 82 relapse tumors, including 547 tumors profiled by whole-genome sequencing, 414 exome-sequenced tumors, and additional samples assessed by low-coverage genome sequencing.
The team estimated that roughly half of primary childhood tumors contain potentially targetable somatic mutations, based on the cases profiled, while more than 7 percent of the cases appeared to involve pathogenic germline variants in cancer risk genes such as TP53.
When the researchers compared the recurrently mutated genes, mutation signatures, and mutation frequencies detected in the childhood tumors with those in adult tumors profiled for the Cancer Genome Atlas project (TCGA), they saw molecular distinctions between the pediatric and adult cases.
For example, primary tumors from the childhood or young adult cases had far lower mutation frequencies, on average, though more highly mutated or hyper-mutated tumors appeared with increasing age and in specific pediatric cancer subtypes.
The team also narrowed in on at least one mutational signature — found in some pediatric ependymoma and atypical teratoid rhabdoid tumors — that has not been described in adult tumors so far.
"The multiple differences found compared to previous studies of adult tumors emphasize the need to consider pediatric cancers separately, further demonstrating a need for mechanism-of-action driven drug development for pediatric indications," the authors noted.
In a related Nature study, a team led by St. Jude Children's Research Hospital researchers described genomic and transcriptomic profiles identified in 1,699 matched tumor and normal samples from children diagnosed with pediatric leukemia or solid cancer before the age of 21. These cases, enrolled through Children's Oncology Group clinical trials, represented half a dozen cancer histotypes: B-lineage acute lymphoblastic leukemia (B-ALL), T-lineage ALL, acute myeloid leukemia, neuroblastoma, Wilms tumor, and osteosarcoma.
Based on whole-genome sequence, exome sequence, transcriptome sequence, and SNP array data, the researchers highlighted 142 driver genes across that childhood cancer set. Again, though, they saw differences between pediatric and adult cancer cases: along with reduced somatic mutation rates overall in the pediatric samples, they found that more than half of the pediatric driver genes were distinct from those described in prior pan-cancer analyses of adult tumors.
Some 62 percent of driver events in childhood cases involved copy number or structural variant shifts, the team reported. It also unearthed two new and nine known mutational signatures, including an ultraviolet light-linked signature found in a handful of B-ALL samples.
In their subsequent analyses, the researchers uncovered 21 recurrently altered biological pathways in a subset of 682 leukemia samples and 236 solid tumors, assessed by whole-genome sequencing or by exome sequencing in combination with SNP array profiling. By bringing together point mutation and RNA sequence data, meanwhile, they narrowed in on apparent allele-specific expression profiles in the pediatric tumors.
The researchers are making the somatic variants used for this study available through the National Cancer TARGET Data Matric and their ProteinPaint portal, which has an interactive heat map viewer for exploring mutations, genes, and pathways across the six histotypes, noting that the site also houses somatic alterations identified in the 961 tumors profiled for the ICGC-led pediatric pan-cancer analysis.
"We anticipate that these complementary pan-cancer datasets will be an important resource for investigations of functional validation and implementation of clinical genomics for pediatric cancer," they concluded.