NEW YORK (GenomeWeb News) – A study in Nature Genetics suggests alterations to two chromatin-remodeling genes — ARID1A and ARID1B — are linked to particularly poor neuroblastoma outcomes.
The ARID1A/ARID1B association with survival time was discovered during a large-scale genomic analysis of the childhood cancer, which affects a set of neural crest cells destined to be part of the peripheral sympathetic nervous system.
"These two genes function in a group of genes that seems to play an important role in neural cell behavior," Children's Hospital of Philadelphia pediatric oncologist Michael Hogarty, the study's co-senior author, said in a statement, "and we will now work to discover if this insight may open up new treatments for children with tumors having these mutations."
Hogarty and the Johns Hopkins Kimmel Cancer Center's Victor Velculescu led a research group that relied on high-coverage genome sequencing, low-coverage genome sequencing, exome sequencing, and/or more targeted approaches to unearth genetic abnormalities ranging from point mutations to chromosomal changes in tumors from more than 70 children with neuroblastoma.
While some of the recurrently mutated genes they identified had been linked to neuroblastoma before, others were new. The ARID1A and ARID1B alterations proved to be among the most tantalizing new candidates, study authors noted, because these appeared to coincide with significantly worse treatment responses and reduced survival times.
"The median survival of individuals with ARID1 gene alterations was lower than that observed for any other genetic alteration assessed … providing a potential marker for early therapy failure and disease progression," they wrote.
Neuroblastoma cases represent roughly 7 percent of cancers diagnosed in childhood, usually showing up before a child's fifth birthday. Though prognoses tend to be better for children diagnosed before the age of 18 months old, the researchers explained, other cases are far more difficult to treat. Consequently, neuroblastoma cases make up a disproportionately high proportion — up to 15 percent — of childhood cancer deaths.
Past research has uncovered copy number changes in the germline that can bump up neuroblastoma risk. And a few genes with recurrent somatic mutations have been detected in neuroblastoma tumors, including ALK, an apparent oncogene activated in 8 percent or so of primary neuroblastoma tumors, and MYCN, which is amplified in around one-fifth of tumors and tends to point to more advanced or aggressive disease.
Earlier this year, members of the St. Jude Children's Hospital-Washington University Pediatric Cancer Genome Project published a study in the Journal of the American Medical Association delineating recurrent neuroblastoma-related mutations or structural changes to a gene called ATRX. Results from that study indicated that ATRX alterations were especially common in individuals diagnosed with neuroblastoma during adolescence or early adulthood, but absent in those diagnosed during infancy.
Even so, many questions remain about the full set of somatic glitches that can occur in neuroblastoma and their potential prognostic significance.
For the current study, researchers combined multiple genomic approaches in an effort to get as comprehensive a view of neuroblastoma changes as possible, using samples from children diagnosed with neuroblastoma before the age of six years old.
Using Illumina's HiSeq or GAIIx instruments, the team did high-coverage whole-genome sequencing (31-fold coverage, on average) on six tumor-normal pairs — data that was used to find sequence-level mutations as well as chromosomal changes.
Those samples were also assessed by exome sequencing, as were 10 more tumor-normal pairs. After nabbing coding sequences with Agilent's SureSelect system, researchers again used Illumina instruments to sequence the exomes, generating enough sequence to cover each at a depth of 94-fold, on average.
In addition, the team did low-coverage (10-fold) whole-genome sequencing on another 26 tumors to unearth even more structural rearrangements and copy number changes in the neuroblastoma genomes, along with more targeted analyses on dozens of tumor samples.
Within protein-coding sequences, researchers reported, each tumor contained 13 somatic substitutions or small insertion/deletions, two focal amplifications, and four structural rearrangements, on average.
Genes that were recurrently affected by mutations, small insertions/deletions, or larger rearrangements included some with past ties to neuroblastoma or other cancers, such as MYCN or ALK. But new candidates turned up, too, including ARID1A, ARID1B, VANGL1, and ZHX2.
Among the most frequently mutated new candidates were ARID1A and ARID1B, the team noted. The chromatin modifying genes were altered in some 11 percent of neuroblastomas and appeared to correspond with tumors that were more aggressive.
For instance, when they considered clinical outcomes for cases involving ARID1A glitches, the researchers saw that those children had average overall survival time of 386 days. In contrast, those free from both ARID1A and ARID1B alterations survived 1,689 days, on average.
For a few cases, the team went a step further, sorting through the repertoire rearrangements in tumors to find patient-specific alterations to track over time in circulating tumor DNA from the patient's blood samples.
And when four of the individuals went on to get immunotherapy for minimal residual diseases, researchers used these circulating tumor biomarkers as a window into each individual's treatment response or recurrence, using analyses similar to those that Velculescu and his colleagues described in Science Translational Medicine last week.