NEW YORK (GenomeWeb News) – In a paper appearing online today in Nature, researchers from the Children's Hospital of Philadelphia and elsewhere reported that they have found germline copy number variations linked to the childhood nervous system cancer neuroblastoma.
The team looked for CNVs by genotyping DNA in blood samples from nearly 850 children with neuroblastoma and about 800 healthy controls at more than a half a million SNPs. In so doing, they found four parts of the genome that were prone to deletions in neuroblastoma cases — results that they verified in a pair of follow-up studies.
One of these deletions — a chromosome 1 deletion containing a previously unrecognized neuroblastoma break point family gene called NBPF23 — appeared to be a true germline variant. The team's subsequent experiments indicate that the gene is active in early nervous system development, offering clues about neuroblastoma biology.
"We knew very little about the genetic predisposition to neuroblastoma just a few years ago," senior author John Maris, director of the Children's Hospital of Philadelphia's Cancer Center, told GenomeWeb Daily News, noting that a great deal of information has come out of research studies done over the past year and a half or so.
Last spring, he and his team published a study implicating the anaplastic lymphoma kinase gene ALK in inherited neuroblastoma. And earlier this year they reported on risk alleles for aggressive neuroblastoma in a gene called BARD1.
Neuroblastoma, a sympathetic nervous system cancer, is the most common type of solid cancer in children. It usually arises in children who are less than five years old, although the age of onset, disease progression, and tumor biology are heterogeneous.
About one percent of neuroblastoma cases are inherited in an autosomal dominant manner, but most arise with no family history of the disease. Maris and his team speculated that this apparently spontaneous form of the disease may involve many common DNA variations, each incrementally increasing neuroblastoma risk.
"This current publication is part of our effort to understand the sporadic form of the disease," Maris said. "We've always suspected that it's going to be a combination of SNPs and CNVs."
In the past, when researchers studied copy number alterations in cancer they typically focused on somatically acquired aberrations in tumor DNA, lead author Sharon Diskin told GenomeWeb Daily News. But she was interested in understanding how germline CNVs influence disease risk as well.
To investigate whether germline variants detected in the blood can predispose individuals to neuroblastoma, the researchers first isolated DNA from blood samples taken from 1,032 individuals with neuroblastoma and 2,043 healthy controls and genotyped the samples at about 550,000 SNPs using the Illumina Infinium II HumanHap550 BeadChip. After tossing out data that didn't meet their quality control criteria, they were left with a discovery set of 846 cases and 803 controls.
When they looked at copy number changes associated with the SNPs, they pinpointed four deletions that were significantly associated with neuroblastoma: one each on chromosomes 1 and 14 and two on chromosome 7.
The chromosome 1 deletion spanned 143,000 base pairs or so of the 1q21.1 region — an area previously implicated in a range of other diseases from schizophrenia to congenital heart defects. For instance, a New England Journal of Medicine paper published last fall by University of Washington researcher Evan Eichler and his team linked copy number changes in 1q21.1 to developmental diseases such as mental retardation, autism, and congenital abnormalities.
The 1q21.1 loci associated with neuroblastoma are relatively far from those implicated in autism, schizophrenia, and other conditions, Maris noted. Still, he added, it is curious that many chromosome 1q21.1 CNVs seem to influence nervous system function.
"We're very interested in this gene family and what is their role in development," Maris said.
The other three deletions affected T-cell receptor genes loci.
All of the associations held up in follow-up studies involving two different groups of patients — one with 363 Caucasian cases and 1,139 controls and another 232 Caucasian cases and 2,218 controls. The team did not detect any duplications reaching genome-wide significance.
For a subset of 226 neuroblastoma patients, the researchers evaluated matched blood and tumor samples from the same individual to gauge which CNVs represented true germline variants. Chromosome 1q21.1 deletions were detected in both blood and tumor samples, Diskin said, suggesting they are germline variants.
In contrast, deletions in the T-cell receptor genes, which had the most significant associations in the GWAS, were not present in tumor DNA. Based on these results, the researchers speculated that changes in TCR genes may be indicative of an immune response related to neuroblastoma.
To determine whether the 1q21.1 deletion is inherited, the researchers genotyped 713 parent-child trios, again using the Illumina HumanHap550 array. The 1q21.1 deletions showed up in 125 children and 123 parents, suggesting the deletion can be passed from one generation to the next. While none of the individuals in the trios had a history of cancer, more than half of the children had a history of autism.
Within the 1q21.1 region, the researchers detected a new transcript — dubbed NBPF23 based on its similarity to the neuroblastoma break point family genes NBPF3, NBPF1, and NBPF15. The team's subsequent experiments supported the notion that NBPF23 expression varies with copy number and indicated that the gene is usually expressed during early neurodevelopment.
The 1q21.1 deletions tended to be detected more often in patients with aggressive neuroblastoma, though the association did not reach statistical significance. On the other hand, the authors noted, TCR loci deletions "showed a marked over-representation in the less aggressive subset of neuroblastoma."
Even so, Maris explained, the current research is more useful for offering insights into neuroblastoma biology than for potential prognostic applications, since researchers already have powerful prognostic markers for tumor samples.
The team is currently evaluating other CNVs implicated in the neuroblastoma GWAS and is continuing to discover and validate neuroblastoma SNPs. They have now genotyped more than 3,000 cases as part of their ongoing neuroblastoma research, Maris noted.