NEW YORK (GenomeWeb News) – A National Human Genome Research Institute-led team reported online today in Nature Genetics that it has sequenced and characterized more than a dozen metastatic melanoma exomes, along with the exomes for matched normal blood samples.
From these tumor-normal exome sequences, researchers tracked down known and previously unrecognized mutations affecting nearly 70 genes. Among them: mutations in an ionotropic glutamate receptor gene called GRIN2A that turned up in between a quarter and a third of melanoma samples screened subsequently and which point to a role for glutamate signaling processes in skin cancer.
The team also found a change in a specific amino acid coding sequence of the TRRAP gene that was shared in four percent of skin tumors tested, as well as a range of other mutations in genes not found in past melanoma studies.
"Surprisingly, we found all these new, interesting genes that have not been implicated previously," senior author Yardena Samuels, a cancer geneticist at NHGRI, told GenomeWeb Daily News.
While the glutamate pathway has been tied to melanoma and other types of cancer before, she added, this study marks the first time certain genes within this pathway have been linked to melanoma.
Previous research — including a genome sequencing study of malignant melanoma published in 2009 — has shown that skin cancers tend to have a particularly high DNA mutation rate. That, in part, reflects the environmental component of these cancers, since UV light can cause extensive DNA damage.
Some genes have been linked to the disease already, Samuels explained, notably the BRAF gene, which can be targeted therapeutically. But much remains unknown about the repertoire of mutations that can underlie melanoma.
"We believe that we need to understand the disease more and try to find new cures for the disease," she said, explaining that melanoma mortality rates remain high for late-stage disease, while the annual incidence of the disease is on the rise.
In an effort to narrow in on pertinent, and perhaps treatable, new skin cancer mutations, Samuels and her co-workers decided to focus on mutations within coding sequences in 14 individuals with metastatic melanoma — samples that "allow us to look at the compilation of all the mutations that a tumor obtained throughout its evolution," Samuels noted.
To do this, collaborators at the National Institutes of Health Intramural Sequencing Center captured 37 million bases of coding sequence — representing roughly 20,000 genes — with the Agilent SureSelect system for each tumor and matched normal sample. They then sequenced the tumor and normal exomes to a mean depth of 180 times or more using the Illumina GAII.
Even after they'd filtered variants in tumor exomes against those in matched normal exomes, the dbSNP database, and 1000 Genomes Project data, though, researchers were still left with nearly 5,200 apparent somatic mutations affecting 3,568 genes in the tumors.
By doing targeted Sanger sequencing, incorporating information on sequence coverage, gene size, and extent of mutation in affected genes, and applying a so-called MPG algorithm, Samuels explained, the team was able to narrow in on 68 genes containing somatic changes. Of these, 16 genes are suspected to contain driver mutations.
Among the genes identified in the study were known skin cancer genes, such as BRAF and ERBB4, along with genes not previously reported for the disease, including GRIN2A, which was mutated in 33 percent of melanoma samples tested in initial follow-up screening and in 25 percent of samples from a broader collection of melanomas.
In four percent of the exome sequences and screened samples, meanwhile, researchers found a recurrent substitution in the transformation/transcription domain-associated gene TRRAP that's predicted to produce a specific amino acid change in the resulting protein.
Given the nature of this mutation, combined with TRRAP's role in a complex with histone acetyltransferase activity, Samuels and her colleagues suspect that the newly detected TRRAP alteration has been under selection and is oncogenic in melanoma — a hypothesis supported by their subsequent RNA interference experiments in melanoma cell lines.
"The main challenge now is figuring out which of these alterations are the most important and putting that information back into patient care," Samuels said, explaining that functional studies are needed to determine the consequences of driver mutations and more.
"We need to do some more basic biology to understand what's going on," she said. "I expect it to be pretty complex."
For its part, her team is continuing to do whole-exome sequencing on additional melanoma samples and collaborating on whole-genome studies of the disease as well.
This study "is really the beginning of deciphering the melanoma landscape," Samuels said. "It's clear that we need more exomes and we need more genomes."
Down the road, the researchers also hope to compare the sorts of mutations found in primary tumors with those in metastatic samples from the same individual to get a better idea of which mutations occur earliest and to learn more about whether mutations are shared within different metastases.