This story was published on July 2.
A whole-genome sequencing study by St. Jude Children's Research Hospital and Washington University has confirmed the genetic underpinnings of four subtypes of childhood medulloblastoma and uncovered potential drug targets in the disease.
The study, part of the St. Jude/Washington University Pediatric Cancer Genome Project, compared tumor and matched normal DNA in a total of 93 subjects and identified a total of 41 genes associated for the first time with medulloblastoma. The results appeared June 20 in Nature.
The group found a number of recurrent somatic mutations specific to the four medulloblastoma subgroups and discovered that several mutations are in components of the epigenetic machinery, offering potential subgroup-specific drug targets.
Giles Robinson, the study's lead author, told Clinical Sequencing News that the main achievement of the study was identifying mutations that correlate with disease subgroups that have been defined through other measures, like gene expression profiling.
Robinson said the disease subgroups — sonic hedgehog, or SHH; WNT, or wingless; subgroup three; and subgroup four — differ significantly in their prognoses. For example, almost all patients in the WNT subtype survive while only about 60 percent of those with subtype three medulloblastoma live three years after diagnosis, the group reported.
"Our mutation results are clinically important because these groups are shown to do better or worse depending on what subgroup you fall into, but currently all medulloblastoma is treated the same, [with] … therapy schemes involving cranio-spinal radiation and quite high-dose chemo."
Now, he said, "we can start to think about targeting therapy toward those individual subgroups."
According to Robinson, the potential epigenetic drug targets highlighted by the sequencing results were surprising and unanticipated by the researchers. "We found mutations highlighted on some epigenetic pathways we'd never considered to be part of this disease," he said.
In the study, the group sequenced tumor and normal DNA from 37 samples as a discovery set, identifying mutations recurrent in at least two samples, which the group then looked for in a second set of 56 medulloblastomas. Overall, the researchers found recurrent somatic mutations in 41 genes that had not previously been implicated in medulloblastoma. Delving further, they discovered that several of these genes were epigenetic regulators, connected to pathways that are potentially targetable with new inhibitor therapies or those in development.
"We didn't really know what these genes were when we started, but we saw a collection of genes that were epigenetic regulators," Robinson said. "And when we looked in further detail, they started pointing toward specific epigenetic marks, like H3K27 and H3K4."
These mutations in subgroups three and four — which include the most aggressive forms of the disease — affected several genes that the group believes are involved in epigenetic chromatin marking involving H3K27 and H3K4.
The group also identified another set of mutated genes enriched in the WNT subgroup of medulloblastomas that are linked to different influences on the epigenetic machinery of the disease, like CTNNB1-associated chromatin remodelers. This suggests that epigenetic disruptions appear to be subgroup-specific.
"In the big picture," he said, "some of these, specifically H3K27, control the transition from an early progenitor stem-cell-like cell toward a terminally differentiated cell. So when the cancers in those subgroups were disrupting that mark, the cells were unable to transition into a terminally differentiated cell."
The researchers highlighted one epigenetic area, H3K27 methylation — associated with mutations in the aggressive subgroups three and four — as being a promising target for inhibitor therapies in development.
Robinson said the team is now trying to work with companies and other research groups to investigate further the potential of targeted inhibitors associated with maintenance of H3K27 trimethylation. "It's given us a whole new insight into what to look for therapeutically," he said.
Additionally, the researchers reported that mutations that activate PIK3CA and DDX3X in the WNT subgroup may also be potential therapeutic targets.
Robinson said the group's whole-genome sequencing approach likely gave a deeper picture of the disease than would have been possible with a targeted or exome-based strategy. "We had hoped we would get information, but digging through all the data [has] unveiled stuff we never thought we would find, and unveiled pathways we never thought about before," he said.
Because none of the mutations the team found are directly actionable with current therapies, the researchers do not plan to release sequencing results to study participants. "We just don't have therapy to give them right now based on these genes," he said.
The medulloblastoma results are the latest findings out of the Pediatric Cancer Genome Project, which has also identified potential treatment targets in retinoblastoma (CSN 1/18/2012).
Last month, the project released data representing 520 genome sequences from matched normal and tumor tissue samples of 260 pediatric cancer patients — the largest-ever public access release of comprehensive human cancer genome data by the scientific community, according to the project.
James Downing, PCGP project site director and scientific director for St. Jude Children’s Research Hospital, told CSN in an e-mail that the project has now completed sequencing and primary analysis on more than 350 pediatric cancers, or 700 cancer and normal genomes, exceeding its first-year goal. Downing said the project expects to complete its first phase before February 2013.