Abstract
A recent study integrating proteomic and genomic analyses uncovered unexpected similarities in some types of pediatric brain tumors, such as certain craniopharyngiomas and low-grade gliomas, suggesting potential treatments and more refined prognoses.
When it comes to analyzing pediatric brain tumors, the more “-omics,” the better. According to a recent Study, adding proteomics to standard genomic methods can provide insights into the tumors' biology, refine prognoses, and identify potential treatment targets.
Pediatric brain tumors harbor relatively few mutations, so genomic analyses haven't uncovered many options for personalized therapies. To obtain a clearer picture of the molecular landscape of tumors and discover potential treatment targets, researchers have started integrating proteomics into genomic studies. However, they have not previously performed a large-scale proteogenomic study of pediatric brain tumors.
In the new work, researchers at Children's Hospital of Philadelphia (CHOP) in Pennsylvania and colleagues at more than a dozen institutions analyzed 218 tumor samples from 199 children. The samples represented seven histologic types of brain tumors—low-grade glioma, ependymoma, high-grade glioma, medulloblastoma, ganglioglioma, craniopharyngioma, and atypical teratoid rhabdoid tumor. As part of the comprehensive proteogenomics investigation, the researchers performed whole-genome and RNA sequencing, as well as proteomic and phosphoproteomic profiling.
The results divided the tumors into eight categories that didn't always match the established histologic classification. For example, although medulloblastomas and ependymomas retained their “identities,” the analysis split craniopharyngiomas into two groups. Some clustered with low-grade gliomas with wild-type BRAF, whereas others were more like low-grade gliomas carrying the BRAFV600E mutation, even though they lacked this alteration. The existence of these two subtypes was not apparent from the RNA data.
This finding points to a potential therapy. A clinical trial reported in 2019 discovered that the MEK inhibitor selumetinib (Koselugo; AstraZeneca) triggered responses in pediatric low-grade gliomas with the BRAFV600E mutation, so the new findings suggest that MEK inhibitors might effectively treat pediatric craniopharyngiomas that are similar to these gliomas.
The research also indicates that proteomics could help physicians provide more accurate prognoses. The team found that in children with high-grade glioma, increased abundance of IDH proteins correlated with longer overall survival (OS)—but only if the patients carried wild-type genes for the H3 histone. In patients with mutated H3 histone genes, OS was shorter when the abundance of IDH1 or IDH2 was elevated. Further analysis of high-grade gliomas suggested that inhibitors of MEK and CDK might be effective against these tumors.
Another question the researchers addressed is whether tumors that have recurred or progressed are similar enough to the initial tumors to receive the same treatment. Using 18 pairs of samples, the scientists determined that although the “later” tumors fell into the same histologic groups as their predecessors, they typically belonged to different proteomic groups. One recurrent low-grade glioma clustered with other low-grade gliomas with wild-type BRAF, for instance, but its forerunner was more like low-grade gliomas that carried the BRAFV600E mutation. Such discrepancies occurred often enough, the researchers concluded, that recurrent and progressing tumors should be reanalyzed and not automatically considered the same as the initial tumors when making treatment decisions.
“By combining both proteomics and genomics in the molecular analyses of pediatric cancer, we can not only newly define proteomic-specific features for individual pediatric brain tumor types, but additional therapeutic rationales can emerge,” says co-author Adam Resnick, PhD, of CHOP.
“They did a Herculean job,” says Keith Ligon, MD, PhD, of Dana-Farber Cancer Institute in Boston, MA, who wasn't connected to the research. “This is a huge, really valuable resource.” Proteomics is still more challenging to perform than genomics, he says, but compiling this dataset demonstrates the technology's value and could help speed its improvement. –Mitch Leslie
