Building translational understanding of pediatric radiation-induced gliomas, an unmet need in childhood cancer

Pediatric radiation-induced gliomas (RIGs) are high-grade brain tumors caused by radiation kids received for a previous cancer. These incurable tumors have had little research attention. We conducted a study of drivers and potential treatments of RIG in the largest patient sample set ever assembled.

Like Comment
Read the paper
Cranial radiation therapy is used to treat children with several different types of pediatric cancer, including brain tumors and some forms of high-risk leukemia. Unfortunately, this type of radiation treatment can result in the formation years later of a secondary brain tumor called a radiation-induced glioma (RIG). RIG is an aggressive brain tumor that occurs in children or young adults 5-12 years after they have received brain radiation for another cancer. RIG is a tragic consequence of the previous successful treatment for childhood cancers that would have otherwise been fatal.
RIGs are incurable and respond poorly to treatment. Children seem to have a greater risk of developing a RIG than adults after receiving cranial radiation. RIGs occur in approximately 3% of children who receive cranial radiation and account for about 4% of childhood brain tumor deaths. RIGs may be responsible for even more fatalities because they can be misdiagnosed as a late recurrence of the previous cancer rather than a new cancer altogether.
Researchers from our group at Children’s Hospital Colorado and the University of Colorado Anschutz Medical Campus previously published findings showing that gene expression of RIGs differs from that of primary pediatric high-grade gliomas (1). Through a collaboration with researchers at St. Jude Children’s Research Hospital and the Childhood Cancer Survivor Study, we recently assembled the largest group of RIG samples to date with the goal of performing comprehensive molecular profiling in order to advance our understanding of RIG tumor biology. We also performed drug screening of patient-derived RIG cell lines to determine possible treatment targets (see figure below). 
We performed DNA methylation profiling, whole genome/whole exome DNA sequencing (WGS/WES) and bulk RNA-Seq on our RIG samples. Based upon DNA methylation profiling, the RIG samples clustered primarily with other high-grade glioma tumors of pediatric receptor tyrosine kinase I (RTK I) subtype (2). From the WGS/WES data, we identified common copy-number alterations, including Chromosome (Ch.) 1p loss/1q gain, and Ch. 13q and Ch. 14q loss. We also found focal alterations, including PDGFRA and CDK4 gain and CDKN2A and BCOR loss. In several samples, PDGFRA or CDK4 gain resulted from extrachromosomal circular DNA-mediated amplification associated with chromothripsis. PDGFRA gain is a common characteristic of RTK I subtype tumors. Based upon our analysis of the RNA-Seq data, RIGs comprised two transcriptomic subgroups. One subgroup had proneural, or stem-like characteristics, with relatively greater DNA repair capability (versus the other subgroup), lesser mutation burden and Ch. 1p loss. The second subgroup had mesenchymal or pro-inflammatory characteristics, with depleted DNA repair gene expression, and a greater mutation burden. Drug screens performed on two RIG cell lines, one from each of the transcriptomic subgroups, identified microtubule inhibitors/stabilizers, DNA-damaging agents, MEK inhibitors, and, in the inflammatory subgroup, proteasome inhibitors as potentially effective therapies.
Radiation-induced gliomas are a significant source of childhood cancer mortality. They have received little research attention and have had no clinical trials of experimental treatments, making RIG an unmet need in the brain tumor field. Our findings, made possible by a multi-institutional collaboration, identify molecular characteristics of RIGs not previously apparent in smaller studies. We also provide the first preclinical drug screening data from RIG cell lines. Our study, now published in Nature Communications, represents an early step in developing urgently needed RIG-specific therapies. It is also co-published with a similar study from our colleagues at the German Cancer Research Center that showed complementary findings; together, this work is a powerful step forward in understanding RIG. Further preclinical work, including development and testing of animal models, is required to progress to clinical trials of promising therapeutics for this challenging and deadly disease.
1. Donson, A.M. et al. Unique molecular characteristics of radiation-induced glioblastoma. J Neuropathol Exp Neurol 66, 740-9 (2007).
2. Capper, D. et al. DNA methylation-based classification of central nervous system tumours. Nature 555, 469-474 (2018).

Adam Green

Associate Professsor of Pediatrics, University of Colorado School of Medicine