Tackling 1q+ PFA ependymomas

See the article by Pierce and Witt et al. in this issue, pp. 1540–1551.

Ependymomas are central nervous system tumors of glial origin. Ependymomas account for approximately 10% of all pediatric CNS tumors, and the frequency is much higher (~30%) in children under the age of 3 years.^1–3 Of the 9 DNA methylation molecular subgroups of ependymoma, posterior fossa type A tumors (PFAs) are lethal and lack any persistent genomic alterations, with a few exceptions such as alterations in chromosome X open reading frame 67 (CXORF67; enhancer of zeste homolog inhibitory protein) and gains in chromosome 1q.^4-6 As such, PFAs are largely considered an epigenetic driven disease with alterations resulting in DNA cytosine-phosphate-guanine island promoter hypermethylation, gene body hypomethylation, and loss of histone trimethylation (H3K27me3) and active oncogenic super-enhancers (H3K27Ac).^7–9 In particular, gains of chromosome 1q have been associated with poor prognosis of PFA patients. Lack of 1q-positive PFA models has hampered testing of therapeutics and discernment of the mechanisms of this disease. In this issue, Foreman and colleagues have generated 2 PFA 1q+ patient-derived xenograft (PDX) models that faithfully recapitulate histologic features of ependymoma, such as small round blue cells with a striped chromatin pattern forming ependymal rosettes.¹⁰ Moreover, these models have high expression of CXORF67, lose H3K27me3 expression, and retain 1q gains through serial in vivo passaging. These models also subgroup as ependymomas using a DNA methylation classifier, unlike high-passage cell lines that can drift at the molecular level. Moreover, the authors use these PDX models to demonstrate preclinical testing of radiation and 5-fluorouracil, with radiation extending survival. These PDX models will also allow for exciting possibilities such as establishing relapsed or treatment resistance in vivo. One limitation of this model is the inability to fully study the immune component and its contribution to tumor biology, something that is common for many brain tumor models. The authors will make their models available to the brain tumor community to accelerate research in this field, which will provide an invaluable resource to reliably test new investigational therapies.