Ependymomas are tumors of the brain and spinal cord and are fatal in up to 40% of patients. Although ependymomas from different regions of the central nervous system are histologically similar, they are distinct on a molecular and genomic level, thereby justifying different treatments. Nevertheless, since chemotherapy is largely ineffective in treatment of ependymomas, they typically are treated with surgery and radiotherapy. Therefore, there is an urgent need to better understand and model the pathophysiology of ependymomas to design more effective and unique therapies. While it is tempting to speculate that human supratentorial ependymoma (ST-EPN), a location-specific ependymoma, may be a consequence of just one damaging event in the DNA, such as the well-known and clinically actionable BCR-ABL fusion gene driven chronic myelogenous leukemia, genetically accurate animal models and a deeper understanding of the oncogenic function of the fusion gene would be required. Due to a relatively stable genome and presence of a fusion gene (C11orf95-RELA) that can induce tumors in mice, the study of ST-EPN can be effectively conducted in mice.
Detected in 72% of ST-EPN patients, C11orf-RELA fusions (RELAFUS) have been previously shown to transform mouse embryonic neural stem cells ex vivo and activate NF-kB signaling. Dr. Eric Holland and members of his laboratory in the Human Biology Division, along with their collaborators in Seattle, Memphis, Japan and the United Kingdom, created a de novo mouse model of C11orf95-RELA fusion, and showed that this fusion gene induced ependymomagenesis that may depend on other cell signaling pathways in addition to nuclear factor kB (NF-kB). The results of this study were published in a recent issue of Cell Reports. To determine if RELAFUS is required for transformation, the authors created a de novo mouse model by using a system previously developed for glia-specific gene transfer in transgenic mice, known as RCAS/tv-a system. The RELAFUS, along with variants containing mutations in critical regions of RELA, were delivered into specific brain cell types including Nestin-, glial fibrillary acidic protein (GFAP)-, and BLBP-expressing cells. Mice with the RELAFUS developed tumors within two months, in contrast to control mice receiving only RELA or C11orf95 genes alone. Notably, these RELAFUS-induced tumors not only recapitulated the growth pattern and morphological and immunohistochemical features of human ST-EPN, they also mirrored the transcriptome of human ependymomas. Further analysis of mRNA expression data showed that following de novo tumor formation in vivo, and in vitro within neural stem cells, RELAFUS activated NF-kB target gene expression.
Since NF-kB is similarly activated by the overexpression of wild-type RELA within neural stem cells and does not promote malignant transformation, the authors reasoned that there would be non-NF-kB target genes that would be dysregulated in their model. Indeed, by Gene Set Enrichment Analysis (GSEA) of transcriptomic data, dysregulation of cell-cell adhesion and ion transport genes were found to be selectively associated with RELAFUS-driven transformation. To further investigate the transforming functions of RELAFUS in both NF-kB and non-NF-kB related ways, the authors created in vivo models of mutant forms of wild-type RELA and RELAFUS using the RCAS/tv-a system. They found that constitutively active RELA does not result in ependymoma formation. They also found that Ser-486 and acetylation of key residues played a critical role in RELAFUS driven tumorigenesis. Using the de novo RELAFUS-induced tumor model, the authors also investigated the expression of the fusion gene in neonatal mice and found the ventricular wall to be the putative location of the cells of origin of ependymoma.
In summary, the authors created a genetically accurate mouse model of human ST-EPN and found that RELAFUS drives several other biological processes in addition to NF-kB. They showed the oncogenic-driving potential of a subset of residues critical for RELA activity. Nevertheless, the molecular mechanism underlying RELAFUS driven ependymomagenesis in humans may be more complicated than the mouse model, given that these RELA fusions were generated by a gene rearrangement event that may affect other genes on the chromosome. However, it is evident from this study that ependymomas can be clearly distinguished from other glial neoplasms, setting the stage for the development of more effective diagnostic tests and eventual therapies.
Ozawa T, Arora S, Szulzewsky F, Juric-Sekhar G, Miyajima Y, Bolouri H, Yasui Y, Barber J, Kupp R, Dalton J, Jones TS, Nakada M, Kumabe T, Ellison DW, Gilbertson RJ, Holland EC. 2018. A de novo mouse model of C11orf95-RELA fusion-driven ependymoma identifies driver functions in addition to NF-kB. Cell Reports 23:3787—3797.
Funding was provided by the National Institutes of Health, Cancer Research UK, the Mathile Foundation and CureSearch.