A multi-institutional group of researchers, led by investigators at Children's Hospital Los Angeles and the University of Michigan, have identified a simple and inexpensive tool for assessing the prognosis of pediatric brain tumors called ependymomas. Their study, which demonstrates the epigenetic mechanism behind these tumors, may offer future opportunities for novel therapeutic options. It will be published online by Science Translational Medicine on November 23.
Childhood posterior fossa ependymomas (PF) are tumors found largely in the hind brain (consisting of the cerebellum, pons and the brainstem) of children. Routine assessment of tumor grade and other markers in PF ependymomas do not correlate well with outcomes in these tumors, highlighting the need for new prognostic markers. Genomic sequencing efforts have not identified mutations in these tumors, and the origin of PF ependymomas remains obscure.
While lacking recurrent genetic mutations, a subset of these tumors exhibit alterations in DNA methylation. In this study, the researchers looked at modification of histones - protein components of the chromatin around which DNA winds, and which play a role in gene regulation - in particular, histone H3.
Co-lead investigator, Sriram Venneti, MD, PhD, of the Department of Pathology at the University of Michigan, observed that histone H3 is modified differently in pediatric posterior fossa ependymoma. Specifically, 80 percent of these tumors exhibited loss of the H3K27me3 a repressive mark, while 20 percent of tumors retained H3K27me3. Researchers went back and looked at MRIs and outcomes of children treated for these tumors and identified that tumors with loss of H3K27me3 tumors behaved more aggressively and showed poor overall survival. This suggests that reduced H3K27me3 may be a prognostic indicator in PF ependymomas.
"Detection of H3K27me3 by immunohistochemical staining is a widely available and cost effective surrogate molecular marker. This test can be readily implemented in most departments of pathology and provides a much-needed tool to risk stratify and identify ependymoma patients who would potentially benefit from epigenetic therapies," said co-lead investigator Alexander R. Judkins, MD, of the Department of Pathology and Laboratory Medicine at CHLA and Keck School of Medicine of the University of Southern California.
This loss in H3K27me3, along with other epigenetic changes, was similar to that observed in another type of pediatric brain tumor of the hind brain region termed diffuse intrinsic pontine gliomas (DIPGs). This suggests that both of these tumors arise from similar epigenetic states. Intriguingly, researchers found that certain progenitor cells in this part of the brain also showed low H3K27me3, suggesting - as both tumors share epigenetic similarities - that low methylation of H3K27me3 is important to the development of tumors in this region of the brain.
According to the Collaborative Ependymoma Research Network (CERN) Foundation, the current standard therapy for ependymomas includes total surgery that may be followed by radiation therapy. Complete surgical resection is often not possible due to tumor location and concerns about damaging the surrounding brain during surgery.
Because such treatment can come with a cognitive cost, especially to pediatric patients, there has been a lot of interest in what drives ependymomas so that they may be treated more effectively. "By demonstrating the epigenetic mechanism - that we theorize likely goes awry during brain development - we will be better able to identify these tumors, determine a more accurate prognosis and - importantly -perhaps develop better therapeutic options," said Venneti.
Additional contributors include Jill Bayliss, Chan Chung, Pooja Panwalkar, Abhijit Parolia, Arul M. Chinnaiyanm, Richard C McEachin and Marcin Cieslik, University of Michigan; Piali Mukherjee and Ari Melnick, Weill Medical College of Cornell University; Chao Lu, C. David Allis and Benjamin Sabari, The Rockefeller University, Siddhant U. Jain and Peter W. Lewis, University of Wisconsin-Madison; Daniel Martinez, Children's Hospital of Philadelphia; Ashley S. Margol and Benita Tamrazi, Children's Hospital Los Angeles; Melike Pekmezci, UC San Francisco; Benjamin A. Garcia, Perelman School of Medicine, University of Pennsylvania; Gaspare La Rocca and Craig B. Thompson, Memorial Sloan Kettering Cancer Center; Mariarita Santi, Children's Hospital of Philadelphia; and Cynthia Hawkins, The Hospital for Sick Children, Toronto.
This work was supported by grants from the National Cancer Institute of the NIH (K08 CA181475, the Mathew Larson Foundation, the Sidney Kimmel Foundation and the Doris Duke Foundation (SV); and RO1GM110174 and P01CA196539 (BAG). RCM is supported by the University of Michigan Bioinformatics Core. C.L. is the Kandarian Family Fellow supported by the Damon Runyon Cancer Research Foundation (DRG-2195-14). CBT is supported by the Cancer Center Support Grant (CCSG) of Memorial Sloan Kettering Cancer Center (CBT) - a NIH P30 CA008748.