Comparative Biology of Glioma Stem Cells and Embryonic Neural Stem Cells in Dogs

In spite of accounting for less than 1.5% of human cancers in the United States, glioma tumors are a leading cause of cancer related mortality, with a dismal median survival time of only 12 months in patients suffering from glioblastoma multiforme (GBM). These tumors are driven by glioma stem cells (GSCs), a subpopulation of cells with similar molecular and functional characteristics to physiologic neural stem cells. Widespread and intensive investigation regarding the biology of these GSCs has nonetheless yielded disappointingly little progress in our understanding of the formation or progression of glioma tumors to date. Confounding attempts at dissecting the complex genomic aberrations in human GBMs are the often-large regions of chromosomal amplification or deletion, frequently involving entire chromosomal arms, in which we are only able to ascribe significant tumor suppressor or oncogene function to a single gene within this large region. In the face of such a complex, heterogeneous genome, efforts at modeling glioblastoma biology in genetically engineered mouse models by deleting or amplifying single genes or small clusters of genes often fails to recapitulate the behavioral, phenotypic, or genomic heterogeneity of spontaneous human glioblastoma tumors. Currently, there is no validated, naturally occurring model for human gliomagenesis. The domestic dog recapitulates every human histologic grade and develops glioma at an equivalent incidence to humans and as such, represents the only feasible model for comparative study of spontaneously occurring glioma tumors.;Here, we have characterized the dramatic similarities between glioma stem cells isolated from a canine anaplastic astrocytoma and our human GSC lines, and follow the canine GSCs as they form serial, orthotopic xenografts in immunocompromised mice. Serial xenotransplantation of our canine GSCs results in a progressive increase in tumor malignancy, expansion of the GSC subpopulation, and progressive genomic alterations strikingly similar to those associated with human secondary GBM formation. Chiefly, canine GSCs exhibit deletions of CDKN2A/p16Ink4a, deletion of PTEN , and loss of p53 function through amplification of MDM2 and MDM4 and loss of ARF. Importantly, these three pathways are identified as the major alterations driving human gliomagenesis, suggesting significant similarities between the two species at a molecular or genomic level.;In addition to the identification of these highly conserved copy number alterations (CNAs) present in our canine GSCs over serial xenotransplantation, analysis of the comparative genomic alterations between canine and human GSCs identifies numerous additional, putative tumor suppressor genes across both species. The canine genome is organized into 38 autosomes compared to 22 in the human, resulting in the dispersion of human chromosomal regions across multiple syntenic canine genomic segments occupying numerous individual chromosomes and allowing us to focus comparisons on small, highly conserved regions containing tumor suppressor or oncogenes we believe may be relevant to human glioma biology.;Additionally, we have isolated and characterized canine physiologic neural stem cells throughout gestation. Embryonic neural stem cells early in brain development exhibit increased clonogenic growth and proliferation, express higher markers associated with stem cell biology and perhaps most important, are refractory to differentiation cuesall of which are features shared by glioma stem cells. Global gene expression analysis of these embryonic neural stem cells identifies shared signaling networks to other mammalian neural stem cells. Future analysis of transcriptional networks governing canine neural stem cell self-renewal may reveal novel genes or gene interactions governing the proliferation of canine and human GSCs.