GE-40 * COMPARATIVE GENOMIC ANALYSES IDENTIFIES NON-SILENT MUTATIONS PRESENT IN BOTH MURINE AND HUMAN MEDULLOBLASTOMA

Abstract

Mouse models of human cancers are now widely used in cancer biology and therapeutics. However, these models have, more recently, been scrutinized regarding the degree with which they faithfully recapitulate their human counterparts. To determine whether mouse and human tumors are driven by similar genetic alterations we chose medulloblastoma, the most common malignant pediatric brain tumor, and performed whole-exome sequencing on 12 animals representing models of three distinct medulloblastoma subgroups: WNT, SHH and G3 (C-MYC). Without sequencing data from matching normal tissue, we developed a robust germline-filtering pipeline that removes known single nucleotide polymorphisms (SNPs) in commonly used mouse strains. We rigorously compared putative mutation calls to raw control mouse binary alignment map (BAM) files to filter out remaining polymorphisms. Sequencing tumor and non-tumor DNA from ancestral mouse strains validated the presence of a mutation with somatic success rates of 61% and 14% for high quality and low quality calls, respectively. Using this approach, we identified 75 somatic mutations (73 single nucleotide variation [SNVs] and 2 small insertions & deletions [INDELS]), including 40 with predicted amino acid changes. Of these 40 mouse MB amino-acid change mutations, the human ortholog of eight genes also harbored amino acid change mutations. Lrfn2, Smyd1 and Wdr11 were mutated in a mouse G3 medulloblastoma and tested in vivo using intracranial allografts. Each gene was overexpressed in mouse G3 tumor spheres by lentiviral gene transfer followed by cell sorting and intracranial stereotactic implants. Wdr11 significantly repressed tumor progression, and increased the survival of tumor-bearing mice by ten and eleven days compared to tumor cells uninfected or infected with empty vector. Our pipeline successfully filtered out most germ-line SNPs even without matching normal tissue data, and identified Wdr11 as a novel target in medulloblastoma.