Abstract A46: Transcriptome analysis on tumor and its normal cerebellar counterpart from mouse models reveals molecular mechanisms underlying evolution of SHH and Group3 medulloblastoma

Abstract

Abstract Human medulloblastoma (MB) consists of four molecular subgroups. In spite of enormous effort majority of current MB mouse models are SHH subgroup, represented by Patched heterozygous mice (designated as SHH-A). GTML MB is the only published spontaneous Group3 MB mouse model (designated as G3), which was induced by expressing MycN under the regulation of GLT1 promoter. By using Sleeping Beauty transposon system we have created and published two SHH mouse models: Math1SB11/T2Onc/Ptch+/- (designated as SHH-B) and Math1SB11/T2Onc/P53mut (designated as SHH-C). To dissect the molecular mechanism of SHH and Group3 MB during tumor evolution (initiation, maintenance and maturation), we profiled transcriptomes on SHH-A, SHH-B, SHH-C and G3 MBs along with normal mouse cerebellum (CB) of different developmental stage (E15, E18, P0, P60, P95 and P122). Fetal and/or adult human CBs are routinely used as normal controls for human pediatric MBs. One major advantage of profiling mouse tumor samples over human tumor samples is the availability of age-matched normal mouse CB counterparts. This is particularly true in case of developmental diseases such as pediatric MBs. Indeed, our study showed that transcriptomes of normal mouse CB from E15, E18 and P0 (designated as EdCB, representing early developmental CB), are similar to each other but distinct from mouse CB at P60, P95 and P122 (designated as LdCB, representing late developmental CB). Importantly P60 to P120 are the ages at which most of spontaneous MBs are harvested from a tumor-bearing mouse. We therefore analyzed global gene expression level of MBs and EdCB in comparison with LdCB, the mouse MB's normal counterparts. As a developmental disease, MB is presumably initiated from a single cell or a single cell clone that is transformed at an early developmental stage of a normal CB. In fact, it is well known that a SHH MB is initiated from transformation of granule cell precursors at EGL, the external granule layer. However, it is unclear about the initiation of Group3 MB. In this study we showed that compared to a normal LdCB, a particular MB transriptome harbors four major types of abnormally expressed genes: 1. Genes not only specific to one subgroup but also shared with EdCB; representing tumor initiation events. 2. Genes specific to a subgroup only; representing subgroup-specific tumor promotion events. 3. Genes shared not only by both subgroups but also by EdCB, representing non-subgroup specific tumor maintenance events; 4. Genes shared by both subgroups only; representing non-subgroup specific late tumor maturation events. We confirmed that SHH MBs from SHH-A, B, and C are all initiated at Math1/Atoh1 expressing cells, which are further transformed by up-regulation of Olig3 and MycN among other genes. Rxrg-expressing cells are identified as Group3 MB initiating cells, which are further transformed by over-expression of Crx (Otx3) along with other genes such as Otx2, and down-regulation of cerebellum-specific Zic1 and others such as Lgi1. Both subgroups of MBs may be maintained and mature as genes associated with cell cycle, mitosis and chromosome integrity, such as KIFs, CDK1, PLK1/4, AURKA/B, CENP family and MCMs are abnormally overexpressed. Our study reveals distinct molecular mechanisms underlying tumor evolution of SHH and Group3 MBs. We propose a MB evolution model, illustrating how a normal cell at specific cell of origin can be transformed to a fully developed subgroup specific MB. Furthermore, our results demonstrate that despite the fact that they are initiated from distinct cell of origin, different subgroup of MBs do share common abnormal molecular events for tumor maintenance and maturation, suggesting that subgroup specific target therapies alone may be not as effective as their combination with therapies targeting abnormalities common to different subgroup of MBs. Citation Format: Xiaochong Wu, Adrian Dubuc, William A. Weiss, Robert J. Wechsler-Reya, Michael D. Taylor. Transcriptome analysis on tumor and its normal cerebellar counterpart from mouse models reveals molecular mechanisms underlying evolution of SHH and Group3 medulloblastoma. [abstract]. In: Proceedings of the AACR Special Conference: Advances in Brain Cancer Research; May 27-30, 2015; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2015;75(23 Suppl):Abstract nr A46.