Medulloblastoma (MB) is the most common malignant brain tumor in children. About 350 new cases occur each year in the US with a peak incidence around 5 years of age. Cerebellar granule neuronal precursors (CGNPs) are thought to be the cells of origin of MBs. Proliferation of CGNPs is controlled by growth factors such as the mitogen Sonic Hedgehog (Shh), Wnt, and Notch ligands, and mutations in these signaling pathways are found in a subset of MB patients. The cyclin D-dependent kinase inhibitory protein, p18Ink4c, is transiently expressed in mouse CGNPs as they exit the cell cycle. Co-inactivation of p53 and Ink4c in mice results in MB with up to 25% penetrance, whereas mice heterozygous for Ptc1 (encoding the receptor for Shh) and lacking either one or two Ink4c alleles develop MBs with 50% incidence. Importantly, CDKN2C/INK4C expression is extinguished in a significant subset of MB patients, implicating this cell cycle regulator in the genesis of MBs in humans. The protracted tumor onset in these mouse models (∼2-5 months) implies that loss of Ink4c together with p53 or Ptc1 is not sufficient to induce MB, and that additional mutations must occur for tumor formation. One goal has been to identify genes that collaborate with inactivation of Ink4c, p53, or Ptc1 in MB formation. By use of Affymatrix Gene Chip microarrays, we have identified a panel of ∼30 genes that are consistently overexpressed in tumor cells, as compared to their expression levels in postnatal day 5 (P5) cerebellum. We have now purified CGNPs from the developing postnatal cerebella of tumor-prone knock-out mice, infected these cells with retroviruses encoding candidate genes together with a green fluorescence protein (GFP), and reinjected the marked cells into the brains of recipient mice under a cranial window. This enables us to follow tumor growth by observing GFP-marked cells in living mice. Our preliminary results show that CGNPs from Ink4c,p53 doubly-deficient mice infected with genes that are overexpressed in MB, such as N-Myc, not only exhibit a proliferative advantage in culture but also induce MBs in the cerebra of recipient mice. These primary tumors are transplantable subcutaneously in immunocompromised mice. This approach will allow the characterization of genes important for MB formation, will define the genetic networks involved in the development of this disease, and will generate new preclinical mouse models for testing new modes of therapy.

[Proc Amer Assoc Cancer Res, Volume 47, 2006]