Medulloblastomas represent the most frequent malignant brain tumour in children, and are thought to develop in the posterior fossa of the cerebellum. Some patients with medulloblastomas have a deficiency in the tumour suppressor gene Apc (Turcot’s syndrome). Although the majority of medulloblastomas arise sporadically, people with Apc mutations are 92 times more likely to develop medulloblastomas. Apc encodes a very large protein known to function as a regulator of the Wnt signalling pathway. Activation of the canonical Wnt pathway leads to the stabilisation of beta-catenin. In response to Wnt signals, beta-catenin translocates to the nucleus where it interacts with the LEF/TCF family of transcription factors to activate transcription of target genes such as c-myc and cyclinD1. Mutations of Apc that cause an increase in beta-catenin are found to be tumourgenic, whereas other mutations are not. Therefore it is thought that the main tumour suppression function of Apc is in its ability to destabilise and hence reduce cytoplasmic beta-catenin. The central hypothesis of this thesis is that the loss of Apc can lead to the development of medulloblastoma. Work from other groups has reported activation of Wnt signalling in a proportion of primary medulloblastomas. We undertook a study to assess this by using the cre-loxP recombination system to mutate Apc in a temporal and spatial manner. This approach is necessary as Apc has many functions in development and Apc mutant mice (Apcmin) do not develop past embryonic day 6.5 (E6.5). To date, there are no known cre-strains available to mutate Apc specifically in the cerebellum at early postnatal stages, so we combined the creloxP method with an avian retrovirus mediated method for tissue specific gene delivery (RCAS/tv-a system), in an attempt to create a strain of mice which carried the genotype Ntv-a +;ApcLoxP/LoxP. This would allow us to infect an RCAS-cre virus directly into the hindbrain at postnatal day 4 (P4). However subsequent genotyping of these animals showed that none carried the desired genotype of Ntv-a +;ApcLoxP/LoxP, making it impossible for both copies of Apc to be mutated in a mouse most likely because both the Ntv-a and Apc transgenes were located on the same chromosome. Consistent with this, out of a total of 265 mice none were found to have the Ntv-a +;ApcLoxP/LoxP genotype. We then adopted an alternative method for mutating Apc by infecting ApcLoxP/LoxP mice directly with an AdCre virus. PCR analysis showed that Apc was mutated, however the AdCre virus did not infect cells of the cerebellum, and instead only infected the choroid plexus. In these animals, 7 of 94 (7%) developed hydrocephalus indicating that losing Apc in the choroid plexus may promote hydrocephalus. Finally, to address the role of Apc in normal hindbrain development, we crossed our ApcLoxP/LoxP mice to an En1cre strain which caused mutation in Apc from E8.5 in the midhindbrain region. The resulting En1cre+;ApcLoxP/LoxP mutants displayed hydrocephalus in all ventricles and an in-growth of mesenchyme tissue at the mid-hindbrain border, closely associated with a tumour-like area of cells showing activated Wnt signalling. No mice were found to live past E18.5. In conclusion, the role of Apc in medulloblastoma remains unclear. Future studies could use a different technique to mutate Apc such as crossing ApcLoxP/LoxP mice to the new nestin-creER strain and inducting cre with administration of tamoxifen.