Identification of cancer genes using Sleeping Beauty mutagenesis of the Ptch1 murine tumour model

Abstract

Medulloblastoma (MB) is a paediatric tumour of the cerebellum which is responsible for 15-20% of all childhood brain tumours. Mortality due to this disease is high (~40%) and successful treatment is associated with significant neurological and cognitive consequences, making new therapies desirable. Disruption of the Sonic Hedgehog (SHH) signaling pathway, including mutations in PTCH1, define a major subset of human MB. Mice heterozygous for loss of function mutations in the Ptch1 ortholog develop MBs at low frequency. To identity genes that co-operate with Ptch1 in MB development, a Sleeping Beauty insertional mutagenesis screen in this murine model was performed. Mutagenesis significantly increased the frequency of MB formation in Ptch heterozygote mice from ~3% to ~25% after 8 months (p<0.0001). To identify the genes responsible for this enhanced tumour formation, Splinkerette-PCR and 454-FLX sequencing were applied to define SB insertion sites within 40 tumours. Statistical analysis of these data using Monte Carlo and Kernel Convolution methods identified 18 candidate medulloblastoma genes defined by common insertion sites (CISs), including 11 identified by both methods used. Many of these are genes known to be involved in neuronal development and cell fate determination. Subsequent ARACNe network analysis of the candidate genes in human gene expression datasets has established that seven (Nfia, Nfib, Tead1, Tgif2, Myt1l, Fgf13 and Crebbp) lie within a single network which is enriched for both neuronal genes and transcription factors, and includes genes known to interact with the SHH pathway. Bioinformatic analysis confirmed that the tumours generated from SB mutagenised mice were similar in their gene expression pattern to human SHH subgroup tumours, while microarray expression analysis of SB induced tumours identified Igf2, a gene previously implicated in MB development, as a key output of network activity. Finally, analysis of gene expression within granule neuron precursor cells (GNPCs), the cell of origin of SHH subgroup MB, identified Tgif2 and Myt1l as prime candidate genes for downstream functional analysis. Lentiviral based modulation of their gene expression was therefore attempted in primary GNPCs. shRNA based knock-down of Myt1l was achieved, and shown to be associated with significantly increased expression of both Gli1 (the downstream effector of the Shh pathway) and Math1 (a marker of undifferentiated neuronal cells), consistent with a reduction in MYT1L expression playing a role in human MB development. Further analysis of this gene, and others defined here, should both improve our understanding of MB and define potential targets for therapeutic intervention.