Discovery of Synthetic Lethal Interactions with CDH1 in Medulloblastoma

Abstract

Medulloblastoma is the most frequently occurring malignant brain tumour in children, affecting just under two individuals per million annually. Medulloblastomas have an aggressive, invasive growth pattern, with a high risk of metastasis to structures of both the brain and spine. The five-year survival rate for medulloblastoma is currently 70-75%, however, due to the high rate of recurrence observed in this cancer, the 20-year survival rate drops to approximately 50%. Therapeutic options for medulloblastoma involve neurosurgery, cranial radiation therapy and chemotherapy. This highly aggressive therapy leaves surviving patients with long-term debilitating side effects, such as deficits in neurocognitive and endocrine function.

CDH1 is a tumour suppressor gene involved in cell-cell adhesion, differentiation and proliferation which is commonly lost across various cancer types, including medulloblastoma. Tumour suppressor gene inactivation results in a lack of expressed protein, and therefore cannot be exploited by conventional drugs as there is no protein to target. The concept of synthetic lethality has important implications in this context, as it allows for the exploitation of inactivated tumour suppressor genes in cancerous cells. Any two genes are said to be in a synthetic lethal relationship if mutation of either gene maintains cell viability, but simultaneous mutation of both genes causes cell death. This allows for the development of novel treatment strategies which target the synthetic lethal partner of a silenced tumour suppressor gene, selectively killing tumour cells and leaving healthy tissue relatively unharmed. This is particularly significant in medulloblastoma where the toxicity of current treatment regimens is high, and provides a large margin for improvement. The aim of this study was to make use of publicly available microarray expression data to first identify a CDH1-low subgroup of medulloblastomas. This group of tumours was then used to identify potential synthetic lethal interactions with CDH1, using a novel bioinformatic strategy. Following this, experimental validation of the putative synthetic lethal candidates, RARB and PDGFD, involved short hairpin RNA (shRNA) knockdown of identified candidate genes within a cell line pair representing a model of E-cadherin loss. This was accompanied by quantitative real-time PCR (qPCR) to confirm gene knockdown. Both of the investigated putative synthetic lethal candidate genes were validated as synthetic lethal partners of CDH1, as knockdown induced selective killing of E-cadherin deficient cells. To further characterise these synthetic lethal partnerships, drug inhibition of candidate genes within the same model of E-cadherin loss was performed. For RARB, drug inhibition was successful in inducing a synthetic lethal phenotype, and to a greater degree than observed through shRNA knockdown. For PDGFD, drug inhibition was unsuccessful in producing a synthetic lethal phenotype. However, the drug used was not specific to the synthetic lethal candidate, and inhibition of other biological pathways was likely deleterious to E-cadherin expressing cells as well as E-cadherin deficient cells. In summary, this study has demonstrated the effectiveness of a novel bioinformatic strategy for the identification of synthetic lethal partnerships. Two synthetic lethal partners of CDH1 in a medulloblastoma context have been identified and validated by this study, adding further support to the development of novel therapeutic techniques utilising the concept of synthetic lethality.