A System Approach to Astrocytoma Progression

Astrocytomas are brain cancers that account for 75% of neuroepithelial tumors. Accurate diagnosis is required for adequate treatment and assessing prognosis for patients. We used both gene-based and network-based approaches to find the key processes involved in astrocytoma progression from pilocytic astrocytoma (WHO grade I) to glioblastoma (WHO grade 4). Out of 20985 genes in Affymetrix 133A array, we identified a core of 494 differentially expressed genes (DEGs) that appeared to be central in astrocytoma. DEGs exhibiting coherent expression profiles were mapped to subnetworks using Expander and enriched GO categories within these subnetworks were identified. On the other hand, our network-based approach identified 193 most perturbed expressed pathways that are essential to astrotytoma pathophysiology. These pathways were characterized into different biological processes involving regulation of immune system process, regulation of signal transduction, regulation of MAPK activity, regulation of cell cycle and cell death, among others. A majority of these processes showed deregulation, while immune system related pathways became more tightened as the disease became more aggressive. Patients with glioblastoma, the most aggressive form of astrocytoma, showed heterogenity in their gene expression profiles and we identified a pathway, MTA3 which could effectively classify GBM patients into two subtypes with significant differences (p<0.05) in survival time. Our systems approach provides a comprehensive basis for developing models for astrocytomas and suggests possible therapeutic approaches.