Introduction: The SOX4 gene encodes a high-mobility group (HMG)-box transcription factor involved in embryonic differentiation and development and, later, is involved in signal transduction through the TGF-beta, IL5-R, and WNT pathways. Abnormal expression of SOX4 is associated with numerous malignancies, including those of the central nervous system. Its role as a potential etiologic agent and therapeutic target in glioblastoma is of particular interest given the limited treatment options and universally poor prognosis of these tumors. Notwithstanding, the fundamental patterns of transcriptome and proteome-level expression of SOX4 remain inadequately characterized across the glioma spectrum. We believe that this may be critically important to understanding the development, maintenance, and progression of gliomas; and, more importantly, represents a necessary step toward exploring the potential of SOX4 as a target in translational glioma therapy.

Methods: We used microarray analysis to analyze transcriptome-level expression of the SOX4 gene using tumor tissue from 43 patients with World Health Organization (WHO) grade I-IV gliomas and validated these results using RT-PCR. We also conducted qualitative characterization and semi-quantitative analysis of the protein-level expression of SOX4 in a subset of this population and in primary and transformed glioma cell lines using immunofluorescence and mass spectrometry. Subgroup analysis was performed in a supervised and unsupervised fashion using an integrated data analysis model incorporating the results of the transcriptomic and proteomic investigations.

Results: Transcriptome analysis demonstrates SOX4 mRNA overexpression in all glioma specimens. The magnitude of overexpression varies between patients, but there is insufficient evidence to suggest an independent correlation between the magnitude of differential expression and either WHO tumor grade or survival phenotype in low- or high-grade gliomas. Protein-level investigation indicates that, while transformed cell lines demonstrate almost exclusive localization of SOX4 to the nucleus, primary cell lines and surgical tissue specimens of from tumors of all WHO grades exhibit both nuclear and cytoplasmic localization of the SOX4 protein. The relative distribution of SOX4 protein between these compartments is variable and is not related to WHO grade.

Conclusions: The SOX4 gene is overexpressed in all grades of gliomas, and the magnitude of overexpression differs between patients. The gene product localizes variably to the nucleus and cytoplasm in clinical specimens, the functional significance of which remains to be determined. We find insufficient evidence from our integrated genomic and proteomic analysis to suggest that the magnitude of SOX4 gene expression is an independent correlate or a direct predictor of survival in patients with gliomas. Rather than defining genotypic tumor subgroups, our findings suggest that SOX4 overexpression may be a common feature of gliomas, independent of tumor grade or clinical phenotype. This raises the possibility that abnormal expression of SOX4 may play a fundamental role in the development of all gliomas, a finding which may be logical given its role in early development and differentiation of embryologic glia. Ongoing investigation is required to more thoroughly characterize the contribution of SOX4 overexpression to the pathophysiology of gliomas, which is essential to understanding the translational therapeutic potential of SOX4.

Citation Information: Clin Cancer Res 2010;16(7 Suppl):A23