Functional characterization of TUSC2 for its role as a novel tumor suppressor for malignant glioma

Abstract

Glioblastoma (GBM) is the most frequently diagnosed brain tumor in adults. Even with the standard-of-care therapy, most GBM patients have an abysmal survival of less than 15 months. Multiple tumor suppressors have been identified as critical factors in controlling gliomagenesis, and GBM growth and progression. Recently, Tumor Suppressor Candidate 2 (TUSC2) has been identified as a critical tumor suppressor lost in GBM. However, the mechanisms that drive TUSC2-mediated tumor suppression in GBM are not well understood. Therefore, the goal of my thesis was to elucidate mechanisms involved in TUSC2-mediated tumor suppression and its role in gliomagenesis.Through a biased approach to identify TUSC2-interacting proteins, we identified ALDH1A3 as a novel TUSC2-interacting protein. We found TUSC2 knockout (TUSC2-KO) to significantly increase overall ALDH activity in GBM, and TUSC2 overexpression had the opposite effect. Furthermore, addition of ALDH1A3 significantly increased ALDH activity in both cell-free and in vitro assays, whereas TUSC2 inhibited ALDH1A3 and nullified these effects. Consistent with the role of ALDH1A3 in RA biosynthesis, TUSC2-KO significantly increased RA synthesis, enhanced RARα nuclear localization, and increased both mRNA and protein expression of downstream RA target genes. Conversely, TUSC2 overexpression had the opposite effect. Furthermore, we observed that TUSC2 loss is correlated with increased ALDH1A3 and RAR activity in GBM patients, and that TUSC2 loss and increased ALDH1A3 or RAR activity was predictive of worse overall patient survival. To investigate the role of TUSC2 in gliomagenesis, we generated a conditional TUSC2-KO transgenic mouse model to induce TUSC2 loss specifically in mouse astrocytes. We successfully generated and maintained a TUSC2-KO mouse colony and confirmed successful loss of TUSC2 in the mouse brain through IVIS imaging and immunohistochemical staining. However, even with TUSC2 loss in mouse astrocytes, we did not see any signs of brain malformation or gliomagenesis. Lastly, since TUSC2 protein has never been investigated for phosphorylation, we investigated TUSC2 phosphorylation and its impact on TUSC2 function in GBM. Through an unbiased approach, we discovered that TUSC2 is phosphorylated at serine 50 (S50). We subsequently identified PKAcβ and PKCα as the TUSC2/S50-phosphorylating kinases. Interestingly, activation of PKAcβ and PKCα significantly increased p-TUSC2/S50 levels and decreased total TUSC2 expression in GBM. Furthermore, phosphorylation of residue S50 significantly reduced TUSC2 protein stability as compared to a phospho-dead TUSC2 mutant, TUSC2/S50A. We discovered that TUSC2/S50 phosphorylation was necessary for NEDD4 binding and subsequent TUSC2 poly-ubiquitination and proteasomal degradation in GBM. Together, these studies conducted in my thesis project provide novel insights into TUSC2 tumor suppressive pathways, the role of TUSC2 loss in gliomagenesis, and TUSC2 phosphorylation and protein stability.