Tumor gene expression in vitro and in vivo

Division Of Basic Sciences - Nci

Investigators

Linked publications & trials

Abstract

Our growth regulation research has been concerned with oncogenes as positive regulators and tumor suppressor genes as negative regulators of normal and neoplastic growth. The main current project is focused on the molecular biology of the tumor suppressor gene DLC1, including the targets that regulate it and the targets that it regulates. DLC1 is inactivated frequently in a wide range of tumors, but many aspects of its mechanism of action remain incompletely understood. DLC1 negatively regulates Rho, via its Rho-GAP activity, and also encodes other activities. A major hypothesis is that DLC1 is frequently inactivated in cancer because it encodes a biologically important multifunctional protein. In support of this possibility, we have previously determined that DLC1 interacts with: 1) the tensin proteins, via an N-terminal region of DLC1 for which no function had been previously identified; 2) focal adhesion kinase (FAK) and talin, via a shared 8 amino acid motif; and 3) caveolin-1 (CAV-1), via the StAR-related lipid transfer (START) domain near the C-terminus of DLC1. DLC1 mutant analysis indicated that the binding sites for each of these interactions contributed to the growth suppressor function of DLC1, but that reduced binding seen with the mutants did not affect in vivo Rho-GAP activity of DLC1. These studies validate the hypothesis that DLC1 is a multifunctional protein whose biological activity depends both on its Rho-GAP activity and its ability to bind a variety of signaling molecules. We have also evaluated 12 cancer types in TCGA for the presence of point mutations in DLC1 and 9 other Rho-GAPs. The results indicate that DLC1 is the Rho-GAP that is mutated most frequently, with 5-8% of tumors in five of the tumor types evaluated having DLC1 missense mutations. Furthermore, 20-26% of the tumors in four of these five tumor types harbored missense mutations in at least one of the 10 Rho-GAPs. Experimental analysis of the DLC1 mutants indicated seven of nine mutants whose lesions were located in the Rho-GAP domain were deficient for Rho-GAP activity and for suppressing cell migration and anchorage-independent growth. Point mutations in DLC1 were scattered throughout the coding sequences of the gene, implying that important biological functions are present throughout the protein. These results indicate that point mutation in DLC1 is an important mechanism for its decreased function in cancer, in addition to gene deletion and epigenetic or other non-genetic changes. We have characterized how post-translational modifications of DLC1, especially phosphorylations, may affect its activity. We first identified CDK5, a cytoplasmic kinase whose physiologic role promotes differentiation, as a major activator of DLC1, which occurs via CDK5 phosphorylation of 4 serines in DLC1. When these serines, which are located N-terminal to the RHO-GAP domain of DLC1, are not phosphorylated, the N-terminus binds to the Rho-GAP domain, which places DLC1 in a closed, inactive conformation. When the serines are phosphorylated, it decreases the interaction of the N-terminus with the Rho-GAP domain, which places DLC1 in an open, active conformation. AKT is another kinase that directly phosphorylates DLC1, on 3 serines located N-terminal to the Rho-GAP domain. The effects of AKT antagonize those of CDK5, in that the AKT phosphorylations attenuate the Rho-GAP and tumor suppressor activities of DLC1 by changing DLC1 from an open, active configuration to a closed, inactive configuration. AKT lies downstream from receptor tyrosine kinases (RTKs), and our data indicate that increased Rho-GTP following stimulation of cells with RTK ligands is attributable to the activation of AKT and its attenuation of DLC1. We also identified SRC family kinases (SFKs) as direct targets for phosphorylating DLC1 on two tyrosines, which attenuate the RHO-GAP and tumor suppressor activities of DLC1. One tyrosine is located in the RHO-GAP domain, and its phosphorylation by SFKs reduces the RHO-GAP function of DLC1. The other tyrosine is located near the sequence required for binding tensin, and its phosphorylation reduces tensin binding. The findings with AKT and SFKs may have therapeutic potential, as treatment of tumor xenografts with AKT and/or SFK inhibitors has much greater antitumor activity against tumors that express DLC1 compared with isogenic tumors that do not express DLC1. The high antitumor activity from AKT and/or SFK inhibition is associated with loss of phosphorylation of the DLC1 by AKT and/or SFK together with reactivation of the RHO-GAP and tumor suppressor activities of DLC1. We have identified methylation of DLC1 as yet another biologically important post-translational modification. The methylation is attributable to cytoplasmic EZH2, a protein methylase that is overexpressed in many cancers and known mainly as a nuclear protein that regulates mRNA expression because it methylates Histone-H3. DLC1 methylation enables ubiquitnation of the methylated residue and proteasome-dependent degradation of DLC1. This multistep process means DLC1 has a shortened half-life in many tumors and is even absent in some. This process is reversible, and inhibition of EZH2, ubiqutination, or the proteasome can stabilize the protein. For EZH2, this is conceptually distinct from using its inhibition to change mRNA expression. By itself, increasing DLC1 protein levels does not have a major effect on growth regulation, but adding an AKT inhibitor and/or SRC inhibitor strongly reduces growth. This effect depends on DLC1 protein and is more than a correlation, as disrupting the DLC1 alleles abolishes most of the antitumor effect.