Targeting EDC3 Phosphorylation to Regulate Tumor Growth and Invasion Through Controlling P-Body Dynamics

Abstract

Processing bodies (P-bodies) are cytoplasmic mRNA granules that serve as hubs of mRNA repression, degradation, and processing. Initially described as membrane-less organelles, P-bodies stand at a junction between protein translation and RNA decay. P-bodies, while present at low levels within normal cells, increase in response to common stressors such as hypoxia and nutrient starvation within the cell. As part of the integrated stress response, P-bodies form within the cell, and mRNA localize within them, which can ultimately change the fate of these RNA localized within. Tumors undergo many stresses within the course of the disease. Hypoxia, reactive oxygen species, chemotherapy, and many other stresses can lead to markedly increased P-body numbers, suggesting their importance in tumorigenic stress responses. P-bodies likely regulate cell fate by regulating mRNA levels and, thus, protein expression contributing to tumor initiation, growth, and metastasis. Despite their crucial role in controlling the fate of RNA messages within the cell, it is still unclear how P-body formation and activity are regulated. Within tumors, the dysregulation of oncogenic kinases has been shown to drive many factors like growth, invasion, and metastasis. Pim and AKT kinases are dysregulated within many cancer types, including both solid and hematologic malignancies. These two kinases have significant overlap in their targets and often work together to enhance or diminish signaling cascades within the cell. We find a core P-body component enhancer of mRNA-decapping protein 3 (EDC3), that functions to enhance mRNA decapping and degradation, to be a substrate for the known drivers of cancer growth Pim and AKT protein kinases. These kinases phosphorylate EDC3 on serine 161 (EDC3 S161, Ser161), which we show to be heavily phosphorylated in every cell line tested across multiple tumor types. Interestingly when compared to non-tumorigenic, immortalized cell lines, we see little to no phosphorylation of EDC3. Dual immunohistochemical staining of a resected breast tumor sample stained with a phospho-specific antibody for EDC3 S161 and antibodies against the total protein demonstrate that within the tumor, EDC3 is almost entirely phosphorylated. In contrast, the surrounding normal breast tissue within the same patient demonstrated very low levels of phosphorylation despite having roughly similar levels of protein. Within the tumor, it appears that EDC3 phosphorylation is controlling its localization into P-bodies under both normal and stress conditions. Using immunofluorescence (IF) staining of EDC3, we tested the effect of kinase inhibition on the cells' ability to form P-bodies. Under normal conditions, which show heavily phosphorylated EDC3, we see few P-bodies in each cell. When phosphorylation is inhibited following dual kinase inhibition, we see a dramatic increase in P-body number. This increase is consistent in times of cellular stress, again showing an increase in P-body numbers following inhibition of the kinases. When EDC3 is knocked out, or in a phospho-mimetic state, we fail to see any increase in the amount of P-bodies following kinase inhibition. As EDC3 is an RNA binding protein and core P-body component, we wanted to examine how kinase inhibition might influence the RNA content within P-bodies. To determine which RNAs were regulated by P-bodies, RNA-seq was performed on isolated P-bodies following affinity purification in PC3-LN4 cells using GFP-EDC3. P-body purification was performed in the setting of both Pim and AKT inhibitor treatment and DMSO control-treated cells. P-bodies contained roughly 7,000 differentially regulated mRNAs, 3711 enriched and 3381 decreased relative to normalized levels of total cellular mRNA. In cells treated with Pim and AKT inhibitors, the treatment caused significant changes in 313 specific mRNAs, 133 being enriched into P-body's, while 180 were depleted within the granules. Among these changes were genes involved in metabolic processes, transcription, migration, and cell growth. Additionally, we see shifts in every class of RNA within the P-body following drug treatment with a slight decrease in protein-coding RNA, and increased localization of long-non-coding RNA when compared to DMSO treated P-bodies. We also saw changes in the specific RNA binding motifs within the P-body, suggesting that phosphorylation of EDC3 can change the RNA content within P-bodies. High-throughput screening of potential protein-protein interaction partners showed that EDC3 and Pim were predicted to interact. We went on to test this interaction and show tight binding between EDC3 and Pim1 and 3 while failing to interact with Pim2. Separating EDC3 into its functional domains showed the site of interaction occurring in the intrinsically disordered region (IDR) of EDC3, which contains the phosphorylation site. Protein NMR was conducted to determine the exact interaction motifs, and it was discovered to bind 30 peaks within the EDC3 IDR, that we identified to be residues 137 to 167, which include the phosphorylation site at 161. Computer modeling shows that EDC3 binds in a previously uncharacterized pocket of Pim1 kinase that is distinct from the ATP-binding domain. Interestingly, Pim1 and 3 both contain this pocket, while Pim2 does not. Consistent with the fact that Pim2 failed to interact with EDC3. Mutating 4 residues predicted to interact with EDC3 in this pocket, we see a failure of EDC3 to interact. Importantly this protein still has kinase activity, despite losing the tight-binding seen with EDC3. Converting serine 161 to alanine by CRISPR-Cas9 in PC3-LN4 human prostate cancer cells inhibited tumor growth and migration both in tissue culture and mouse models, suggesting that this protein modification is an essential regulator of cell growth. These cells also failed to invade and migrate when compared to the naïve cells. Conversion of this site to aspartic acid to mimic a phosphorylated state, we witnessed a dramatic increase in the cells' ability to migrate and invade, suggesting the overall importance in the fate of the cell. Interestingly we also see a reduction in protein stability when mutations in the backside pocket are present. Because of this, and the potential to make a targeted Pim1 inhibitor, we set about testing potential "backside" pocket inhibitors. Surface Plasmon Resonance (SPR) screening showed several hits; whose activity was tested and shown to reduce EDC3 phosphorylation and binding, as well as cause cell death similar to that of Pan-Pim kinase inhibitors. Together these data demonstrate that the phosphorylation of EDC3 by the Pim and AKT kinases, in tumors, regulates the translation of specific mRNAs via potentially changing their storage and degradation, which in turn alters cell growth, migration, and tumor motility and that the phosphorylation of EDC3 is essential in the cells ability to form P-bodies. We also show that EDC3 and Pim physically interact, that this interaction occurs in the IDR of EDC3. Further, we predict that EDC3 binds to a newly characterized backside pocket in Pim1 and 3 and that this serves as a potentially druggable site (Fig. 0-1).