Header Image


1. Switching of TGFbeta signaling outcomes in normal and disease states (Collaborators: Peter Gray - Salk Insitute, Julia Tchou - UPenn, Joan Brugge - Harvard)

Progression of solid tumors to a metastatic stage accounts for over 90% of cancer mortality. Thus, it is critical to identify therapeutic strategies that target both primary and metastatic tumors. Epithelial-mesenchymal transition (EMT) negatively correlates with therapy response, contributing to intratumoral heterogeneity and systemic dissemination in breast cancer. We previously reported that pseudopodium-enriched atypical kinase one (PEAK1) promotes breast cancer cell EMT and metastasis by potentiating fibronectin - transforming growth factor beta (TGFβ) signaling cross-talk. Since eukaryotic initiation factor five A (eIF5A), a unique translation factor that is activated by post-translational hypusination, is required for PEAK1 expression, we hypothesized that TGFβ may directly regulate eIF5A activity as a novel means of promoting EMT, and that targeting this pathway may inhibit or reverse metastatic progression. In this regard, we provide evidence of an active eIF5A-EMT program in undifferentiated breast cancer tissue. Notably, blockade of eIF5A hypusination (via deoxyhypusine synthase, DHPS, inhibition) reduces PEAK1 translation, cell viability and TGFβ-induced EMT. Conversely, we demonstrate that TGFβ induces post-translational hypusination of eIF5A in metastatic breast cancer cells. TGFβ is known to activate histone deacetylase six (HDAC6) and HDAC6 was independently reported to promote eIF5A deacetylation and nuclear export, supporting its translation activity. When delivered in combination, HDAC6 and DHPS inhibitors synergize to sequester eIF5A to the nucleus, suppress eIF5A-dependent translation and potently kill metastatic breast cancer cells. To identify candidate pathways downstream of the eIF5A/PEAK1 axis during EMT, we generated a Cytoscape interactome using eIF5A signaling and PEAK1-induced EMT genes as search terms. All interactome component genes were then analyzed across two breast cancer patient studies available on the Cancer Bio Portal. Interestingly, SOX2, PIK3CA and EIF4A2 were the interactome nodes that exhibited copy number amplifications among patients harboring genomic alterations in the initial interactome search genes, and SOX2 amplification significantly and independently associated with decreased patient survival (p = 0.0476). Taken together, our results establish a novel signaling pathway by which TGFβ stimulates HDAC6 and DHPS function to activate cytoplasmic eIF5A and promote EMT and survival in breast cancer cells. This work also identifies new therapeutic reagents and strategies that may improve cancer patient survival.



2. Biomarker identification in pancreatic cancer (Collaborators: Michael Bouvet - UCSD, Claremont BioSolutions)

Pancreatic ductal adenocarcinoma (PDAC) has single-digit 5-year survival rates at <7%. There is a dire need to improve pre-malignant detection methods and identify new therapeutic targets for abrogating PDAC progression. To this end, we mined our previously published pseudopodium-enriched (PDE) protein/phosphoprotein datasets to identify novel PDAC-specific biomarkers and/or therapeutic targets. We discovered that integrin alpha 1 (ITGA1) is frequently upregulated in pancreatic cancers and associated precursor lesions. Expression of ITGA1-specific collagens within the pancreatic cancer microenvironment significantly correlates with indicators of poor patient prognosis, and depleting ITGA1 from PDAC cells revealed that it is required for collagen-induced tumorigenic potential. Notably, collagen/ITGA1 signaling promotes the survival of ALDH1-positive stem-like cells and cooperates with TGFβ to drive gemcitabine resistance. Finally, we report that ITGA1 is required for TGFβ/collagen-induced EMT and metastasis. Our data suggest that ITGA1 is a new diagnostic biomarker and target that can be leveraged to improve patient outcomes.

PDAC screening



3. Influence of the microenvironment on cancer progression and therapy resistance (Collaborators: Joan Brugge - Harvard, Julia Tchou - UPenn, Ben Spike - Huntsman Cancer Center)

4. Tissue regeneration and wound healing (Collaborators: Athanasia Panopoulos - Notre Dame, Peter Gray - Salk Insititute)

Stem cells are necessary for proper development, tissue homeostasis and regeneration while dysregulation of their activity leads to diseases such as diabetes and cancer. We hypothesized that the stem cell marker, Cripto (or TDGF1), is a regulator of tissue regeneration. Using the zebrafish model of caudal fin wound healing, we show that expression of the zebrafish Cripto homolog, one-eyed pinhead (or oep), is increased in blastemal stem cell tissue at 96 hours post amputation. We further demonstrate that Cripto is necessary and sufficient for stem cell-mediated regeneration in this in vivo model. We identified non-muscle Myosin IIs (MYH9/10) as novel Cripto-binding proteins using proteomics and confirmed these interactions by co-IP/Western blot and immunofluorescence (IF) in mesenchymal stem cells. Notably, dual pharmacological inhibition of Cripto and Myosin IIs did not reduce caudal fin regeneration beyond that of either inhibitor alone, suggesting that these proteins function in the same pathway during this process. Our previous work demonstrated that soluble (GPI anchor-cleaved) Cripto promotes tissue-specific stem cell functions via cell-surface binding to GRP78 (glucose-regulated protein 78). Here, we report that Myosin II inhibition reduces cell surface localization and production of soluble Cripto. Since GRP78 primarily functions as an ER chaperone that responds to cellular stresses, such as nutrient or oxygen depletion, we tested whether nutrient deprivation affects Cripto expression, sub-cellular localization or binding to Myosin IIs in mesenchymal stem cells. Interestingly, we found that glutamine is necessary, but not sufficient, to sustain normal Cripto expression and its Myosin II-associated functions. Using the BioGrid database, we discovered Rab11A as a co-binder of GRP78 and Myosin IIs. Given the endosomal/exosomal trafficking functions of Rab11A, we hypothesized that Cripto may in turn help traffic GRP78 or other cell surface proteins in a Myosin II-dependent manner. Inhibition of Myosin IIs in mesenchymal stem cells transfected with wild-type, constitutively active or dominant negative GFP-tagged Rab11A revealed that Myosin IIs act upstream of and in concert with Rab11A to regulate Cripto localization to Myosin II-rich membrane protrusions. Based upon these data, we propose a new model of Cripto function whereby Myosin II/Rab11A activities promote membrane localization of Cripto to facilitate a unique endosomal/exosomal transport processes. Furthermore, reduced glutamine metabolism limits cellular Cripto expression, restricting protein sorting and autocrine/paracrine signaling during stem cell-mediated tissue regeneration.



Trainees in our group will gain experience with the following techniques: Immunohistochemistry/Immunofluorescence, qPCR/RT-PCR, Site-Directed Mutagenesis, Cell Biology, Intravital Cancer Imaging, RNAi Techniques, Viral Gene Delivery, Confocal/TIRF/Epi Microscopy, Chick-CAM and Orthotopic Xenograft Tumor Models, Cell Proliferation/Migration/Cycle/Transformation Assays, Radio-Immunoassays of Ligand-Receptor Interactions, Luciferase-Response Assays, Phosphoprotein Analysis, Kinase Assays, iPSC/ESC Culturing and Characterization, Adipocyte/Neuronal Cell Differentiation, Bioinfomatics, TGF-beta Superfamily Signaling, Analysis of Stem Cell Properties in Cancer Cells, Primary Cell Isolation and Characterization from Normal and Tumor Tissue, Zebrafish Husbandry and Developmental Studies, and FACS