The branched actin regulator cortactin is a central element connecting signaling pathways with the actin cytoskeleton and plays a role in diverse cellular processes. Cortactin-mediated actin assembly is known to contribute to cancer aggressiveness by regulating secretion of molecules that promote cell motility, invasion, and tumor growth. To better understand the role of cortactin in protein secretion, I focused on molecular and cellular mechanisms underlying cortactin-mediated regulation of protein transport in this dissertation. I demonstrated that cortactin-mediated branched actin assembly regulates late endosomal/lysosomal compartment maturation and trafficking, and subsequent retrograde transport to the Golgi complex. In addition, I identified a novel molecular mechanism by which cortactin control of actin dynamics on late endosomes is regulated by the phosphoinositide PI(3,5)P2. Using a combination of biochemical and cellular imaging approaches I demonstrated that PI(3,5)P2 removes cortactin from late endosomal branched actin networks by competing with actin filaments for binding to cortactin. This interaction promotes net disassembly of branched actin networks at late endosomal membranes. Overall, these findings suggest that cortactin-PI(3,5)P2 interactions regulate actin dynamics on late endosomal membranes by promoting net actin turnover, thereby promoting late endosomal/lysosomal compartment maturation and trafficking.

Lung cancer remains the leading cause of cancer related deaths in the United States despite a significant number of advancements in the molecular diagnosis and treatment of this disease. Recent genome wide expression analyses of human lung cancer has identified a number of receptor tyrosine kinases (RTKs) as overexpressed and potentially representing molecular drivers of lung cancer. Among these RTKs identified was EPHA2, which is highly expressed in lung cancer correlating to poor clinical outcomes. The role of EPHA2 in lung cancer, specifically in distinct molecular subtypes, is largely unknown. In this study, we dissected the role of EPHA2 in a variety of molecular subtypes of lung cancer and discovered that KRAS and EGFR(T790M) mutant lung cancers were most vulnerable to inhibition of EPHA2 by either genetic or pharmacological methods. We demonstrated functional evidence in vivo that the EPHA2 receptor is required for tumor growth and survival in both Kras(G12D) and EGFR(L858R+T790M) transgenic mouse models. We also showed that in lung cancer EPHA2 controls cell viability through a PI3K/mTOR dependent mechanism regulating apoptosis. Additionally, we identified a novel, ATP competitive EPHA2 RTK inhibitor, ALW-II-41-27, which was capable of inhibiting KRAS and EGFR mutant lung cancer cell viability both in vitro and in vivo. Overall these findings provide genetic, functional, mechanistic, and pharmacologic evidence that EPHA2 promotes the progression and survival of lung tumors. In addition, these findings provide rationale for the development of EPHA2 targeted therapeutics for clinical use.

CANCER BIOLOGY The role of mTORC2 in mammary morphogenesis and HER2-mediated tumorigenesis Meghan Melinda Morrison Dissertation under the direction of Professor Rebecca Cook The phosphatidyl inositol-3-kinase (PI3K)/Akt signaling pathway is aberrantly activated in nearly 60% of breast cancers, through HER2 amplification, PIK3CA mutation, PTEN inactivation, and other alterations. The mTOR complexes mTORC1 and mTORC2 operate as activators (mTORC2) or effectors (mTORC1) of Akt and control key cellular processes, including growth, survival and metabolism. While most studies of mTOR in breast cancer have focused on mTORC1, little is known about the distinct role of mTORC2 in breast cancer. As molecular pathways that are essential for normal development are often hijacked by cancer cells, we sought to test the requirement of Rictor/mTORC2 in untransformed mammary epithelial cells (MECs) and transformed MECs. We show that genetic disruption of Rictor decreased Akt phosphorylation, ductal length, secondary branching, cell motility, and cell survival, effects that were recapitulated with a pharmacological dual inhibitor of mTORC1/mTORC2, but not upon genetic disruption of mTORC1 function via Raptor deletion. Surprisingly, Akt re-activation was not sufficient to rescue cell survival, invasion, or branching of mTORC2-impaired MECs. However, PKCα-mediated signaling through the small GTPase Rac1 was necessary for mTORC2-dependent mammary epithelial development during puberty, revealing a novel role for Rictor- PKCα-Rac1 signaling axis in MEC survival and motility during branching morphogenesis. These data lead us to hypothesize that Rictor/mTORC2 drives PI3K/Akt-mediated cell survival of HER2-amplified breast cancer cells. RICTOR gene ablation in a transgenic mouse model of HER2-amplified breast cancer delayed tumorigenesis and decreased lung metastasis, Akt-S473 phosphorylation, cell growth and survival. In human HER2-amplified breast cancer cells, Rictor loss, but not Raptor loss, decreased Akt-S473 phosphorylation, reducing cell survival and motility/invasion. Interestingly, Rictor/mTORC2 loss or treatment with a dual mTORC1/2 inhibitor improved lapatinib-induced cell killing in parental and resistant tumor cells to a greater extent than mTORC1 inhibition. Akt re-activation rescued cell survival, but not motility/invasion, in Rictor-depleted cells. However, Rictor loss caused accumulation of the Rac inhibitor RhoGDI2, thus impairing Rac1-dependent invasion. We conclude that HER2-amplified breast cancers use Rictor/mTORC2 to drive Akt-mediated cell survival and Rac1-mediated cellular invasion. These studies support additional studies into mTORC2-specific inhibitors. Approved Rebecca Cook 6/26/15

Intestinal homeostasis relies on complex interactions between the intestinal epithelium, microbiota, and host immune system, all of which cooperate to maintain homeostasis in an environment colonized and challenged by an estimated 100 trillion bacteria. It is the disruption and dysregulation of these interactions that is currently thought to underlie intestinal pathologies such as inflammatory bowel disease (IBD) and colorectal cancer (CRC). The Myeloid translocation gene (MTG) family of transcriptional corepressors and the tight-junction associated protein Blood vessel epicardial substance (BVES) have been shown to be important regulators of the intestinal epithelium. I hypothesized that loss of MTG or BVES function play important roles in IBD and CRC. Using genetic mouse models, cell lines, and organoid cultures, we clarified the role MTGs and BVES play in intestinal pathologies by identifying their binding partners and what signaling pathways they regulate. Overall, the primary objective of my work was to further elucidate the functional role of MTGs and BVES in the intestinal epithelium and carcinogenesis.

Cancer is the second leading cause of death in the United States. It is important to identify novel drivers of cancer in order to understand these diseases so we can improve patient prognosis. This dissertation focuses on two types of cancer, colorectal adenocarcinoma (CRC) and head and neck squamous cell carcinoma (HNSCC), and the discovery of novel regulators in each using different methods. In CRC, publically available data from The Cancer Genome Atlas (TCGA) was utilized. A bioinformatics analysis was used to connect protein expression data to clinical outcome, thus identifying proteins that are associated with poor prognosis in CRC. This analysis identified both known and novel regulators of CRC. Two of these molecules were further studied: insulin-like growth factor binding protein 2 (IGFBP2) and GATA3. These initial findings were validated by follow-up experiments. High levels of IGFBP2, a known but understudied molecule in CRC progression, were associated with poor prognosis. Using a tissue microarray (TMA) stained with IGFBP2, I found that IGFBP2-high staining patients had significantly decreased survival and recurrence-free survival. Low levels of GATA3, a transcription factor never associated with CRC, were associated with poor prognosis. With in vitro experiments using GATA3-manipulated cell lines, I showed that GATA3 overexpression decreases three-dimensional colony growth and invasiveness but not intrinsic proliferation rates. Cortactin is overexpressed through the 11q13 amplicon in 30-40% of HNSCC cases and controls a number of actin-based cellular phenotypes important for tumor progression. I found the cortactin SH3 domain was critical for Golgi morphology and in vivo tumor growth. Shank2, a cortactin SH3 binding protein that is also overexpressed through the 11q13 amplicon, has never been studied in any type of cancer. Using Shank2-manipulated cells, I found Shank2-knockdown decreased invasion, and preliminary evidence suggested a decrease in MMP secretion and invadopodia activity. Identification and characterization of these novel regulators, IGFBP2, GATA3, and Shank2, gives us mechanistic insight into the progression of CRC and HNSCC.

EGFR-mutant lung cancers are highly sensitive to EGFR tyrosine kinase inhibitors (TKIs; erlotinib/gefitinib/afatinib), but tumors develop drug resistance within 9-16 months. Resistance to gefitinib/erlotinib commonly occurs via a second-site EGFR mutation, T790M. Two strategies to overcome T790M+ resistance are mutant-specific EGFR TKIs, such as AZD9291, and dual inhibition of EGFR with afatinib plus the anti-EGFR antibody, cetuximab (A+C). Unfortunately, second-line acquired resistance to A+C and AZD9291, after first-line acquired resistance to erlotinib/gefitinib/afatinib, also occurs. To prevent/delay resistance to AZD9291, the combination of AZD9291 plus selumetinib (MEK1/2 inhibitor; AZD6244/ARRY-142886) is also currently being tested in a Phase I clinical trial (NCT02143466). The effects of sequential and combination treatment with various anti-EGFR agents on tumor evolution and drug resistance are largely unknown. In these studies, we modeled drug resistance pre-clinically to: 1. Assess the heterogeneity of mechanisms of first-line resistance to erlotinib and afatinib 2. Determine the optimum order of treatment with A+C vs. AZD9291 in the setting of T790M+ EGFR-mutant lung tumors 3. Elucidate mechanisms of first- and second-line acquired resistance to AZD9291 and 4. Elucidate mechanisms of resistance to AZD9291 plus selumetinib. Next-generation sequencing of genomic DNA from cell line models of resistance to erlotinib/afatinib revealed multiple potentially functional genomic changes within a given pool of resistant cells (including T790M). We also found that AZD9291 is more potent than A+C at inhibiting cell growth in the setting of T790M+ resistance to erlotinib. A+C-resistant cell lines remain sensitive to AZD9291, but AZD9291-resistant cell lines are cross-resistant to A+C. Resistance to AZD9291 is associated with dysregulation of MAPK signaling and can be overcome by addition of the MEK 1/2 inhibitor, selumetinib. Finally, AZD9291 plus selumetinib-resistant cell lines display increased baseline phospho-MEK/ERK and are sensitive to in vitro treatment with an ERK inhibitor, SCH772984 or alternative MEK inhibitor, trametinib. These studies provide a more comprehensive understanding of how EGFR-mutant tumors undergo rewiring of their signaling circuitry in response to single-agent EGFR- and combined EGFR+MEK-inhibition. This work, emphasizing a mechanistic understanding of the effects of therapies on tumor evolution, provides a framework for future clinical trials testing different treatment sequences.

Lung cancer is the leading cause of cancer-related death in the United States and worldwide. Lung cancers that are driven by mutations in the Epidermal Growth Factor Receptor (EGFR) are treated with targeted anti-EGFR therapy, such as first-generation tyrosine kinase inhibitors (TKIs) erlotinib or gefitinib. Unfortunately, all patients whose tumors initially respond to erlotinib or gefitinib will eventually go on to develop primary acquired resistance, most commonly through a second-site EGFR mutation, T790M. The combination of a second-generation EGFR TKI, afatinib, plus the anti-EGFR antibody, cetuximab, may overcome primary acquired resistance. Unfortunately, secondary acquired resistance is also seen with the new drug combination, and the underlying mechanisms have not yet been determined. Using tumor sample from a patient with secondary acquired resistance, we identified two unexpected NF2 mutations, neither of which was detected in the patients pre-treatment sample, suggesting that mutations in NF2 were acquired during treatment. The NF2 gene encodes merlin, a known tumor suppressor that is thought to regulate EGFR. Using cell line and animal models, we demonstrate that loss of NF2 is sufficient to induce resistance to TKI in EGFR-mutant cells. Co-treatment with the mTORC1 inhibitor, rapamycin, re-sensitizes cells to TKI. These studies suggest a novel potential mechanism of acquired resistance to targeted therapy in EGFR-mutant lung cancer. To further explore acquired resistance, we also evaluated a novel, irreversible, mutant-specific third-generation EGFR TKI, AZD9291. Pre-clinically, this drug potently inhibits signaling pathways and cellular growth in both EGFR-mutant and EGFR-mutant/T790M mutant cell lines in vitro, with lower activity against wild-type EGFR lines, translating into profound and sustained tumor regression in EGFR mutant tumor xenograft and transgenic models. We use cell line and mouse models to characterize AZD9291, providing the foundation for a recently-initiated clinical trial. In sum, this work evaluates mechanisms of acquired resistance to targeted therapy in EGFR-mutant lung cancer and identifies potential strategies to overcome such resistance.

There is a major need to better understand the molecular basis of triple negative breast cancer (TNBC) in order to develop effective therapeutic strategies. Using gene expression data from 587 TNBC patients we previously identified six subtypes of the disease, among which a mesenchymal-stem like (MSL) subtype. The MSL subtype has significantly higher expression of the transforming growth factor beta (TGF-β) pathway-associated genes relative to other subtypes, including the TGF-β receptor type III (TβRIII). We hypothesized that TβRIII is tumor promoter in mesenchymal-stem like TNBC cells. Representative MSL cell lines SUM159, MDA-MB-231 and MDA-MB-157 were used to study the roles of TβRIII in the MSL subtype. We stably expressed short hairpin RNAs specific to TβRIII (TβRIII-KD). These cells were then used for xenograft tumor studies in vivo; and migration, invasion, proliferation and three dimensional culture studies in vitro. Furthermore, we utilized human gene expression datasets to examine TβRIII expression patterns across all TNBC subtypes. TβRIII was the most differentially expressed TGF-β signaling gene in the MSL subtype. Silencing TβRIII expression in MSL cell lines significantly decreased cell motility and invasion. In addition, when TβRIII-KD cells were grown in a three dimensional (3D) culture system or nude mice, there was a loss of invasive protrusions and a significant decrease in xenograft tumor growth, respectively. In pursuit of the mechanistic underpinnings for the observed TβRIII-dependent phenotypes, we discovered that integrin-α2 was expressed at higher levels in MSL cells after TβRIII-KD. Stable knockdown of integrin-α2 in TβRIII-KD MSL cells rescued the ability of the MSL cells to migrate and invade at the same level as MSL control cells. We have found that TβRIII is required for migration and invasion in vitro and xenograft growth in vivo. We also show that TβRIII-KD elevates expression of integrin-α2, which is required for the reduced migration and invasion, as determined by siRNA knockdown studies of both TβRIII and integrin-α2. Overall, our results indicate a potential mechanism in which TβRIII modulates integrin-α2 expression to effect MSL cell migration, invasion, and tumorigenicity.

LIM domain Only-2 (LMO2) is a T cell oncogene whose molecular mechanism for T-cell transformation still remains to be elucidated. There are two hypotheses that could explain how the enforced expression of LMO2 induces T-cell Acute Lymphoblastic Leukemia (T-ALL). The first hypothesis is that LMO2 enforced expression results in the functional deficiency of E2A proteins. LMO2 overexpression correlates with the loss of E2A proteins, or Heb, suggesting that LMO2 functions by creating a deficiency of E2A either by redirection, sequestration, or turnover. To test this hypothesis, we enforced the expression ofE47Lmo2-induced T-ALL cell lines using an E47/estrogen receptor fusion construct that could be forced to homodimerize with 4-hydroxytamoxifen (4HT). We discovered that forced homodimerization triggered a G1 cell cycle growth arrest in 2 of the 4 lines tested. Transcriptome analysis suggested that E47 has remarkably diverse effects in T-ALL but that functional deficiency of E47 is not a universal feature of Lmo2-induced T-ALL. The second hypothesis, suggests that Lmo2 enforced expression results in transcriptional regulation of genes needed for T-cell transformation. The gene of interest in our studies is Hematopoietically expressed homeobox (Hhex). Hhex is a T-cell oncogene that is frequently deregulated in murine retroviral insertional mutagenesis screens and its enforced expression induces T-cell leukemia in bone marrow transduction and transplantation experiments. To further understand Hhexs function, we induced a conditional knockout in floxed Hhex mice with the Vav-iCre transgene. Mice were viable and showed normal blood cell counts with highly efficient deletion of Hhex in all hematopoietic tissues. Most impressively, Hhex conditional knockout markedly prolonged T-ALL onset in CD2-Lmo2 transgenic mice. Hhex conditional knockouts also had a significant decrease in mature B cells in the spleen and bone marrow. Bone marrow transplant and thymic repopulation studies revealed that loss of Hhex is important not only B cell development but also early T cell development. Our experiments show that Hhex is a critical transcription factor in lymphoid development and in LMO2-induced T-ALL.

Breast cancer is the leading cancer diagnosis in pre-menopausal women in the U.S. Of these breast cancers, 25% are diagnosed 2-5 years post-partum. Unfortunately, post-partum breast cancers (ppBCs) are highly metastatic. The reasons underlying the exaggerated lethality of ppBCs are unclear, but relate to stromal remodeling events that occur during post-partum involution, when milk-producing mammary epithelial cells (MECs) undergo widespread cell death. Upon engulfment of apoptotic cells (efferocytosis), immune suppressive wound healing cytokines are produced including Tgfβ1, IL4, and IL10, which increase macrophage polarization towards an M2 phenotype. Wound healing/TH2-like cytokines and M2 macrophages each correlate with decreased disease-free survival in breast cancer patients. Transforming growth factor (TGF)-β1 is a pleiotropic cytokine that has gained interest for its role in both tumor progression and metastasis. TGFβ1 operates to suppress cytotoxic immunity, increase fibroblast activation and collagen deposition, increases migration and invasion of tumor epithelial cells, and may enhance cancer stem cell-like properties in some tumor cells. TGFβ1 is abundantly up-regulated in response to efferocytosis during post-partum involution of the mammary gland. We recently showed that TGFβ1 is induced upon macrophage-mediated efferocytosis in the mammary tumor microenvironment (TME). Secreted TGFβ1 also acts on mammary tumor epithelial cells to increase epithelial-mesenchymal transition (EMT), cell adhesion, motility, and invasion. We utilized an immune-competent mouse model of ppBC, which recapitulates many of the clinical aspects of ppBCs in patients, including increased metastasis in response to post-partum events to show that inhibition of efferocytosis for the first 7 days of involution prevents M2 macrophage polarization, blocks induction of TGFβ1, and prevents metastasis of ppBCs through involution day 40. Our preliminary results show that ppBCs treated with the neutralizing anti-TGFβ antibody 1D11 for the first 14 d of involution produce fewer lung metastases as compared to control IgG-treated ppBCs.