The ability to make transgenic animals has been a great tool for biologists to study living organisms. In C. elegans, the way transgenes are generated makes them problematic in many circumstances, and there is no single, simple, reliable approach that circumvents all of the problems with current methods of introducing transgenes into C. elegans. In Chapter 2, I discuss my attempt to develop a transgenic system in C. elegans using the bacteriophage phiC31 integrase system. I show evidence that phiC31 integrase is active in C. elegans somatic tissue. I have successfully integrated a transgene into the C. elegans genome in single-copy using phiC31 dependent recombination-mediated cassette exchange. However, attempts to repeat phiC31-mediated integration has been unsuccessful. In Chapter 3, I use genetic analysis to test many genes that were reported to be associated with the gamma-secretase complex in a mammalian tissue culture system. The gamma-secretase complex is an important component in the Notch signaling pathway. Not only is the gamma-secretase complex essential in the Notch pathway, it is also implicated in the pathology of familial Alzeheimer's disease (FAD). As gamma-secretase complex components show a Notch loss-of-function phenotype in C. elegans, a reverse genetic approach, using genes encoding proteins that associate with Presenilin was used to identify putative new Notch modulators. Several genes were identified that suppress a glp-1(gf) allele and one gene that suppress a gfp-1(lf) allele. These genes are unlikely to be core components of the Notch signaling pathway.

Cell motility is an essential process that depends on a coherent, cross-linked cytoskeleton that physically coordinates the actions of numerous structural and signaling molecules. In culture, a common feature of cells is the coherent movement of the endoplasmic reticulum and membranous organelles toward the periphery during substrate adhesion and spreading. The actin cross-linking protein, filamin (Fln), has been implicated in the support of three-dimensional cortical actin networks capable of both maintaining cellular integrity and withstanding large forces. Although numerous studies have examined cells lacking one of the multiple Fln isoforms, compensatory mechanisms can mask novel phenotypes only observable by further Fln depletion. Indeed, shRNA-mediated knockdown of FlnA in FlnB-/- mouse embryonic fibroblasts (MEFs) causes a novel endoplasmic spreading deficiency as detected by endoplasmic reticulum markers. Microtubule (MT) extension rates are also decreased but not by peripheral actin flow, because this is also decreased in the Fln-depleted system. Additionally, Fln-depleted MEFs exhibit decreased adhesion stability that leads to increased ruffling of the cell edge, reduced adhesion size, transient traction forces, and decreased stress fibers. FlnA-/- MEFs, but not FlnB-/- MEFs, also show a moderate defect in endoplasm spreading, characterized by initial extension followed by abrupt retractions and stress fiber fracture. FlnA localizes to actin linkages surrounding the endoplasm, adhesions, and stress fibers. Thus I suggest that Flns have a major role in the maintenance of actin-based mechanical linkages that enable endoplasmic spreading and MT extension as well as sustained traction forces and mature focal adhesions. I also report that treatment with the calpain inhibitor N-[N-(N-Acetyl-L-leucyl)-L-leucyl]- L-norleucine (ALLN) restores endoplasmic spreading and focal adhesion (FA) maturation in the absence of Fln. Further, expression of calpain-uncleavable talin, but not full-length talin, also rescues endoplasmic spreading in Fln-depleted cells and indicates a crucial role for stable, mature FAs in endoplasmic spreading. Because FA maturation involves the vimentin intermediate filament (vIF) network, I also examined the role of vIFs in endoplasmic spreading. Wild-type cells expressing a dominant-negative vimentin variant incapable of vIF polymerization exhibit deficient endoplasmic spreading as well as defects in FA maturation. ALLN treatment restores FA maturation despite the lack of vIFs, but does not restore endoplasmic spreading. Consistent with a role for vIFs in endoplasmic spreading, adhesive structures do not contain vIFs when the endoplasm does not spread. Fln-depleted cells also exhibit a microtubule-dependent mistargeting of vIFs. Thus, I propose a model in which cellular force generation and interaction of vIFs with mature FAs are required for endoplasmic spreading. Additionally, I discuss future lines of investigation concerning the role of FlnA in the endoplasmic spreading mechanism as well as mechanosensitive functions of FlnA. Finally, I speculate on a potential application of endoplasmic spreading deficiencies as hallmarks of metastatic breast cancer.

MYC (or c-myc) is a bona fide "cancer driver" oncogene that is deregulated in up to 70% of human tumors. In addition to its well-characterized role as a transcription factor that can directly promote tumorigenic growth and proliferation, MYC has transcription-independent functions in vital cellular processes including DNA replication and protein synthesis, contributing to its complex biology. MYC expression, activity, and stability are highly regulated through multiple mechanisms. MYC deregulation triggers genome instability and oncogene-induced DNA replication stress, which are thought to be critical in promoting cancer via mechanisms that are still unclear. Because regulated MYC activity is essential for normal cell viability and MYC is a difficult protein to target pharmacologically, targeting genes or pathways that are essential to survive MYC deregulation offer an attractive alternative as a means to combat tumor cells with MYC deregulation. To this end, we conducted a genome-wide synthetic lethal shRNA screen in MCF10A breast epithelial cells stably expressing an inducible MYCER transgene. We identified and validated FBXW7 as a high-confidence synthetic lethal (MYC-SL) candidate gene. FBXW7 is a component of an E3 ubiquitin ligase complex that degrades MYC. FBXW7 knockdown in MCF10A cells selectively induced cell death in MYC-deregulated cells compared to control. As expected, cellular MYC levels are stabilized when FBXW7 expression is attenuated. Notably, stabilization of MYC is more pronounced compared to other FBXW7 targets. FBXW7 knockdown with MYC deregulation results in cell cycle defects, as well as CDC45 accumulation on chromatin, suggesting DNA replication stress. Intriguingly, FBXW7 and MYC expression correlate most strongly in the luminal A-subtype of breast cancer associated with low to normal MYC expression. Together, our results suggest that knockdown of FBXW7 increases cellular MYC levels and promotes cell death possibly through accumulation of MYC-dependent genomic stress, and that FBXW7 inhibition may be selectively synthetic lethal with breast cancers that retain MYC-dependence. We also identified UVSSA and ERCC8, two genes involved in transcription-coupled repair (TCR), as MYC-SL candidates from our genome-wide screen. TCR is a DNA damage repair pathway associated with active RNA polymerase II-transcription complexes. We show that both UVSSA and ERCC8 knockdown confer increased lethality selectively in MYC-deregulated cells. This MYC-SL interaction is not exacerbated by exogenous UV irradiation, suggesting that TCR may be required for survival upon MYC deregulation independently of its role in UV damage repair. UVSSA knockdown with MYC deregulation results in cell cycle defects and CHK2 activation, suggesting genomic stress. Intriguingly, we observe that lethality associated with UVSSA down-regulation in cells expressing MYCER is alleviated by inhibiting transcription. This suggests that transcription-dependent aberrant genomic structures generated during MYC deregulation may require TCR for maintaining survival. Taken together, our results suggest that increased levels of transcription-dependent genomic stress may accumulate with MYC deregulation, and that TCR may have functions outside of repairing UV-induced damage in resolving these lesions or structures.

It is well established that the p53 tumor suppressor plays a crucial role in controlling cell proliferation and apoptosis upon various types of stress. There is increasing evidence showing that p53 is also critically involved in various non-canonical pathways, including metabolism, autophagy, senescence and aging. Through a ChIP-on-chip screen, we identified a novel p53 metabolic target, pantothenate kinase-1 (PANK1). PanK1 catalyzes the rate-limiting step for CoA synthesis, and therefore, controls intracellular CoA content; Pank1 knockout mice exhibit defect in β-oxidation and gluconeogenesis in the liver after starvation due to insufficient CoA levels. We demonstrated that PANK1 gene is a direct transcriptional target of p53. Although DNA damage-induced p53 upregulates PanK1 expression, depletion of PanK1 expression does not affect p53-dependent growth arrest or apoptosis. Interestingly, upon glucose starvation, PanK1 expression is significantly reduced in HCT116 p53 (-/-) but not in HCT116 p53 (+/+) cells, suggesting that p53 is required to maintain PanK1 expression under metabolic stress conditions. Moreover, by using p53-mutant mice, we observed that PanK activity and CoA levels are lower in livers of p53-null mice than that of wild-type mice upon starvation. Similar to the case in Pank1 knockout mice, β-oxidation and gluconeogenesis are impaired in p53-null mice. Together, our findings show that p53 is critical in regulating energy homeostasis through transcriptional control of PANK1. Our study on PANK1 led us to the question of how p53 can differentially regulate a diverse array of downstream targets in a context-dependent manner. Studies have shown that p53 acetylation at K120 and K164 lysine residues contribute to p53-mediated apoptosis and growth arrest functions, which was further supported by the 3KR mouse model (K117/161/162R) that mirrors the K120/164R mutations in human p53. These studies also suggest that a potentially large number of p53 targets can still be regulated by p53 in the absence of K120/164 acetylation (K117/161/162R in mouse). To investigate whether additional modifications of p53 can further contribute to promoter-specific transactivation, we conducted a screen using mass spectrometry and identified a novel acetylation site at K101. Our data demonstrated that K101 in human p53, as well as the homologous K98 lysine residue in mouse p53, can be acetylated by acetyltransferase CBP. Acetylation at this novel site does not contribute to p53 stability or DNA-binding capabilities. Ablation of K98 acetylation in mouse p53 alone does not affect the transcriptional activity of p53. However, simultaneous loss of K98 acetylation with the previously characterized K117/161/162 acetylations (4KR98 p53) significantly abrogates p53-mediated activation of TIGAR and MDM2 genes. The 3KR mouse model, although cannot elicit canonical p53-mediated apoptotic and cell cycle arrest responses, still retains the ability to suppress tumor formation. We, therefore, investigated whether other non-canonical targets of p53 could potentially mediate tumor suppression. By RNA-seq profiling of gene expression in cells expressing 3KR p53, we identified TNFRSF14 (tumor necrosis factor receptor superfamily, member 14) as a novel p53 target. The TNFRSF14 receptor has been shown to be frequently mutated in follicular lymphoma and diffuse large B cell lymphoma, and stimulation by its ligand LIGHT leads to cell death in many cancer cells. We report that TNFRSF14 is a novel p53 target that can be activated by 3KR p53. Interestingly, transactivation of TNFRSF14 is defective by 4KR98 p53. Furthermore, LIGHT ligand stimulates cell death in TNFRSF14-expressing cells and cells expressing 3KR p53, but not those expressing 4KR98 p53. Altogether, our findings in these studies underscore the extensive scope of p53 functions and provide new insights into the versatility of non-canonical pathways. Not only does p53 mediate tumor suppression through both canonical and non-canonical downstream effectors, p53 can also contribute to cellular homeostasis and energy balance.

The development of insulin resistance (IR) in the liver is one of the key pathophysiologic events in the development of type 2 diabetes mellitus, but most patients do not become uniformly resistant to the hepatic actions of insulin. Although insulin loses its ability to blunt glucose production, it largely retains its capacity to drive lipogenesis. This "selective IR" results in the characteristic hyperglycemia and dyslipidemia of type 2 diabetes. In this thesis, we take two approaches to better understand the mechanisms underlying selective IR. First, the compensatory chronic hyperinsulinemia (CHI) of insulin resistance downregulates levels of the insulin receptor (InsR). We have therefore modeled CHI in primary hepatocytes to demonstrate that the reduction in InsR number results in insufficient signaling capacity to halt glucose production while still leaving enough residual signaling capacity to promote lipogenesis. That is, the two processes are inherently differentially sensitive to insulin. Second, we hypothesize that FoxO1, a key insulin-inhibited transcription factor, coordinately regulates both hepatic glucose and lipid homeostasis. We have developed a transgenic mouse model heterozygous for a knocked-in allele of DNA binding-deficient FoxO1 and have proceeded to dissect the mechanisms by which FoxO1 differentially regulates glucose and lipid handling. We found that while the former requires FoxO1 to bind to its consensus sequences in target-gene promoters, the latter proceeds via a co-regulatory action of FoxO1. Taken together, these findings reveal novel connections between the glucose and lipid "arms" of the insulin-signaling pathway and how they may go awry in the run-up to diabetes.

Prostate cancer is a common malignancy with nearly one million annual diagnoses worldwide. Among a subset of patients, primary disease eventually progresses to disseminated castration resistant prostate cancer (CRPC). In recent years, treatment modalities that improve survival in CRPC have emerged including taxane chemotherapy and second generation androgen signaling inhibitors, among others. Indeed, today the first line chemotherapeutic docetaxel as well as the second line agent cabazitaxel are mainstays of treatment. However, CRPC inexorably progresses to a chemotherapy resistant state that ultimately precedes lethality. Elucidating the molecular determinants of aggressiveness in chemotherapy resistant CRPC may therefore stimulate new therapeutic strategies that improve clinical outcomes. We used laboratory models and clinical databases to identify GATA2 as a regulator of chemotherapy resistance and tumorigenicity in this context. Whole genome expression profiling, clinical validation and genetic screening approaches revealed that GATA2 regulates a signature of cancer progression associated genes. Mechanistically, direct upregulation of the growth hormone IGF2 emerged as a significant mediator of the aggressive properties regulated by GATA2. IGF2 in turn activated IGF1R and INSR as well as a downstream polykinase program. The characterization of this regulatory axis prompted a combination strategy whereby dual IGF1R/INSR inhibition restored the efficacy of chemotherapy and improved survival in preclinical models. These studies reveal a GATA2-IGF2 aggressiveness axis in chemotherapy resistant prostate cancer and identify a therapeutic opportunity in this challenging disease.

The PHB-d protein family is an evolutionarily ancient family of integral membrane proteins with members in all taxa that are involved in a wide variety of biological process but share several common molecular properties: oligomerization, detergent resistant membrane (DRM) association, and regulation of other proteins. To better understand the biological roles of the PHB-d gene family and provide a starting point for analysis of their individual functions, I determined the expression patterns of all of the uncharacterized PHB-d genes in the nematode C. elegans. All eight of the proteins similar to mammalian stomatin are detectably expressed in neurons. The specificity of expression varies from 34 neuron types that express sto-4 to only one that expresses sto- 3. STO-1 protein is localized to amphid sensory cilia and is needed for optimal chemotaxis to diacetyl. This phenotype is likely to affect chemotaxis mediated by the AWA sensory neurons. Involvement of PHB-d proteins in ! chemosensation may be conserved in mammals. Two murine PHB-d genes are highly expressed in olfactory sensory neurons, but no olfactory phenotype has been described in mice lacking either gene. C. elegans senses gentle body touch via a mechanosensory channel complex that contains two Degenerin/Epithelial sodium channel (DEG/ENaC) subunits and two prohibitin homology domain (PHB-d) proteins. One of the PHB-d proteins (UNC-24) has a C-terminal sterol carrier protein 2 (SCP-2) domain and only a subtle mechanosensory abnormal (Mec) phenotype. The other PHB-d (MEC-2) is essential for channel function. I show here that unc-24 is needed for proper localization of MEC-2 to the channel complex, but that this function is unlikely to underlie the unc-24 phenotype. This result supports the hypothesis that PHB-d/SCP-2 proteins have a conserved role in localizing other PHB-d proteins and demonstrates that UNC-24 must have a second, unknown, role in mechanosensation. MEC-2 is embedded in the plasma membrane via a non-spanning hydrophobic hook. Mutation of a conserved proline (P134S) in the hydrophobic hook leads to complete touch insensitivity. Mutation of the equivalent proline in human stomatin converts the hook to a transmembrane domain and expels the PHB-d and C-terminus of the protein. We show here that the P134S mutation has the same effect on MEC-2 topology in HEK293T cells. This result has implications for MEC-2 function because the P134S mutation has previously been shown to affect some but not all MEC-2 biochemical properties. The conserved proline is found in all C. elegans and human stomatins and its mutation in human podocin leads to kidney failure.

TDP43 is an RNA and DNA binding protein that has been shown to play an integral role in disease mechanisms that underlie ALS. In fact, a common feature of the vast majority of ALS cases is the presence of TDP-43 aggregates in postmortem tissue from the brain and spinal cord. This finding has spurred research to understand the physiological roles of TDP-43 in the absence of disease, and how these roles are affected by disease. Current TDP-43 mouse models fail to faithfully and reproducibly recapitulate key aspects of ALS, possibly due to the transgenic approaches used. To address these concerns, we generated a targeted, conditional mouse model, and embryonic stem cell lines expressing either human WT or M337V mutant TDP-43 at equivalent levels. We show that expression of mutant hTDP-43 in mice with a mixed genetic background leads to selective motor neuron loss, muscle weakness and premature death. However this disease phenotype is not observed with the same TDP43 mutation in a pure Bl6 background. We next sought to identify alterations in the biochemistry of the mutant protein that may underlie its toxicity, such as its interactions with RNA and protein. By creating a library of RNAs bound by TDP-43 in the mouse spinal cord, we found that the M337V mutation does not compromise the ability of TDP43 to bind to target mRNA transcripts, although the mutation does lead to changes in expression of genes known to be involved in inflammation. In addition, we identified 22 proteins that bind to TDP43 in an RNA-dependent manner, and found that the M337V mutation does not alter these interactions. This work establishes novel mouse and cellular models that provide insights into the functions of normal and ALS-causing mutant TDP43 protein.

We investigated the role of Notch signaling in the vascular microenvironment, with particular attention paid to the vascular consequences of Notch signaling disruption in myeloid cells. We adapted an established in vitro model of angiogenesis to recreate interactions between endothelial sprouts and vascular support cells, including macrophages and vascular pericytes. We found that inflammatory polarization of macrophages increased their ability to foster angiogenesis, and that intact Notch signaling was essential to this phenomenon. We also demonstrated a role for Notch/Jagged1 signaling in the interaction between vascular pericytes and endothelial sprouts, the disruption of which limits the growth and maturation of vessel networks. We have also investigated the role of myeloid Notch signaling in vivo, using a number of developmental and pathological models of angiogenesis. We found that Notch inhibition leads to decreased myeloid cell recruitment to a broad variety of functionally distinct angiogenic sites. Importantly, we observed that myeloid Notch disruption has vascular consequences in both physiological and pathological angiogenesis. Myeloid Notch- inhibited mice exhibit decreased vascular complexity in the deep retinal plexus during development. Additionally, these mice show significantly increased vascular tuft formation in the setting of oxygen-induced retinopathy, suggestive of a heretofore- undescribed role for myeloid Notch signaling in the pathogenesis of this significant human disease. This body of work increases our understanding of the role of Notch signaling both in the dynamics of myeloid cells and in their specific contribution to angiogenesis in multiple disparate contexts. It also contributes to our understanding of a number of key models of human disease, and may prove useful in the development of novel therapies to treat those diseases. Further, we are confident that our new experimental methodology will allow continued fruitful reductive study of the complex intercellular interactions within the vascular microenvironment.

T-cell acute lymphoblastic leukemia (T-ALL) is an aggressive hematological tumor resulting from the malignant transformation of immature T-cell progenitors. Originally associated with a dismal prognosis, the outcome of T-ALL patients has improved remarkably over the last two decades as a result of the introduction of intensified chemotherapy protocols. However, these treatments are associated with significant acute and long-term toxicities, and the treatment of patients presenting with primary resistant disease or those relapsing after a transient response remains challenging. Oncogenic activation of NOTCH1 signaling plays a central role in the pathogenesis of T-cell acute lymphoblastic leukemia (T-ALL), with mutations on this signaling pathway affecting more than 60% of patients at diagnosis. However the transcriptional regulatory circuitries driving T-cell transformation downstream of NOTCH1 remain incompletely understood. Here we identify HES1, a transcriptional repressor controlled by NOTCH1 as a critical mediator of NOTCH1 induced leukemogenesis strictly required for tumor cell survival. Mechanistically, we demonstrate that HES1 inhibits leukemia cell death by repressing BBC3, the gene encoding the PUMA BH3-only proapototic factor. Finally, we identify perhexiline, a small molecule inhibitor of mitochondrial carnitine palmitoyltransferase-1, as a HES1-signature antagonist drug with robust antileukemic activity against NOTCH1 induced leukemias in vitro and in vivo.