Epithelial-cadherin downregulation enables cancer cells to escape from the primary mass; however, E-cadherin has been found to be expressed on metastatic foci, bringing into question the role of this molecule in tumor progression. We define a novel role for the cellular adhesion molecule E-cadherin, in which the proteins re-emergence promotes carcinoma-parenchymal interactions in ectopic sites. Non-metastatic E-cadherin positive MCF7 breast cancer cells form heterotypic cohesions mediated by E-cadherin, and in invasive and metastatic MDA-MB-231 cells, the E-cadherin promoter hypermethylation that prevents endogenous E-cadherin expression is reversed when these cells are cultured with hepatocytes. The function of this re-expression is suggested by the E-cadherin-dependent sustained activation of Erk-MAP kinase and Akt in these breast carcinoma cells. Thus, we propose that E-cadherin expression and subsequent heterocellular interactions direct cell fate decisions that may ultimately enable colonization of a secondary site by an invasive cancer cell.

Environmental triggers, such as viral infection and environmental toxins, have been proposed to initiate the autoimmune disease of Type 1 Diabetes (T1D), however, the mechanism is unknown. The identification of novel autoantigens may provide insight to the mechanism of environmental triggers and pathogenesis of T1D. I identified the antigen recognized by the diabetogenic BDC- 2.5 T cell clone using a novel in vivo reconstitution system, Restricted Immune System via Adoptive Transfer (RISAT). In RISAT, immunodeficient mice are adoptive transferred with a single T cell clone and an open repertoire of B cells. Reconstituted mice are immunized with an antigenic protein preparation. This system will drive an antibody response to the cognate antigen for the T and B cell through the co-stimulatory pathways involved in linked recognition. For the BDC-2.5 RISAT, non-obese diabetic (NOD).Rag-/- mice were adoptive transferred with the diabetogenic BDC-2.5 T cells and NOD B cells and then immunized with an antigenic beta cell membrane preparation (βmem) to drive an antibody response. The resulting antibodies recognized the endoplasmic reticulum (ER) stress associated protein glucose regulated protein 78 (GRP78) from βmem. To determine if ER stress plays a role in the antigenic response of the BDC-2.5 T cell clone, the non-antigenic NOD insulinoma cell line, NIT-1, were treated with thapsigargin, which induces ER stress. The treatment of NIT-1 with thapsigargin led to increased GRP78 synthesis, correlating with antigenic recognition by the BDC-2.5 T cell clone. The antibodies from the BDC-2.5 TCR-Tg recognizes a subset of GRP78 which is modified with phosphoserine. The data presented in this thesis demonstrates a mechanistic link between ER stress and environmental triggers leading to the initiation of TID through the novel autoantigen, GRP78. Also the technique, RISAT, can be used to identify additional potential autoantigens of isolated T cell clones in both T1D and other autoimmune diseases.

Cardiac allograft arteriosclerosis (CAA) is the leading cause of death in cardiac transplantation recipients. Nitric oxide (NO) produced by NO donors or nitric oxide synthase (NOS) inhibits smooth muscle cell proliferation and migration, promotes endothelial cell proliferation, blocks endothelial cell apoptosis, prevents platelet and leukocyte adhesion and platelet aggregation. NO has been shown to inhibit intimal hyperplasia and suppress the development of cardiac allograft arteriosclerosis. Conversely, induction of NO in cardiomyocytes leads to myocardial apoptosis and suppresses cardiac function, mainly represented by decreased cardiac contractility. Based on these studies, we hypothesize that cell type-specific overexpression of eNOS or iNOS, in smooth muscle or endothelial cells, will inhibit the development of cardiac allograft arteriosclerosis without suppressing cardiac function. To test this hypothesis, we developed two gene delivery systems that specifically express NO in smooth muscle or endothelial cells. One system we constructed is adenoviral vectors that employed the fms-like tyrosine kinase-1 (Flt-1, -748 ~ +284 bp), or the intercellular molecule 2 (ICAM2, -292 ~ +44 bp) promoters to drive the human iNOS specific expression in endothelial cells. Another system we developed is vascular-specific iNOS inducible transgenic mice that employed the GeneSwitch system (Invitrogen) and vascular specific promoter. We obtained two iNOS transgenic lines, in which iNOS can be induced by the ligand mifepristone with the supply of the Switch protein. To determine the efficacy of eNOS endothelial-specific expression on suppression of cardiac allograft arteriosclerosis, we employed eNOS transgenic mice in which the human eNOS promoter was driving the human eNOS expression. In a mouse heterotopic transplantation model, the male transgenic and female wild-type counterparts were used as donor and recipient, respectively. The donor grafts were harvested 60 days after transplantation and the percentage of vessel occlusion was determined. The vessel occlusion developed in the transegenic donor did not show significant difference compared to the wild-type control. As an alternative approach, the eNOS/GTPCH double transgenic mice will be used as donor and its efficacy will be determined. We hope that these proof-of -principle studies will serve as the basis for the development of an effective treatment strategy to prevent cardiac allograft arteriosclerosis.

Phosphatidylinositol-3-kinase (PI3K) is a well-known regulator of cell division, motility, metabolism and survival in most cell types. Proper liver function and development highly depend on intact PI3K signal transduction. Aberrant PI3K pathway signaling in the liver is associated with hepatocellular carcinoma (HCC). In addition, PI3K signaling is involved in the homeostasis of lipid and glucose metabolism. Activation of the PI3K pathway induces lipogenesis and glycogenesis in the liver, since both Akt overexpressing transgenic mice and PTEN knockout mice develop fatty liver and hypoglycemia. Our laboratory characterized a novel protein that we call PI3K Interacting Protein 1 (PIK3IP1) which binds to the p110 catalytic subunit of PI3K and reduces its activity in vitro. Little is known about PIK3IP1s role in tumorigenesis and metabolism in vivo. Therefore we constructed PIK3IP1 transgenic mice (TG) which overexpress PIK3IP1 in hepatocytes under an albumin promoter in the C3H mouse strain to investigate the effect of PIK3IP1 on hepatocyte growth and metabolism, as well as HCC tumorigenesis. We detected a high expression level of PIK3IP1 in the livers from TG animals. The PI3K pathway was successfully suppressed both in liver tissues and isolated hepatocytes, which was confirmed by Western blots and phospho-protein array studies. Given the fact that PI3K signaling is associated with liver tumorigenesis, our next objective was to determine whether PIK3IP1 inhibits HCC development through PIK3IP1- mediated downregulation of the PI3K pathway. In vivo, spontaneous liver tumorigenesis was significantly dampened in the transgenic animals. This was accompanied by decreased hepatic PI3K activity and reduced hepatocyte proliferation in the transgenics as compared to controls. Isolated PIK3IP1 transgenic mouse hepatocytes showed blunted PI3K signaling, DNA synthetic activity, motility and survival as compared to controls. We then investigated the effect of PIK3IP1 on the maintenance of whole-body glucose and fat homeostasis. We observed that mice overexpressing PIK3IP1 have increased body weight, hyperglycemia, as well as increased visceral fat deposition. This suggests PIK3IP1 is an important regulator of metabolism. In conclusion, we successfully generated a transgenic mouse model with PIK3IP1 overexpression in hepatocytes to assess the biological functions of PIK3IP1, an important negative regulator of PI3K, in liver tumorigenesis and insulin signaling. A high level expression of PIK3IP1 suppressed PI3K signaling pathway in vivo and in vitro, which curbs hepatic tumorigenesis. Furthermore, we show that PIK3IP1 overexpression can contribute to glucose homeostasis and fatty deposition.

Vascular Endothelial Growth Inhibitor (VEGI) is an endothelial cell autocrine factor and a member of the tumor necrosis family of ligands. VEGI is able to specifically inhibit endothelial cell growth and is an efficient inhibitor of angiogenesis. The molecular mechanisms of VEGI activity on endothelial cells remain undefined. Here we focused on two important steps in the signal transduction of VEGI. We first determined a role of NF-κB in VEGI-induced apoptosis. We found that inhibition of the NF-κB pathway resulted in an increased apoptotic potential of VEGI. We conclude that the NF-κB pathway plays a role in suppressing the apoptotic potential of VEGI. We next investigated the receptor responsible for VEGI-induced endothelial cell apoptosis. DR3 is a receptor for VEGI and thus we first focused on confirming if DR3 is the receptor responsible for VEGI-mediated endothelial cell death. We determined VEGI had diminished apoptotic activity in endothelial cells that are depleted of DR3 by siRNA. However, it was determined that the apoptotic stimuli, LPS and TNFα, were also unable to mediate cell death in DR3-depleted endothelial cells. We conclude that DR3 is mediating an intracellular event that is involved in controlling the apoptotic pathway. This is a novel role of DR3 that is yet to be described. However, this role of DR3 interferes with our analysis of the ligand/receptor relationship and therefore we were unable to confirm that DR3 is the receptor responsible for VEGI-induced apoptosis. We also provide preliminary evidence that VEGI is utilizing an unknown receptor to mediate NF-κB activation. We therefore provide several mechanisms to control VEGI-mediated endothelial cell death, one being the activation of NF-κB to suppress the apoptotic potential of VEGI and the needed presence of DR3 for VEGI to initiate apoptosis, a role that is possibly independent of ligand binding.

The class of proteins known as immunophilins, that are cis-trans prolyl isomerases perform diverse chaperone roles. The immunophilins FKBP52 and FKBP51 (FK506 Binding Proteins) are adapter proteins involved in the trafficking of the glucocorticoid receptor (GR), in which FKBP52 facilitates binding of retrograde molecular motor protein dynein and the GR, while FKBP51 binds only the GR. This body of work presents: 1. An analysis of the FKBP family of proteins and their potential for involvement of neuropathogenesis and identifies FKBP52 and FKBP51 as an evolutionarily divergent duo in mammals, 2. The changes in gene and protein levels of these immunophilins in the frontal cortex of patients with HIV and Major Depressive Disorder (MDD), and 3. A role for FKBP52 in the ligand-activated redistribution of GR in neurons. Using primary human mixed neuron-glia cultures, we tested the hypothesis that immunophilin ligands, like FK506, may alter the kinetics of FKBP52 or FKBP51-mediated trafficking of the GR in neurons. We treated the neuron-glia cultures with cortisol with or without FK506 pretreatment, and found that FK506 altered the distribution of GR. By knocking down expression of FKBP52 using siRNA in a differentiated neuroblastoma cell line, hydrocortisone-mediated nuclear translocation of GR was slowed. Treatment of neuroblastoma cells media supplemented with 10% conditioned media of HIV-infected microglia lead to increased expression of both immunophilins. In a parallel study, we assessed the transcriptional and postranscriptional levels of the GR adapter proteins FKBP52 and FKBP51 in autopsy tissues from the frontal cortex of patients with MDD with and without HIV. We found increased expression of both proteins in HIV infected patients. FKBP51 was increased in MDD while expression of FKBP52 was the highest in the HIV population with MDD. These data support the hypothesis that the immunophilins described here modulate the cellular function of the GR in the brain and expression levels may be related to mood disorders. In general, viral infection and inflammation increase expression, of both immunophilins, which may alter the cortisol-induced trafficking of GR.

Head and neck squamous cell carcinoma (HNSCC) is a commonly occurring malignancy associated with severe morbidity, persistently high mortality rates, frequent recurrence, and the appearance of second primary tumors (SPTs). A great need exists, therefore, for new therapies, including complementary and preventive approaches to treating HNSCC. Signal transducer and activator of transcription (STAT)-3, an oncogenic transcription factor, shows promise as an important therapeutic target in the treatment of HNSCC. The current study focuses on the STAT3-targeting activities of two natural compounds, guggulsterone and honokiol, and investigation of their antitumor activity in HNSCC. Guggulsterone, a compound contained in the resin of the Commiphora mukul plant, used in Indian Ayurvedic medicine, is widely available as a dietary supplement and associated with few side effects. Honokiol is a naturally-occurring compound that has been used in traditional Chinese medicine and is derived from the plant, Magnolia officinalis. Both compounds have been shown to have anticancer activity in various models and to inhibit nuclear factor kappa B (NF kappa B), an oncogenic transcription factor. NF kappa B and STAT3 interact with one another in various ways. Therefore, we hypothesized that guggulsterone and/or honokiol might be useful in targeting STAT3. Both compounds inhibited growth and invasiveness and induced apoptosis in HNSCC cell lines, in addition to decreasing levels of phosphotyrosine STAT3, and, for guggulsterone, total STAT3. Guggulsterone was also found to cause cell cycle arrest and to target hypoxia-inducible factor (HIF)-1 alpha, a potential therapeutic target whose expression is correlated with poor clinical outcome in HNSCC. Guggulsterone-induced growth inhibition relied partly on its ability to inhibit STAT3. Both compounds enhanced the activities of current HNSCC therapies and modestly inhibited tumor growth in the xenograft model of HNSCC. To test the chemopreventive potential of STAT3 and epidermal growth factor receptor (EGFR) inhibition, a study administering Guggulipid, a guggulsterone-containing nutraceutical, or erlotinib, an EGFR-targeting tyrosine kinase inhibitor (TKI) to mice treated orally with a carcinogen is currently underway. Our results so far suggest that guggulsterone and honokiol-mediated inhibition of STAT3 and guggulsterone-mediated inhibition of HIF-1 alpha provide a biologic rationale for further clinical investigation of these compounds as complementary and preventive treatments for HNSCC.

The process of developing a greater understanding of the fundamental molecular mechanisms involved in prostate carcinogenesis will provide insights into the questions that still plague the field of prostate cancer research. The goal of this study was to identify altered genes that may have utility either as biomarkers, for improved diagnostic or prognostic application, or as novel targets important in the pathobiology of prostate cancer. We hypothesize that an improved understanding of the genomic and proteomic alterations associated with prostate cancer will facilitate the identification of novel biomarkers and molecular pathways critical to prostate carcinogenesis. In order to enhance our knowledge of the molecular alterations associated with prostate cancer, our laboratory performed microarray analysis comparing gene expression in healthy normal prostate to that in prostate cancer tissue. Of the greater than 400 genes with significantly altered expression identified in our study, MT2A, Tacc2, Nell2, FosB, PCP4, and Cyr61 were selected for further evaluation to confirm expression changes and evaluate their potential impact in prostate cancer. Analysis of MT2A, Tacc2, and Nell2 expression patterns failed to demonstrate significant changes between prostate cancer and donor prostate tissue and, therefore, these results do not support their further development as prostate cancer biomarkers. We demonstrated that PCP4 was expressed predominently in the stromal compartment of the prostate and was expressed at similar levels in the stroma of normal and prostate cancer tissue. Interestingly, protein expression of PCP4 in a panel of colon cancer tissues was dramatically higher in adenoma and adenocarcinoma tissues compared to donor and benign colon tissue and, consequently, we feel that PCP4 has more potential as a biomarker in colon cancer than in prostate cancer. We also demonstrated that FosB and Cyr61 were upregulated in prostate cancer tissues over donor prostate tissues. Based on expression analysis of FosB and expression and functional analysis of Cyr61, we believe that these two targets have the greatest potental to be functionally significant in the etiology of prostate cancer.

Oxidative stress and tissue remodeling are involved in the development of fibrosis of the lung and heart. Extracellular superoxide dismutase (EC-SOD) is an important antioxidant enzyme that has been shown to limit inflammation and fibrosis. Idiopathic pulmonary fibrosis is a lung disease characterized by severe, progressive interstitial fibrosis. Cardiac fibrosis has various causes but is a fatal side effect of treatment with the chemotherapeutic doxorubicin. Both of these diseases involve oxidant/antioxidant imbalances and can be studied through animal models. It was hypothesized that one mechanism through which EC-SOD protects the lungs and heart from inflammation and fibrosis is by preventing oxidative shedding of extracellular matrix components, specifically syndecans. In the lung, wild type and EC-SOD KO mice were treated with titanium dioxide, asbestos, or Bleomycin. Over the course of injury, EC-SOD KO mice have significantly higher levels of shed syndecan-1 and -4 in their bronchoalveolar lavage fluid and tissue. By IHC staining, the lung distribution of EC-SOD decreases in areas of fibrosis while syndecan-1 increases. Furthermore, in vitro, EC-SOD prevents syndecan-1 shedding from epithelial cells through its antioxidant activity and by directly binding to syndecan-1. Shed syndecan-1 is chemotactic to neutrophils and inhibits wound healing. In the heart, the signifiance of EC-SOD on normal heart morphology, fibrosis, and cardiac function were evaluated in wild type and EC-SOD KO mice in a doxorubicin-induced LV fibrosis model. The lack of EC-SOD causes LV posterior wall thinning and ventricular dilation without an insult. After an oxidative insult induced by doxorubicin, EC-SOD KO mice lost significantly more cardiac function compared to wild-type mice, had enhanced inflammatory cell recruitment, increased shedding of syndecan-1 from the heart and increased caspase-3 activation or apoptosis. In summary, this study shows that EC-SOD is important in the maintenance of normal cardiac morphology, the development of cardiac fibrosis, inflammation, apoptosis, and the loss of function associated with oxidative cardiac injury. In the lung, this investigation shows that the loss of pulmonary EC-SOD leaves syndecan-1 vulnerable to oxidative stress and that oxidant-induced loss of cell surface syndecan-1 impairs re-epithelialization, induces inflammation, and promotes a fibrotic microenvironment in the lung.

Brain aging, the phenomenon by which the passage of chronological time is associated with reduced brain volume, is important because it is regarded as a key component of increased dementia risk. The main purpose of this thesis is to present a model and supporting data that enhances knowledge of the underlying processes that drive brain aging and dementia risk. This has been done using structural and functional MRI scans from the Cardiovascular Health Study-Cognition Study (CHS-CS), a longitudinal community cohort study of elderly individuals that possesses extensive clinical and neuropsychological data. The model defended in this dissertation states that the relationship between older age and lower gray matter volume is not strictly a function of the passage of chronological time. Rather, older age is correlated with vascular diseases that themselves are a driving force behind brain aging. Most importantly, the three entities of aging, vascular disease, and neurodegeneration overlap in key strategic areas of the brain. Thus, a large factor behind dementia risk is that there are common brain areas that serve as sites of synergy by which age, vascular disease, and neurodegeneration can summate and thus amplify the risk for cognitive impairment and dementia. Using a whole brain method for analyzing structural MRI scans, we have found that the age and vascular disease, as reflected by white matter hyperintense lesions (WMHL) jointly affect areas of the brain known to be targeted by age and neurodegeneration and that they interact in these key strategic brain regions. Finally, results from perfusion MRI imaging will be reported showing that hypertension as the main predictor of lower regional cerebral blood flow. Taken together, these data will be interpreted to support the following model: structural and functional changes in an aging brain are modulated by hypertensive vascular disease. Additionally, the damage exerted by neurodegenerative processes on the brain is also modified by vascular disease. Finally, there are common strategic anatomical sites in which this synergy occurs and they include areas with important cognitive functions such as the hippocampus, cingulate gyrus, and precuneus. This model has several broad implications. First, it suggests that age related brain atrophy and perhaps even neurodegenerative atrophy itself can be reduced in magnitude if underlying vascular diseases are either prevented or better managed. Second, such a reduction in brain aging may lower risk for dementia by providing additional brain reserve. Third, key strategic brain regions provide a basis for further study and therapeutic targets.