Neuroendocrine (NE) prostate cancer develops as an aggressive disease that does not respond to androgen ablation therapy. The paracrine action of NE cells can facilitate progression of androgen dependent adenocarcinoma to an androgen independent state, suggesting a significant role for NE cells during failure of androgen ablation therapy. To investigate the pathways involved in NE differentiation of prostate cancer, we looked at the expression of genes involved in endocrine differentiation of -cells in the pancreas. This study was performed using the NE prostate cancer mouse model (12T-10) and the derivative allograft model (NE-10). Immunohistochemical studies show that neuroendocrine prostate tumors express transcription factors Foxa2, mouse achaete-scute homolog-1 (mash-1), neurogenin3 (Ngn3) and Nkx2.2. These tumors show a loss of hairy/enhancer of split (Hes-1), a gene that inhibits NE differentiation. Human NE prostate cancers also express Foxa2 and human achaete-scute homolog-1 (HASH-1). These genes are expressed in NE prostate tumors in a sequence similar to pancreatic -cell endocrine differentiation. Foxa2 expression is detected in early prostatic intraepithelial neoplasia (PIN). Mash-1 expression is detected in a few clusters within low grade PIN lesions and Nkx2.2 expression is rarely detected in the PIN lesion. Ngn3 and Nkx2.2 frequently appear in the invasive NE cancer. NE metastasis to lung and liver show a distinct gene expression pattern. The lung metastasis expresse Ngn3 but not Nkx2.2 whereas liver metastases express Nkx2.2 but not Ngn3 These results suggest that Ngn3 and Nkx2.2 expression are markers for site-specific metastasis and/or transcriptionally regulated genes required for organ-specific metastasis. This study indicates that a pathway similar to pancreatic -cell differentiation is involved in NE differentiation of prostate cancer. In the second part of this study the significance of Foxa2 expression in the formation of NE tumors was determined by conditionally knocking it out in the NE tumors of the TRAMP mice. The Foxa2 knocked out TRAMP mice could still develop NE prostate tumors that now started expressing mash-1. These results suggest that NE tumors can develop in the absence of Foxa2 and there are at the very least three classes of NE tumors, one that expresses Foxa2, one expresses mash-1 and one that expresses both.

<p>Prostate proliferation is highest during development and tumorigenesis. We postulate that proteins which regulate normal prostate growth also promote prostate cancer (PCa) progression. Two Dimensional Gel Electrophoresis was utilized to compare patterns of protein expression in 12T-7f prostates (LPB-Tag mouse model for PCa) during tumor development and progression with those of normal developing and adult wild type CD-1 prostates. Stathmin expression and phosphorylation patterns were analyzed in mouse and human PCa cell lines as well as in human PCa tissue arrays. Stathmin was identified by two-dimensional gel electrophoresis and mass spectrometry. Stathmin levels increase early during normal mouse prostate development and again during prostate tumor development and progression. In human prostate adenocarcinoma, stathmin increases in Gleason pattern 5. Further, stathmin is differentially phosphorylated in androgen-dependent LNCaP cells compared to androgen-independent PC-3 and DU145 cells. This differential phosphorylation is modulated by androgen and anti-androgen treatment. <p>We have identified stathmin as a regulator of epithelial cell homeostasis. Knocking down stathmin expression using RNAi technology resulted in epithelial-to-mesenchymal transitional cells (EMT) within 24 hours and without TGF-beta treatment. Our results indicate that stathmin blocks p38MAPK phosphorylation, thereby promoting E-cadherin expression and juxta-membrane localization while suppressing vimentin expression. Moreover, blocking phosphorylation of p38MAPK in stathmin siRNA-transfected DU145 or expressing stathmin in stathmin-negative NMuMG cells prevents the emergence of EMT despite TGF-beta1 treatment. <p>In summary, we propose that stathmin functions as a gatekeeper in maintaining a normal epithelial cell phenotype. Stathmin exhibits differential function in normal compared to cancer cells, similar to the duel roles of TGF-beta as a tumor suppressor in normal cells and a tumor promoter at later stages of cancer progression. Stathmin over-expression has been associated with a number of different cancers including leukemia as well as breast, ovarian and prostate cancer. On the other hand, mechanisms, which disrupt stathmin expression, such as activation of TGF-beta1 signaling, result in destabilization of the cytoskeletal framework of the epithelial cell and progression to EMT. Thus, stathmin represents an attractive target for therapeutic intervention aimed at maintaining a normal epithelial phenotype and controlling tumor spread.

Prostaglandin E2 (PGE2), which exerts its functions by binding to four G protein-coupled receptors (EP1-4), is implicated in tumorigenesis. Among the four EP receptors, EP3 is unique in that it exists as alternatively spliced variants, characterized by differences in the cytoplasmic C-terminal tail. Although three EP3 variants á, â and ã have been described in mice, their functional significance in regulating tumorigenesis is unknown. In this study we provide evidence that expressing murine EP3 á, â and ã receptor variants in tumor cells reduces to the same degree their tumorigenic potential in vivo. In addition, activation of each of the three mEP3 variants induces enhanced cell-cell contact and reduces cell proliferation in vitro in a Rho-dependent manner. Finally, we demonstrate that EP3- mediated RhoA activation requires the engagement of the heterotrimeric G protein G12. Thus, our study provides strong evidence that selective activation of each of the three variants of the EP3 receptor suppresses tumor cell function by activating a G12-RhoA pathway.

There is a significant genetic component to prostate cancer (PrCa) risk. Multiple epidemiologic and linkage studies suggest X-linked heritability. Located at chromosome Xq27-28, HPCX has been identified as a PrCa susceptibility locus in multiple independent studies. We hypothesized that HPCX harbors common variants of modest effect size predisposing to risk of PrCa. This dissertation details a comprehensive large-scale haplotype-based investigation within HPCX to test this hypothesis complemented by two prior independent studies. These two complementary studies were designed to test novel methods and to assess the power of our newly ascertained PrCa study population for use in the HPCX investigation. First, we successfully develop our haplotype-based study design using a proven study population investigating a candidate gene of known significance with an undefined causal variant; the study population was the Shanghai Breast Cancer Study, and the breast cancer associated gene was the rate-limiting enzyme of steroid biosynthesis, CYP11A1. Second, through confirmation of a known risk variant on chromosome 8q24, we assess the power of our new PrCa study population, uniquely comprised of cases with a family history of PrCa who were age-matched to controls with no family history of PrCa. Furthermore, in this study we investigate a locus at chromosome 19q12-13 associated with PrCa aggressiveness. This locus harbors transforming growth factor-Ò1 (TGFB1), an ideal candidate gene. We test the hypothesis that the common T29C functional polymorphism within TGFB1 impacts PrCa risk and aggressiveness and report an association specific to indolent disease, the first of its type. Finally, we comprehensively investigate a 352 kb candidate interval within HPCX for association with risk of PrCa in a large-scale genetic association study. We perform de novo SNP discovery and assay non-unique genomic regions disregarded by current high-throughput platforms. We find one variant significantly associated with risk of PrCa in two independent populations, obtained by dividing our PrCa study population into discrete training and test sets. The work presented in this thesis represents part of a worldwide effort to uncover elusive variants predisposing to PrCa risk.

The development of cancer in humans is characterized by the accumulation of genetic alterations that either enhance or diminish activity in signaling pathways mediating cellular growth and proliferation. Over time these alterations gradually transform normal cells into malignant cells with aberrant properties. This process of transformation gives rise to cancerous cells no longer under control by natural growth-regulatory mechanisms within the body. The signaling pathways that govern cell growth and proliferation are mediated in part through the activity of regulatory GTPases, and the R-Ras family of GTPases has been implicated in the promotion of tumorigenesis. Here we investigate the transforming properties of two highly homologous members of the R-Ras family. R-Ras and TC21(R-Ras2) are two Ras-related GTPases with the potential to induce oncogenic transformation in mammalian cells. Despite their similarities, these proteins have distinct functions and differ in their ability to transform cells in vitro and induce tumor formation in vivo. We hypothesized that these differences in tumorigenicity resulted from differential activation of signaling pathways mediating growth and proliferation downstream of R-Ras and TC21. Using mammary epithelial cells we show that TC21 is significantly more transforming than R-Ras, and we demonstrate that distinct signaling events are required for these oncogenes to induce cellular transformation. Transforming growth factor-beta (TGF-â) is a key regulator of cell growth in the body. TGF-â can cooperate with oncogenic members of the Ras superfamily to trigger cell transformation, yet it is unknown whether TGF-â exhibits this type of cooperative behavior with R-Ras and TC21. We investigated this possibility and found R-Ras-transformation to be highly dependent on TGF-â-signaling, while TC21-transformation was less dependent. Through these studies we have generated an in vitro/in vivo model of tumorigenesis that can be used to investigate the molecular events mediating R-Ras and TC21-induced transformation. Importantly, this model will be useful for identifying structural domains that mediate the oncogenic activity of R-Ras proteins, and for testing the efficacy of small molecule inhibitors as potential cancer therapeutics.

Gastric adenocarcinoma is the second leading cause of cancer-related death in the world, and infection with Helicobacter pylori is the strongest known risk factor for this malignancy. Adherence of H. pylori to gastric epithelial cells is critical for induction of gastric injury; therefore, the objective of this dissertation was to investigate the molecular pathways induced by pathogenic H. pylori that lead to epithelial cellular responses with carcinogenic potential. We hypothesized that a host factor that may influence aberrant responses to H. pylori is decay-accelerating factor (DAF), a glycosylphosphatidylinositol (GPI)-anchored protein that primarily functions by dissociating C3/C5 convertases on the host-cell surface. DAF serves as a cellular receptor for several pathogens, and DAF expression is increased within H. pylori-infected human gastric tissue. <p> We show that H. pylori adheres significantly more avidly to cells that stably express human DAF compared to vector alone. In vitro, H. pylori induces DAF up-regulation in gastric epithelial cells, and genetic deficiency of DAF attenuates the development of inflammation in H. pylori-infected mice. Utilizing an in vitro coculture model we demonstrate that DAF is transcriptionally up-regulated by H. pylori, which was attenuated by disruption of a structural component of the cag secretion system (cagE). H. pylori-induced expression of DAF was dependent upon activation of the p38 MAP kinase pathway, but not NF-êB. In vivo, mice infected with wild-type H. pylori demonstrated significantly increased DAF expression in the gastric epithelium versus uninfected controls or mice infected with an H. pylori cagE- isogenic mutant strain. <p> Collectively these studies have defined DAF as a novel receptor for H. pylori that is an important mediator of the inflammatory response in vivo. We also demonstrated that H. pylori cag+ strains induced the up-regulation of this receptor in vitro and in vivo in a cag-dependent manner, representing the first evidence of regulation of an H. pylori host receptor by the cag pathogenicity island. These studies have provided a foundation for further investigation of an H. pylori receptor that has important implications for our understanding of disease progression and clinical outcome.

Colon cancer progression is frequently characterized by activating mutations in Ras and emergence of the tumor promoting effects of TGF-beta signaling. Ras-inducible rat intestinal epithelial cells (RIE:iRas) undergo a well described epithelial to mesenchymal transition (EMT), acquiring an invasive phenotype in response to H-RasV12 expression and TGF-beta treatment, modeling tumor progression. The gene expression profile accompanying Ras and TGF-beta-induced EMT in RIE:iRas cells was characterized by microarray analysis. Ras and TGF-beta cooperatively regulate 329 transcripts during EMT, many of which are of known importance in tumor progression. In particular, several EGF family receptors and ligands, which are shown here to be necessary for TGF-beta-induced EMT in both RIE:iRas and LIM1863 human colorectal cancer cells, were shown to be synergistically regulated by these combined treatments. In addition, enrichment of a class of mRNAs containing 3 AU-rich element (ARE) motifs known to regulate mRNA stability were synergistically regulated by Ras and TGF-beta. Among these, vascular endothelial growth factor (VEGF), a key regulator of tumor angiogenesis, was synergistically regulated in both RIE:iRas cells as well as in an independent cell culture model (YAMC) and was further shown to be regulated by a post-transcriptional mechanism. Expression profiling of human colorectal cancers (CRC) further revealed that many of the ARE-containing genes were differentially expressed in stage 4 human colon adenocarcinomas compared to adenomas. Furthermore, genes differentially expressed in CRC are also significantly enriched with ARE-containing transcripts. These studies demonstrate that oncogenic Ras and TGF-beta globally regulate genes containing AREs in association with EMT by a post-transcriptional mechanism with potential clinical significance.

Colon cancers arise as the consequence of the accumulation of mutations and epigenetic alterations in normal cells that lead to the transformation of these cells into cancer. One way through which these genetic and epigenetic alterations affect the behavior of the cells is by deregulating signaling pathways in the cells. The transforming growth factor beta (TGF-beta) signaling pathway is commonly deregulated in colon cancer. The TGF-beta signaling pathway consists of the secreted ligand, TGF-beta, the heteromeric TGF-beta receptor composed of TGFBR1 and TGFBR2, and the downstream Smad dependent and Smad independent signaling pathways. The TGF-beta signaling pathway regulates proliferation, apoptosis, and differentiation of normal colon epithelium. Mutations in TGFBR2, SMAD4, and SMAD2 have been found commonly in colon cancer and approximately 75% of colon cancers are resistant to TGF-beta. Thus, we sought to determine whether alterations in the activation of the secreted TGF-beta1 ligand or alterations in the TGFBR2 receptor levels could alter the TGF-beta signaling pathways and potentially be another mechanism through which cancers deregulate TGF-beta signaling. We found that the aberrant methylation of thrombospondin 1 (TSP1), which activates secreted TGF-beta1, is an epigenetic mechanism that attenuates the activation of TGF-beta and the activation of the Smad signaling pathway. The identification of methylated TSP1 in colon cancers is the first demonstration of epigenetic alterations promoting colon cancer formation through repressing TGF-beta signaling. Our studies of the effect of TGFBR2 expression levels on TGF-beta signaling pathway activation have shown that the activation of Smad independent pathways is dependent on the receptor level. Using a precisely regulated inducible in vitro system, we observed differential activation of the PI3K-AKT and MAPK pathways that varied depending on TGFBR2 expression levels. Furthermore, TGF-beta mediated induction of p21 also varied depending on TGFBR2 expression levels. These results suggest that modulation of TGFBR2 levels can alter the activation of Smad independent signaling pathways and is a potential mechanism for altering the genes that are induced by TGF-beta. Thus, we have determined that the TGF-beta signaling pathway can be altered through suppression of the TGFBR2 and through suppression of mechanisms that affect TGF-beta ligand activation.

Matrix metalloproteinases (MMPs) are classically associated with late stage metastases, though previous work by our lab and others have expanded the role of MMPs to all stages of tumor development. Using a genetic model of intestinal tumorigenesis, the Min (multiple intestinal neoplasia) mouse, we previously demonstrated a role for MMPs, in particular MMP-7, in the development of intestinal adenomas. To further explore the role of proteinases in the development of intestinal neoplasia, I created a novel microarray to examine more than 500 proteases for differential expression in intestinal tumors. Relative microarray analysis found that MMPs-10, -13, and -14 have the highest fold change in expression in tumor compared to normal samples, while absolute microarray analysis indicated that MMPs-9, -12, -15 and -19 are present in tumor samples but not normal intestine. As a result of this screen, I generated Min mice genetically deficient for various proteinases that were differentially detected in Min adenomas (MMP-2, -9, -12, and -19). Genetic ablation of MMP-2, -12, or -19 did not affect tumor number or size, however, MMP-9 deficient Min mice developed 25% fewer tumors than littermate controls. Further, in the context of intestinal tumors, the major cellular source of MMP-9 is neutrophils thus suggesting a pro-tumorigenic role of inflammation in early tumorigenesis. Additionally, our screen detected several mast cell produced proteinases in tumor tissue that were absent from normal intestine. To examine the effect of mast cell proteinases in tumorigenesis, I generated mast cell-deficient Min mice. Surprisingly, these mice developed 50% more tumors than mast cell Min mice, thus indicating that mast cells function in an anti-tumorigenic role early during intestinal tumorigenesis. Thus, these studies demonstrate that proteinases and infiltrating inflammatory cells function in both pro- and anti-tumorigenic roles during early intestinal tumor development.

We have identified a novel signaling pathway, AKT/IKKα/VAV1, which induces endothelial cell survival and motility. Ang1/Tie2 and VEGF/VEGFR signaling activates Akt and induces cell survival. Akt and IKKÑ both induce endothelial EMT and endothelial cell motility. This can be blocked with co-expression of IÛB-Ñ, suggesting that endothelial EMT is mediated through the NF-ÛB canonical pathway. Vav1 and Ô-catenin are both upregulated by IKKÑ. Vav1 is required for induction of endothelial EMT, and induces endothelial motility and tumor angiogenesis. Ô-catenin also induces endothelial cell motility and tumor angiogenesis, through regulation of RhoA and Cdc42 activity. The proposed model advances the study of angiogenesis. We demonstrate a novel mechanism for endothelial cell survival and endothelial cell motility. More importantly, we show that the Akt/IKKÑ/Vav1 signaling pathway can induce endothelial EMT and that this process plays a role in tumor angiogenesis.