The emergence of drug-resistance in cancer cells during chemotherapy remains a major obstacle in the treatment of neoplasia. Multidrug resistance (MDR) to a group of unrelated cytotoxic compounds can be conferred to eucaryotic cells by the expression of P-glycoprotein (Pgp), a putative plasma membrane transporter believed to mediate the efflux of these agents out of cells. A variety of agents are able to reverse this MDR phenotype by inhibiting the Pgp transporter. Blocking the action of this protein increases the effectiveness of cancer chemotherapeutic agents and, hence, has significant clinical implications. A mutant Pgp1 cDNA containing the substitution (Gly{dollar}\sp{lcub}338{rcub}{dollar}Ala{dollar}\sp{lcub}339{rcub}{dollar} to Ala{dollar}\sp{lcub}338{rcub}{dollar}Pro{dollar}\sp{lcub}339{rcub}{dollar}) within the sixth transmembrane domain (tm6) has been cloned. The expression of this mutant confers an MDR phenotype preferentially resistant to actinomycin D. In this thesis we report that this MDR phenotype also has a decreased sensitivity toward reversal by cyclosporin A, while the sensitivity toward verapamil is unaltered. The accumulation of {dollar}\rm\lbrack\sp3H\rbrack{dollar} vincristine in cells expressing the wild-type Pgp1, not the mutant, increases dramatically in the presence of cyclosporin A, which correlates well with the reversal profile. We have altered only one amino acid residue at this location (Gly{dollar}\sp{lcub}338{rcub}{dollar} to Ala{dollar}\sp{lcub}338{rcub}{dollar} or Ala{dollar}\sp{lcub}339{rcub}{dollar} to Pro{dollar}\sp{lcub}339{rcub}{dollar}). The transfectants expressing the Pgp1 containing the proline substitution, rather than the alanine, demonstrate an MDR phenotype which is preferentially resistant to actinomycin D, and insensitive to reversal by cyclosporin A. Modeling the whole tm6 domain (with the Quanta modeling program and energy minimization by the CHARMm program) reveals that the proline substitution at position 339 rather than the alanine at 338 drastically changes the local {dollar}\alpha{dollar}-helice conformation, especially the polar side chain alignment along the hydrophilic side of this amphipathic {dollar}\alpha{dollar}-helice. We hypothesize that the Ala{dollar}\sp{lcub}339{rcub}{dollar} to Pro{dollar}\sp{lcub}339{rcub}{dollar} substitution, rather than the Gly{dollar}\sp{lcub}338{rcub}{dollar} to Ala{dollar}\sp{lcub}338{rcub}{dollar}, is the primary contributor to the aforementioned altered phenotype. We have also attempted to determine the functions of two spliced variants of Pgp1, previously cloned in this laboratory, by expressing them in an in vitro system. The biogenesis of one of the variants, ADX124, has also been investigated. We conclude that it is derived from a splicing event that involves the internal splicing signals that are maintained in the mature full-length Pgp1 transcript.

The Wnt-1 proto-oncogene is one of the most frequently activated genes in mouse mammary tumor virus (MMTV)-derived breast tumors. Wnt-1 is a member of a family of genes that display complex patterns of expression during embryogenesis and have been shown to be required for central nervous system, kidney, limb, and placenta development. Several Wnt genes are expressed in the developing mouse mammary gland. The Wnt-1 product is a secreted glycoprotein that has been implicated in cell signaling. The identification of the Drosophila Wnt-1 homologue, wingless (wg ), a segment polarity gene, has enabled the components of the Wnt/wg intracellular pathway to be identified. To date, dishevelled (dsh) has been identified as the most upstream component of wg signaling subsequent to receptor activation and encodes a cytoplasmic protein required for interpretation of the wg signal. Three mammalian homologues of dsh, Dvl1, Dvl2 and Dvl3 , have been identified and characterized. Each exhibits expression patterns in embryonic and adult mice that are consistent with a role in Wnt signal transduction.;Based on the observation that Wnt-1 expressed ectopically in the mammary glands of transgenic mice causes extensive hyperplasia and mammary adenocarcinomas, we have generated three lines of FVB/N mice that harbor an MMTV LTR Dvl2-myc expression construct in their genome. Concomitantly, we have generated an additional transgenic FVB/N line with a putative dominant-negative MMTV Dvl2-myc construct (MMTVDeltaPDZ Dvl2-myc). We anticipated that replication of the Wnt-1 mammary phenotype in the MMTV Dvl2-myc animals, or, alternatively, modification of the phenotype in double transgenic MMTV Wnt-1/MMTVDeltaPDZDvl2-myc mice, would confirm a role for murine Dvl in Wnt signaling. The resultant MMTV Dvl2-myc and MMTV DeltaPDZ Dvl2-myc transgenic mice are viable, fertile, and exhibit no lactation defects. Western blot and RT-PCR analyses have confirmed that heterozygous transgenic mice express the Dvl2-myc-encoded protein and mRNA in their mammary and salivary glands; we have been unable to demonstrate expression in the DeltaPDZDvl2-myc line. Gross and histological examinations of glands isolated from virgin, pregnant and multiparous MMTV Dvl2-myc females have revealed no significant differences between these tissues and those harvested from wild-type littermates. A single mammary lipoma and two neuroectodermal tumors involving the uterus have been detected in aged multiparous MMTV Dvl2-myc females; no transgene expression has been associated with the uterine tumors, however, and all have been considered sporadic events. While we have been unable to document changes in the mammary gland upon ectopic expression of our transgene, the lack of a phenotype does not exclude a role for Dvl2 in normal mammary development and Wnt-mediated tumorigenesis.

The genes CDKN2 (MTS1, CDK4I, p16{dollar}\rm\sp{lcub}INK4{rcub}){dollar} and CDKN2B (MTS2, CDK4Ib, p15{dollar}\rm\sp{lcub}INK4{rcub}){dollar} encoding the proteins p16 and p15 respectively, are both located on chromosomal band 9p21, a locus at which frequent homozygous and heterozygous deletions occur in many primary human tumors, including esophageal carcinoma. CDKN2 and CDKN2B belong to a family of cyclin-dependent kinase 4 inhibitors (INK4) and control cell proliferation by preventing entry into the S phase of the cell cycle. Their inactivation may contribute to uncontrolled growth in human cancers. To investigate whether CDKN2 and CDKN2B are involved in esophageal tumorigenesis, intragenic mutations of CDKN2 and CDKN2B in primary esophageal cancers were examined. Sixty primary esophageal cancers were analyzed for mutations in both exons 1 and 2 of CDKN2 and CDKN2B by direct sequencing of PCR-amplified genomic DNAs. Six sequence alterations of CDKN2 were observed in five squamous cell carcinomas and in one adenocarcinoma. All six nucleotide changes resulted in marked alterations in amino acid sequence. Four were nonsense mutations leading to premature termination codons; nucleotide substitutions identical to two of these stop codons were previously reported in other tumor types. One CDKN2B nonsense mutation which truncates the protein product by 75% and one silent CDKN2B mutation occurred somatically. In order to determine the tumor suppressive role of these genes in vitro, homozygous deletion, intragenic mutation, and messenger RNA (mRNA) expression of CDKN2 and CDKN2B in nine esophageal squamous cancer cell lines were studied. To establish whether CDKN2 and CDKN2B exert growth suppressive effect on esophageal cancer cell lines, CDKN2, CDKN2B, and a combination of these two genes were transfected into esophageal cancer cells that were homozygously deleted for CDKN2, CDKN2B or both genes. Polymerase chain reaction (PCR) amplification revealed that five of nine cell lines (55%) manifested homozygous deletions of CDKN2, CDKN2B, and/or flanking loci on chromosomal band 9p21. Reverse transcriptase-PCR (RT-PCR) was used to examine CDKN2 and CDKN2B mRNA in the nine cell lines. Lack of CDKN2 and CDKN2B mRNA correlated perfectly with homozygous deletion involving these genes. No subtle intragenic mutations of CDKN2 and CDKN2B were detected by DNA sequencing of their entire coding sequences in any cell lines lacking homozygous deletion. Two of the cell lines manifested homozygous deletions excluding CDKN2; one of these two deletions also excluded CDKN2B. Restoration of CDKN2 and/or CDKN2B into the cancer cell lines demonstrated marked growth suppression by CDKN2 and/or CDKN2B. Taken together, the data obtained about CDKN2 and CDKN2B in primary tumors and cell lines suggest that both genes are involved in the malignant phenotype in esophageal cells. Inhibition of cancer cell growth or proliferation by expression of transfected CDKN2 and CDKN2B is functional evidence that CDKN2 and CDKN2B are tumor suppressors in esophageal epithelium. The relative varity of point mutations and frequency of deletions suggest that homozygous deletion is the predominant mechanism of inactivation of CDKN2 and CDKN2B.

Our laboratory have designed and synthesized a number of cytochrome P450 17alpha-hydroxylase/C17,20-lyase (P450c17) inhibitors with the aim of inhibiting androgen synthesis and as potential agents to treat prostate cancer. VN/63-1, VN/85-1, VN/87-1, VN/107-1, VN/108-1, and VN/109-1 inhibitor were chosen for study. To select the most potent inhibitors, a new method for evaluating inhibitors of the P450c17 enzyme was designed. Bacteria expressing the P450c17 were used as a source of the enzyme. HPLC was replaced by a rapid radiometric assay, acetic acid releasing assay. This new analytical system will be a powerful tool to accelerate the process of new compound selection. Utilizing this system, I showed that VN/85-1, VN/87-1, and VN/108-1 are the most potent inhibitors of P450c17. The effects of the new inhibitors on androgen dependent LNCaP and androgen independent PC-3 prostatic carcinoma cells proliferation were studied. I have shown that VN/85-1, VN/87-1, and VN/108-1 are the most potent inhibitors of LNCaP cell growth in vitro . VN/85-1 also has weak inhibitory effects against androgen independent PC-3 cells. The effects of azolyl steroids on LNCaP androgen receptor (AR) were determined. VN/63-1, VN/85-1, and VN/87-1 were full antagonists of LNCaP AR. Although, VN/107-1 , VN/108-1, and VN/109-1 were partial agonists of LNCaP AR, all of the evaluated inhibitors were antiandrogens against the wild-type AR. The new inhibitors were also evaluated in vivo. The male SCID mice bearing LNCaP tumor xenografts were utilized for new compounds evaluation. I showed that VN/85-1 and VN/87-1 have potent antitumor effects in this in vivo model. The ability of pregnenolone to stimulate LNCaP cell proliferation was discovered and investigated. I showed that this pregnenolone stimulation is mediated through the mutated AR and can be inhibited by our novel compounds, especially by VN/63-1, VN/85-1, and VN/87-1. This findings suggests that pregnenolone has a significant role in stimulating prostate cancer and that our new compounds by inhibiting not only androgens but also pregnenolone action may be more effective for treating prostate cancer than currently available agents.

Overexpression of erbB-2 is associated with a poor prognosis in lung cancer (Kern, 1990). We have found that overexpression of erbB-2 alone in a human bronchial epithelial cell model system is not sufficient for tumor production (Hamburger, 1998). Characterization of our cell model system revealed the presence of constitutive production of the erbB family of ligands: TGF-alpha, amphiregulin, betacellulin, heparin-binding EGF and heregulin. Of these endogenously-expressed ligands, only TGF-alpha expressed and secreted at high levels in the presence of overexpressed erbB-2 was found to correlate with tumorigenicity. Of the four constitutively expressed erbB receptors, erbB-1 and erbB-2 were the major species, while erbB-3 and erbB-4 represented a minor fraction of the population. Endogenously produced TGF-alpha was found to induce formation of erbB-1/erbB-2, erbB-1/erbB-3, erbB-1/erbB-4, erbB-2/erbB-4, erbB-2/erbB-3 and erbB-3/erbB-4 heterodimer complexes. However, only the TGF-alpha-induced erbB-1/erbB-2 heterodimer was formed in tumorigenic E6TM cells but not in non-tumorigenic E6TMA cells (Hamburger, 1998). Activation of the PI-3-kinase signaling pathway was comparable in both cell lines, however, activation of the ras-MAPK signaling pathway, however, was found to correlate with formation of the TGF-alpha induced erbB-1/erbB-2 heterodimer complex and tumorigenicity. We demonstrated that both Stats 1 and 3 co-immunoprecipitate with the constitutively produced erbB-1/erbB-2 heterodimer complex in tumorigenic E6TM cells. Activation of Stats 1 and 3 was shown to be dependent on high levels of TGF-alpha. We demonstrated that the erbS-1/erbB-2 heterodimer and erbB-2 kinase activity was required for Stat activation, while the erbB-1 homodimer was not sufficient for the TGF-alpha-induced activation of Stats 1 and 3. Our results support the hypothesis that the erbB-1/erbB-2 heterodimer, formed and activated in response to TGF-alpha, is critical to the tumorigenic conversion of these lung epithelial cells through the activation of signaling cascades including the MAPK and Stat pathways, but not PI-3-kinase.

Growth factor receptors of the tyrosine kinase subfamily I, ECFR/ErbB receptor family members, EGF receptor, erbB2, erbB3, and erbB4, mediate specific biological signals via heterodimerization among family members after ligand stimulation. The overall aim of this study was to examine erbB family member interaction mediating the biological effect of Heregulin (HRG), an erbB3 ligand. C-erbB2, an orphan receptor, was involved in the concentration-dependent pleiotropic responses of a breast cancer cell line to HRG, via heterodimerization with erbB3. HRG-mediated proliferation was maintained in erbB2-nonexpressing, antisense-erbB2 stable transfectants. In contrast, c-erbB2 was required for induction of HRC-mediated differentiation at high concentrations of HRG. To further study proteins that modulate the actions of erbB3, a receptor for HRG with an impaired tyrosine kinase activity, we isolated interacting proteins using a yeast two hybrid system screening. EBP1 and EBP2 have been identified as erbB3 interacting proteins using the yeast two hybrid system. EBP1 and EBP2 interacted with the juxtamembrane domain of erbB3 in the absence of ligand activation or intrinsic tyrosine kinase activity, in vitro. EBP1 and EBP2 mRNA was expressed in normal human epithelial tissues including brain, heart, and skeletal muscle, and in human carcinomas including breast, lung, and sarcoma. Overexpression of EBP1 or EBP2 suppressed colony growth and induced differentiation in a human breast carcinoma cell line, AU565. Treatment with HRG, but not with EGF, caused dissociation of EBP1 from erbB3 in vivo. EBP1 translocated from the cytoplasm into the nucleus following HRG stimulation. These findings suggest that HRG may regulate erbB3 function in part by directly linking receptor activation to nuclear activation through the translocation of an erbB3 interacting proteins. C-erbB2 was also partly involved in 17beta-estradiol-HRG induced cell growth inhibition of ER-negative erbB2-expressing breast carcinoma cell lines.

Murine polyomavirus is a DNA tumor virus that transforms cells in culture and induces tumors in mice. The major viral oncoproteins responsible for these phenomena are large T (LT) and middle T (MT) antigen. This thesis examines the ability of the T antigens to abrogate specific cell cycle check-points, which is a fundamental process in neoplastic transformation. We examined the ability of the T antigens to overcome the growth suppression function of p53. As a model system, a temperature sensitive p53 gene was introduced into mouse embryo fibroblasts from a p53 null mouse. At the permissive temperature (32C) a functionally wild-type p53 protein is expressed and arrests the growth of these cells in G1/G0 phase of the cell cycle. We introduced LT and MT antigen into these cells and demonstrated that LT overcame the growth arrest while MT did not. A LT mutant that was defective for binding the retinoblastoma tumor suppressor protein (pRB), was unable to overcome the growth arrest. This suggests that p53 directs the growth arrest through regulation of pRB and inactivation of pRB overcomes the arrest. Although MT did not overcome the arrest, mitogenic signals stimulated by MT were not inhibited by p53. We show that MT still associated with the signaling molecules PI3K and SHC, and increased the transcriptional activity of c-myc and AP1 in the p53 arrested cells. These results suggest that the growth arrest mediated by p53 is dominant over the proliferative signals generated by MT. Serum deprivation of NIH 3T3 cells results in growth arrest in G 0. Expression of MT in serum deprived cells was sufficient to reinitiate cell cycle progression. Following MT expression, levels of cyclin D and E increased and the level of p27 decreased resulting in increased G1 cyclin dependent kinase activity and hyperphosphorylation of pRB. These results indicate that MT abrogates the mitogen dependent check-point in G0. These studies demonstrate that LT overcomes the growth suppressive effects of p53 and MT substitutes for serum growth factors and overcomes growth arrest in G0. We conclude that the ability of the T antigens to overcome these distinct check-points is critical for cell transformation and presumably is also important for virus growth.

During carcinogenesis, the ability to control cell growth or death is lost. Transforming growth factor beta-1 (TGFbeta1) inhibits epithelial cell growth and plays a role in multistage carcinogenesis. To better understand TGFbeta1 in carcinogenesis, we have studied TGFbeta1 effects on cell shape and its mediators, which are critical to cell growth. We have also characterized a cell model appropriate for our TGFbeta 1 studies. The NRK-52E rat kidney epithelial cell line and two H-ras oncogene-transfected NRK-52E cell lines, H/1.2 and H/6.1, were characterized. NRK-52E cells were differentiated, anchorage-dependent, non-tumorigenic, and contact-inhibited. H/1.2 and H/6.1 cells had increased growth rates, increased plasma membrane ruffling, altered cytoskeletal structure, anchorage-independent growth, and formed renal cell carcinoma-like tumors in vivo. H/6.1 cells formed more aggressive tumors in vivo than H/1.2, and were selected for further studies. Since Ras GTPases regulate cytoskeletal remodeling and we demonstrated H-ras oncogene-induced morphologic alterations, our cell model was validated for investigating the effects of TGFbeta1 on normal and tumorigenic cells. NRK-52E and H/6.1 cells had similar TGFbeta1 receptor expression and growth inhibition by TGFbeta1. NRK-52E treated with TGFbeta 1 showed cell flattening, actin protein expression, stress fiber formation, and increased vinculin fluorescence and expression. TGFbeta1 effects on NRK-52E cells induced talin fluorescence but not its expression. Biochemical analysis of F- and G-actin in NRK-52E cells showed significant TGFbeta1-induced increases. All of these effects were delayed, diminished, or absent in H1/6.1 cells. Since Rho GTPases are known to mediate stress fiber formation and TGFbeta1-induced transcription, we microinjected C3 transferase, a Rho inhibitor. C3 transferase depolymerized stress fibers in control cells but not in TGFbeta1-treated cells. Very high levels of C3 transferase eventually depolymerized all stress fibers in both treated and control cells. These results indicate that TGFbeta1 signaling in renal epithelial cells involves alterations in actin and focal adhesion proteins that differ between normal and H-ras-transformed cells. We suggest that TGFbeta1 modifies aspects of Rho GTPase signal transduction. Elucidating the mechanism by which TGFbeta1 induces cytoskeletal remodeling in tumor cells will reveal potential targets for cancer therapy.

A complete understanding of mammary cell differentiation is required in order to fully understand the molecular basis of breast cancer. Members of the ErbB receptor tyrosine kinase family have been implicated in the etiology and progression of many types of human cancers, including breast cancer. Heregulin (HRG), a ligand for the ErbB3 and ErbB4 receptors, plays a key role in mammary gland development. During pregnancy, HRG-mediated ErbB receptor signaling is responsible for the formation of lobuloalveoli followed by terminal mammary cell differentiation. In some mammary carcinoma cell lines, such as the AU565 cell line, HRG has been shown to induce cellular differentiation. In order to further characterize HRG-ErbB3 signaling, our laboratory previously used a yeast two-hybrid interaction screen to isolate ErbB3 binding proteins. EBP1(ErbB3 Binding Protein 1) was identified and has been demonstrated to dissociate from the ErbB3 receptor upon HRG stimulation. We further characterized this signaling protein by generating a polyclonal antibody against GST-EBP1. Protein and mRNA expression profiles were examined in a variety of cell lines. EBP1 expression was ubiquituous, and not limited to ErbB3-expressing cell lines. We also determined the biological effects of transfecting ebp1 cDNA into breast carcinoma cells. The EBP1 protein is basally phosphorylated in vivo on serine and threonine residues and this phosphorylation is enhanced after thirty minutes of heregulin stimulation. Both serine and threonine residues of GST-EBP1 fusion protein are phosphorylated by PKC in vitro. In vivo, we demonstrated that whereas HRG-induced phosphorylation of EBP1 occurs predominantly in a PKC-independent manner, basal EBP1 phosphorylation is dependent upon the activation of PKC. In conclusion, sustained activation of the ERK signaling pathway is essential for HRG-induced differentiation of the AU565 mammary carcinoma cell line. The ErbB3 binding phospho-protein EBP1 has also been implicated as a potential signaling molecule involved in this process. (Abstract shortened by UMI.)

The recently discovered candidate tumor suppressor gene located at chromosome 10q23.3, designated PTEN or MMAC1, was previously found to be mutated in a small percentage of several types of cancer. A high rate of loss of heterozygosity (LOH) has been reported at chromosome 10q23--26 in endometrial tumors. We performed mutational analysis of PTEN gene in genomic DNA samples from endometrial, gastric, colorectal, and pancreatic cancers. Results revealed that PTEN was mutated in 21 (55%) of 38 endometrial carcinomas: 14/18 (78%) of microsatellite instability-positive (MI+) and 7/20 (35%) of MI- cases. Eighteen (86%) of 21 tumors with mutations are predicted to produce truncated protein products. Two different mutations were present in each of five cases, while loss of heterozygosity (LOH) was observed in additional 8 tumors possessing only one mutation. No mutations were found in 9 pancreatic or 24 colorectal cancers, even in six colorectal cancers that were 10q-LOH-positive. Only one of 14 gastric cancers had a mutation. Six (27%) of 22 tumors with mutation had either insertions or deletions involving a poly(A6) tract in exon 8, suggesting a possible mutational hot spot in this gene. These results suggest that PTEN is an important tumor suppressor gene for endometrial cancers, particularly with those that are microsatellite instability positive (MI+). PTEN may also constitute a coding region target of microsatellite instability. Tumor suppressor PTEN is a dual-specificity phosphatase that is mutated in multiple advanced, metastatic human cancers. In order to discover potential mechanisms underlying suppression of cancer cell invasiveness and metastasis by PTEN, endogenously PTEN-mutant tumor cell lines were transfected with wild-type PTEN cDNA, invasiveness was analyzed using a modified invasion chamber assay, and several candidate targets of PTEN were evaluated. Exogenous wild-type PTEN reduced tyrosine phosphorylation of catenins, phosphatidylinositol 3-kinase (PI3-K) and focal adhesion kinase (FAK), and simultaneously inhibited the invasiveness of PTEN-mutant tumor cells. Wild-type PTEN also diminished both tyrosine and serine/threonine phosphorylation in c-Src and suppressed its kinase activity. These results suggest that c-Src is an important functional target for the phosphatase activity of PTEN which may be related to suppression of cancer cell invasiveness and metastasis by PTEN.