A common mechanism used by cancer cells to over-ride normal restraints on cellular proliferation is abrogation of the tumour suppressive functions of the INK4a/ARF locus. This can be achieved either through genetic changes to the locus or by dysregulation of the molecular pathways that operate to mediate its function or transcriptional regulation. The INK4a/ARF locus encodes two structurally unrelated proteins which have a common exon translated in alternative reading frames. These proteins are named p16INK4a and p14ARF and they both have antiproliferative effects mediated by the Rb and p53 pathways respectively. In this thesis, proteomic approaches have been used to map out some of the protein-protein interaction networks that are involved in function or regulation of INK4a/ARF. Multiprotein complexes are involved both in function (D cyclin-Cdk) and regulation (Cbx7) of the locus and a major focus of this work has been the determination of their molecular composition. The main findings of this thesis relate to the protein-protein interactions of Cbx7, a known transcriptional repressor of INK4A/ARF, which has been implicated as an oncogene. In an analogous manner to other members of the Polycomb group (PcG) of proteins, it was found to participate in a large multiprotein complex. Constituents of a Cbx7 complex isolated from human cells were identified by mass spectrometry. Analysis revealed that it was made up of a subset of the known human PcG proteins and some novel interacting proteins, including an RNA helicase, which had not previously been reported in PcG complexes. The specificity of these interactions was then validated by other biochemical methods. The impact of some of these interactions on the repressive function of Cbx7 has been evaluated in primary human fibroblasts, with a view to understanding how the complex silences transcription.

The CDKN2a locus encodes two important tumour suppressors, pl61NK4a and ARF. The two genes share a common exon which is translated in different reading frames. pl6,NK4a binds to CDK4 and CDK6, preventing them from forming active complexes with D cyclins. As a result, pRb does not undergo the phosphorylation necessary for the transition from the G1 to S phase of the cell cycle. ARF inhibits the ubiquitination of p53 by MDM2, thereby causing the accumulation of p53. There is a growing awareness that the CDKN2a locus plays a central role in the cellular defences against transformation, and in the cellular response to stress. For example, pl6INK4a is involved in senescence, a permanent cell cycle arrest triggered in primary human fibroblasts in response to many stresses, including the overexpression of oncogenes. However, little is known about the regulation of pl6INK4a under these circumstances, and work in this thesis investigates this issue using overexpression of Myc as a model. The thesis also describes the characterisation of human diploid fibroblasts (Milan cells) from a patient homozygous for the R24P mutation of pl6INK4a. As this mutation is in exon la, ARF is unaffected. The mutant pl6INK4a cannot bind to CDK4, but retains some capacity to bind to CDK6. Milan cells have also been used in combination with shRNA targeting ARF to investigate the relative roles of pl6INK4a and ARF in the prevention of transformation. A panel of Milan cells were produced expressing telomerase, with combinations of Myc, Ras and shRNA targeting ARF, and the ability of the cells to grow in soft agar was assessed. A similar panel of Milan expressing p53 shRNA was also built up. These cells were used to investigate whether ablation of ARF can substitute for the loss of p53 function often associated with transformation, and to help identify which aspects of the p53 pathway are activated in the defence against transformation.

An increasing body of evidence highlights inflammation as playing a relevant contribution in the establishment and progression of several diseases including cancer. Indeed environmental factors can determine in different organs a chronic inflammatory status, which in turn may be responsible for increased incidence of cancer. A clear relationship between smoking habits, consequent pro-inflammatory cytokines production, and lung cancer has been extensively demonstrated; a strikingly correlation between western diet consumption leading to obesity-dependent inflammation and colon cancer is becoming evident. Typical inflammation-related diseases of the gastrointestinal (GI) tract are inflammatory bowel diseases (IBOs), comprising Chron's Disease (CD) and Ulcerative Colitis (UC). Even though IBOs affect millions of people all over the world, their etiology is poorly understood yet. Furthermore particular effort is needed to treat and care for IBOs patients as they show an increased risk of developing tumours after 40 years from 10.8% (Ekbom, Helmick et al. 1990) to 25%-40% (Rutter, Saunders et al. 2006) compared to the unaffected population. Secreted protein acidic and rich in cysteine (SPARC) is involved in tissue remodelling, repair and organogenesis; it has been shown to have roles in regulating extracellular matrix composition (Lane and Sage 1994; Van and Sage 1999), tissue permeability, endothelial cell shape and barrier function (Goldblum, Ding et al. 1994) and enterocytes migration (Sansom, Mansergh et al. 2007), thus suggesting it could also play a role in maintaining or regenerating intestinal mucosa under pathogenic conditions, an aspect not investigated yet. The colonic mucosa of SPARC KO mice has a less robust stroma component, thick basement membranes and collagen network; although in physiological conditions these structural features do not cause any intestinal abnormalities they could instead contribute to a major predisposition to crypt damage in in vivo models of UC. By applying the Dextran Sodium Sulfate (DSS)-induced model of experimental colitis in SPARC KO and WT mice, the aim of my work is to reveal a potential role in IBDs (namely UCs) for extracellular matrix proteins and characterize the molecular mechanisms behind the observed phenotype. Colonic inflammation has been triggered in SPARC KO and WT mice by the administration in drinking water of the irritant DSS at two different concentrations (i.e. 2.5% and 3.5%) for 2 weeks. Samples have been collected either immediately (acute phase) or 10 days after treatment suspension (restitution phase) and characterized by mean of immunohistochemistry, RT-PCR, flow cytometry and western blotting analysis. Reciprocal bone marrow (BM) chimeras have been generated (i.e. WT BM > SPARC KO recipients and SPARC KO BM > WT recipients) and both groups treated for 1 week at 3.5% DSS. Mice have been analyzed either immediately or 14 days after treatment withdrawall, while clinical parameters were monitored along all the experimentation. DSS induced a more severe colitis-associated phenotype in SPARC KO mice compared to WT ones both at acute and restitution time points. Interestingly although under the milder experimental condition used (2.5% DSS), in both SPARC KO and WT mice the same level of tissue damage has been observed, still the former showed delayed recovery after treatment withdrawal. RT-PCR analysis revealed a different modulation of the TGF-I3/SMAD pathway in SPARC KO mice that, consequently, affected T regulatory cell (Tregs) levels. BM chimera experiments on the other hand pointed out the importance of stromal cells produced SPARC compared to that produced by epithelial cells (ECs) in determining signs of colitis. These findings highlight a new emerging role of extracellular matrix proteins like SPARC in integrating signals from wounded colonic epithelial tissues to gut immune cells.

There are currently a few bioinformatics tools, such as dbEST, DDD and GEPIS to name a few, which have been widely used to retrieve and analyse EST data for gene expression levels. The Cancer Genome Anatomy Project (CGAP, run by NCBI) cDNA xProfiler and eDNA DGED tools can be used to examine EST to compare gene expression levels between cancer and normal tissue. However, neither COAP nor other similar tools provide an easy way to compare expression in normal and cancerous tissue with e.g. expression levels in related or proximal tissues at the same time while also presenting that data for study separately. Furthermore, the expression data are often assumed to be correct and no quality control tools are made available at eGAP, dbEST and GEPIS. In this study the CGAP tools were recreated with the aim of enabling a wider range of tissues to be searched and compared in a single search. The CGAP tools were found to contain many errors in their library and gene parsing algorithms, for which solutions were implemented in the recreated algorithms. A method was also devised for the tissue origin of EST libraries to be verified and for the uncharacterised libraries to be annotated with a likely tissue of origin using EST data alone. An initial list of tissue-specific genes was optimised to create gene expression matrices which could be used to determine the tissue origin of a library. The matrices were demonstrated to show potential for cancer staging and for the indication of the degree of normalisation of a library in addition to tissue typing, making tissue-specific expression a suitable quality control method for expression data. Together the improved expression profiling algorithm and the expression matrices provide new tools to assess the quality of EST data and their suitability for expression profiling.

Gene expression data form a rich source of information for elucidating the biological function of cellular systems on the pathway level. For this reason, various pathwaybased methods have been developed for analyzing gene expression data from highthroughput experiments. However, in order to utilize the full potential the data can offer, e.g., for cancer research, a more thorough understanding of such methods is required. This thesis consists of two major parts, which contain the results. In the first results part of the thesis we investigate the statistical characteristics of five competitive gene set methods. One major finding shows that three of these five methods, namely, GSEA, GSEArot and GAGE, are negatively influenced by the number of background genes, and, hence, the filtering of the data, in the sense that these methods become more sensitive for expression changes despite the fact that the number of samples remains constant This counter intuitive behavior leads to principle problems for the application of these methods to biological data making the results from these methods no longer reconcilable with the principles of statistical inference rendering the obtained results in the worst case inexpressive. In order to avoid these problems, we suggest an experimental design that helps preventing such issues. Further, we present a new assessment method that allows a power analysis of competitive but also self-contained gene set methods. More precisely, due to the general lack of a sufficient sample size in real data sets, simulated expression data are required in order to investigate statistical methods thoroughly. However, the simulation of pathway-based methods is challenging due to the presence of nontrivial correlation structures within pathways the simulations need to account for. For this reason, we investigated new simulation methods in order to identify commonalities and differences with respect to their biological characteristics. In the second results part we present an R software package we developed, called NETBlOV. NETBlOV enables the visualization of large biological networks and to highlight structural features that are of biological relevance, e.g., the modularity of these networks.

In this study, a number of human cancer cell lines were characterised for their genomic instability phenotypes: using a panel of centromeric probes utilising the fluorescence <i>in situ</i> hybridisation (FISH) technique to detect CIN, and a microsatellite assay using radioactive PCR labelling to detect MIN.  This characterisation allowed comparison of hBUB1 expression data, and assessed the differences between phenotypic cell line groups for <i>hBUB1</i> mutations and polymorphisms.  Having demonstrated the potential of radioactive-microsatellite PCR to display changes between drug sensitive and selected daughter drug resistant cell lines, steps to assess the actual <i>hBUB1</i> 2q region in primary resected human cancers, with corresponding normal tissue pairs, was taken. NSCLC and CRC tumours were used as models for evaluation of the <i>hBUB1 </i>locus.  NSCLC and CRC subtypes were chosen due to the high incidence of CIN observed in these tumours.  The <i>hBUB1</i> locus on chromosome 2q was assessed for genomic instability using a panel of seven microsatellite markers.  In addition, two novel CA dinucleotide repeats: BUBCA18 a 175bp CA₂₀ product, and BUBCA19 a 119bp CA₁₂ product each located within <i>hBUB1</i> intronic sequence, were cloned and evaluated in the 32 CRC and 20 NSCLC samples.  The microsatellite assessment demonstrated that the 2q locus was highly unstable in CRC, with 62.5% (n=20/32) patient sample pairs displaying MIN in one or more marker.  NSCLC had no instability at any locus, suggesting that aneuploidy due to changes around the hBUB1 chromosome 2 locus was unlikely in this cancer subtype. The known <i>hBUB1</i> mutations and selected polymorphisms were assessed in the two sets of patient groups, and in the human cancer cell lines characterised previously.  Sequencing of the hBUB1 exons using flanking intronic primers and restriction fragment length polymorphism (RFLP) assays, developed for specific coding changes, were utilised to achieve this.  No alterations were observed for the exon 4 and 13 mutations in all samples tested.  Polymorphism analysis similarly revealed low frequencies in CRC and NSCLC. The novel hBUB1 microsatellite repeat, BUBCA18, displayed potential for the use as a clinical marker in detecting CRC within the general population.  Results from our BUBCA18 pilot study demonstrated that heterozygosity for BUBCA18 was significantly associated with disease predisposition (OR 8.2 95%CI 28.24-2.68).  This preliminary finding, and hypothesis that BUBCA18 could be used to test risk, was investigated further in a CRC case-control population: comprising 189 CRC cases and 251 matched controls.  BUBCA18 profiles were assessed using the radioactive-PCR technique.  The BUBCA18 case/control investigation showed that the initial hypothesis was incorrect, as heterozygosity of BUBCA18 was insignificant when used to predict CRC risk, within the larger sample group. The studies reported in this thesis have generated novel data regarding the impact of hBUB1 in human cancer.  Initial evidence suggesting that hBUB1 has a significant role for CIN initiation seems unlikely.

No description available.

No description available.

No description available.

The current model for cancer development envisions cells under going a series of genetic mutations and/or alterations which result in their inability to respond normally to intracellular and extracellular signals that control proliferation, differentiation and death. The number of required genetic alterations varies for different types of cancer and it is likely that further changes occur during its progression to increased malignancy. Thus, cancer is not a static disease but during the development and progression of tumour, multiple changes occur in two kinds of genes: oncogenes and tumour suppressor genes. Oncogene-products can be classified as growth factors, growth factor receptors, Ras oncoproteins, cytoplasmic protein kinases, transcription factors, anti-apoptotic proteins. In particular, the <i>ras</i> oncogene family includes three members: <i>N-ras, K-ras, H-ras</i>. In non-transformed cells, Ras protein, belonging to G-protein family, transduces growth signals from external to the internal environment. In fact, when activated, Ras exchanges GDP with GTP and this allosteric change allows binding of Ras effector molecules and transduction of signalling cascades. Ras activity is required for cell cycle progression. In cancer it has been observed that this oncogene is constitutively activated by mutations and induces the cell to enter into cell-cycle also in the absence of growth signals. Among the transcription factors, a gene involved in many tumours is myc. This transcription factor plays a key role in cell proliferation as its target proteins include many positive regulators of the cell-cycle. In tumour cells the protein product of this oncogene is overexpressed. The cooperationb etweenm ultiple oncogenesa nd/orl oss of tumour suppressors from different functional classes is necessary for transformation to proceed. In fact, it was observed that, although overexpression of a single oncogene does not transform wild-type mouse embryonic fibroblasts, combinations of <i>myc</i> and <i>H-ras</i>^VAL12, can induce cellular transformation and the cells expressing both oncogenes displayeda markedp roliferative advantage. In thyroid, neoplastic transformation generates several different histotypes of tumours, ranging from poorly aggressive and well-differentiated, to highly malignant and undifferentiated anaplastic cancers. The aim of my thesis was to study the tumorigenesis induced by oncogenes and the oncogene cooperation in vivo during the gradual passage from a poorly aggressive to a much more aggressive tumour. To this end a mouse model expressing the two oncogenes <i>H-ras</i>^VAL12 and c-<i>myc</i> (referred as <i>ras</i> and <i>myc</i>) in a tissue-specific as well as in a conditional manner was generated. For this purpose, the coding sequences of the two oncogenes were fused in a bicistronic construct and an IRES (Internal Ribosome Entry Sequence or Site) was inserted between them, to ensure the expression of the second oncogene. The construct was inserted under the control of the promoter of the ubiquitously expressed genes ROSA26 and Eefl<i>a</i>l. In order to express these oncogenes in a tissue-specific manner, the transcription of the two oncogenesis prevented by a STOP sequence flanked by two LoxP sites. Such a STOP sequence can be removed by Cre recombinase protein. The transgenic mice were crossed with mice expressing Cre in a tissue-specific manner. Two strains of transgenic mice expressing Cre in thyroid cells were used: 1. transgenic mice for TgCre, in which Cre is expressed under the control of Tg promoter after the development of the thyroid; 2. Pax8Cre, in which the Cre sequence is inserted in the Pax8 locus and is expressed during the early stages of the thyroid development. In such a manner the oncogenes were expressed only in thyroid cells, but were still inactive. In particular, <i>ras</i> was fused to the mutated ligand binding domain of the estradiol receptor that is sensitive to tamoxifen and not to endogenous estradiol; while <i>myc</i>was fused to the mutated ligand binding domain of the progesterone receptor (hPR891) that is sensitive to RU486 and not to endogenous progesterone. With these fused oncogenes it is possible to activate only Ras (with tamoxifen) or only Myc (with RU486) or both (providing both tamoxifen and RU486). Moreover the activity of two oncogenes might be used to immortalize mouse cell lines in culture.