Cancer is one of the major causes of death in the western world. Resistance to anti-cancer drugs and diagnostic difficulties are major obstacles to successful treatment. This thesis describes studies based on microarray expression analysis and high-throughput compound screening for identification of resistance mechanisms, drug targets and diagnostic markers. In paper I-IV, we applied global expression analysis and measurements of drug response in a human tumor cell line panel to identify drug targets and resistance mechanisms. In paper I, we identified gene transcript levels that correlate with drug resistance and sensitivity. Both well known and new potential markers and mechanisms were identified. In paper II, we showed that STAT1 activity is associated with cross-resistance to both doxorubicin and radiation in vitro and that fludarabine can counteract STAT1 activity and reduce resistance. In Paper III-IV, cell lines were exposed to a compound library consisting of more than thousand different substances in a high-throughput screening effort. These studies revealed that cell line models of squamous cell carcinoma (Paper III) and drug resistant myeloma (Paper IV) are sensitive to phosphodiesterase inhibitors and glucocorticoids respectively. The target molecules for these drugs were over-expressed at the mRNA level and constitute likely explanations for the observed drug potency. In paper V, we identified mRNA markers for the distinction between two types of thyroid tumors, thyroid follicular adenomas and thyroid follicular carcinomas, by means of microarray expression analysis. Our results indicated that distinction between the two tumor types is possible with a small number of markers.

Mikromatriser handlar numera inte bara om att alstra genuttrycksdata i snabb takt, utan det är minst lika viktigt att effektivt ta hand om informationen efteråt. I den här avhandlingen presenteras ett arbetsflöde för att mäta, lagra och analysera genuttrycksdata i asp och poppel (Populus spp.). En Populus} mikromatrisdatabas - UPSC--BASE - tillgänglig för alla intresserade, utvecklades i syfte att samla in och lagra genuttrycksdata. Flertalet analysverktyg gjordes samtidigt tillgängliga, för att möjliggöra ett smidigt arbetsflöde från rådata till biologiska slutsatser. En av de stora utmaningarna i analys av mikromatriser är att kunna särskilja bruset från värdefull biologisk information. Att studera träd som växer utomhus är komplext eftersom de interagerar med omgivningen i mycket större utsträckning än vad som är fallet i växthusets kontrollerade miljö. Det här arbetet visar att det är möjligt, med hjälp av avancerad statistik och god försöksplanering, att följa och jämföra genuttrycket i blad från aspar utomhus under flera år för att dra värdefulla slutsatser om geners reglering. Den lagrade biologiska informationen i UPSC-BASE är avsedd att vara en värdefull tillgång för växtfältet i stort. I databasen finns nästan hundra olika experiment som innefattar alltifrån unga blad till ved, insektsangrepp, kyla och torka samt studier av genmodifierade växter. Informationen kan användas både för jämförande studier inom olika aspförsök, men också för att jämföra med andra växter. För att illustrera möjligheterna, studerades och grupperades gener i blad baserat på hur de uppträder över alla dessa experiment. Dessa grupperingar användes sedan för att definiera gener som är viktiga i bladutvecklingen. Sammanfattningsvis ger arbetet som presenteras i denna avhandling tillgång till verktyg och kunskap för storskaliga studier av genuttryck och den lagrade informationen har bevisats vara en värdefull tillgång för mer ingående studier av geners reglering. / DNA microarray analysis today is not just generation of high-throughput data, much more attention is paid to the subsequent efficient handling of the generated information. In this thesis, a pipeline to generate, store and analyse Populus transcriptional data is presented A public Populus microarray database - UPSC--BASE - was developed to gather and store transcriptomic data. In addition, several tools were provided to facilitate microarray analysis without requirements for expert-level knowledge. The aim has been to streamline the workflow from raw data through to biological interpretation. Differentiating noise from valuable biological information is one of the challenges in DNA microarray analysis. Studying gene regulation in free-growing aspen trees represents a complex analysis scenario as the trees are exposed to, and interacting with, the environment to a much higher extent than under highly controlled conditions in the greenhouse. This work shows that, by using multivariate statistics and experimental planning, it is possible to follow and compare gene expression in leaves from multiple growing seasons, and draw valuable conclusions about gene expression from field-grown samples. The biological information in UPSC-BASE is intended to be a valuable transcriptomic resource also for the wider plant community. The database provides information from almost a hundred different experiments, spanning different developmental stages, tissue types, abiotic and biotic stresses and mutants. The information can potentially be used for both cross-experiment analysis and for comparisons against other plants, such as Arabidopsis or rice. As a demonstration of this, microarray experiments performed on Populus leaves were merged and genes preferentially expressed in leaves were organised in to regulons of co-regulated genes. Those regulons were used to define genes of importance in leaf development in Populus. Taken together, the work presented in this thesis provides tools and knowledge for large-scale transcriptional studies and the stored gene expression information has been proven to be a valuable information resource for in-depth studies about gene regulation.

Hindsight (HNT) is a zinc finger transcription factor that is required for morphogenesis of the Drosophila embryo, having roles in germ band retraction (GBR) as well as dorsal closure (DC). HNT expression is also found in sensory organ precursors (SOP) of the developing pupal peripheral nervous system, and muscle progenitor cells, but the role of HNT in neurogenesis and myogenesis during embryogenesis has not been investigated in any depth. Microarray analysis of embryos over-expressing HNT during GBR and DC identified 1290 genes with significant changes in expression. This data set included many potential HNT targets, including genes associated with myogensis, and a disruption of muscle development was observed in embryos over-expressing HNT. It is possible that HNT may function to repress muscle identity genes in muscle founder cells. In addition, HNT over expressing embryos were found to resemble the neurogenic class of mutants. Among the potential target genes, D-Pax2 (shaven, sparkling, CG11049) expression, which is known to be expressed in the developing peripheral nervous system, was confirmed to be up-regulated following HNT over-expression. Interestingly, D-Pax2 and HNT expression were found to co-localize at the onset of their expression at stages 10-12 in embryos, but were not co-localized in later stages of embryogenesis. The up-regulation of D-Pax2 by HNT over-expression was further characterized and was found to be associated with strong ectopic HNT expression. The relevance of HNT to the regulation of D-Pax2 during normal development remains to be determined, but it is possible that endogenous expression of HNT is involved in D-Pax2 repression.

Non-small cell lung cancer (NSCLC) constitutes a clinically, histologically, and genetically heterogeneous disease entity that represents a major cause of cancer-related death. Early-stage patients, who undergo surgery with curative intent, experience high recurrence rates and the effect of adjuvant treatment is modest. Prognostic biomarkers would be of particular relevance to guide intensified treatment depending on expected outcome and moreover often infer a biological role in tumourigenesis. This thesis presents a translational study approach to establish a well-characterised NSCLC frozen-tissue cohort and to obtain a profile of each specimen with regard to genome-wide copy number alterations, global gene expression levels and somatic mutations in selected cancer-related genes. Furthermore, the generation of a formalin-fixed, paraffin-embedded tissue microarray enabled validation of findings on the protein level using immunohistochemistry. The comprehensive molecular characterisation, combined with data on clinical parameters, enabled the analysis of biomarkers linked to disease outcome. In Paper I, single nucleotide polymorphism arrays were applied to assess copy number alterations in NSCLC and associations with overall survival in adenocarcinoma and squamous cell carcinoma were described. In Paper II, we evaluated expression levels of selected stromal proteins in NSCLC using immunohistochemistry and the adhesion molecule CD99 was identified as an outcome-related biomarker in two independent cohorts. Paper III presents a strategy for prognostic biomarker discovery based on gene expression profiling, meta-analysis, and validation of protein expression on tissue microarrays, and suggests the putative tumour suppressor CADM1 as a candidate biomarker. In Paper IV, we propose a prognostic role for tumour-infiltrating IGKC-expressing plasma cells in the local tumour microenvironment, indicating an involvement of the humoral immune response in anti-tumor activity. In Paper V, we combined next-generation deep sequencing with statistical analysis of the TP53 database to define novel parameters for database curation. In summary, this thesis exemplifies the benefits of a translational study approach, based on a comprehensive tumour characterisation, and describes molecular markers associated with clinical outcome in NSCLC.

This research used classical and molecular epidemiology tools to assess the potential importance of feedlot cattle as Campylobacter reservoirs. The project was conducted from November 2004 to September 2005 in southern Alberta.<p> Fresh pen-floor fecal samples were collected from commercial feedlot cattle near slaughter weight in seven feedlots. Overall, 87% of 2,776 fecal samples were culture positive for Campylobacter species (86% of 1,400 in winter, 88% of 1,376 in summer), and 69% of 1,486 Campylobacter positive isolates were identified as <i>Campylobacter jejuni</i>. After accounting for clustering within pen and feedlot, the number of days-on-feed and feedlot size were associated (p ¡Ü 0.05) with Campylobacter species isolation rates.<p> Retail ground beef was collected from 60 grocery stores (four chains, three cities). None of the 1,200 packages were culture positive for Campylobacter species. Polymerase chain reaction (PCR) results from a subset of samples (n=142) indicated that 48% of packages were positive for Campylobacter DNA. By species, 14.8% (21/142), 26.8% (38/142) and 1.4% (2/142) of packages were PCR positive for <i>C. jejuni</i>, <i>C. coli</i> and <i>C. hyointestinalis</i> DNA, respectively. The collection period (1, 2, 3 or 4) was associated (p ¡Ü 0.05) with the odds of detecting Campylobacter species DNA using PCR.<p> Oligonucleotide DNA microarrays were used as a platform for comparative genomic hybridization (CGH) analysis of 87 C. jejuni isolates (46 bovine, 41 human) obtained within the same geographical regions and time frame. Of the 13 CGH clusters identified based on overall comparative genomic profile similarity, nine contained human and cattle isolates, three contained only human isolates, and one contained only cattle isolates. In addition, human clinical and feedlot cattle C. jejuni isolates were compared on a gene-by-gene basis and only a small number of the 1,399 genes tested were unequally distributed between the two groups (p ¡Ü 0.05).<p> The high isolation rates of Campylobacter species and <i>C. jejuni</i> reported here may have implications for food safety, public health and environmental contamination. Our findings suggest that feedlot cattle and human <i>C. jejuni</i>strains are very similar and may be endemic within southern Alberta.

Detection of cellular messenger RNA is a useful diagnostic strategy for the detection of patho-gens. A rapid and sensitive method for on-site detection of specific pathogens would be of great use in a number of fields. For example, a simple and inexpensive method for the detection of harmful biological agents in train stations and airports is useful for national security. Rapid detection of pathogenic E. coli strains in food production would also be of great benefit in ensuring the safety and quality of our food supply. Here we present a method for the rapid de-tection of cellular mRNA. This system is based on the 3’-labeling approach in which targeted RNA is simultaneously extended and labeled with the use of biotin labeled-dNTPs and DNA po-lymerase on an immobilized nucleic acid probe. The biotin is subsequently converted to an enzymatic label, which produces a detectable chemiluminescent reaction in the presence of substrate. Detection time of this system is short (approximately 20 minutes) because there is no need for amplification by PCR, transcription, or fluorophore labeling. This novel methodology has been successfully demonstrated by selective detection of lac Z mRNA in a total RNA sample from E. coli.

The type II tumour suppressor ING3 has been shown to modulate transcription, cell cycle control, and apoptosis. To investigate the putative role of ING3 in melanoma development, we examined the expression of ING3 in 58 dysplastic nevi, 114 primary melanomas, and 50 metastatic melanomas with tissue microarray and immunohistochemistry. Overall ING3 was reduced in metastatic melanomas compared with dyslastic nevi and primary melanomas. Reduced nuclear ING3 staining also correlated with melanoma progression, increased cytoplasmic ING3 level, tumour location at sun-exposed sites, and a poorer disease-specific 5-year survival of patients with primary melanoma. Multivariate analysis revealed that nuclear ING3 staining can independently predict patient outcome in primary melanomas. In melanoma cells, ING3 expression was rapidly induced by UV irradiation. Using stable clones of melanoma cells overexpressing ING3, we showed that ING3 significantly promoted UV-induced apoptosis. Unlike its homologues ING1b and ING2, ING3-enhanced apoptosis upon UV irradiation was independent of functional p53. Furthermore, ING3 did not affect the expression of mitochondrial proteins but increased the cleavage of Bid and caspases. Moreover, ING3 upregulated Fas expression and ING3-mediated apoptosis was blocked by inhibiting caspase-8 or Fas activation. Knockdown of ING3 expression decreased UV-induced apoptosis remarkably, suggesting that ING3 plays a crucial role in cellular response to UV radiation. To explore how ING3 is deregulated in advanced melanomas, we examined ING3 expression in metastatic melanoma cells and found that ING3 was downregulated due to a rapid protein turnover in these cells. Further studies demonstrated that ING3 undergoes degradation via the ubiquitin-proteasome pathway. We also demonstrate that ING3 interacts with the SCF (Skp1/Cul1/Roc1/Skp2) E3 ligase complex. Knockdown of Cul1 or Skp2 significantly stabilized ING3 in melanoma cells. In addition, lysine residue 96 is essential for ING3 ubiquitination as its mutation to arginine completely abrogated ING3 turnover and enhanced ING3-stimulatd apoptosis upon UV irradiation. Taken together, ING3 is deregulated in melanomas as a result of both nucleus-to-cytoplasm shift and rapid degradation. The level of ING3 in the nucleus may be an important marker for human melanoma progression and prognosis. Restoration of ING3 expression significantly sensitizes melanoma cells to UV radiation through the activation of Fas/caspase-8 pathway.

Notch signaling is important for development in Caenorhabditis elegans and the REF-1 family pro-teins, a set of the bHLH transcription factors, are the first targets of Notch signaling. Little is known about the molecular mechanisms employed by the REF-1 family to regulate development. In this project, I iden-tified novel targets of three REF-1 family proteins, HLH-25/HLH-28/HLH-29, and determined which target genes are activated and which are repressed by the REF-1 proteins. These targets were identified by gene expression microarray and were functionally categorized by Gene Oncology analysis. A systems biology approach was performed to identify networks associated with those targets. In addition to the mo-lecular genetics studies, I identified and better characterized the range of phenotypes induced by muta-tions in ref-1 family genes.

We consider the problem of testing the equality of two sample means, when the number of tests performed is large. Applying this problem to the context of gene expression data, our goal is to detect a set of genes differentially expressed under two treatments or two biological conditions. A null hypothesis of no difference in the gene expression under the two conditions is constructed. Since such a hypothesis is tested for each gene, it follows that thousands of tests are performed simultaneously, and multiple testing issues then arise. The aim of our research is to make a connection between Bayesian analysis and frequentist theory in the context of multiple comparisons by deriving some properties shared by both p-values and posterior probabilities. The ultimate goal of this work is to use the posterior probability of the one-sided alternative hypothesis (or equivalently, posterior probability of the half-space) in the same spirit as a p-value. We show for instance that such a Bayesian probability can be used as an input in some standard multiple testing procedures controlling for the False Discovery rate.

This research used classical and molecular epidemiology tools to assess the potential importance of feedlot cattle as Campylobacter reservoirs. The project was conducted from November 2004 to September 2005 in southern Alberta.<p> Fresh pen-floor fecal samples were collected from commercial feedlot cattle near slaughter weight in seven feedlots. Overall, 87% of 2,776 fecal samples were culture positive for Campylobacter species (86% of 1,400 in winter, 88% of 1,376 in summer), and 69% of 1,486 Campylobacter positive isolates were identified as <i>Campylobacter jejuni</i>. After accounting for clustering within pen and feedlot, the number of days-on-feed and feedlot size were associated (p ¡Ü 0.05) with Campylobacter species isolation rates.<p> Retail ground beef was collected from 60 grocery stores (four chains, three cities). None of the 1,200 packages were culture positive for Campylobacter species. Polymerase chain reaction (PCR) results from a subset of samples (n=142) indicated that 48% of packages were positive for Campylobacter DNA. By species, 14.8% (21/142), 26.8% (38/142) and 1.4% (2/142) of packages were PCR positive for <i>C. jejuni</i>, <i>C. coli</i> and <i>C. hyointestinalis</i> DNA, respectively. The collection period (1, 2, 3 or 4) was associated (p ¡Ü 0.05) with the odds of detecting Campylobacter species DNA using PCR.<p> Oligonucleotide DNA microarrays were used as a platform for comparative genomic hybridization (CGH) analysis of 87 C. jejuni isolates (46 bovine, 41 human) obtained within the same geographical regions and time frame. Of the 13 CGH clusters identified based on overall comparative genomic profile similarity, nine contained human and cattle isolates, three contained only human isolates, and one contained only cattle isolates. In addition, human clinical and feedlot cattle C. jejuni isolates were compared on a gene-by-gene basis and only a small number of the 1,399 genes tested were unequally distributed between the two groups (p ¡Ü 0.05).<p> The high isolation rates of Campylobacter species and <i>C. jejuni</i> reported here may have implications for food safety, public health and environmental contamination. Our findings suggest that feedlot cattle and human <i>C. jejuni</i>strains are very similar and may be endemic within southern Alberta.