The focus of this dissertation was leptin and the leptin receptor, and the role of these genes (OB and OB-R) in the development of obesity and type 2 diabetes in humans and Psammomys obesus, a polygenic rodent model of obesity and type 2 diabetes. Studies in humans showed that circulating leptin concentrations were positively associated with adiposity, and independently associated with circulating insulin and triglyceride concentrations. Analysis of two leptin receptor sequence polymorphisms in a Caucasian Australian population and a population of Nauruan males, with very high prevalence rates of obesity, showed no associations between sequence variation within the OB-R gene and obesity- or diabetes-related phenotypic measures. In addition, these two OB-R polymorphisms were not associated with longitudinal changes in body mass or composition in either of the populations examined. A unique analysis of the effects of multiple gene defects in the Nauruan population, demonstrated that the presence of sequence alterations in both the OB and OB-R genes were associated with insulin resistance. Psammomys obesus is regarded as an excellent rodent model in which to study the development of obesity and type 2 diabetes in humans. Examination of circulating leptin concentrations in Psammomys revealed that, as in humans, leptin concentrations were associated with adiposity, and independently associated with circulating insulin concentrations. This animal model was utilised to examine expression of OB-R, and the regulation of expression of this gene after dietary manipulation. OB-R is known to have several isoforms, and in particular, OB-RA and OB-RB gene expression were examined. OB-RB is the main signalling isoform of the leptin receptors. It has a long intracellular domain and has previously been shown to play an important role in energy balance and body weight regulation in rodents and humans. OB-RA is a much shorter isoform of OB-R, and although it lacks the long intracellular domain necessary to activate the JAK/STAT pathway, OB-RA is also capable of signalling, although to a lesser degree than OB-RB. OB-RA is found to be expressed almost ubiquitously throughout the body, and this isoform may be involved in transport of leptin into the cell, although its role remains unclear. OB-RA and OB-RB were both found to be expressed in a large number of tissues in Psammomys obesus. Interestingly, obese Psammomys were found to have lower levels of expression of OB-RA and OB-RB in the hypothalamus, compared to lean animals. This finding raises the possibility that decreased leptin signalling in the brain of obese, hyperleptinemic Psammomys obesus may contribute to the leptin resistance previously described in this animal model. However, the primary defect is unclear, as alternatively, increased circulating leptin concentrations may lead to down-regulation of leptin receptors. The effect of fasting on leptin concentrations and gene expression of OB-RA and OB-RB was also examined. A 24-hour fast resulted in no change in body weight, but a reduction in circulating leptin concentrations, and an increase in hypothalamic OB-RB gene expression in lean Psammomys. In obese animals, fasting again did not alter body weight, but resulted in an increase in both circulating leptin concentrations and hypothalamic OB-RB gene expression. In the liver, fasting resulted in a large increase in OB-RA gene expression in both lean and obese animals. These results highlighted the fact that regulation of leptin receptor gene expression in polygenic models of obesity and type 2 diabetes is complex, and not solely under the control of circulating leptin concentrations. Sucrose-feeding is an established method of inducing obesity and type 2 diabetes in rodents, and this experimental paradigm was utilised to examine the effects of longer term perturbations of energy balance on the leptin signalling pathway in Psammomys obesus. Addition of a 5% sucrose solution to the diet of lean and obese Psammomys resulted in increased body weight in both groups of animals, however only obese Psammomys showed increased fat mass and the development of type 2 diabetes. The changes in body mass and composition with sucrose-feeding were accompanied by decreased circulating leptin concentrations in both groups of animals, as well as a range of changes in leptin receptor gene expression. Sucrose-feeding increased hypothalamic OB-RB gene expression in obese Psammomys only, while in the liver there was evidence of a reduction in OB-RA and OB-RB gene expression in both lean and obese animals. The direct effects of sucrose on the leptin signalling pathway are unclear, however it is possible to speculate that the effect of sucrose to decrease leptin concentrations may have been involved in the exacerbation of obesity and the development of type 2 diabetes in obese Psammomys, From these studies, it appears that sequence variation in the OB and OB-R genes is unlikely to be a major factor in the etiology of obesity in human populations. The ability to examine regulation of expression of these genes in Psammomys obesus, however, has demonstrated that the effects of nutritional modifications on leptin receptor gene expression need closer attention. The role of the OB and OB-R genes in metabolism and the development of type 2 diabetes also warrants further examination, with particular attention on the differential effects of dietary modifications on leptin receptor gene expression across a range of tissues.

Gene expression and its regulation is a highly coordinated system, involved in many biological processes such as cell growth, division and differentiation. Transcriptional regions, involved in gene regulation, consist of a heterogeneous collection of smaller regulatory elements. In some cases, co-regulated genes contain a common set of transcription factor binding sites (TFBS). Analysis of promoter regions is the major approach in understanding the transcriptional regulatory mechanisms. It is also useful for interpretation of mammalian gene expression studies, where co-expressed genes may share motifs representing putative TFBS. Motif identification also has the advantage that it can predict control regions in genes that have not been measured experimentally. However, a common problem is incomplete genomic sequence for the experimental species of interest. The approach here is to identify and use orthologous gene promoter sequences from a related and well-characterised species. The primary aim of this study was to identify and predict regulatory TFBS in species where promoter sequence does not exist or is incomplete. The MEME programme was employed for the motif prediction step. The predicted elements were subsequently compared to known TFBS using TRANSFAC and JASPAR databases for identification. A methodology based on relative entropy was used. The validity of the method was confirmed as the predicted motifs in the training set were the expected sites involved in regulation of muscle development. The technique was applied to two data sets, generated from expressed sequence tag (EST) clustering analysis and microarray experiments. All data sets, software and results are available on the accompanying CD. Bovine expression data was analysed for cardiac-specific expression using two separate approaches, combining bovine library EST frequency and human gene expression ratios. For each approach, the orthologous human and bovine promoter sequences were analysed for common motifs. Across all comparisons, 37% of motifs were identified as known TFBS using the TRANSFAC and JASPAR databases. As the human comparison had more promoter sequences available, this was the main limiting factor for the corresponding bovine analysis, rather than cross-species divergence or accuracy of gene expression measurement. Results from this study demonstrate that using promoter sequences from a related species is a viable approach when studying gene expression in species with limited amount of genomic sequence. As the bovine genome becomes more complete, it can in turn serve as the reference genome for other agriculturally important ruminants, such as sheep, goat and deer. The second application concerned in silico analysis of gene regulation patterns in response to stimuli. Recently it has been shown that a mutation in the bone morphogenetic receptor IB leads to an increased ovulation rate in sheep. The objective of this study was to analyse gene expression patterns in cultured cells in response to four members of the BMP family, i.e. BMP2, BMP4, BMP6 and BMP7 and the control TGFβ. Microarray data was provided by J. Young. Twelve highly upregulated genes were stimulated by all BMPs, seven of which are known BMP target genes. Analysis of the predicted motifs identified four elements that may be involved in the regulation process. Cross-species comparison for one of the genes, ID1, showed high conservation of one of the motifs across 11 mammalian genomes. This particular motif had not been identified as a known binding site. In summary, the analysis of the expression data suggest an extension of the list of BMP targets. The proposed method is relatively robust when sufficiently co-expressed (co-regulated) sequences can be identified, whether from the same or another species.

Correct development of multicellular organisms requires the programmed removal of supernumerary, redundant, or damaged cells, a process achieved by apoptosis. Apoptosis, or Programmed Cell Death (PCD) is executed by caspases, a highly conserved family of cysteine proteases. The removal of redundant larval tissues during metamorphosis is controlled by the steroid hormone ecdysone. Ecdysone signalling is mediated by the nuclear receptor heterodimer EcR/Usp, which in turn transcriptionally activates a host of transcription factors which then go on to regulate genes essential for PCD, like caspases. The apical caspase dronc is upregulated in the larval midgut and salivary glands prior to their destruction and is dependent on the BRC and E93 transcription factors. To further understand the role of dronc in development, a dronc mutant fly was generated and a transgenic promoter-reporter strategy was employed to investigate dronc regulation. Larval organs from dronc mutants lack dying cells and, when irradiated, fail to show a radiation-induced PCD response. The midguts from dronc mutants undergo apoptosis and have high caspase activity. These data indicate that a droncindependent caspase activation pathway is active in the midgut. Salivary glands from dronc mutants failed to be removed and have reduced caspase activity. Consequently it is clear that the role of DRONC differs significantly between the midgut and salivary glands. The employment of a transgenic dronc promoter-LacZ reporter system identified promoter regions essential for the correct temporal and spatial expression of dronc. A region of the dronc promoter between 1.1kb and 2.8kb has elements essential for LacZ expression in salivary glands. This region was also dependent on the BR-C and E93 transcription factors for salivary gland expression. A functional ecdysone receptor binding element (EcRBE) was identified in the dronc proximal promoter. The EcR-B1 isoform directly binds this EcRBE and is necessary for correct dronc expression in the larval salivary glands. This work revealed some novel findings regarding the role of DRONC in development and the availability of a specific dronc mutant now makes it possible to explore some of the recently published non- poptotic roles of dronc. This work aids in understanding how nuclear hormones control transcription and shows dronc to be an ideal model gene to explore these molecular and genetic processes. / Thesis (Ph.D.)--Department of Medicine, 2004.

Thesis (Ph. D.)--University of Hong Kong, 2006. / Title proper from title frame. Also available in printed format.

Thesis (M. Phil.)--University of Hong Kong, 2007. / Title proper from title frame. Also available in printed format.

Thesis (Ph. D.)--University of Hong Kong, 2007. / Title proper from title frame. Also available in printed format.

The identification of specific and reproducible biomarkers is critical for the early diagnosis of cancer, which has a profound effect the survival rate of patients. Comprehensive laboratory and clinical evidence needs to be collected to confirm the accuracy of the biomarkers prior to their clinical use. Heterogenous nuclear ribonucleoproteins (hnRNPs) A2 and B1 have been suggested as biomarkers for cancer since 1988 when hnRNP A2/B1 overexpression was first linked with the occurrence of lung cancer. Later studies established a correlation between the expression levels of these hnRNPs and other cancers, such as breast, pancreatic, and lymphatic tumours. In this study, the expression of hnRNPs A1, A2, A3, and B1 has been investigated in various cancer cell lines. hnRNPs A1 and A3, in addition to A2 and B1, were found to be overexpressed in some cancer types. However, the overexpression of none of the hnRNP A/B proteins was universal, and their upregulation may be limited to a few cell types, suggesting they may be effective biomarkers for a subset of cancers. The upregulation of hnRNP A/B proteins in tumours and cancer cell lines led to the hypothesis that they are involved in the uncontrolled cell growth in cancer. According to our Western blot analysis, expression of the hnRNP proteins, A1, A2, and B1, is dependent on the cell cycle whereas no significant change was detected for hnRNP A3, implying that the former three are needed during certain cell cycle stages. The results, together with the transcription factor analysis of the promoter regions of the HNRPA1, HNRPA2, and HNRPA3 genes, suggest that hnRNPs A1, A2, and A3 may have distinct regulatory machineries and cellular functions although they have high amino acid sequence identity. However, their mRNA levels were unchanged across the cell cycle, suggesting the cell-cycle-dependent expression of hnRNPs A1, A2, and B1 is modulated at the translational level. Previous studies showed higher expression of hnRNPs A1 and A2 in rapidly proliferating cells than in quiescent cells, suggesting a role of these proteins in cell proliferation. Though interruption of hnRNP A1 expression did not result in significant change in the viability of murine CB3 cells, simultaneous suppression of hnRNPs A1 and A2 caused apoptosis in a few cell lines. Consistent with this, suppression of hnRNP A1 or A3 expression in our study in Colo16 squamous cells using RNA interference did not affect cell proliferation, but simultaneous suppression of both caused slow cell proliferation. By contrast, reduction of the hnRNP A2 level alone slowed the proliferation of Colo16 cells. These results suggest that although hnRNPs A1, A2, and A3 share some roles in cell proliferation, each of them may have distinct tasks. This conclusion is supported by the data from the comparative analysis of the downstream targets of hnRNPs A1, A2, and A3, which has shown that these three proteins share a limited number of common downstream proteins. The observed impact on cell proliferation of suppressing hnRNP A2 subfamily proteins is in accord with our finding that the downstream targets of hnRNP A2 are overrepresented by genes involved in proliferation regulation, as shown in microarray and real-time PCR analysis. These include cyclin-dependent kinase (CDK) inhibitors, p21 and p27, and their regulatory proteins, such as Skp2 and Rpn10. Skp2 controls the ubiquitination of p21 and p27, and Rpn10 links them to the 26S proteasome, the complex that degrades these two CDK inhibitors. hnRNP A2 also regulates the transcription of securin and separin, which are essential for sister chromatid separation during late anaphase. In addition, hnRNP A2 can also influence cell proliferation through cell growth factors, including fibroblast, vascular endothelial, transforming, and insulin growth factors. Our gene array and real-time PCR analysis have shown that hnRNP A2 regulates the expression of these factors, their receptors, or associated proteins such as IGFBP7 and TGFBR2. The data presented in this thesis link the overexpression of hnRNP A/B proteins, in particular the A2/B1 subfamily, in cancer with their regulatory roles in cell division and cell proliferation. Our findings provide mechanistic evidence that these proteins may be a driving force for the uncontrolled cell growth in cancer, suggesting that some of hnRNP A/B proteins may be potential therapeutic targets for cancer. However, further studies are needed to obtain a global view of the roles of these proteins in cancer.

The identification of specific and reproducible biomarkers is critical for the early diagnosis of cancer, which has a profound effect the survival rate of patients. Comprehensive laboratory and clinical evidence needs to be collected to confirm the accuracy of the biomarkers prior to their clinical use. Heterogenous nuclear ribonucleoproteins (hnRNPs) A2 and B1 have been suggested as biomarkers for cancer since 1988 when hnRNP A2/B1 overexpression was first linked with the occurrence of lung cancer. Later studies established a correlation between the expression levels of these hnRNPs and other cancers, such as breast, pancreatic, and lymphatic tumours. In this study, the expression of hnRNPs A1, A2, A3, and B1 has been investigated in various cancer cell lines. hnRNPs A1 and A3, in addition to A2 and B1, were found to be overexpressed in some cancer types. However, the overexpression of none of the hnRNP A/B proteins was universal, and their upregulation may be limited to a few cell types, suggesting they may be effective biomarkers for a subset of cancers. The upregulation of hnRNP A/B proteins in tumours and cancer cell lines led to the hypothesis that they are involved in the uncontrolled cell growth in cancer. According to our Western blot analysis, expression of the hnRNP proteins, A1, A2, and B1, is dependent on the cell cycle whereas no significant change was detected for hnRNP A3, implying that the former three are needed during certain cell cycle stages. The results, together with the transcription factor analysis of the promoter regions of the HNRPA1, HNRPA2, and HNRPA3 genes, suggest that hnRNPs A1, A2, and A3 may have distinct regulatory machineries and cellular functions although they have high amino acid sequence identity. However, their mRNA levels were unchanged across the cell cycle, suggesting the cell-cycle-dependent expression of hnRNPs A1, A2, and B1 is modulated at the translational level. Previous studies showed higher expression of hnRNPs A1 and A2 in rapidly proliferating cells than in quiescent cells, suggesting a role of these proteins in cell proliferation. Though interruption of hnRNP A1 expression did not result in significant change in the viability of murine CB3 cells, simultaneous suppression of hnRNPs A1 and A2 caused apoptosis in a few cell lines. Consistent with this, suppression of hnRNP A1 or A3 expression in our study in Colo16 squamous cells using RNA interference did not affect cell proliferation, but simultaneous suppression of both caused slow cell proliferation. By contrast, reduction of the hnRNP A2 level alone slowed the proliferation of Colo16 cells. These results suggest that although hnRNPs A1, A2, and A3 share some roles in cell proliferation, each of them may have distinct tasks. This conclusion is supported by the data from the comparative analysis of the downstream targets of hnRNPs A1, A2, and A3, which has shown that these three proteins share a limited number of common downstream proteins. The observed impact on cell proliferation of suppressing hnRNP A2 subfamily proteins is in accord with our finding that the downstream targets of hnRNP A2 are overrepresented by genes involved in proliferation regulation, as shown in microarray and real-time PCR analysis. These include cyclin-dependent kinase (CDK) inhibitors, p21 and p27, and their regulatory proteins, such as Skp2 and Rpn10. Skp2 controls the ubiquitination of p21 and p27, and Rpn10 links them to the 26S proteasome, the complex that degrades these two CDK inhibitors. hnRNP A2 also regulates the transcription of securin and separin, which are essential for sister chromatid separation during late anaphase. In addition, hnRNP A2 can also influence cell proliferation through cell growth factors, including fibroblast, vascular endothelial, transforming, and insulin growth factors. Our gene array and real-time PCR analysis have shown that hnRNP A2 regulates the expression of these factors, their receptors, or associated proteins such as IGFBP7 and TGFBR2. The data presented in this thesis link the overexpression of hnRNP A/B proteins, in particular the A2/B1 subfamily, in cancer with their regulatory roles in cell division and cell proliferation. Our findings provide mechanistic evidence that these proteins may be a driving force for the uncontrolled cell growth in cancer, suggesting that some of hnRNP A/B proteins may be potential therapeutic targets for cancer. However, further studies are needed to obtain a global view of the roles of these proteins in cancer.

<p>Cancer is a common disease and due to problems with resistance against cancer drugs and the limited benefit from chemotherapy in many diagnoses, there is a need to develop new cancer drugs. In this thesis new methods to screen for cancer drugs and to evaluate their mechanism of action are discussed. </p><p>In Paper I, it was found that by studying the gene expression of a cell line panel and combining the data with sensitivity data of a number of cytotoxic drugs, it was possible to cluster compounds according to mechanism of action as well as identifying genes associated with chemosensitivity.</p><p>In Paper II, studies of compounds with selective activity in drug-resistant cell lines revealed the glucocorticoids as a group of interesting compounds. The glucocorticoid receptor was overexpressed in 8226/Dox40 and the difference in sensitivity was abolished when the cells were treated with a glucocorticoid receptor antagonist.</p><p>In Paper III, an image-based screening method for new proteasome inhibitors was successfully developed and the compounds disulfiram, PDTC and NSC 95397 were identified as inhibitors of the proteasome.</p><p>In Paper IV, disulfiram and PDTC were shown to induce cytotoxic activity, to inhibit the activation of the transcription factor NFkappaB and to inhibit the degradation of proteins normally degraded by the proteasome.</p><p>In Paper V, NSC 95397 was shown to be cytotoxic to all cells in the resistance-based cell line panel as well as to patient samples from a variety of cancer diagnoses. Connectivity Map was successfully used as a tool to propose a new mechanism of action of NSC 95397. The gene expression induced by NSC 95397-treatment was similar to that induced by several proteasome inhibitors not present in the Connectivity Map.</p>

The overall objective of this research was to define the cellular basis underlying heritable subfertility in roosters homozygous for the rose comb allele (R/R). Fertilization in the hen is preceded by the ascension of motile sperm through the vagina and sperm sequestration within sperm storage tubules (SST). The objective of the first set of experiments was to determine if reduced sperm sequestration could account for subfertility. Sperm sequestration differed between genotypes following intravaginal insemination (p<0.0001). However, sperm sequestration did not differ between genotypes when sperm were incubated with SST in vitro (p>0.05). Therefore, subfertility was attributed to reduced sperm transport within the vagina. To test this hypothesis, an assay was developed to evaluate fowl sperm motility in vitro. Based upon this assay, ejaculates from subfertile males contained smaller subpopulations of highly motile sperm than the ejaculates from controls (p<0.001). The objective of the next set of experiments was to characterize the motility of individual sperm and to identify a mechanism that could account for the genotypic difference in sperm cell motility. Computer-assisted sperm motion analysis evaluation revealed that ejaculates from controls contained 91% motile sperm whereas ejaculates from subfertile males contained 62% motile sperm (p<0.001). The ATP concentration in sperm from subfertile males was 63% less than that of sperm from controls (p<0.001). A link between sperm ATP concentration and immotility was investigated. First, sperm metabolism was evaluated using motility as an endpoint. The genotypic difference in sperm motility persisted when ATP synthesis was limited to glycolysis (p<0.001). Consequently, mitochondrial respiration could not account for the genotypic difference in sperm motility. In contrast, sperm uptake of [1,2-��H] 2-deoxy-D-glucose did differ between genotypes (p<0.001). The activity of key glycolytic enzymes, creatine kinase, and dynein ATPase did not differ between genotypes (p>0.05). Therefore reduced sperm motility did not appear to be due to ATP synthesis, allocation of high energy phosphate bonds along the axoneme, or ATP consumption (p>0.05). In conclusion, subfertility of roosters homozygous for the rose comb allele was attributed to decreased spermatozoal glucose transport. / Graduation date: 1997