Heterozygous mutations and deletions of the gene that encodes the transcription factor hepatocyte nuclear factor 1β (HNF1B) are the commonest known monogenic cause of developmental kidney disease. However, diagnosis remains challenging due to phenotypic variability and frequent absence of a family history. There is also no consensus as to when HNF1B genetic testing should be performed. This thesis includes work looking at the identification of HNF1B-associated disease. An HNF1B score was developed in 2014 to help select appropriate patients for genetic testing. The aim in chapter 2 was to test the clinical utility of this score in a large number of referrals for HNF1B genetic testing to the UK diagnostic testing service for the HNF1B gene. An HNF1B score was assigned for 686 referrals using clinical information available at the time of testing; performance of the score was evaluated by receiver-operating characteristic curve analysis. Although the HNF1B score discriminated between patients with and without a mutation/deletion reasonably well, the negative predictive value of 85% reduces its clinical utility. HNF1B-associated disease is due to an approximate 1.3 Mb deletion of chromosome 17q12 in about 50% of individuals. This deletion includes HNF1B plus 14 additional genes and has been linked to an increased risk of neurodevelopmental disorders, such as autism. The aim in chapter 3 was to compare the neurodevelopmental phenotype of patients with either an HNF1B intragenic mutation or 17q12 deletion to determine whether haploinsufficiency of the HNF1B gene is responsible for this aspect of the phenotype. Brief behavioural screening showed high levels of psychopathology and impact in children with a deletion. 8/20 (40%) patients with a deletion had a clinical diagnosis of a neurodevelopmental disorder compared to 0/18 with a mutation, P=0.004. 17q12 deletions were also associated with more autistic traits. Two independent clinical geneticists were able to predict the presence of a deletion with a sensitivity of 83% and specificity of 79% when assessing facial dysmorphic features as a whole. These results demonstrate that the 17q12 deletion but not HNF1B intragenic mutations are associated with neurodevelopmental disorders; we conclude that the HNF1B gene is not involved in the neurodevelopmental phenotype of these patients. Extra-renal phenotypes frequently occur in HNF1B-associated disease, including diabetes mellitus and pancreatic hypoplasia. Faecal elastase-1 levels have only been reported in a small number of individuals, the majority of which have diabetes. In chapter 4 we measured faecal elastase-1 in patients with an HNF1B mutation or deletion regardless of diabetes status and assessed the degree of symptoms associated with pancreatic exocrine deficiency. We found that faecal elastase-1 deficiency is a common feature of HNF1B-associated renal disease even when diabetes is not present and pancreatic exocrine deficiency may be more symptomatic than previously suggested. Faecal elastase-1 should be measured in all patients with a known HNF1B molecular abnormality complaining of chronic abdominal pain, loose stools or unintentional weight loss. Hypomagnesaemia is a common feature of HNF1B-associated disease and is due to renal magnesium wasting. The aim in chapter 5 was to measure both serum and urine magnesium and calcium levels in individuals with an HNF1B molecular defect and compare to a cohort of patients followed up in a general nephrology clinic in order to assess their potential as biomarkers for HNF1B-associated disease. The results of this pilot study show that using a cut-off for serum magnesium of ≤0.75 mmol/L was 100% sensitive and 87.5% specific for the presence of an HNF1B mutation/deletion. All individuals in the HNF1B cohort had hypermagnesuria with fractional excretion of magnesium >4%; a cut-off of ≥4.1% was 100% sensitive and 71% specific. This suggests serum magnesium levels and fractional excretion of magnesium are highly sensitive biomarkers for HNF1B-associated renal disease; if these results are confirmed in a larger study of patients with congenital anomalies of the kidneys or urinary tract they could be implemented as cheap screening tests for HNF1B genetic testing in routine clinical care.

Primary human hepatocytes are a scarce resource with variable function, which diminishes with time in culture. As a consequence their use in tissue modelling and therapy is restricted. Human embryonic stem cells (hESCs) could provide a stable source of human tissue due to their properties of self-renewal and their ability to give rise to all three germ layers. hESCs have the potential to provide an unlimited supply of hepatic endoderm (HE) which could offer efficient tools for drug discovery, disease modelling and therapeutic applications. In order to create a suitable environment to enhance HE formation, hESC culture needed to be standardised. As such, a media trail was carried out to define serum free media capable of maintaining hESC in a pluripotent undifferentiated state. We also ensured hESC cultured in the various media could be directly differentiated to HE in a reproducible and efficient manner. The project then focused on the effect of post-translational modifications (PTMs), specifically SUMOylation, in hepatocyte differentiation and its subsequent manipulation to enhance HE viability. SUMOylation is a PTM known to modify a large number of proteins that play a role in various cellular processes including: cell cycle regulation, gene transcription, differentiation and cellular localisation. We hypothesised that SUMO modification may not only regulate hESC self renewal, but also maybe required for efficient hESC differentiation. We therefore interrogated the role of SUMOylation in hESC differentiation to hepatic endoderm (HE). hESC were differentiated and the cellular lysates were analysed by Western blotting for key proteins which modulate the conjugation and de conjugation of SUMO. We demonstrate that peak levels of SUMOylation were detectable in hESC populations and during cellular differentiation to definitive endoderm (DE), day 5. Following commitment to DE we observed a decrease in the level of SUMO modified proteins during cellular specialisation to a hepatic fate, corresponding with an increase in SENP 1, a SUMO deconjugation enzyme. We also detected reduced levels of hepatocyte nuclear factor 4 α (HNF4α), a critical regulator of hepatic status and metabolic function, as SUMOylation decreased. As a result, we investigated if HNF4α was SUMOylated and if this process was involved in modulating HNF4α’s critical role in HE. HNF4α is an important transcription factor involved in liver organogenesis during development and is a key regulator for efficient adult liver metabolic functions. We observed a decreasing pattern of HNF4α expression at day 17 of our differentiation protocol in conjunction with a decrease in SUMO modified proteins. In order to further investigate and validate a role of SUMOylation on HNF4α stability Immunoprecipitation (IP) was employed. HNF4α protein was pulled down and probed for SUMO 2. Results show an increase in the levels of SUMO2 modification as the levels of HNF4α decrease. Through deletion and mutation analysis we demonstrated that SUMO modification of HNF4α was restricted to the C-terminus on lysine 365. Protein degradation via the proteasome was responsible for the decrease in HNF4α, demonstrated by the use of a proteasome 26S inhibitor MG132. Additionally, a group at the University of Dundee has shown that polySUMOylation of promyelocytic leukaemia protein (PML) leads to its subsequent ubiquitination via RNF4, an ubiquitin E3 ligase, driving its degradation. Using an in vitro ubiquitination assay, we show that polySUMOylated HNF4α is preferentially ubiquitinated in the presence of RNF4. Overall polySUMOylation of HNF4α may reduce its stability by driving its degradation, hence regulating protein activity. In conclusion, polySUMOylation of HNF4α is associated with its stability. HNF4α is subsequently important for HE differentiation both driving the formation of the hepatocytes and in maintaining a mature phenotype, in agreement with a number of different laboratories. Creating the ideal environment for sustaining mature functional hepatocytes, primary and those derived from hESCs and iPSCs, is essential for further use in applications such as drug screening, disease modelling and extracorporeal devices.

The UDP-glucuronosyltransferases (UGTs) are a superfamily of enzymes that glucuronidate small, lipophilic molecules, thereby altering their biological activity and excretion. In humans, important examples of UGT substrates include molecules of both endogenous and xenobiotic origin; thus, UGTs are considered essential contributors to homeostatic regulation and an important defence mechanism against chemical insult. In keeping with both roles, UGTs are most strongly expressed in the liver, a predominant organ involved in detoxification. Rates of glucuronidation in humans are neither uniform among individuals, nor constant in an individual over time. Genetic determinants and non-endogenous signals are both known to influence the expression of UGTs, which in turn may affect the efficacy of certain pharmaceutical treatments or alter long-term risk of developing disease. Thus, this thesis focuses on the transcriptional regulation of UGT genes in humans, particularly on mechanisms that are likely to be relevant to their expression and variation in the liver. Two major approaches were used: firstly, extensive studies of several UGT promoters were performed to identify and characterise transcriptional elements that are important for UGT expression; and secondly, important hepatic transcription factors were investigated as potential regulators of UGT genes. UGT1A3, UGT1A4 and UGT1A5 are a subset of highly related, but independently regulated, genes of the human UGT1 subfamily. UGT1A3 and UGT1A4 are expressed in the liver, whereas UGT1A5 is not. The presented analysis of the UGT1A3, UGT1A4 and UGT1A5 proximal promoters demonstrates that a hepatocyte nuclear factor (HNF)1-binding site common to all three promoters is important for UGT1A3 and UGT1A4 promoter activity in vitro, but is insufficient to drive UGT1A5 expression. Two additional elements required for the maximal activity of the UGT1A3 promoter were also identified that may distinguish this gene from UGT1A4. UGT1A3 was investigated further, focusing on mechanisms that may contribute to interindividual variation in UGT1A3 expression. Polymorphisms in the UGT1A3 proximal promoter were identified and their functional consequences tested. Known variants of HNF1alpha were also tested for altered activity towards the UGT1A3 gene. UGT1A9 is the only hepatic member of the UGT1A7-1A10 subgroup of UGT1 enzymes. Previous work had identified HNF1-binding sites in all four genes, and HNF4alpha as an UGT1A9-specific regulator. The work presented herein extends these findings to show that HNF1 factors and HNF4alpha synergistically regulate UGT1A9, and that HNF4alpha is not the only transcription factor responsible for the unique presence of UGT1A9 in the liver. Liver-enriched transcription factors screened as potential UGT regulators were chosen from the HNF1, HNF4, HNF6, FoxA and C/EBP protein families. Functional interactions newly identified by this work were HNF4alpha with UGT1A1 and UGT1A6, HNF6 with UGT1A4 and UGT2B11, FoxA1 and FoxA3 with UGT2B11, UGT2B15 and UGT2B28 and C/EBPalpha with UGT2B17. Observations were also made regarding different patterns of interaction between each UGT and the transcription factors tested, particularly HNF1alpha.

My first project was to investigate the role of Hepatocyte Nuclear Factor 1 (HNF1) and Nuclear Factor I (NFI) on alpha-fetoprotein (AFP) promoter activity during liver development. AFP is highly expressed in the fetal liver, silenced at birth, and remains at very low levels in the adult liver. A GA substitution located at -119 of the human AFP promoter is associated with hereditary persistence of AFP (HPAFP) expression in the adult liver (Hum Molec Genet, 1993, 2:379). The -120 region harbors overlapping binding sites for HNF1 and NFI. While it has been shown that the GA substitution increases HNF1 binding, the role of NFI in AFP regulation has not been investigated. This overlapping HNF1/NFI site is conserved in other mammals, including mice. In this study, I used a combination of biochemical, tissue culture, and animal studies to explore further the role of this HNF1/NFI site in AFP regulation. Transient co-transfections in Hep3B hepatoma cells indicate that HNF1 activates while NFI represses the mouse AFP promoter. EMSAs indicate that HNF1 and NF1 compete for binding to this site. Transgenes regulated by the wild-type AFP promoter are expressed at low levels in the adult liver. Transgenes with a GGAA mutation (similar to the G-A human mutation) are more active in the adult liver. My data indicate that HNF1 and NFI compete for binding to the -120 region of the AFP promoter and this competition is involved in postnatal AFP repression. My second project was to study the control of Elongation of very long chain fatty acids like 3 (Elovl3) in the liver by Zinc fingers and homeoboxes 2 (Zhx2). The Zhx2 gene was originally characterized in our lab based on its ability to control the developmental repression of several hepatic genes, including AFP (PNAS, 102:401). Zhx2 is a member of a small family of proteins found only in vertebrates that also includes Zhx1 and Zhx3. These proteins all contain two zinc fingers and four homeodomains, suggesting that they function as regulators of gene expression. My study shows that Zhx2 regulates Elovl3 expression in female liver. Mouse strain-specific differences in adult liver Elovl3 mRNA levels and transgenic mouse data indicate that Zhx2 activates Elovl3 expression in the female adult liver. I also demonstrate that Elovl3 is repressed in the regenerating liver and that the level of Elovl3 repression is controlled by alpha-fetoprotein regulator 2 (Afr2). In addition, I show that Elovl3 expression is reduced in liver tumors, fibrotic livers and fatty livers, raising the possibility that Elovl3 can serve as a marker for HCC and liver damage.

The overall objective of this research was to understand the mechanims by which growth hormone (GH) regulates insulin-like growth factor-I (IGF-I) gene expression in liver. Previous studies have suggested that GH regulation of IGF-I gene expression in liver is mediated by binding of the transcription factor signal transducer and activator of transcription (STAT) 5 to four binding sites located distantly from the IGF-I promoter. The first specific objective of this research was to determine whether additional STAT5 binding sites were involved in GH stimulation of IGF-I gene expression in liver. Sequence analysis of 170 kb of mouse genomic DNA revealed nineteen consensus STAT5 binding sequences corresponding to fourteen ~200 bp chromosomal regions that were conserved in the corresponding human DNA sequence. Eight of these chromosomal regions were able to mediate STAT5 activation of reporter gene expression in cotransfection experiments. Two of these chromosomal regions corresponded to those previously identified. Gel-shift assays indicated that the eight new STAT5 binding sites and three of the four previously identified STAT5 binding sites could bind GH-activated STAT5 from mouse liver. Together, these results suggest that GH stimulation of IGF-I gene transcription in the mouse liver may be mediated by at least eleven STAT5 binding sites located distantly from the IGF-I promoter. In a previous study, I found that liver expression of liver-enriched transcription factor hepatocyte nuclear factor 3γ (HNF-3γ) was increased by GH in cattle. Therefore, the second specific objective of this research was to determine how GH stimulates HNF-3γ gene expression and whether the increased HNF-3γ mediates GH stimulation of IGF-I gene expression in bovine liver. Sequence analysis of the bovine HNF-3γ promoter revealed the presence of two putative binding sites for STAT5. The proximal putative STAT5 binding site appears to be conserved in other mammals. Chromatin immunoprecipitation (ChIP) assays demonstrated that GH increased the binding of STAT5 to the HNF-3γ promoter in bovine liver and that this binding was associated with increased HNF-3γ expression. Gel-shift assays demonstrated that the proximal STAT5 binding site in the HNF-3γ promoter could bind GH-activated STAT5 from bovine liver. Cotransfection analyses showed that the proximal STAT5 binding site was necessary for the HNF-3γ promoter to be activated by GH. The promoter of the bovine IGF-I gene contains three putative HNF-3 binding sites that seem to be evolutionarily conserved. ChIP assays indicated that GH stimulated the binding of HNF-3γ to the IGF-I promoter in bovine liver. Gel-shift assays showed that one of the putative HNF-3 binding sites could bind HNF-3γ protein from bovine liver. Co-transfection analyses demonstrated that this HNF-3 binding site was necessary for HNF-3γ activation of reporter gene expression from the IGF-I promoter. In summary, the results of this dissertation research suggest that GH-activated STAT5 directly stimulates IGF-I gene transcription in liver by binding to at least eleven distantly located STAT5 binding sites in the IGF-I locus and indirectly stimulates IGF-I gene transcription by enhancing HNF-3γ gene expression in the liver. / Ph. D.

Abstract Cytochrome P450 (CYP) enzymes are the major metabolizers of xenobiotics, e.g. drugs, and environmental toxins. Thus, changes in CYP expression have an important impact on drug metabolism and susceptibility to chemical toxicity. In the present study, the transcriptional mechanisms of both constitutive and inducible regulation of the Cyp2a5 gene in mouse liver were investigated. Mouse primary hepatocyte cultures were used as the main model system together with cell and molecular biology methods. The key activation regions of the Cyp2a5 5' promoter were determined using reporter gene assays. Two major transcription activation sites of the Cyp2a5 5' promoter, called the proximal and the distal, were found. Transcription factors hepatocyte nuclear factor-4 (HNF-4) and nuclear factor I were shown to bind to the proximal promoter. Aryl hydrocarbon receptor nuclear translocator (ARNT) and upstream stimulatory factor bound to a common palindromic E-box element in the distal promoter region. All three response elements were shown to be essential for constitutive expression of CYP2A5 in murine hepatocytes. ARNT appeared to control Cyp2a5 transcription without a heterodimerization partner suggesting active involvement of the ARNT homodimer in mammalian gene regulation. Aryl hydrocarbon receptor (AHR) ligands were shown to induce Cyp2a5 transcriptionally by an AHR-dependent mechanism, and established Cyp2a5 as a novel AHR-regulated gene. The AHR response element and the E-box, identified in these studies, were located near to each other and close to a separately defined nuclear factor (erythroid-derived 2)-like 2 binding site in the distal region of the Cyp2a5 promoter, suggesting cooperation between these elements. Peroxisome proliferator-activated receptor gamma coactivator-1α was shown to up-regulate Cyp2a5 transcription through coactivation of HNF-4α. This indicates that xenobiotic metabolism can be regulated by modification of co-activation. The present results show that CYP2A5 is regulated by several different cross-regulatory pathways. The regulatory mechanisms involved in the transcription of the Cyp2a5 gene may also control other CYP genes, especially the human ortholog CYP2A6, and may explain some of the individual variations in the metabolism of xenobiotics.