Although astrocytomas account for 40% of brain tumours in children, little is known about the genetics of these paediatric tumours. Indeed, 85% of low-grade (WHO grade I and II) and 50% of high-grade (WHO grade III and IV) paediatric astrocytoma have a normal karyotype. The aim of this study was to identify non-random genetic aberrations in different grades of paediatric astrocytoma at both the genomic and expression levels. Affymetrix genechip technology and array comparative genomic hybridisation (aCGH) have been used to generate gene expression profiles of 35 paediatric astrocytoma short-term cell cultures of all grades and 19 pilocytic (PA, grade I) biopsies and identify copy number alterations (CNA) in 32 paediatric astrocytoma short-term cell cultures of all grades and 11 pilocytic biopsies. The PAI biopsy samples have a distinct expression profile compared to normal brain with 1844 genes being differentially expressed in all samples. The KEGG pathway most influenced by these genes is antigen processing and presentation, with the majority of genes being up- regulated. Addition pathways altered include PI3K signalling and MAPK signalling. Only single clone CNAs were detected in PAI including alterations at Ip36.32-p36.3, 14ql2 and 22q33.33 which were lost or gained in the majority of samples. The clone at 14ql2 is located in a region of large-scale copy number alteration (LCV). Alterations at this site have been linked to increased risk of paediatric solid tumour development. No known genes are located within the clone site. However, FOXG1B is adjacent to the clone region and is significantly down-regulated in PAI compared to normal brain. Hierarchical clustering of the short-term cultures according to expression profile similarity demonstrated that paediatric astrocytoma can be grouped into low and high-grade tumours by molecular signature. Furthermore, approximately half of paediatric glioblastoma multiforme (GBMIV, grade IV) clustered with 7 adult GBMIV cultures, suggesting that some paediatric GBMIV are genotypically similar to those arising in adults. KEGG pathways influenced by differential gene expression include Wnt signalling and the cell cycle pathway, with the finding that same pathways are being disrupted to varying extents in low and high-grade paediatric astrocytoma. The frequency of CNAs was similar to those previously reported, with gain of all or part of chromosome 7 as the most common alteration. Correlations between gene expression and CNAs were also identified in the short-term cultures.

Deletion at ITPR1 underlies a young onset autosomal recessive ataxia in mice and a late onset autosomal dominant ataxia (SCA15) in humans. Data presented show the utility of investigating spontaneous mouse mutations in understanding human disease. Through linkage and sequence analysis a novel mutation in the gene encoding inositol 1,4,5-triphosphate receptor type 1 was identified to underlie a severe movement disorder in mice. The 18bp in frame deletion in Itpr1 exon 36 was shown to be allelic to that of another model, opisthotonos (Lane 1972). The Itpr1Δ18 mutation leads to a decreased to almost total lack in the normally high level of ITPR1 expression in cerebellar Purkinje cells. In addition, high density genome wide SNP genotype data in humans showed a SUMF1-ITPR1 deletion to segregate with spinocerebellar ataxia 15 (SCA15), a late-onset autosomal dominant disorder, which was previously mapped to the genomic region containing ITPR1; however, no causal mutations had been identified (Knight et al. 2003). With this deletion not observed in a control population, decreased ITPR1 protein levels in individuals carrying the deletion, and subsequent identification of similar deletions in additional spinocerebellar ataxia families, the data provide compelling evidence that heterozygous deletion in ITPR1 underlies SCA15. As demonstrated, high density genome wide SNP analysis can facilitate rapid detection of structural genomic mutations that may underlie disease when standard sequencing approaches are insufficient. The data suggest genetic alterations at ITPR1 underlie approximately over 1% of autosomal dominant SCA type III (ADCA III) cases for which currently no genetic cause has been identified. Data described herein add weight to a role for aberrant intracellular Ca2+ signaling in Purkinje cells in the pathogenesis of spinocerebellar ataxia.

Primary Ciliary Dyskinesia (PCD) is a recessively inherited disorder caused by cilia and sperm flagella dysmotility associated with axoneme ultrastructural defects. Symptoms include recurrent respiratory tract infections, sinusitis, bronchiectasis, subfertility, and laterality defects due to defective embryonic nodal cilia. PCD is genetically heterogeneous and two genes, DNAI1 and DNAH5, account for 38% of cases. To identify new PCD genes, genome wide linkage screens were undertaken in consanguineous PCD families using homozygosity mapping: (1) five Pakistani families with missing inner and outer dynein arms (2) two Arabic families with central pair agenesis and no dextrocardia. Three disease loci were mapped on chromosome llq23.3-24.3 and 17q21.31-22 (Pakistani), and chromosome 6p21.31-21.1 (Arabic), with peak multipoint LOD scores of 3.6, 6.0 and 6.7, respectively. Comparative bioinformatic analysis identified 7 positional candidate genes which were subjected to mutational analysis. A 3 bp deletion in C6ORF206 at 6p21.31-21.1 was revealed in affected Arabic PCD individuals, resulting in a predicted in-frame loss of a C-terminal lysine residue, K268. The protein encoded by C6ORF206 was identified as homologous to the Chlamydomonas reinhardtii radial spoke head protein, RSP9. Functional work was undertaken to determine if the K268del mutation was pathogenic. RSP9 is mutated in the paralysed flagella Chlamydomonas mutant, pfl 7, and transformation of pfl 7 with wild-type RSP9 rescued motility. However transformation of pfl 7 with RSP9 R261del (equivalent to human K268del) rescued motility to a lower level, resulting in an ineffective swimming stroke. Expression of RSP9 was investigated by in situ hybridisation, and zebrafish RSP9 knock-down morphants were created to investigate its role in vertebrate ciliary function. Although expressed at the vertebrate embryonic node, knockdown of RSP9 function did not affect laterality in zebrafish, however it did cause dysfunction of the nasal cilia. This data suggests that C6ORF206/RSP9 functions as a ciliary protein in ciliated organisms and that the K268del mutation likely causes PCD without situs inversus.

Bipolar disorder is a common and serious mental illness. The occurrence of mania is central to the diagnosis, but affected individuals typically also suffer episodes of depression. The results of family, twin and adoption studies argue convincingly for genetic susceptibility to bipolar disorder. Linkage studies conducted at the Molecular Psychiatry Laboratory, UCL have previously implicated the regions 12q24, 21q22, lq42 and 11 pi4- 15 as harbouring susceptibly loci for bipolar disorder. In this thesis I report fine mapping of the 12q24, 21q22 and lq42 regions by linkage disequilibrium methods, employing a case-control design. For the llpl4-15 region association with the candidate gene BDNF was tested. I also present attempts to replicate findings of association at the genes DAOA and COMT, located in regions implicated by meta-analysis of the linkage data. I have attempted to put these investigations in context, necessitating consideration of the conceptual developmental of bipolar disorder, the classical techniques for assessing the genetic contribution to aetiology, and mapping strategies. Fine mapping of the UCL linkage regions implicated two novel susceptibility loci and provided support for two previously identified loci. Association of multiple markers within a 180 kb region of 12q24.3 was found, implicating Slynar and LOC387895. Association was also found with two markers in the more centromeric gene P2RX7, previously implicated in a Canadian sample. Multiple associated markers were found on 21q22.3. Two candidate genes - C21orf29 and TRPM2 - were identified from this region. Initial efforts to fine map the lq42 region suggested the involvement of the previously implicated DISCI gene. However association was only found with a single marker. Although haplotypic association was found with BDNF, the complex structure of the microsatellite marker hindered interpretation of the results. Partial replication of the association with DAOA was achieved but the involvement of COMT was not supported.

CHARGE and DiGeorge (DGS) are multiple malformation haploinsufficiency syndromes that show significant phenotypic overlap, especially when considering derivatives of the pharyngeal apparatus. Patients exhibit characteristic cardiovascular malformations including anomalies associated with defective patterning of the pharyngeal arch arteries (paa). CHD7, on chromosome 8q12.1, was recently identified as a candidate gene for CHANGE and is mutated in ~60% of patients. Most DGS patients carry a 3Mb heterozygous deletion of chromosome 22q11, encompassing ~30 genes including TBX1. Animal models of DGS have demonstrated the importance of Tbx1 during pharyngeal and heart development. The critical role of TBX1 in DGS is supported by the identification of TBX1 mutations in DGS patients without 22q11 deletions. CHARGE and DGS exhibit overlapping genetic etiologies. Patients diagnosed with CHARGE can have 22q11 deletions, while complete DGS can be found in patients with CHD7 mutations. However, despite the genetic overlap between the two human conditions, work presented here failed to identify CHD7 mutations in a cohort of non-deleted patients. Results presented in this thesis show that insertion of a gene trap cassette, disrupting CHD7, produced a mouse model of CHARGE consistent with the human condition. Chd7^{+/-} embryos exhibited heart defects resulting from aberration of the paa, including interrupted aortic arch type-B. Chd7 is only the second gene reported, after TBx1, heterozygous mutations of which cause fourth paa defects. Furthermore, the phenotype of Chd7^{+/-} animals demonstrated epistasis between the two genes during formation of the fourth paa and thymus. Contrary to hypotheses suggesting CHARGE is the result of genetic deficiencies in neural crest cells (NCC), conditional Wnt1-Cre driven restoration of Chd7 expression in NCC, did not prevent paa defects. AP2 \propto –Cre driven restoration of Chd7 expression in the ectoderm and NCC simultaneously, rescued the paa defects, identifying a requirements for Chd7 expression in the ectoderm during development of these vessels.

Bardet-Biedl Syndrome (BBS) is a rare genetic disease comprising obesity, retinal dystrophy, Polydactyly and renal abnormalities. When this study began there were six known loci for BBS, only one of which, BBS6, had been identified. The initial aims of the project were to find BBS genes by screening of candidate genes within the known intervals and performing a genome wide screen using previously unlinked patients. During the time of this project three of the known BBS loci {BBSI, 2 and 4) have been identified and a further two novel genes (BBS7 and 8) have also been cloned. BBS was previously thought to be a recessive disease, but through mutation screening of all BBS genes in our cohort of 160 patients we have found evidence for complex inheritance involving the requirement for three mutations a homozygous mutation in one BBS gene and a further heterozygous mutation at a second BBS locus, to manifest disease in some families. As a quick and cheap alternative to sequencing for mutation detection, I developed the technique of multiplex capillary heteroduplex analysis (MCHA) which is now in routine use for screening new BBS cases. To determine the possible function of the BBS4 protein, a yeast-two-hybrid screen was undertaken to identify interactors of BBS4. Pericentriolar Material 1 (PCM1), one of the potential interactors, also co-localises with the protein at the centriolar satellites of centrosomes and basal bodies of primary cilia. The BBS8 protein, which shares homology with BBS4 also localises to the basal body and, from immunohistochemistry experiments using the BBS8 antibody, has been shown to be expressed in ciliated tissues such as kidney, retina, brain and spermatids, implicating basal body dysfunction in the aetiology of BBS. During the course of my studies, we have progressed from a handful of genetic loci to the determination of the putative protein function which may pave the way to the design of future therapies.

Mutation analysis of the recognized ARVC genes and of further candidate genes was performed on a large cohort of ARVC patients. Several novel mutations were identified and three further desmosomal genes were linked to the disease: plakophilin2, desmocollin2 and desmoglein2. Heart and skin samples from ARVC patients were subjected to microscopic examination and immunohistochemistry to study the effect of the newly-identified mutations on the structure of cell adhesion complexes.;The functional effects of a particular novel mutation were thoroughly examined in vitro. S39_K40insS is the first dominant ARVC-causing plakoglobin mutation to be reported. Yeast-two hybrid analysis was used to investigate the effect of S39_K40insS on the proteins interactions established by plakoglobin. A HEK293 cell line stably expressing the mutant protein was generated and used to study the effects of S39_K40insS on desmosomal structure, cell proliferation, cell death, subcellular localization and expression levels of proteins involved in adhesion and signalling and cellular responses to defined mechanical load. A recombinant adenovirus expressing the mutant protein was generated and used to transfect neonatal rat ventricular cardiomyocytes, whose behaviour and responses were subsequently analysed. The functional consequences of S39_K40insS were compared with those of PK215del2, a previously reported recessive plakoglobin mutation known to underlie Naxos disease, a syndromic form of ARVC.;These results point towards novel mechanisms of disease pathogenesis, that apart from weakened cell-cell adhesion involve altered protein turnover kinetics and defects in signalling pathways. Similar studies should improve our understanding of ARVC and provide a more accurate diagnostic algorithm.

Acute leukaemia is the most common form of childhood cancer and is the primary cause of cancer-related mortality in children. The major initiating event in infant leukaemia is rearrangement of the Mil gene at 1 lq23, and such leukaemias have a poor survival rate. Chromosomal translocations occur commonly in leukaemia and an in-frame fusion gene is created with altered properties to the original genes involved. Translocations of chromosome 1lq23 result in the fusion of Mil to a variety of partners genes. Mll-Afl results from t(911), Mll-A/4 from t(411) and Mll-Enl from t(1119) and are thought to cause aberrant transcriptional regulation. To identify candidate target genes of t(1119), MLL-ENL has been over-expressed in 32D cells, an immortalised myeloid progenitor cell line. The actions of MLL-ENL on cell survival following growth factor withdrawal, and on G-CSF mediated differentiation have been studied. Potential transcriptional targets of MLL-AF9 and MLL-ENL have been identified using microarray technology. Ten genes were identified as possible candidate target genes of the MLL-ENL protein that might be involved in leukaemogenesis. Of particular interest were Gatal and HTm4 which have a role in haematopoiesis and cell cycle regulation. shRNA constructs were generated to knock down MLL-ENL and allow validation of the potential target genes. The mechanism of leukaemic transformation by MLL-ENL has also been investigated. In leukaemia, fusion proteins may contribute to haematological malignancy in various ways, though the main routes are thought to be either a gain-of-function or dominant negative action. In order to determine the mechanism by which a MLL fusion protein immortalizes murine haematopoietic progenitor cells (HPC), a system was developed where MLL-ENL was expressed in murine HPCs in the absence of both alleles of Mil. The experiments were carried out using murine HPCs from conditional Mil knockout mice generated in our laboratory where the wild- type Mil allele is flanked by LoxP sites at exons 9 and 10, and can be inactivated following Cre-mediated recombination. The ability of cells to undergo sequential re-plating in methylcellulose was used as an indication of immortalization. These experiments now definitively show that the MLL-ENL fusion protein acts in a gain-of-function manner.

Bipolar disorder (BP) and schizophrenia are severe neuropsychiatric conditions that are among the leading causes of morbidity and chronic disability world-wide. Both conditions are characterised by a substantial genetic heterogeneity, which has complicated the search for susceptibility loci. One strategy to tackle this difficulty lies in the study of population isolates that are characterised by a reduced genetic heterogeneity. In this thesis, I have therefore conducted genetic studies of BP and schizophrenia in the well-characterised South American population isolate of Antioquia, Colombia. Our group has recently reported the results of a linkage scan of six Antioquian families segregating severe BP. Here, I performed a follow-up study of a candidate region on chromosome 5q33. I sequenced the CLINT 1 gene, a functional candidate that has also been implicated in schizophrenia, in affecteds from four BP pedigrees from the original linkage study and identified three single base pair variants, all of which had been previously described. A transmission distortion test of one of these variants, rs 11955293, in a sample of 176 unrelated BP patients from Antioquia and their parents found no evidence of association with BP. Although these results do not rule out a minor effect of the CLINT1 gene on susceptibility to the disorder in Antioquia, other loci are likely to be of greater significance. This includes other genes on chromosome 5q33, but also other candidate regions in the genome. To further explore the latter possibility, I conducted a whole-genome linkage scan in an additional nine pedigrees with severe BP from Antioquia and analysed the obtained genotype data jointly with that of the initial linkage scan. Using parametric and non-parametric linkage approaches, I explored three different diagnostic models: a narrow model including only BP type I (BPI) as affected a model including BPI and II and major unipolar depression and a third model including only individuals who had experienced psychosis as affected. This second linkage scan found evidence for a number of candidate regions, including chromosome 13q33 for BPI, chromosomes lpl3-31 and lq25-31 for mood disorders, chromosome 12ct-ql4 for mood disorders, and chromosomes 2q24-31 and 16pl2 for psychosis. Encouragingly, many of these loci had previously been pinpointed as BP susceptibility loci in other populations on the other hand, we also identified a novel locus on chromosome 12q. While the use of population isolates can help decrease the genetic heterogeneity of a complex disease, complementary strategies can be used to reduce this heterogeneity even further. In studying the NOS1AP gene, a functional candidate on chromosome lq23 that is involved in glutamatergic neurotransmission, in a sample of 102 unrelated Antioquian schizophrenia patients and their parents, I have therefore used both categorical and dimensional approaches to the disease phenotype. In the categorical approach, I conducted an analysis for association between the NOS1AP gene and DSM-IV schizophrenia by TDT. For the dimensional approach, two clinical scales measuring positive and negative symptoms, SANS and SAPS, were applied to all patients and dimensional scores were obtained from these scales by factors analysis. I then performed quantitative TDT analysis of the dimensional scores. My analyses found association to both DSM-IV schizophrenia and a clinical dimension capturing negative symptoms, in line with a role of NOS1AP in glutamatergic neurotransmission. The results of these analyses also underline the usefulness of a dimensional approach in psychiatric genetics.

This thesis describes the identification and characterisation of genes important in development of the pharyngeal apparatus and heart, the major structures affected in DiGeorge syndrome (DGS). DGS is characterised by craniofacial, cardiovascular, thymus and parathyroid defects. It is most commonly caused by heterozygous deletion of a 3Mb region of chromosome 22ql 1 encompassing at least 30 genes. Haploinsufficiency of the TBX1 transcription factor is considered to be the major underlying cause of this syndrome. Animal models of DiGeorge syndrome have demonstrated the importance of Tbxl in pharyngeal and heart development and therefore, identifying the downstream targets of Tbxl is of vital importance in understanding the development of these systems. This project was aimed at identifying cell autonomous effects of Tbxl by isolating Tbxl-lacZ expressing cells and comparing the gene expression profiles of Tbxl null and heterozygous cells by microarray analysis. Validation of the downregulated gene, Hesl has been further investigated by the characterisation of pharyngeal and heart defects in mice carrying null alleles of this potential Tbxl target. In addition, BAC recombineering was also conducted to generate a transgenic mouse carrying a GFP-labelled, reversible, mutant Tbxl allele. This modified Tbxl allele should provide the basis for further enrichment of Tbxl targets by allowing isolation of Tbxl-expressing cells from transgenic mice and subsequent restoration of Tbxl function by cre-mediated recombination. Furthermore, in order to identify other pathways important in heart development, DGS patients with atypical chromosome rearrangements were analysed. This approach led to the identification of a potentially disrupted gene, HIC2, whose function was analysed using gene trap mouse models and which was shown to play a role in heart development. Overall these experiments have led to the elucidation of novel genes and genetic pathways affected in DGS and have contributed to a better understanding of the mechanisms controlling morphogenesis of the pharyngeal apparatus and heart.