Fragile X Syndrome (FXS) is a neurodevelopmental disorder and the most common heritable single gene cause of Autism Spectrum Disorder (ASD). The Fragile X (FMR1-KO) mouse model has been used to understand the pathophysiology of the disease. However, the majority of studies have been done in adult mice and early life outcomes have yet to be explored. Therefore, in order to contribute to the knowledge of the neurodevelopmental processes associated with brain disorders, this thesis examines postnatal outcomes in the Fragile X Mouse Model: early life behaviours, the developmental trajectory of a set of ASD risk genes, and neuroanatomical phenotype. The first study examined ultrasonic vocalizations in pups and showed a transient increase in calls in FMR1-KO mice. To understand the relationship between early life behaviours, the second study examined outcomes in the pre-pubertal period in these mice when challenged with lipopolysaccharide and maternal separation. The results showed genotype and treatment interactions affecting sexually dimorphic behavioural outcomes and developmental milestones. In the third study, possible underpinnings of behavioural differences were explored by examining mRNA expression of the neuroligins and neurexins. In FMR1-KO mice, changes were transient and sex-specific, suggesting these as molecular effectors in the disease. Lastly, using structural brain imaging, the fourth study examined regional volume differences that may be related to behavioural differences. Differences in regions affected in FXS patients were observed and genetic background was shown to affect the neuroanatomical phenotype. Overall, this thesis demonstrates that the FXS model recapitulates some outcomes in other ASD mouse models and shows that this single gene has multiple interactions with sex, strain, and postnatal challenge which manifests at specific ages at molecular, brain structure and behavioural levels. This work contributes to the efforts elucidating the neurobiology of ASD and reverse translation approaches to identify therapeutic targets for neurodevelopment disorders. / Dissertation / Doctor of Philosophy (PhD) / Autism spectrum disorder (ASD) is a diagnosis based on observed behaviours: impaired communication and repetitive actions. However, there are genetic and other behavioural differences in ASD patients that are not shared among the group. It is important to tease apart this group since current treatments for ASD do not target the biological problems or the core impairments. This thesis focuses on Fragile X Syndrome, the leading genetic condition that results in ASD in order to understand the biological basis of ASD. Using a mouse model, compared to healthy mice, these studies report changes in behaviours, in the size of different brain regions, and in molecules involved in connecting brain cells during development. These findings shed light on the molecular story underlying ASD. By understanding the nature of influences on the developing brain, the type and timing of interventions can be designed to keep the brain on a healthy trajectory.

Mdm2 and Mdm4 are critical negative regulators of the p53 tumor suppressor. Mdm4-null mutants are severely anemic and exhibit impaired proliferation of the fetal liver erythroid lineage cells. This phenotype may indicate a cell-intrinsic function of Mdm4 in erythropoiesis. In contrast, red blood cell count was nearly normal in mice engineered to express low levels of Mdm2, suggesting that Mdm2 might be dispensable for red cell production. In the first part of the thesis, we further explore the tissue-specific functions of Mdm2 and Mdm4 in the erythroid lineage by crossing the conditional Mdm4 and Mdm2 alleles to an erythroid-specific-cre (EpoRGFP-Cre ) knock-in allele. Our data show that Mdm2 is required for rescuing erythroid progenitors from p53-mediated apoptosis during primitive erythropoiesis. In contrast, Mdm4 is only required for the high erythropoietic rate during embryonic definitive erythropoiesis. Thus, in this particular cellular context, interestingly, Mdm4 only contributes to p53 regulation at a specific phase of the differientation program. Moreover, a large body of evidence indicates that aberrant expression of either MDM2 or MDM4 impairs p53 tumor suppression function and consequently favors tumor formation. Overexpression of MDM2 was observed in 10% of 8000 human cancers from various sites, including lung or stomach, and MDM4 was found amplified and/or overexpressed in 10-20% of over 800 diverse tumors including lung, colon, stomach and breast cancers. Remarkably, selective MDM4 amplification occurs in about 65% of human retinoblastomas. In contrast, MDM2 amplifications are relatively rare (about 5%) in retinoblastomas, indicating that depending on the tumor context (cell type, initiating oncogene, …), MDM4, rather than MDM2, overexpression might be selected for as a more efficient mean of suppression of p53 function. As part of a large effort to better understand why different cell types require distinct combinations of mutations to form tumours, we will examine the molecular basis for selective up-regulation of Mdm4 in retinoblastomas. In this context, we have successfully generated 2 conditional transgenic mouse lines expressing either mycMdm2 or mycMdm4 driven by the PCAGGs promoters in the ROSA26 locus. Since a cassette containing a floxed transcriptional stop element is inserted upstream of the transgenes, we can achieve tissue-specific expression and spatio-temporal regulation of the transgenes by using different Cre and CreER. By the use of N-terminal myc-tag fused with the transgenes, we are able to compare the expression levels of the transgenes. Finally, due to C-terminal IRES-GFP element, we can easily identify transgene expressing cells. One of our aims is to use this Mdm4 conditional transgenic mouse line as the first, non-chimeric, mouse model of retinoblastoma that can be used as an appropriate preclinical model to improve treatment of this disease.

Colorectal cancer (CRC) is the third most common cancer globally with around 1.3 million cases diagnosed annually. In cases of inherited CRC where none of the rare, high-risk mutations associated with familial syndromes are observed it has been theorised that the heritable risk is due to common, low-risk variants in the genome. Identifying these variants of modest or small effect size has become possible due to the use of large genome-wide association studies (GWAS). Through a meta-analysis of five previous GWAS, Dunlop et al. identified three novel susceptibility loci at 6p21, 11q13.4 and Xp22.2. The aims of this study were to further characterise the Xp22.2 risk locus and investigate the function of the putative risk gene SHROOM2. The variant originally identified in the Dunlop et al. study as showing an association with CRC risk rs5934683 is located between SHROOM2 and GPR143. Genome–wide expression analysis has shown that the variant is an eQTL (expression quantitative locus), affecting expression of SHROOM2, but not GPR143. An important caveat to this analysis was that the commercial array used to measure gene expression does not detect all predicted GPR143 transcripts. Hence it was important to understand whether GPR143 might be involved at the locus and whether there was altered expression in normal colonic mucosa. I have analysed the expression of GPR143 and shown that it is poorly expressed in normal mucosa and that expression of the alternative transcripts is rare. This provided further evidence for the gene of interest at Xp22.2 being SHROOM2 and thus became the focus for further investigation. In order to understand SHROOM2 function a knockout mouse was generated allowing studies of gene function beyond the previously used in vitro systems. Embryonic stem cells containing a Shroom2 knockout first allele were obtained from the International Knockout Mouse Consortium and a novel mouse line established. This mouse line was found to be an incomplete knockout and a Cre recombination strategy was employed to remove the critical exon and create a true null allele for Shroom2. This model was validated as being a true knockout of Shroom2 at both the RNA and protein level and the model subjected to initial phenotyping focusing on tissues where the gene has previously been identified as expressed. To investigate the role of Shroom2 as a CRC susceptibility gene preliminary data has been gathered from crosses to the ApcMin/+ CRC model, and analysis of the intestines of the Shroom2KO line has been undertaken. Two spontaneously occurring anorectal adenomas have been identified in Shroom2 null mice, and an additional mid-colonic polyp phenotype identified when crossed onto the ApcMin/+ background. Additionally, embryonic fibroblasts have been used in growth and wound healing assays to determine what effect total loss of Shroom2 has at a cellular level. Proteomics analysis to identify significantly altered pathways associated with Shroom2 loss has also been carried out and has highlighted a number of interesting targets for further investigation. In summary, a novel Shroom2 knockout mouse model has been developed to investigate the CRC susceptibility locus identified at Xp22.2. Preliminary data from this mouse model appears to confirm SHROOM2 as having a role in tumour development in the large intestine.

Glucocorticoids - cortisol and corticosterone - are steroid hormones synthesised in the adrenal gland that are important mediators of the stress response. Glucocorticoids are also vital in development to aid in organ maturation. Endogenous glucocorticoid levels rapidly rise before birth in all mammals to promote fetal organ maturation. Because preterm birth occurs before this natural rise in glucocorticoid levels, pregnant women at risk of preterm delivery are administered synthetic glucocorticoids to mature the fetal lung and aid neonatal survival. Mice that globally lack the glucocorticoid receptor (GR) die at birth, attributed to lung immaturity. Effects on tissues other than the lung remain less well characterised. Previous work has shown endogenous glucocorticoid action is also essential to mature the mouse fetal heart. Mice globally lacking GR have small, functionally and structurally immature hearts. Mice with tissue-specific deletion of GR in cardiomyocytes and vascular smooth muscle cells (SMGRKO mice; generated using Sm22α-Cre) also have an increased risk of death around the time of birth, suggesting that glucocorticoid maturation of the cardiovascular system is important for neonatal survival. GR expression within the fetal mouse heart initiates at E10.5 but GR in the myocardium is not activated and localised to the nucleus until E15.5. This suggests that mice can respond to glucocorticoid from E10.5. Here, it was hypothesised that antenatal glucocorticoid exposure, prior to the increase in endogenous glucocorticoid levels, would advance fetal heart maturation and this will depend on cardiovascular GR. To investigate the effects of antenatal glucocorticoid treatment on fetal heart maturation in mid-gestation and identify effects mediated by GR, mice with a conditional deletion of GR in cardiomyocytes and vascular smooth muscle cells were studied (SMGRKO mice). Pregnant mice received dexamethasone (dex) in the drinking water from E12.5-E15.5. Levels of Fkbp5 mRNA (a marker of glucocorticoid action) were unchanged between control and SMGRKO mice at E15.5 or following dex treatment. This suggested a lack of response to dex treatment. However, liquid chromatography mass spectrometry measurement confirmed the presence of dex and its active metabolite 6- hydroxydexamethasone (6OHDex) in livers of E15.5 fetuses from dex treated dams (fetal: Dex 0.46 ± 0.1 ng/g, 6OHDex 13.6 ± 0.35 ng/g; dam: Dex 7.96 ± 3.65 ng/g, 6OHDex 4.75 ± 1.2 ng/g). Livers of fetuses exposed to dex had lower levels of the naturally occurring active glucocorticoid, corticosterone, compared to vehicle treated fetuses. This suggests HPA axis suppression in dex exposed fetuses. Maternal liver showed no significant difference in corticosterone levels between dex and vehicle treated mice, suggesting that whilst dex suppressed the HPA axis in fetuses, it did not in the dams. To determine any persistent effects of early antenatal dex treatment on fetal heart, a later time point in gestation, E17.5, was also assessed. At E17.5, 2-days following cessation of dex treatment, dex and its metabolites were undetectable in the fetal and maternal liver. However, corticosterone levels remained reduced in fetal liver at E17.5 in dex exposed animals (vehicle treated: 4.31 ± 0.47 ng/g, Dex treated: 1.72 ± 0.42 ng/g, p < 0.01), whilst levels in the dam liver did not differ from vehicle treated controls. This suggests prolonged HPA axis suppression following dex treatment, which reduced the natural late-gestation rise in glucocorticoids required for fetal organ maturation. To determine whether early antenatal dex treatment could advance fetal heart function, Doppler imaging with a Vevo 770 high frequency ultrasound imager was used. Isovolumetric contraction time, isovolumetric relaxation time and ejection time of the left ventricle were unaltered by dex treatment. However, at E15.5 the mitral deceleration index (MDI), a measure of diastolic function that takes into account loading conditions, was 1.5 fold lower in vehicle treated SMGRKO mice than control (Cre-) littermates (p < 0.05). This reduction in SMGRKO mice suggests glucocorticoids are required within the fetal cardiomyocytes and/or vascular smooth muscle cells to mature the diastolic function of the fetal heart. Dex exposure had no effect on MDI in SMGRKO fetuses, but reduced the MDI by 1.5 fold in control mice to similar levels as in SMGRKO mice (p < 0.05). RNA analysis revealed a trend (p=0.09) for reduced levels of Nr3c1 mRNA (encoding GR) in hearts of E15.5 control (Cre-) fetuses following dex treatment. Although this requires confirmation at the level of GR protein, this finding together with the lack of induction of the GR target, Fkbp5, suggests dex may cause glucocorticoid resistance through down-regulation of GR. At E17.5, 2-days following cessation of dex there were no changes in systolic parameters and the reduction in MDI found at E15.5, following dex, had normalised. Litter size was reduced (close to a 50% reduction) at E17.5 in dex treated mice. This was similar between SMGRKO and control fetuses. The cause of death was not established, but potentially could be due to the reduction in the natural rise in glucocorticoids at E17.5, previously shown to be important for fetal heart maturation. It is therefore possible that mice with more immature hearts may die before reaching E17.5. RNA analysis was undertaken to determine any mechanistic alterations following dex treatment, which could support fetal heart functional alterations found at E15.5. In contrast to expectation, dex also decreased expression of mRNA encoding the calcium handling proteins SERCA2a, NCX1, and CaV1.2 in E15.5 fetal mouse hearts in both control and SMGRKO mice (p < 0.05), compared with the respective vehicle treated mice. These proteins had previously shown to be induced by glucocorticoid action in cardiomyocytes. However, the similar down-regulation in both genotypes indicates this effect is not dependent on GR in cardiomyocytes. Lowered SERCA2a activity following dex treatment could contribute to the changes in MDI observed in control mice. Similarly, Scnn1a and Kcnj12 mRNA levels, previously found to be induced by glucocorticoids in cardiomyocytes, were down-regulated in the E15.5 fetal heart in vivo following dex. Collectively, these data are consistent with glucocorticoid resistance or down-regulation of glucocorticoid action in E15.5 fetal hearts following dex administration. Mutations in KCNJ12 are associated with long QT syndrome, which is characterised by a delayed repolarisation of the heart following each contraction. An altered relaxation of the fetal heart found in control mice following dex could therefore be due to a prolongation of the cardiac action potential, particularly with a delayed repolarisation, because of lower Kcnj12 expression. At E17.5, there were no significant differences in expression of calcium handling genes or ion channel mRNAs between genotypes or following earlier dex exposure. Thus, effects of dex on mRNA expression level may not persist, which could account for the lack of functional changes observed 2-days following cessation of treatment. Because effects seen in vivo with dex treatment were contrary to those predicted, and to further investigate the effect of dex upon calcium content, an in vitro model of primary fetal E15.5 cardiomyocytes was used. Cardiomyocytes were treated with dex for 24 hours and effects on membrane potential voltage changes and calcium transients measured. Following dex, isolated fetal cardiomyocytes showed an elongated repolarisation phase of the action potential (untreated: 120.45 ± 13.81 ms, Dex: 142.34 ± 12.97 ms, p < 0.01), and duration of calcium transients (untreated: 103.31 ± 13.78 ms, Dex: 120.43 ± 23.36 ms, p < 0.05). This assessment of fetal cardiomyocytes was preliminary work to aid in the understanding of mechanisms of fetal heart functional alterations associated with glucocorticoid regulation. The results suggest glucocorticoids may be important in regulating calcium levels. In summary, dex treatment in mice from E12.5-E15.5 did not advance fetal heart maturation. It reduced litter size at E17.5, irrespective of whether GR was expressed in cardiomyocytes or not. The normal late-gestation increase in endogenous glucocorticoid levels in the fetus was reduced by dex, even after treatment finished. / The suppression of corticosterone levels following antenatal dex may reduce maturation of the heart at E15.5 and could be responsible for the reduction in litter size. Downregulation of GR in the fetal heart, may be a mechanism that results in glucocorticoid resistance following antenatal dex treatment, which could explain the lack of beneficial effects of antenatal dex upon fetal heart maturation in these experiments in mice.

Human eyes with exfoliation syndrome (XFS) exhibit a distinctive pattern of iris transillumination defects that are recapitulated in Lyst mutant mice carrying the beige allele. Here I present the identification and characterization of the B6-Lystbg-J mouse model of XFS, modifiers of Lyst mediated ocular phenotypes, mechanisms of intraocular pressure (IOP) pathology related to circadian rhythms, and mechanisms of iris transillumination defects in the B6-Lystbg-J mice. Clinical and histological analysis shows that the B6-Lystbg-J mice have multiple similarities to human XFS including: iris transillumination defects, production of an exfoliative-like material, and pronounced pigment dispersion. Despite these insults, Lyst mutation does not cause increased IOP or optic nerve damage within the context of a C57BL/6J genetic background. However, defects in the circadian rhythm regulation of IOP were identified. Sequence analysis identifies that the beige mutation is predicted to delete a single isoleucine from the WD40 domain of the LYST protein. I identified CSNK2B as a binding partner of LYST and showed that LYSTbg-J completely disrupts the interaction. CSNK2B function in regulating E-cadherin and β-catenin binding is subsequently disrupted. These results lead to a working hypothesis that aspects of the XFS phenotype involve LYST and CSNK2B pathways, likely influencing cell-cell adherens junctions. Epistasis experiments were used to test for genetic modifiers of Lyst, which demonstrated that albino Lyst mutant mice exhibited complete rescue of Lyst-dependent iris phenotypes. In a genetic background-driven approach, a DBA/2J strain of congenic mice was created. The DBA/2J background, which harbors multiple mutations influencing melanosomal-proteins, enhanced Lyst dependent iris phenotypes. Thus, both experimental approaches implicated melanosomes, a potential source of oxidative stress, as mechanistically contributory. Supporting a contributory role of oxidative damage, Lyst mutation resulted in genetic context sensitive changes in iris lipid hydroperoxide levels, being lowest in albino and highest in DBA/2J mice. These results identified an association between oxidative damage to lipid membranes and severity of Lyst-mediated phenotypes, uncovering a new mechanism contributing to pathophysiology involving LYST. In conclusion these results demonstrate that mutation of the Lyst gene can produce ocular features of human XFS and suggests that LYST or LYST-interacting genes may contribute to XFS.

Epithelial ovarian cancer (EOC) is the leading cause of death from gynecological malignancy in Western countries. A better understanding of the etiology and risk factors associated with this disease is crucial for the development of early detection protocols as well as more effective therapies. Epidemiological data has shown that the risks of EOC are highest among peri- or post- menopause women, while increased parity or the use of oral contraceptives is preventive. These data suggest that alterations in reproductive factors are associated with ovarian cancer risks; however, the molecular mechanisms underlying such a link remain to be understood. For decades, EOC was believed to arise from the epithelium that surrounds the ovarian surface, yet this concept fails to explain the morphological resemblance of ovarian epithelial neoplasms with the epithelial cells of the Müllerian-derived female reproductive tract. Alternative ideas have argued that EOC may originate from extra- or para-ovarian tissues such as the fallopian tube and ovarii rete. Studies of the origin of EOC will provide a better understanding of the disease and advance the protocols for early diagnosis. The aims for this thesis are to establish in vivo ovarian tumor models based on the germ cell deficient Wv/Wv mice that mimicking menopausal physiology. The Wv mice harbor a point mutation in c-Kit, which reduces its tyrosine kinase activity to about 1%, resulting in a premature loss of ovarian germ cells and follicles that recapitulates the initiation of menopause in human. We have developed ovarian tumor models by deleting the tumor suppressor genes p53 or p27kip1 in Wv/Wv mice. We found that both Wv/Wv:p27+/- and Wv/Wv :p27 -/- mice developed ovarian epithelial tumors, which consist of papillary structures lined by hyperchromatic neoplastic cells. Positive Cytokeratin 8 (CK8) staining indicated the epithelial origin of these tumors. In vitro primary cultures of mouse ovarian surface epithelial (MOSE) cells from wildtype, p27+/- and p27 -/- mice further confirmed the growth advantage caused by p27 deficiency. However, neither p27 +/- nor p27 -/- MOSE cells were transformed in vitro, probably due to the compensatory increase of cyclin dependent kinase inhibitor (CKI) proteins including p21, p16, p19. When p53 was deleted unilaterally in the ovarian surface epithelial cells of Wv/Wv:p53 loxP/loxP mice by single administration of Adenovirus containing Cre activity (Ad-Cre), ovarian tumors developed after long latency. The ovarian tumors were significantly enlarged when compared with the uninfected ovary from the same mouse. However, most of the lesions in Wv:p53 conditional knockout tumors was negative for epithelial and follicular markers. In vitro deletion of p53 in MOSE cells significantly increased the proliferation and passage numbers of these cells. A compensatory increase of the CKI protein p16, as well as the cellular senescence level was also observed in p53 deleted MOSE cells, suggesting that p53 deletion alone was not sufficient to bypass p16- mediated tumor defense mechanisms in MOSE cells. Taken together, single deletion of p27 and p53 significantly amplified the phenotype of benign tubular adenomas in Wv/Wv mouse. However, neither p27 nor p53 deletion was sufficient to induce the development of malignant ovarian carcinomas in Wv/Wv mice, probably due to the up-regulation of CKI family proteins such as p21, p16 or p19.

p53 functions as an important tumour suppressor through its ability to regulate a number of important cellular processes such as cell cycle arrest, apoptosis, DNA repair, senescence, and angiogenesis. An in vivo genetic modifier screen performed using Drosophila melanogaster resulted in the identification of D. melanogaster brain tumour (brat) as a putative modifier of of the p53 small eye phenotype. Mammalian homologs of brat are members of the tripartite motif family that contain a c-terminal NHL domain. We focus on elucidating the in vivo role of one such homolog, BERP, through the generation and characterization of a classical gene-deletion mouse mutant. We report that BERP-deficient mice exhibit enhanced learning/memory, increased fear, impaired motor coordination, and increased resistance to PTZ -induced seizures. Electrophysiological and biochemical studies show a decrease in mIPSC amplitude along with a decrease in cell surface expression of gamma2 subunit-containing GABA A receptors in the brains of BERP-deficient mice. In addition, no effect of genotype is apparent when examining BERP mRNA levels in the brain. This suggests that the decreased cell surface expression of gamma2 subunit-containing GABA A receptors is likely a posttranscriptional phenomenon and supports the possibility that BERP may be involved in the intracellular trafficking of GABA A receptors. In investigating the possible relationship between BERP and p53, we identify the presence of a transcriptionally competent p53 response element within the first intron of the human BERP genomic locus and demonstrate that the BERP expression is up regulated in a p53-dependent manner both in vitro and in vivo. These results support the interpretation that BERP is a novel p53-regulated gene and suggest a new role for p53 in the regulation of GABA A receptor trafficking and epileptogenesis.

Systemic lupus erythematosus (SLE) is a multisystem autoimmune disease with a strong and complex genetic basis. To dissect the function of the lupus susceptibility loci on New Zealand black (NZB) mouse chromosome 1, the lab had previously generated congenic mice with an introgressed homozygous NZB chromosome 1 intervals extending from ~35 or ~82 to 106 cM on the C57BL/6 background. Although both mouse strains made IgG anti-nuclear antibodies (ANAs), ANA titres and cellular activation were significantly higher in mice with the longer interval. These studies suggest the presence of two susceptibility genes. In this thesis I have sought to further characterize the cellular abnormalities and underlying genetic polymorphisms that produce them in these mice. Using mixed hematopoietic chimeric mice, with a mixture of tagged-B6 and congenic bone marrow I demonstrate that there are intrinsic B and T cell functional defects in chromosome 1 congenic mice. I further show that an intrinsic B cell defect is required for efficient recruitment of B cells into the spontaneous germinal centres and differentiation of autoantibody producing cells in these mice. To more precisely localize the susceptibility loci, I produced and characterized a number of additional subcongenic mouse strains. This revealed surprising genetic complexity with the presence of at least four lupus susceptibility loci and a suppressor locus on chromosome 1, several of which appeared to impact on T cell function. Finally, I generated bicongenic mice carrying both NZB chromosome 1 and 13 intervals, hypothesizing that since these were two of the major intervals associated with autoimmune disease in NZB mice they would fully recapitulate the autoimmune phenotypes. Although this hypothesis was incorrect, several novel phenotypes developed including marked expansion of the plasmacytoid and myeloid dendritic cell compartments and increased BAFF and IgA autoantibody production. Although this expansion was associated with TLR hyper-responsiveness, disease severity remained mild, possibly due to the lack of IFN- production, which appeared to be inhibited in these mice. Thus, lupus arises from immune defects affecting several cellular populations, which are the product of multiple genetic polymorphisms that interact in a complex fashion to produce the autoimmune phenotype.

p53 functions as an important tumour suppressor through its ability to regulate a number of important cellular processes such as cell cycle arrest, apoptosis, DNA repair, senescence, and angiogenesis. An in vivo genetic modifier screen performed using Drosophila melanogaster resulted in the identification of D. melanogaster brain tumour (brat) as a putative modifier of of the p53 small eye phenotype. Mammalian homologs of brat are members of the tripartite motif family that contain a c-terminal NHL domain. We focus on elucidating the in vivo role of one such homolog, BERP, through the generation and characterization of a classical gene-deletion mouse mutant. We report that BERP-deficient mice exhibit enhanced learning/memory, increased fear, impaired motor coordination, and increased resistance to PTZ -induced seizures. Electrophysiological and biochemical studies show a decrease in mIPSC amplitude along with a decrease in cell surface expression of gamma2 subunit-containing GABA A receptors in the brains of BERP-deficient mice. In addition, no effect of genotype is apparent when examining BERP mRNA levels in the brain. This suggests that the decreased cell surface expression of gamma2 subunit-containing GABA A receptors is likely a posttranscriptional phenomenon and supports the possibility that BERP may be involved in the intracellular trafficking of GABA A receptors. In investigating the possible relationship between BERP and p53, we identify the presence of a transcriptionally competent p53 response element within the first intron of the human BERP genomic locus and demonstrate that the BERP expression is up regulated in a p53-dependent manner both in vitro and in vivo. These results support the interpretation that BERP is a novel p53-regulated gene and suggest a new role for p53 in the regulation of GABA A receptor trafficking and epileptogenesis.

Systemic lupus erythematosus (SLE) is a multisystem autoimmune disease with a strong and complex genetic basis. To dissect the function of the lupus susceptibility loci on New Zealand black (NZB) mouse chromosome 1, the lab had previously generated congenic mice with an introgressed homozygous NZB chromosome 1 intervals extending from ~35 or ~82 to 106 cM on the C57BL/6 background. Although both mouse strains made IgG anti-nuclear antibodies (ANAs), ANA titres and cellular activation were significantly higher in mice with the longer interval. These studies suggest the presence of two susceptibility genes. In this thesis I have sought to further characterize the cellular abnormalities and underlying genetic polymorphisms that produce them in these mice. Using mixed hematopoietic chimeric mice, with a mixture of tagged-B6 and congenic bone marrow I demonstrate that there are intrinsic B and T cell functional defects in chromosome 1 congenic mice. I further show that an intrinsic B cell defect is required for efficient recruitment of B cells into the spontaneous germinal centres and differentiation of autoantibody producing cells in these mice. To more precisely localize the susceptibility loci, I produced and characterized a number of additional subcongenic mouse strains. This revealed surprising genetic complexity with the presence of at least four lupus susceptibility loci and a suppressor locus on chromosome 1, several of which appeared to impact on T cell function. Finally, I generated bicongenic mice carrying both NZB chromosome 1 and 13 intervals, hypothesizing that since these were two of the major intervals associated with autoimmune disease in NZB mice they would fully recapitulate the autoimmune phenotypes. Although this hypothesis was incorrect, several novel phenotypes developed including marked expansion of the plasmacytoid and myeloid dendritic cell compartments and increased BAFF and IgA autoantibody production. Although this expansion was associated with TLR hyper-responsiveness, disease severity remained mild, possibly due to the lack of IFN- production, which appeared to be inhibited in these mice. Thus, lupus arises from immune defects affecting several cellular populations, which are the product of multiple genetic polymorphisms that interact in a complex fashion to produce the autoimmune phenotype.