Paediatrics (RCH) - Theses

Permanent URI for this collection

Search Results

Now showing 1 - 10 of 10
  • Item
    Thumbnail Image
    Epigenetic dysregulation underpinning immune cell dysfunction in food allergy
    Imran, Samira ( 2022)
    Food allergy is a major health concern, particularly in Australia, where some of the highest prevalence has been recorded. Although food allergy is most common in younger age groups (between years 0 to 4), most of these allergies resolve within the first few years of life. However, allergies to peanuts and tree nuts tend to be lifelong, and are more likely to cause severe allergic reactions such as anaphylaxis. These severe allergic reactions occur most frequently in the adolescent age group, the highest risk group for fatalities from such reactions. Moreover, up to 30 percent of food allergic adolescents suffer from multiple food allergies, and there is no current effective treatment for this variable phenotype. Rates of food allergy have risen rapidly in the last few decades, at a pace that cannot be explained by genetics alone. While there is a genetic predisposition in food allergy, numerous studies have shown correlations between specific lifestyle and early life factors and food allergy development. Therefore, the origins of food allergy are multi-factorial, and the study of epigenetic factors may allow us to establish the interface between the role of these environmental exposures and our genomes. Indeed, past studies have shown that immune cells from food allergic individuals show aberrant phenotypes even in the absence of allergens, thereby indicating that these immune cells are fundamentally wired to respond differently upon non-specific activation, and these differences have been linked to epigenetic variation. This thesis aimed to identify immune pathways implicated in adolescent food allergy, using samples from individuals enrolled in the SchoolNuts study, a study profiling the prevalence of food allergy in adolescents aged 10 to 14 years in Melbourne, Australia. This was achieved in three parts, first by establishing the baseline immune response to activation in CD4 T cells from non-allergic adolescents. To do this, we compared the DNA methylation (epigenomic) profiles (DNAme) of purified naive CD4 (nCD4) T cells at quiescence (72 hour culture in media alone) or following activation (72hour culture in media with antiCD3 and antiCD28 beads) in healthy infants (aged 12 months, n 18) and adolescents (aged 10 to 15 years, n 15); using the Infinium MethylationEPIC array (EPIC array). This data was then integrated with transcriptomic (bulk RNAseq) and cytokine profiles obtained from culture media from the same samples (quiescent and activated), and publicly available naive CD4 open chromatin (ATACseq) signatures from unrelated infants. By showing that T cell activation induced widespread changes in DNA methylation, I revealed that T cell activation is intrinsically linked to DNA methylation reprogramming. Multi-omic data analysis of transcriptomic and open chromatin datasets validated these findings by showing that this level of reprogramming was reflected in gene expression patterns and altered chromatin conformation, with upregulated gene expression and increased DNA accessibility. While I found similar responses to activation in nCD4T cells from infants and adolescents, there were selected genomic sites showing age specific methylation patterns following activation. Many of these CpG sites were located within the T cell receptor gene locus, suggesting that age influences activation-associated remodelling of DNA methylation at this region. Analysis of cytokine production of these cells at quiescence and following activation indicated enhanced production of IFNgamma, IL2 and IL10 in adolescents, relative to infants. Correlation of these cytokine levels, and DNA methylation level of probes from the EPIC array located near genes encoding for them, highlighted probes (cg01720533, cg05646531, cg14284394, cg16247264) potentially associated with this altered response. I then characterised the epigenomes of naive CD4 T cells from adolescents with food allergy, by generating genome-wide DNAme data in purified nCD4 T cells at quiescence and following activation in adolescents (aged 10 to 15 years old) with peanut allergy (peanut only or peanut and one or more additional food allergy) (FA, n 29), and age-matched non-food allergic controls (NA, n 18). As with the previous study, I also analysed transcriptome-wide gene expression and cytokine production in these cells following activation. I showed that adolescents with FA exhibit unique DNA methylation signatures at quiescence and post-activation at key genes involved in Th1 and Th2 differentiation (RUNX3, RXRA, NFKB1A, IL4R). Combined analysis of DNA methylation, transcriptomic data and cytokine output in the same samples identified an attenuated interferon response in nCD4T cells from FA individuals following activation, with decreased expression of several interferon genes, including IFNgamma and a DMR at a key downstream gene, BST2. The third part of this thesis involved performing the first combined epigenetic and transcriptomic analysis of purified CD19 B cells from adolescents with FA, comparing single-food allergic (peanut only) (SA; n 10), multi-food allergic (peanut and one or more other food) (MA; n 7) and non-allergic (controls) (NA; n 9) individuals. Using bulk RNAseq data and genome wide DNA methylation data generated using the EPIC array, I identified 116 differentially expressed genes (DEGs) distinguishing B cells of FA (both SA and MA) adolescents from NA controls. Additionally, I showed higher methylation at PM20D1, a key regulator of several inflammatory processes, in FA individuals. This analysis also identified, for the first time, distinct epigenetic variation specific to the MA phenotype, which included differential methylation at a number of S100 genes (S100A1, S100A13, S100A14), which are involved in pathways such as inflammation, energy metabolism and cell cycle regulation. This thesis is the first to characterise nCD4 T cell responses across multiple age groups and highlight the levels of epigenetic regulation involved in T cell activation, which marks the first step in T cell differentiation to a range of downstream lineages, such as Th1 and Th2, previously implicated in FA. In doing so, this presents important findings in this field of research by identifying potential timepoints for intervention. Moreover, characterisation of age-specific responses to activation deepens our understanding of immune responses within each age group, which will help us determine timings for interventions relating to age-related decline in immune function. My thesis also presents novel findings in the field of adolescent food allergy, by showing epigenetic dysregulation at several genes implicated in T cell differentiation, as well as at a gene encoding an inflammatory regulator in B cells. Collectively, these results indicate that the immune dysfunction observed in adolescents with FA is likely mediated by epigenetic variations across multiple cell types, potentially in an age-specific manner. Moreover, I identified common epigenetic signature distinguishing MA adolescents from SA adolescents; an important finding given the heterogeneity of this MA clinical group, and point the way forward towards exploring targets for further research, which may allow us to identify potential therapeutic approaches to address peanut allergy in adolescents, and work towards establishing a molecular therapy for the multi food allergic phenotype.
  • Item
    Thumbnail Image
    Combined genetic and epigenetic analysis to identify early life determinants of complex phenotype
    Mansell, Toby Edward ( 2019)
    There is now considerable evidence indicating that risk of many complex diseases in adulthood may be influenced by exposure to environmental exposures in utero. A growing number of studies suggest epigenetic markers, including DNA methylation, are involved in this process. Understanding how DNA methylation is impacted by pregnancy exposures, and related to later health, may both contribute to unravelling the aetiology of complex disease risk in later life and provide a potential early-life biomarker for risk prediction. However, current evidence is limited. There has been a predominance of small, poorly powered studies, failure to consider the effects of genetic variation, and limited replication of previous findings. In addition, previous studies investigating the relationship between DNA methylation and offspring health have been primarily cross-sectional. For these reasons, I investigated the associations between pregnancy exposures (in particular, maternal smoking, nutrition and metabolic health, psychosocial stress, and adverse pregnancy conditions), birth outcomes, and offspring blood DNA methylation of the insulin-like growth factor 2 (IGF2) and H19, hypoxia-inducible factor 3A (HIF3A), leptin (LEP) genes. I also considered how genetic variation impacted on these associations. I then investigated the longitudinal relationship between early life methylation and anthropometry, as well as the association between early life methylation and later childhood measures of weight, adiposity, and cardiovascular health. To do this, the large, population-based longitudinal Barwon Infant Study pre-birth cohort (n=1,074) was used, with clinical and questionnaire measures from 28 weeks pregnancy, birth, 12 months post-birth and 4 years post-birth time points. DNA methylation of candidate regions was measured using the Sequenom EpiTyper mass-spectrometry platform in cord (birth) and peripheral (12-month) blood. Infant genetic variation in and near the candidate genes was considered. Infant adiposity was assessed as sum of triceps and subscapular skinfold thicknesses in infancy, and with DEXA scanning at 4 years of age. We found evidence that exposure to maternal psychosocial stress, gestational diabetes, and pre-eclampsia was associated with differences in offspring methylation at the candidate regions, as was infant sex. Genetic variation showed strong effects on DNA methylation levels, with some evidence for the associations of pre-eclampsia and infant adiposity with LEP methylation differing by infant genotype. Early life methylation of HIF3A and LEP showed modest associations with four-year blood pressure and BMI, respectively. While these associations persisted with adjustment for potential confounding factors, they explained relatively little variance in the four-year phenotypes compared to traditional predictors, such as weight. These findings suggest that offspring DNA methylation of these candidate genes involved in regulation of growth and metabolism are sensitive to several environmental exposures and genetic factors. While there is modest evidence for methylation in infant blood associating with later phenotypes, methylation of these genes appears unlikely to have useful predictive utility in isolation. This study is the first to perform early life longitudinal analysis to investigate the association between anthropometry and methylation in infancy. It is also the first to report evidence of earlier methylation associating with later cardiovascular phenotypes. However, as gene expression data was not available, the functional consequences of the altered methylation observed in blood is unclear. Further work is required to replicate these findings in independent cohorts, to determine the nature of expression of these genes in blood, and to investigate if the relationship between early life methylation and later health persists into adulthood.
  • Item
    Thumbnail Image
    Environment versus genetics in the development of health-related problems in childhood obesity
    Kao, Kung-Ting ( 2017)
    Introduction: Childhood obesity is a major health concern affecting a quarter of all Australian youth. Obese youth are at risk of multi-systemic obesity-related comorbidities (ORC), such as Prediabetes and Type 2 Diabetes Mellitus (T2DM), Hypertension, Hyperlipidaemia, Non-alcoholic fatty liver disease (NAFLD), Polycystic ovarian syndrome (PCOS), Obstructive sleep apnoea (OSA), mental health and orthopaedic issues. However, within the obese population, the severity of obesity is a poor predictor of these comorbidities. As such, current guidelines advise screening obese youth for all comorbidities, which is both time consuming and costly. ORC, as with obesity itself, are complex polygenic disorders involving interaction between genetic and environmental risk factors. Furthermore, many of these interactions are likely mediated by epigenetic variation, linking environmental exposures with genetic predisposition. Therefore, clarification of the role of environmental exposures, genetic risk factors and epigenetic profile in obese youth may provide valuable insight into determining which obese youth are at risk of developing ORC. The aim of this study was to characterise and investigate the associations between environmental, genetic and epigenetic factors with obesity-related comorbidities in obese youth. This information obtained will allow clinicians to better predict and profile the comorbidity burden in obese youth for targeting interventions to those most likely to benefit. Materials and Methods: The Childhood Overweight BioRepository of Australia collected comprehensive clinical and anthropometric data, as well as blood samples for genetic, metabolic and hormonal profile testing from patients (age 1 to 18 years) attending the Weight Management Clinic at the Royal Children’s Hospital, Melbourne. Anthropometric measures included standard measurements of weight, height and waist circumference (converted into BMI z-score (BMI-Z) and waist-height ratio for analyses) as well as total percentage body fat (%FM) and percentage truncal body fat (%truncal) using bioimpedance. Dietary intake was assessed using the Australian Child and Adolescent Eating Survey and physical activity was measured objectively using accelerometry. Genetic typing and epigenetic profiling of single nucleotide polymorphisms (SNP) and measures of DNA methylation were performed using the SEQUENOM MassARRAY platform. Results: 349 participants were included in the study. Both metabolic and non-metabolic ORC are common in obese youth. The most commonly used measure of adiposity, BMI-Z, was the anthropometric marker least associated with ORC, while important ORCs such as T2DM and hyperlipidaemia were not significantly associated with any of the anthropometric measures. The presence of two or more significantly elevated anthropometric markers was associated with the presence of a higher number of ORCs. Almost all obese youth reported poor vegetable intake and high sedentary behaviour (TV-watching and total media) time. Although initial analyses observed associations between sedentary behaviour (TV-watching and total media use time) with prediabetes and hyperlipidaemia, unhealthy food options with mental health issues, and high fruit intake with NAFLD, these associations lost significance after correction for multiple testing by false discovery procedure (FDR). However, low vitamin D levels were significantly associated with hypertension, independent of adiposity and season. The Fat Mass and Obesity (FTO) gene has the largest influence on adiposity amongst obesity-risk genes to date and has been associated with ORCs. In our cohort, the FTO risk allele was associated with poorer longitudinal weight outcomes in pre-pubertal children, but not pubertal adolescents. The FTO SNP however was not associated with any ORCs. We also examined the association between the top 22 T2DM risk SNPs identified from previous studies and Prediabetes or T2DM status. The SLC30A8 rs13266634 SNP appeared to be protective of prediabetes/T2DM status in unadjusted analysis but lost significant on FDR testing. The addition of T2DM SNPs profile to TV-watching time improved the predictive model for Prediabetes/T2DM status. We were not able to replicate the previously reported association of FTO CpG hypomethylation with prediabetes/T2DM in obese children and adolescents. However FTO hypomethylation was seen in obese youth with NAFLD, independent of age, sex and total body adiposity. FTO methylation was independent of FTO expression, which was not associated with any ORC. Conclusion: Anthropometry is not associated with comorbidities in childhood obesity. Environmental factors do not appear to be associated with comorbidities. Low vitamin D level is associated with hypertension. Genetic factors impact on weight gain but not comorbidities, and epigenetic factors may contribute to NAFLD. Combining these environmental and genetic elements with established clinical risk factors in the assessment of each individual ORC may assist in enhancing current clinical models of care for obese children and adolescents.
  • Item
    Thumbnail Image
    Investigating DNA methylation in paediatric central nervous system tumours
    Sexton-Oates, Alexandra Louise ( 2016)
    Each year in Australia, approximately 600 children will be diagnosed with cancer, 140 of these with a central nervous system (CNS) tumour. With a five year overall survival rate of just 50%, CNS tumours remain the second most common cause of death of Australian children. In addition to their relatively high mortality, paediatric CNS tumours are also associated with numerous morbidities in survivors, though the cellular and molecular mechanisms driving tumour origin and behaviour remain to be fully understood. One of the major emerging features of all human cancers is the disruption of epigenetic, including DNA methylation, profile. This study investigated DNA methylation in paediatric CNS tumours with the aim of increasing our understanding of tumour development and biology, with a view to identifying potential novel targets for treatment and biomarkers to better predict tumour behaviour at diagnosis. Experiments were conducted using archival patient tissue removed for diagnostic purposes, and stored as formalin-fixed paraffin-embedded blocks in the Department of Anatomical Pathology at the Royal Children's Hospital, Melbourne. As these samples were not collected with informed consent for research – a fundamental tenet of research on human subjects, the first aim of this study was to investigate the practicability of obtaining consent for the use of archival tissue in research, and to test familial views on the use of such samples for research purposes. Using tracing letters, clinic visits and phone calls, 107 of 178 eligible families were able to be contacted with a significant negative correlation between time since diagnosis and ability to make contact. Of those families that returned a completed consent form, 74 of 75 agreed to the use of their child’s archival tissue for future research. This was the first study of its kind in Australia, and importantly identified no differences in opinion between the parents of living and deceased children. This work formed an important part of the arguments underpinning the granting of a waiver of consent for the use of all archival CNS tumour material for research purposes, enabling the establishment of a cohort amenable for molecular investigation. Once established, the study cohort underwent genome-wide DNA methylation profiling using the Illumina Infinium HumanMethylation450 BeadChip array. The second aim of the project was to profile DNA methylation in the most common paediatric CNS tumour, pilocytic astrocytoma (PA). Despite often identical histology, children with PA have diverse outcomes, associated with both tumour location and tumour behaviour. In this study we identified specific DNA methylation variants in PAs arising from three different regions of the brain, with enrichment of genes involved in ‘embryonic nervous system development’ pathways. The study also identified an accurate method of predicting tumour behaviour at diagnosis, by combining clinical and DNA methylation features. The third aim of the project was to determine whether MGMT promoter methylation, known to have prognostic utility in adult high-grade glioma, was also a feature of paediatric high-grade glioma with similar prognostic utility. A total of 19% of the paediatric cases examined were found to have MGMT methylation, however this was not associated with prolonged survival. The final aim of this study was to examine the DNA methylation profile of paediatric high-grade gliomas arising in association with an autosomal recessive genetic disorder, constitutional mismatch repair deficiency (CMMRD) syndrome. The majority of CMMRD-associated tumours were found to show distinct DNA methylation profiles, and were generally hypomethylated, relative to their sporadic, histologically similar counterparts. In conclusion, this study has used DNA methylation profiling to improve our understanding of paediatric CNS tumour biology, highlighting its potential for predicting tumour behaviour and improving patient outcome.
  • Item
    Thumbnail Image
    DNA methylation patterns in paediatric acute lymphoblastic leukaemia
    CHATTERTON, ZAC ( 2014)
    Introduction: Disruption of DNA methylation is the most common molecular alteration in human cancers. Paediatric Acute Lymphoblastic Leukaemia (ALL) is the most prevalent childhood cancer and strong evidence indicates that DNA methylation alterations exist within this disease. Several genetic mutations have been described that contribute to the malignant transformation within the B-cell subtypes of ALL (B-ALL), however many of the malignant phenotypes are unexplained by genetic mutations alone. DNA methylation has the ability to alter gene expression and thus DNA methylation alterations may contribute to observed malignant phenotypes, potentially activating oncogenes or inactivating tumour suppressor genes analogous to genetic mutations. Furthermore, DNA methylation alterations represent viable clinical biomarkers for disease diagnosis, prognosis and disease tracking. At the start of this project, preliminary genome-scale DNA methylation profiling had been performed on paediatric B-ALL with appropriate B-cell controls to identify contributing DNA methylation alterations and only limited studies had investigated techniques, thresholds and assays for the clinical implementation of DNA methylation biomarkers. Materials and Methods: Two approaches were used to characterise genome-scale DNA methylation alterations in 69 paediatric B-ALL patients; the Illumina Infinium HumanMethylation BeadChip arrays HM27 and HM450. Validation of B-ALL DNA methylation alterations was conducted using the SEQUENOM MassARRAY EpiTYPER. Genome-scale analysis of gene expression (Affymetrix microarray) was also performed in 17 B-ALL cases and integrated with B-ALL methylome data. The study also developed novel techniques for the analysis of DNA methylation using MALDI-TOF Mass Spectrometry (SEQUENOM). Results: Genome-scale disruptions in DNA methylation were characterised in paediatric B-ALL, validating a number of previous small scale experiments and identifying hundreds of genes with associated DNA methylation disruption. DNA methylation alterations were found to be prevalent in all paediatric B-ALL subtypes and stable biomarkers of disease. Two highly differentially methylated sites in the gene promoters of FOXE3 and TLX3 were used as targets to establish new MALDI-TOF Mass Spectrometry techniques that could 1) analyse multiple DNA methylation regions in single reaction and 2) sensitively detect rare DNA methylation events. The techniques were applied to patient samples and enabled high sensitivity and specificity measurements for disease diagnosis. Furthermore, these techniques enabled sensitive disease tracking and insights into the detection of minimal residual disease by DNA methylation analysis. Integration of genome-scale DNA methylation and gene expression data identified common and subtype-specific epigenetic disruption in paediatric B-ALL effecting known tumour suppressors and genes implicated in apoptosis, cellular proliferation and cell signalling. Furthermore, this study uncovered prognostic DNA methylation signatures associated with B-ALL relapse, present across several B-ALL subtypes. Conclusions: The findings of this study have revealed common alterations to DNA methylation across the genomes of paediatric B-ALL that establish a mechanism for clonal inheritance of gene deregulation integral to malignant phenotype. Additionally, the study establishes targets, techniques and thresholds for clinical implementation of DNA methylation loci as biomarkers for disease diagnosis, prognosis and tracking.
  • Item
    Thumbnail Image
    Influence of maternal diet and lifestyle on twins DNA methylation at birth and the changes that occur within the first 18 months of life
    LOKE, YUK JING ( 2013)
    The Developmental Origins of Heath and Disease (DOHaD) concept describes the importance and impact of prenatal environment on later health and disease risk. One of the mechanisms of how prenatal environment can influence the development of the offspring is epigenetics, defined as the study of mitotically and/or meiotically heritable changes in the gene function independent of DNA sequence. The term “epigenetics” was coined by Conrad Waddington as an important mechanism that shapes the development of an organism. In this PhD, DNA methylation was used as a representative of epigenetic status, as it is the most robust epigenetic mark. There is mounting evidence for associations between maternal factors and DNA methylation at gene-specific, global, and genome-wide levels. However, most studies to date have been limited in scope, examining DNA methylation in specific regions in one or two cell types. Newborn twins’ samples from the Peri/postnatal Epigenetic Twins Study (PETS) cohort were used for all analyses in this thesis. The benefit of using twins is the ability to investigate shared and non-shared maternal factors (e.g., the fetal ‘supply line’) on DNA methylation, which until now, have not been studied. The major aim of my PhD was to investigate associations of various shared and non-shared maternal factors with DNA methylation at a gene-specific (Chapter 3 and 4), global (Chapter 4) and genome-wide level (Chapter 5) in various cell types. At a gene-specific level, IGF2 and H19 were the chosen as candidate genes because they are known to be important in embryonic growth. Repeat elements Alu and LINE-1 were used as surrogates of global methylation. Epigenetic profile was hypothesised to be dynamic in the first few months of life, due to its critical developmental stages in early childhood. Previous studies have used a cross-sectional model to investigate DNA methylation changes over time at various ages, but very few studies have used a longitudinal model for this purpose from birth to infancy. My PhD also aimed to fill this gap by using a longitudinal study of methylation in twins from birth to 18 months of age at a genome-wide level (Chapter 6). Using the Sequenom MassARRAY EpiTYPER platform to measure DNA methylation of genomic regions controlling expression at IGF2/H19, and in the Alu and LINE-1 interspersed repeats in five different cell types of newborn twins, it was found that both shared and non-shared maternal factors associate with DNA methylation at birth, and often in a cell type- and region-specific manner, even with the surrogates of global methylation. These are valuable findings, as they further inform that associations of specific maternal factors with certain regions and cell types cannot always be extrapolated to other regions or cell types. This study further reports that false positives can occur, most likely due to small sample size. Methylation analysis at a genome-wide level found that maternal factors were more likely to influence genes involved in metabolic pathways, especially amino acid metabolism. It was also found that maternal factors were less likely to affect regions of functional importance. These findings are important for future studies of prenatal environment. Finally, it was reported that one third of the genome in buccal epithelium rapidly changes over time in the first 18 months of life. This study has revealed the complexity of the epigenome in the newborn in response to the influence of maternal factors, and the dynamic changes over time in the epigenome from birth to infancy. Validation of associations in larger sample sizes in various cell types, and identification of the legacy of such influences in the long-term health of the offspring warrant further investigation. Furthermore, expression analysis on these findings would further solidify the clinical relevance of the associations seen.
  • Item
    Thumbnail Image
    Epigenetics of human placental development and pregnancy-associated disease
    NOVAKOVIC, BORIS ( 2013)
    INTRODUCTION: Epigenetics literally means ‘above DNA’ and refers to the study of molecular modifications that control gene expression and chromatin structure. DNA methylation, the most extensively studied epigenetic modification, is involved in both the maintenance of chromosome stability and gene expression. Due to its role in gene expression, tissue specific DNA methylation patterns are assumed to reflect the function of a specific gene in a particular tissue. The human placenta facilitates the interaction between the mother and the fetus, including nutrient and oxygen exchange, waste removal and the protection of the fetus from the maternal immune response. Due to its role at the feto-maternal interface, the placenta is exposed to several environmental factors with the capacity to alter placental function and fetal development. Many of these effects are likely to be mediated by epigenetic change. Linking specific environmental exposures, genetic, and epigenetic variation to maternal and neonatal outcomes may provide valuable mechanistic insights into the role of placental dysfunction in pregnancy-associated disease and later health. Therefore, DNA methylation studies in healthy and disease placentas have the potential to identify new genes associated with placental function. The aim of this PhD was to take a genome-scale approach to characterise gene promoter methylation in the normal human placenta. MATERIALS AND METHODS: Several different tissues, primary cells and cell lines were used in this study. These included placental villi from first, second and third trimester, purified first trimester villous and extravillous cytotrophoblasts, choriocarcinoma and trophoblast-derived cell lines. Placental tissue, neonatal cord blood and maternal peripheral blood serum from twin births, collected as part of the Peri/post-natal Epigenetic Twins Study (PETS) cohort, were used for two aims of this project. Environmental data on maternal nutrition and supplementation during pregnancy were collected through questionnaires or measured in maternal blood serum. DNA methylation levels were analysed on the genome-scale level using the Illumina Infinium HumanMethylation27 BeadChip, and at the gene-specific level using the SEQUENOM MassARRAY EpiTYPER platform. RESULTS: Genome-scale DNA methylation analysis of normal human placenta from first to third trimester identified dynamic changes in DNA methylation patterns in response to increasing gestation and environmental/stochastic factors. Most of the changes were observed at genes involved in immune cell communication and signalling, which likely reflects the change in cell composition as well as the differing immunological interactions between the mother and the fetus as the pregnancy progresses. Furthermore, increasing inter-individual variation in methylation level at certain CpG sites over gestation suggests an accumulation of environmental and/or stochastic influences during intrauterine development. Next, the twin model was employed to quantify the relative influence of the underlying genetic and environmental/stochastic factors on placental methylation at term. Genome-scale methylation analysis of placentas from 8 monozygotic (MZ) and 6 dizygotic (DZ) twin births identified widespread differences in methylation within MZ twin pairs, supporting a role of the intrauterine environment in shaping the placental methylation patterns at term. In general MZ twins were more epigenetically similar than DZ pairs, underlining the influence of DNA sequence on methylation patterns. In the subsequent attempt to tease out the association between a specific environment (maternal and neonatal vitamin D) and placental CYP24A1 methylation in 32 MZ and 54 DZ pairs, no link was identified. Finally, a comparison between first trimester cytotrophoblasts and several widely used trophoblast-derived and choriocarcinoma cell lines identified widespread differences in DNA methylation patterns. Almost all gene families tested showed significant differences in methylation between primary cells and transformed cell lines, with choriocarcinoma lines showing the largest differences. This information may be useful when deciding which cell line to use for functional analysis. CONCLUSIONS: This study revealed that placental DNA methylation patterns are dynamic during pregnancy, likely reflecting placental function at specific points in gestation. Furthermore, the intrauterine environment was shown to shape the placental DNA methylation profile through a combination of environmental and stochastic influences. The identification of environmentally sensitive CpG sites across gestation and within MZ twin pairs warrants further investigation.
  • Item
    Thumbnail Image
    Human centromere organisation and function
    Northrop, Emma Louise ( 2012)
    Centromeres are essential for correct chromosome segregation during cell division. Whilst centromeric function is conserved throughout eukaryotes, the profile of centromeric DNA, although generally repetitive, is not, and neocentromeres have been reported at various euchromatic sites devoid of satellite DNA. The lack of DNA sequence conservation has led to the notion that centromere identity is maintained by epigenetic mechanisms. The repetitive nature of canonical centromeres has made studying the organization of the centromere difficult, although the discovery of neocentromeres has provided a useful model for studying the function and organization of the centromere domain, and several studies have used neocentromeres to investigate several proteins required for successful centromere formation and function. One such neocentromere model is the 10q25 neocentromere of Mardel(10), which has previously been used to map the binding domains of several centromere-associated proteins. This study utilized the 10q25 neocentromere model to map the binding domains of three global proteins reported to be enriched at canonical centromeres. The linker histone, histone H1 and the chromatin assembly protein, HMGA1 are enriched at the 10q25 domain following neocentromere formation, suggesting a possible role for each protein at active centromeres. The transcriptional regulator, CTCF, was present at the 10q25 domain both prior to and after neocentromere formation. However, neocentromere formation altered the binding profile of CTCF at the 10q25 domain, with an additional binding cluster observed overlapping the core CENP-A containing domain on the Mardel(10) neocentromere, suggesting a role for CTCF at active centromeres unrelated to the DNA sequence. This study also looked at the role CTCF may be playing at the centromere. Depletion of CTCF resulted in a decrease in transcription at canonical mouse centromeres and at the 10q25 neocentromere, and also resulted in an increase in mitotic segregation defects. Whilst it is evident that the presence of CTCF is important at the centromere for correct centromeric function, the exact mechanism by which CTCF acts at the centromere is still uncertain. Neocentromere-derived mini-chromosomes (NC-MiCs) were developed by telomere associated chromosome truncation (TACT) of the arms of the Mardel(10) marker chromosome. These NC-MiCs provide a useful model for studying the effects large-scale chromosomal rearrangements have on the size and organization of chromatin domains as they can be compared to the progenitor Mardel(10). This study used NC-MiC6, a 1.4Mb mini-chromosome derived from Mardel(10) and still containing the entire Mardel(10) CENP-A binding domain, to investigate the effect of severe chromosomal alterations on the size and organization of the centromere by analysis of the binding domain of CENP-A on NC-MiC6. The CENP-A binding domain on NC-MiC6 was reduced to one-third the size of the Mardel(10) CENP-A binding domain, despite the chromosome size being reduced by 98%. This reduction in size did not dramatically alter the stability of the chromosome, and no correlation was found between the chromosome size and the CENP-A domain size for neocentromeres. Whilst the ratio between chromosome size and centromere size is unlikely to be the trigger for the shrinkage and reorganization of the chromatin domain observed in the 10q25 CENP-A binding domain, the disruption of the chromatin scaffold matrix (S/MAR) or the flanking pericentric heterochromatin in the formation of NC-MiC6 may be responsible for this alteration. Whilst the evidence presented in this study provide further insights into the organization and function of the centromere, our knowledge is still far from complete and further studies are required before we will fully understand this complex and essential domain. Neocentromeres devoid of α-satellite DNA provide a useful model to obtain this information as they can be used to generate a linear ‘road map’ of protein binding at the neocentromere. This information can be used to further understand the organization of the centromere, and can be used in conjunction with three dimensional studies to understand the higher-order chromatin organization of the centromere, thus shedding some light on this complex domain.
  • Item
    Thumbnail Image
    Epigenetic discordance in newborn monozygotic and dizygotic twin pairs
    Joo, Ji-Hoon Eric ( 2012)
    Introduction: There is strong evidence that the intrauterine environment can program the long-term health outcomes of the developing fetus. Adverse fetal programming is often also accompanied by low birth weight and this can act as a predictor for later health complications (e.g. hypertension). Although substantiated by numerous animal studies and a small number of human studies, the mechanisms underlying this phenomenon (known as “fetal programming”), remain to be elucidated. Interestingly, epigenetic marks are reprogrammed during early development and subject to change more frequently than genetic mutations. Additionally, epigenetic marks are sensitive to a myriad of environmental influences, suggesting that environmentally mediated epigenetic change during early development may underpin the phenomenon of fetal programming. In order to increase our understanding of this potential mechanistic link, the current study measured aspects of intrauterine environment and epigenetic profile in Human Umbilical Vascular Endothelial Cells (HUVECs) collected from healthy twins at birth as a part of the recently established Pre/Post-natal Epigenetic Twins Study (PETS). HUVECs provide an insight into the fetal programming hypothesis because this cell type is an important mediator in both controlling fetal growth and maintaining cardiovascular health. Furthermore, this study utilised a twin design, controlling for genetic influences (monozygotic twins) or major shared environmental factors (dizygotic twins) on epigenetic profile. Epigenetic profile was measured on a genome-scale using a recently developed DNA methylation microarray and gene expression arrays (as a proxy sum of all epigenetic marks). In addition, the H19/IGF2 imprinted region was examined at a high resolution as an example of a genomic region subject to epigenetic control, also implicated in fetal growth. Materials and Methods: Three approaches were employed to measure within-twin-pair epigenetic discordance in this study: 1. Genome-scale gene expression analysis of 10 MZ pairs; 2. Genome-scale DNA methylation analysis of 13 MZ and 11 DZ pairs; and 3. DNA methylation analysis of 33 MZ and 26 DZ pairs on H19/IGF2 imprinted locus. Genome-scale analyses of gene expression and DNA methylation were performed using Illumina BeadChip expression and Infinium methylation microarrays, whilst our H19/IGF2 locus methylation analysis was performed using the Sequenom MassARRAY EpiTYPER platform. Results: Both genome-scale and locus specific analyses identified a range of within-pair epigenetic discordance within MZ twin pairs at birth, indicating epigenetic drift in utero most likely due to subtle differences in the in utero environment together with stochastic factors. However, evidence of a genetic influence on epigenetic profile was also found, as within twin pair discordances were generally lower for MZ twins relative to DZ twins and unrelated individuals. By regressing within-pair discordance for gene expression and DNA methylation with birth weight discordance, we were able to identify a number of genes which may play an important role in fetal growth and which provide a potential mechanism for the fetal programming hypothesis. In addition, we show common involvement of genes which are discordantly expressed (i.e. hypervariable genes) in immune reponse and response to external signals and differently methylated genes in cell death and proliferation. This study also shows a greater variation in DNA methylation in regions distant from CpG islands than the islands themselves, providing compelling evidence in support of the important role of DNA methylation at CpG dinucleotides proximal to CpG islands (CpG island shores and shelves). We also utilised publically available gene expression microarray data of twins of different ages and compared their gene expression discordance with those detected at birth in our twins and found an increasing epigenetic discordance associated with the age. Finally, the data from our concurrent studies of additional tissues (cord blood mononuclear cells, buccal, placental cells) revealed a highly tissue specific DNA methylation pattern in the H19/IGF2 region. Conclusions: The findings of this study have revealed multiple levels of regulation of epigenetic profile occurring in humans prior to birth. It supports a role for the in utero period specifying the epigenetic profile in response to maternal nutrition and other environmental exposures (in addition to other stochastic influences), with implications for the fetus’ immediate, as well as long-term health outcomes.
  • Item
    Thumbnail Image
    Whole-genome expression profiling of cord blood mononuclear cells from monozygotic twin pairs
    Andronikos, Roberta Helen ( 2010)
    Substantial variation in gene expression levels exists between individuals, within specific tissues or cell types. With gene expression being the primary mechanism through which genetic information is translated into phenotype, the extent, nature and sources of this variation constitutes an important aspect of human biology. Variation in gene expression levels reflects a complex interplay of genetic and environmental factors. Certain environmental factors and exposures can modify gene expression through epigenetic modifications of DNA and chromatin, thus regulating transcription in a manner largely independent of genetic variation. The sensitivity of epigenetic mechanisms to these factors offers a means through which the environment can modulate expression of the genotype, with effects upon gene expression and ultimately, the phenotype. Studies of variation in gene expression in monozygotic (genetically identical) twins support a substantial environmental contribution to variation in gene expression levels. It is known that the epigenetic and gene expression profiles of monozygotic twins diverge throughout life. Mounting evidence suggests that the period of pre-natal development represents a particularly sensitive one for the occurrence of environmentally induced changes to epigenetic status and gene activity. The current study forms part of a larger research program investigating epigenetic variation in twins and its association with birth weight, maternal nutrition and foetal genotype. The Peri-/Post-natal Epigenetic Twins Study (PETS) builds upon the ‘developmental origins of adult disease’ hypothesis, based on the association between low birth weight and increased risk of cardiovascular and metabolic disease in later life, and focuses on epigenetic changes occurring in utero as the basis of the ‘foetal programming’ phenomenon. The current study is based on the hypothesis that divergence of epigenetic and gene expression profiles occurs from conception in monozygotic twins, in response to differing environments as experienced in utero. This study investigates the gene expression profiles of the cord blood mononuclear cells (CBMCs) of twelve newborn monozygotic twin pairs, including six pairs with birth weight discordance at greater than 15%. Genome-wide expression profiling was performed using the Illumina® Human-6 v2 BeadChip system. Gene expression discordance within twin pairs was assessed using three measures. Of these, the measure of Euclidean distance was considered to be the most systematic and useful. Expression discordance was found to vary substantially across pairs in our sample, with expression discordance being generally lower within twin pairs than between unrelated individuals. A significant correlation was identified between expression discordance and chorionicity, with greater expression discordance in dichorionic pairs compared to monochorionic pairs. The measure of Euclidean distance was also applied to publicly available datasets from genome-wide expression profiling of comparable tissues from adult twin pairs, revealing higher levels of expression discordance within the adult pairs relative to the newborn pairs. All genes surveyed by the microarray analysis were ranked according to the degree of within-pair variation shown across twin pairs. This ranked gene list was subjected to gene ontology analysis to identify gene ontology (GO) terms for which the corresponding ranks were higher than expected. Of the 27 GO terms ranked significantly higher than expected, one third related to immune response or response to other external signals. This data supports our hypothesis that divergence of gene expression profiles occurs from conception in monozygotic twins, and is reflected in differential expression phenotypes detectable at birth. Taken together, these results highlight the role of environment in determining gene expression profiles, and the contribution of environmentally induced changes in gene expression to expression discordance within monozygotic twin pairs. The increased variation observed within dichorionic twin pairs, coupled with the prominence of genes involved in immune/external signal response amongst those showing increased variation across pairs, implies that this variation may arise in response to subtly differing environments experienced by co-twins in utero. Birth weight is a phenotype of particular interest in the Peri-/Post-natal Epigenetic Twins Study (PETS), due to the association of low birth weight with an elevated risk of cardiovascular and metabolic disease in later life. In this study, linear modelling identified 342 genes whose expression levels showed a significant association with birth weight in dichorionic twin pairs. Gene ontology analysis of these genes revealed significant over-representation of GO terms relating to protein dephosphorylation, a process intrinsic to many forms of signal transduction. These data imply a link between the environmental modulation of gene activity via signal response/transduction and the phenotype of birth weight. Due to the myriad number of signal transduction pathways and physiological processes regulated by protein phosphorylation and dephosphorylation, it is not possible to pinpoint with certainty those that may be linked to birth weight or implicated in the association between low birth weight and elevated disease risk from our data. However, these results do provide a basis for further investigation of the specific environmental factors involved in the determination of gene expression variants associated with birth weight. It is to be hoped that future analyses will assist in the identification of the mechanisms underlying the correlation between low birth weight and an elevated risk of cardiovascular and metabolic disease in later life.