Rozprawy doktorskie na temat „DNA methylation”

DNA methylation is a postreplicative covalent modification of the DNA which is catalysed by the DNA methyltransferase enzyme. DNA methylation plays an important role in controlling the gene expression profile of mammalian cells. The hypothesis presented in this thesis is that the expression of the DNA methyltransferase gene is upregulated by cellular oncogenic pathways, and that this induction of MeTase activity results in DNA hypermethylation and plays a causal role in cellular transformation. Novel DNA methyltransferase inhibitors may inhibit the excessive activity of DNA methyltransferase in cancer cells and induce the original cellular genetic program. These inhibitors may also be used to turn on alternative gene expression programs. Therefore specific DNA methyltransferase antagonists might provide us with therapeutics directed at a nodal point in the regulation of genetic information.

One of the main questions of modern biology is how our cells interpret our genetic and epigenetic information. DNA methylation is a covalent modification of the genome that is essential for mammalian development and plays an important role in the control of gene expression, genomic imprinting and X-chromosome inactivation (Bird and Wolffe, 1999; Szyf et al., 2000). Furthermore, changes in DNA methylation and DNA methyltransferase 1 (DNMT1) activity have been widely documented in a number of human cancers (Szyf, 1998a; Szyf et al., 2000).
In Escherichia coli, timing and frequency of initiation of DNA replication are controlled by the levels of the bacterial methyltransferase and by the methylation status of its origin of replication (Boye and Lobner-Olesen, 1990; Campbell and Kleckner, 1990). In mammalian cells, however, the importance of methyltransferase activity and of DNA methylation in replication is only now starting to emerge (Araujo et al., 1998; Delgado et al., 1998; DePamphilis, 2000; Knox et al., 2000).
The work described in this thesis focuses mainly on understanding the functional relationship between changes in DNA methylation and DNMT1 activity on mammalian DNA replication. In higher eukaryotes, DNA replication initiates from multiple specific sites throughout the genome (Zannis-Hadjopoulos and Price, 1999). The first part of the thesis describes the identification and characterization of novel in vivo initiation sites of DNA replication within the human dnmt1 locus (Araujo et al., 1999). Subsequently, a study of the temporal relationship between DNA replication and the inheritance of the DNA methylation pattern is presented. We also demonstrate that mammalian origins of replication, similarly to promoters, are differentially methylated (Araujo et al., 1998). We then tested the hypothesis that DNMT1 is a necessary component of the replication machinery. The results presented indicate that inhibition of DNMT1 results in inhibition of DNA replication (Knox et al., 2000). Finally, results are presented, demonstrating that the amino terminal region of DNMT1 is responsible for recognizing hemimethylated CGs, DNMT1's enzymatic target. Taken together, the results presented in this thesis demonstrate that DNMT1 is necessary for proper DNA replication and that DNA methylation may modulate origin function.

Myogenesis is the differentiation process which encompasses the formation of skeletal muscle during development, regeneration and tissue homeostasis throughout life. Arising from embryonic or adult stem cells, the myogenic process comprehends the acquisition of a specialized cell identity and the loss of pluri/multipotent and proliferative capacities. Starting with the hypothesis that DNA methylation, together with other epigenetic mechanisms and the transcription factors, orchestrates the transcriptional program, this thesis provides a comprehensive picture of DNA methylation dynamics during murine myogenic progression, addresses their regulatory implications, and identifies relevant differentially methylated regions that define muscle cell identity. Initially, we performed a genome-scale DNA methylation study comparing embryonic stem cells (ESCs), primary myoblasts (MBs), differentiated myotubes (MTs), and mature myofibers (MFs) using AIMS-seq method. We identified 1,000 differentially methylated regions during muscle-lineage determination and terminal differentiation, mainly located in gene bodies and intergenic regions. As a whole, muscle lineage commitment was characterized by a major gain of DNA methylation, while muscle differentiation was accompanied by loss of DNA methylation in CpG-poor regions. Notably, hypomethylated sequences were enriched in enhancer-type chromatin regions, suggesting the involvement of DNA methylation in the regulation of cell-type specific enhancers. Importantly, we detected a demethylated region overlapping the super-enhancer of the cell-identity factor Myf5. We showed that the activation of My5 super-enhancer took place upon DNA demethylation exclusively in muscle-committed cells resulting in gene expression. ChIP analyses showed that the binding of the Upstream stimulatory factor 1 (Usf1) to Myf5 locus was DNA demethylation-dependent in myogenic committed cells. Moreover, Usf1 binding site contained an embedded CpG conserved in humans and demethylated in human MBs but not in human ESCs, altogether reinforcing the hypothesis that DNA methylation regulates gene expression by modulating transcription factor binding accessibility. Next, we analyzed by sodium bisulphite sequencing the DNA methylation state of reported regulatory regions (with and without CpG island) of key genes implicated in myogenesis. After analyzing myogenic and non-myogenic cells we concluded that the muscle cell identity comprehends DNA demethylation of lineage-specific CpG-poor regulatory regions leading to a transcriptionally poised or activated state, while myogenic CpG island promoters are totally unmethylated during myogenesis and are regulated by histone modifications. A collaborative work with Charles Keller’s Lab (Oregon Health & Science University, USA) allowed us to conclude that Rhabdomyosarcoma cell lines present a spurious methylation pattern at usually unmethylated CpG islands, consequently with the aberrant methylation associated to tumorigenesis. Furthermore, the study of pluripotency gene promoters during myogenesis pointed that CpG-poor promoters are repressed during differentiation by DNA methylation and by Polycomb complex at CpG island promoters. Interested in deepen in the DNA demethylation dynamics, we started a collaboration with Rita Perlingeiro’s Lab (University of Minnesota, USA) to study the DNA methylation changes in the myogenic inducible Pax7 ESC-derived model. We showed that the ESC-derived myoblast precursors recreated the DNA methylation signature of in vivo isolated muscle stem cells, supporting this model as a bona fide strategy to generate myogenic precursors in vitro with therapeutic purposes. Finally, we addressed the involvement of an active demethylating mechanism during myogenesis. Apobec2 down-regulation in inducible ESC-derived myoblast precursors with shRNA strategies affected dramatically the myogenic differentiation by impairing DNA demethylation of the Myogenin promoter and abolishing the expression of Myogenin and MHC proteins. Based on these results, we proposed that Apobec2 might be involved in the active muscle specific DNA demethylation along myogenesis.
Partint de la hipòtesi que la metilació de l’ADN, junt amb altres mecanismes epigenètics i els factors de transcripció, orquestra el programa transcripcional, aquesta tesi ofereix un estudi exhaustiu de les dinàmiques de l'ADN durant la progressió miogènica, aborda les seves possibles implicacions reguladores i identifica regions diferencialment metilades que defineixen la identitat de la cèl·lula muscular. L’anàlisi a escala genòmica els perfils de metilació en diferents estadis de la miogènesi va permetre identificar 1000 regions diferencialment metilades, localitzades principalment en regions intergèniques i intragèniques. La majoria de canvis observats eren guanys de metilació i ocorrien durant la determinació de llinatge. D’altra banda, certes regions amb perfils de cromatina associats a enhancers esdevenien desmetilades, suggerint que la metilació de l’ADN pot estar implicada en la regulació de enhancers específics de teixit. L’estudi de gens específics de múscul va mostrar que la identitat de la cèl·lula muscular requereix la desmetilació de l'ADN de regions reguladores pobres en CpGs, alhora que els gens miogènics amb illes CpG a la regió promotora es troben sempre desmetilats i són regulats per modificacions d’histones. Un exemple de la desmetilació específica de múscul és la regió super-enhancer de Myf5. Els assajos d'immunoprecipitació de la cromatina van demostrar que la unió del factor de transcripció Usf1 al locus Myf5 només es donava quan l’ADN estava desmetilat, reforçant la hipòtesi que la metilació de l'ADN regula l'expressió gènica mitjançant la modulació de l'accessibilitat dels factors de transcripció al seu lloc d’unió. Mitjançant l’estudi el perfil de metilació de l'ADN de gens miogènics en un model miogènic derivat de cèl·lules embrionàries, es va observar el mateix perfil observat en mioblasts primaris, indicant que aquest model és una bona estratègia per obtenir mioblasts in vitro amb finalitats terapèutiques. Finalment, el bloqueig de la deaminasa Apobec2 va afectar severament la diferenciació miogènica i la desmetilació de l'ADN del promotor de la Miogenina, indicant que la deaminasa Apobec2 podria estar implicada en la desmetilació activa de l'ADN.

Follicular Lymphoma (FL) is a common B cell Non-Hodgkin Lymphoma with a median survival of 8-10 years. Patients frequently undergo transformation to a more aggressive lymphoma and this is associated with drastically reduced survival. The hallmark of FL is the t(14;18) translocation yet this alone is insufficient for lymphomagenesis. While a number of secondary genetic changes have been described, epigenetic studies have lagged behind. Epigenetics refers to mechanisms that alter gene expression without a change in the primary DNA sequence. DNA methylation was quantitatively profiled at 1505 CpG loci in 164 untreated FL as well as 10 pairs of pre- and post-transformation samples and 24 benign haematopoietic controls. Tumour-specific methylation occurred in >100 genes, preferentially occurring within CpG rich areas known as CpG islands and in genes marked by a repressive histone modification in embryonic stem (ES) cells, trimethylated Lysine 27 of Histone H3 (H3K27Me3). Significant inverse correlation with gene expression was identified for a small number of genes. Significant changes in methylation were not seen in FL samples upon transformation. These findings suggested that widespread DNA methylation occurred as an early 'pre-programmed' event in lymphomagenesis rather than being due to silencing of individual tumour suppressor genes. Methylation profiling of a subset of these samples at >27,000 CpG loci revealed aberrant methylation in FL in >700 CpG islands. These hypermethylated genes were enriched for high-density CpG promoters and for a key set of genes with developmental function which were marked by both repressive (H3K27Me3) and activating (H3K4Me3) marks in ES cells. Examination of H3K27me3 expression by immunohistochemistry in pre- and post-transformation biopsy samples showed wide variation in expression. Furthermore, mutations were identified in the histone methylase EZH2 that catalyses H3K27Me3, in 7 of 20 patients. We can conclude therefore that deregulation of DNA and histone methylation are critical inter-related events for FL pathogenesis.

Daphnia magna is gaining interest as a model for epigenetic research. It is easy to maintain under laboratory conditions and has low genetic diversity due to parthenogenetic reproduction. The D. magna genome is responsive to a wide range of stimuli and genomics resources are being developed for this species. Despite these great advantages, information regarding the epigenome of D. magna and its regulation is still lacking. Thus, the main aim of this work was to describe the methylome of D.magna and investigate its regulation and responsiveness to environmentally relevant exposure conditions. Despite the low levels of global DNA methylation, a defined profile could be identified. DNA methylation in D. magna is sporadic and mainly found at coding regions. These data suggest that D. magna encodes a complete set of genes for DNA methylation reactions. Evidence of direct effects on the DNA methylation profile were found in animals exposed to the DNA methylation inhibitor 5-azacytidine and these changes were persistent after the removal of the stressor. Acute and chronic exposures to environmentally relevant concentrations of stressors (arsenic and hypoxia) also induced changes in gene transcription levels and concentrations of onecarbon pathway metabolites. These findings indicate that the epigenome of D. magna is responsive to changes in the environment, supporting its use as an environmentally relevant model organism for epigenetics research. Furthermore, the maintenance of some of the epigenetic changes in the absence of the initial stressor supports the concept of ‘epigenetic memory’ and its potential use in chemical risk assessment.

Chemical modification of the cytosine base via the addition of a methyl group to form 5-­‐methylcytosine (5-­‐mC) is a well-­‐studied example of an epigenetic mark, which contributes to regulation of gene expression, chromatin organisation and other such cellular processes without affecting the underlying DNA sequence. In recent years it was shown that 5-­‐mC is not the only DNA modification found within the vertebrate genome. 5-­‐hydroxymethylcytosine (5-­‐hmC) was first described in 1952 although it wasn’t until 2009 when it was rediscovered in mammalian tissues that it sparked intense interest in the field. Research has found that unlike the 5-­‐mC base from which it is derived, 5-­‐hmC displays variable levels and patterns across a multitude of tissue and cell types. As such the patterns of these DNA modifications can act as an identifier of cell state. This thesis aims to characterize the methyl and hydroxymethyl profiles of induced pluripotent stem cells (iPSCs), derived from control mouse embryonic fibroblast cell line (p53-­‐/-­‐) as well as and methylation hypomorphic (p53-­‐/-­‐, Dnmt1 -­‐/-­‐) mutant cell lines. As such both somatic cells were subject to reprogramming with Yamanaka factors (Oct4, cMyc, Klf4 and Sox2) via the piggyback transposition technique. Successful reprogramming was confirmed by a number of techniques and outcomes, including the de novo expression of a number of key pluripotency related factors (Nanog, Sall4 and Gdf3). Reprogrammed cells were then analysed for transcriptomic changes as well as alterations to their methyl and hydroxymethyl landscapes that accompany reprogramming. Through this work I have shown that the reprogramming of MEF derived cell lines results in a global increase in 5-­‐hmC for both p53-­‐/-­‐ and (p53-­‐/-­‐, Dnmt1 -­‐/-­‐) hypomorphic mutant cell lines – possibly through the reactivation of an alternative form of DNMT1. I demonstrate by both antibody based dot blot assay and genome wide sequencing that the reprogramming of the (p53-­‐/-­‐, Dnmt1 -­‐/-­‐) somatic cells towards a pluripotent state brings about an increase in methylation levels within the cells. This latter observation may indicate that the reprogramming of the cells is driving them towards a more wild type phenotypic state. My studies suggest that lack of DNMT1 function is not a barrier to reprogramming of somatic cells.

Rhabdomyosarcoma (RMS) is a highly aggressive pediatric soft-tissue sarcoma. It is mainly classified into two major subtypes characterized by alveolar (ARMS) and embryonal (ERMS) histologies. ARMS are characterized by a more aggressive behavior with a higher tendency to present metastasis at diagnosis and to relapse after treatment. Approximately 80% of ARMS harbour the reciprocal chromosomal translocation t(2;13)(q35;q14) and, less commonly, the variant translocation t(1;13)(p36;q14), in which PAX3 and FOXO1, or PAX7 and FOXO1 genes, respectively, are juxtaposed. Unfortunately, no such specific genetic aberrations are known in ERMS, and myogenic factors as myogenin and MyoD1 are the only diagnostic indicators that can be used. Despite aggressive multimodal therapies, the prognosis of high-risk RMS patients has not been improved, with a 5-year overall survival rate less than 20-30%, which prompts a need for new therapeutic strategies. In the last decade many scientific studies have demonstrated that gene expression signature distinguishes PAX3-FOXO1 positive RMS from PAX3-FOXO1 negative, but the reasons of the different expression are still unknown. Aberrant DNA methylation patterning is a hallmark of cancer and could be responsible for the different gene expression of RMS tumor subtypes. We performed genome-wide methylation profile by microarray experiments followed by Reduced-Representation Bisulfite Sequencing (RRBS). Microarray analysis demonstrated a different methylation profile between PAX3-FOXO1 positive and negative RMS, besides among metastatic and non-metastatic RMS. We confirmed HOXC11 as one of the gene differentially methylated between PAX3-FOXO1 positive and negative RMS cell lines using in vitro demethylating agents and bisulfite sequencing. Unfortunately, we did not validate the result in the cohort of RMS biopsies. Moreover, we performed another analysis on microarray data comparing metastatic vs non-metastatic RMS. We found an elevated numbers of differentially methylated regions (DMRs) and many of them map to promoter regions of genes implicated in tumors development. In particular we found DMRs linked to clustered protocadherins, known as tumor suppressor genes. We confirmed a different expression pattern of PCDHA4, as well as a different methylation level of its promotorial region, comparing metastatic and non-metastatic RMS samples. Nevertheless, the methylation status and the expression level of PCDHA4 did not have significant correlation with clinical features and are not a predictor of poor prognosis in RMS. Then, we performed an RRBS sequencing, in order to validate data obtained with microarray platforms. We observed a very low concordance between the two approaches, probably caused by a low quality DNA used in microarray experiments. The RRBS sequencing had demonstrated again that PAX3-FOXO1 positive and negative RMS have a different methylation pattern. Moreover, we demonstrated that GADD45G and NELL1, already described as tumor suppressors in other cancers and often downregulated by methylation processes, had also an involvement in RMS biology. Our experiments confirmed an epigenetic regulation by DNA methylation for GADD45G and NELL1 and that their expression were correlated to RMS histology, presence of fusion status and IRS group staging. Furthermore, GADD45G and NELL1 expression levels affect the progression free survival of RMS patients suggesting their association with a poor prognosis. In conclusion, we demonstrated that GADD45G and NELL1 could be novel potential biomarkers in RMS and we evidenced that the DNA methylation pattern in RMS could be interesting for new therapeutic strategies. We hope that our efforts could contribute to a better molecular classification of RMS tumors and to the identification of new targets for improving standard therapy.
Il rabdomiosarcoma (RMS) è una sarcoma pediatrico dei tessuti molli altamente aggressivo. Viene classificato principalmente in due sottotipi, caratterizzati da istologia alveolare (RMSA) o embrionale (RMSE). Nei RMSA si osserva un comportamento più aggressivo e una maggiore tendenza a presentare metastasi alla diagnosi e alla ricaduta dopo trattamento. Circa l'80% dei RMSA presentano la traslocazione cromosomica reciproca t(2; 13) (q35; q14) e, meno comunemente, la variante t(1; 13) (p36; q14), in cui i geni PAX3 e FOXO1, o PAX7 e FOXO1, rispettivamente, sono giustapposti. Purtroppo, non si conoscono aberrazioni genetiche specifiche nei RMSE e i fattori miogenici, come miogenina e MyoD1, sono gli unici indicatori diagnostici che possono essere utilizzati. Nonostante l’applicazione di terapie aggressive multimodali, la prognosi dei pazienti affetti da RMS, della categoria alto rischio, non è migliorata, con un tasso di sopravvivenza a 5 anni inferiore al 20-30%. Questo dato indica la necessità di sviluppare nuove strategie terapeutiche. Nell’ultimo decennio molti studi scientifici hanno dimostrato che in base al profilo di espressione genica è possibile distinguere RMS PAX3-FOXO1-positivi e PAX3-FOXO1-negativi, ma le ragioni di questa diversa espressione sono ancora sconosciute. L’anomala metilazione del DNA è un indicatore di neoplasia e potrebbe essere la causa responsabile della diversa espressione genica dei due sottotipi di tumore. In questo studio, per mezzo di esperimenti di microarray, abbiamo realizzato un’analisi dello stato di metilazione del DNA su tutto il genoma, proseguendo poi con esperimenti di sequenziamento sfruttando la tecnica Reduced-Representation Bisulfite Sequencing (RRBS). L’analisi dei risultati ottenuti con gli esperimenti di microarray ha dimostrato, non solo un profilo di metilazione diverso tra i RMS PAX3-FOXO1-positivi e negativi, ma anche tra i RMS metastatici e non metastatici. Abbiamo confermato che il gene HOXC11 risulta essere differenzialmente metilato tra linee cellulari di RMS PAX3-FOXO1-positive e negative, sfruttando trattamenti con agenti demetilanti in vitro e sequenziamento del DNA dopo conversione con bisolfito; purtroppo, non abbiamo confermato il risultato nella coorte di biopsie di RMS. Inoltre, abbiamo effettuato un'ulteriore analisi sui dati di microarray confrontando i RMS metastatici con i non metastatici. Abbiamo trovato un elevato numero di regioni differenzialmente metilate (DMR) e molte di queste sono risultate coincidere con le regioni promotoriali di geni implicati nello sviluppo di tumori; in particolare, abbiamo trovato DMR connesse alla famiglia delle clustered protocaderine, note come geni soppressori di tumore. Abbiamo poi confermato un diverso profilo di espressione del gene PCDHA4, così come un diverso stato di metilazione a livello della sua regione promotoriale, confrontando campioni di RMS metastatici e non metastatici. Tuttavia, lo stato di metilazione e il livello di espressione di PCDHA4 non hanno dimostrato una correlazione significativa con le caratteristiche cliniche del RMS. Il gene PCDHA4 non risulta quindi essere un predittore prognostico nel RMS. Successivamente, abbiamo effettuato un sequenziamento RRBS, al fine di validare i dati ottenuti con le piattaforme dei microarray. Ne è risultata una bassa concordanza tra i due approcci, probabilmente a causa della bassa qualità del DNA utilizzato negli esperimenti di microarray. Il sequenziamento RRBS ha dimostrato ancora una volta che i RMS PAX3-FOXO1-positivi hanno un profilo di metilazione diverso dai RMS PAX3-FOXO1-negativi. Inoltre, abbiamo dimostrato che GADD45G e NELL1, già descritti come soppressori tumorali in altri tipi di tumore e spesso regolati in maniera negativa da processi di metilazione, sono anche coinvolti nella biologia del RMS. Con i nostri esperimenti abbiamo confermato una regolazione epigenetica, mediata dalla metilazione del DNA ,per i geni GADD45G e NELL1, e come la loro espressione sia correlata alla istologia del RMS, alla presenza dei geni di fusione e alla stadiazione in gruppi IRS. Inoltre, abbiamo dimostrato che i livelli di espressione di GADD45G e NELL1 influenzano la sopravvivenza libera da progressione di malattia nei pazienti affetti da RMS, suggerendo la loro associazione con una prognosi sfavorevole. In conclusione, il nostro lavoro ha dimostrato che GADD45G e NELL1 potrebbero essere nuovi potenziali biomarcatori nel RMS, evidenziando come il profilo di metilazione del DNA nel RMS potrebbe favorire lo sviluppo di nuove strategie terapeutiche. Ci auguriamo che i nostri sforzi possano contribuire ad una migliore classificazione molecolare dei tumori nei pazienti affetti da RMS e alla identificazione di nuovi bersagli farmacologici per una terapia più mirata.

The mammalian epigenome is globally reprogrammed at two stages of development; this involves the erasure and re-establishment of DNA methylation by both passive and active mechanisms, including DNA repair pathways, and occurs concurrently with an increase in developmental potency. In addition to Uhrf1 and the Tet enzymes, the interplay between activation induced cytidine deaminase (AID) and the DNA repair machinery has been implicated in epigenetic reprogramming of various in vivo and in vitro systems including mouse primordial germ cells, zygotes and induced pluripotent stem cells. AID deaminates cytosine to uracil and can also deaminate methylcytosine, whereas the primary role of UNG is to maintain the integrity of the genome through erasure of uracil. In this thesis, I have aimed to investigate the role of DNA repair in demethylation. To do this I have focused on the specific role of AID and UNG in the demethylation of a static system – primed serum ESCs and a dynamic system – serum to 2i (naïve) to epiblast-like ES cells. As the role of both AID and UNG involves genomic uracil, the central theme of my thesis is the impact of accumulation of uracil on DNA methylation levels in the genome. Therefore, my first aim was to develop a quantitative method to detect low levels of genomic uracil in DNA firstly, by mass spectrometry and secondly, by whole genome sequencing. In Chapter Three, I show that the impact of deamination during DNA preparation can be minimised, such that the level of genomic ESC uracil can be accurately determined as around 12,000 uracil per genome and that, as anticipated, Ung null ESCs have almost twice the genomic uracil content of wildtype ESCs. Secondly, I address the main question which is the impact of uracil accumulation on methylation levels. In order to do this, I generate two cell lines: Ung knockout and Aid over expressing, both of which should result in an increase in genomic uracil. I demonstrate that while over expression of Aid stimulates demethylation in static system and in a dynamic demethylating system, the impact of Ung knockout is less clear. In (static) serum ESCs, loss of Ung results in hypomethylation however, in order to transition to 2i (naïve) ESCs, a process which involves demethylation of the genome, it appears the Ung is required as loss of this gene inhibits proper demethylation. As such, I conclude that UNG-mediated DNA repair functions alongside passive demethylation, by reduction of UHRF1 levels, to demethylate 2i ESCs. To probe the mechanism by which accumulation of uracil in the genome alters methylation levels, I investigate the impact of Ung KO and Aid OE on global levels of DNA damage. I show that both cell lines have a greater incidence of double strand breaks compared to a wild type cell line, and accordingly, upregulate their DNA damage response pathway and the expression of certain repair genes. I suggest that increasing genomic levels of uracil causes genomic instability and that DNA demethylation occurs as a consequence of the repair of extensive DNA damage. More broadly, I suggest that ESCs are uniquely poised, due to their heightened DNA damage response, to use uracil as an intermediate of DNA demethylation. Interestingly, I also note that the biological impact on serum ESCs of loss of Ung appears to be an increase in pluripotency.

DNA methylation is essential for the survival and development of vertebrates. Methylated cytosine in the context of CpG-dinucleotides within the genome is recognized by proteins from the methyl-CpG DNA binding domain (MBD) family. When bound to methyl-CpG DNA, most MBD proteins can recruit histone deacetylase (HDAC) silencing complexes to the site of chromatin to remodel its structure, and this causes transcription repression. Loss of CpG-DNA methylation results in embryonic lethality in mice, but loss of methyl-CpG DNA recognizing MBD proteins produce a viable phenotype. Our interest in MBD proteins arises from our discovery that a subset of them interact with RNA. Two MBD proteins (MBD2 and MeCP2) bind their RNA partners using a domain that contains arginine and glycine (RG) rich motifs. As proteins with RG rich motifs are often substrates of post-translational modifications catalyzed by Protein Arginine Methyltransferase (PRMT), I asked whether the two MBD proteins are methylated at their arginines, and the consequences of such modification. By in vitro and in vivo labeling assays, I ascertained that MBD2 and MeCP2 are substrates of PRMT, and that PRMT1 and PRMT5 are responsible for catalyzing different forms of methylation on the MBD2 protein. The relationship between the two PRMTs with regard to MBD2 methylation was characterized. Subsequently, I identified the significance of MBD2 arginine methylation. Biochemically, methylated species of MBD2 protein have less affinity for the HDAC silencing complex and methyl-CpG DNA. In cells, the loss in repression activity of arginine methylated MBD2 was demonstrated. As MBD2 mediated repression activity can be relieved by arginine methylation, I propose that this mechanism might possibly explains the discrepancy between the phenotypes of CpG-DNA methylation null and MBD null mice. This study provides the first evidence that PRMT participate in the DNA methylation system of chromatin control.

DNA methylation is a crucial epigenetic modification involved in numerous biological processes. However, despite its functional importance, the evolutionary history of this modification and the mechanisms diving such changes are poorly understood. The aim of this thesis is to provide a better understanding of DNA methylation in the context of human recent evolution. We identified and described hundreds of regions presenting a human-specific DNA methylation pattern compared to great apes. We also analyzed for the first time the relationship between DNA methylation changes and sequence evolution at both nucleotide and protein level. In summary, this research reveals new insights into the evolutionary properties of DNA methylation and the interpretation of inter-species non-coding variation
La metilación del ADN es una modificación epigenética implicada en numerosos procesos biológicos. Sin embargo, a pesar de su relevancia funcional, se sabe muy poco sobre su historia evolutiva y los mecanismos que generan estos cambios. El objetivo de esta tesis es proporcionar una mejor compresión de la metilación del ADN en el contexto de la evolución humana reciente. Hemos identificado y descrito cientos de regiones que presentan un patrón de metilación especifico de humanos. Así mismo, hemos analizado por primera vez la relación entre los cambios en metilación y la evolución de la secuencia tanto a nivel nucleotídico como proteico. En resumen, esta investigación revela nuevos conocimientos sobre las propiedades evolutivas de la metilación del ADN y la interpretación de la variación no codificante entre especies.

Assessment of disease status in fish is used as an indicator of the biological effects of contaminants in the marine environment. At some UK offshore sites the prevalence of liver tumours in Limanda limanda (dab) exceeds 20%. However, the molecular mechanisms of tumour formation and the causative agents are not known. The contribution of epigenetic mechanisms, although well-established in human tumourigenesis, is under-studied in tumours of aquatic species. In this thesis, alteration in the DNA methylation patterns in tumours of two fish species, the model species zebrafish (Danio rerio) and the un-sequenced marine flatfish dab, were investigated. The data presented provided a comprehensive characterisation of DNA methylation pattern in zebrafish liver and the first evidence of alterations in DNA methylation profiles of key genes in tumourigenesis pathways in any aquatic species. A statistically significant lower level of global DNA methylation was demonstrated in hepatocellular adenoma (HCA) and non-cancerous surrounding liver tissue (ST) compared to liver of non-cancer bearing dab. The evidence presented in this thesis suggests that chronic exposure to a mixture of pollutants contribute to global DNA hypomethylation followed by further epigenetic and genomic changes, leading to the development of tumours in dab. These findings suggest a link between the environment, epigenome and cancer in fish tumours in the wild.

Thesis (Ph. D.)--Massachusetts Institute of Technology, Computational and Systems Biology Program, 2013.
Cataloged from PDF version of thesis.
Includes bibliographical references.
All the cells in the body contain the same genome yet showcase drastically different phenotypes. This is the result of different transcriptional programs, which are partly controlled by epigenetic modifications, including DNA methylation. In this thesis, I analyze genome-scale DNA methylation profiles across pre-implantation development to identify the targets and characterize the dynamics of global demethylation that lead to totipotency and the subsequent changes to embryonic specification. In Chapter 1, I validate and refine the decades old model for DNA methylation in mouse embryogenesis, identify many retrotransposons with active DNA methylation signatures at fertilization, and discover many, novel differentially methylated regions between the gametes that exist transiently during early development. Notably, the majority of epigenetic events unique to mammalian pre-implantation development are characterized in mouse. In Chapter 2, 1 describe the DNA methylation dynamics in human preimplantation development and show that the regulatory principles that operate in mouse are conserved, though some of their targets are species-specific and define regions of local divergence. Finally, in Chapter 3, I compare DNA methylation dynamics of fertilization to an artificial reprogramming process, somatic cell nuclear transfer, in mouse, and find that most dynamics are conserved but occur at a smaller magnitude after artificial reprogramming. I conclude this thesis with a summary of the chapters and a brief discussion of ongoing and future work.
by Michelle M. Chan.
Ph.D.

There is evidence for a role for epigenetic mechanisms in Alzheimer's disease (AD), the most common age-dependent neurodegenerative disorder. The most studied epigenetic mark DNA methylation -the addition of a methyl group to cytosines located in CpG dinucleotides (5mC) - is known to change with aging and may reflect subtle changes in gene expression. Recently a second type of modified cytosine - a hydroxylated and methylated form (5hmC) - has been detected in the brain and maybe linked to the regulation of gene expression. Case-control differences in post-mortem brain DNA methylation were sought by examining both global DNA methylation and DNA methylation of two candidate genes relating to AD risk factors. Simultaneous assessment of 5mC and 5hmC methylation at a global level indicate hypomethylation of 5mC and hypermethylation of 5hmC in AD brain relative to controls, consistent with the notion that 5hmC serves as an intermediary form for demethylation of 5mC. Age was separately associated with a decrease in LINEl methylation and an increase in 5hmC methylation. The comorbidity of depression in AD was explored by assessing the methylation status of the serotonin transporter (SERT) gene promoter across several brain areas and showed tentative associations of disease with SERT CpG methylation. These measurable differences are very small and unlikely to represent any biological plausibility. In a subset of AD patients with additional clinical and behavioural measures there was no effect of SERT 5HTTLPR genotype on DNA methylation. The hypothesis that amyloid- deposition in brain is a consequence of amyloid precursor protein (APP) gene over-expression was examined by measuring DNA methylation across the APP gene region. AD status associates with methylation levels of several CpG sites within the 5' region CGI shore and exon 5 of the APP gene. However there are no co-occurring separate associations of total APP protein levels at these CpG sites. This study demonstrates the utility of the Fluidigm microfluidics platform to generate highly parallel bisulphite sequencing/base­ pair resolution CpG data.

DNA methylation appears to be involved in the regulation of gene expression. Transcriptionally inactive (silenced) genes normally contain a high proportion of 5-methyl-2'-deoxycytosine residues whereas transcriptionally active genes show much reduced levels. There appears good reason to believe that chemical agents capable of methylating 2'-deoxycytosine might affect gene expression and as a result of hypermethylating promoter regions of cytosine-guanine rich oncogenic sequences, cancer related genes may be silenced. This thesis describes the synthesis of a number of `electrophilic' S-methylsulphonium compounds and assesses their ability to act as molecules capable of methylating cytosine at position 5 and also considers their potential as cytotoxic agents. DNA is methylated in vivo by DNA methyltransferase utilising S-adenoxylmethionine as the methyl donor. This thesis addresses the theory that S-adenoxylmethionine may be replaced as the methyl donor for DNA methytransferase by other sulphonium compounds. S-[3H-methyl]methionine sulphonium iodide was synthesised and experiments to assess the ability of this compounds to transfer methyl groups to cytosine in the presence of DNA methyltransferase were unsuccessful. A proline residue adjacent to a cysteine residue has been identified to a highly conserved feature of the active site region of a large number of prokaryotic DNA methyltransferases. The thesis examines the possibility that short peptides containing the Pro-Cys fragment may be able to facilitate the alkylation of cytosine position 5 by sulphonium compounds. Peptides were synthesised up to 9 amino acids in length but none were shown to exhibit significant activity. Molecular modelling techniques, including Chem-X, Quanta, BIPED and protein structure prediction programs were used to assess any structural similarities that may exist between short peptides containing a Pro-Cys fragment and similar sequences present in proteins. A number of similar structural features were observed.

In mammalian genomes, the base cytosine is frequently modified to S-methylcytosine by the action of DNA methyltransferases. DNA methylation may affect gene expression within the cell by acting to repress transcription, and has been suggested to be involved in numerous cellular processes including tissue-specific gene control, repression of transposon activity, X-chromosome inactivation, genomic imprinting and tumourigenesis. While the pattern of mammalian DNA methylation is typically stable in the adult, widespread changes occur to genomic methylation levels during embryonic development, with a generalised demethylation of the genome at implantation followed by extensive de novo methylation in the postimplantation embryo. Embryonic changes to genomic methylation levels are important in determining the adult methylation state, however limitations in the methods used to detect DNA methylation have prevented a detailed study of DNA methylation in the embryo. Bisulfite sequencing is a technique for the detection of 5—methylcytosine that offers considerable advantages over previously developed methods, and may be used to characterise DNA methylation in detail. However, bisulfite sequencing has not previously been used for the detection of DNA methylation in embryonic samples. In this project, the bisulfite sequencing technique is adapted for the detection of DNA methylation from mouse embryos, and the methylation state of single—copy genes determined throughout embryonic development. The various parameters affecting DNA methylation analysis in mouse embryos are examined. The CpG island within the promoter of the Rb gene was found to be completely unmethylated through all stages of development, supporting a model in which the CpG island is protected from de novo methylation. The methylation state of an allele-specific methylation imprint located 5‘ to the imprinted H19 gene was determined, and found to be methylated only at the paternal allele for all developmental stages examined. However, unlike the Rb CpG island, the control of allele—specific methylation was not absolute, with a low level of demethylation at the paternal allele and a low level of methylation at the maternal allele observed within the area bounded by the methylation imprint. The methylation of a boundary region for the H19 methylation imprint was also characterised. The behaviour of the H19 methylation imprint in mutant ES cells deficient in the DNA methyltransferase Dnmt-I was examined, compared to wild—type ES cells and "rescued" ES cells in which Dnmt—I function had been restored. The H19 methylation imprint in mutant ES cells was found to be completely erased, and not reinstated in the rescued ES cells. The area containing the H19 methylation imprint was found to be relatively resistant to de novo methylation. Methylation at the promoter of the skeletal Ot-actin gene, which exhibits a tissuespecific pattern of expression, was analysed in both embryos and differentiated tissues. The embryonic methylation profile was found to be broadly similar to previous models of tissue-specific genes. In the adult, no consistent correlation was found between promoter demethylation and tissue—specific gene expression. The implications of these analyses of embryonic DNA methylation are discussed, in terms of both the technical aspects of DNA methylation deteCtion techniques and for the regulation of gene expression in the cell.

Adipositas ist eine polymorphe chronische Erkrankung mit epidemischer Prävalenz. Im katabolen Leptin-Melanocortin-Signalweg ist das Proopiomelanocortin Gen (POMC) ein zentrales Element, das bei Dysfunktion massive Adipositas bewirken kann. Auch eine kürzlich identifizierte intragenische Methylierungsvariante des POMC wurde mit Adipositas assoziiert und deutet somit auf eine mögliche epigenetische Modulation des Gewichtsphänotyps hin. Zur Aufklärung der Relevanz, Stabilität und Entwicklung dieser epigenetischen Modifikation wurden die Funktionalität, Ontogenese und Phylogenese der POMC DNA-Methylierung untersucht. In vitro Analysen zeigten DNA-Methylierungsabhängige Promoteraktivität beider CpG-Inseln (CGIs) des POMC. Diese hier erstmals beschriebene Transkriptionsaktivität der intragenischen CGI weist auf einen alternativen Promoter des POMC hin. Hinsichtlich der Ontogenese konnten in Mensch und Maus postnatal stabile DNA-Methylierungsmuster mit interindividueller Konservierung für beide CGIs des POMC identifiziert werden. Zusätzlich erwiesen sich Gewebeunabhängigkeit der DNA-Methylierungsmuster und ihre pränatale Ausbildung zwischen dem Blastocystenstadium und der frühen Organogenese in der Maus. Die POMC DNA-Methylierungsmuster upstream des Exon3 unterscheiden sich in Mensch und Maus. Der mögliche Einfluss von primatenspezifischen Alu-Elementen im Intron2 des POMC hierauf wurde in verschiedenen Primatenfamilien analysiert. Die Ergebnisse zeigen eine bedingte Assoziation der Alu-Elemente mit der DNA-Methylierung in der entsprechenden Region, lassen jedoch auch weitere Einflussfaktoren vermuten. Insgesamt zeigt diese Arbeit, dass die POMC DNA-Methylierung artspezifisch konserviert ist und in der frühen Embryogenese, vermutlich Alu-abhängig, ausgebildet wird. Dabei könnten stochastische Variationen der DNA-Methylierung die POMC-Aktivität beeinflussen und somit das Risiko für Adipositas erhöhen.
Obesity is a polymorphic chronic disease with epidemic prevalence. Within the catabolic leptin-melanocortin signaling pathway pre-proopiomelanocortin (POMC) is a pivotal element. Dysfunction of POMC, e.g. due to mutations, can cause severe obesity. Moreover, a recently identified intragenic methylation variant of POMC was found to be associated with obesity. Therefore, this indicates potential epigenetic modulation of the weight phenotype. To gain further insight into the relevance, stability, and origin of this epigenetic modification, the functionality, ontogenesis, and phylogenesis of the POMC DNA methylation patterns were analyzed. In vitro analyses revealed DNA methylation-dependent promoter activity of both CpG islands (CGIs) of POMC. Thereby, the intragenic CGI was identified as a potential alternative promoter of POMC, which has not been described before. Regarding the ontogenesis, postnatally stable POMC DNA methylation patterns with interindividual conservation were detected for both CGIs in humans and mice. In addition, it was observed that the POMC DNA methylation patterns are non-tissue-specific, stable upon long time administration of a high fat diet, and develop prenatally between the blastocystal stage and the early organogenesis. The POMC DNA methylation pattern upstream of exon3 differs in humans and mice. A possible influence of primate-specific Alu elements within the intron2 region of POMC was analyzed in various primate families. Results evince a partial association of the Alu elements with the DNA methylation pattern in this particular region, but also suggest an influence of additional factors. Overall, this work demonstrates that DNA methylation of the POMC locus is species-specific highly conserved, and that it is established during early embryogenesis, possibly Alu-triggered. In the course of this, stochastic variances of the POMC DNA methylation might influence the POMC activity and consequently alter the risk to develop obesity.

Die DNA Methyltransferasen sind verantwortlich für die spezifische Methylierung von DNA-Basen. Mehrere DNA Methyltransferasen sind bekannt, wobei die Dnmt1 das hauptsächlich vorkommende Enzym ist. Bei Säugetieren korreliert die DNA-Methylierung mit der Genaktivität und ist essentiell für die Embryonalentwicklung. Eine beeinträchtigte Funktion oder Verfügbarkeit des Enzyms kann zu pathologisch veränderten Zuständen führen. Die Regulation der Dnmt1 und die damit verbundene Bedeutung bei der Entstehung von Krankheiten ist bisher nur unvollständig untersucht. In der Frühphase der Embryonalentwicklung von Säugetieren ändert sich das Methylierungsmuster des Genoms dramatisch. In zeitlich aufeinander folgenden Phasen wird die DNA demethyliert (Verlust der Methylgruppen) und neu methyliert (De-Novo Methylierung). Die Hypothese dieser Arbeit ist, dass verschiedene Isoformen der Dnmt1 in spezifischen Entwicklungsstadien exprimiert werden und zu Veränderungen des Methylierungsmusters der DNA beitragen. Um diese Regulation zu untersuchen, wurde die Struktur der Maus Dnmt1-Gens bestimmt. Außerdem wurde in verschiedenen Gewebetypen die Transkriptionsgröße und die Transkriptionsintensität der mRNA mit Hilfe von Northern-Blots quantifiziert. Mit diesen Experimenten konnte im Hoden- und Skelettmuskelgewebe ein längeres Dnmt1-Transkript als in anderen Geweben identifiziert werden. Dieses neue Dnmt1-Transkript wurde mit Hilfe von RT-PCR und RACE-Techniken kloniert und ist in beiden Geweben identisch. Es unterscheidet sich auf DNA-Ebene in der Sequenz des 5'-Endes von der bisher bekannten Form der Dnmt1 und besitzt einen anderen Startpunkt für die Transkription. Darüber hinaus besitzt das neue Dnmt1-Transkript ein 800 Basenpaar großes erstes Exon, welches sich von dem des bekannten Dnmt1-Transkripts unterscheidet. Die spezifische zelluläre Lokalisation des neuen Transkripts wurde mit Hilfe der In-Situ-Hybridisierung analysiert. Mit dieser Technik wurde das alternative Transkript in stärker spezialisierten, haploiden spermatogenen Zellen (Spermatiden) und zu einem geringen Maß im Skelettmuskel nachgewiesen. Während der Differenzierung von Muskelzellen wurde eine verminderte Expression des bereits bekannten mRNA-Transkripts und eine verstärkte Expression des neu identifizierten mRNA-Transkripts festgestellt. Obwohl die mRNA der alternativen Isoform verschiedene, kurze offene Leserahmen enthält, welche die Translation eines spezifischen Dnmt1 Proteins verhindern könnten, wurde durch Immunofluoreszenz- und Western-Blot Analysen ein Translationsprodukt nachgewiesen. Nach den hier aufgezeigten Ergebnissen werden alternative Dnmt1 Isoformen in vivo exprimiert, welche eine aktive Rolle bei der Regulation der DNA-Methylierung spielen könnten.
DNA methyltransferases (DNA MTases) are enzymes responsible for DNA methylation (transfer of methyl groups to a base in the DNA) and are vital for the development of mammals. Several MTases have been identified in eukaryotes but the most abundant is Dnmt1. Furthermore, many pathological conditions are often attributed to an altered availability or function of this enzyme, however the understanding of the regulation of Dnmt1 and the concomitant relationship to diseases is far from being complete. In mammals the methylation of DNA correlates with gene activity, and methylation patterns change dramatically during early development when the genome of the mammalian embryo undergoes consecutive waves of demethylation (loss of methylation) and de novo methylation (methylation of DNA sites that have not been previously methylated). The hypothesis of this study was that alternative Dnmt1 isoforms are expressed at specific developmental stages and thus contribute to changes in the DNA methylation pattern. To study this regulation the structure of the Dnmt1 gene was determined. In this work, the tissue distribution and abundance of Dnmt1 mRNA was analyzed by Northern blot and a new, longer transcript was identified that is present in testis and skeletal muscle tissue. The novel isoform was cloned by a combination of RT-PCR and RACE techniques and found to be identical in both tissues. This new isoform differs from the ubiquitous cDNA in the 5' end, utilizing a new transcriptional start site and an 800 bp long alternative first exon. The cellular localization of this new transcript was determined by in situ hybridization and found to be present in the more specialized haploid spermatogenic cells, spermatids and at lower level in skeletal muscle. During muscle differentiation, the ubiquitous isoform is downregulated while the alternative isoform is upregulated. Although this mRNA codes for several short upstream ORFs which could prevent translation of the Dnmt1-specific ORF, it was found by immunofluorescence and Western blot analyses that this transcript can be translated in vivo producing a shorter Dnmt1 protein. The results shown here indicate that alternative Dnmt1 isoforms are expressed in vivo and might play an active role in the regulation of DNA methylation.

Chez les mammifères, la méthylation est une marque épigénétique ciblant la cytosine principalement dans un contexte CpG pour produire une 5mC. 5mC est très sensible à une déamination spontanée ou enzymatique, conduisant à la formation d'un mésappariement G/T. La 5mCpeut également être oxydée pour former successivement la 5hmC, la 5fC et la 5caC. Ces modifications de la 5mC participent aux processus actifs de déméthylation de l’ADN. Chez les mammifères, la thymine, dans le mésappariement G/T, est clivée par TDG et MBD4. TDG est également en mesure d'exciser 5fC et 5caC. Cette thèse avait pour but de clarifier la fonction de TDG et MBD4 dans la dynamique de la 5mC. Nous avons montré que MBD4 est associée aux protéines de réparation des mésappariements. Les tests enzymatiques, in vitro, montrent que le complexe MBD4/MMR a une activité bifonctionnelle (glycosylase/lyase) spécifique pour G/T, qui est régulée par la méthylation. Pour TDG, nous avons ciblé cette enzyme dans les cellules MEF et caractérisé la distribution des cytosines modifiées. Les résultats montrent des profils de méthylation/oxydation d'ADN qui sont régulés par TDG et surviennent principalement au niveau des répétitions de CA et dans les rétroéléments spécifiques de la lignée souris
In mammals, methylation is an epigenetic mark targeting cytosine mainly in a CpG context, producing 5mC. 5mC is highly sensitive to a spontaneous or enzymatic deamination leading to G/Tmismatch. 5mC can also be oxidized to 5- 5hmC, 5fC and 5caC. These modifications of 5mC participate in the active demethylation processes. In mammals, the thymine in G/T mismatch is cleaved by TDG and MBD4 glycosylases. TDG is able also to excise the 5fC and 5caC.This thesis was to clarify the function of TDG and MBD4 in the dynamics of 5mC. We showed that MBD4 is associated with PMS2, MLH1, MSH2 and MSH6 proteins, four proteins involved in DNA mismatch repair (MMR). The in vitro enzymatic tests show that MBD4/MMR complex has a bifunctional glycosylase/lyase activity specific for G/T and is regulated by methylation.For TDG, we targeted this enzyme in MEF cells and characterized the distribution of modified cytosines. The results show that DNA methylation/oxidation patterns are regulated by TDG and occur mainly at CA repeats and at the mouse-lineage specific retro-elements

Thesis (Ph.D.)--University of Missouri-Columbia, 2007.
The entire dissertation/thesis text is included in the research.pdf file; the official abstract appears in the short.pdf file (which also appears in the research.pdf); a non-technical general description, or public abstract, appears in the public.pdf file. Title from title screen of research.pdf file (viewed on March 23, 2009) Vita. Includes bibliographical references.

Cancer cells have aberrant DNA methylation patterns which are characterized by hypomethylation of a large set of promoters and hypermethylation of tumor suppressor genes. The dynamic nature of the epigenome makes it a valuable target for therapeutic interventions. This thesis focuses on understanding the use of various inhibitors towards DNA methylation-related proteins and their respective anti-cancer activities at both global and gene-specific levels. The widely used demethylating agent 5-azacytidine and 5-aza-2'-deoxycytidine (5-azaCdR) are FDA-approved drugs for the treatment of myelodysplastic syndrome. However, these nucleoside analogs which trap the DNA methyltransferases (DNMTs) are non-specific. Studies have shown that 5-azaCdR induced pro-metastatic genes and caused long distance metastasis. This raises serious safety concerns for their clinical use. On the contrary, targeting the DNMTs individually or in combination did not result in dramatic induction of pro-metastatic genes as with 5-azaCdR treatment. In particular, single DNMT1-specific inhibition resulted in maximum growth suppression when compared to inhibition of all three major DNMTs, while not increasing cell invasiveness. DNMT1 has been shown to be important for cancer growth. Our study supports the idea that DNMT1 has a major role in cancer over the other DNMTs and that DNMT1 inhibitors could be effective anti-cancer drugs. 5-azaCdR has nevertheless been proven to be a potent suppressor of cancer growth. We tested the idea of a combinatorial treatment that may minimize its side-effects on cell invasion while maintaining its growth suppressor effects. The methyl-CpG binding protein 2 (MBD2) protein has been shown to demethylate pro-metastatic genes. Its inhibition in concurrent with 5-azaCdR treatment synergistically suppressed cancer growth, while reversed the 5-azaCdR-induced invasion. In order to have a deeper understanding of the impact of the treatments, microarrays studies on the methylome and transcriptome of the treated cells were carried out. Bioinformatics analysis indicated that the combined treatment suppressed gene networks that were involved in cell mobility, while synergistically enhanced gene networks that were involved in cell death. This data indicate that combining 5-azaCdR treatment with MBD2 inhibition results in more potent anti-cancer effects than either treatment alone. In order to explore the currently available drugs that inhibit MBD2, we tested the combination of S-adenosylmethionine (SAM) with 5-azaCdR on the same cancer cell lines. SAM remethylated gene promoters of pro-metastatic genes and repressed 5-azaCdR-induced invasion similarly to MBD2 inhibition. We then investigated the relationship between SAM and MBD2 downregulation and observed hypermethylation on both CpG and non-CpG sites in the MBD2 promoter upon SAM treatment. Interestingly, inhibition of MBD2 using short interference RNA also resulted in hypermethylation of its own promoter. This observation suggested that SAM treatment could directly downregulate MBD2 expression, which is further downregulated through a feedback loop. These results also suggested that SAM treatment could have a direct effect on MBD2 promoter, which in turn affects multiple MBD2 targets that are involved in invasion. Together, the data from this thesis support the idea that targeting the epigenome could be a highly efficacious anti-cancer therapy and that combining drugs that target DNA methylation could increase the potency over individual treatments.
Les cellules cancéreuses présentent un profil de méthylation caractérisé par l'hypométhylation d'un grand nombre de promoteurs et l'hyperméthylation de gènes suppresseurs de tumeur. La nature dynamique de l'épigénome en fait une cible de choix pour les interventions thérapeutiques. Cette thèse vise à comprendre l'utilisation de divers inhibiteurs visant des protéines liées à la méthylation de l'ADN et leurs activités anticancéreuses à une échelle génomique globale et au niveau de gènes particuliers. Les agents déméthylants 5-azacytidine et 5-aza-2'-deoxycytidine (5-azaCdR) sont des médicaments pour le traitement du syndrome myélodysplasiqueapprouvés par la FDA. Cependant, ces analogues de nucléosides qui piègent les DNA méthyltransférases (DNMTs) ne sont pas spécifiques. Des études ont montrées que la 5-azaCdR induisait l'expression de gènes pro-métastatiques et l'apparition de métastases. Ceci soulève de sérieuses interrogations quant à leur utilisation en clinique. À l'inverse, le ciblage spécifique des DNMTs ne conduit pas à une induction dramatique des gènes pro-métastatiques. Plus particulièrement, l'inhibition spécifique de DNMT1 résulte en une suppression de la croissance maximale des tumeurs, sans effet sur l'invasion cellulaire, lorsque l'on compare à l'inhibition des trois principales DNMTs. Notre étude supporte l'idée que DNMT1 à un rôle majeur dans le cancer et que le développement d'inhibiteurs de DNMT1 pourraient conduire à des médicaments anti-cancéreux efficaces.Il a néanmoins été montré que la 5-azaCdR était un suppresseur potentiel de la croissance cancéreuse. Nous avons testé l'hypothèse qu'un traitement combiné permettrait de minimiser ses effets secondaires sur l'invasion cellulaire tout en maintenant ses effets suppresseurs de croissance. Il a été montré que la protéine methyl-CpG binding protein 2 (MBD2) participait à la déméthylationde gènes pro-métastatiques. Son inhibition simultanée à un traitement 5-azaCdR abolit de façon synergétique la croissance cancéreuse, tout en inhibant l'invasion induite par la 5-azaCdR. Des analyses du méthylome et du transcriptome ont été réalisées par micropuces à partir de cellules traitées avec un siRNA dirigé contre l'ARNm de MBD2 et la 5-azaCdR afin d'avoir une meilleure compréhension de l'impact de la combinaison des traitements. Les analyses bioinformatiques ont indiqué que le traitement combiné réprimait des réseaux de gènes impliqués dans la mobilité cellulaire tandis que les réseaux de gènes activés étaient impliqués dans la mort cellulaire. Ces données indiquent que le traitement à la 5-azaCdR combiné avec l'inhibition de MBD2 résulte en de plus puissants effets anti-cancéreux que l'un ou l'autre des traitements individuels.Nous avons également testé la combinaison de la S-adenosylmethionine (SAM), un médicament actuellement disponible sur le marché et inhibant l'activité de MBD2, avec la 5-azaCdR sur les lignées cellulaires utilisées précédemment. La SAM, de façon similaire à l'inhibition de MBD2 par un siRNA, permet la méthylation des promoteurs de gènes pro-métastatiques et réprime l'invasion induite par la 5-azaCdR. Nous avons ensuite examiné la relation entre la SAM, la diminution de l'expression de MBD2 et l'hyperméthylation observée à la fois aux sites CpG et non-CpG au niveau du promoteur de MBD2 après traitement avec la SAM. De façon intéressante, l'inhibition de MBD2 par des petits ARN interférant résulte également en une hyperméthylation de son propre promoteur. Cette observation suggère que le traitement avec SAM pourrait directement réduire l'expression de MBD2, qui serait réduite encore plus via une boucle de rétrocontrôle. L'ensemble des données de cette thèse supporte l'idée que le ciblage de l'épigénome pourrait être une thérapie anti-cancéreuse hautement efficace et que la combinaison de médicaments qui ciblent la méthylation de l'ADN pourrait augmenter l'efficacité des traitements individuels.

Germ cell tumours (GeTs)affect both paediatric and adult populations, and can occur either in gonadal or extragonadal regions along the body's ventral midline. These tumours can be broadly categorized into two subgroups, seminomatous (SEM) or nonseminomatous (N-SEM). The latter can be further subcategorized into embryonal carcinoma (EC), teratoma, yolk sac tumour (YST) and choriocarcinoma (eC) according to their differentiation. As in many other tumours, DNA methylation has been proposed to be involved in GCTdevelopment. However to date, most studies were performed using adult testicular GCTs. Furthermore, these studies only include a handful of genes in their analysis. Thus, the roles of DNA methylation in paediatric and extragonadal GeTs have not been explored. Therefore, this project attempted to fill this gap in knowledge by performing methylation analysis in a cohort of paediatric GCT samples and GCTcell lines. Although paediatric GCTsmostly consist of teratomas, seminomas or YSTs, only the latter two were included in the methylation analysis as they were the only samples in the available tumour bank. Using the methylation level of L1NE-l repeat elements as a measurement of global genome methylation, we found that both paediatric seminoma and YST samples displayed global hypomethylation as compared to somatic controls. However, when methylation at gene promoter regions was investigated using Illumina Golden Gate methylation arrays, seminoma and YST exhibited very different methylation features. YSTs were found to be highly methylated at many of the sites investigated. Surprlslnglv, we found that the methylation features in seminoma were similar to the somatic controls. From this analysis, we identified 85 genes that were differentially methylated in the VSTs. However, by correlating our methylation data with the expression array data performed by our collaborators on the same samples, only eight of these genes (PYCARO, CASPB, C02, HOAC9, TFAP2C, ETV1, EV/2A, HLA-F) were differentially expressed. As in previous GCTstudies, our analysis was focused on the methylation at epG islands. During the course of this project technological advancement led to the creation of new methylation arrays that offer wider genome coverage. One example is the Infinium Methylation 450K array that covers more than 450,000 CpG sites and includes regions flanking the CpG islands such as the CpG shores and CpG shelves. Since no previous GCT studies have attempted to investigate methylation in those regions, we utilized this methylation array on four GCTcell lines; TCAM2 (seminoma), NT2Dl (teratocarcinoma), GCT27 (embryonal carcinoma) and GCT44 (yolk sactumour). Similar to previous GCT studies, we found that nonseminomatous GCT cell lines displayed higher methylation at the CpG islands as compared to the seminoma cell lines. Strikingly, expanding our analysis to other regions (CpG shores and shelves etc.) revealed that each GCT subtype exhibited distinct methylation features. Both ECand teratoma cell lines displayed higher methylation than the seminoma and YST cell lines at all regions. Interestingly, the YST cell line only showed higher methylation than the seminoma cell line at the CpG islands and to a lesser extent at the CpG shores while the seminoma cell line exhibited higher methylation at the CpG shelves as compared to the YST cell line. This is the first time such features have been reported for GCTs. From this Infinium methylation data, we have also identified a high number of hypermethylated genes including those that are uniquely methylated for each cell line. By correlating this methylation data with Affymetrix gene expression data, 98 genes that were differentially methylated and differentially expressed in the YST cell line have been identified. However, further analysis needs to be performed to understand the role of these genes in YST development. As in other types of tumour, the hypermethylation observed in the YST cell line might be caused by many epigenetic modifiers. Using real-time RT-PCR on three epigenetic modifiers (DNMT38, EZH2, SUZ12), we found that DNMT38 was highly expressed in the YST samples and cell line as compared to the seminoma samples and cell line. This suggests that DNMT38 might contribute to YST hypermethylation and resulting differences in their biology. However, knockdown of DNA methyltransferases (DNMTs) and DNMT38 using 5-azadeoxycytidine and microRNA-29b respectively, did not seem to have any effect on the response of all four GCT cell lines towards cisplatin. On the other hand, both knockdowns only caused little effect on cell migration; affecting only the seminoma and YST cell lines. Nonetheless, further analysis is still needed to fully assess the role of DNA methylation in regulating cell behaviour. In summary, paediatric YSTs displayed hypermethylation at many promoter regions as compared to seminomas. Meanwhile, methylation analysis at regions outside of CpG islands in GCTcell lines revealed unique methylation features for each GCT subtype which might indicate different underlying mechanisms in their development. Further analysis on genes found to be differentially methylated and differentially expressed in both paediatric and GCTcell lines are now needed to fully establish their role in GCT development.

The initiation of DNA methylation in Arobidopsis is controlled by the RNA-directed DNA methylation (RdDM) pathway that uses 24nt siRNAs to recruit the de novo methyltransferase DRM2 to the target site. The REPETITIVE PETUNIA SEQUENCE (RPS) acts as a hot spot for de novo methylation, in particular, hypermethylation is found at a 62 nt region containing an 11 nt palindromic sequence with an 18 nt spacer that forms a putative stem loop structure. The analysis of deletion/substitution derivatives of RPS showed that de novo methylation did not depend on the DNA sequences in the loop region, but did require the 11nt stem region. In addition, a 10nt region directly downstream of the loop region contributed to the efficiency of de novo methylation. Where hypermethylation occurs, it is associated with histone 3 lysine 9 dimethylation. A mutant construct where the palindromic region had been interrupted, produced two distinct types of either almost fully hypomethylated or fully hypermethylated transgenes. This suggests that the RPS stem loop provides a structural de novo methylation target. Analysis of the RPS derivatives in mutant backgrounds identified three distinct stages of DNA methylation which included an RdDM-independent initiation step, DCL-dependent but RDR2-independent spreading. and RDR2-dependent amplification. The initial step therefore represents a potentially novel mechanism for de novo DNA methylation.

Mismatch repair is a vital DNA repair mechanism whose absence leads to a tolerance towards mutations and a predisposition to colon cancer. MLHl is one of the main proteins involved and is highly conserved from E. coli to human. It not only plays a role in repair, but can signal the cell to die if damage levels become too high. The mechanism by which MLHl triggers cell death in response to damage is not entirely clear, and is likely to differ between normal and cancerous cells. Previous work in the Walsh lab had generated MLH1-depleted subclones of a telomerase- immortalised normal human fibroblast cell line. These had been generated by transfecting the parental line hTERT-1604 with an shRNA vector against MLHl and selecting subclones which had reduction in MLHl to various degrees. I further characterised these MLHl knockdown cell lines and revealed that they also exhibit resistance to the methylating agent N-Methyl-N-Nitrosourea. Through the use of various assays we were able to determine that the hTERT immortalised cells did not undergo cell cycle arrest, apoptosis or senescence in response to MNU as colon cancer cell do. Instead, they undergo an MLH1-dependant form of programmed cell death mediated by PARP, but independent of caspase, P53 and ATM/ATR. In 2004, DNMTl deficiency was also implicated in causing MMR defects in mouse embryonic stem cells without a causal mechanism being identified. The effects of DNMTl deficiency are not the same in stem cells and differentiated cells. To determine if depletion of DNMTl can also cause MMR defects in normal human cells, I created DNMT1-depleted hTERT-1604 cells using the same shRNA-mediated strategy as above. Subsequent characterisation of the DNMT1-depleted subclones established that they have a reduction in DNA methylation and the most severely reduced cells are arrested at the G2/M checkpoint. Subclones with significant reductions in DNMT1 also exhibited a decrease in MLH1 expression at the protein, but not the mRNA level. This reduction in MLH1 expression was reversed when a protease inhibitor was employed. The subclone with 31% DNMT1 expression exhibited microsatellite instability, providing further evidence for an interaction between mismatch repair and DNMT1 in human cells and suggesting a mechanism by which this may occur. In conclusion, the work presented here demonstrates a novel role for the mismatch repair protein MLH1 in triggering PARP-dependent cell death in response to damage by MNU. It also shows a link between DNMT1 depletion and MMR deficiency through destabilisation of MLH1.

Type 1 diabetes is an autoimmune disease due to the interaction of genetic and non-genetic factors, leading to an immune response against insulin secreting islet cells. Concordance rates for type 1 diabetes in monozygotic twins vary widely and no single environmental factor has been shown to cause the disease. Therefore, epigenetics has been suggested to play a role in diabetes aetiology. Preliminary results identified DNA methylation changes in CD14+ monocytes from childhood-onset type 1 diabetes which antedated the disease. Following on from this work, this present study was carried out to investigate whole-genome DNA methylation profiles in CD14+CD16- monocytes, CD4+ T cells, CD19+ B cells and buccal cells from 24 monozygotic twin pairs discordant for type 1 diabetes. DNA methylation was profiled using Illumina Infinium HumanMethylation450K BeadChip and analysed using the ChAMP pipeline. Bisulfite sequencing was also carried out on CD4+ cells from four monozygotic twin pairs also discordant for type 1 diabetes. Through bioinformatics analyses, 258 cell-type specific differentially-methylated positions were identified from the 450K BeadChip and 125 differentially-methylated regions from bisulfite sequencing. DNA methylation was also shown to be stable, as similar methylation differences found in the preliminary study, were again detected in the same twin pairs sampled years later. As DNA methylation is a stable marker, it could be used as a biomarker. β-cell death in diabetes releases DNA with unmethylated CpG sites in the insulin promoter region into the blood circulation. To detect these differences, an assay was also developed testing serum samples from monozygotic twin pairs. The data presented provided comprehensive DNA methylation profiles in type 1 diabetes from this discovery cohort. The methylation signature found will then be validated in diabetic, pre-diabetic and control singletons. This in turn will provide data for later functional analyses to identify genes associated with type 1 diabetes risk.

Epigenetics is the study of factors that can change DNA and passed to next generation without change to DNA sequence. DNA methylation is one of the categories of epigenetic change. DNA methylation is the attachment of methyl group (CH3) to DNA. Most of the time it occurs in the sequences that G is followed by C known as CpG sites and by addition of methyl to the cytosine residue. As science and technology progress new data are available about individual’s DNA methylation profile in different conditions. Also new features discovered that can have role in DNA methylation. The availability of new data on DNA methylation and other features of DNA provide challenge to bioinformatics and the opportunity to discover new knowledge from existing data. In this research multiple data series were used to identify classes of methylation DNA to CpG sites. These classes are a) Never methylated CpG sites,b) Always methylated CpG sites, c) Methylated CpG sites in cancer/disease samples and non-methylated in normal samples d) Methylated CpG sites in normal samples and non-methylated in cancer/disease samples. After identification of these sites and their classes, an analysis was carried out to find the features which can better classify these sites a matrix of features was generated using four applications in EMBOSS software suite. Features matrix was also generated using the gUse/WS-PGRADE portal workflow system. In order to do this each of the four applications were grid enabled and ported to BOINC platform. The gUse portal was connected to the BOINC project via 3G-bridge. Each node in the workflow created portion of matrix and then these portions were combined together to create final matrix. This final feature matrix used in a hill climbing workflow. Hill climbing node was a JAVA program ported to BOINC platform. A Hill climbing search workflow was used to search for a subset of features that are better at classifying the CpG sites using 5 different measurements and three different classification methods: support vector machine, naïve bayes and J48 decision tree. Using this approach the hill climbing search found the models which contain less than half the number of features and better classification results. It is also been demonstrated that using gUse/WS-PGRADE workflow system can provide a modular way of feature generation so adding new feature generator application can be done without changing other parts. It is also shown that using grid enabled applications can speedup both feature generation and feature subset selection. The approach used in this research for distributed workflow based feature generation is not restricted to this study and can be applied in other studies that involve feature generation. The approach also needs multiple binaries to generate portions of features. The grid enabled hill climbing search application can also be used in different context as it only requires to follow the same format of feature matrix.

Epigenetic mechanisms respond to both genetic and environmental factors, but these underlying effects are not yet fully characterized or completely understood. In this thesis I used twins as a tool for evaluating the effect of genetic and environmental impacts on DNA methylation, a well-known epigenetic mechanism. DNA methylation was studied on a genome-wide scale in human blood samples with a combination of bioinformatics, computational, and statistical approaches. First, genetic influences on DNA methylation profiles were assessed by estimation of the heritability of DNA methylation and identification of methylation-quantitative trait loci in identical and non-identical twins profiled with the commonly used Infinium HumanMethylation450 BeadChip and the new enhancer-enriched Infinium MethylationEPIC Beadchip. Strong genetic effects (heritability > 0.4) were detected for 10% of sites and common genetic variants were identified to affect methylation levels at 22% of sites, from the 771,169 interrogated CpG sites across the genome. Second, I explored influences on the early-life methylome by comparing genome-wide DNA methylation profiles in naturally-conceived twins and twins conceived by in vitro fertilisation. Analysis of epigenetic profiles obtained by methylated DNA immunoprecipitation coupled with deep sequencing detected small changes at a gene previously associated with infertility, TNP1. Finally, I explored the effect of intrinsic factors on adult DNA methylation profiles, performing epigenome-wide studies of menopause and related phenotypes such as the use of hormone replacement therapy, which have metabolic consequences in middle-aged women. Epigenetic changes at seven CpG sites were associated with hormone replacement therapy. In summary, the results presented in this thesis give insights into genetic and specific environmental influences on the human epigenome.

Although the many cells within a mammal share the same DNA sequence, their gene expression programmes are highly heterogeneous, and their functions correspondingly diverse. This heterogeneity within an isogenic population of cells arises in part from the ability of each cell to respond to its immediate surroundings via a network of signalling pathways. However, this is not sufficient to explain many of the transcriptional and functional differences between cells, particularly those that are more stable, or, indeed, differences in expression between parental alleles within the same cell. This conundrum lead to the emergence of the field of epigenetics - the study of heritable changes in gene expression independent of DNA sequence. Such changes are dependent on “epigenetic modifications”, of which DNA methylation is one of the best characterised, and is associated with gene silencing. The establishment of correct DNA methylation patterns is particularly important during early development, leading to cell type specific and parental allele specific gene regulation. Besides DNA methyltransferases, various other proteins have recently been implicated in DNA methylation. The absence of these proteins leads to defects in DNA methylation and development that can be even more severe than those in DNA methyltransferase knockouts themselves. In this study I focus on three such accessory proteins: LSH (a putative chromatin remodelling ATPase), G9a (a histone lysine methyltransferase) and SmcHD1 (a structural maintenance of chromosomes protein). To compare DNA methylation between WT cells and cells knocked out for each of these proteins, I used whole genome methylated DNA affinity purification and subsequent hybridization to promoter microarrays. This enabled me to compare the requirement for each protein in DNA methylation at specific genomic regions. The absence of LSH in mouse embryonic fibroblasts (MEFs) resulted in the loss of DNA methylation at 20% of usually methylated promoters, and the misregulation of associated protein coding genes. This revealed a requirement for LSH in the establishment of DNA methylation at promoters normally methylated during pre-implantation as well as post-implantation development. Secondly, I identified hypomethylation at 26% of normally methylated promoters in G9a-/- compared to WT ES cells. Strikingly, this revealed that G9a is required for maintenance of DNA methylation at maternal as well as paternal imprinting control regions (ICRs). This is accompanied by expression defects of imprinted genes regulated by these ICRs. Finally, in collaboration with the Brockdorff lab at the University of Oxford I identified a role for SmcHD1 in establishing DNA methylation at promoters on the X chromosome normally methylated slowly during X chromosome inactivation. Interestingly, SmcHD1 was also required for DNA methylation at autosomal gene promoters, contrary to previous reports that it is mainly involved in X chromosome methylation. I conclude that different accessory proteins are required to facilitate correct DNA methylation and gene repression at distinct regions of the genome, as well as at different times during development. This function of accessory proteins may be in part dependent on the prior establishment of specific chromatin signatures and developmental signals, together comprising a precisely regulated system to establish and maintain appropriate DNA methylation throughout development.

Genomic imprinting directs the allele‐specific expression of a subset of loci according to their parental origin. Monoallelic expression of these genes is regulated by imprinting control regions (ICRs) and is established in the embryonic germ line through differential DNA methylation. Differentiated cells can be reprogrammed to pluripotency by several strategies including the ectopic expression of specific ‘inducers’ and by transfer of nuclei into enucleated eggs. Cellular fusion of somatic cells with a pluripotent stem cell partner can also lead to dominant pluripotent reprogramming. Although ES cells (derived from the inner cell mass) and embryonic germ cells (EG, derived from primordial germ cells) can both reprogram, EG cells are unique in being able to erase genomic imprints from the somatic partner. In order to characterize the earliest events in successful reprogramming, as well as EG‐specific DNA demethylation, I generated experimental heterokaryons between B‐lymphocytes and mouse stem cell lines. I showed that ES cells that lack Polycomb Repressor Complex 2 (PRC2) failed to reprogram B cells and were unable to induce two early events that characterise successful B cell reprogramming; a global redistribution of HP1α and an increased serine 10 phosphorylation. In the second part of my study I confirmed that EG cells were able to induce DNA demethylation at several ICR in B cells following fusion. I present evidence that this reprogramming of the somatic genome requires Tet1 and Tet2 and is accomplished through a two‐step process involving both DNA synthesis and conversion of 5methylcytosine into 5hydroxymethylcytosine.

The Paired box gene 6 (PAX6) is a tissue-specific transcription factor, which controls proliferation and differentiation processes of embryonic- and adult progenitor cells. Consequently, dysregulation of PAX6 can cause cancer and a tight regulation is essential. The present thesis focuses on two mechanisms for regulation of PAX6, which are microRNAs (Study 1) and DNA methylation (Study 2 and Study 3), respectively. Study 1 examines three predicted microRNA-7 (miR-7) target sites within the 3'untranslated region (3'UTR) of the human PAX6 gene. A Luciferase reporter assay demonstrates that two of the predicted miR-7 target sites are functionally active and transient transfection of miR-7 mimics reduces PAX6 protein levels in a human cell line. Study 2 examines the DNA methylation profile of the PAX6 gene, specifically focusing on promoter CpG islands (Study 2a), exons (Study 2b) and intronic CpG islands (Study 2c). In neural induction of human pluripotent stem cells, PAX6 gene expression is first activated, and then repressed. Using targeted bisulfite sequencing and methylation-specific PCR it is revealed that the PAX6 gene is differentially methylated in a region-specific and differentiation stagedependent manner. Promoter CpG islands and 5'located exons remain predominantly non-methylated, whilst 3'located exons are constitutively methylated. In contrast, central exons and intronic CpG islands are differentially methylated with the highest level of DNA methylation at the neuroectoderm stage when PAX6 expression also peaks. Study 3 focuses on a potential link between DNA methylation and differential exon usage (DEU). The alternative exon 5a of PAX6 contains a noncanonical CpG dinucleotide (a common target site for DNA methylation) at its 5'donor splice site and a high CpG content on immediate up- and downstream exons. Further based on previous genome-wide studies suggesting that the DNA methylation status of alternative exons may influence exon inclusion and that exon 5a of PAX6 does not contain a single CpG site, it is hypothesized that the DNA methylation status of gene-specific CpG profiles on- and around alternative splice sites is associate with DEU. A novel approach for detection of DEU from whole-transcriptome sequencing (RNA-seq) data is developed and future integrative studies aim to examine a potential association with DNA methylation.

Endogenous retroviruses (ERVs) are found in genomes of all higher eukaryotes. As retrotransposition is deleterious, pathways have evolved to repress these retroelements. While DNA methylation transcriptionally represses ERVs in differentiated cells, this epigenetic mark is dispensable for maintaining proviral silencing during early stages of mouse embryogenesis and in embryonic stem cells (mESCs). Studies in diverse species have found histone H3K9 methylation and DNA methylation to function together to repress retrotransposons. However, until recently, little was known about the role of this histone modification in proviral silencing in mESCs. Interestingly, our analysis of mESCs lacking G9a, an H3K9-specific lysine methyltransferase (KMTase) revealed that although ERVs lost H3K9 di-methylation (me2) and DNA methylation, they remained silent. Strikingly, the levels of H3K9 tri-methylation (me3) remained unaltered, suggesting that this mark may instead be responsible for maintaining these parasitic elements transcriptionally inactive. The first stage of my research focused on identifying the enzyme depositing H3K9me3 at ERVs and on determining its role in proviral silencing. I discovered that Setdb1, another H3K9-specific KMTase, was indeed depositing H3K9me3 at a subset of ERVs and was required for maintaining transcriptional repression. Interestingly, this silencing pathway operated independently of DNA methylation. Through collaboration, we also discovered that this pathway played a diminished role in differentiated cells. Taken together, these findings indicate the existence of a DNA methylation-independent proviral silencing pathway in mESCs. The second stage of my research focused on the establishment of transcriptional repression of newly integrated proviruses. By employing an exogenous retroviral construct, I discovered a dramatic silencing defect in mESCs lacking G9a, which phenocopied cells depleted of the de novo DNA methyltransferases. Furthermore, efficient DNA methylation of proviruses required G9a-mediated H3K9me2. These findings reveal that histone modifications and DNA methylation function in concert to defend the genome against invading retroviral elements in mESCs. Taken together with discoveries regarding the mechanism of DNA demethylation in early embryos, I propose that cells undergoing DNA methylation reprogramming predominantly employ histone modification-based pathways to maintain these parasitic elements in a silent state; however, the establishment of transcriptional repression for newly integrated elements also requires de novo DNA methylation.

Background Dementia, including Alzheimer’s disease (AD), is a significant global health challenge, ranked as the seventh leading cause of death with nearly 10 million new cases annually. Despite recent advancements in anti-amyloid therapies such as aducanumab, lecanemab and donenamab, targeting AD in its symptomatic stages has often yielded limited success due to advanced neurodegeneration. Early intervention is crucial for effective treatment, making precise early diagnosis imperative. The current diagnostic process for AD often lacks biomarker support, leading to misdiagnosis rates of around 25-30% in primary care settings. Blood based biomarkers are emerging as a promising solution to enhance AD diagnosis and optimise clinical trial design. These biomarkers can detect core AD pathological indicators like A and phosphorylated tau, as well as neurodegeneration markers such as neurofilament light chain (NfL) and markers of neuroinflammation, such as Glial Fibrillary Acidic Protein (GFAP). Blood-based biomarkers offer the advantages of being minimally invasive, cost-effective, and accessible compared to costly amyloid PET and CSF biomarkers. DNA methylation has gained attention as a potential biomarker due to its association with various diseases, including some neurodegenerative diseases, including AD. While brain tissue studies have identified associations between DNA methylation and AD pathology, translating these findings to peripheral blood remains a challenge, meaning the relationship between brain and blood methylation patterns has not yet been fully elucidated. Aims: The overarching aim of this thesis was to identify markers of DNA methylation associated with AD risk and related phenotypes, such as cognition and neuroimaging modalities. Specifically, this thesis can be divided into two broad aims. Chapters 2 and 3 aim to investigate the association of DNA methylation clocks with AD-related traits and Chapters 4 and 5 aim to investigate the relationship between DNA methylation sites and AD-related traits, using both a targeted (Section 4.2) and an exploratory approach (Section 5.2). Initially, our focus was on evaluating the relationship between methylation age and AD-related phenotypes, with a specific emphasis on the use of DNA methylation clocks, as covered in Chapters 2 and 3. Chapter 2 focussed on using first- (Section 2.2) and second-generation (Section 2.3) clocks to assess whether accelerated ageing is associated with cross-sectional measures of cognition and pathological changes in the brain and determine if an individual’s current methylation age is an indicator of future changes in cognition and brain pathological features. It was then aimed to assess if methylation profiles within a priori candidate AD associated risk genes are associated with cross-sectional measures of cognition and pathological changes in the brain (Chapter 4). Finally, we aimed to undertake epigenome wide association studies with respect to AD-related phenotypes and to develop trait-specific methylation risk scores and test their association with AD-risk and related phenotypes (Section 5.2). Methods: All studies presented in this thesis utilised data from two well-characterised AD cohorts; the discovery cohort; the Australian Imaging, Biomarkers and Lifestyle Study of Ageing (AIBL) and the validation cohort; the Alzheimer’s Disease Neuroimaging Initiative (ADNI). Both the AIBL and ADNI are longitudinal cohort studies that have collected data at intervals over an extended period of greater than 15 years of follow up. All individuals in this thesis, across both cohorts, have available DNA methylation data as generated on the Illumina Infinium HumanMethylation EPIC 850k Chip. Chapters 2 and 3 consisted of assessing the relationship between first- and second-generation clocks with Positron emission topography (PET) derived measures of amyloid- (A) burden, MRI volumetric measures (white matter volume, grey matter volume, hippocampal volume and ventricle volume), cognition (the Pre-Alzheimer’s Cognitive Composite (PACC)) and age at onset of pathological levels of A in the brain (Α ΑAO ). Linear regressions were utilised in Chapter 2 to assess the association between accelerated age and trait outcomes. Cox proportional hazards regressions were utilised in Chapter 3 to assess the time taken to reach an event, where the event of interest was the occurrence of A AAO . For both Chapter 4 and 5, linear regressions were used to assess the relationship between CpG sites and AD-related traits. Chapter 4 relied on a targeted approach for analyses, whereas Chapter 5 utilised an unbiased discovery approach. Further, in Chapter 5, trait-specific methylation risk scores (MRSs) were developed, which relied on utilising the summary statistics from the EWAS to generate MRSs at 6 p-value thresholds and one machine learning based risk score, trained using Elastic Net. Results: Accelerated biological ageing shows no associations with cross-sectional and longitudinal measures of cognition, brain A burden, or longitudinal measures of brain volume; however, it is associated with cross-sectional measures of brain volume. In cognitively unimpaired individuals with a high amyloid- burden (CU A high ), an inverse association was observed with lower hippocampal volumes seen in those who were experiencing accelerated ageing, measured by the Hannum clock (Chapter 2; Section 2.2). Chapter 2, Section 2.3, focussed on the same analyses with the addition of second-generation clocks and one extra outcome variable, A AAO . Here, it was observed that accelerated ageing, measured by the PCPhenoAge clock, was associated with hippocampal volume in the whole cohort, reflecting the association seen in Section 2.2. The results were replicated in ADNI, however the effect observed was in the opposite direction. Further, when investigating the utility of second-generation clocks in the prediction of A AAO , associations in AIBL after FDR correction were observed, however these were not validated in ADNI (Chapter 3). When investigating individual CpG sites from genes previously associated with AD (Chapter 4), we identified numerous associations in the AIBL cohort that were significant after correction for the FDR, although only one novel site was validated in the ADNI cohort. This was located within the gene body region of CR1 (cg00416522) and was associated with ventricle volume in CU A high . Finally, when considering the genome in a set of exploratory epigenome wide association analyses, no individual CpG sites remained significant after correction for the FDR, however, as expected, several MRSs were associated with disease risk. Conclusions: The work presented in this thesis provides an in-depth investigation of peripheral DNA methylation patterns, methylation age and AD risk and related phenotypes. The results here suggest that while DNA methylation potentially has a role in AD development, further investigation is required to fully elucidate its mechanistic role. Whilst this thesis provides some evidence to support the involvement of DNA methylation in AD risk and related traits and adds to a growing body of literature aiming to describe the utility of DNA methylation as a peripheral biomarker, it also highlights several limitations and proposes future areas for research focus.

Cancer Biology & Genetics
M.S.
DNA methylation is an epigenetic mark with profound impact on gene expression and regulation. It is known to be altered both in cancer and throughout aging. These aging and cancer related changes are characterized by hyper-methylation of normally unmethylated CpG islands, and global DNA hypomethylation. Microbiota-free mice are known to live longer than their normal counterparts, and microbial dysbiosis is known to be a hallmark of colorectal cancer. In order to determine the microbiota’s ability to impact DNA methylation patterns, that in turn may influence aging phenotypes and cancer development, the DNA methylomes of germ-free (GF; no microbiome) and specific pathogen free (SPF; controlled microbiome) mice were analyzed with a DREAM assay. This was done in wild-type mice and in IL-10 KO mice, with or without the addition of the colitis-associated cancer inducing compound, azoxymethane (AOM). We found that individually, inflammation and the microbiota induce moderate changes in the methylation profiles of the large intestine. However, their effects on DNA methylation seem to synergize; in the presence of inflammation, SPF mice have highly different methylation profiles than GF mice. In addition, inflammation causes large methylation changes in specific-pathogen-free mice, but only moderate changes in germ-free mice. The inflammation and microbiota induced changes were characterized by hyper-methylation of sites with low methylation (CpG islands), and hypomethylation of sites with high methylation; these patterns resembled the DNA methylation drift seen during aging and in transformed cells. All sites subject to age-related methylation drift were vulnerable to the microbiome, but the converse was not true. A subset of sites was vulnerable to only the microbiome/inflammation, indicating multiple mechanisms of action. Overall, this research indicates the microbiota plays a key role in determining host DNA methylation state.
Temple University--Theses

Epigenetic regulation of gene expression is critical for normal human development and cellular differentiation. Although each somatic cell in the human body is genetically identical, epigenetic marks including the DNA methylation pattern are tissue-specific and critical for determining the vast array of cellular phenotypes. For studying respiratory disease, the airway epithelium is the ideal target tissue since it is the first point of contact for inhaled particles, viruses and airborne allergens. Cultured airway epithelial cells of asthmatic children show striking phenotypic differences compared to those from non-asthmatic individuals such as poor wound healing and enhanced expression of inflammatory cytokines. The altered state of the asthmatic epithelium could be due to underlying differences in gene expression associated with changes in their DNA methylation profile. Therefore, in this study, the purpose was to determine the extent to which DNA methylation, a reportedly stable epigenetic mark, is altered by standard cell culture conditions over passage for primary and immortalized bronchial epithelial cells. We analyzed four individual primary bronchial epithelial cell lines and a 1HBEO- immortalized bronchial epithelial cell line in triplicate. CpG DNA methylation was characterized during stages of cell propagation for 1505 CpG sites in promoter regions and/or first exons of 807 genes and compared with its gene expression. A comparison of 1HBEO- cells and primary bronchial epithelial cell lines revealed that 1HBEO- cells have significantly higher levels of overall methylation and more hyper- and heterogeneously-methylated CpG loci. In addition, there were a large number of CpG sites that had variable DNA methylation over passage in the primary cell lines but not in the immortalized cell lines. In summary, there were extensive differences in DNA methylation profiles between primary and immortalized bronchial epithelial cell lines during cell propagation, revealing the importance of this epigenetic modification in cell culture. This work is critical because, although little is known about the CpG specific changes that occur in cell culture, passaged cells are still regularly grouped together and phenotyped for major biological studies.

Arginine methylation is a post-translational modification occurring in higher eukaryotes that results in the addition of one or two methyl group on the nitrogen in the side chain of arginines. The enzymes responsible for protein arginine methylation have been classified in three groups. Type I enzymes promote the formation of both NG-monomethylated and asymmetric o-NG,NG-dimethylated arginines (aDMA). Type II enzymes catalyze the formation of monomethylated and symmetrical o-N G,N'G-dimethylated arginines (sDMA). The type III enzyme found in yeast catalyzes the monomethylation of the delta-guanidino nitrogen atom of the arginine residue. Although some abundant proteins have been described as being substrates for arginine methyltransferases for some time, there are still few known proteins to bear this modification. The primary goal of the work presented in this thesis was to identify new arginine methylated proteins and functionally characterize the roles of arginine methylation in new cellular processes. First, we generated four arginine methyl-specific antibodies: ASYM24 and ASYM25 are specific for aDMA whereas SYM10 and SYM11 recognize sDMA. Cell extracts were used to purify the protein complexes recognized by each of the four antibodies and the proteins were identified by microcapillary reverse-phase liquid chromatography coupled on line with electrospray ionization tandem mass spectrometry (LC/MS/MS). The analysis of 2 tandem mass spectra for each methyl-specific antibody resulted in the identification of 247 proteins, of which 197 are putatively arginine methylated.
The DNA repair MRE11/RAD50/NBS1 (MRN) complex was purified using one of the aDMA specific antibody. Since a role of protein arginine methylation in DNA damage checkpoint control and DNA repair had not been previously reported we chose to investigate the consequence of MRE11 methylation in DNA damage. Our results show that the MRE11 checkpoint protein is arginine methylated as determined by mass spectrometry and methylarginine-specific antibodies. The glycine-arginine rich (GAR) domain of MRE11 was specifically methylated by protein arginine methyltransferase 1 (PRMT1). Mutation of the arginines within MRE11 GAR domain severely impaired the exonuclease activity of MRE11. Cells treated with methyltransferase inhibitors displayed a DNA damage-resistant DNA synthesis phenotype and prevented the re-localization of the MRN complex to sites of DNA damage. Downregulation of PRMT1 with small interfering RNAs (siRNA) also yielded a damage-resistant DNA synthesis phenotype that was rescued with the methylated MRE11 complex. Taken together, the work presented in this thesis allowed the identification of many new potentially arginine methylated proteins and demonstrated a novel role for arginine methylation in the regulation of DNA repair enzymes and of the intra-S phase DNA damage checkpoint.

Maintaining appropriate patterns of gene expression in the gametes and during early embryogenesis is essential for normal development. DNA methylation is an epigenetic means of regulating gene expression and is an important molecular mark regulating the sex-specific expression of genes subject to genomic imprinting. Imprinted genes are expressed from only one of two inherited chromosomes and are differentially marked during gametogenesis to allow for their parental allele specific expression. These genes affect embryo growth, placental function, behavior after birth and are implicated in the etiology of a number of human diseases. The primary objective of this thesis was to gain a better understanding of the developmental dynamics and origins of DNA methylation profiles regulating maternally methylated imprinted genes during mouse oocyte development. Studies revealed that maternally methylated imprinted genes acquire methylation within their DMRs during postnatal oocyte growth and that this acquisition occurs in a gene and allele specific manner. It was also observed that maternal methylation imprint acquisition is related to oocyte diameter and that a repetitive parasitic element also acquires methylation during this period. DNA methylation is catalyzed by DNMTs and investigations into the developmental expression profiles of Dnmt3a, Dnmt3b and Dnmt3L indicated that transcript accumulation of these enzymes during oocyte development coincided with the timing of maternal methylation imprint establishment. Moreover, expression analysis in DNMT-depleted oocytes suggested these enzymes to be coordinately regulated. Additional studies aimed at developing another model of oocyte imprinting lead to the identification and characterization of a putative bovine Snrpn DMR. Its DNA methylation profile was found to be conserved with that of mouse and human. Snrpn DNA methylation analysis in bovine IVF and SCNT embryos revealed slight loss of methylatio

Methylation is the only known modification of DNA and in animals it mainly occurs at cytosines in a CpG context. The pattern of DNA methylation varies among organisms; some invertebrates are totally devoid of it, while others have densely methylated regions embedded in an otherwise unmethylated genome. The genome of mammals on the other hand, is very rich in DNA methylation with the exception of regions with high CpG frequency, known as CpG islands, that are often found devoid of methylation. Little is known about the factors that determine the genome-wide pattern of DNA methylation. Moreover, although there appears to be a specific developmental program for the establishment of methylation in specific genomic regions, the molecular events that lead to methylation establishment remain unknown. The establishment of methylation in the regulatory region of the murine Oct4 gene as well as the occurrence and establishment of methylation in mouse CpG islands are investigated in this study. The promoter of Oct4, which encodes an important developmental regulator, is known to gain methylation as the gene becomes silenced during early development. An in vitro model of murine early development has been used to recapitulate the events that lead to the gene’s silencing. In accordance to other reports, detailed methylation analysis of the gene’s entire upstream region and expression analysis showed that DNA methylation establishment follows the gene’s downregulation. Moreover, establishment of methylation at the Oct4 locus seems to start from the gene’s proximal enhancer and then spread towards the distal enhancer and the promoter. Although the initial establishment of methylation in the distal enhancer was not impaired in G9a -/- cells, methylation in these cells was unable to spread and accumulate. These findings demonstrate that the promoter of the gene is not the primary target for methylation as previously assumed and give rise to two possible mechanisms for DNA methylation establishment at this gene; one possibility is that methylation is actively targeted to the proximal enhancer, while the other is that the promoter and the distal enhancer are resistant to methylation, perhaps because of transcription factors bound to them. Moreover, the finding that G9a is not necessary for DNA methylation establishment but appears to have a role in methylation spreading, together with observations on the kinetics of the downregulation and the timing of methylation establishment, allowed the formation of a possible model for the role of DNA methylation in this gene’s downregulation. According to this model, DNA methylation acts to accelerate the gene’s downregulation ensuring its coordinated repression in the developing organism. For the study of methylation in CpG islands, first a novel algorithm was applied for the identification of CpG islands in the mouse genome. Approximately 21,000 CpG islands were identified in the mouse genome, half of which localised at the 5’ of genes, while the majority of the remaining was equally distributed in intragenic and intergenic regions. Only a very small proportion of the CpG islands localised at the 3’ of genes. When the gene ontology terms related with the CpG island-associated genes where interrogated, two main gene functions emerged as being preferentially associated with CpG islands, development and cell maintenance. Then, an affinity purification method, together with microarray hybridisation was applied for the identification of methylated CpG islands from mouse brain. Approximately 18% of all CpG islands were methylated in brain, with the big majority localised at 5’ and intragenic regions. When the gene ontology of the methylated CpG island-associated genes was analysed, developmental but not housekeeping genes were overrepresented in the methylated fraction. In order to further investigate the relationship of CpG islands with developmental genes, the same methodology was applied for the identification of CpG islands that become methylated after the in vitro induction of differentiation of ES cells. Although this approach failed to produce genome-wide data, it enforced the idea of a developmental program for CpG island methylation.

Preeclampsia is a leading cause of maternal and fetal death throughout the world. It is caused by placental dysfunction and clinically characterized by hypertension and other adverse outcomes. Early-onset preeclampsia (EOPET) is a severe form of the disorder. Despite much investigation, the underlying biology of EOPET is unclear. It is known that disrupted oxygen delivery and altered cellular differentiation are characteristics of preeclampsia placentas, and that this likely has an effect on the placental molecular profile. This thesis primarily investigates DNA methylation, a key component in regulating gene expression, in placentas and cellular states related to EOPET. Investigating placental cells exposed to hypoxia, we found 147 CpG sites in cytotrophoblast whose DNA methylation was significantly altered by exposure to hypoxia for 24 hours. Many of these sites overlapped with the 223 CpG sites that were altered between normoxic cytotrophoblast and syncitiotrophoblast, however the change was in the opposite direction (hypomethylated vs. hypermethylated), implying hypoxia can molecularly prevent differentiation in trophoblast cells. Expanding on these findings to look at DNA methylation in placental tissue from preeclampsia pregnancies, we found significant differences at 282 CpG sites. Several of these differences occurred in genes that have functional relevance for the development of EOPET. Many of the candidate genes also showed differential gene expression in preeclampsia placentas. To investigate the utility of these candidate CpGs as 1st trimester EOPET biomarkers, placentas with increased susceptibility to preeclampsia (trisomy 16) were investigated across gestational ages. There were many DNA methylation differences in 3rd trimester trisomy 16 placentas that were shared with chromosomally normal 3rd trimester EOPET placentas, suggesting a common molecular profile of preeclampsia prone placentas, regardless of etiology. Comparing 1st trimester trisomy 16 against 3rd trimester trisomy 16, we found 77 CpG sites differentially methylated in both conditions, and further found 3 changes in first trimester trisomy 16 shared with 3rd trimester EOPET. Overall, these studies have identified several molecular changes in EOPET and related conditions that provide insight into the biology of the disorder while also providing novel candidates to investigate further in a clinical setting.