|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
1 EPIGEN E TIC EFFECTS OF DIETARY SUPPLEMENTATION AND NUTRITION By JASON ORR BRANT A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2009
2 2009 Jason Orr Brant
3 To my Mom
4 ACKNOWLEDGMENTS I would like to thank my parents for all their love and support throughout this endeavor I would also like to thank Dr. Yang for the opportunity to perform my graduate research is his lab, and for all his help and mentoring these past years. This work would not have been possible without all the advice and support from friends and colleagues too numerous to list. You know who you all are, and know that I am grateful for your help and friendship.
5 TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................................... 4 LIST OF TABLES ................................................................................................................................ 7 LIST OF FIGURES .............................................................................................................................. 8 ABSTRACT ........................................................................................................................................ 10 CHAPTER 1 INTRODUCTION AND LITERATURE REVIEW ................................................................. 12 Epigenetics ................................................................................................................................... 12 DNA Methylation ................................................................................................................ 12 Role of Methylation ............................................................................................................. 13 Repetitive Elements .................................................................................................................... 15 Global DNA Methylation Levels ............................................................................................... 16 DNA Methylation and Development ......................................................................................... 18 Imprinting .................................................................................................................................... 18 Nutrition and DNA Methylation ................................................................................................ 21 Developmental Origins of Health and Disease Hypothesis .............................................. 24 Folic Acid and DNA Methylation ...................................................................................... 25 Maternal Nutrition and Fetal Health ................................................................................... 27 2 EPIGENTETIC EFFECTS OF A LOW PROTEIN DIET IN UTERO ................................... 36 Introduction ................................................................................................................................. 36 Results .......................................................................................................................................... 38 Pair -fed Animal Feedings .................................................................................................... 38 Methylated DNA Immunoprecipitation (MeDIP) ............................................................. 38 MeDIP on Experimental Animals ...................................................................................... 41 Amplification and Labeling of DNA .................................................................................. 41 Analysis of Promoter Array Data ....................................................................................... 43 Locus -Specific DNA Methylation Analysis .............................................................................. 46 DNA Methylation Analysis by Sodium Bisulfite Genomic Sequencing ......................... 47 mRNA Expression ............................................................................................................... 49 Discussion: ................................................................................................................................... 49 3 ANALYSIS OF DNA METHYLATION IN RESPONSE TO CHRONIC FOLIC ACID SUPPLEMENTATION AND WITHDRAWAL IN CHINESE WOMEN ............................. 70 Introduction ................................................................................................................................. 70 Results .......................................................................................................................................... 72
6 Methylation Specific PCR ................................................................................................... 72 Bisulfite Genomic Sequencing Analysis of Repetitive DNA Methylation ...................... 73 Bisulfite Genomic Sequencing Ana lysis of SNRPN Promoter ......................................... 75 Pyrosequencing Analysis of L1 Elements .......................................................................... 77 Bisulfite Genomic Sequencing Analysis of Subjects with a Large Decrease in DNA Methylation after Washout .............................................................................................. 77 Discussion .................................................................................................................................... 79 4 MATERIALS AND METHOD S ............................................................................................... 91 Animals ........................................................................................................................................ 91 Genomic DNA Extraction .......................................................................................................... 91 Methylated DNA Immunoprecipitation (MeDIP) ..................................................................... 92 Sonication ............................................................................................................................. 92 Immunocapture .................................................................................................................... 93 Quantitative RT PCR of MeDIP ................................................................................................ 94 Amplification of Immunoprecipitated DNA ............................................................................. 95 Uracil Incorporation, DNA Fragmentation and Labeling ......................................................... 95 Electrophoresis Mobility Shift Assay (EMSA) ......................................................................... 96 Hybridization and Scanning Mouse Promoter Arrays .............................................................. 97 Data Analysis using Parteks Genomic Suite Software Package ............................................. 98 High Resolution Sodium Bisulfite Genomic Sequencing ........................................................ 98 RNA Purification ......................................................................................................................... 99 Reverse Transcriptase Real Time PCR for Expression .......................................................... 100 5 DISCUSION AND FURTURE DIRECTIONS ...................................................................... 104 LIST OF REFERENCES ................................................................................................................. 111 BIOGRAPHICAL SKETCH ........................................................................................................... 125
7 LIST OF TABLES Table page 3 1 List of subjects analyzed by sodium bisulfite sequencing (BGS) at the SNRPN promoter. ................................................................................................................................. 81 4 1 Methylated DNA immunoprecipitation (Me DIP) primers. ............................................... 101 4 2 Sodium bisul fite genomic sequencing primers. ................................................................. 102 4 3 Reverse Transcriptase qRT PCR primers. .......................................................................... 103
8 LIST OF FIGURES Figure page 1 1 Parent of origin specific changes in DNA methylation during embryogenesis. ................ 31 1 2 Erasure of DNA methylation during gametogenesis .......................................................... 32 1 3 Schematic representation of the H19 / Igf2 imprinted domain. ............................................. 33 1 4 Schematic representation of the role of CT CF in long range transcriptional regulation in imprinted domains. ............................................................................................................ 34 1 5 The role of folate in one -carbon metabolism. ...................................................................... 35 2 1 Schematic representation of pair -feeding study. .................................................................. 52 2 2 Weight gain of pregnant dams throughout pregnancy. ........................................................ 53 2 3 Pair Feeding Study. ................................................................................................................ 54 2 4 Outline of methylated DNA immunoprecipitation (MeDIP) ............................................. 55 2 5 Verification of enrichment of methylated DNA by qRT PCR. ........................................... 56 2 6 qRT -PCR results of MeDIP enrichment of NPD and LPD samples. .................................. 57 2 7 Electro -mobility shift assay.. ................................................................................................. 58 2 8 Heat map of Man1a Partek analysis. .................................................................................... 59 2 9 Heat map of Cdca8 Partek analysis. ..................................................................................... 60 2 10 Bisulfite Genomic Sequencing Data of H19 DMR. ............................................................. 61 2 11 Bisulfite Genomic Sequencing Data of Igf2 DMR1. ........................................................... 62 2 12 Bisulfite Genomic Sequencing Data of Igf2 DMR2. ........................................................... 63 2 13 Bisulfite Genomic Sequencing Data of Snrpn. .................................................................... 64 2 14 Bisulfite Genomic Sequencing Data of Mkrn3 ................................................................... 65 2 15 Bisulfite Genomic Sequencing Data of Man1a. ................................................................... 66 2 16 Bisulfite Genomic Sequencing Data of Cdca8. .................................................................... 67 2 18 mRNA Expression. ................................................................................................................ 69
9 3 1 Methylation Specific PCR results for Tumor Suppressor Genes ....................................... 82 3 2 Bisulfite Genomic Sequencing of L1. ................................................................................... 83 3 3 Bisulfite Genomic Sequencing of SNRPN at 0 months. ...................................................... 84 3 4 Bisulfite Genomic Sequencing of SNRPN at 0 months. ...................................................... 85 3 5 Bisulfite Genomic Sequencing of SNRPN at 9 months. ...................................................... 86 3 6 Bisulfite Genomic Sequencing of SNRPN ........................................................................... 87 3 7 Pyrosequencing Analysis of L1 Repetitive Element. ........................................................... 88 3 8 Bisulfite Genomic Sequencing of subject 2624. .................................................................. 89 3 9 Bisulfite Genomic Sequencing of subject 3126. .................................................................. 90
10 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy EPIGEN E TIC EFFECTS OF DIETARY SUPPLEMENTATION AND NUTRITION By Jason Orr Brant May 2009 Chair: Thomas P. Yang Major : Medical Sciences --Biochemistry and Molecular Biology DNA methylation is associated with longterm repression of transcription, and has generally been considered a fairly stable epigenetic mark C hanges in DNA methylation patterns were thought to normally only occur during embryonic and germ cell development, however recent reports have indicated nutritional insults, exposure to environmental toxins and aging can in fact alter D NA methylation patterns. Two experimental approaches were utilized i n order to gain a better understanding of how nutritional insults result in altered DNA methylation. The first approach focused on the effects on DNA methylation due to exposure to a low p rotein diet in utero on both a genome -wide scale and at specific loci. The results indicate that there were subtle changes in DNA methylation occurring at differentially methylated regions within the H19 / Igf2 imprinted domain in animals exposed to a low p rotein diet during development. Additionally, it was demonstrated that the expression of H19 and Igf2 was also altered. These findings are important in that they indicate that exposure to nutritional insults in utero can induce epigenetic changes in offspr ing, however the genome -wide analysis indicates that that was no major changes occurring in DNA methylation in low protein c ompared to normal protein animals. T he second approach involved examining the effects of folic acid supplementation and withdrawal i n Chinese women of child bearing age on DNA methylation Although the success
11 in prevention of neural tube defects has been dramatic, and folic acid is generally c onsidered to be safe, the long -term consequences of increased folate levels have yet to be extensively studied in long term clinical trials Folate plays a major role in one carbon metabolism and is involved in the methylation of DNA. In order to examine the effects of increased folate levels on DNA methylation, blood samples were analyzed from a population -based, randomized trial of folic acid supplementation and withdrawal The results indicate that folic acid supplementation and withdrawal produce d changes in DNA methylation in a locus -specific manner. Methylation Specific PCR of the promoter s of tumor suppressor genes (TSGs) indicate that there was an observable increase in DNA methylation after 6 months of folic acid supplementation in two out of ten subjects analyzed which was no longer detectable after 3 months of withdrawal N o wide spre ad hypermethylation of TSGs were detected There were also dramatic changes in DNA methylation at the maternally imprinted SNRPN promoter with a near complete loss of DNA methylation after 6 months of folic acid supplementation, and a complete loss after 3 months of folic acid withdrawal. A nalysis of the L1 repetitive element determined that there was no major change in DNA methylation occurring at this element Overall, this study has demonstrated that exposure to nutritional insults in utero can induce b oth changes in DNA methylation and in gene expression levels, and that dietary supplementation and withdrawal in an adult population can induce changes in DNA methylation. Th ese results warrant further stud ies into the biological significance of these obse rved changes.
12 CHAPTER 1 INTRODUCTION AND LITERATURE REVIE W Epigenetics Epigenetics is a process or mechanism that affect s heritable changes in gene expression independent of the underlying linear DNA sequence. Epigenetic mechanisms include modification s of chromatin that can shape the transcriptional memory of a cell. This collection of modifications complements the genetic information to determine which genes are transcribed and at what levels (Jaenisch and Bird, 2003; Wu and Morris, 2001) Epigenetic modifications can be s table and heritable through cell mitosis and meiosis. However, these epigenetic marks are also dynamic and may undergo both global and locus -specific changes, especially during development and differentiation, to establish appropriate l evels of transcripti on in a cell -type specific manner, and throughout specific stages of development (Li, 2002; Reik et al., 2003) Once established, these epigenetic modifications can then be maintained in terminally differentiated cells. DNA Methylation Methylation of DNA is an important epigenetic modification in mammalian genomes. The covalent addition of a methyl group to the 5 carbon position of cytosine is catalyzed by a reaction involving of family of proteins known a s the DNA m ethyltransferases (DNMTs) and in mammals occurs predominately in the context of CpG dinucleotides (Robertson, 2002) which are underrepresented and non randomly distribut ed in the genome (Jones and Takai, 2001; Wilson et al., 2007) The mammalian genome consists mainly of CpG poor regions punctuated by CpG rich regions, termed CpG isla nds (Miranda and Jones, 2007) which were first defined by Jones et al. as being longer than 500 bases, having a GC content greater than 55% and an observed CpG/expected CpG ratio of 0.65 (Jones and Takai, 2001) The majority of CpG islands
13 are found within promoter regions, and about 40% of genes contain CpG islands within the 5 region of the gene (promoter, untranslated region (UTR), and first exon) (Jones and Baylin, 2002) The majority of CpG islan ds are hypomethylated whereas CpG poor regions, such as intergenic and intronic regions tend to be hypermethylated (Miranda and Jones, 2007) The mammalian genome contains only a small amount of protein coding DNA, with the overwhelming majority of the genome consisting of introns, repetitive elements, and parasitic DNA. Repetitive and parasitic DNA contains pote ntially active transposable elements that need to be stably silenced to ensure genomic stability and maintain integrity of the genome (Jones and Baylin, 2002; Jones and Laird, 1999) The DNMT family of proteins can be loosely categorized into two main types, based upon their preferred DN A substrate and timing of methylati on De novo DNA methylation which is catalyzed by DMNT3a and DNMT3b (Okano et al., 1999) establishes the methylation patterns during gametogenesis, post -fertilization fetal development and in the differentiation of cells. Once established, these methylation patterns need to be stably maintained thr oughout mitosis and are done so through the actions of the maintenance methyltransferase DNMT1. DNMT1 preferentially recognizes hemimethylated DNA and remethylates newly synthesized daughter strands of DNA during replication, thus ensuring the methylation patterns are maintained during mitosis (Bacolla et a l., 1999; Flynn et al., 1996; Glickman et al., 1997; Pradhan et al., 1997; Yokochi and Robertson, 2002) Role of M ethylation Although DNA methylation is generally associate d with transcriptional repression, the exact mechanism of how DNA methylation leads to repression is unclear. DNA methylation has been implicated in a diverse array of pro cesses, including transcriptional regulation X chromosome inactivation, silencing of repetitive DNA and transposable elements, genomic imprinting, chromatin structure and genomic stability (Baylin et al., 2001; Jones and Laird, 1999;
14 Robertson, 2001) Although the mechanism of repression is still not clear, numerous models for how DN A methylation represses transcription have been proposed. One model proposes that DNA methylation directly interferes with the binding of regulatory proteins and transcription factors to their target sites. This inability of transcription factor binding le ads to repression (Deng et al., 2001; Eden and Cedar, 1994; Rhodes et al., 1994) However, most models suggest that DNA methylation changes the interactions between DNA and prote ins, thus inhibiting transcription initiation or that DNA methylation leads to conformational changes in chromatin structure which lead to chromatin condensation and t o transcriptional repression (Harikrishnan et al., 2005; Jones et al., 1998; Nan et al., 1998) The methyl -binding domain (MBD) proteins (Hendrich and Bird, 1998; Sansom et al., 2007) are a family of proteins that can bind to methylated DNA and recruit other repressive complexes via association with a transcription re pressor domain (TRD) (Nan et al., 1998) This allows MBPs to target chromatin remodeling proteins, such as histone deacetylases (HDAC), to methylated DNA (Feng and Zhang, 2001) as well as to form complexes with other MBD family members and DNMT1 to ensure repression is maintained after replication (Feng et al., 2002) The role of DNA methylation in transcriptional repression of repetiti ve elements is of critical importance. The majority of the mammalian genome is comprised of repetitive elements, many of which contain long terminal repeats ( LTRs ) capable of function ing as promoters if not silenced, allowing the transcription and spreading of these parasitic DNA elements. (Yoder et al., 1997) The movement of these parasitic elements can have devastating effects on the integrity of the genome, as their expression can lead to nonallelic recombination, as well a s insertion into and subsequent disruption of genes, altering their transcription and or function (Kazazian and Moran, 1998) Cells in which DNMT1 h ave been knocked out contain only about 30% of their
15 global DNA methylation levels and have a ten -fold increase in chromosomal rearrangements (Chen et al., 1998) Repetitive Elements Repetitive sequences make up about 45% of the human genome (Rollins et al., 2006) (Lander et al., 2001) Repetitive sequences are divided into 4 categories: DNA transposons, which a re about 3% of the human genome, r etrotransposons, and endogenous retroviruses which range from 40 42% of the human genome and satellite DNA which are simple repeat sequences, with repetitive units ranging from 2 70 bases aligned in tandem. R etrotransposons and endogenous retroviruses can be further classifi ed based on the presence or absence of a LTR. Among the non -LTR repetitive elements, there are both autonomous elements, capable of integrating copies of itself without assistance, and nonautonomous elements, which require the action of other elements in order to become integrated into the genome. Non LTR autonomous elements are also known as Long Interspersed Nuclear Elements, or LINE 1, or simply L1 elements. The most promin ent member of the non autonomous elements is comprised of Short Interspersed Nucl ear Elements, or SINEs, of which the Alu family of repeats is the most abundant member (Reviewed by Wilson et al. ) (Wilson et al., 2007) With the discovery that tumors show global hypomethylation of repetitive elements in relation to tumorogenesis and cancer (Riggs and Jones, 1983) a nalyzing the methylation status has led to important insights. For instance, both LINEs and SINE s h ave been found to be hypomethylated in cancer cells. Additionally, s atellite DNA has also been reported to become hypomethylated in a variety of cancers (Jeanpierre, 1994) Satellite 2 (Sat2) is predominately localized in the peri c entromeric heterochromatin of select human chromosomes. Additionally Sat the major component of human centromeres (Lee et al., 1997) also becomes
16 hypomethylated in cancers (Ehrlich, 2002) It is estimated that 35 40% of all methylation in the genome occurs in repetitive elements (Bestor, 1998; Kochanek et al., 1993; Schmid, 1998) For this reason, examining the methylation levels of various types of repeat DNA is an excellent surrogate marker for analyzing total genomic methylation levels (Yang et al., 2004) Global DNA M ethylation Levels In healthy somatic human cells, CpG methylation ranges from 7090% (Miranda and Jones, 2007) These methylated dinucleotides comprise approximately 0.75-1% of the total bases in the human genome (Ehrlich et al., 1982; Tuck-Muller et al., 2000) As discussed earlier, most CpG islands are unmethylated, while the remaining CpG dinucleotides found within intergenic and intronic regions are hypermethylated. Some CpG islands within gene promoters become methylated as a normal process of development and tissue specific differentiation, genomic imprinting, as well as during the process of X -chromosome inactivation in females and as a consequence of the aging process (Hellman and Chess, 2007; Jones, 1999; Reik et al., 2001; Richardson, 2003) However, it is becoming increasingly clear that abnormal methylation, be it locus -specific hyper or hypomethylation, or global hypomethylation can lead to aberrant gene expression and contribute to disease formation. Hypermethylation induced ge ne silencing at the promoters of tumor suppressor genes can greatly increase the risk of initiation of cancer development. For example, methylation -mediated silencing of the DNA repair protein MLH1 has been shown to be a factor involved in colorectal cance r and an increase in microsatellite instability (Cunningham et al., 1998) A now well recognized hallmark of cancer is global hypomethylation with locus -specific hypermethylation, prompting Feinberg and colleagues to state Although individual genes vary in hypomethylation, all tumors examined so far, both benign and malignant, have shown global reduction of DNA methylation This is a striking feature of neoplasia. (Feinberg et al., 2006)
17 Transposable elements, which are normally silenced by DNA methylation, are DNA sequences with the ability to integrate into the genome at different sites. As discussed above, DNA transposons, retrotransposons and endogenous retroviruses account for almost half of the mammalian DNA content (Kazazian, 2004) T he two most abundant types of retrotransposons are auto nomous LINE family and the non autonomous SINE fam ily LINE, or L1 elements possess a strong internal promoter and contains 2 open reading frames ( ORFs ) which encode for an RNA binding protein and a reverse transcriptase with endonuclease activity (Feng et al., 1996; Gama Sosa et al., 1983) which enable them to integrate anywhere in the genome. M ost L1 elements exist in a state where the 5 end is truncated (Ostertag and Kazazian, 2001) the ORFs contain mutat ions (Skowronski et al., 1988) or both, rendering them essentially inactive However some subfamilies can still be transcribed when activated, with an estimated 80 to 100 active L1 elements per diploid genome (Brouha et al., 2003; Deininger et al., 2003; Kazazian, 2004) LINE hypomethylation can occur early in cancer in itiation, and hypomethylation has been observed in a number of cancers relative to adjacent healthy tissue (Hoffmann and Schulz, 2005; Suter et al., 2004; Takai et al., 2000) The SINE family is a non autonomous el ement which relies on the action of the L1 proteins to facilitate transposition (Dewannieux et al., 2003) The most abundant SINE s in hum ans is the Alu family of repetitive element s Alu does not encode any proteins, but has expanded to over 1.1 million copies and now accounts for 11 13% of the human genome (Esnault et al., 2000; Wei et al., 2001) Although Alu elements are methylated in somatic cells, the maintenance of methylation appears to vary greatly between individuals (Sandovici et al., 2005) Alu demethylation has not been extensively studie d in disease and does not seem to correlate well with global hypomethylation (Weisenberger et al., 2005)
18 DNA Methylation and D evelopment Although DNA methylation is generally considered to be a fairly stable epigenetic mark there are two developmental periods when dynamic changes in genome -wide DNA methylation pa tter ns occur; during embr yonic and germ cell development. E mbryogenesis begins with fertilization to form a single -celled zygote, which progresses to form a multicellular organism with over 200 functionally distinct and diverse cell types (Mann and Bartolomei, 2002) E ach differentiated cell type has its own epigenetic signature which reflects genotype, developmental history and environmental influences, which is then manifested in the cell s phenotype (Nafee et al., 2008) During normal development, these cells must undergo major epigenetic reprogramming. During germ cell development, a significant part of the genome is demethylated, and becomes remethylated in a cell or tissue -specific manner (Reik et al., 2001) The first phase of methylation reprogramming occurs after fertilization and before the formation of the blastocyst. Post -fertilization, a rapid paternal -specific loss of methylation is observed (Dean et al., 2003; Mayer et al., 2000; Oswald et al., 2000) T his putative active demethylation is complete before replication begins in the paternal pronucleus but spares paternally imprint ed genes, heterochromatin around centromeres and some repetitive elements (Morgan et al., 2005) D uring subsequent replication in the embryonic tissue, there is a passive loss of methylation due to a lack of maintenance methylation during DNA replication (Bestor, 2000) U pon implantation, de novo methylation begins to reestablish DNA methylation patterns, thus ensuring proper patterns of gene e xpression within the developing embryo (Li, 2002; Reik et al. 2001; Santos and Dean, 2004) ( figure 1 1 ). Imprinting In diploid organisms, the overwhelming majority of autosomal genes are expressed from or repressed on both parental copies, in a biallelic manner. There is however a subset of genes,
19 known as imprinted genes, in which the transcriptional competence of the gene is determined by parental inheritance. This results in functionally different alleles of the same gene within the same cell. As the DNA sequence of these alleles is essentially identical, the functional differences must be imparted through an epigenet ic modification of one or both alleles (Reik and Walter, 1998; Tilghman, 1999) Genomic imprinting was first discovered almost 25 years ago as a result of nuclear transplantation experiments in mouse (McGrath and Solter, 1983, 1984; Surani and Barton, 1983; Surani et al., 1984, 1986) These elegant experiments demonstrated that mammalian development requires genetic information from both the paternal and maternal genomes. Both diploid androgenetic and diploid gynogenetic embryos failed to thrive suggesting that there were genes exclusively expressed from one parental genome and the failure to thrive of uniparental embryos was due to loss of function of these genes The first imprinted gene to be discovered was the Insulinlike growth factor 2 gene ( Igf2 ), which is a fetal -specific growth facto r. Target ed mutations in this gene resulted in a heterozygous fetal undergrowth phenotype, but only when the mutated gene was inherited paternally (DeChiara et al., 1991) Additionally, the exten t of the growth inhibition was identical in paternal heterozygotes and homozygotes, indicating that the entire extent of Igf2 activity was contributed from the paternal genome (DeChiara et al., 1991) In the last decade and a half close to 200 genes have been identified as imprinted or are predicted to be imprinted. (Information on the current state of imprinted genes can be found at geneimprint.com) Imprinted genes tend to be clustered together and organized in large chromosomal domains (Delaval and Feil, 2004; FergusonSmith and Surani, 2001; Reik and Walter, 2001) These imprinted domains generally contain both paternally and maternally imprinted genes, as well as both protein coding genes and non -coding RNA genes. This chromosomal organization
20 of imprinted genes is often conserved among mammalian species (Reik and Walter, 1998; Verona et al., 2003) Each i mprinted cluster is generally under the control of a cis acting element, termed the imprinting control region (ICR), which coordinately regulates gene expression throughout the imprinted domain (Reik et al., 2003) The ICR acquires differential DNA methylation patterns in a parent of origi n manner in the developing germ cells, when the parental genomes are separated, thus allowing differential modification (F igure 1 2 ). These methylation marks are then stably maintained during development in all tissues w h ere the imprint is recognized. This differential methylation mark must also be able to be erased and reset to a parent of origin specific imprint during germ cell development (Delaval and Feil, 2004) The H19 / Igf2 imprinted domain is probably the most well -studied imprinted domain. The imprinted locus is located on distal mouse chromosome 7 (Figure 1 3) and has been implicated in a number of congenital growth abnormalities (Henry et al., 1991) Insulin like growth factor 2 (Igf2 ) encodes a fetal -specific growth factor that is widely expressed d uring mouse fetal development from the paternal allele, and is of particular importance in placenta growth (Constancia et al., 2002) In mice, Igf2 is highly expressed in the embryo, but expression is essentially nonexistent in the adult (DeChiara et al., 1991) In adult humans, IGF2 is expressed biallelically from an adult -specific promoter (de Pagter Holthuizen et al., 1987) which appears to have evolved to become inactivated in mice (Rotwein and Hall, 1990) Downstream of Igf2 is the maternally imprinted H19 gene which encodes a non-translated RNA (Bartolomei et al., 1991) H19 is also highly expressed during fetal development, particularly in tissues of mesoderm and endoderm origin (Bartolomei and Tilghman, 1997; Poirier et al., 1991) Human and mouse H19 have extensively conserved secondary structure characteristics, suggesting that
21 although the RNA is not translated, it may p lay an important biological role (Brannan et al., 1990) Located 2 4 kb upstream from the maternally imprinted H19 gene is a differentially methylated region (DMR) which is hypermethylated on the paternally inherited allele (Brandeis et al., 1993; Tremblay et al., 1997) This DMR acts as an imprinting control region (ICR) for a rather complicated hierarchy of multiple DMRs that are responsible for maintaining proper imprinted gene expression throughout this domain. Igf2 contains four separate promoters and three d ifferent DMRs (Lopes et al., 2003) T he 5 most promoter and DMR (DMR0) of Igf2 has been shown to be placenta specific, and is the only maternally methylated DMR in the region (Moore et al., 1997) The paternally methylated DMR1 is located upstream of the fetal Igf2 promoters and contains a methylation-sensitive silencer (Constancia et al., 2000; Eden et al., 2001) DMR2 is located in the last exon of Igf2 and contains a methylation -sensitive activator (Murrell et al., 2001) (F igure 1 3 ). Expression of both H19 and Igf2 appears to be under the control of multiple downstream enhancer element s that can modulate expression of either H19 or Igf2 T he H19 DMR, or ICR, contains binding sites for CCCTC binding factor (CTCF). Methylation of the DMR prevents binding of CTCF to the paternal allele thus allowing the enhancer to act on the Igf2 promoter. On the maternal allele, CTCF binds to the unmethylated maternal allele and acts as an insulator, blocking enhancer access to Igf2 and driving expression at the unmethylated maternal H19 promoter (F igure 1 4 ) (Bell and Felsenfeld 2000; Hark et al., 2000) Nutrition and DNA Methylation Throughout most western countries, cardiova scular and respiratory diseases and cancer account for fully three quarters of all mortalities in adults (Murray and Lopez, 1994) For this reaso n, chronic diseases are becoming the major focus of health care related problems. T here is a growing body of evidence suggesting that chronic diseases may originate in response to
22 nutrition al insults during in utero development (Roseboom et al., 2001b) (Lillycrop et al., 2 005; Waterland and Jirtle, 2003) In both human and animal studies, both epidemiological and experimental, evidence suggest nutrient deprivation in utero can have adverse long -term effects on the metabolic and ph ysiological states of offspring Epidemiological studies looking at standard obstetric birth records show that l ow birth weight has been li n ked with increased incidence of hypertension (Law et al., 1991) non insulin-dependent diabetes (Phillips et al., 1994) chronic bronchitis and coronary heart disease (CHD) (Barker et al., 1990). C ancer has been linked with a high birth weight in sim ilar epidemiological studies (Hjalgrim et al., 2003; Michels et al., 1996; Michels and Xue, 2006) Unfortunately the intrauterine c onditions that led to these results cannot be determined by these types of studies, and many factors could contribute to low birth weight. The association between small birth weight and CHD ha s been studied in several countries (Barker et al., 1990; Barker and Osmond, 1986; Barker et al., 1993; Barker et al., 1989; Hales et al., 1991) T hese studies depended on size at gestational age (SGA) rather than prematurity (Eriksson et al., 1999; Leon et al., 1998; Osmond et al., 1993) De ath rates among men who were thin at birth as recorded by a low ponderal index but had accelerated weight gain during childhood had the highest death rates from CHD. The ponderal index determines an individuals leanness and is similar to the body mass index (BMI). Th ese men had a 5 -fold increase in mortality compared to men with high birth weights that were lean in childhood. This effect rate is the highest observed in cardiovasc ular epidemiology. The authors suggested that the increased death rate may have been due to poor prenatal nutrition followed by improved postnatal nutrition (Eriksson et al., 1999) The results of these epidemiological studies have led t o the Barker Hypothesis, or the fetal origins hypothesis which states that the fetus
23 adapts to a limited supply of nutrients in utero which in turn permanently alters its physiology and metabolism, leading to an increased risk of adult onset disease (Barker, 1995). T he Dutch famine of 194445 has provided a wealth of epidemiological data relating malnutrition during development to adult onset chronic degenerative diseases. During World War II, due to a ban on all food transport in the Netherlands and an unusually harsh winter which blocked passage from the rural east to the urban west the daily food rations fell below 1000 calories per day, and at the height of the famine, between 400 and 800 calories per day i n the western cities of the Netherlands During this disastrous famine, women we re still able to conceive and give birth, and these offspring were studied throughout their lives, and have provided an enormous opportunity to study the effects of malnutrition during gestation and correlate this to health related issues in adult life (Roseboom et al., 2001a) For the study, the obstetric records of 2424 off spring were included and 741 adults agreed to attend the cl inic for extensive measurements. T hree periods of 16 weeks were used to distinguish between babies exposed during earl y, m id and late gestation Babies born before the famine, or those conceived after the famine were termed unexposed and used as comparison. People exposed to famine in early gestation appeared to h ave a higher risk of coronary heart disease (Ravelli et al., 1999) had a higher BMI (Roseboom et al., 2000a; Roseboom et al., 2000b) and a more atherogenic lipid profile (Roseboom et al., 2001a) and those exposed to famine in later stages of development showed in creases in the occurrence of obstructive airways disease (Lopuhaa et al., 2000) People who were exposed to famine in early gestation were also more likely to rate their overall health as poor (Roseboom et al., 2001b) Critics have pointed out that many variables, besides extreme shortage of food, must also be taken into accountl (Huxley et al., 2002; Huxley and Neil, 2004) The famine coincided with an unusually harsh winter This, combined with the stress of war,
24 absence of their spouses, and a general decline in basic services, coinciding with widespread infection are confounding fac tors that need to be considered However, the mechanism by which early nutritional insult leads to an increase i n disease susceptibility and altered metabolism in adult life, remains, at best, poorly understood. One possibility is that epigenetic changes in the placenta and/or fetus due to exposure to maternal malnutrition during development may lead to stable, heri table, aberrant changes in the regulation of genes important for proper placental and fetal development. As discussed earlier, m ethylation of CpG dinucleotides is a well characterized epigenetic modification generally involved in silent chromatin. As DNA m ethylation patterns are reprogrammed in the early embryo and maintained throughout adult life, it is possible that early deprivation of nutrients could lead to alterations within the establishment of proper methylation imprints (Waterland and Garza, 1999) Many genes responsible for proper placental and fetal development are imprinted in placental mammals, (Constancia et al., 2002) and since methylation imprints are established during gametogenesis (Reik et al., 2001) it has been pr o posed that exposure to nutritional insults during gametogenesis and placental and fetal development will lead to alterations in the epigenetic states of imprinted genes, specifically aberrant DNA methylation at differentially methylated regions (Lillycrop et al., 2005; Waterland and Garza, 1999) This hypothesis will be discussed in greater detail below. Developmental Origins of Health and Disease Hypothesis As early as the time of Hippocrates the concept that adult diseases may originate during the process of development has been proposed Since then, epidemiological studies and experimental data have contributed to an ever changing hypothesis. As recently as 40 years ago, Rose published observations of familial patterns of coronary heart disease (Rose, 1964) Shortly thereafter, Fordsdahl showed that poor living conditions in early life were important risk factors for arteriosclerotic heart disease (Forsdahl, 1977) These hypotheses were greatly expanded by
25 the work of Barker et al. (Barker, 1995; Barker and Osmond, 1986) which would later le a d to the concept of the fetal origins of adult disease. Barker termed this the thrifty phenotype, which states that environmental cues during development may influence development in such a way as to prepare the fetus for a predicted environment in adult life. This concept of developmental plasticity suggested that these adaptations may alt er metabolism in a way as to be detrimental in later life. At the same period as Baker and colleagues were performing their work, Trichopoulos proposed a similar hypothesis, the fetal origins of cancer, for the origination of breast cancer in utero (Trichopoulos, 1990) Realization that developmental plasticity extends into the postnatal period led to the ter m developmental origins hypothesis (Gluckm an and Hanson, 2004) Waterland and Garza demonstrated that epigenetic changes can occur in a limited period of opportunity during development and that these changes persist into adulthood and this hypothesis was termed metabolic imprinting (Waterland and Garza, 1999) All of these refinements in nomenclature of these hypothes e s were intended to highl ight specific biological mechanisms that could be unified into a single hypothesis, now termed the developmental origins of health and disease (DOHaD) (Waterland and Michels, 2007) Thus when conducting epidemiological studies of adult disease, it is important to recognize that both the genome and the epigenome interactively influence sensitivity to disease in adult life (Dolinoy et al., 2007) Folic A cid and DNA M ethylation F olate is a water soluble B vitamin that is essential for the synthesis of nucleotides and plays a major role in one carbon metabolism (Mackenzie, 1984) F olate, in its various coenzyme forms acts as both a methyl donor and acceptor in o ne -carbon metabolism, with its most prominent role being involved in the remethylation of homocysteine to methionine (Figure 1 5.) Methionine is a precursor of S -Adenosylmethionine (SAM), which serves as the universal one carbon donor involved in methylat ion o f DNA, RNA, lipids and proteins (Lamprecht and Lipkin,
26 2003; Lucock, 2000) (Figure 1 5) Because of this essential role, perturbations in the levels of folate can have profound effects on both nucleotide synthesis and methylation of DNA both of which are critical in maintaining the integrity of DNA and the proper regulation of gene expression (Dolinoy and Jirtle, 2008; Robertson, 2005) Folate plays an important role in the pathogenesis of several disorders, including anemia, cardiovascular disease, and development al abnormalities such as neural tube defect s (NTD s ) (Kim, 2003) and has been linked to an increased risk of developing several types of cancer including colon, pancreas and possibly breast cancers (Kim, 2007; Larsson et al., 2006; Ulrich, 2007) T here is an overwhelming body of evidence in support of the benefits of periconceptional supplementation with folic acid, a synthetic, oxidized form of folate, in the reduction in neural tube defects, and it is recommended that women of child-bearing age consume 400 g of folate daily. As this level of intake was generally no t being achieved in the United States population a program of mandatory, nation -wide fortification of flour and uncooked cereal grains with folic acid was implemented in 1998 in the US and shortly thereafter in Canada. After the fortification program beg an, plasma folate concentration increased by 100% along with a reduction in homocysteine levels, resulting in a 20 50% decrease in the incidence of NTD s (Honein et al., 2001; Jacques et al., 1999; Ray, 2004) Interestingly, it was observed that following the implementation of this program, the re was a temporal association between folic acid fortification and an increase in colorectal cancer rates in both the US and Canada (Mason et al., 2007) Although it is impossible to draw a causal link between these two events, it has been suggested that an excess of folic acid could have promoted the growth of previously undetected preneoplasia (Song et al., 2000a; Song et al., 2000b) Although the success in prevention of NTDs has been dramatic, and folic acid is generally considered to be safe, there are concerns
27 that the levels of folic acid intake may be far higher than initially expected. This is causing some concern as the long term effects of folic acid exposure and the longterm consequences of increased folate levels have yet to be extensively studied in any long term clinical trials. In human studies conducted in metabolic units involving folate depletion of volunteers, global DNA hypomethylation was observed in circulating peripheral blood lymphocytes After repletion of folate levels, DNA methylation levels returned to normal, indicating that the effect on DNA me thylation was transitory and persisted only during the folate deplete period (Jacob et al., 1998) Studies involving folate depletion in rodents indicate similar effects. Folate deficient rats had significantly lower global DNA methylation levels in liver tissue compared to folate replete control rodents (Balaghi et al., 1993) It has also bee n demonstrated that rats fed a diet supplemented with folic acid showed increased levels of global DNA methylation in liver tissue (Choi et al., 1998) These experimental and observational studies of folate status a r g u e strongly for the need to perform a long term clinical trial to assess the effects of folic acid supplementation. Maternal Nutrition and Fetal Health Poor maternal nutrition in the form of protein or protein -calorie restriction can also have a deleterious effect on offspring. In humans, intrauterine growth ret ardation (IUGR) effects approximately 3% to 5% of pregnancies in the United Sates, and this number is increased dramatically with poor maternal nutrition, general health, and environmental factors such as ni cotine, alcohol, and drug abuse Although complex in origin, poor maternal nutrition is the most likely candidate for IUGR, and worldwide, millions of pregnancies are affected by maternal low protein or calorie restriction, making the causal understanding of IUGR of great medical importance. In studies i nvolving an animal model of IUGR utilizing a maternal low protein during pregnancy, the offspring exposed to a low protein diet had smaller birth weights as
28 compared to control diet offspring. In addition to low birth weight, these animals developed increa sed risk for hypertension and type II diabetes in adulthood, and had a larger adult BMI than control offspring in adulthood. Interestingly, these deleterious traits were passed on at least two generations even in the absence of poor nutrition during their pregnancies (Novak et al, personal communication ), suggesting a heritable, epigenetic change is occurring. There are several examples of experiments testing this hypothesis using various animal models ; they are briefly summarize d below These studies have proved that nutritional insults during pregnancy and/or lactation have a major impact on tissue development and function, which leads to an increased risk of adult disease. In a study performed by Rees et al., female rats were fed a normal protein diet containing 18% casein or low protein diet containing 8% casein for two weeks prior to mating. Female rats continued on their respective diet until 21 dpc, when they were euthanized along with the fetuses, and fetal tissues, liver, heart and kidney, were r emoved (Rees et al., 2000) The amount of food consumed during p regnancy did not differ significantly between groups of animals. At 21 dpc, the fetuses from the dams on the 8% casein diet were 13.7% smaller than the control group, and fetal livers from the 8% casein group were ~24% smaller In addition to fetal and tis sue weights, the authors examined the global methylation levels in the fetal tissues using the methyl acceptance assay. The results indicate a greater than 25% increase in methylation in the livers of the fetuses from the 8% casein diet. No significant cha nges were no ted for heart or kidney tissues. The authors summarize by stating that protein restriction in utero causes genome wide changes in DNA methylation, at least in the liver, and these changes have the potential to alter the regulation of important genes in the offspring (Rees et al., 2000)
29 In another set of ex periments, Lillycrop et al. use a similar animal model to show locus specific decreases in DNA methylation in liver tissue (Lillycrop et al., 2005) For this study, timed -pregnant rats were put into one of three diet groups at time of conception. One group was fed an 18% casein diet (C ontrol) another a 9% casein diet (R estricted) and a third group consumed a 9% casein diet that was fortified with 5mg/kg folic acid ( Restricted F ortified) (5 times the folic acid of the other diets). Each group was fed this diet until spontaneous birth at day 21 dpc, at which point the dams were fed a lactating diet that the pups were weaned onto for 28 days. At day 34 the pups were euthanized and the livers were removed. The methylation status and expression levels wer e examined for two hepatic genes, glucocorticoid receptor (GR) and peroxisomal proliferator activated receptor (PPAR ) through the use of methylation -specific PCR The results showed that offspring exposed to a restricted diet in utero had 2 6 % less methylation at the PPAR promoter and 23% less methylation at the GR promoter as compared to offspring in the control diet group. Methylation levels of GR and PPAR promoters were essentially the same for offspring of both the control group and the restr icted fortified diet group. The authors also examined whether the changes in DNA methylation had a corresponding change in expression levels. mRNA levels of the PPAR gene increased by 945% and the expression of GR increased by 300% in the restricted diet group as compared to the control diet group. There was no statistical difference in the expression levels between the control and restricted fortified diet groups (Lillycrop et al., 2005) In a follow up study, it was shown through high resolution so dium bisulfite genomic sequencing that the hypomethylation observed at the promoter of the PPAR gene was occurring at specific CpG nucleotides (Lillycrop et al., 2008) The CpG sites that were affected were within the putative binding site of various transcription
30 factors, and the authors suggest that changes in DNA methylation at these sites may alter transcription factor binding and lead to changes in gene expression (Lillycrop et al., 2008) The focus of this diss ertation will be on analyzing the effects of a low protein diet in utero on DNA methylation levels at both the global level through analysis of repetitive elements, and the locus -specific level. The DNA methylation statuses of imprinted domains, especiall y Igf2 / H19 are analyzed in great detail due to their importance in proper fetal development. Additionally, the effects of folic acid supplementation and withdrawal in a folic acid nave human population on DNA methylation are studied.
31 Figure 1 1. Par ent of origin specific changes in DNA methylation during embryogenesis. The paternal genome (blue) is demethylated first by a putative active mechanism, followed by passive demethylation of the maternal genome (red). Both genomes become remethylated to dif ferent extents in the extraembryonic and embryonic lineages around the time of implantation. Figure adapted from Reik et al. 2001. (Reik et al., 2001)
32 Figure 1 2. Erasure of DNA methylation during gametogenesis During spermatogenesis and oogenesis, parental imprints are established at the ICRs of imprinted domains. DNA methylation imprints are depicted by lollypops in two different imprinted domains. One paternally derived (ICR1) and the other maternally derived (ICR2). Figure adapted from Delaval et al. (Delaval and Feil, 2004)
33 Figure 1 3. Schematic representation of the H19 / Igf2 imprinted domain showing the 3 paternally methylated DMRs and the placenta -specific maternally methylated DMR0. Adapted from Lopez 2003 (Lopes et al., 2003)
34 Figure 1 4. Schematic representation of the role of CTCF in long range transcriptional regulation in imprinted domains. Methylation of the H19 DMR prevents CTCF from binding, allowing down stream enhancers to promote transcription of the paternal allel e of Igf2 CTCF binds to the maternally unmethylated H19 DMR, blocking access of enhancers to Igf2 and allowing enhancers to drive expression of maternal H19 .Figure adapted from Chao et al. (Chao and D'Amore, 2008)
35 Figure 1 5. The role of folate in one -carbon metabolism. Schem atic showing biochemical pathways and the role of folate in one carbon metabolism, particularly the role of folate in DNA methylation, circled in red. Figure adapted from Lamprecht et al. (Lamprecht a nd Lipkin, 2003)
36 CHAPTER 2 E PIGENTETIC EFFECTS OF A LOW PROTEIN DIE T IN UTERO Introduction There is an increasing awareness of the potential for deleterious effects o f maternal malnutrition during pregnancy on the offspring, such as an increased susceptibility to metabolic syndrome and cardiovascular disease (CVD) in humans (Godfrey and Barker, 2001) In animal studies involving rodents, maternal malnutrition re inforces the observations that early environmental conditions can have adverse effects on the offspring throughout their life (Bertram and Hanson, 2001) There is a growing body of evidence, both epidemiological a nd experimental, in which nutrient deprivation in utero has been demonstrated to have adverse longterm effects on the metabolic and physiological states of offspring (Malandro et al., 1996; Osmond et al., 1993; Roseboom et al., 2001b) However, the mechanisms by which altered metabolism and disease susceptibility may arise as a consequence of dietary restriction during gestation remains unclear. One possibility is that epigenetic changes in the placenta and/or fetus in response to maternal malnutrition could lead to stable aberrant regulation of genes important for normal placental and / or fetal development. Recent experimental data from the literature has demonstrated that exposure to a low protein diet (LPD) in utero can induce changes in DNA methylation at both the global and locus -specific scale (Lillycrop et al., 2005; Lillycrop et al., 2007; Rees et al., 2000) These findings are im portant in that they indicate that exposure to nutritional insults in utero can induce epigenetic changes in offspring However, until recently, the focus has been limited to either a global view of genomic DNA methylation or a locus specific analysis of D NA methylation of candidate genes only. In this study, we set out to expand the results of these initial findings to a much broader, genome wide view of potential changes in DNA m ethylation. We have used a modified experimental approach first developed by Weber et
37 al. (Weber et al., 2005) in which the authors utilized an immu nocapture approach using antibodies against methylated DNA followed by microarray analysis. Using this method, the authors compared the DNA methylation levels of normal fibroblasts to a transformed colon cancer cell line on a n 80 kb resolution bacterial ar tificial chromosome ( BAC ) array for all human chromosomes, and a CpG island array with an average resolution of 760 bases. The results of their study showed that they were able to detect large regions of hypomethylation in gene -poor genomic areas in the tr ansformed cells compared to normal fibroblasts using the BAC array, and were able to detect 30 unique CpG island sites that showed hypermethylation in the transformed cells that were hypomethylated in normal fibroblasts on the CpG island array (Weber et al., 2005) We wished to expand this approach to the use of a high resolution Affymetrix Mouse Promoter array. This has been accomplished through the use of antibodies that specifically recognize 5 -methylcytosine to obtain DNA fractions that are highly enriched for methylated DNA sequences from genomic DNA of experimental and control animals. Th e highly enriched DNA was then hybridize d to Affymetrix Promoter Array Gene Chips to analyze changes in methylation on a n array containing most of the known mouse promoter s The Affymetrix Mouse Promo ter Array contains 4.6 million probes allowing the interrogation of 28,000 mouse promoters on a single gene -chip platform. These chips cover approximately 6 k b upstream and 2.5 kb downstream of the transcription start site at each promoter which should include many regulatory elements and DMRs, as well as ~70% of all CpG islands in the mouse genome (Affymetrix). The probes are 25 mers that are tiled at an average 35 base resolution, allowing interrogation of genomic regions at high resolutions. Results fr om the gene chip experiment s were then validated using high resolution sodium bisulfite genomic sequencing.
38 Results Pair -fed Animal Feedings Timed -pregnant C57BL/6J mice were received on day 4 dpc, weight matched and put into either a control (NPD) or restricted protein (LPD) feeding group (Figure 2 1) The mice were weighed every 24 hours and their weight recorded starting from day 5 until day 18 (Figure 2 2 ). The amount of food consumed by the LPD animals was weighed every 24 hours and the corresponding amount of food consumed was given to the pair -fed control animal the following day to ensure equal consumption of calories (Figure 2 1) Although there was no significant statistical difference in the amount of weight gained du ring pregnancy between the NPD and LPD animals, there was a trend for the NPD animals to gain more weight compared to the LPD animals (Figure 2 3 A) However, there was no significant statistical difference in the amount of daily food consumption between the NPD and LPD animals (Figure 2 3 B). The difference in weight gains appears to be related to the average litter size, as the LPD group tended to have smaller litters compared to the N PD group (Figure 2 3 C). As the number of pups would be expec ted to have the greatest impact on weight gain, it is not surprising that the NP D group gained more weight. When weight gain per pup was analyzed there was no statistical difference between the two groups (Figure 2 3 D). These results indicate that the ave rage weight gained, food consumed and numbers of pups did not differ significantly between the experimental and control animal groups. Methylated DNA Immunoprecipitation (MeDIP) In order to analyze DNA methylation using promoter tiling arrays, the DNA wa s first enriched for methylated DNA sequences. This was performed using a modified methylated DNA immunoprecipitation (MeDIP) procedure (Figure 24 ). Genomic DNA was first sheared in length to about 250 700 bases through sonication. These DNA fragments wer e then denatured and
39 single -stranded methylated fragments are immunocaptured using antibodies specific to 5 methylcytosine along with a no antibody control IP reaction The 5 Me thylcytosine and no antibody control IP DNA was then captured using secondary antibodies attached to magnetic beads to purify the IP from unbound DNA. This complex was then washed and eluted. These highly enriched fractions were then labeled and hybridized to Affymetrix promoter arrays (Figure 2 4 ). To validate the specificity of e nrichment of methylated DNA fragments versus no antibody and non -specific antibody control s quantitative RT PCR was performed on methylated and unmethylated genomic sequence fractions To test the efficiency of the MeDIP assay, several controls were perfo rmed. During the immunocapture, a non-specific control antibody (normal mouse IgG) reaction and a no antibody control reaction were always run in parallel to the 5 MeC IP reaction. Three genomic sequences known to be methylated were chosen as enrichment c ontrols : Xist, Hprt and the H19 DMR. X inactive specific transcript ( Xist) is an X -linked gene that is only expressed from the inactive X chromosome in female mammals, and is hyper methylated on the active X chromosome in females and males. Hypoxanthine phosphoribosyl transferase 1 (Hprt ) is an X -linked housekeeping gene that is expressed from the active X chromosome and is hypermethylated on the inactive X chromosome in females T he H19 DMR is a paternally imprinted differentially methylated region with hypermethylation on the paternal allele and hypomethylation on the maternal allele As an additional test of efficiency, we performed the initial control experiments using genomic DNA from both male and female mouse livers. This was performed in order to observe the expected methylation differences between Xist and Hprt in male and female cells. The promoter region of Xist is hypermethylated on the active X chromosome and hyp o methylated on the inactive X. Therefore we would expect
40 males to have approximatel y twice the enrichment of methylated DNA at this region as females. This is due to the fact that the only genomic contribution of Xist comes from the single, active X chromosome in males. Additionally, Hprt is a house keeping X linked gene that is hypermeth ylated only on the inactive X chromosome in female cells so we would expect to see enrichment for methylation only in females. The H19 DMR is differentially methylated and we would expect to see equal enrichment from both male and female cells. Two genomi c sequences known to be unmethylated were chosen as negative controls. Adenine phosphoribosyl transferase (Aprt) and -Actin (ActB) are housekeeping genes and contain CpG islands in their promoter regions that are known to be hypomethylated. A c ontrol genomic sequence (CSa) was chosen that contains no CpG dinucleotides as an additional negative control (Figure 2 5 ). The results from our experiments to determine the efficiency of MeDIP enrichment of methylated DNA confirm that we were, in fact obtaining a highly enriched fraction of methylated fragments as compared to non -methylated controls. The methylation of Xist in male DNA was ~30% of input compared to ~10% of input in females (Figure 2 5 ). This is in agreement with expected results. In males, whic h contain no inactive X chromosome, methylation of Hprt was undetectable, while females showed an enrichment of methylation of approximately 5% compared to input. Enrichment at the H19 DMR was essentially equal between males and females. The unmethylated control sequences Aprt, -Actin and CSa, had undetectable enrichment of methylated sequences in both males and females (Figure 2 5). These results indicate that the MeDIP protocol used was efficiently enriching methylated DNA fragments.
41 MeDIP on Experimental Animals As MeDIP has demonstrated, we now can obtain highly enriched fractions of methylated DNA fragments from genomic DNA. We next wished to combine the detection of DNA methylation with the larg e scale capabilities of a micro array analysis. Fetal l ivers were harvested and pooled (four each) from litters obtained from four independent pair -fed animal studies, genomic DNA extracted, and MeDIP performed. The specificity and efficiency of enrichment for methylated DNA was again determined through the use of quantitative RT -PCR (qRT PCR) using the same controls as before, plus the 5 LTR of I AP, which is known to be hypermethylated. The qRT -PCR results for I AP show similar levels of enrichment between all four NPD and LPD samples (Figure 2 6). Although there are significant differences in enrichment of methylated DNA at Xist between NPD and LPD samples, the samples are pooled fetal liver s from pups that were not sexed, therefore the ratio of male to females is unknown, and no meaningful conclusion can be drawn. However, there appears to be a slight decrease in the enrichment of methylated DNA at the H19 DMR in the LPD samples compared to the NPD samples. The unmethylated control sequences showed negligible enrichment for all samples (Figure 2 6). These results again confirm that the MeDIP protocol produced highly enriched fractions of methylated DNA. Amplification and Labeling of DNA The amount of DNA recovered from a typical MeDIP is extremely low (~40ng/IP). In order to generate quantities of DNA necessary for use on a micro array, the MeDIP DNA must first be amplified in an unbiased manner To achieve this, we performed a whole genome am plification (WGA) using a kit from Sigma. The Whole Genome Amplification kit (WGA2) is designed to generate an amplifiable fragment library from the MeDIP DNA, and amplify the
42 sample ~600 fold without introducing bias. For each sample, 25 ng of MeDIP DNA w as amplified using WGA2, with an average yield after amplification of between 4 6 g. Gel electrophoresis of amplified products indicate no noticeable change in the size distribution of the original MeDIP input DNA. Xist and Aprt remained highly enriched i n randomly chosen WGA amplified samples according to a qRT -PCR analysis (data not shown). To prepare the WGA a mplified DNA samples for hybridization on the Affymetrix promoter arrays, the samples were biotin labeled using a standard random -primed labeling reaction. Briefly, uracil was incorporated into the amplified DNA using a random primed labeling reaction containing a mixture of dNTPs (10 M dATP, 10 M dCTP, 10 M dGTP, 8 M dTTP, 2 M dUTP). After uracil incorporation, DNA was fragmented through the a ctions of u racil DNA glycosylase (UDG) and apurinic/apyrimidinic (AP) endonuclease (APE 1) and end labeled with a biotinylated ATP analog using Affymetrixs Double -Stranded DNA Labeling Mix. To test for the efficiency of uracil incorporation, fragmentatio n and ultimately, biotin incorporation, we performed an Electro Mobility Shift Assay (EMSA). The high affinity binding properties of avidin to biotin were exploited in this assay. Samples of fragmented and labeled DNA were either pre -incubated with avidin, or not, and run on a 4% to 20% gradient acrylamide TBE gel. The binding of biotin to avidin retards gel migration, and efficient labeling can be visualized by a shift in molecular weight of labeled fragments as compared to fragments without a vidin The re sults of the EMSA indicate that all samples were efficiently labeled (Figure 2 7 ). The labeled DNA was then hybridized to Affymetrix Mouse Promoter Arrays in Dr. Henry Bakers laboratory in collaboration with Dr. Baker and Cecilia Lopez. Briefly : Fragmente d and labeled DNA was hybridized to the promoter arrays at 45oC for 16 hours. Arrays were washed and then stained with a fluorescently labeled antibody conjugated to avidin which binds to
43 biotinylated nucleotides in the labeled samples. Fluorescent signal intensity was then recorded using the GeneChip Scanner 3000. A more thorough description of scanning the arrays can be found chapter 4. Analysis of Promoter Array Data The results of the gene chip sca n were analyzed using Parteks Genomic Suite Software package. This software allows sophisticated statistical analysis of multiple gene chips from multiple experimental groups. For our experiment, we had a total of 16 promoter arrays, four NPD and four LPD chips with 5 MeC, along with paired control chips using input DNA from each corresponding sample. The signal intensity for each chip was normalized using Robust Multi -chip Averaging (RMA), and average signal intensities were scaled across all chips. Afte r normalization, the signals from input chips were subtracted from the corresponding antibody gene -chips to reduce noise due to background. The remaining eight data sets, representing four NPD and four LPD arrays with input subtracted were then compared S ignal intensities were compared between each experimental (LPD) array and the corresponding control (NPD) array Differences in signal intensity were analyzed using a 300 base window (approximately 8 consecutive probe sets), using a threshold p-value of 0 .001. Probe sets that met this threshold value in three out of four arrays were scored as a positive differential signal for subsequent analysis. Parteks software identified 176 positive genomic regions that showed changes in signal intensity meeting the above mentioned criteria. The genomic location of these probe sets were then used to annotate genes associated with them. A fter analyzing positively identified probe sets within the UCSC G enome B rowser, it was observed that the overwhelmingly vast majority of these positive regions were directly adjacent to a repetitive element in an intronic region of a gene. The sequence encompassing the positively
44 identified probe sets were analyzed for 30 regi ons with the lowest p -value for an increase and 30 regions with the lowest p-value for a decrease in signal intensity. The genomic sequence of the regions analyzed had very little to no CpG dinucleotides within them, suggesting that the resulting signal wa s due to the IP of a DNA fragment containing a repetitive element along with enough adjacent sequence to bind to the probes on the array. Two probe sets were identified that were in intronic regions of single copy genes that did contain a moderate density of CpGs, Cdca8 and Man1a. These genes will be discussed in further detail later. The Partek software generates a heat map to visualize the difference in signal intensities between the control and experimental samples for individual probe sets, and also shows the location of individual probes within a genome browser. For Mannosidase 1, alpha, ( Man1a) a protein involved in the degradation of terminally misfolded proteins in the endoplasmic reticulum, the heat map shows a strong increase in signal intensity fo r the LPD samples, indicating an enrichment of methylated DNA fragments hybridizing to these probe sets (Figure 2 8). Cell division cycle associated 8 ( Cdca8) is a protein that is critical for the proper segregation of chromosomes during mitosis (Yamanaka et al., 2008) In this case, the heat map indicates that there is increased signal intensity for the NPD samples compared to the LPD samples (Figure 2 9) A third gene that was also identified in the Partek screen, albeit with less stringent parameters, was RIO Kinase 1 ( RIOkI), which is a ubiquitously expressed serine kinase (LaRonde -LeBlanc and Wlodawer, 2005) (heat map not shown). After reviewing all the data from the promoter arrays, we proposed that there are two possible explanations for not finding significant changes in DNA methylation between the NPD and LPD groups. The first explanation is that the degree of any change s in DNA methylation between the NPD and LPD groups was below the limit of the array to detect. The design of the
45 tiling array, with 25 -mer probes tiled at a 35 base resolution inherently has a n elevated background and thus a reduced signal to noise ratio which could have masked small changes in DNA methylation between the NPD and LPD groups Additionally the 5 -MeC antibody precipitates a large fraction of the genome as compared to an IP performed with an antibody against a specific DNA -binding prot ein. This is due to the fact that a large portion of the genome consists of methylated repetitive DNA (Lander et al., 2001; Rollins et al., 2006) which would IP along with single copy genes, and would also have contribute d to a n elevated background signal and a low er signal to noise ratio. These two factors would make it more difficult to detect small changes in DNA methylation. Thi s seems to be a reasonable explanation as similar experiments in the literature have shown locus -specific changes in DNA methylation (Lillycrop et al., 2005; Lillycrop et al., 2007; Waterland et al., 2006) although the changes in methylation observed were modest at best, around 20% change. After the promoter array analysis had been performed, Lillycrop et al. (Lillycrop et al., 2008) demonstrated that the observed changes in methylation at the hepatic PPAR gene promoter in response to a low protein diet in utero were at individual CpG sites thus confirming our notion that these types of changes would be difficult to detect using a tiling array approach. This led us to propose that any changes in DNA methylation between the two feeding groups would be minor, which is consistent with the observation that there is very little phenotypic difference between NPD and LPD offspring at birt h. The second explanation is that there was a technical issue in the implementation of the experiment As the MeDIP and qRT -PCR showed highly enriched fractions of methylated DNA and the EMSA results indicate d that the labeling reaction was successful, it may be that the either the hybridization to the arrays was unsuccessful or the chips themselves were defective in some manner. This explanation seems unlikely as statistical evidence and pair -wise matching of
46 the arrays during the quality control steps indicate that the arrays hybridized well and the collection of signal intensity was successful. Locus -Specific DNA Methylation Analysis In order to test our hypothesis that there may be modest or CpG site -speci fic differences in DNA methylation between the NPD and LPD groups we analyzed the methylation status of two well studied and characterized imprinted domains whose regulation are known to be controlled by DNA methylation. The first domain was the H19 / Igf2 imprinted domain. This region was chosen because Igf2 is a critical growth factor involved in fetal and placental development, and both genes are highly expressed in fetal tissues (Constancia et al., 2002) The imprinted domain is under the regulation of an ICR located 24 kb upstream of H19 whose methylation status is critical for interaction with two DMRs in Igf2 which establish proper imprinted express ion of these genes. The second imprinted domain was the An gelman /P rader Willi syndrome (AS/PWS) imp rinted domain. This is another well characterized imprinted domain whose imprinting control center (ICR) at the promoter of the Snrpn gene controls proper imprinted gene expression of genes as far away as 2 Mb including the Mkrn3 locus Both Snrpn and Mkrn3 expression are regulated by DNA methylation. Additionally, we analyze d the DNA methylation status of Man1a, Cdca8 and ROIkI the three genes identified using Partek Genomic Suite Software from the promoter array analysis In order to determine if a ny changes in DNA methylation were occurring at these genes, h igh resolution sodium bisulfite genomic sequencing was performed in a region overlapping the location of the positive probe sets for Man1a, Cdca8, and ROIkI For Snrpn and Mkrn3 GpG islands wit hin the promoters of these genes were analyzed For the H19 / Igf2 imprinted domain, differentially methylated regions were analyzed for the H19 DMR, and Igf2 DMRs 1 and 2. For each locus, two NPD samples and all four LPD samples were analyzed.
47 DNA Methylation Analysis by Sodium Bisulfite Genomic Sequencing The results for H19 DMR (Figure 2 10 A) indicate a moderate, but statistically significant reduction in DNA methylation in the LPD group as compared to the NPD group (Figure 2 10 B). Methylation dropped from 37% in the NPD group to 2 5 % in the LPD group (Figure 2 10 C) The relative risk (RR) of demethylation is 1.44 times greater in the LPD group compared to the control group (Figure 2 10 D ). However, it should be noted that these dat a are from one bisulfite treatment only, and attempts to repeat these experiments yielded highly variable results. In order to determine changes in the methylation status with any degree of certainty, a second technique, such as pyrosequencing should be e mployed to verify the initial findings. At the Igf2 locus, both DMR1 ( Figure 2 11 A) and DMR2 (Figure 2 12 A) were analyzed for changes in DNA methylation. T here was no detectable change in DNA methylation at DMR1 between NPD (Figure 2 11 B) and LPD (Figu re 2 11 C) and the RR had a value of 1 .0 (Figure 2 11 D ), indicating no increased risk to changes in DNA methylation based on exposure to the low protein diet. T he re was however, a statistically significant increase in DNA methylation at DMR2 in the LPD group as compared to the NPD group Methylation increased from 20% in the NPD samples (Figure 2 12 B) to 2 9 % in the LPD samples ( Figure 2 12 C) The relative risk of increased methylation was 1.32 times greater in the LPD group compared to the NPD group (F igure 2 12 D ). T his is the combin ed data from tw o independent bisulfite treatments which showed a large degree of variability between individual experimental results, though the combined general trend is towards an increase in methylation in the experiment al group. To further verify these data by an independent means pyrosequencing should be performed at DMR2. The results for the AS/PWS imprinted domain showed no statistically significant change in DNA methylation levels at either the Snrpn or Mkrn3 promot er s. At the promoter of Snrpn,
48 w hich also functions as an ICR for the imprinted domain (Figure 2 13 A) DNA methylation increased from 17% in the NDP samples (Figure 2 13 B) to 22% in the LPD samples (Figure 2 13 C). This change in DNA methylation is not si gnificant though, with a p -value of 0.06 and a RR of 0.91 (Figure 2 13 D ). The results of the DNA methylation analysis at the promoter of Mkrn3 (Figure 2 14 A) indicate a slight decrease in methylation, from 17% in the NPD samples (Figure 2 14 B) to 13% in the LPD samples Figure 2 14 C) This change was also not statistically significant with a p value of 0.03 and a RR of 1.05 (Figure 214 D ). The DNA methylation analysis of the regions overlapping the probes identified by the Partek software analysis for Man1a, CdCa8 and ROIkI all indicate that there is no statistically significant changes in methylation. At Man1a DNA methylation decreased from 85% in the NPD samples (Figure 2 15 A), to 83% in the LPD samples (Figure 2 15 B), with a p-value of 0.5 and a RR of 0.91 (Figure 2 15 C). Cdca8 DNA methylation was essentially identical between the NPD samples, 79% (Figure 2 1 6 A), and the LPD samples, 80% (Figure 2 1 6 B), with no statistical significance to the change (Figure 2 1 6 C). Both the NPD samples (Figure 2 1 7 A) and the LPD samples (Figure 2 1 7 B) were hypomethylated at ROIkI with DNA methylation levels less than 1% Figure 2 1 7 C). Man1a, Cdca8, and RIOkI were all identified by Parteks software analysis as regions indicating change in DNA methylation ye t at all three regions, no changes in DNA methylation were observed The most likely explanation for their detection on the promoter array is their close genomic proximity to repetitive elements that may be undergoing differential methylation between the N PD and LPD samples. Due to the fragment size of the sonicated DNA, the labeled DNA will hybridize to the promoter array if the probe sequence is located within a few hundred bases of the repetitive element being immunoprecipitated. For Man1a, t he locations of the
49 positives probes are within the first intron and located ~500 bases from a SINE and the probes for Cdca8 are within the 5 UTR and within 200 bases of a micro-satellite simple repeat. This proximity to a repetitive element is well within t he size range of the sonicated D NA fragments used in the MeDIP. mRNA Expression As the H19 and the Igf2 DMRs physically interact (Murrell et al., 2004; Wall ace and Felsenfeld, 2007) during proper imprinted expression, we next wanted to analyze whether the observed changes in DNA methylation at the H19 DMR and the Igf2 DMR2 produced any effect on expression of these genes. Since CTCF binds the unmethylated maternal allele at H19 DMR, blocking enhancer access to the upstream Igf2 promoter and driving H19 expression could a decrease in DNA methylation of the H19 DMR allow for more CTCF binding, thus decreasing Igf2 and increasing H19 expression ? Another unknown is whether a n inc rease of DNA methylation at the Igf2 DMR2 has any effect on the transcription of Igf2 Igf2 DMR2 acts as a methylation sensitive activator, and an in crease in methylation leads to incr ease d Igf2 expression in vitro (Murrell et al., 2001) although this has not been tested in vivo To investigate these questions expression of Igf2 and H19 were analyzed using reverse transcriptase qRT -PCR. RNA from two NPD and 4 LPD samples were analyzed for Igf2 and H19 expression using G apdh as a reference gene to calculate fold change. The results of the expression analysis show that there is a strong induction of H19 expression in the LPD samples as compared to the NPD samples. Furthermore, the expression of Igf2 was reduced by ten -fold in the LPD samples as compared to the NPD samples (Figure 2 1 8 ). D iscussion: Research involving the effects of maternal malnutrition and the associated increased risk of adult onset diseases has until relatively recently relied on epidemiological data Animal models
50 of maternal malnutrition, including protein restricted diets, have revealed intriguing evidence implicating epigenetic changes as a possible explanation of increased disease risk in adulthood (Lillycrop et al., 2005; Waterland and Garza, 1999; Waterland et al., 2006) However, these studies have been somewhat limited in their analysis due to the technologies available to perform these types of analyses. The analysis of DNA methylation has been either limited to methods that observe changes in global DNA methylation or to the site -specific analysis of a limited number of candidate genes. While site -specific analysis of DNA methylation provides much more informative data, it is a time consum ing and laborious process, which severely limits the number of loci that can be analyzed. In the proceeding work, we have adopted and adapted new technologies that allow us to expand our abilities to analyze DNA methylation in a site -specific manner on a genome -wide level. We have optimized techniques that enable us to obtain highly enriched fractions of methylated DNA fragments from genomic DNA, which can then be applied to subsequent analysis applications, such as various gene -chip platforms that are avai lable, or by utilizing massively parallel sequencing technologies to generate data concerning changes in DNA methylation based on various experimental conditions. In this particular study, we have utilized the Affymetrix Mouse Promoter Array Gene Chip plat form to analyze changes in DNA methylation in fetal tissue from offspring exposed to a low protein diet in utero. In the preceding work, we have demonstrated that there appears to be no wide -spread large scale changes in DNA methylation within promoter re gions in fetal livers of offspring exposed in utero to a maternal low protein diet. However, we have demonstrated that exposure to a low protein diet in utero can produce modest changes in DNA methylation at differentially methylated regions within certain imprinted domains specifically the Igf2 / H19 imprinted domain. At the H19 DMR, there was an observable, and statistically significant decrease in DNA
51 methylation, and at Igf2 DMR2, there was an observable and statistically significant increase in DNA methylation in fetal livers Analysis of the AS/PWS imprinted domain revealed no statistically significant changes in DNA methylation at either the imprinted Snrpn or Mkrn3 loci. Genomic regions identified as being differentially methylated between NPD and LPD samples in the promoter array analysis revealed no observable changes in DNA methylation. The most likely explanation for this is due to the close genomic proximity of these probe sets to repetitive elements. The repetitive sequences are not included on the promoter array, but these elements may contain adjacent sequences that when immunocaptured by MeDIP, can hybridize to the array, giving a false positive. We were also able to demonstrate altered expression of both H19 and Igf2 between the NPD and LPD groups. There was a strong induction in the expression of H19 in the LPD samples, and a 10-fold decrease in expression of Igf2 in the LPD samples compared to the NPD samples. T he concomitant increase in DNA methylation at igf2 DMR2 and decrease in expression of Igf2 is inconsistent with the role of DMR2 as a methylation-sensitive activator. A more detailed examination and possible explanation for this will be discussed in greate r detail in Chapter 5.
52 Figure 2 1. Schematic representation of pair -feeding study. A) Timed pregnant females are weight matched and separated into pair fed groups. B) Amount of food consumed by experiment group animals is recorded and the corresponding amount is fed to control group animals the following day. C) This feeding procedure continues until day 19 of pregnancy, when dams and pups are euthanized and fetal liver tissue is harvested.
53 Figure 2 2 Weight gain of pregnant dams throughout preg nancy. Dams were measured every 24 hours from day 5 until day 18.5. Weight was recorded in grams. All four weight matched pair -fed animals are represented by matching colored lines, with solid points indicating NPD animals, and open points representing LPD animals. 15.00 17.00 19.00 21.00 23.00 25.00 27.00 29.00 31.00 33.00 35.00 E 5 E 6 E 7 E 8 E 9 E 10 E 11 E 12 E 13 E 14 E 15 E 16 E 17 E 18 Weight (g) Embryonic Day (dpc) NPD1 LPD1 NPD2 LPD2 NPD3 LPD3 NPD4 LPD4
54 Figure 2 3 Pair -Feeding Study. A) Average weight gain per dam for each feeding pair. B) Average daily food consumption per dam for each feeding pair. C) Number of pups per litter for each feeding pair. D) Average weight gain per pup for each dam for each feeding pair.
55 Figure 2 4. Outline of methylated DNA immunoprecipitation (MeDIP) and schematic of work flow
56 Figure 2 5 Verification of enrichment of methylated DNA by qRT -PCR. Xist is an X linked gene that is hypermethylated only on the active X chromosome. Hprt is an X linked gene that is hypomethylated on the active X chromosome. H19 DMR is a paternally imprinted gene. Aprt and ActB are housekeeping genes known to be unmethylated at their CpG islands within their promoters. CSa is a control sequence containing no CpG dinucleotides.
57 Figure 2 6. qRT -PCR results of MeDIP enrichment of NPD and LPD samples. Regions analyzed are the same as described in figure 2 4, except for the addition of I AP, which is hypermethylated. Each of the four feeding pairs is represented at each region as four pairs of bars. Black bars represent NPD samples and grey bars represent LPD sa mples. No antibody control and IgG were removed from this graph for the sake of clarity in presentation
58 Figure 2 7 Electro -mobility shift assay. Representative EMSA gel verifying efficient biotin labeling of DNA. Fragmented and labeled DNA from 4 normal protein diet samples were incubated with or without Avidin prior to gel electrophoresis. Efficient labeling can be visualized as a shift in molecular weight in the Avidin treated lanes.
59 Figure 2 8. Heat map of Man1a Partek analysis The gray rectangle at top represents the region view in Partek s genome browser of Man1a. The points in the dot plot represent individual probes (NPD in blue, LPD in red), whose height represents signal intensity. The rows at bottom indicate probes in ind ividual samples, labeled at left. Red indicates increased signal intensity and blue decreased signal intensity.
60 Figure 2 9 Heat map of Cdca8 Partek analysis Figure is labeled as in figure 2 8
61 Figure 2 10. B isulfite G enomic S equencing D ata of H 19 DMR A) Schematic representation of the H19 / Igf2 imprinted domain (Lopes et al., 2003) .Location of H19 DMR indicated by black arrow B) Percent CpG methylation for NPD samples. C) Percent CpG methylatio n of LPD samples. D) Statistical analysis of changes in CpG methylation between NPD and LPD samples.
62 Figure 2 11. Bisulfite G enomic S equencing D ata of Igf2 DMR 1. A) Schematic representation of the H19 / Igf2 imprinted domain (Lopes et al., 2003) .Locati on of Igf2 DMR1 indicated by black arrow. B) Percent CpG methylation for NPD samples. C) Percent CpG methylation of LPD samples. D) Statistical analysis of changes in CpG methylation between NPD and LPD samples.
63 Figure 2 12. Bisulfite G enomic S equencing D ata of Igf2 DMR 2. A) Schematic representation of the H19 / Igf2 imprinted domain (Lopes et al., 2003) .Location of Igf2 DMR2 indicated by black arrow. B) Percent CpG methylation for NPD samples. C) Percent CpG methylation of LPD samples. D) Stati stical analysis of changes in CpG methylation between NPD and LPD samples
64 Figure 2 13. Bisulfite G enomic S equencing D ata of Snr pn. A) Schematic representation of the AS/PWS imprinted domain. Blue ovals represent paternally expressed genes and Pink oval represent maternally expressed genes. Location of Snrpn indicated by black arrow. B) Percent CpG methylation for NPD samples. C) Percent CpG methylation of LPD samples. D) Statistical analysis of changes in CpG methylation between NPD and LPD samples
65 Figure 2 14. Bisulfite Genomic Sequencing Data of Mkrn3 A) Schematic representation of the AS/PWS imprinted domain. Blue ovals represent paternally expressed genes and Pink oval represent maternally expressed genes. Location of Mkrn3 indicated by black a rrow. B) Percent CpG methylation for NPD samples. C) Percent CpG methylation of LPD samples. D) Statistical analysis of changes in CpG methylation between NPD and LPD samples
66 Figure 2 15. Bisulfite Genomic Sequencing Data of Man1a. A ) Percent CpG methylation for NPD samples. B) Percent CpG methylation of LPD samples. C ) Statistical analysis of changes in CpG methylation between NPD and LPD samples
67 Figure 2 16. Bisulfite Genomic Sequencing Data of Cdca8. A ) Percent CpG methylation for NPD samples. B) Percent CpG methylation of LPD samples. C ) Statistical analysis of changes in DNA methylation between NPD and LPD samples
68 Figure 2 17. Bisulfite Genomic Sequencing Data of ROIkI A ) Percent CpG methylation for NPD samples. B) Percent CpG methylation of LPD samples. C ) Statistical analysis of changes in CpG methylation between NPD and LPD samples
69 Figure 2 18. mRNA Expression. Expression levels are fold change normalized to reference gene, Gapdh. H19 expression are displa yed as blue bars, Igf2 expression is displayed as red bars. Error bars are the calculated standard error of the mean of two independent duplicate reactions.
70 CHAPTER 3 ANALYSIS OF DNA METHYLATION IN RESPONSE TO CHRONIC FOLIC ACI D SUPPLEMENTATION AN D WITHDRAWAL IN CHINESE WOMEN Introduction The benefits of periconceptional supplementation with folic acid, a synthetic, oxidized form of folate, in the reduction in neural tube defects (NTDs) have been well studied, and it is recommended that women of c hild bearing age consume 400 g of folate daily (Berry et al., 1999) Although the success in prevention of NTDs has been dramatic, and folic acid is generally considered to be safe, there are concerns that the levels of folic acid intake may be far higher than recommended (Pfeiffer et al., 2005). A s the long term consequen ces of increased folate levels have yet to be extensively studied in long term clinical trials this is becoming an area of increased concern and research Folate plays a major role in one carbon metabolism (Mackenzie, 1984) with its most prominent role s being involved in nucleotide synthesis and the remethylation of hom ocysteine to methionine a precursor of S -Adenosylmethionine (SAM), which serves as the universal one carbon donor involved in methylation of DNA, RNA, lipids and proteins (Lamprecht and Lipkin, 2003; Lucock, 2000) Because of this essential role, perturbations in the levels of folate can influence methylation of DNA which is critical in maintaining the integrity of DNA the stability of the genome, and the proper regulation of gene expression (Dolinoy and J irtle, 2008; Robertson, 2005) In order to examine the effects of increased folate levels on D NA methylation, we analyzed blood samples from a populationbased, randomized trial of folic acid supplementation and withdrawal in Chinese women of child bearing age (Hao et al., 2008) This large populationbased study set out to determine the effects of folic acid supplementation and withdrawal on homocysteine levels in a folic acid deplete population, based on dosages of folic acid ranging
71 from 100 g to 4000 g daily. Blood samples were collected at baseline (0 months) 1, 3, and 6 months, at which time folic acid supplementation ceased. Blood samples were again collected 3 months after cessation of supplementation (Hao et al., 2008) In addition to examining homocysteine levels, another study utilizing these blood samples was performed to determine changes in global DNA methylation levels based on dosage of folic acid and the presence of a common single nucleotide polymorphism (C677 T) within the methylenetetrahydrofolate reductase (MTHFR) gene. The C to T polymorph ism at position 677 codes for an enzyme with reduced activity, and individuals homozygous for the TT genotype have elevated levels of blood serum homocysteine and decreased levels of genomic DNA methylation (Friso et al., 2002) Preliminary data obtained from the laboratory of Dr. Lynn Bailey at the University of Floridas Food Science and Human Nutrition Department, including work by Dr. Eoin Quinlivan and David Maneval, d emonstrate a 13% reducti on in global DNA methylation after 6 months of supplementation, with an additional 23% decrease after 3 months of withdrawal T he global reduction in DNA methylation appeared to be independent of dosage and genotype, but with a TT genotype dependent recovery in DNA methylation status after withdrawa l. To further characterize this decrease in global DNA methylation, a subset of subjects, those receiving 400 g of folic acid once daily, were chosen randomly and analyzed for changes in DNA methylation of tumor suppressor genes, imprinted genes and repetitive elements in response to folic acid supplementation and withdrawal. For this particular study, a total of ten subjects were analyzed, five with the CC genotype, and five with the TT genotype.
72 Results Methylation Specific PCR As global DNA hypomethylation and site -specific hypermethylation are a hallmark of the cancer genome, the initial findings that there was a substantial decrease in global DNA methylation was a concern. In order to investigate the possibility of site -specific DNA hypermeth ylation occurring simultaneously with global DNA hypomethylation, the promoter regions of five tumor suppressor genes (TSG) were analyzed by methylation -specific PCR (MSP) These genes were chosen based on a va riety of factors, including the fact all these genes are known to be silenced by promoter hypermethylation and these genes are all known to be silenced in a high proportion of colon cancer tumor cells (Ohm et al., 2007) This was an important determinant as the link between folate and colon cancer has been well studied (Kim, 2003) These genes were also found to be silenced in the broadest range of tumor types (Ohm et al., 2007) thus making them ideal candidate genes to be screened for possibly increased promoter methylation within our subjects The goal of our initial screen was to observe if folic acid supplementation and subsequent withdrawal had any impact on the DNA methylat ion status of tumor s uppressor genes, as methylation-mediated silencing of these genes could lead to increased risk of cancer development. Methylation -specific PCR allows us to perform a rapid screen of several different tumor suppressor genes in order to find candidate genes for a more detailed and in depth analysis. The results of the MSP analysis indicate that there were two subjects with the CC genotype 3150 and 3309, that did show methylation at the death associated protein kinase 1, (DAPK1) and tiss ue inhibitor of metalloproteinase 3 (TIMP3) promoter s, respectively, at 6 months (Figure 3 1A and 3 1 B. ) This increase in DNA methylation compared to 0 month samples appears to be transient, as it is no longer detectable at 9 months. This raises the
73 inter esting possibility that there may be an increase in site -specific DNA methylation occurring during the folic acid supplementation phase which is not maintained during the washout period. Although DNA methylation-mediated silencing of DAPK and TIMP3 may le d to an increase in the development of cancer (Ohm et al., 2007) the biological significance of hypermethylation of these TSG promoters in circulating lymphocytes is unknown. There was also another subject, 2588 (CC) that showed hypermethylation at the estrogen receptor 1 (ESR1) gene promoter at all three time points analyzed, including baseline (Figure 3 1 C) suggesting that this subject may have had hypermethylation at the ESR1 promoter before enrollment in the clinical study Cyclin dependent kinase inhibitor 2A ( P16 ) (figure 3 1 D) and hypermethylated in cancer 1 ( HIC1 ) (Figure 3 1 E ) showed no detectable changes in methylation at 6 or 9 months as compared to 0 month samples Bisulfite Genomic Sequencing Analysis of Repetitive DNA Methylation In order to determine where in the genome the reduction in global DNA methylation seen by Bailey et al. is occurring, we chose to focus our analyses on repetitive DNA elements as i t is estimated that 35% to 40% of all DNA methylation is within repetitive elements (Bestor, 1998; Kochanek et al., 1993; Schmid, 1998) T he L1 element is the most abundant non -LTR retrotransposable element, which account for approximately 21% of the human genome (Wilson et al., 2007) This made the L1 element a good candidate for analyzing changes in global DNA methylation levels. The L1 repetitive element was analyzed by high re solution sodium bisulfite genomic sequencing ( BGS ) to determine if any changes were occurring in DNA methylation The initial BGS analysis focused on the 5 CC and 5 TT subjects that were analyzed by MSP. A pproximately ten clones from each subject at 0, 6 a nd 9 months were sequenced for a total of approximately 100 clones analyzed at each time point. For each subject, two bisulfite conversions and two PCR reactions were performed in order to reduce any potential bias from a
74 single reaction. It is critical w hen performing bisulfite sequencing to examine a sufficient number of clones in order to obtain an accurate representation of the population of L1 elements being interrogated. Since L1 elements are evolutionarily quite old (Ostertag and Kazazian, 2001; Skowronski et al., 1988; Yang et al., 2004) there are many CpG sites which have become mutated over time by the spontaneous deamination of methylcytosine to thymine T his type of mutation is indistinguishable from an unmethylated CpG site as analyzed by bisulfite genomic sequencing. If the deamination mutation occurs on the reverse strand, the mutated site will read as a TpA site during bisulfite sequencing (Yang et al., 2004) and will be scored as a reverse strand mutation. For this reason, when performing statistical analyses of changes in DNA methylation, CpG sites determined to be unmethylated or mutated sites during bisulfite ge nomic sequencing were grouped together as not methylated, and compared to methylated CpG sites for all samples analyzed. The DNA methylation status of the L1 repetitive elements was analyzed both stratified by genotype and grouped together for 0, 6 and 9 m onths subjects. The bisulfite sequencing results for individual clones for 0, 6 and 9 months respectively, are represented in figures 3 2A, B and C. When stratified by genotype, there was no statistical difference in the proportion of methylated, unmethyla ted and mutated sites a t all three time points analyzed (F igure 3 2 D). Additionally, when subjects are grouped together by time points, there was no statistical difference in the levels of DNA methylation between samples (Figure 3 2 E). This is a surprisi ng result given the evidence of an overall global reduction in DNA methylation, and the fact that the L1 family of repetitive elements a ccounts for a large proportion of methylated DNA in the human genome (Ehrlich, 2002) The relatively large overall decrease in global DNA methylation
75 observed may be influenced by subjects with an extremely large decrease in global DNA methylation, however it needs to be noted that the global DNA methylation of these 10 randomly selected subjects is unknown at this time Bisulfite Genomic Sequencing Analysis of SNRPN Promoter We next wanted to determine if any there were any changes in DNA methylation occurring in a locus specific manner. In order to determine whether folate supplementation and withdrawal induces any site specific changes in DNA methylation levels, we chose to analyze the SNRPN promoter. SNRPN is a maternally imprinted gene w ithin the PWS/AS domain. Parent -of -origin methylation at the SNRPN promoter is critical to maintain i ng proper imprinted expression across the entire 2.5 Mb domain (Glenn et al., 1993; Zeschnigk et al., 1997) The expression of SNRPN itself is also under the regulation of promoter methylation (Zeschnigk et al., 1997) Our initial analysis of this region was performed using high resolution sodium bisulfite genomic sequencing of the same 5 CC and 5 TT subjects analyzed for MSP and L1 DNA methylation Approximately 10 clones for each subject were analyzed at each ti me point and when possible replicate bisulfite tr eatments and PCR were performed. Due to difficulty in obtaining samples and problems with sample quality, not all subjects were able to be analyzed and or verified by a repeat bisulfite conversion and seque ncing. A list of samples analyzed and whether verification for a particular sample was able to be performed are listed in T able 3 1 As SNRPN is a maternally imprinted gene, the BGS results for the 0 month samples gave the expected ratio of approximately 50% hypermethylated clones and 50% hypomethylated clones. Without a known single nucleotide polymorphism (SNP) for these samples, it is impossible to determine the parent of origin for each clone, but we make the assumption that the hypermethylated clones a re from the maternal allele.
76 When subjects are stratified by genotype, the percent methylation is fairly similar at 0 months, with CC subjects averaging 40% and TT subjects averaging 45% DNA methylation (Figure 3 3 A) and when samples are combined by ge notype for analysis, the 0 months samples averaged 43% DNA methylation (Figure 33 B). The methylation data for 0 months samples are summarized in F igure 3 3 C. At 6 months there is a dramatic decrease in DNA methylation in the CC genotype, with methylatio n levels decreasing to 8% and a near complete loss of DNA methylation in the TT genotype, with methylation levels less than 1%. (Figure 3 4 A). When samples are combined by genotype for analysis, the 6 months samples averaged 4% DNA methylation (Figure 3 4 B, C). There is a further reduction in DNA methylation in the 9 month samples with both CC and TT subjects averaging less than 1% methylation (Figure 3 5 A, B and C). The overall combined data for both stratified CC and TT analysis and grouped analysis f or each time point for DNA methylation at SNRPN are summarized in F igure 3 6 A and B. These results are truly striking, in that we have demonstrated a possible connection between folic acid supplementation and withdrawal with a dramatic decrease in locus -s pecific DNA methylation at an imprinted control region. In order to more thoroughly analyze the global DNA methylation status of these samples, and to analyze a greater proportion of L1 sub-families, we analyzed these samples using pyrosequencing technology. This technique allows for the quantitative analysis of a small number of CpG sites by sequencing a large pool of individual DNA molecules simultaneously. The advantage of this technique is that a large number of DNA molecules are analyzed simultaneously reducing the risk of bias compared to sequencing individual clones as in bisulfite genomic sequencing. Another advantage to this technique is that the DNA methylation levels are measured in a truly quantitative manner.
77 Pyrosequencing Analysis of L1 Elements In order to analyze DNA methylation by pyrosequencing w e used a kit designed by Biotage to specifically analyze global methylation using the L1 element as a surrogate marker. The primers for this L1 kit were specifically designed to amplify the l argest portion of L1 subfamilies as possible, thus allowing us to examine the broadest range of L1 elements and increasing our chan c e of observing a change in DNA methylation if it were to occur with the L1 family. Pyrosequencing was performed on 0 month s amples and on 9 months samples, after 3 months of folic acid withdrawal. The results of the pyrosequencing show that in 0 month samples both CC and TT subjects averaged 72% and 71% DNA methylation respectively (Figure 3 7 A). The results are similar for th e 9 month samples, with both CC and TT subjects averaging 72% DNA methylation (Figure 3 7 A). When all CC and TT subjects were combined the pyrosequencing data indicate d no change in DNA methylation between 0 and 9 month samples, with both groups averagin g 72% methylation (Figure 3 7 B). Additionally, the percentage methylation was similar between the results obtained from the bisulfite genomic sequencing analysis of the L1 repetitive element and the pyrosequencing results using the L1 specific sequencing kit, with only a slight difference between bisulfite sequencing and pyrosequencing in the 9 months samples, which is most likely due to selection bias involved in sequencing individual clones using BGS (Figure 3 7 C). Bisulfite Genomic Sequencing Analysis of Subjects with a Large Decrease in DNA Methylation after Washout Additional data from Bailey et al. reveal ed that there are subsets of subjects that show either a dramatic decrease or a moderate increase in the levels of global DNA methylation in respons e to folic acid supplementation and especially to withdrawal of folic acid. Several of these subjects have global DNA methylation levels that are decreased by greater than 80% after
78 the folic acid washout period as compared to baseline DNA methylation leve ls. With this degree of loss of methylation these subjects would likely have hypomethylation at repetitive elements potentially leading to activation of transposable elements and genomic instability (Wilson et al., 2007) In order to determine if these subjects were indeed experiencing hypomethylation at repetitive elements t wo subjects were chosen for further analysis one CC (3126) and one TT (2624) genotype, which showed a grea ter than 85% reduction in DNA methylation levels. Sodium bisulfite genomic sequencing was performed for L1 elements in order to determine if any changes could be detected using this technique, given the large decrease in global DNA methylation of the se sam ples. For each subject, 24 clones were sequenced for 0 and 9 months each. The clones sequenced by BGS and analyzed for subject 262 4 (TT genotype) at 0 and 9 months are shown in F igure 3 8 A and B. The results for 2624 show a statistically significant decr ease in DNA methylation at L1 between 0 and 9 months (Figure 3 8 C). The 9 month sample is only 73% as methylated as the 0 month sample. The clones sequenced by BGS and analyzed for subject 3126 (CC genotype) at 0 and 9 months are shown in figure 3 9 A and B, and show no statistically significant change in DNA methylation between 0 and 9 month samples This is a surprising result given the greater than 85% decrease in global DNA methylation for this subject Although the degree of loss of DNA methylation observed at the L1 element for subject 2624 is a modest decrease it does not account for the amount of decrease observed globally. It may be that other repetitive elements that represent a smaller fraction of the genome are in fact demethylated to a high er extent than the L1 element is. Another possible explanation is that our sample size is simply not large enough to accurately represent the methylation levels of the L1
79 family of repeats or that our L1 primers may preferentially amplify DNA from a subse t of cells that have not undergone demethylation. This may be due to degradation of DNA in cells with a dramatic loss of global DNA methylation. Discussion In the preceding work, we have demonstrated that folic acid supplementation and withdrawal can produce changes in DNA methylation in a loc us -specific manner. Our analysis of the DNA methylation of the promoters of five tumor suppressor genes by methylation-specific PCR indicated that in at least two subjects, in two different genes, there was a n observable increase in DNA methylation after 6 months of folic acid supplementation. Both of these subjects were homozygous for the MTHFR 677C genotype, indicating a potential genotype dependent association with the observed changes in DNA methylation. I t was also demonstrated that the changes in methylation appeared to be transient, as the increase in DNA methylation was no longer detectable in either subject after 3 months of folic acid withdrawal. We were also able to demonstrate that folic acid supple mentation and withdrawal can produce dramatic changes in DNA methylation in an imprinted gene The DNA methylation of the maternally imprinted SNRPN promoter was shown to experience a near complete loss of DNA methylation after 6 months of folic acid suppl ementation, and to have become completely demethylated after 3 months of folic acid withdrawal. These results were obtained through the use of high resolution sodium bisulfite genomic sequencing at the promoter region of SNRPN In addition to the overall d ecrease in DNA methylation observed at this locus, it was also demonstrated that there appears to be a genotype dependent manner to the loss of methylation, with subjects homozygous for the MTHFR 677CT genotype showing a greater degree of loss of methylat ion as compared to those subjects with the homozygous CC genotype.
80 We have also determined that there is no detectable decrease in global DNA methylation levels caused by folic acid supplementation and withdrawal. This was demonstrated by the use of high r esolution sodium bisulfite genomic sequencing of L1 repetitive DNA elements and verified though the use of pyrosequencing technology. In both experimental endeavors, there was neither a genotype dependent loss of DNA methylation nor an overall loss of DNA methylation in the subjects analyzed, although the global DNA methylation levels of these subjects is unknown. However, we did analyze two subjects whose DNA methylation levels were known to be decreased by more than 85% after folic acid withdrawal as comp ared to their baseline levels of DNA methylation. Through the use of bisulfite genomic sequencing, we were able to demonstrate a statistically significant decrease in DNA methylation in one subject homozygous for the TT genotype, with no change observed for the CC subject. Although the decrease in DNA methylation observed was moderate, it does not account for the greater than 85% reduction in global DNA methylation. A more detailed discussion of potential explanations for this will be included in Chapter 5.
81 Table 3 1. List of subjects analyzed by sodium bisulfite sequencing (BGS) at the SNRPN promoter. Subjects are listed by genotype and month of sample. Check mark in BGS Sequenced column indicates that data was obtained from a single bisulfite treatment, PCR and sequencing reaction. Check mark in Verified column indicates that data has been verified by a second, independent bisulfite treatment, PCR and sequencing reaction. Genotype Subject ID Sample month BGS Sequenced Verified CC 3150 0 months CC 3150 6 months CC 3150 9 months CC 2588 0 months CC 2588 6 months CC 2588 9 months CC 3655 0 months CC 3655 6 months CC 3655 9 months CC 3309 0 months CC 3309 6 months CC 3309 9 months CC 2562 0 months CC 2562 6 months CC 2562 9 months TT 2561 0 months TT 2561 6 months TT 2561 9 months TT 2642 0 months TT 2642 6 months TT 2642 9 months TT 2618 0 months TT 2618 6 months TT 2618 9 months TT 3317 0 months TT 3317 6 months TT 3317 9 months TT 3248 0 months TT 3248 6 months TT 3248 9 months
82 Figure 3 1. M ethylation S pecific P CR results for T umor S uppressor G enes at 0, 6 and 9 months. Genotype is listed above each group of number coded subjects. U represents MSP performed with primers specific for unmethylated sequences. M represents MSP performed with primers specific for methylated sequences. For each tumor suppressor gene, methylati on was analyzed for 0, 6 and 9 months as labeled at left of gel. Red triangles indicate subjects that had observed changes in methylation. F) Representative gel of MSP controls. FB: Normal human fibroblasts. HCT 116: Human male colorectal cancer cell line. HCT DHO: HCT 116 cell line with DNMT3a and DNMT3b knockouts. BSF H2O: Bisulfite treated water negative PCR control.
83 Figure 3 2. Bisulfite Genomic Sequencing of L1. A) 0 month samples grouped by genotype. Percentage values calculated as number of methylated, unmethylated or mutated sites divided by total CpG sites. B) 6 month samples, labeled as in A. C) 9 month samples, labeled as in A. D) Stacked bar chart showing ratio of methylated, unmethylated and mu tated sites stratified by genotype at each month. E) Stacked -bar chart showing ratio of methylated, unmethylated and mutated sites in grouped samples for each month.
84 Figure 3 3. Bisulfite Genomic Sequencing of SNRPN at 0 months A) 0 month samples stratified by genotype. B) Combined 0 month samples. C) Stacked -bar chart of ratio between methylated and unmethylated CpG sites by genotype and combined for 0 month samples See Table 3 1 for samples analyzed.
85 Figure 3 4. Bisulfite Genomic Sequencing of SNRPN at 0 months. A) 6 month samples stratified by genotype. B) Combined 6 month samples. C) Stacked -bar chart of ratio between methylated and unmethylated CpG sites by genotype and combined for 6 month samples See Table 3 1 for samples analyzed.
86 Figure 3 5 Bisulfite Genomic Sequencing of SNRPN at 9 months. A) 9 month samples stratified by genotype. B) Combined 9 month samples. C) Stacked-bar chart of ratio between methylated and unmethylated CpG sites by genotype and combined for 9 month sample s See Table 3 1 for samples analyzed.
87 Figure 3 6. Bisulfite Genomic Sequencing of SNRPN A) Stacked -bar chart representing percent CpG methylation stratified by genotype and by month for all subjects combined. B) Stacked bar chart representing percen t CpG methylation by month. See Table 3 1 for samples analyzed.
88 Figure 3 7. Pyrosequencing Analysis of L1 Repetitive Element. A) Percent methylation for 0 and 9 month subjects stratified by genotype. B) Percent methylation for combined subjects at 0 and 9 months. C) Comparison of bisulfite genomic sequencing (BGS) with pyrosequencing results.
89 Figure 3 8. Bisulfite Genomic Sequencing of subject 2624. A) Bisulfite genomic sequencing of 0 months samples. B) Bisulfite genomic sequencing of 9 months s amples. C) Stacked bar chart representing ratio of methylated CpG sites to unmethylated (combination of unmethylated and mutated sites).
90 Figure 3 9. Bisulfite Genomic Sequencing of subject 3126. A) Bisulfite genomic sequencing of 0 months samples. B) Bisulfite genomic sequencing of 9 months samples. C) Stacked bar chart representing ratio of methylated CpG sites to unmethylated (combination of unmethylated and mutated sites).
91 CHAPTER 4 MATERIALS AND METHOD S Animals Timed pregnant C57BL/6J mice were ordered from The Jackson Laboratory and delivered on or before day 5 of pregnancy. Upon arrival, mice were weight -matched and placed into either control or experimental groups. At day 5 of pregnancy, weight -matched contro l and experimental mice were pair fed to ensure equal consumption between weight -matched pairs. The control group was fed a normal protein diet (19.39% protein) and the experimental group was fed a low protein diet (8% protein); both diets were prepared fr om TestDiet (Purina Mills, LLC/PMI). Both diets are isocaloric by weight, with extra calories being supplied by sucrose in the low protein diet. The experimental group animals were given the low protein diet ad libitum and food consumption was weighed dai ly. The control group animals were then given normal protein diet equal in weight to the amount of food consumed for the pair -fed experimental animal during the previous 24 hours, such that food available to control mice is matched accordingly on a daily b asis. Water was supplied ad libitum and both groups of mice were weighed daily to ensure that pair -feeding has no adverse affect on weight, as per IACUC mandates. All animals were housed in barrier cages (Specific Pathogen Free) in a climate controlled roo m (22C/20% humidity) with 12:12hours light dark cycles. Pair feeding was continue d until sacrifice at day 18 .5 of gestation, at which time individual maternal and fetal liver tissue s were collected and immediately frozen in liquid nitrogen and stored at 80oC. Genomic DNA Extraction To prepare fetal liver tissues for DNA extraction, previously frozen livers from each litter were ground together using a mortar and pestle. Tissues were kept frozen throughout this process by using liquid nitrogen in the mort ar and keeping samples on dry ice. One third of ground liver
92 samples were removed and stored at 80oC for later RNA extractions. DNA was purified from the remaining tissues using standard phenol extractions and ethanol precipitation (Strauss, 2001) Briefly, f rozen, ground tissues were transferred to a 50 m l conical tube to which 10 m l s of DNA Extraction Buffer with Proteinase K (50 mM T ris HCl pH 8.5; 25 mM EDTA pH 8.0; 150 mM NaCl; 300 g /mL proteinase K) were added. Samples were rotated overnight at room temperature. Samples were then extracted sequentially using equal volumes of phenol:chloroform (1:1), followed by chloroform. Extrac ts were then treated with RNase (20 U/ml) (Ambion AM2286), followed by an additional round of sequential organic extractions. DNA was precipitated by addition of NaCl to a final concentration of 300 mM and 2.5 volumes of 100% ethanol (EtOH) Samples were allowed to precipitate overnight at -20oC, followed by centrifugation. DNA was resusp ended in 4 m l of TE (100 mM Tris HCl, pH 8.0; 10 mM EDTA, pH 8.0). Methylated DNA Immunoprecipitation (MeDIP) Sonication Sonication was performed using a Sonic Dismembrator Model 100 sonicator to shear genomic DNA to a size of 300bp to 700bp using the fol lowing parameters: 80 g of genomic DNA in 1 m l total volume IP Buffer (0.05% Triton X 100 in PBS). DNA was sonicated 3 times for 20 seconds at power setting 3 (8 watts power output), in a 4 m l plastic culture tube in an ethanol and ice bath with 1 minute in between pulses to cool samples. The samples were then pulsed an additional 2 times for 20 seconds each at power setting 4 (10 watts) with 1 minute of cooling in between pulses. 20 l of sonicated sample were removed and run on an agarose electrophoresis gel apparatus to verify sonication efficiency and DNA fragment size
93 Immunocapture Immunocapture of methylated DNA was performed using an anti 5 -methylcytosine ( 5 CMe) antibody (Epigentek Catalog # A 1014). For each IP, 4 g of sonicated DNA was denatured at 95oC for 10 minutes and snap cooled on ice for 2 minutes, at which point cold IP Buffer was added to 500 L total volume and 10 l 5 CMe was added to each methylated DNA IP. Normal mouse IgG (Millipore 12 371) and a no antibody negative contr ol were performed for each sample. After addition of appropriate antibody, samples were rotated slowly at 4oC for 16 hours. For each IP sample, 40 l magnetic beads (Dynal Dynabeads M 280 Sheep anti -Mouse IgG) were prepared by washing twice in 500 l Bea d Washing Buffer (0.1% BSA, 0.02% NaN3 in PBS) with 2 minutes on magnetic rack between each wash and resuspended in 30 l Bead Washing Buffer. After antibody binding, each sample mixture was transferred to freshly prepared beads and incubated at 4oC for 6 hours while slowly rotating. Upon binding of DNA/antibody complex to Dynabeads, samples were allowed to settle in a magnetic rack for 5 minutes. The supernatant from the no antibody control sample was removed to a 1.5 m l tube and stored in ice for future u se as input DNA sample. The supernatants were removed from the remaining samples and discarded. Washing of the antibody/DNA/bead complexes was performed sequentially with 1 mL w ash buffer at 4oC with slow rotation for 5 minutes each wash using the following wash buffers: Low Salt Wash (0.1% SDS, 1% Triton X 100, 2mM EDTA, 200mM Tris HCl, pH 8.0, 150mM NaCl) High Salt Wash (0.1% SDS 1% Triton X 100, 2mM EDTA 200mM Tris HCl, pH 8.0 500mM NaCl ), LiCl Wash (0.25M LiCl 1% NP 40, 1% sodium desoxycholate 1mM EDTA 100mM Tris HCl, pH 8.0 ) followed by two washes each of T.E. (10mM EDTA pH 8.0, 100 mM Tris HCl pH 8.0).
94 Beads were then resuspended in Elution Buffer (100mM NaCl, 10mM Tris -HCl, 25mM EDTA pH 8.0, 0.5% SDS, 0.1mg/mL proteinase K) and incubated for 30 minutes at 50oC with 550 RPM shaking in a thermomixer After proteinase K treatment, SDS was added to a final concentration of 1.5% for all samples (including Input sample) and heated to 100oC for 5 minutes and cooled briefly on ice, followed by bead sepa ration on a magnetic rack. Supernatant, containing eluted DNA, was removed and purified using the QIAquick PCR Clean up Kit (Qiagen Cat alog # 28106), following the manufacturers provided protocol. Briefly: Five volumes of Binding Buffer PB was added to eluted DNA and applied to the clean up column. After drawing sample through the column using a vacuum manifold, columns were washed with 750 l Wash Buffer PE and then dried by centrifugation. DNA was eluted off the column with 150 l 0.5 X TE pH 8.0 and st ored at 20oC. Quantitative RT PCR of MeDIP Enrichment of methylated DNA in the meDIP samples was determined by RT qPCR using primers for the highly methylated LAP LTR sequences (Weber et al., 2005) the differentially methylated H19 DMR, Hprt and Xist sequences, and for the unmethylated sequences of ActB Aprt, and the contro l sequence (no CpG sites) CSa with both the input and immunoprecipitated samples. The p rotocol was adapted from Weber et al. (Weber et al., 2005) Real Time quantitative PCR was performed using the SYBR GREEN PCR Master Mix from Applied Biosystems (Part No.: 4309155) and a MJ Research DNA E ngine Opticon 2 Continuous Fluorescence Detector For each reaction, 1/60th of the input and immunoprecipitated DNA, and 5 pmol of each primer were used in a 20 l reaction volume. A 5 -point standard curve was generated from a serial dilution of DNA (100% 0.01%) for each primer pair, and percent input was calculated for each loc us using Opticon 2 software. To determine relative enrichment
95 of methylated versus unmethylated sequences after immunoprecipitation, the calculated pe rcent input values for each locus were normalized to the unmethylated sequences ActB or Aprt. Primers used are listed in Table 4 1. Amplification of Immunoprecipitated DNA Immunoprecipitated DNA was amplified using Sigmas Whole Genome Amplification Kit (Sigma Cat. WGA2) The manufa cturers protocol was followed with minor adjustments. The initial fragmentation steps of the protocol were not performed, as the IP DNA was previously sheared by sonication. For each sample, 25 ng of DNA in 10 l volume were used for amplification, to whi ch 2 l Library Preparation Buffer and 1 l Library Stabilization Buffer were added. Samples were vortexed briefly, heated to 95oC for 2 minutes, and snap cooled in ice. 1 l of Library Preparation Enzyme was added to each sample and placed in a thermocycl er pre chilled to 16oC and run using the following protocol: 16oC for 20 minutes; 24oC for 20 minutes; 37oC for 20 minutes; 75oC for 5 minutes; 4oC for at least 2 minutes. To each library sample, 7.5 l of 10X Amplification Master Mix, 47.5 l Nuclease -fre e H2O and 5 l of WGA DNA Polymerase were added. The samples were then incubated in a thermal cycler according to the following protocol: 95oC for 3 minutes ; 14 cycles of 94oC for 15 seconds, 65oC for 5 minutes; 4oC hold. Amplified samples were then purified using the QIAquick PCR Purification Kit (QIAGEN catalog number 28106). Uracil Incorporation, DNA Fragmentation and Labeling To prepare DNA for subsequent fragmentation and labeling procedures, dUTP was incorporated through a random primed la beling reaction using octadeoxyribonucleotides in 10X Labeling Buffer from New England Bioloabs (NEB catalog number N1501L ). 6 g of WGA amplified DNA in 193.5 L total volume was heated at 95oC for 5 minutes and snap cooled on
96 ice. To this, 22.5 L 10x La beling Buffer, 6 l of dNTPs (10 M dATP, 10 M dCTP, 10 M dGTP, 8 M dTTP, 2 M dUTP) and 3 l of Klenow fragment (NEB Catalog number M0212S) was added to each sample and incubated at 37oC for 2 hours. DNA from each reaction was purified using the QIAqui ck PCR Purification Kit. To fragment the UTP containing DNA for labeling, the GeneChip WT Double -Stranded DNA Terminal Labeling Kit (Affymetrix #900812) was used according to the manufacturers protocol. Briefly, 13 g of DNA were fragmented using 15 Units UDG (Uracil DNA Glycosylase), 225 Units APE 1 (apurinic/apyrimidinic (AP) endonuclease) and the supplied buffer. The r eaction was incubated at 37oC for 1 hour; 93oC for 2 minutes; 4oC for at least 2 minutes. DNA was labeled with a nucleotide analog conjugated to biotin using the Double Stranded DNA Labeling Mix from the Affymetrix GeneChip WT Double -Stranded DNA Terminal Labeling Kit. 45 l of fragmented DNA were incubated with 2 l Terminal Deoxynucleotidyl Transferase (TdT) 1 l DNA Label (Biotin labeled nucleotide analog), and 12 l 5 X TdT Buffer at 37oC for 1 hour; 70oC for 10 minutes; 4oC for 2 minutes. Electrophoresis Mobility Shift Assay (EMSA) In order to determine the efficiency of DNA fragmentation and labeling, a gel -shift assay was performed to qualitatively determine uracil incorporation. 1 L aliquots of labeled DNA (with incorporated biotin residues), were heated to 72oC for 2 minutes and then incubated with or without 5 l of 2 mg/m l NuetrAvidin (Pie rce Catalog # PI 31000) at room temperature for 5 minutes. Samples were then prepared for gel loading by addition of 5X Sucrose Loading Buffer (Amresco Catalog # E 274) to a final concentration of 1X. Samples were then loaded on a 4% 20% gradient TBE acrylamide gel (BioRad Catalog # 1611235) in a Criterion Electrophoresis
97 Cell and run at 40 volts for 20 minutes; 100 volts for 30 minutes; 150 volts for 30 minutes with both a 100 bp and a 1 kb Plus DNA Ladder (Invitrogen Catalog # 15628019 and 10787018). After electrophoresis, the gel was stained in 1X SYBR Gold (Invitrogen Catalog # S 11494) in PBS for 30 minutes and visualized on a standard UV trans illuminator. Efficient incorporation of l abeled DNA was visualized as a slower migrating smear of DNA fragments compared to labeled DNA with no NuetrAvidin added. Hybridization and Scanning Mouse Promoter Arrays Hybridization of samples to Affymetrix Mouse Promoter Array 1.0R and scanning of arrays was performed in the laboratory of Dr. Henry Baker by Cecilia Lopez. Hybridization and scanning was performed following the manufactures protocol for the GeneChip Scanner 7G Briefly s amples were prepared using the GeneChip Hybridization, Wash and St ain Kit from Affymetrix (Affymetrix Cat. # 900720). For each array, 7.5 g of fragmented and labeled DNA target was mixed with 50 pM Control Oligonucleotide B2 (Affymetrix Cat. # 900301), 1X Hybridization Mix and 7% dimethyl sulfoxide (DMSO) in a total volume of 200 l Nuclease free Water. Hybridization Cocktail was then heated to 95oC for 5 minutes, cooled to 45oC for 5 minutes and briefly centrifuged to collect sample. The sample was then injected into the array through the septa on the array case and incubated at 45oC in hybridization oven at 60 RPM for 16 hours. Washing and staining of arrays is performed in the Fluidics Station using reagents from the GeneChip Hybridization, Wash and Stain Kit. The Fluidics Station is operated using protocols in the GeneChip Operating Software. After washing and staining, the arrays were scanned using the GeneChip Scanner 3000 7G also controlled by the GeneChip Operating Software.
98 Data Analysis using Parteks Genomic Suite Software Package Raw data from the Affymetrix GeneChip Scanner was analyzed using Parteks Genomic Suite Software Package (Partek Incorporated 12747 Olive Blvd., Suite 205, St. Louis, Missouri 63141, U.S.A). using the Tiling Array Workflow option. Raw data (.CEL files) files were imported and arrays normalized with Robust Multi chip Averaging (RMA) and quantile normalization with output set to log base 2. After normalization, the no antibody input chip data was subtracted from the corresponding antibody chip. After subtraction, th e genomic data was segmented using the Genomic Segmentation option with the following parameters: minimum genetic markers set to 10, p -value threshold set to 0.001, signal -to -noise set to 0.3, and the Region R eport set to report values above or below 0. Th e resulting segmented, overlapped regions were exported automatically into a new spreadsheet within the software. An observation of Variance (ANOVA) was performed between NPD and LPD data sets, and the resulting data was used to generate a list of positive genomic regions using the Create Region List option with the p value threshold set to < = 0.05. A gene was list was then created from the Region List using the Find Overlapping Genes option and the most recent genome build of the mouse genome (mm9 July 2007). High Resolution Sodium Bisulfite Genomic Sequencing High resolution sodium bisulfite genomic sequencing was performed essentially as described by (Clark et al., 1994; Kang et al., 2003) Briefly ; 2.5 g of genomic DNA was sheared by vortexing for 2 minutes at room temperature DNA was then denatured in the presence of 300 mM NaOH for 30 minutes at 37oC. A solution of sodium bisulfite and hydroquinone pH 5.0 was added to the denatured DNA to a final concentr ation of 1.55 M sodium bisulfite and 0.5 mM hydroquinone. The samples were then incubated in a thermocycler under the following conditions: 95oC for 30 seconds; 55oC for 90 minutes. These conditions were
99 cycled for 16 18 hours. After conversion, DNA was de -salted using the Promegas Wizard DNA Clean -Up System following the manufacturers protocol, desulphonated in 300 mM NaOH at 37oC for 15 minutes and precipitated by addition of 0.5 volumes 7.5 M ammonium acetate (NH4AOC ) and 2.5 volumes 100% EtOH. DNA pr ecipitates were then washed in 70% EtOH and resuspended in 50 l dsH2O. PCR reactions were performed under the following conditions: 1/25th to 1/10th volume purified converted DNA; 1x PCR Buffer; 1 M each primer; 200 M dNTPs and 0.125 units of HotStarTaqTM Polymerase (Qiagen Cat. # 203205) in a 25 l reaction volume. Thermocycler conditions were as follows: 95oC for 15 minutes; 45 cycles of 94oC for 45 seconds, primer specific annealing temperature for 30 seconds, 72oC for 90 seconds; 72oC for 10 minutes and hold at 4oC after completion. PCR primers were designed to amplify the upper stand of bisulfite converted target sequences usin g MethylPrimer Express Software. P rimers are listed in Table 4 2 RNA Purification RNA was extracted using the RNeasy Mini Kit (Qiagen Cat. # 74104) following the manufacturers recommended protocol. Briefly; 30 mg of previously frozen ground fetal liver tissue was lysed in 600 l Buffer RLT Plus, and lysate was centrifuged at 13,000 x g for 3 minutes. Genomic DNA was r emoved by transferring supernatant to gDNA Eliminator column and centrifuging at 8,000 x g for 30 seconds. 600 l of 70% EtOH was added to flow -through, mix ed and sample was transferred to RNeasy mini column and centrifuge d for 15 seconds at 8,000 x g. Sa mples were then washed by adding 700 l of Buffer RW1 to RNeasy mini column and centrifuging for 15 seconds at 8,000 x g, followed by two sequential washes of 500 l Buffer RPE centrifuged at 8,000 x g for 15 seconds each. RNeasy mini column was then trans ferred to
100 an RNase -free 1.5 ml tube and RNA eluted with two sequential elutions using 50 l RNase -free water and centrifuging at 8,000 x g for 1 minute each. RNA was stored at 80oC. Reverse -Transcriptase Real Time PCR for Expression First -strand cDNA was generated using SuperScriptTM III Reverse Transcriptase and Random Primers from Invitrogen (Invitrogen Cat. # 18080093, #48190011 ) following the manufacturers r ecommended protocol. Briefly, 30 ng of RNA was mixed with 4.5 ng random primers (hexamers), 1 .5 l 10 mM dNTPs in a total volume of 21 l with diethyl pyrocarbonate (DEPC) treated H2O. Sample was heated to 65oC for 5 minutes, and then snap cooled on ice for 1 minute. To the sample 6 l of 5X Buffer, 3 units of RNasin Plus RNase Inhibitor (Promega Cat. # 9PIN261) and 300 units of SuperScript III were added for a total volume of 30 l. Sample was mixed gently and incubated at 25oC for 5 minutes, 50oC for 1 hour, followed by heat inactivation of the reverse transcriptase by incubation at 70oC for 15 minutes. RNA was then digested by addition of 20 units of RNase Cocktail (Ambion AM2286) incubated at 37oC for 20 minutes. For quantitative Real Time PCR, 1 l of cDNA was used for each reaction using the RT qPCR protocol described earlier. A no reverse transcriptase and a no template control were also added. Expression data was expressed as fold increase or decrease normalized to the reference gene Gapdh.
101 Table 4 1. Methylated DNA immunoprecipitation (MeDIP) primers. All primers listed 5 to 3. P rimer name Primer sequence Annealing temperature mLAP mDIP Upper CTCCATGTGCTCTGCCTTCC 59 o C mLAP mDIP Lower CCCCGTCCCTTTTTTAGGAGA 59 o C mXist mDIP Upper CGCGGATCAGTTAAAGGC GT 59 o C mXist mDIP Lower AACCACGGAAGAACCGCA C 59 o C MeDIP Hprt Upper GCAGCGTTTCTGAGCCATTG 59 o C MeDIP Hprt Lower AAAAGCGGTCTGAGGAGGAA 59 o C mH19 DMR mDIP Upper GCATGGTCCTCAAATTCT GCA 59 o C mH19 DMR mDIP Lower GCATCTGAACGCCCCAAT TA 59 o C mAPRT mDip Upper TGCTGTTCAGGT GCG GTC AC 59 o C mAPRT mDip Lower AGATCCCCGAGGCTGCCT AC 59 o C mActB mDIP Upper AGCCAACTTTACGCCTAG CGT 59 o C mActB mDIP Lower TCTCAAGATGGACCTAATACG GC 59 o C mCSa mDIP Upper TGGTTGGCATTTTATCCCTAG AAC 59 o C mCSa mDIP Lower GCAACATGGCAACTGGAA ACA 59 o C
102 Table 4 2. Sodium bisulfite genomic sequencing primers. All primers listed 5 to 3. Primer name Primer sequence Primer sequence mH19 DMR BSF Upper GAGTATTTAGGAGGTATAAGAATTTTGTAA 51 o C mH19 DMR BSF Lower AAAACTAACATAAACCCCTAACCTC 51 o C mIgf2 DMR1 BSF Upper AGGTGAAGGTTTTGTGGGTAG 51 o C mIgf2 DMR1 BSF Lower CTCTACCTTTCCCCAAAAAAAA 51 o C mIgf2 DMR2 BSF Upper TGATGGAATTGTTTTTGTTTAA 51 o C mIgf2 DMR2 BSF Lower TAACACCTCCTCTCCAAAAC 51 o C mSnrpn BSF Upper TATTTGGGTTGTTAAAAATTTTAA 51 o C msnrpn BSF Lower TCCATTATTCCAAATTAACAAT 51 o C mMkrn3 BSF Upper AAGTAGTAGAYGGTAAAGGTAATGTGTGTA 51 o C mMkrn3 BSF Lower ACCTCAATAAAAACTATAAACTCTTCCAT 51 o C Hu SN R PN BSF Upper GGAATTGGTTTTTTAGAATAAAGGATTTTAGGG 57 o C Hu SN R PN BSF Lower CCCCCTCTCATTACAACAATACTATAAAACCC 57 o C HuLINE1 BSF Upper ATTTTATATTTGGTTTAGAGGG 55 o C HuLINE1 BSF Lower ATCAAAAATCAAAAACCCACTT 55 o C
103 Table 4 3. Reverse Transcriptase qRT -PCR primers. All primers listed 5 to 3. Primer name Primer sequence Annealing temperature mGAPDH RT Forward GCCTTCCGTGTTCCTACCC 60 o C mGAPDH RT Reverse CCTCAGTGTAGCCCAAGATGC 60 o C mIgf2 RT Forward GTGCTGCATCGCTGCTTAC 60 o C mIgf2 RT Reverse ACGTCCCTCTCGGACTTGG 60 o C mH19 RT Forward GCCTCAATAACTGGAGAATGGAA 60 o C mH19 RT Reverse CTCATGGGAATGGTGTGTCTG 60 o C
104 CHAPTER 5 DISCUSION AND FURTURE DIRECTIO NS This dissertation has focused on the epigenetic effects of exposure to a low protein diet in utero in a rodent model, and the epigenetic effects of folic acid supplementation and withdrawal in Chinese women of child bearing age. The overall focus of our la boratory is to gain insight into the mechanism by which environmental influences, particularly diet and nutrition, can affect the epigenome in both developing and adult organisms. DNA methylation has generally been considered a fairly stable epigenetic mar k, associated with longterm repression of transcription. Changes in DNA methylation patterns were thought to normally only occur during embryonic and germ cell development. During embryogenesis, a single celled zygote progresses to a multicellular organis m with over 200 functionally distinct and diverse cell types (Mann and Bartolomei, 2002) Each of these distinct cell typ es has undergone epigenetic changes to form a unique transcriptional memory that is then stably maintained throughout mitosis. (Nafee et al., 2008) During germ cell development, the DNA methylation marks at imprinted genes must be erased and reestablished in a parent -of origin manner (Reik et al., 2001) It has been demonstrated that nutritional insults during these critical developmental periods can affect DNA methylation and influence gene expression levels (Lillycrop et al., 2005; Waterland et al., 2006) In order to investigate this further, we set out to explore changes in DNA methylation in fetal tissues of m ic e exposed to a maternal low protein diet in utero Our experimental approach was to utilize newly developed techniques to enrich for methylated DNA fractions from genomic DNA (Weber et al., 2005) and combine them with the genome -wide scale analysis capabilities of microarrays. The results of our study demonstrate d that there may be subtle changes in DNA methylation occurring at the H19 / Igf2 imprinted domain, however, there were no large wide spread changes in DNA methylation occurring in fetal livers due to maternal malnutrition. Given
105 the fact that there were no major phenotypic differences at birth between pups from a norma l protein diet compared to a low protein diet, it is not surprising that there would be no dramatic changes in DNA methylation occurring. H owever, it may be that there are more subtle changes occurring in DNA methylation which could have a biological effe ct on the offspring The idea of subtle changes in DNA methylation due to maternal malnutrition has been demonstrated at least for the PPAR gene in fetal rats Lillycrop et al. demonstrated an approximately 20% decrease in DNA methylation at the PPAR pr omot er, with an increase in expression 10.5 fold higher in rats exposed to a low protein diet in utero as compared to normal protein diet control animals (Lillycrop et al., 2005; Lillycrop et al., 2008) Modest changes in DNA methylation may be difficult to detect using a high resolution promoter array combined with an antibody that immunoprecipitates such a large fraction of the genome. As discussed earlier, the Affymetrix Mouse Promoter Array inherently has an elevated background and the 5 MeC antibody will immunoprecipitate a large proportion of the genome due to the high content of methylated repetitive DNA in mammalian genomes (Lander et al., 2001; R ollins et al., 2006) These two factors might make it more difficult to detect small changes in DNA methylation. A second method to generate genomic representations for DNA methylation is the HpaII tiny fragment enrichment by ligation-me diated PCR (HELP) assay. This technique takes advantage of differential digestion using methylation -sensitive restriction enzymes, followed by ligation mediated PCR. The resulting library represents both hypomethylated loci and hypermethylated loci, which can then be used as probes to interrogate microarray platforms (Oda and Greally, 2009) For future studies perhaps a better suited microarray platform would be the NimbleGen arrays. These arrays ar e tiled with 50 to 75 -mer probes with 100 base spacing, resulting in a lower background and an increased signal to noise ratio. A second advantage to this platform i s
106 that the arrays support two color scanning, allowing analysis of both input and IP samples on the same chip, reducing variations among sample sets. A second approach would be to utilize massively parallel sequencing to sequence an entire pool of highly e nriched methylated DNA fractions This technique has the advantage that the entire population of fragments that are immunoprecipitated or generated through other means, can be interrogated, without the limitation of array design. In this manner, repetitive DNA elements can be analyzed as well, an advantage over microarrays in which repetitive sequences have been removed from the arrays. The information obtained from massively parallel sequencing has the ability to reveal data concerning locus specific chang es as well as global changes in DNA methylation. In the preceding work, we were also able to demonstrate significant changes in expression of H19 and Igf2 The LPD group showed a strong induction of expression of H19 and a 10 -fold decrease in expression o f Igf2. These expression data are seemingly inconsistent with the known roles of DNA methylation at the Igf2 DMR2. Methylation on the maternal allele of the H19 DMR facilitates the bind ing of CTCF which blocks enhancer access to Igf2 (Bell and Felsenfeld, 2000; Hark et al., 2000) A decrease in methylation could potentially increase CTCF binding, leading to an increase in enhancer availability to H19 and a decrease of enhancer availability at Igf2 consistent with our results demonstrating t hat an increase in DNA methylation at the H19 DMR induces H19 expression and reduces Igf2 expression. H owever, Igf2 DMR 2 is a methylation sensitive activator, and increased methylation has been shown to increase Igf2 expression in vivo (Murrell et al., 2001) Our results indicate an increase of DNA methylation of Igf2 DMR2, which would be expected to increase Igf2 expression. However, it has been reported that Igf2 DMR2 functions at the level of transcription initiation, and it may be that with enhancer
107 competition occurring due to increased CTCF binding at the H19 DMR, there is not enough enhancer activity at Igf2 DMR2 to promote any increase in transcription of Ig f2 In order to gain a better understanding of the mechanisms leading to increased H19 and decreased Igf2 expression we would like to repeat this experiment by crossing C57BL/6J and Cast/Ei mice. This would allow us to examine both changes in methylation and expression in a parent -of -origin allele -specific manner by utilizing single nucleotide polymorphisms between these two related but divergent mouse strains. This approach would allow us to determine if the increase in H19 expression is occurring due to activation of the paternal allele, and whether the decrease in expression of Igf2 in due solely to down regulation of the paternal allele, or if loss of imprinting is occurring at this locus leading to overall dysregulation of expression with transcriptio n occurring from both alleles, but at greatly reduced levels due to lack of enhancer access. There remains the possibility that the observed changes in expression may be unrelated to changes in DNA methylation, and any changes in expression levels may be a respon se to upstream signals due to maternal and/or fetal sensing of amino acid deprivation during the critical period of rapid growth associated with the developing embryo. The DNA methylation status in somatic cells has been generally regarded as being stably maintained throughout the life of the organism. However, DNA methylation can in fact be labile and decreases in methylation have been observed to occur in somatic tissues as part of the aging process (Bjornsson et al., 2008; Richardson, 2003) In the preceding work, we have demonstrated that folic acid supplementation and withdrawal can produce changes in DNA methylation in somatic cells in a locus -specific manner at the promoters of DAPK and TIMP3 tumor suppressor genes Methylation -specific PCR indicated that there was an observable increase in DNA methylation after 6 months of folic acid supplementation. It was also demonstrated that the
108 changes in methylation appeared to be transient, as the increase in DNA methylation was no longer detectable in either subj ect after 3 months of folic acid withdrawal. We were also able to demonstrate that folic acid supplementation and withdrawal can produce dramatic changes in DNA methylation in an imprinted gene. The DNA methylation of the maternally imprinted SNRPN promote r was shown to experience a near complete loss of DNA methylation after 6 months of folic acid supplementation, and to have become completely demethylated after 3 months of folic acid withdrawal. Th ese results demonstrate that the DNA methylation status in somatic cells can undergo changes in response to nutritional supplementation and withdrawal in adult organisms. The biological significance of loss of methylation at either tumor suppressor genes or at the SNRPN locus in circulating leukocytes is unknown, and future studies should include different cell types if possible. A rodent model recapitulating this study would allow the examination of various tissue types that would be otherwise unavailable in a human study. We have also demonstrated that there is no detectable decrease in DNA methylation levels at the L1 repetitive elements caused by folic acid supplementation and withdrawal in ten subjects randomly chosen from the group receiving 400 g of folic acid daily. It should be noted that the global methylation status of these subjects is unknown, so it is possible that these subjects did not experience significant changes in DNA methylation. However, we did analyze two subjects whose DNA methylation levels were known to be decreased by more than 85% after folic acid withdrawal as compared to their baseline levels of DNA methylation. Through the use of bisulfite genomic sequencing, we were able to demonstrate a statistically significant decrease in DNA methylation in one subject homozygous for the TT genotype, with no change observed for the CC subject. Although the re was a modest 27% decrease in DNA methylation observed it does not account for the greater than 85% reduction in global DNA methylation observed in
109 these two subjects Repetitive elements make up about 45% of the human genome (Ro llins et al., 2006) with the LINE elements account ing for 21% of the human genome (Wilson et al., 2007) and it is estimated that 35% to 40% of all DNA methylation occurs at repetitive elements (Bes tor, 1998; Kochanek et al., 1993) A subject that has experienced an 85% reduction in global methylation would conceivably have had to come from loss of methylation at the L1 elements. One possibility is that the L1 primers we are utiliz ing for bisulfite genomic sequencing analysis are preferentially amplifying a subfamily of L1 elements that are protected from demethylation or that DNA from cells that have not undergone demethylation are being preferentially amplified Another possibili ty is that degradation of the DNA sample due to handling and shipping of samples from China, or to incomplete purification during the extraction steps, is limiting the available pool of amplifiable fragments, and that we are not able to amplify a truly re presentative population of DNA molecules from these samples. The levels of global methylation w ere determined using LC/MS/MS ( Laboratory of Dr. Bailey; University of Floridas Food Science and Human Nutrition Department; work by Dr. Eoin Quinlivan and Davi d Maneval) on DNA that has been digested to nucleosides, so degradation of the DNA sample would not interfere with this type of assay The quantity of DNA that is available to us for this study are exceedingly small, and therefore precious, precluding the p ossibility of gel electrophoresis to determine the extent of sample degradation. A second approach would be to utilize pyrosequencing, which conceivably amplifies a greater proportion of L1 subfamilies tha n our BGS primers do. Additionally, in future studi es, the analysis of DNA methylation at other repetitive elements, such as satellite repeats, will be performed. Overall future studies will involve the continuing optimization of new technologies to gain a better understanding of the
110 mechanisms involved in epigenetic alterations occurring due to nutritional or environmental insults.
111 LIST OF REFERENCES Bacolla, A., Pradhan, S., Roberts, R.J., and Wells, R.D. (1999). Recombinant human DNA (cytosine 5) methyltransferase. II. Steady -state kinetics reveal allosteric activation by methylated dna. The Journal of biological chemistry 274, 3301133019. Balaghi, M., Horne, D.W., and Wagner, C. (1993). Hepatic one -carbon metabolism in early folate deficiency in rats. The Biochemi cal journal 291 ( Pt 1) 145149. Barker, D.J. (1995). Fetal origins of coronary heart disease. Bmj 311, 171174. Barker, D.J., Bull, A.R., Osmond, C., and Simmonds, S.J. (1990). Fetal and placental size and risk of hypertension in adult life. Bmj 301, 2 59262. Barker, D.J., and Osmond, C. (1986). Infant mortality, childhood nutrition, and ischaemic heart disease in England and Wales. Lancet 1 10771081. Barker, D.J., Osmond, C., Simmonds, S.J., and Wield, G.A. (1993). The relation of small head circum ference and thinness at birth to death from cardiovascular disease in adult life. Bmj 306, 422426. Barker, D.J., Winter, P.D., Osmond, C., Margetts, B., and Simmonds, S.J. (1989). Weight in infancy and death from ischaemic heart disease. Lancet 2 577580. Bartolomei, M.S., and Tilghman, S.M. (1997). Genomic imprinting in mammals. Annual review of genetics 31, 493525. Bartolomei, M.S., Zemel, S., and Tilghman, S.M. (1991). Parental imprinting of the mouse H19 gene. Nature 351, 153155. Baylin, S.B., E steller, M., Rountree, M.R., Bachman, K.E., Schuebel, K., and Herman, J.G. (2001). Aberrant patterns of DNA methylation, chromatin formation and gene expression in cancer. Human molecular genetics 10, 687692. Bell, A.C., and Felsenfeld, G. (2000). Methyl ation of a CTCF -dependent boundary controls imprinted expression of the Igf2 gene. Nature 405, 482485. Berry, R.J., Li, Z., Erickson, J.D., Li, S., Moore, C.A., Wang, H., Mulinare, J., Zhao, P., Wong, L.Y., Gindler, J., et al. (1999). Prevention of neura l tube defects with folic acid in China. China U.S. Collaborative Project for Neural Tube Defect Prevention. The New England journal of medicine 341, 14851490. Bertram, C.E., and Hanson, M.A. (2001). Animal models and programming of the metabolic syndrome. British medical bulletin 60, 103121. Bestor, T.H. (1998). The host defence function of genomic methylation patterns. Novartis Foundation symposium 214, 187195; discussion 195189, 228132.
112 Bestor, T.H. (2000). The DNA methyltransferases of mam mals. Human molecular genetics 9 23952402. Bjornsson, H.T., Sigurdsson, M.I., Fallin, M.D., Irizarry, R.A., Aspelund, T., Cui, H., Yu, W., Rongione, M.A., Ekstrom, T.J., Harris, T.B., et al. (2008). Intra individual change over time in DNA methylation w ith familial clustering. Jama 299, 28772883. Brandeis, M., Kafri, T., Ariel, M., Chaillet, J.R., McCarrey, J., Razin, A., and Cedar, H. (1993). The ontogeny of allele -specific methylation associated with imprinted genes in the mouse. Embo J 12, 36693677. Brannan, C.I., Dees, E.C., Ingram, R.S., and Tilghman, S.M. (1990). The product of the H19 gene may function as an RNA. Molecular and cellular biology 10, 28 36. Brouha, B., Schustak, J., Badge, R.M., Lutz -Prigge, S., Farley, A.H., Moran, J.V., and Kaz azian, H.H., Jr. (2003). Hot L1s account for the bulk of retrotransposition in the human population. Proceedings of the National Academy of Sciences of the United States of America 100, 52805285. Chao, W., and D'Amore, P.A. (2008). IGF2: epigenetic regul ation and role in development and disease. Cytokine & growth factor reviews 19, 111 120. Chen, R.Z., Pettersson, U., Beard, C., Jackson Grusby, L., and Jaenisch, R. (1998). DNA hypomethylation leads to elevated mutation rates. Nature 395, 89 93. Choi, S. W., Kim, Y.I., Weitzel, J.N., and Mason, J.B. (1998). Folate depletion impairs DNA excision repair in the colon of the rat. Gut 43, 93 99. Clark, S.J., Harrison, J., Paul, C.L., and Frommer, M. (1994). High sensitivity mapping of methylated cytosines. Nuc leic acids research 22, 29902997. Constancia, M., Dean, W., Lopes, S., Moore, T., Kelsey, G., and Reik, W. (2000). Deletion of a silencer element in Igf2 results in loss of imprinting independent of H19. Nature genetics 26, 203206. Constancia, M., Hemb erger, M., Hughes, J., Dean, W., Ferguson Smith, A., Fundele, R., Stewart, F., Kelsey, G., Fowden, A., Sibley, C., et al. (2002). Placental -specific IGF -II is a major modulator of placental and fetal growth. Nature 417, 945948. Cunningham, J.M., Christen sen, E.R., Tester, D.J., Kim, C.Y., Roche, P.C., Burgart, L.J., and Thibodeau, S.N. (1998). Hypermethylation of the hMLH1 promoter in colon cancer with microsatellite instability. Cancer Res 58, 34553460. de Pagter Holthuizen, P., Jansen, M., van Schaik, F.M., van der Kammen, R., Oosterwijk, C., Van den Brande, J.L., and Sussenbach, J.S. (1987). The human insulin-like growth factor II gene contains two development -specific promoters. FEBS letters 214, 259264.
113 Dean, W., Santos, F., and Reik, W. (2003). Epigenetic reprogramming in early mammalian development and following somatic nuclear transfer. Seminars in cell & developmental biology 14, 93 100. DeChiara, T.M., Robertson, E.J., and Efstratiadis, A. (1991). Parental imprinting of the mouse insulin lik e growth factor II gene. Cell 64, 849859. Deininger, P.L., Moran, J.V., Batzer, M.A., and Kazazian, H.H., Jr. (2003). Mobile elements and mammalian genome evolution. Current opinion in genetics & development 13, 651658. Delaval, K., and Feil, R. (2004) Epigenetic regulation of mammalian genomic imprinting. Current opinion in genetics & development 14, 188195. Deng, G., Chen, A., Pong, E., and Kim, Y.S. (2001). Methylation in hMLH1 promoter interferes with its binding to transcription factor CBF and i nhibits gene expression. Oncogene 20, 71207127. Dewannieux, M., Esnault, C., and Heidmann, T. (2003). LINE -mediated retrotransposition of marked Alu sequences. Nature genetics 35, 41 48. Dolinoy, D.C., and Jirtle, R.L. (2008). Environmental epigenomics in human health and disease. Environmental and molecular mutagenesis 49, 4 8. Dolinoy, D.C., Weidman, J.R., and Jirtle, R.L. (2007). Epigenetic gene regulation: linking early developmental environment to adult disease. Reproductive toxicology (Elmsford, N Y 23, 297307. Eden, S., and Cedar, H. (1994). Role of DNA methylation in the regulation of transcription. Current opinion in genetics & development 4 255259. Eden, S., Constancia, M., Hashimshony, T., Dean, W., Goldstein, B., Johnson, A.C., Keshet, I., Reik, W., and Cedar, H. (2001). An upstream repressor element plays a role in Igf2 imprinting. Embo J 20, 35183525. Ehrlich, M. (2002). DNA methylation in cancer: too much, but also too little. Oncogene 21, 54005413. Ehrlich, M., Gama -Sosa, M.A., Hua ng, L.H., Midgett, R.M., Kuo, K.C., McCune, R.A., and Gehrke, C. (1982). Amount and distribution of 5 -methylcytosine in human DNA from different types of tissues of cells. Nucleic acids research 10, 27092721. Eriksson, J.G., Forsen, T., Tuomilehto, J., Wi nter, P.D., Osmond, C., and Barker, D.J. (1999). Catch up growth in childhood and death from coronary heart disease: longitudinal study. Bmj 318, 427431. Esnault, C., Maestre, J., and Heidmann, T. (2000). Human LINE retrotransposons generate processed ps eudogenes. Nature genetics 24, 363367.
114 Feinberg, A.P., Ohlsson, R., and Henikoff, S. (2006). The epigenetic progenitor origin of human cancer. Nature reviews 7 21 33. Feng, Q., Cao, R., Xia, L., Erdjument Bromage, H., Tempst, P., and Zhang, Y. (2002). Identification and functional characterization of the p66/p68 components of the MeCP1 complex. Molecular and cellular biology 22, 536546. Feng, Q., Moran, J.V., Kazazian, H.H., Jr., and Boeke, J.D. (1996). Human L1 retrotransposon encodes a conserved end onuclease required for retrotransposition. Cell 87, 905 916. Feng, Q., and Zhang, Y. (2001). The MeCP1 complex represses transcription through preferential binding, remodeling, and deacetylating methylated nucleosomes. Genes Dev 15, 827832. Ferguson -Smith, A.C., and Surani, M.A. (2001). Imprinting and the epigenetic asymmetry between parental genomes. Science (New York, NY 293, 10861089. Flynn, J., Glickman, J.F., and Reich, N.O. (1996). Murine DNA cytosine C5 methyltransferase: pre -steady and steady-state kinetic analysis with regulatory DNA sequences. Biochemistry 35, 73087315. Forsdahl, A. (1977). Are poor living conditions in childhood and adolescence an important risk factor for arteriosclerotic heart disease? British journal of prev entive & social medicine 31, 91 95. Friso, S., Choi, S.W., Girelli, D., Mason, J.B., Dolnikowski, G.G., Bagley, P.J., Olivieri, O., Jacques, P.F., Rosenberg, I.H., Corrocher, R., et al. (2002). A common mutation in the 5,10 methylenetetrahydrofolate reduc tase gene affects genomic DNA methylation through an interaction with folate status. Proceedings of the National Academy of Sciences of the United States of America 99, 56065611. Gama Sosa, M.A., Wang, R.Y., Kuo, K.C., Gehrke, C.W., and Ehrlich, M. (1983 ). The 5 methylcytosine content of highly repeated sequences in human DNA. Nucleic acids research 11, 30873095. Glenn, C.C., Nicholls, R.D., Robinson, W.P., Saitoh, S., Niikawa, N., Schinzel, A., Horsthemke, B., and Driscoll, D.J. (1993). Modification of 15q11-q13 DNA methylation imprints in unique Angelman and Prader Willi patients. Human molecular genetics 2 13771382. Glickman, J.F., Flynn, J., and Reich, N.O. (1997). Purification and characterization of recombinant baculovirus -expressed mouse DNA met hyltransferase. Biochemical and biophysical research communications 230, 280284. Gluckman, P.D., and Hanson, M.A. (2004). Developmental origins of disease paradigm: a mechanistic and evolutionary perspective. Pediatric research 56, 311317.
115 Godfrey, K.M ., and Barker, D.J. (2001). Fetal programming and adult health. Public health nutrition 4 611 624. Hales, C.N., Barker, D.J., Clark, P.M., Cox, L.J., Fall, C., Osmond, C., and Winter, P.D. (1991). Fetal and infant growth and impaired glucose tolerance at age 64. Bmj 303, 10191022. Hao, L., Yang, Q.H., Li, Z., Bailey, L.B., Zhu, J.H., Hu, D.J., Zhang, B.L., Erickson, J.D., Zhang, L., Gindler, J. et al. (2008). Folate status and homocysteine response to folic acid doses and withdrawal among young Chinese women in a large -scale randomized double -blind trial. The American journal of clinical nutrition 88, 448457. Harikrishnan, K.N., Chow, M.Z., Baker, E.K., Pal, S., Bassal, S., Brasacchio, D., Wang, L., Craig, J.M., Jones, P.L., Sif, S. et al. (2005). Brahma links the SWI/SNF chromatinremodeling complex with MeCP2 dependent transcriptional silencing. Nature genetics 37 254264. Hark, A.T., Schoenherr, C.J., Katz, D.J., Ingram, R.S., Levorse, J.M., and Tilghman, S.M. (2000). CTCF mediates met hylation -sensitive enhancer blocking activity at the H19/Igf2 locus. Nature 405, 486489. Hellman, A., and Chess, A. (2007). Gene body-specific methylation on the active X chromosome. Science (New York, NY 315, 11411143. Hendrich, B., and Bird, A. (1998 ). Identification and characterization of a family of mammalian methyl CpG binding proteins. Molecular and cellular biology 18, 65386547. Henry, I., Bonaiti -Pellie, C., Chehensse, V., Beldjord, C., Schwartz, C., Utermann, G., and Junien, C. (1991). Unipa rental paternal disomy in a genetic cancer -predisposing syndrome. Nature 351, 665667. Hjalgrim, L.L., Westergaard, T., Rostgaard, K., Schmiegelow, K., Melbye, M., Hjalgrim, H., and Engels, E.A. (2003). Birth weight as a risk factor for childhood leukemia : a meta analysis of 18 epidemiologic studies. American journal of epidemiology 158, 724735. Hoffmann, M.J., and Schulz, W.A. (2005). Causes and consequences of DNA hypomethylation in human cancer. Biochemistry and cell biology = Biochimie et biologie ce llulaire 83, 296321. Honein, M.A., Paulozzi, L.J., Mathews, T.J., Erickson, J.D., and Wong, L.Y. (2001). Impact of folic acid fortification of the US food supply on the occurrence of neural tube defects. Jama 285, 29812986. Huxley, R., Neil, A., and Col lins, R. (2002). Unravelling the fetal origins hypothesis: is there really an inverse association between birthweight and subsequent blood pressure? Lancet 360, 659665. Huxley, R.R., and Neil, H.A. (2004). Does maternal nutrition in pregnancy and birth w eight influence levels of CHD risk factors in adult life? The British journal of nutrition 91, 459 468.
116 Jacob, R.A., Gretz, D.M., Taylor, P.C., James, S.J., Pogribny, I.P., Miller, B.J., Henning, S.M., and Swendseid, M.E. (1998). Moderate folate depletion increases plasma homocysteine and decreases lymphocyte DNA methylation in postmenopausal women. The Journal of nutrition 128, 12041212. Jacques, P.F., Selhub, J., Bostom, A.G., Wilson, P.W., and Rosenberg, I.H. (1999). The effect of folic acid fortifica tion on plasma folate and total homocysteine concentrations. The New England journal of medicine 340, 14491454. Jaenisch, R., and Bird, A. (2003). Epigenetic regulation of gene expression: how the genome integrates intrinsic and environmental signals. Na ture genetics 33 Suppl 245 254. Jeanpierre, M. (1994). Human satellites 2 and 3. Annales de genetique 37, 163171. Jones, P.A. (1999). The DNA methylation paradox. Trends Genet 15, 34 37. Jones, P.A., and Baylin, S.B. (2002). The fundamental role of epigenetic events in cancer. Nature reviews 3 415428. Jones, P.A., and Laird, P.W. (1999). Cancer epigenetics comes of age. Nature genetics 21, 163167. Jones, P.A., and Takai, D. (2001). The role of DNA methylation in mammalian epigenetics. Science (N ew York, NY 293, 10681070. Jones, P.L., Veenstra, G.J., Wade, P.A., Vermaak, D., Kass, S.U., Landsberger, N., Strouboulis, J., and Wolffe, A.P. (1998). Methylated DNA and MeCP2 recruit histone deacetylase to repress transcription. Nature genetics 19, 187191. Kang, S.H., Kiefer, C.M., and Yang, T.P. (2003). Role of the promoter in maintaining transcriptionally active chromatin structure and DNA methylation patterns in vivo. Molecular and cellular biology 23, 41504161. Kazazian, H.H., Jr. (2004). Mobile elements: drivers of genome evolution. Science (New York, NY 303, 16261632. Kazazian, H.H., Jr., and Moran, J.V. (1998). The impact of L1 retrotransposons on the human genome. Nature genetics 19, 19 24. Kim, Y.I. (2003). Role of folate in colon cancer development and progression. The Journal of nutrition 133, 3731S 3739S. Kim, Y.I. (2007). Folate and colorectal cancer: an evidence-based critical review. Molecular nutrition & food research 51, 267292. Kochanek, S., Renz, D., and Doerfler, W. (1993). DN A methylation in the Alu sequences of diploid and haploid primary human cells. Embo J 12, 11411151.
117 Lamprecht, S.A., and Lipkin, M. (2003). Chemoprevention of colon cancer by calcium, vitamin D and folate: molecular mechanisms. Nat Rev Cancer 3 601614. Lander, E.S., Linton, L.M., Birren, B., Nusbaum, C., Zody, M.C., Baldwin, J., Devon, K., Dewar, K., Doyle, M., FitzHugh, W., et al. (2001). Initial sequencing and analysis of the human genome. Nature 409, 860921. LaRonde -LeBlanc, N., and Wlodawer, A. ( 2005). A family portrait of the RIO kinases. The Journal of biological chemistry 280, 3729737300. Larsson, S.C., Giovannucci, E., and Wolk, A. (2006). Folate intake, MTHFR polymorphisms, and risk of esophageal, gastric, and pancreatic cancer: a meta anal ysis. Gastroenterology 131, 12711283. Law, C.M., Barker, D.J., Bull, A.R., and Osmond, C. (1991). Maternal and fetal influences on blood pressure. Archives of disease in childhood 66, 12911295. Lee, C., Wevrick, R., Fisher, R.B., Ferguson -Smith, M.A., and Lin, C.C. (1997). Human centromeric DNAs. Human genetics 100, 291304. Leon, D.A., Lithell, H.O., Vagero, D., Koupilova, I., Mohsen, R., Berglund, L., Lithell, U.B., and McKeigue, P.M. (1998). Reduced fetal growth rate and increased risk of death from ischaemic heart disease: cohort study of 15 000 Swedish men and women born 191529. Bmj 317, 241245. Li, E. (2002). Chromatin modification and epigenetic reprogramming in mammalian development. Nature reviews 3 662673. Lillycrop, K.A., Phillips, E.S., Jackson, A.A., Hanson, M.A., and Burdge, G.C. (2005). Dietary protein restriction of pregnant rats induces and folic acid supplementation prevents epigenetic modification of hepatic gene expression in the offspring. The Journal of nutrition 135, 13821386. Lillycrop, K.A., Phillips, E.S., Torrens, C., Hanson, M.A., Jackson, A.A., and Burdge, G.C. (2008). Feeding pregnant rats a proteinrestricted diet persistently alters the methylation of specific cytosines in the hepatic PPAR alpha promoter of the offspring. The British journal of nutrition 100, 278282. Lillycrop, K.A., Slater Jefferies, J.L., Hanson, M.A., Godfrey, K.M., Jackson, A.A., and Burdge, G.C. (2007). Induction of altered epigenetic regulation of the hepatic glucocorticoid receptor in the offspring of rats fed a protein restricted diet during pregnancy suggests that reduced DNA methyltransferase 1 expression is involved in impaired DNA methylation and changes in histone modifications. The British journal of nutrition 97, 10641073. Lopes, S., Lewis, A., Hajkova, P., Dean, W., Oswald, J., Forne, T., Murrell, A., Constancia, M., Bartolomei, M., Walter, J. et al. (2003). Epigenetic modifications in an imprinting cluster are
118 controlled by a hierarchy of DMRs suggesting long range chromatin interactions. Human molecular genetics 12, 295305. Lopuhaa, C.E., Roseboom, T.J., Osmond, C., Barker, D.J., Ravelli, A.C., Bleker, O.P., van der Zee, J.S., and van der Meulen, J.H. (2000). Atopy, lung function, and obstructive airways disease after prenata l exposure to famine. Thorax 55, 555561. Lucock, M. (2000). Folic acid: nutritional biochemistry, molecular biology, and role in disease processes. Molecular genetics and metabolism 71, 121 138. Mackenzie, R. (1984). Biogenesis and Interconversion of Su bstituted tetrahydofolates. Folate and Pterins 1 255306. Malandro, M.S., Beveridge, M.J., Kilberg, M.S., and Novak, D.A. (1996). Effect of low -protein diet induced intrauterine growth retardation on rat placental amino acid transport. The American journal of physiology 271, C295303. Mann, M.R., and Bartolomei, M.S. (2002). Epigenetic reprogramming in the mammalian embryo: struggle of the clones. Genome biology 3 REVIEWS1003. Mason, J.B., Dickstein, A., Jacques, P.F., Haggarty, P., Selhub, J., Dallal, G., and Rosenberg, I.H. (2007). A temporal association between folic acid fortification and an increase in colorectal cancer rates may be illuminating important biological principles: a hypothesis. Cancer Epidemiol Biomarkers Prev 16, 13251329. Mayer, W., Smith, A., Fundele, R., and Haaf, T. (2000). Spatial separation of parental genomes in preimplantation mouse embryos. The Journal of cell biology 148, 629634. McGrath, J., and Solter, D. (1983). Nuclear transplantation in mouse embryos. The Journal of experimental zoology 228, 355362. McGrath, J., and Solter, D. (1984). Completion of mouse embryogenesis requires both the maternal and paternal genomes. Cell 37, 179183. Michels, K.B., Trichopoulos, D., Robins, J.M., Rosner, B.A., Manson, J.E., Hunte r, D.J., Colditz, G.A., Hankinson, S.E., Speizer, F.E., and Willett, W.C. (1996). Birthweight as a risk factor for breast cancer. Lancet 348, 15421546. Michels, K.B., and Xue, F. (2006). Role of birthweight in the etiology of breast cancer. International journal of cancer 119, 20072025. Miranda, T.B., and Jones, P.A. (2007). DNA methylation: the nuts and bolts of repression. Journal of cellular physiology 213, 384390. Moore, T., Constancia, M., Zubair, M., Bailleul, B., Feil, R., Sasaki, H., and Reik, W. (1997). Multiple imprinted sense and antisense transcripts, differential methylation and tandem repeats in
119 a putative imprinting control region upstream of mouse Igf2. Proceedings of the National Academy of Sciences of the United States of America 94, 1 250912514. Morgan, H.D., Santos, F., Green, K., Dean, W., and Reik, W. (2005). Epigenetic reprogramming in mammals. Human molecular genetics 14 Spec No 1, R47 58. Murray, C.J., and Lopez, A.D. (1994). Global and regional cause -of -death patterns in 1990. Bull World Health Organ 72, 447480. Murrell, A., Heeson, S., Bowden, L., Constancia, M., Dean, W., Kelsey, G., and Reik, W. (2001). An intragenic methylated region in the imprinted Igf2 gene augments transcription. EMBO reports 2 11011106. Murrell, A ., Heeson, S., and Reik, W. (2004). Interaction between differentially methylated regions partitions the imprinted genes Igf2 and H19 into parent -specific chromatin loops. Nature genetics 36, 889893. Nafee, T.M., Farrell, W.E., Carroll, W.D., Fryer, A.A., and Ismail, K.M. (2008). Epigenetic control of fetal gene expression. Bjog 115, 158168. Nan, X., Ng, H.H., Johnson, C.A., Laherty, C.D., Turner, B.M., Eisenman, R.N., and Bird, A. (1998). Transcriptional repression by the methyl CpG -binding protein MeC P2 involves a histone deacetylase complex. Nature 393, 386389. Oda, M., and Greally, J.M. (2009). The HELP assay. Methods in molecular biology (Clifton, NJ 507, 77 87. Ohm, J.E., McGarvey, K.M., Yu, X., Cheng, L., Schuebel, K.E., Cope, L., Mohammad, H.P ., Chen, W., Daniel, V.C., Yu, W., et al. (2007). A stem cell like chromatin pattern may predispose tumor suppressor genes to DNA hypermethylation and heritable silencing. Nature genetics 39, 237242. Okano, M., Bell, D.W., Haber, D.A., and Li, E. (1999). DNA methyltransferases Dnmt3a and Dnmt3b are essential for de novo methylation and mammalian development. Cell 99, 247257. Osmond, C., Barker, D.J., Winter, P.D., Fall, C.H., and Simmonds, S.J. (1993). Early growth and death from cardiovascular disease i n women. Bmj 307, 15191524. Ostertag, E.M., and Kazazian, H.H., Jr. (2001). Biology of mammalian L1 retrotransposons. Annual review of genetics 35, 501538. Oswald, J., Engemann, S., Lane, N., Mayer, W., Olek, A., Fundele, R., Dean, W., Reik, W., and Wa lter, J. (2000). Active demethylation of the paternal genome in the mouse zygote. Curr Biol 10, 475478. Pfeiffer, C.M., Caudill, S.P., Gunter, E.W., Osterloh, J., and Sampson, E.J. (2005). Biochemical indicators of B vitamin status in the US population a fter folic acid fortification: results from the
120 National Health and Nutrition Examination Survey 19992000. The American journal of clinical nutrition 82, 442450. Phillips, D.I., Barker, D.J., Hales, C.N., Hirst, S., and Osmond, C. (1994). Thinness at bi rth and insulin resistance in adult life. Diabetologia 37, 150154. Poirier, F., Chan, C.T., Timmons, P.M., Robertson, E.J., Evans, M.J., and Rigby, P.W. (1991). The murine H19 gene is activated during embryonic stem cell differentiation in vitro and at t he time of implantation in the developing embryo. Development (Cambridge, England) 113, 11051114. Pradhan, S., Talbot, D., Sha, M., Benner, J., Hornstra, L., Li, E., Jaenisch, R., and Roberts, R.J. (1997). Baculovirus -mediated expression and characteriza tion of the full length murine DNA methyltransferase. Nucleic acids research 25, 46664673. Ravelli, A.C., van Der Meulen, J.H., Osmond, C., Barker, D.J., and Bleker, O.P. (1999). Obesity at the age of 50 y in men and women exposed to famine prenatally. T he American journal of clinical nutrition 70, 811816. Ray, J.G. (2004). Folic acid food fortification in Canada. Nutrition reviews 62, S35 39. Rees, W.D., Hay, S.M., Brown, D.S., Antipatis, C., and Palmer, R.M. (2000). Maternal protein deficiency causes hypermethylation of DNA in the livers of rat fetuses. The Journal of nutrition 130, 18211826. Reik, W., Dean, W., and Walter, J. (2001). Epigenetic reprogramming in mammalian development. Science (New York, NY 293, 10891093. Reik, W., Santos, F., and D ean, W. (2003). Mammalian epigenomics: reprogramming the genome for development and therapy. Theriogenology 59, 21 32. Reik, W., and Walter, J. (1998). Imprinting mechanisms in mammals. Current opinion in genetics & development 8 154164. Reik, W., and Walter, J. (2001). Genomic imprinting: parental influence on the genome. Nature reviews 2 21 32. Rhodes, K., Rippe, R.A., Umezawa, A., Nehls, M., Brenner, D.A., and Breindl, M. (1994). DNA methylation represses the murine alpha 1(I) collagen promoter by an indirect mechanism. Molecular and cellular biology 14, 59505960. Richardson, B. (2003). Impact of aging on DNA methylation. Ageing research reviews 2 245261. Riggs, A.D., and Jones, P.A. (1983). 5-methylcytosine, gene regulation, and cancer. Advanc es in cancer research 40, 1 30.
121 Robertson, K.D. (2001). DNA methylation, methyltransferases, and cancer. Oncogene 20, 31393155. Robertson, K.D. (2002). DNA methylation and chromatin unraveling the tangled web. Oncogene 21, 53615379. Robertson, K.D. (2005). DNA methylation and human disease. Nature reviews 6 597610. Rollins, R.A., Haghighi, F., Edwards, J.R., Das, R., Zhang, M.Q., Ju, J., and Bestor, T.H. (2006). Large -scale structure of genomic methylation patterns. Genome Res 16, 157163. Rose, G. (1964). Familial Patterns in Ischaemic Heart Disease. British journal of preventive & social medicine 18, 75 80. Roseboom, T.J., van der Meulen, J.H., Osmond, C., Barker, D.J., Ravelli, A.C., and Bleker, O.P. (2000a). Plasma lipid profiles in adults af ter prenatal exposure to the Dutch famine. The American journal of clinical nutrition 72, 11011106. Roseboom, T.J., van der Meulen, J.H., Osmond, C., Barker, D.J., Ravelli, A.C., and Bleker, O.P. (2001a). Adult survival after prenatal exposure to the Dut ch famine 1944-45. Paediatric and perinatal epidemiology 15, 220225. Roseboom, T.J., van der Meulen, J.H., Ravelli, A.C., Osmond, C., Barker, D.J., and Bleker, O.P. (2000b). Plasma fibrinogen and factor VII concentrations in adults after prenatal exposure to famine. British journal of haematology 111, 112117. Roseboom, T.J., van der Meulen, J.H., Ravelli, A.C., Osmond, C., Barker, D.J., and Bleker, O.P. (2001b). Effects of prenatal exposure to the Dutch famine on adult disease in later life: an overvie w. Mol Cell Endocrinol 185, 93 98. Rotwein, P., and Hall, L.J. (1990). Evolution of insulinlike growth factor II: characterization of the mouse IGF II gene and identification of two pseudo -exons. DNA and cell biology 9 725735. Sandovici, I., Kassovsk a Bratinova, S., Loredo Osti, J.C., Leppert, M., Suarez, A., Stewart, R., Bautista, F.D., Schiraldi, M., and Sapienza, C. (2005). Interindividual variability and parent of origin DNA methylation differences at specific human Alu elements. Human molecular g enetics 14, 21352143. Sansom, O.J., Maddison, K., and Clarke, A.R. (2007). Mechanisms of disease: methyl -binding domain proteins as potential therapeutic targets in cancer. Nature clinical practice 4 305315. Santos, F., and Dean, W. (2004). Epigenetic reprogramming during early development in mammals. Reproduction (Cambridge, England) 127, 643651.
122 Schmid, C.W. (1998). Does SINE evolution preclude Alu function? Nucleic acids research 26, 45414550. Skowronski, J., Fanning, T.G., and Singer, M.F. (1988). Unit length line 1 transcripts in human teratocarcinoma cells. Molecular and cellular biology 8 13851397. Song, J., Medline, A., Mason, J.B., Gallinger, S., and Kim, Y.I. (2000a). Effects of dietary folate on intestinal tumorigenesis in the apcMin m ouse. Cancer Res 60, 54345440. Song, J., Sohn, K.J., Medline, A., Ash, C., Gallinger, S., and Kim, Y.I. (2000b). Chemopreventive effects of dietary folate on intestinal polyps in Apc+/ -Msh2 / mice. Cancer Res 60, 31913199. Strauss, W.M. (2001). Prepar ation of genomic DNA from mammalian tissue. Current protocols in molecular biology / edited by Frederick M Ausubel [et al Chapter 2, Unit2 2. Surani, M.A., and Barton, S.C. (1983). Development of gynogenetic eggs in the mouse: implications for parthenoge netic embryos. Science (New York, NY 222, 10341036. Surani, M.A., Barton, S.C., and Norris, M.L. (1984). Development of reconstituted mouse eggs suggests imprinting of the genome during gametogenesis. Nature 308, 548550. Surani, M.A., Barton, S.C., and Norris, M.L. (1986). Nuclear transplantation in the mouse: heritable differences between parental genomes after activation of the embryonic genome. Cell 45, 127136. Suter, C.M., Martin, D.I., and Ward, R.L. (2004). Hypomethylation of L1 retrotransposons in colorectal cancer and adjacent normal tissue. International journal of colorectal disease 19, 95 101. Takai, D., Yagi, Y., Habib, N., Sugimura, T., and Ushijima, T. (2000). Hypomethylation of LINE1 retrotransposon in human hepatocellular carcinomas, b ut not in surrounding liver cirrhosis. Japanese journal of clinical oncology 30, 306309. Tilghman, S.M. (1999). The sins of the fathers and mothers: genomic imprinting in mammalian development. Cell 96, 185193. Tremblay, K.D., Duran, K.L., and Bartolom ei, M.S. (1997). A 5' 2kilobase -pair region of the imprinted mouse H19 gene exhibits exclusive paternal methylation throughout development. Molecular and cellular biology 17, 43224329. Trichopoulos, D. (1990). Hypothesis: does breast cancer originate in utero? Lancet 335, 939940. Tuck Muller, C.M., Narayan, A., Tsien, F., Smeets, D.F., Sawyer, J., Fiala, E.S., Sohn, O.S., and Ehrlich, M. (2000). DNA hypomethylation and unusual chromosome instability in cell lines from ICF syndrome patients. Cytogenetic s and cell genetics 89, 121128.
123 Ulrich, C.M. (2007). Folate and cancer prevention: a closer look at a complex picture. The American journal of clinical nutrition 86, 271273. Verona, R.I., Mann, M.R., and Bartolomei, M.S. (2003). Genomic imprinting: int ricacies of epigenetic regulation in clusters. Annual review of cell and developmental biology 19, 237259. Wallace, J.A., and Felsenfeld, G. (2007). We gather together: insulators and genome organization. Current opinion in genetics & development 17, 400407. Waterland, R.A., and Garza, C. (1999). Potential mechanisms of metabolic imprinting that lead to chronic disease. The American journal of clinical nutrition 69, 179197. Waterland, R.A., and Jirtle, R.L. (2003). Transposable elements: targets for e arly nutritional effects on epigenetic gene regulation. Molecular and cellular biology 23, 52935300. Waterland, R.A., Lin, J.R., Smith, C.A., and Jirtle, R.L. (2006). Post -weaning diet affects genomic imprinting at the insulin-like growth factor 2 (Igf2) locus. Human molecular genetics 15, 705716. Waterland, R.A., and Michels, K.B. (2007). Epigenetic epidemiology of the developmental origins hypothesis. Annual review of nutrition 27, 363388. Weber, M., Davies, J.J., Wittig, D., Oakeley, E.J., Haase, M ., Lam, W.L., and Schubeler, D. (2005). Chromosome -wide and promoter -specific analyses identify sites of differential DNA methylation in normal and transformed human cells. Nature genetics 37, 853862. Wei, W., Gilbert, N., Ooi, S.L., Lawler, J.F., Ostert ag, E.M., Kazazian, H.H., Boeke, J.D., and Moran, J.V. (2001). Human L1 retrotransposition: cis preference versus trans complementation. Molecular and cellular biology 21, 14291439. Weisenberger, D.J., Campan, M., Long, T.I., Kim, M., Woods, C., Fiala, E ., Ehrlich, M., and Laird, P.W. (2005). Analysis of repetitive element DNA methylation by MethyLight. Nucleic acids research 33, 68236836. Wilson, A.S., Power, B.E., and Molloy, P.L. (2007). DNA hypomethylation and human diseases. Biochimica et biophysica acta 1775, 138162. Wu, C., and Morris, J.R. (2001). Genes, genetics, and epigenetics: a correspondence. Science (New York, NY 293, 11031105. Yamanaka, Y., Heike, T., Kumada, T., Shibata, M., Takaoka, Y., Kitano, A., Shiraishi, K., Kato, T., Nagato, M., Okawa, K. et al. (2008). Loss of Borealin/DasraB leads to defective cell proliferation, p53 accumulation and early embryonic lethality. Mechanisms of development 125, 441450.
124 Yang, A.S., Estecio, M.R., Doshi, K., Kondo, Y., Tajara, E.H., and Issa, J .P. (2004). A simple method for estimating global DNA methylation using bisulfite PCR of repetitive DNA elements. Nucleic acids research 32, e38. Yoder, J.A., Walsh, C.P., and Bestor, T.H. (1997). Cytosine methylation and the ecology of intragenomic paras ites. Trends Genet 13, 335340. Yokochi, T., and Robertson, K.D. (2002). Preferential methylation of unmethylated DNA by Mammalian de novo DNA methyltransferase Dnmt3a. The Journal of biological chemistry 277, 1173511745. Zeschnigk, M., Schmitz, B., Dit trich, B., Buiting, K., Horsthemke, B., and Doerfler, W. (1997). Imprinted segments in the human genome: different DNA methylation patterns in the Prader Willi/Angelman syndrome region as determined by the genomic sequencing method. Human molecular genetics 6 387 395.
125 BIOGRAPHICAL SKETCH Jason Orr Brant was born in Jupiter, Florida, to Garry and Christine Brant. He attended Florida Culinary Institute and had a short lived but successful career as a chef in California and New York. Jason then decided to expand his love of learning, and after graduating magn a cum laude with a degree in environmental s cience and forest b iology from SUNY -ESF, he entered graduate school at the University of Florida. He studied epigenetic gene regulation in the Labo ratory of Dr. Thomas P. Yang, where he will continue his education as a postdoctoral fellow. Jason ultimately wants to continue his research career as an officer in the United States Army.