Drug Discovery in
Cancer Epigenetics


Translational Epigenetics Series
Trygve O. Tollefsbol, Series Editor
Transgenerational Epigenetics
Edited by Trygve O. Tollefsbol, 2014
Personalized Epigenetics
Edited by Trygve O. Tollefsbol, 2015
Epigenetic Technological Applications
Edited by Y. George Zheng, 2015
Epigenetic Cancer Therapy
Edited by Steven G. Gray, 2015
DNA Methylation and Complex Human Disease
By Michel Neidhart, 2015
Epigenomics in Health and Disease
Edited by Mario F. Fraga and Agustin F. Fern´andez, 2016
DNA Biomarkers and Diagnostics
Edited by Jos´e Luis Garcı´a-Gim´enez, 2016
Drug Discovery in Cancer Epigenetics
Edited by Gerda Egger and Paola Arimondo, 2016


Drug Discovery in
Cancer Epigenetics

Edited by

Gerda Egger
and

Paola Arimondo

AMSTERDAM • BOSTON • HEIDELBERG • LONDON
NEW YORK • OXFORD • PARIS • SAN DIEGO
SAN FRANCISCO • SINGAPORE • SYDNEY • TOKYO
Academic Press is an imprint of Elsevier


Academic Press is an imprint of Elsevier
The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB
225 Wyman Street, Waltham MA 02451
Copyright r 2016 Elsevier Inc. All rights reserved
No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical,
including photocopying, recording, or any information storage and retrieval system, without permission in writing from
the publisher. Details on how to seek permission, further information about the Publisher’s permissions policies and our
arrangements with organizations such as the Copyright Clearance Center and the Copyright Licensing Agency, can be
found at our website: www.elsevier.com/permissions.
This book and the individual contributions contained in it are protected under copyright by the Publisher (other than as may
be noted herein).
Notices
Knowledge and best practice in this field are constantly changing. As new research and experience broaden our
understanding, changes in research methods, professional practices, or medical treatment may become necessary.
Practitioners and researchers may always rely on their own experience and knowledge in evaluating and using any
information, methods, compounds, or experiments described herein. In using such information or methods they should be
mindful of their own safety and the safety of others, including parties for whom they have a professional responsibility.
To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume any liability for any
injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or
operation of any methods, products, instructions, or ideas contained in the material herein.
Library of Congress Cataloging-in-Publication Data
A catalog record for this book is available from the Library of Congress

British Library Cataloguing-in-Publication Data
A catalogue record for this book is available from the British Library
ISBN: 978-0-12-802208-5
For information on all Academic Press publications
visit our website at
Printed and bound in the United States of America

Publisher: Mica Haley
Acquisition Editor: Catherine Van Der Laan
Editorial Project Manager: Lisa Eppich
Production Project Manager: Melissa Read
Designer: Mark Rogers


List of Contributors
Paola Arimondo
Unite´ de Service et de Recherche, CRDPF, Toulouse, France
Dina Arvanitis
INSERM, UMR1048, Institute of Cardiovascular and Metabolic Diseases, University-Paul
Sabatier, Toulouse, France
´
Frederic
Ausseil
Unite´ de Service et de Recherche, CRDPF, Toulouse, France
Mina Bekheet
Laboratory of Cancer Biology, Department of Oncology, University of Oxford, Oxford, UK
Christopher G. Bell
Department of Twin Research & Genetic Epidemiology, St Thomas’ Hospital, King’s College
London, London, UK; MRC Lifecourse Epidemiology Unit, University of Southampton,
Southampton, UK; Academic Unit of Human Development and Health, University of

Southampton, Southampton, UK; Epigenomic Medicine, Centre for Biological Sciences, Faculty
of Environmental and Natural Sciences, University of Southampton, Southampton, UK
Magdalena Benetkiewicz
´
Groupe Cooperateur
Multidisciplinaire en Oncologie, Paris, France
Kate H. Brettingham-Moore
School of Medicine, University of Tasmania, Hobart, TAS, Australia
Peter J. Brown
Structural Genomics Consortium, University of Toronto, Toronto, ON, Canada
Corey Carter
John P. Murtha Cancer Center, Walter Reed National Military Medical Center/National Cancer
Institute, Bethesda, MD, USA
Christophe Cisarovsky
Research Laboratory in Oncology, University of Lausanne, Centre Hospitalier Universitaire
Vaudois, Lausanne, Switzerland
Pierre Cordelier
INSERM, UMR1037, Cancer Research Center of Toulouse, University of Toulouse-Paul
Sabatier, Toulouse, France
Armand de Gramont
New Drug Evaluation Laboratory, Centre of Experimental Therapeutics and Medical Oncology,
Department of Oncology, University of Lausanne, Centre Hospitalier Universitaire Vaudois,
Lausanne, Switzerland

xv


xvi

List of Contributors


Frank J. Dekker
Pharmaceutical Gene Modulation, Groningen Research Institute of Pharmacy, University of
Groningen, Groningen, The Netherlands
Yannick Delpu
Skirball Institute of Biomolecular Medicine, New York University Langone Medical Center,
New York, NY, USA
Helmut Dolznig
Institute of Medical Genetics, Medical University of Vienna, Vienna, Austria
Marle`ne Dufresne
INSERM, UMR1037, Cancer Research Center of Toulouse, University of Toulouse-Paul
Sabatier, Toulouse, France
Gerda Egger
Clinical Institute of Pathology, Medical University of Vienna, Vienna, Austria
Chantal Etievant
Unite´ de Service et de Recherche, CRDPF, Toulouse, France
Sandrine Faivre
Medical Oncology, Department of Oncology, University of Lausanne, Centre Hospitalier
Universitaire Vaudois, Lausanne, Switzerland
Gary R. Fanger
EpicentRx, Mountain View, CA, USA
Panagis Filippakopoulos
Structural Genomics Consortium, Nuffield Department of Clinical Medicine, University of
Oxford, Oxford, UK; Ludwig Institute for Cancer Research, Nuffield Department of Clinical
Medicine, University of Oxford, Oxford, UK
Alexandre Gagnon
´
´
´ Montreal,
´ QC, Canada

Departement
de Chimie, Universite´ du Quebec
a` Montreal,
Marion Gayral
INSERM, UMR1037, Cancer Research Center of Toulouse, University of Toulouse-Paul
Sabatier, Toulouse, France
Melanie R. Hassler
Clinical Institute of Pathology, Medical University of Vienna, Vienna, Austria
Markus Hengstschla¨ger
Institute of Medical Genetics, Medical University of Vienna, Vienna, Austria
Quanah J. Hudson
CeMM Research Center for Molecular Medicine of the Austrian Academy of Sciences, Vienna,
Austria


List of Contributors

xvii

Jean-Pierre J. Issa
Fels Institute for Cancer Research and Molecular Biology, Temple University School of
Medicine, Philadelphia, PA, USA
Manfred Jung
Institute of Pharmaceutical Sciences, Albert-Ludwigs-Universita¨t Freiburg, Freiburg, Germany
Stefan Knapp
Structural Genomics Consortium, Nuffield Department of Clinical Medicine, University of
Oxford, Oxford, UK; Institute for Pharmaceutical Chemistry, Johann Wolfgang Goethe-University,
Frankfurt am Main, Germany
Nina Kramer
Institute of Medical Genetics, Medical University of Vienna, Vienna, Austria

Stefan Kubicek
CeMM Research Center for Molecular Medicine of the Austrian Academy of Sciences,
Vienna, Austria; Christian Doppler Laboratory for Chemical Epigenetics and Antiinfectives,
CeMM Research Center for Molecular Medicine of the Austrian Academy of Sciences,
Vienna, Austria
Nicholas B. La Thangue
Laboratory of Cancer Biology, Department of Oncology, University of Oxford, Oxford, UK
Dorian Larrieu
INSERM, UMR1037, Cancer Research Center of Toulouse, University of Toulouse-Paul
Sabatier, Toulouse, France
Maxime Leroy
´
´
´ Montreal,
´ QC, Canada
Departement
de Chimie, Universite´ du Quebec
a` Montreal,
Niek G.J. Leus
Pharmaceutical Gene Modulation, Groningen Research Institute of Pharmacy, University of
Groningen, Groningen, The Netherlands
Marco P. Licciardello
CeMM Research Center for Molecular Medicine of the Austrian Academy of Sciences,
Vienna, Austria
Marie Lopez
Unite´ de Service et de Recherche, CRDPF, Toulouse, France
Michelle Lybeck
EpicentRx, Mountain View, CA, USA
Wolfgang J. Miller
Laboratories of Genome Dynamics, Department of Cell and Developmental Biology, Center of

Anatomy and Cell Biology, Medical University of Vienna, Vienna, Austria


xviii

List of Contributors

Heidi Olzscha
Laboratory of Cancer Biology, Department of Oncology, University of Oxford, Oxford, UK
Arnold L. Oronsky
InterWest Partners, Menlo Park, CA, USA
Bryan T. Oronsky
EpicentRx, Mountain View, CA, USA
Neil C. Oronsky
CFLS, San Jose, CA, USA
Thomas Prebet
Departement d’Hematologie et Unite d’Evaluation Therapeutique en Oncologie/Hematologie,
Institut Paoli Calmettes, Marseille, France; Hematology Department, Smilow Cancer Center at
Yale University, New Haven, CT, USA
Eric Raymond
Medical Oncology, Department of Oncology, University of Lausanne, Centre Hospitalier
Universitaire Vaudois, Lausanne, Switzerland
Noe¨l J.-M. Raynal
´
´ and Sainte-Justine University Hospital
Departement
de Pharmacologie, Universite´ de Montreal
´ QC, Canada
Research Center, Montreal,
Elisa Redl

Clinical Institute of Pathology, Medical University of Vienna, Vienna, Austria
Tony R. Reid
Moores Cancer Institute, University of California, San Diego, CA, USA
Dina Robaa
Institut fu¨r Pharmazie, Martin-Luther-Universita¨t Halle-Wittenberg, Halle, Germany
Martin Scherzer
Institute of Medical Genetics, Medical University of Vienna, Vienna, Austria
Jan J. Scicinski
EpicentRx, Mountain View, CA, USA
Semira Sheikh
Laboratory of Cancer Biology, Department of Oncology, University of Oxford, Oxford, UK
Wolfgang Sippl
Institut fu¨r Pharmazie, Martin-Luther-Universita¨t Halle-Wittenberg, Halle, Germany
Mira Stadler
Institute of Medical Genetics, Medical University of Vienna, Vienna, Austria


List of Contributors

xix

Phillippa C. Taberlay
Chromatin Dynamic Group, Genomics and Epigenetics Division, Garvan Institute of Medical
Research, Darlinghurst, NSW, Australia; Epigenetics Program, Genomics and Epigenetics
Division, Garvan Institute of Medical Research, Darlinghurst, NSW, Australia; St. Vincent’s
Clinical School, Faculty of Medicine, University of New South Wales, Darlinghurst, NSW,
Australia
Tirza Timmerman,
Pharmaceutical Gene Modulation, Groningen Research Institute of Pharmacy, University of
Groningen, Groningen, The Netherlands

´ ˆ me Torrisani
Jero
INSERM, UMR1037, Cancer Research Center of Toulouse, University of Toulouse-Paul
Sabatier, Toulouse, France
Christine Unger
Institute of Medical Genetics, Medical University of Vienna, Vienna, Austria
Thea van den Bosch
Pharmaceutical Gene Modulation, Groningen Research Institute of Pharmacy, University of
Groningen, Groningen, The Netherlands
Norbert Vey
Departement d’Hematologie et Unite d’Evaluation Therapeutique en Oncologie/Hematologie,
´ Marseille, France
Institut Paoli Calmettes, Marseille, France; Aix-Marseille Universite,
Tobias Wagner
Institute of Pharmaceutical Sciences, Albert-Ludwigs-Universita¨t Freiburg, Freiburg, Germany
Stefanie Walter
Institute of Medical Genetics, Medical University of Vienna, Vienna, Austria
Angelika Walzl
Institute of Medical Genetics, Medical University of Vienna, Vienna, Austria
Daniel J. Weisenberger
USC/Norris Comprehensive Cancer Center, University of Southern California, Los Angeles,
CA, USA


Preface
The causal involvement of epigenetic pathways in tumor development and progression has been
widely accepted and recent findings have nourished our knowledge of tumor biology and opened
potential clinical applications dependent on epigenetic aberrations. Deregulation and mutation of
epigenetic enzymes as well as global and local changes of epigenetic chromatin modifications are
implicated in a variety of malignancies and provide novel therapeutic and diagnostic targets for

oncology. There is great interest in developing novel epigenetic drugs targeting chromatin
modifiers as well as chromatin reader proteins and some successful recent studies have confirmed
the validity of these drugs for cancer therapy.
We were fortunate to gather many experts and renowned authors to contribute to this book and
are very thankful for their efforts and valuable contributions. This book intends to provide (i) an
introduction into cancer epigenetics and to give a comprehensive overview on, (ii) methods and
tools for epigenetic drug development, (iii) classes of epigenetic drugs, (iv) development of diagnostic tools, and (v) clinical implications of epigenetic therapy. Emerging concepts, such as “episensitization,” are also presented alongside with indications beyond cancer.
The intended audience includes both basic scientists as well as clinicians. The book is directed
toward scientists of the academic and industrial sector, who are aiming to test and develop epigenetic cancer drugs. Importantly, the book will increase the awareness level of epigenetic drugs for
oncologists. We envision that this might foster increased and more rapid translation of epigenetic
drugs into the clinics.
Gerda Egger and Paola Arimondo

xxi


CHAPTER

BASIC EPIGENETIC MECHANISMS
AND PHENOMENA

1

Melanie R. Hassler1, Elisa Redl1, Quanah J. Hudson2, Wolfgang J. Miller3 and Gerda Egger1
1

Clinical Institute of Pathology, Medical University of Vienna, Vienna, Austria 2CeMM Research Center for
Molecular Medicine of the Austrian Academy of Sciences, Vienna, Austria 3Laboratories of Genome Dynamics,
Department of Cell and Developmental Biology, Center of Anatomy and Cell Biology,
Medical University of Vienna, Vienna, Austria


CHAPTER OUTLINE
1.1 Introduction ......................................................................................................................................4
1.2 Basic Epigenetic Mechanisms ...........................................................................................................6
1.2.1 DNA Methylation .......................................................................................................... 7
1.2.2 DNA Demethylation ...................................................................................................... 8
1.2.3 Histone Modifications ................................................................................................... 8
1.2.3.1 Histone Acetylation and Deacetylation....................................................................... 8
1.2.3.2 Histone Phosphorylation .........................................................................................10
1.2.3.3 Histone Methylation and Demethylation ..................................................................10
1.2.3.4 Chromatin-Remodeling Complexes and Histone Variants .........................................11
1.2.4 Noncoding RNAs ........................................................................................................ 11
1.3 Epigenetic (Re)Programming........................................................................................................... 12
1.3.1 Epigenetic Asymmetry in the Zygote............................................................................. 12
1.3.2 Reprogramming in the Germline .................................................................................. 14
1.3.3 Induced Pluripotency.................................................................................................. 14
1.4 Genomic Imprinting as a Model of Epigenetic Silencing ................................................................... 15
1.5 Dosage Compensation in Mammals ................................................................................................. 17
1.6 PEV in Drosophila .......................................................................................................................... 19
1.7 Transgenerational and Intergenerational Epigenetic Inheritance ....................................................... 20
1.8 Epigenetics and Disease................................................................................................................. 23
1.8.1 Selected Monogenetic Diseases ................................................................................... 23
1.8.2 Selected Neurodegenerative Diseases........................................................................... 26
1.8.3 Selected Autoimmune Diseases ................................................................................... 27
References ........................................................................................................................................... 28

G. Egger & P. Arimondo (Eds): Drug Discovery in Cancer Epigenetics. DOI: />© 2016 Elsevier Inc. All rights reserved.

3



4

CHAPTER 1 BASIC EPIGENETIC MECHANISMS AND PHENOMENA

1.1 INTRODUCTION
The history of epigenetics dates back to the early Greek philosopher Aristotle, who considered
development as a process that provides form from unformed material within male and female
germ cells and involves a dynamic course based on intrinsic factors [1]. During the last few centuries vivid debates were fueled by rather extreme scientific views as to the development of an
organism. Preformation as a predictable predetermined and stable process was opposed to epigenesis, which involves morphogenesis and differentiation based on regulatory response to the environment, cellular communication, and dynamic processes [2]. The term epigenetics was coined by
Conrad Hal Waddington to describe causal “mechanisms by which the genes of a genotype bring
about phenotypic effects” during development [3]. He introduced the concept of “epigenotype” to
indicate the complex processes and networks of genetic control linking the genotype and the phenotype [4]. Thus, disruption of genes at early stages of development could affect this epigenotype
and have far-reaching consequences on different organs and tissues. He modified this name from
Valentin Haecker, who created the term “phenogenetics” to describe visible stages of ontogeny
[5]. Being both an embryologist and geneticist he felt there was a need to combine genetics and
experimental embryology and the expression “epigenetics” for him was in accordance with the
classical concept of epigenesis. In the years and decades to follow, the term “epigenetics” was
used to describe a variety of phenomena and mechanisms [6] and up to now no clear definition as
to the use of the term has been adopted. The persistence of epigenetic systems to implement different phenotypes and a potential heredity or cellular inheritance was tied to epigenetics by Nanney
and Ephrussi, respectively [7,8]. With his new concept of carcinogenesis, Holliday [9] proposed
that malignant transformation not only results from genetic mutations but also from epigenetic
changes in DNA methylation resulting in heritable alterations in gene expression. This put DNA
methylation at center stage as an epigenetic mechanism. The discovery of the pivotal role of DNA
methylation for imprinting and X chromosome inactivation (XCI), plus the discovery of histonemodifying enzymes provided a revival of the term “epigenetic” in the 1990s [6,10À12]. Histone
modifications were suggested to provide an epigenetic code allowing for both transient regulation
of gene expression and long-term epigenetic memory [13,14]. Among epigenetic mechanisms,
RNAi was shown to induce not only posttranscriptional gene silencing but was also found to confer chromatin alterations including DNA methylation and histone modifications in plants and in
yeast [15]. Another noncoding RNA component, long noncoding RNA (lncRNA), was shown to
bridge long-range chromatin silencing and chromatin modification by recruitment of histonemodifying complexes both at the inactive X (Xi)-chromosome or gene-specific loci to promote

stable chromatin states [16,17]. Thus, epigenetics, which is in fact a transdisciplinary field, has
moved from observing phenomena to elucidating its underlying mechanisms, which include DNA
methylation, chromatin modification/remodeling, and noncoding RNA. All three are interrelated
and cooperate to maintain the stable epigenome of a given cell lineage in response to internal or
external stimuli (Figure 1.1).
A generally well-accepted definition of “modern” epigenetics was coined by Arthur D. Riggs
describing epigenetics as “The study of mitotically and/or meiotically heritable changes in gene
function that cannot be explained by changes in DNA sequence” [18]. Currently, the term
“epigenetics” is en vogue and used in very different scientific communities, including ethics and


1.1 INTRODUCTION

Genetics

Environment

Phenomena
X-chromosome inactivation
Imprinting
Position effect variegation
Mating type switching
Heterochromatin formation

5

Development
DNA
methylation


Histone modification/
remodeling

Epigenetic reprogramming
Cellular memory
Differentiation
Lineage commitment

Noncoding RNA

Disease
Loss of imprinting
Mutation in genes regulating epigenetics
Loss/gain of epigenetic marks
Deamination of MeC
Genomic instability

FIGURE 1.1
Epigenetic mechanisms and their role in different processes. Epigenetic mechanisms include DNA methylation,
histone modification, and remodeling, as well as noncoding RNA. All three are interrelated and are involved in
different phenomena and developmental processes, and also in a variety of diseases. Changes in epigenomic
signatures can be due to genetic alterations in epigenetic pathway proteins and to exposure to different
environmental stimuli.

social sciences or religious studies and the debate on “nature versus nurture” has been newly
sparked by discussions of whether environmental factors can alter the epigenome and be
transmitted to subsequent generations. Such discussions and studies are usually based on unresolved
questions and often lack biological evidence. Clearly, epigenetic mechanisms play a central role for
development, cellular differentiation, and homeostasis and their deregulation is associated with
diverse disease states including cancer. As discussed in more detail in Chapter 2, epigenetic



6

CHAPTER 1 BASIC EPIGENETIC MECHANISMS AND PHENOMENA

alterations in cancer can be the result of mutations in genes involved in epigenetic regulation and
very often occur as a consequence of exposure to harmful environmental factors.
One prime aspect of epigenetic regulation is the reversibility of epigenetic marks that can be
modulated by chemical compounds, which makes it an important target for developing new
therapeutic drugs, which will be addressed in great detail in the following chapters of this book.
Starting in the 1960s, antineoplastic activity of 5-azacytidine (5-azaCR) and 5-aza20 deoxycytidine (5-azaCdR) was shown against leukemia in mice and the first clinical studies
were initiated in 1967 in Europe [19À21]. Initially, these compounds were used as cytotoxic
agents as they were incorporated into RNA (5-azaCR) and DNA (5-azaCR, 5-azaCdR) and inhibit
DNA synthesis. Jones and Taylor then showed that cytidine analogs induced (trans)differentiation
of mouse embryonic and fibroblast cells into myotubes, striated muscle cells, adipocytes, and
chondrocytes, and that this was due to inhibition of DNA methylation of newly synthesized DNA
[22À24]. These two inhibitors were also the first epigenetic drugs to be approved by the FDA
and are currently in clinical use for patients with myelodysplastic syndromes (MDSs), acute myeloid leukemia, and chronic myelomonocytic leukemia [25,26]. At approximately the same time,
N-butyrate was identified to alter histone acetylation [27À29], and interestingly, HDAC inhibitors
were the second class of epigenetic drugs to be approved by the FDA for the treatment of cutaneous T-cell lymphoma [30]. The list of inhibitors targeting epigenetic machineries is ever-growing
and has recently attracted the interest of the pharmaceutical industry to develop programs for
epigenetic drug discovery.
This chapter will give an overview on basic epigenetic mechanisms including DNA
methylation, histone modification, and remodeling as well as noncoding RNA-based mechanisms.
Furthermore, we will address some classic epigenetic phenomena such as dosage compensation,
imprinting, and position effect variegation (PEV). Although epigenetic mechanisms were
suggested to maintain gene expression patterns persistently through subsequent cell generations it
was also suggested that environmental factors could cause dynamic epigenetic alterations and
result in diseases. Epidemiological studies have suggested that the environment can cause

phenotypic effects and directly impact on epigenetics, and that these changes can be heritable not
only during mitotic cell divisions but through subsequent generations. We will discuss some of
the current data available analyzing potential inter- and transgenerational epigenetic effects and
disease.

1.2 BASIC EPIGENETIC MECHANISMS
In eukaryotic cells DNA is not “naked” but exists as an intimate complex with specialized proteins
called histones, which together with DNA comprise chromatin. Nuclear DNA is spooled around
nucleosomal units consisting of small histone proteins in order to fit into the nucleus [31].
Chromatin exists in two forms: the less condensed euchromatin that is associated with
transcriptionally active regions and the highly condensed, usually transcriptionally inactive heterochromatin. Chromatin can be modified by posttranslational histone modifications, histone variants,
energy-dependent chromatin-remodeling steps that mobilize or alter nucleosome structures, and
noncoding RNA that can influence chromatin structure. Furthermore, DNA itself can be modified
covalently by methylation of cytosines, usually of CpG dinucleotides.


1.2 BASIC EPIGENETIC MECHANISMS

7

1.2.1 DNA METHYLATION
DNA methylation is an epigenetic modification that is correlated with gene repression and is known
to play an important role in gene regulation, development, and tumorigenesis. It consists of the
addition of a methyl group to the carbon at position 5 of cytosine residues in the DNA template. In
mammals it usually occurs at CpG dinucleotides but symmetric, asymmetric and non-CpG methylation is also known in embryonic stem cells (ESC), Neurospora crassa, and plants [32À34]. In
mammalian somatic tissues approximately 70% of all CpG sites within the DNA are methylated
and DNA methylation distribution shows enrichment in noncoding regions and interspersed repetitive elements [35] but not in CpG islands of active genes [36]. CpG islands were first defined in
1987 by Gardiner-Garden and Frommer as 200 bp stretches of DNA with a G 1 C content of at
least 50% and an observed CpG/expected CpG excess of 0.6 [37]. CpG islands are often associated
with promoters and usually lack DNA methylation, thus allowing gene expression. Approximately

60% of human genes have CpG island promoters. However, methylation of CpG islands can reinforce silencing of genes, for example, for genes on the Xi chromosome or for some imprinted
genes. Additionally, in cancer cells, genes are often aberrantly silenced by CpG island methylation.
DNA methylation is catalyzed by DNA methyltransferases (DNMTs), which can either catalyze de
novo methylation (DNMT3A/B) at novel sites or maintenance methylation (DNMT1) following DNA
replication [38]. DNMT1 is specific for CpG and its preferred substrate is hemimethylated DNA, DNA
methylated at CpG on one strand. Inactivation of Dnmt1 in mouse ESC results in genome-wide loss of
CpG methylation, indicating that this DNMT is necessary for stable maintenance of DNA methylation
[39]. However, Lei et al. [40] showed that de novo methylation of proviral DNA introduced into ES
cells was not catalyzed by DNMT1 but another unknown DNMT. Besides DNMT1, three candidate
proteins that could potentially encode additional DNMTs were found. DNMT2 had minimal DNMT
activity in vitro and deletion of Dnmt2 did not alter the level of DNA methylation [41]. However,
DNMT2 acts as an RNA methyltransferase that methylates tRNAs [42]. In contrast to DNMT1,
DNMT3A and DNMT3B did not show preference for hemimethylated DNA in vitro [43] and both
genes were necessary for de novo methylation of proviral genomes and repetitive elements in embryos
and ESC [44]. In mice, inactivation of Dnmt3a and Dnmt3b results in early embryonic lethality and
loss of one gene causes postnatal or embryonic lethality. Furthermore, in humans, mutations of
DNMT3B are associated with the ICF syndrome, a rare condition characterized by immunodeficiency,
centromeric instability, and facial abnormalities [45].
But how does DNA methylation interfere with gene expression? First, the presence of methyl
groups interferes with the binding ability of transcription factors that are crucial for transcriptional
activation. Transcription factors often recognize CG-rich motifs and several of these are unable to
bind to methylated CpG sequences [46]. Second, proteins that bind to methyl-CpGs can also repress
gene expression at methylated sites. These proteins were found by performing band-shift assays using
random methylated DNA sequences as probes [47]. MeCP1, a DNAÀprotein complex that is specific
for methylated DNA, was discovered in a variety of mammalian cell types. However, the first individual methyl-CpG-binding protein to be purified and cloned was MeCP2. Upon the presence of a
methyl-CpG-binding domain four other members of this family, MBD1, MBD2, MBD3, and MBD4,
could be identified [48]. Furthermore, two other structural domains are currently known to bind methylated DNA: the SET and RING finger-associated domain, found in UHRF1 and UHRF2 and zinc
fingers, found in Kaiso and Kaiso-like proteins [49]. Methyl-CpG-binding proteins recruit a variety of



8

CHAPTER 1 BASIC EPIGENETIC MECHANISMS AND PHENOMENA

proteins including histone deacetylases (HDACs) and chromatin-remodeling factors that are responsible for transcriptional repression [50] and can also interact with transcription factors, such as MeCP2
with TFIIB [51].

1.2.2 DNA DEMETHYLATION
Studies of cellular reprogramming have demonstrated that differentiated cellular states can be
radically altered, suggesting that DNA methylation may be reversible in mammalian cells. Today it
is known that not only passive DNA demethylation due to reduction in activity or absence of
DNMTs, but also active demethylation catalyzed by special enzymes takes place within a cell.
Active DNA demethylation plays an important role in early mammalian development as well as in
tissue-specific differentiation but is also observed in adult cells. One of the most prominent processes that involve rapid active DNA demethylation in adult cells is the activity-dependent
DNA demethylation of brain-derived neurotrophic factor and fibroblast growth factor 1 promoters
in postmitotic neurons [52]. 5-Methylcytosine (5mC) can be hydroxylated to
5-hydroxymethylcytosine (5hmC) by members of the tenÀeleven translocation (TET) enzyme family. 5hmC can further be oxidized to 5-formylcytosine or 5-carboxylcytosine. Furthermore, 5mC or
5hmC can be deaminated to 5-methyluracil or 5-hydroxymethyluracil by members of the AID/
APOBEC enzyme family. Finally, the intermediates produced by TET and AID/APOBEC are
replaced by the uracil DNA glycosylase [53] family of base excision repair glycosylases, that mediate DNA repair [54].

1.2.3 HISTONE MODIFICATIONS
The basic chromatin unit, or nucleosome, consists of a protein octamer containing two molecules
of each canonical (or core) histone (H2A, H2B, H3, H4), around which 147 bp of DNA are
wrapped. The core histones consist of a globular domain and flexible aminoterminal “histone tails”
[55]. These tails, particularly, those of H3 and H4, are accessible for modifications, such as
phosphorylation, acetylation, methylation, and ubiquitylation, which can be correlated to both
transcriptional activation or repression. In general, active marks include acetylation, arginine
methylation, and lysine methylation, such as H3K4 and H3K36, while repressive marks include
methylation of H3K9, H3K27, and H4K20. However, the globular domains of core histones can

also be modified [56] (Figure 1.2).
But how do histone modifications work in general? First, histone modifications can act in cis by
directly altering the chromatin structure by disrupting the contact between adjacent nucleosomes
or between histones and DNA, for example, by charge changes [59]. Second, histone modifications
can also act in trans and can be recognized by specific binding proteins, so-called
“reader” proteins, which can further recruit chromatin-modifying complexes and modulate chromatin structure [60].

1.2.3.1 Histone Acetylation and Deacetylation
Histone acetylation, catalyzed by histone acetyltransferases (HATs), can directly alter chromatin
structure through the loss of positive charges within histones, resulting in reduced interactions of histones with negatively charged DNA and increased accessibility of DNA-binding sites [61]. Reader


1.2 BASIC EPIGENETIC MECHANISMS

TSS

DRE

E
Enhancer

9

Promoter

NDR

E

E


Gene body

NDR

H3K4me1
H3K27ac
H3K9ac

H3K4me3
H4K20me1
(at TSS)

H2AZ
H3.3

H2AZ
H3.3

H3K4me1
H3K27me3

H3K27me3
H3K9me2

Downstream

NDR
H3K36me3
H3K79me2

H4K20me1 (5' exon)
H2BK5me1 (5' exon)

Less modified

H3.3

H3K27me3
H3K9me2

H3K27me3
H3K9me2

(BLOCs and LOCKs)

Intergenic

H3K9me3
H4K20me3

Heterochromatin

H3K9me3
H3.3

FIGURE 1.2
Histone modification and nucleosome occupancy determine different genomic regions. Dependent on their activity
state, genes are marked by characteristic chromatin modifications, nucleosome density, and histone variants, which
together reflect the on/off state of a gene (top). Note that only selected histone modification marks are illustrated.
NDR indicates nucleosome-depleted regions in enhancers, promoters and 30 regions of active genes. BLOCs and

LOCKs designate large silent gene-rich regions marked by repressive histone marks H3K27me3 and H3K9me2,
respectively [57,58]. (Bottom) Chromatin composition including histone modifications, grade of compaction, and
histone variants is indicated for intergenic regions and heterochromatin. Inverted arrows on top indicate inverted
repeats. These regions show generally high levels of DNA methylation and are associated with repressive histone
marks. Red, nucleosomes of active genes; blue, nucleosomes found in repressed genes, intergenic regions or
heterochromatin; DRE, distal response element; TSS, transcription start site; E, exon.

proteins with bromo-like domains recognize and bind acetylated histones. These proteins can either
be basal transcription factors, HATs, such as p300 or GCN5 or part of large chromatin-associating/
altering complexes such as the ATP-dependent remodeling complex SWI/SNF, which further alter
the chromatin structure [62]. The same principle applies to histone methylation, which is recognized
by chromo-like domains, and phosphorylation, which is recognized by 14-3-3 proteins [63,64].


10

CHAPTER 1 BASIC EPIGENETIC MECHANISMS AND PHENOMENA

Histone deacetylation is catalyzed by HDACs that remove acetyl groups [65]. HDACs are
grouped into four classes according to their homology with yeast proteins: type I, type II, and type
IV HDACs are referred to as “classical” HDACs and are zinc-dependent enzymes, whereas type III
or SIR2-related HDACs require the cofactor NAD. HDACs are frequently part of large multisubunit complexes, which target the enzyme to genes. This targeting often requires other histone interacting proteins, for example, the Sin3 corepressor, which is targeted by Rpd3 to H3K36me sites
and suppresses transcription in yeast [66,67].

1.2.3.2 Histone Phosphorylation
Histone phosphorylation was the first characterized histone modification since it has long been
understood that kinases regulate signal transduction pathways in the cell. In 1991, Mahadevan et al.
were able to show that stimulation of proliferation and transcription of the so-called “immediateearly” genes in cells correlate with histone H3 phosphorylation. Phosphorylation of H3S10 has
been shown to be required for chromosome condensation and segregation during mitosis [68].
H3S10 phosphorylation occurs at the onset of mitosis, interferes with HP1-H3K9me3 binding and

leads to release of heterochromatin protein 1 (HP1), which might be necessary for full mitotic chromatin condensation [69]. Furthermore, phosphorylation of the histone variant H2AX has been
shown to be associated with double-strand DNA breaks, where it helps to recruit repair proteins to
the site of the break [70].

1.2.3.3 Histone Methylation and Demethylation
Histone methylation is the most complex histone modification since it can occur on either lysines
or arginines and is associated with either transcriptional activation or repression. Furthermore,
lysines can be mono- (me1), di- (me2), or tri- (me3) methylated and arginines can be monomethylated or symmetrically or asymmetrically dimethylated.
Enzymes of three distinct families can catalyze lysine methylation of histones: the PRMT1 family, whose substrate is arginine, the SET domain containing protein family, and the non-SET
domain proteins DOT1/DOT1L, which methylate lysine residues [71,72].
In general, methylation of H3K4, H3K36, and H3K79 is linked to transcriptional activation,
however, depending on the methylation state and the genomic location, the same modification may
lead to different effects [59,73]. In contrast, methylation of H3K9, H3K27, H3K64, and H4K20, as
well as methylation of the linker histone H1, has been implicated in transcriptional repression [73].
Specific reader proteins bind to the methylation sites and alter gene expression. These proteins contain one of three distinct methyl lysine recognition domains: the so-called chromo, tudor or PHD
repeat domains [74]. Furthermore, H3K79me3 and H4K20me are associated with DNA repair [75].
For a long time it was not clear whether histone demethylation takes place in the cell or not.
Shi et al. [76] discovered the protein LSD1, an enzyme that removes methyl groups specifically
from H3K4. It can associate with different complexes and depending on the complex it can also
change its demethylation specificity [77]. Today several families of histone demethylases are
known, which act on various substrates and therefore play different roles in gene activation and
repression [78].


1.2 BASIC EPIGENETIC MECHANISMS

11

1.2.3.4 Chromatin-Remodeling Complexes and Histone Variants
Another major mechanism to alter chromatin and nucleosome composition in a noncovalent manner

is the recruitment of ATP-dependent chromatin-remodeling complexes. These protein complexes
can be generally categorized into two families: the SNF2H or ISWI family, which mobilizes
nucleosomes and the Brahma or SWI/SNF family, which alters the structure of nucleosomes [79].
Furthermore, core histones can be exchanged in an ATP-dependent manner by specialized histone variants. Histone variants are associated with specific expression states, genomic localization,
and species-distribution patterns. They change the structural and functional properties of nucleosomes by affecting chromatin remodeling and histone modifications. There are two important histone
variants of H3: H3.3, which marks transcriptionally active genes [80] and CenpA, which is found in
centromeric chromatin and is essential for centromeric function and chromosome segregation [81].
Histone H2A shows three histone variants. The presence of H2A.Z correlates with transcriptional activity, while H2A.X senses DNA damage and is therefore crucial for DNA repair [82].
MacroH2A specifically associates with the Xi chromosomes in mammals and is an epigenetic
regulator of key developmental genes.

1.2.4 NONCODING RNAs
Within the last few years different classes of noncoding RNAs have been identified, which can
either cause transcriptional or posttranscriptional gene silencing or directly impact on chromatin
structure by recruiting chromatin-modifying complexes.
RNA interference was originally discovered in Caenorhabditis elegans, where exogenously
introduced double-stranded RNA molecules are able to silence the expression of homologous
sequences within the genome [83]. RNAi and RNAi-mediated transcriptional and posttranscriptional silencing is best understood in Schizosaccharomyces pombe, where deletion of one
of the genes encoding for Argonaute, Dicer or RNA-dependent RNA polymerase proteins leads to
loss of heterochromatic gene silencing, reduced H3K9 methylation at centromeric repeats and accumulation of noncoding RNAs [84À86]. Furthermore, RNAi has also been shown to be required for
heterochromatin assembly in Schizosaccharomyces pombe and it was shown that pericentromeric
siRNA accumulation requires the H3K9 methyltransferase Clr4 [87]. In higher organisms, for
example, Drosophila melanogaster mutations of Argonaute proteins lead to a decrease of H3K9me
as well as HP1 binding [88] and mouse cell lines with mutations in Dicer show loss of silencing at
centromeres [89]. Furthermore, gene silencing triggered by RNAi can be long term and heritable,
as demonstrated in Caenorhabditis elegans, where four chromatin-remodeling factors are necessary
for maintaining the silent state of genes [90].
In higher organisms microRNAs (miRNA), a class of small noncoding RNAs, are essential for
the regulation of early differentiation programs. For example, HOX genes encode a family of transcriptional regulators important for patterning along body axes. Expression of HOX genes is regulated by different molecular mechanisms including nuclear dynamics, RNA processing,
translational regulation, and miRNAs [91]. Most of the miRNAs that regulate HOX gene expression

are encoded within the HOX cluster [92]. In mammals there are four HOX gene clusters, which
contain five genes encoding miRNA-10 and miRNA-96 [93].


12

CHAPTER 1 BASIC EPIGENETIC MECHANISMS AND PHENOMENA

An additional class of noncoding RNAs constitute lncRNAs, which are able to recruit histonemodifying enzymes using RNAi-independent strategies to particular genomic sites [94,95]. The
most prominent lncRNA in mammals so far is XIST that mediates XCI and was shown to target the
Polycomb repressive complex 2 (PRC2) including the H3K27 methyltransferase EZH2 to chromatin
[96,97]. Other lncRNAs that recruit PRC2 and confer silencing include HOTAIR, RepA, and
Kcnq1ot1 amongst others [94]. Generally, lncRNAs appear to work as scaffolds to enable complex
formation and recruitment to distinct genomic loci. A group of lncRNAs transcribed from enhancers can act as transcriptional activators in cis via DNA looping and recruitment of coactivators.
An example for this is the lncRNA HOTTIP, which recruits a H3K4 methyltransferase complex to
activate transcription of HOXA homeobox genes [98].

1.3 EPIGENETIC (RE)PROGRAMMING
Generating a broad range of different cell types in a highly ordered and reproducible manner is one
of the most remarkable hallmarks of complex genomes [99]. In mammals B25,000 genes contain
the information for the development of B200 different cell types. During development of multicellular organisms different cells and tissues acquire specific gene expression programs, which are
mainly regulated by epigenetic modifications such as DNA methylation and histone modification
[36,100,101]. These epigenetic mechanisms propagate appropriate patterns of gene expression and
can be heritable but potentially reversible [102].
Cells of multicellular organisms can be functionally divided into two major groups: totipotent
reproductive germ cells, necessary for the transmission of genetic information to the next generation, and differentiated somatic cells. For somatic cells epigenetic marks become fixed once the
cell differentiates [100] and developmental processes in these cells are usually associated with a
progressive loss of developmental potential. However, in cells undergoing dedifferentiation (e.g.,
cancer cells), or transdifferentiation, reprogramming can also take place in differentiated somatic
cells [103].

Major reprogramming and global erasure of epigenetic marks take place at two time points in
the mammalian life cycle: during early embryonic development after fertilization and in primordial
germ cells (PGCs), where totipotency or pluripotency is restored and parental imprints are erased
(Figure 1.3).

1.3.1 EPIGENETIC ASYMMETRY IN THE ZYGOTE
At fertilization, the parental genomes are in different stages of the cell cycle and contain divergent
epigenetic marks and chromatin composition [100]. The paternal genome is single copy (1C),
packed densely with protamines instead of histones, whereas the maternal genome is arrested in
metaphase II (2C) and contains histones. Upon fertilization and zygote formation protamines in the
paternal genome are rapidly replaced with histones that lack H3K9me2 and H3K27me3, followed
by an active, genome-wide loss of DNA methylation. In contrast, the maternal pronucleus contains
histones with H3K9me2 and H3K27me3 marks that were acquired during oocyte growth and does
not lose 5mC at this early stage, but rather by a passive mechanism during subsequent


1.3 EPIGENETIC (RE)PROGRAMMING

13

FIGURE 1.3
Epigenetic reprogramming. After fertilization, the paternal genome (bottom blue line) becomes actively
demethylated rapidly in the zygote, while the maternal genome (bottom red line) is demethylated by a passive
mechanism. Both genomes are remethylated at the time of implantation. During embryonic development, another
reprogramming event is occurring in germ cell development in PGCs, were genome-wide DNA methylation is
followed by remethylation in male germ cells (top blue line) in prospermatogonia and in female germ cells (top
red line) after birth. Both reprogramming events are associated also with changes in repressive histone
modification patterns as indicated by methylation changes in histone H3K27 and H3K9.

cell divisions. This asymmetry in paternal and maternal epigenomes can be detected up to the

four-cell state. However, the functional importance of this asymmetry in the zygote is not clear
[99]. Gill et al. [104] hypothesized that chromatin modifications established in the gametes may be
required for proper embryonic development, and it could be shown that loss of early zygotic paternal demethylation perturbs the activation of several pluripotency-associated genes and impairs
development [105].
During preimplantation development a passive loss of 5mC can be detected in the maternal
genome until the blastocyst state when the inner cell mass acquires high levels of 5mC, H3K9me2,
and H3K27me3. In contrast to the paternal active DNA demethylation catalyzed by TET enzymes,
it has been hypothesized that DNA demethylation in the maternal genome occurs passively through
exclusion of the maintenance methyltransferase DNMT1 from the nucleus [106]. During active
paternal demethylation the maternal genome is protected by DPPA3 (also called STELLA) from
TET3 oxidation [107]. DPPA3 is essential for embryonic development during preimplantation and
loss of Dppa3 leads to embryonic lethality in mice [108]. Furthermore, although DNA demethylation is a global process during preimplantation development, some genomic sequences, like


14

CHAPTER 1 BASIC EPIGENETIC MECHANISMS AND PHENOMENA

imprinted loci, centromeric heterochromatin, and intracisternal A particle (IAP) retrotransposons
maintain DNA methylation during preimplantational development, which is also mediated by
DPPA3.
All in all, an elaborate temporal and spatial epigenetic modification program seems to be
necessary for zygote development and to exploit the entire potential of the genome. However
totipotency is lost as cleavage progresses [99]. Histone modifications as well as DNA methylation
are necessary for regulation of lineage induction and defects in DNA and histone methyltransferases can lead to impaired differentiation in ESC [109]. Polycomb repressive complexes are
required for the bivalency of key regulator developmental genes in a yet transcriptionally poised
state, where activating and repressing histone marks are present at the same time, ensuring
lineage flexibility.

1.3.2 REPROGRAMMING IN THE GERMLINE

A second major epigenetic reprogramming event takes place during development of germ cells
in the early embryo. PGCs are formed from epiblast cells by signaling molecules that are
produced by extraembryonic ectoderm and primary endoderm [110]. This process is associated
with a loss of 5mC and H3K9me2 and an increase of H3K27me3, a more flexible histone
modification that allows rapid activation of poised developmental regulator genes [111]. Global
demethylation of PCGs occurs between E11.5 and E12.5 of mouse development and leads to
erasure of methylation in imprinted genes and single copy genes [112] and to reactivation of the
X chromosome in females [113]. However, there is evidence that not all epigenetic marks
are erased in imprinted genes in PGCs. The paternal H19 and the maternal Snrpn alleles are
demethylated during PGC development but later during spermatogenesis and oogenesis they
need to be methylated again. This de novo methylation of formerly methylated genes seems to
happen at an earlier stage than de novo methylation of originally unmethylated genes [114,115],
indicating that other epigenetic marks are not completely erased during PGC development and
provide signals for de novo methylation [100]. To distinguish between somatic cells and PGCs
several markers can be used, which reflect the different genetic programs of PGC and somatic
cell specification. PGCs were found to express several pluripotency genes, including Sox2 and
Oct4, which were not expressed in the neighboring somatic cells, indicating that PGCs exhibit
pluripotency, which is lost in somatic cells. On the other hand some genes, including Hoxb1
and Hoxa1, which are important for cell differentiation and morphogenesis are significantly
upregulated in somatic cells, but not expressed in PGCs, indicating that PGCs repress somatic
cell fate [116].

1.3.3 INDUCED PLURIPOTENCY
Briggs and King [117] performed the first nuclear transfer experiments with cells isolated from late
blastula stage frog embryos from Rana pipiens that developed into complete embryos when
transplanted into enucleated oocytes. Subsequently, Gurdon and colleagues succeeded in producing
sexually mature frogs from nuclear transplantation of adult nuclei in Xenopus laevis [118].
Meanwhile, it could be demonstrated that adult somatic cell nuclei can be reprogrammed by
nuclear transfer from different species including mammals [119]. These studies indicated on one



1.4 GENOMIC IMPRINTING AS A MODEL OF EPIGENETIC SILENCING

15

hand that developmental restrictions are due to reversible epigenetic modifications [119] and on the
other hand that the oocyte must contain factors that mediate the reprogramming of adult cells into
an embryonic state [103].
However, the big breakthrough was the discovery of transcription factors that can induce pluripotency in somatic cells to generate induced pluripotent stem (iPS) cells without the need for
oocytes. Takahashi and Yamanaka [120] identified four transcription factors, OCT4 (O), SOX2 (S),
KLF4 (K), and cMYC (M) that were able to reprogram adult mouse fibroblasts into ES-like iPS
cells. The advantages of transcription-factor-induced reprogramming lies in its simplicity and
robustness. By ectopic expression of OSKM a broad range of different cell types can easily be
reprogrammed to pluripotency. In their ideal state iPS cells are functionally indistinguishable from
ESC, can form chimeras and teratomas, and are able to differentiate into cells of all three germ
layers [121]. However, the induction of pluripotency upon OSKM expression requires a latency
period of 1À2 weeks and occurs only in 1% of the starting cells. Furthermore, it has been shown
that the differentiation state of a somatic cell can influence the efficiency of iPS cell generation
and that compounds that inhibit DNA methylation and histone modification can increase the
efficiency [122].

1.4 GENOMIC IMPRINTING AS A MODEL OF EPIGENETIC SILENCING
Monoallelic expression can occur randomly for specific clusters of genes, for example, for B- or
T-cell receptor genes [123], and for up to 10À24% of individual genes per cell [123,124]. Genetic
differences between alleles that affect the activity of promoters or cis-regulatory elements can also
lead to monoallelic or strongly biased allelic expression that may affect many genes in human
[125]. In contrast, parental-specific monoallelic or imprinted expression is relatively rare, with only
approximately 0.5% of genes (125 in mouse) showing imprinted expression in mammals (http://igc.
otago.ac.nz, [126]). In spite of their small number, many imprinted genes play important roles in
development and growth, as was already demonstrated in the 1980s by experiments that showed

early embryonic lethality of androgenic and gynogenic embryos, indicating that both the maternally
and paternally inherited genomes together are required for normal development [127,128].
Following this, a number of developmental disorders have been shown to be caused by defects at
specific imprinted gene loci, such as the chromosome 15q11-13 imprinted region, where paternal
mutations can cause PraderÀWilli syndrome and maternal mutations can cause Angelman syndrome [129]. In such diseases a nonmutated allele is present, but epigenetically silenced by the
imprinting mechanism, opening the possibility for treatment by reactivation of the silent allele with
epigenetic drugs.
Genomic imprinting is a classical epigenetic process in mammals controlled by DNA methylation that is deposited in genomic regions called imprint control elements (ICEs) in either the male
or female germline by the de novo DNA methyltransferases DNMT3A and DNMT3L [130À132]
(Figure 1.4). Differential DNA methylation is then maintained on the same parental allele in all
somatic cells in the organism by DNMT1 [133]. The majority of reported imprinted genes in mice
lie near to one of the 24 reported gametic differentially methylated regions (gDMRs) [134,135],
seven of which have been demonstrated to be ICEs by genetic deletion that resulted in loss of


16

CHAPTER 1 BASIC EPIGENETIC MECHANISMS AND PHENOMENA

FIGURE 1.4
Regulation of imprinted expression. Imprinted silencing is controlled by differential DNA methylation on the ICE
that is established during either oogenesis or spermatogenesis by the DNMT3A/3L de novo DNMTs complex and
maintained on the same parental chromosome by DNMT1. The unmethylated ICE then acts to silence a cluster of
imprinted genes in cis, in the most common mechanism by acting as a promoter for a lncRNA that then causes
imprinted silencing. Although, differential DNA methylation of the ICE is present in all somatic cells, and the
lncRNA is also expressed in most cell types, many imprinted genes are only susceptible to imprinted silencing in
some cell types or developmental stages, and hence show tissue-specific imprinted expression.

imprinted expression of nearby genes [10]. This indicates that most imprinted genes are in coregulated clusters controlled by an ICE, that ranges in size from 3 to 12 genes spread over
100À3700 kilobases (kb) [10]. DNA methylation is associated with gene silencing, but paradoxically many imprinted genes are expressed from the same allele as where the ICE is methylated.

This is explained by the mechanism by which imprinted silencing is achieved. Different
mechanisms of imprinted silencing via the ICE have been shown, but the most common mechanism
described so far for four of the seven defined ICEs (the Igf2r, Kcnq1, Gnas, and PraderÀWilli clusters) is by lncRNA-mediated silencing [136À139]. In these cases the ICE is associated with the
promoter of the lncRNA and silenced on the maternal allele by DNA methylation. The lncRNA is
then expressed and silences imprinted genes in cis on the paternal allele. Similarly, in the Igf2 cluster where imprinted silencing is not controlled by an lncRNA, the imprinted protein-coding genes
Igf2 and Ins2 are expressed from the paternal allele where the ICE is methylated. In this cluster the


1.5 DOSAGE COMPENSATION IN MAMMALS

17

insulator protein CTCF binds the unmethylated maternal ICE blocking Igf2 from accessing enhancers, while CTCF binding is prevented by methylation on the paternal allele allowing the enhancers
to activate Igf2 expression [140,141]. Hence, in imprinted silencing DNA, methylation acts to
repress the repressor allowing the expression of imprinted genes. For a limited number of imprinted
genes DNA methylation can then play a secondary role to reinforce silencing directly on the silent
promoter, forming a somatic DMR [10].
Genomic imprinting is a phenomenon with a clear phenotype, making it a useful model to
investigate aspects of epigenetic gene regulation. With the advent of high-throughput sequencing
technologies, thousands of lncRNAs have been recently identified [142], but imprinted lncRNAs
remain among the few lncRNAs that have so far been demonstrated to be functional by genetic
experiments. lncRNAs could act by their transcription alone or via their RNA product to regulate
gene expression. Posttranscriptional knockdown by RNA interference-based approaches will
not affect the function of lncRNAs that act via their transcription, while genetic deletions of
lncRNAs could also remove cis-regulatory elements such as enhancers, making neither strategy
ideal for investigating lncRNA function. Therefore, in the imprinting field lncRNAs have been
truncated by introducing a polyadenylation cassette to investigate their function and mechanism
of action [136À139]. Using this approach, the lncRNA Airn that silences Igf2r in mouse was
truncated to various lengths, demonstrating that Airn need only be transcribed across the Igf2r
promoter to cause silencing due to transcriptional interference [143]. In contrast, the Airn

lncRNA product was shown to be necessary to silence the nonoverlapped gene Slc22a3 by
recruiting the repressive EHMT H3K9 dimethyltransferase to its promoter [144]. Similarly, the
Kcnq1ot1 lncRNA responsible for silencing genes in the Kcnq1 cluster is associated with
EHMT2 and the PRC2 H3K27trimethyltransferase complex, and imprinted expression of some
genes in the Kcnq1 cluster is lost when these complexes are disrupted, indicating that Kcnq1ot1
may also cause silencing by recruiting repressive histone-modifying complexes [145]. The experimental approaches developed by studying imprinted lncRNAs, and the silencing mechanisms
uncovered, may be applied to investigate the function and mechanism of action of lncRNAs
outside of imprinted regions.
The majority of imprinted genes show tissue-specific imprinted expression [146], indicating that
they may have a dose-dependent function in specific tissues. This indicates that the imprinted
silencing mechanism may act tissue-specifically, which has been hypothesized to occur, for example, by imprinted lncRNAs blocking tissue-specific enhancer activity [147]. Open questions in
well-studied imprinted clusters, and the likely possibility that other imprinted clusters may achieve
imprinted silencing in different ways, makes genomic imprinting a fertile field of study to further
elucidate mechanisms of epigenetic gene silencing.

1.5 DOSAGE COMPENSATION IN MAMMALS
The difference in the distribution of the sex chromosomes in mammals requires a mechanism of
gene dose balancing to adjust for the double amount of X-linked genes in females. H.J. Muller was
the first to develop the concept of dosage compensation in Drosophila following his studies on
X-linked eye pigment genes [148]. While in Drosophila the dosage difference is equaled by