Content uploaded by Gangning Liang
Author content
All content in this area was uploaded by Gangning Liang on Jan 27, 2015
Content may be subject to copyright.
insight review articles
T
he term ‘epigenetics’ defines all meiotically and
mitotically heritable changes in gene expression
that are not coded in the DNA sequence itself.
Three systems, including DNA methylation,
RNA-associated silencing and histone modifica-
tion, are used to initiate and sustain epigenetic silencing.
Unravelling the relationships between these components
has led to surprising and rapidly evolving new concepts,
showing how they interact and stabilize each other (Fig. 1).
Disruption of one or other of these interacting systems can
lead to inappropriate expression or silencing of genes,
resulting in ‘epigenetic diseases’. Here we discuss potential
causes of some of these diseases, and suggest how they might
be treated in the future.
Methylation of the C
5
position of cytosine residues in
DNA has long been recognized as an epigenetic silencing
mechanism of fundamental importance
1,2
. The methyla-
tion of CpG sites within the human genome is maintained
by a number of DNA methyltransferases (DNMTs) and
has multifaceted roles for the silencing of transposable
elements, for defence against viral sequences and for the
transcriptional repression of certain genes. 5-Methylcyto-
sine is highly mutagenic, causing C:G to T:A transitions and
resulting in a strong suppression of the CpG methyl-
acceptor site in human DNA (Box 1). CpG islands, which
are regions of more than 500 base pairs in size and with a GC
content greater than 55% (ref. 3), have been conserved
during evolution because they are normally kept free of
methylation. These stretches of DNA are located within the
promoter regions of about 40% of mammalian genes and,
when methylated, cause stable heritable transcriptional
silencing. Aberrant de novo methylation of CpG islands is
a hallmark of human cancers and is found early during
carcinogenesis
4
.
Histone modifications have also been defined as epi-
genetic modifiers. Post-translational modifications of his-
tones, including acetylation and methylation of conserved
lysine residues on the amino-terminal tail domains, have
been studied closely over the past few years. Generally, the
acetylation of histones marks active, transcriptionally com-
petent regions, whereas hypoacetylated histones are found
in transcriptionally inactive euchromatic or heterochro-
matic regions. Histone methylation can be a marker for
both active and inactive regions of chromatin. Methylation
of lysine 9 on the N terminus of histone H3 (H3-K9) is a
hallmark of silent DNA and is globally distributed through-
out heterochromatic regions such as centromeres and
telomeres. It is also found on the inactive X chromosome
and at silenced promoters
5
. In contrast, methylation of
lysine 4 of histone H3 (H3-K4) denotes activity and is
found predominantly at promoters of active genes
5
.
Because lysine methylation can be monomeric, dimeric or
trimeric, and histones may also be subject to other post-
translational modifications such as phosphorylation
6
, this
enormous variation leads to a multiplicity of possible com-
binations of different modifications. This might constitute
a ‘histone code’
7
, which can be read and interpreted by
different cellular factors.
Links between histone modifications and DNA
methylation have been found in plants and fungi (Fig. 1),
where H3-K9 methylation is a prerequisite for DNA
methylation
8–10
. DNA methylation can also trigger H3-K9
methylation
11–13
, and this has also been shown in mammals,
Epigenetics in human disease and
prospects for epigenetic therapy
Gerda Egger, Gangning Liang, Ana Aparicio & Peter A. Jones
Departments of Biochemistry and Molecular Biology and Urology, USC/Norris Comprehensive Cancer Center, Keck School of Medicine of the
University of Southern California, 1441 Eastlake Avenue, Room 8302L, Los Angeles, California 90089-9181, USA
(e-mail: jones_p@ccnt.hsc.usc.edu)
Epigenetic mechanisms, which involve DNA and histone modifications, result in the heritable silencing of
genes without a change in their coding sequence. The study of human disease has focused on genetic
mechanisms, but disruption of the balance of epigenetic networks can cause several major pathologies,
including cancer, syndromes involving chromosomal instabilities, and mental retardation. The development
of new diagnostic tools might reveal other diseases that are caused by epigenetic alterations. Great potential
lies in the development of ‘epigenetic therapies’ — several inhibitors of enzymes controlling epigenetic
modifications, specifically DNA methyltransferases and histone deacetylases, have shown promising anti-
tumorigenic effects for some malignancies.
?
?
DNA
methylation
Histone
modification
Heritable silencing
RNA
?
?
Figure 1 Interaction between RNA, histone modification and DNA
methylation in heritable silencing. Histone deacetylation and other
modifications, particularly the methylation of lysine 9 within histone H3
(H3-K9) residues located in the histone tails, cause chromatin
condensation and block transcriptional initiation. Histone modification can
also attract DNA methyltransferases to initiate cytosine methylation, which
in turn can reinforce histone modification patterns conducive to silencing.
Experiments in yeast and plants have clearly shown the involvement of
RNA interference in the establishment of heterochromatic states and
silencing. RNA triggering of heritable quiescence might therefore also be
involved in higher organisms.
NATURE
|
VOL 429
|
27
MAY 2004
|
www.nature.com/nature 457
JONES new 19/5/04 1:33 pm Page 457
© 2004
Nature
Publishing
Group
insight review articles
458 NATURE
|
VOL 429
|
27 MAY 2004
|
www.nature.com/nature
where H3-K9 methylation directs DNA methylation to peri-
centromeric heterochromatin
14
. Interactions between histone
deacetylases (HDACs), histone methyltransferases and methyl-
cytosine-binding proteins
15,16
lead to the recruitment of DNA
methyltransferases
17
, although it is not yet clear what initiates the
recruitment of the different epigenetic modifiers to their specific
target sequences.
The role of RNA in post-transcriptional silencing has attracted
much interest. However, RNA, in the form of antisense transcripts,
noncoding RNAs (such as Xist) or RNA interference (RNAi), can
also lead to mitotically heritable transcriptional silencing by the
formation of heterochromatin. In Schizosaccharomyces pombe, for
example, the deletion of different components of the RNAi
machinery results in an impairment of centromere function, a
derepression of transgenes integrated at centromeres, and a loss of
the characteristic H3-K9 methylation within the same region
18,19
. A
similar connection between RNAi, DNA methylation and H3-K9
methylation has been demonstrated in Arabidopsis thaliana
20
.
Although RNAi-directed silencing of heterochromatic regions has
not yet been shown in mammals, the involvement of RNA in different
silencing mechanisms has been described. For example, antisense
RNAs are involved in the silencing of some mammalian imprinted
genes
21
and in dosage compensation in mammals
22
. A recent report of
a case of -thalassaemia showed how antisense transcription could
lead to DNA methylation and stable silencing of a globin gene
23
. RNA
might therefore be a key trigger to direct histone modifications (for
example, H3-K9 methylation) and DNA methylation to specific loci
(for example, pericentromeric heterochromatin), thereby evoking
heritable and stable silencing. The therapeutic activation of abnor-
mally silenced genes thus requires drugs that can target the inherent
stability of these multifaceted changes.
Epigenetic diseases
Multicellular organisms require mutually reinforcing mechanisms
permitting heritable patterns of gene silencing. Mutations in genes
that affect global epigenetic profiles can give rise to human diseases,
which can be inherited or somatically acquired (Table 1). Interestingly,
many of these epigenetic abnormalities result in chromosomal alter-
ations and learning disabilities. For example, mutations in the ATRX
gene result in consistent changes in the pattern of methylation of
ribosomal DNA, Y-specific repeats and subtelomeric repeats. Fragile
X syndrome results when a CGG repeat in the FMR1 5 untranslated
region expands and becomes methylated de novo, causing the gene to
be silenced and creating a visible ‘fragile’ site on the X chromosome
under certain conditions. On a more global level, the ICF (immuno-
deficiency, centromeric region instability and facial anomalies)
Box 1
Evolution and methylation of CpG islands
Early genomes did not contain 5-methyl-
cytosine, and CpG sites (cream circles in the
figure) occurred as frequently as expected on
a statistical basis. As genomes become
methylated at CpG sites (pink circles),
promoters of about half of human genes
were somehow protected from this
modification in the germ line and remained as
CpG islands. Because of the well-known
enhanced mutability of methylated cytosine,
CpGs were converted to TpGs and depleted
from the rest of the genome, particularly in
the transcribed regions of genes in which
most of the methylation occurs. About 40%
of the human genome is made up of
transposable elements (four cream circles in
a group) or their relics, and the active
elements are capable of inserting themselves
both into genes and into heterochromatic
regions. Methylation of the CpG sites within
these transposed elements results in the
silencing of their promoters at the same time
as it does not hinder the transcription of the
host gene. In this way, mammals are different
from Neurospora crassa, in which
methylation of cytosine residues in the
transcribed regions blocks elongation by
polymerase II. Organisms such as Drosophila
melanogaster, which have very little cytosine
methylation, show a higher degree of
transposition of transposable elements
91
.
The methylation of transposable elements
such as Alus, LINEs and LTR sequences
increases the rate of C to T transition
mutations at these sites, so these sequences eventually become
depleted of CpGs. The interaction between transposable elements
and the cytosine methylation system results in genomic expansion
and CpG depletion. Physiological methylation of CpG islands in the
promoters of X-linked or imprinted genes, for example, leads to their
long-term silencing. Pathological methylation can result in the
silencing of tumour-suppressor or other cancer-relevant genes.
Epigenetic therapy seeks to reverse these changes.
Methylation
Deamination and mutation
Retrotransposition, methylation
and genome expansion
Deamination and mutation
De novo
methylation and silencing
JONES new 19/5/04 1:33 pm Page 458
© 2004
Nature
Publishing
Group
insight review articles
NATURE
|
VOL 429
|
27 MAY 2004
|
www.nature.com/nature 459
syndrome is caused by mutations in the DNMT3b gene, which is an
essential enzyme required for the establishment of DNA methylation
patterns
24
. The fact that all of these diseases show gross chromo-
somal anomalies points to a central role for epigenetic mechanisms
in chromosome architecture.
Several inherited syndromes are due to faulty genomic imprinting
— defined as parent-specific, monoallelic expression of a gene — such
as Angelman’s syndrome, Prader–Willi syndrome and Beckwith–
Wiedemann syndrome (BWS). In these conditions, an abnormal
phenotype is established as a result of the absence of the paternal or
maternal copy of an imprinted gene or because of deregulation of an
imprinted gene.
For example, a cluster of imprinted genes at 11p15.5 is involved in
the pathology of BWS, a syndrome characterized by organ over-
growth and association with embryonal tumours such as Wilms’
tumour. Loss of methylation in imprinting control regions can cause a
deregulation of imprinting and either biallelic expression (such as
IGF2) or silencing (such as CDKN1C) of imprinted genes, which is
found in most sporadic cases of BWS.
Particular interest has recently been generated by the finding that
one of the most common forms of intellectual disability in young
girls, namely Rett syndrome, is due to germline mutations of MeCP2
(ref. 25). The MeCP2 protein binds to methylcytosine residues
26
and
the disease pathogenesis might be linked to the derepression of genes
normally suppressed by DNA methylation. No direct evidence has yet
been found for this, because human cells with MeCP2 mutations do
not show global gene derepression
26
. However, MeCP2 might have a key
role in the control of neuronal gene activity resulting in the pathology of
Rett syndrome
27,28
.
As mentioned above, a fascinating recent discovery was that the
transcription of an antisense RNA led to gene silencing and to the
methylation of a structurally normal -globin gene in a patient with
thalassaemia
23
. Other examples of inappropriate gene silencing might
contribute to disease yet might be missed as part of conventional
diagnosis. This case of thalassaemia might well be the tip of the iceberg
and indicates that several other kinds of human disease might be the
result of genomic rearrangements that cause epigenetic silencing.
Epigenetic changes can also have a major role in the development
of human cancer. For example, a high percentage of patients with
sporadic colorectal cancers with a microsatellite instability phenotype
show methylation and silencing of the gene encoding MLH1 (ref. 29).
Thus, epigenetic silencing can result directly in genetic instability. In
some cases, promoter-associated methylation of MLH1is found not only
in tumour but also in normal somatic tissues, including spermatozoa.
These germline ‘epimutations’ predispose individuals carrying aberrant
methylation patterns to multiple cancers
30,31
. Disruption of path-
ways that lead to cancer is often caused by the de novo methylation of
the relevant gene’s promoters
4
. Epigenetic silencing has been recog-
nized as a third pathway satisfying Knudson’s hypothesis that two
hits are necessary for the silencing of tumour-suppressor genes
32
.
Chromatin-modifying enzymes have also been associated with
the aetiology of different haemopathologies. A characteristic of
human leukaemias is the presence of various chromosomal
translocations, leading to the expression of fusion proteins. Both
histone acetyltransferases and histone methyltransferases can be
part of such fusions and cause the upregulation of target genes
33
. In
acute promyelocytic leukaemia, the oncogenic fusion protein
PML–RAR (promyelocytic leukaemia–retinoic acid receptor-)
recruits an HDAC to repress genes essential for the differentiation
of haematopoietic cells
34
. Similarly, in acute myeloid leukaemia,
AML1–ETO fusions recruit the repressive N-CoR–Sin3–HDAC1
complex and inhibit myeloid development
35
.
The importance of correct chromatin composition is further
underlined by the roles of ATP-dependent chromatin remodelling
complexes in disease. These are multisubunit complexes that are
capable of moving and shifting nucleosomes, thereby regulating
transcription. Several members of the highly conserved SWI–SNF
complex have been implicated in cancer
36
. For example, a loss of
SNF5 is observed in paediatric cancers, and the ATPase subunits
BRM and BRG1 are mutated in a variety of cancer cell lines and pri-
mary tumours; this is associated with a poorer prognosis in patients
with non-small-cell lung cancer
36
.
Epigenetic drift
A distinguishing feature of epigenetic changes in comparison with
genetic changes is that they tend to be acquired in a gradual rather
than an abrupt process. For example, a generalized decrease in
genomic 5-methylcytosine concentrations occurs as mammalian
cells age
37,38
, and this gradual loss of DNA methylation can result in
aberrant gene activation. The decrease in 5-methylcytosine occurs
at the same time as there is a localized hypermethylation of CpG
islands at gene promoters
39
. The accumulation of methylated CpGs
within the promoter might be acting as a ‘rheostat’ rather than a
‘switch’ for gene silencing
40
. These alterations are therefore targets
for prevention strategies. It has already been shown that a lowering
of DNA methylation or DNA methyl-binding proteins can decrease
the number of intestinal polyps in cancer-prone mice
41
. Although
the full extent of epigenetic changes is largely unknown, there are
obvious advantages to designing strategies that can prevent the
aberrant regulation of genes as a function of age.
Table 1 Epigenetic diseases
Disease Symptom Aetiology References
ATR-X syndrome Intellectual disabilities, -thalassaemia Mutations in ATRX gene, hypomethylation of certain repeat and satellite 82
sequences
Fragile X syndrome Chromosome instability, intellectual disabilities Expansion and methylation of CGG repeat in FMR1 5 UTR, promoter 83
methylation
ICF syndrome Chromosome instability, immunodeficiency DNMT3b mutations, DNA hypomethylation 84
Angelman’s syndrome Intellectual disabilities Deregulation of one or more imprinted genes at 15q11–13 (maternal) 85
Prader–Willi syndrome Obesity, intellectual disabilities Deregulation of one or more imprinted genes at 15q11–13 (paternal) 86
BWS Organ overgrowth Deregulation of one or more imprinted genes at 11p15.5 (e.g. IGF2)87
Rett syndrome Intellectual disabilities MeCP2 mutations 25,26
-Thalassaemia (one case) Anaemia Methylation of 2-globin CpG island, deletion of HBA1 and HBQ1 23
Various cancers Microsatellite instability De novo methylation of MLH1 29
Disruption of Rb, p53 pathway, uncontrolled proliferation De novo methylation of various gene promoters 4
Disruption of SWI–SNF chromatin remodelling complex Mutations in SNF5, BRG1, BRM 36
Overexpression of IGF2, silencing of CDKN1C Loss of imprinting 88, 89
Leukaemia Disturbed haematopoiesis Chromosomal translocations involving HATs and HMTs 62
Rubinstein–Taybi syndrome Intellectual disabilities Mutation in CREB-binding protein (histone acetylation) 90
Coffin–Lowry syndrome Intellectual disabilities Mutation in Rsk-2 (histone phosphorylation) 90
ATR-X syndrome, -thalassaemia, mental retardation syndrome, X linked; BWS, Beckwith–Wiedemann syndrome; CREB, cAMP-response-element-binding protein; HAT, histone acetyltransferase;
HMT, histone methyltransferase; ICF, immunodeficiency, centromeric region instability and facial anomalies syndrome; UTR, untranslated region.
JONES new 19/5/04 1:33 pm Page 459
© 2004
Nature
Publishing
Group
Initial clinical trials used increasingly cytotoxic doses of these
agents, even though the induction of gene expression shows a bell-
shaped response curve
44
. Exciting new clinical trials have shown that
low doses of these agents might be efficacious in treating myeloid
dysplastic syndrome and other leukaemias
47
. In addition, azanucleo-
sides are being tested for the treatment of haemoglobinopathies
48
, in
which the aim is to cause the demethylation of the promoters of the
fetal globin genes that have become silenced in the patient as part of
normal development, leading to an increase in haemoglobin F to
correct the anaemia that characterizes these diseases.
A disadvantage of the azanucleosides is their instability in aque-
ous solutions, but this might be overcome by the use of other ana-
logues, such as zebularine or 5-fluoro-2-deoxycytidine, which also
inhibit DNA methylation after incorporation into DNA and might
be orally active
49
. Procainamide, which is used to treat cardiac
arrhythmias, is also an inhibitor of DNA methylation
50
. In addition,
natural products derived from tea
51
and from sponges
52
have shown
activity in vitro. Clinical trials with antisense oligonucleotides that
target the DNA methyltransferases are also underway
53
.
Epigenetic silencing is almost universally associated with histone
deacetylation, which is catalysed by at least three classes of HDACs in
human cells. The HDACs are partly redundant in function, and a
growing series of small molecules has been designed to inhibit their
activities either globally or more specifically (Table 2). HDAC
inhibitors can induce differentiation, growth arrest and/or apoptosis
in transformed cells in culture and in tumours. The driving hypothe-
sis is that accumulation of acetylated proteins, particularly histones,
results in the induction of genes and the upregulation of others that
have become epigenetically silenced. In particular, the gene encoding
p21, which is a cell-cycle kinase inhibitor, is commonly upregulated
in tumour cells treated with these agents in the absence of p53 (ref.
54). This is important in terms of cancer therapy, because many can-
cers have no functional p53 and are therefore unable to arrest cells in a
p53-dependent fashion. Different HDAC inhibitors are being used
intravenously or orally in several phase I and II clinical trials
55
, in
which changes in histone acetylation have been documented.
Because there are many different HDACs, it will be important in the
future to design therapies that can target individual enzymes and
thus increase the precision of this approach.
Coupling therapies
The links between histone modification and DNA methylation (see
Fig. 1) have encouraged investigators to think about dual therapies
insight review articles
460 NATURE
|
VOL 429
|
27 MAY 2004
|
www.nature.com/nature
Table 2 Epigenetic drugs
Target Drug Clinical trials
DNA methylation 5-Azacytidine Phase I/II/III
5-Aza-2-deoxycytidine Phase I/II/III
FCDR
Zebularine
Procainamide
EGCG Phase I
Psammaplin A
Antisense oligomers Phase I
Histone deacetylase Many
55
, including:
Phenylbutyric acid Phase I/II
SAHA Phase I/II
Depsipeptide Phase I/II
Valproic acid Phase I/II
EGCG, epigallocatechin-3-gallate; FCDR, 5-fluoro-2-deoxycytidine; SAHA, suberoylanilide
hydroxamic acid.
Figure 2 Mechanism of action of nucleoside
analogue inhibitors. Deoxynucleoside analogues
such as 5-aza-2-deoxycytidine (depicted by Z) are
converted into the triphosphate inside S-phase cells
and are incorporated in place of cytosine into DNA.
Ribonucleosides such as 5-azacytidine or
zebularine are reduced at the diphosphate level by
ribonucleotide reductase for incorporation (not
shown). Once in DNA, the fraudulent bases form
covalent bonds with DNA methyltransferases
(DNMTs), resulting in the depletion of active
enzymes and the demethylation of DNA. Pink
circles, methylated CpG; cream circles,
unmethylated CpG.
Z
Z
Z
ZZ
Z
-MP
Z
-DP
Z
-TP
Z
Z
Z
DNA replication
DNMTs
DNMTs
Strand separation
Z
Z
Epigenetic therapy
The fact that many human diseases, including cancer, have an epi-
genetic aetiology has encouraged the development of a new thera-
peutic option that might be termed ‘epigenetic therapy.’ Many agents
have been discovered that alter methylation patterns on DNA or the
modification of histones, and several of these agents are currently
being tested in clinical trials (Table 2).
Inhibitors of DNA methylation rapidly reactivate the expression
of genes that have undergone epigenetic silencing, particularly if this
silencing has occurred in a pathological situation. The prototype
inhibitors, 5-azacytidine (5-aza-CR) and 5-aza-2-deoxycytidine
(5-aza-CdR), were initially developed as cytotoxic agents
42
, but it was
subsequently discovered that they are powerful inhibitors of DNA
methylation and induce gene expression and differentiation in cultured
cells
43,44
. Both nucleoside analogues are converted to the deoxy-
nucleotide triphosphates and are then incorporated in place of cyto-
sine into replicating DNA (Fig. 2). They are therefore active only in
S-phase cells
43
, where they serve as powerful mechanism-based
inhibitors of DNA methylation
44
. DNA methyltransferases get trapped
on DNA containing modified bases such as azacytosine, 5-fluoro-
cytosine, pseudoisocytosine or zebularine, resulting in the formation
of heritably demethylated DNA
44,45
. Covalent attachment of the
various DNA methyltransferases to DNA might well be responsible
for the cytotoxicities of these agents, particularly at high doses
46
.
JONES new 19/5/04 1:33 pm Page 460
© 2004
Nature
Publishing
Group
insight review articles
NATURE
|
VOL 429
|
27 MAY 2004
|
www.nature.com/nature 461
combining DNA methylation inhibitors with HDAC inhibitors
(Box 2). Jahangeer et al.
56
first showed that 5-aza-CR and butyrate (an
HDAC inhibitor) acted synergistically to upregulate the -adrenergic
receptor in HeLa cells. Subsequently, Ginder et al.
57
showed that the
two agents acted synergistically in anaemic chickens to increase the
level of embryonic p-type globin messenger RNA in the haematopoi-
etic cells. The synergy of demethylation and HDAC inhibitors was
investigated in greater detail by Cameron et al.
58
. Suzuki et al.
59
and
Yamashita et al.
60
have used the approach of treating cultured cells
with 5-aza-CdR and trichostatin A to isolate new tumour-suppressor
genes efficiently. In addition, a recent study has shown a strong
synergy between 5-aza-CdR and phenyl butyrate (an HDAC
inhibitor) in the prevention of murine lung cancer
61
. This is therefore
an exciting time, as clinical trials are developed that can target these
two epigenetic modifications in patients
62
. In addition to these
approaches, it might be advantageous to sensitize cells by epigenetic
therapy followed by treatment with chemotherapy
63
, interferon
64
or
immunotherapy
65
, among others. In terms of cancer therapy, it will
be essential to include both genetic (see reviews in this issue by
Strausberg et al. (page 469) and Bell (page 453)) and epigenetic
markers
66
, which will permit an individually targeted therapy.
Potential pitfalls of epigenetic therapy
Despite the promise of epigenetic therapy, there are several concerns
regarding the clinical applications of these agents. These relate mainly
to the nonspecific activation of genes and transposable elements in
normal cells, and also to potential mutagenicity and carcinogenicity.
Unfortunately, few studies have examined the effects of azanucleo-
sides on completely normal cells as opposed to cell lines. The drugs
have profound effects on immortal lines
44
, yet only 0.4% of 6,600
genes (compared with 1% in tumour cell lines) analysed were
upregulated more than fourfold in normal human fibroblasts
exposed to 5-aza-CdR
67
. Early studies showed that 5-aza-CR could
activate a human X chromosome in a rodent somatic cell hybrid
68
but
not in normal human cells
69
. These data suggest that DNA methyla-
tion is only one of the mechanisms enforcing silencing in normal cells
(Fig. 1) and that they are therefore less sensitive to drug-induced gene
activation. Imprinted genes can be activated by 5-aza-CdR
70
,
implying a need for caution. Azanucleosides have been shown to be
mutagenic
71
and possibly carcinogenic in rats
72
, and might be able
to activate silenced oncogenes
73
, although they can clearly act as cancer-
prevention agents
41,61
.
Azanucleosides have shown some promising results in clinical
trials without much evidence of adverse effects. For example, treat-
ment of 41 leukaemia patients with 5-aza-CdR showed only mild
effects on global genomic demethylation, as measured by changes
in Alu methylation
74
. Original methylation levels were regained
within two weeks after therapy, and no development of a secondary
malignancy was observed in follow-up studies. Furthermore,
administration of a low dose of 5-aza-CdR induced cytogenetic
remissions in a substantial number of patients with myelodysplastic
syndrome with pre-existing chromosomal abnormalities
75
. No
increase in chromosomal instability was observed after therapy,
which argues against a strong effect of 5-aza-CdR on chromosomal
integrity in patients.
The therapeutic mechanism of action of DNA methylation and
HDAC inhibitors is by no means straightforward. Evidently, both
classes of agents can activate genes — but is this the way that they
work in patients? HDAC inhibitors and DNA methylation
inhibitors are cytotoxic agents and induce cell-cycle arrest and
apoptosis by upregulating p21 and/or p53 (refs 55, 76). Loss of
genomic methylation causes p53-dependent apoptosis
77
, and p53
represses DNMT1 (ref. 78), suggesting a feedback loop between the
two proteins. In addition, the binding of proteins, including DNA
methyltransferases, to the DNA of cells treated with azanucleo-
sides can result in cytotoxicity
46,79
. These points make it imperative
that surrogate endpoints be examined in patients to gain a better
understanding of the mechanism of action.
Future perspectives
Epigenetic disorders give rise to several significant human diseases,
and the race is on to find therapies that can reverse silencing. Some-
times it might be possible to reuse a wild-type embryonic gene in
place of a mutated adult gene such as in thalassaemia or sickle-cell
anaemia. Rett syndrome is caused by mutations in the MeCP2 gene,
which resides on the X chromosome. Because X-inactivation is mostly
random, about 50% of the cells in these girls harbour a suppressed
wild-type gene, so a valid target might be the reactivation of this good
but unused copy. Because of the potential risks of epigenetic therapy,
it is likely that trials to validate the approach will be based on patients
with life-threatening diseases such as cancer. Here, it is important to
remember that the target is an abnormally methylated CpG island
and that these seem to be particularly sensitive to reactivation by
DNA methylation inhibitors in cancer cells. In addition, because
multiple genes become methylated in individual cancers
80
, there is
the possibility of hitting many targets with one drug. Finally, because
methylation of CpG islands increases with age
39
and could therefore
contribute to the development of chronic diseases in addition to can-
cer, there might be benefits in drugs and lifestyle changes that could
bring about reversion, or slow gradual epigenetic silencing.
It is apparent that we are just at the beginning of understanding
the substantial contributions of epigenetics to human disease, and
there are probably many surprises ahead. For example, the finding
HDAC
X
H3-K9 HMT,
DNMTs
X
HAT
HDAC inhibitor DNMT inhibitor
Low concentration
High concentration
Histone
Histone
X
X
Ac
Me
MBD
Genes containing CpG islands in their promoters (for example, the
p16 tumour-suppressor gene) can be reversibly controlled
(opened or shut) by altered levels of histone acetylation and the
presence or absence of transcription factors at the promoter.
Heritable silencing (locking) is achieved by multiple histone
modification changes, including trimethylation of the H3-K9,
binding of methylated DNA-binding proteins such as MeCP2, and
de novo methylation of the CpG island
92
. The exact order in which
these changes are acquired is not certain, although it seems likely
that H3-K9 methylation precedes cytosine methylation
93
. Once
these changes have occurred, they tend to reinforce each other
and the gene becomes refractory to reactivation (upper panel).
Demethylated, silenced genes can often be activated by histone
deacetylase (HDAC) inhibitors, but these agents are largely
ineffective with methylated promoters
58
. In contrast, DNA
methylation inhibitors are highly effective not only at removing the
cytosine methylation but also at rapidly reversing the chromatin
structural changes
94,95
, and not only in unlocking the gene but
also in opening it for transcription. New clinical trials seek to take
advantage of the synergy between low doses of DNA methylation
inhibitors and HDAC inhibitors (lower panel). DNMT, DNA
methyltransferase; HAT, histone acetyltransferase; HMT, histone
methyltransferase; MBD, methyl-CpG-binding domain protein.
Box 2
Gene silencing and pharmacological reactivation
JONES new 19/5/04 1:33 pm Page 461
© 2004
Nature
Publishing
Group
that loss of imprinting can be seen not only in normal colonic
epithelium but also in the lymphocytes of colorectal patients was
completely unexpected
81
. Elucidating the whole bandwidth of epi-
genetic mechanisms is an exciting challenge and will eventually lead
to a clearer understanding of the development of human disease and
direct therapeutic concepts into new directions.
doi:10.1038/nature02625
1. Holliday, R. & Pugh, J. E. DNA modification mechanisms and gene activity during development.
Science 187, 226–232 (1975).
2. Riggs, A. D. X inactivation, differentiation, and DNA methylation. Cytogenet. Cell Genet. 14,
9–25 (1975).
3. Takai, D. & Jones, P. A. Comprehensive analysis of CpG islands in human chromosomes 21 and 22.
Proc. Natl Acad. Sci. USA 99, 3740–3745 (2002).
4. Jones, P. A. & Baylin, S. B. The fundamental role of epigenetic events in cancer. Nature Rev. Genet. 3,
415–428 (2002).
5. Lachner, M. & Jenuwein, T. The many faces of histone lysine methylation. Curr. Opin. Cell Biol. 14,
286–298 (2002).
6. Fischle, W., Wang, Y. & Allis, C. D. Binary switches and modification cassettes in histone biology and
beyond. Nature 425, 475–479 (2003).
7. Strahl, B. D. & Allis, C. D. The language of covalent histone modifications. Nature 403, 41–45 (2000).
8. Tamaru, H. & Selker, E. U. A histone H3 methyltransferase controls DNA methylation in Neurospora
crassa. Nature 414, 277–283 (2001).
9. Jackson, J. P., Lindroth, A. M., Cao, X. & Jacobsen, S. E. Control of CpNpG DNA methylation by the
KRYPTONITE histone H3 methyltransferase. Nature 416, 556–560 (2002).
10.Malagnac, F., Bartee, L. & Bender, J. An Arabidopsis SET domain protein required for maintenance
but not establishment of DNA methylation. EMBO J. 21, 6842–6852 (2002).
11.Johnson, L., Cao, X. & Jacobsen, S. Interplay between two epigenetic marks. DNA methylation and
histone H3 lysine 9 methylation. Curr. Biol. 12, 1360–1367 (2002).
12.Soppe, W. J. et al. DNA methylation controls histone H3 lysine 9 methylation and heterochromatin
assembly in Arabidopsis. EMBO J. 21, 6549–6559 (2002).
13.Tariq, M. et al. Erasure of CpG methylation in Arabidopsis alters patterns of histone H3 methylation in
heterochromatin. Proc. Natl Acad. Sci. USA 100, 8823–8827 (2003).
14.Lehnertz, B. et al. Suv39h-mediated histone H3 lysine 9 methylation directs DNA methylation to
major satellite repeats at pericentric heterochromatin. Curr. Biol. 13, 1192–1200 (2003).
15.Nan, X. et al. Transcriptional repression by the methyl-CpG-binding protein MeCP2 involves a
histone deacetylase complex. Nature 393, 386–389 (1998).
16.Fuks, F. et al. The methyl-CpG-binding protein MeCP2 links DNA methylation to histone
methylation. J. Biol. Chem. 278, 4035–4040 (2003).
17.Fuks, F., Burgers, W. A., Brehm, A., Hughes-Davies, L. & Kouzarides, T. DNA methyltransferase
Dnmt1 associates with histone deacetylase activity. Nature Genet. 24, 88–91 (2000).
18.Hall, I. M. et al. Establishment and maintenance of a heterochromatin domain. Science 297,
2232–2237 (2002).
19.Volpe, T. A. et al. Regulation of heterochromatic silencing and histone H3 lysine-9 methylation by
RNAi. Science 297, 1833–1837 (2002).
20.Zilberman, D., Cao, X. & Jacobsen, S. E. ARGONAUTE4 control of locus-specific siRNA
accumulation and DNA and histone methylation. Science 299, 716–719 (2003).
21.Rougeulle, C. & Heard, E. Antisense RNA in imprinting: spreading silence through Air. Trends Genet.
18, 434–437 (2002).
22.Panning, B. & Jaenisch, R. RNA and the epigenetic regulation of X chromosome inactivation. Cell 93,
305–308 (1998).
23.Tufarelli, C. et al. Tr anscription of antisense RNA leading to gene silencing and methylation as a novel
cause of human genetic disease. Nature Genet. 203, 157–165 (2003).
24.Okano, M., Bell, D. W., Haber, D. A. & Li, E. DNA methyltransferases Dnmt3a and Dnmt3b are
essential for de novo methylation and mammalian development. Cell 99, 247–257 (1999).
25.Amir, R. E. et al. Rett syndrome is caused by mutations in X-linked MECP2, encoding methyl-CpG-
binding protein 2. Nature Genet. 23, 185–188 (1999).
26.Klose, R. & Bird, A. Molecular biology. MeCP2 repression goes nonglobal. Science 302,
793–795 (2003).
27.Chen, W. G. et al. Derepression of BDNF transcription involves calcium-dependent phosphorylation
of MeCP2. Science 302, 885–889 (2003).
28.Martinowich, K. et al. DNA methylation-related chromatin remodeling in activity-dependent BDNF
gene regulation. Science 302, 890–893 (2003).
29.Kane, M. F. et al. Methylation of the hMLH1 promoter correlates with lack of expression of hMLH1 in
sporadic colon tumors and mismatch repair-defective human tumor cell lines. Cancer Res. 57,
808–811 (1997).
30.Gazzoli, I., Loda, M., Garber, J., Syngal, S. & Kolodner, R. D. A hereditary nonpolyposis colorectal
carcinoma case associated with hypermethylation of the MLH1 gene in normal tissue and loss of
heterozygosity of the unmethylated allele in the resulting microsatellite instability-high tumor.
Cancer Res. 62, 3925–3928 (2002).
31.Suter, C. M., Martin, D. I. & Ward, R. L. Germline epimutation of MLH1 in individuals with multiple
cancers. Nature Genet. advance online publication 4 April 2004 (doi:10.1038/ng1342).
32.Jones, P. A. & Laird, P. W. Cancer epigenetics comes of age. Nature Genet. 21, 163–167 (1999).
33.Hake, S. B., Xiao, A. & Allis, C. D. Linking the epigenetic ‘language’ of covalent histone modifications
to cancer. Br. J. Cancer 90, 761–769 (2004).
34.Grignani, F. et al. Fusion proteins of the retinoic acid receptor- recruit histone deacetylase in
promyelocytic leukaemia. Nature 391, 815–818 (1998).
35.Jones, L. K. & Saha, V. Chromatin modification, leukaemia and implications for therapy. Br. J.
Haematol. 118, 714–727 (2002).
36.Roberts, C. W. & Orkin, S. H. The SWI/SNF complex — chromatin and cancer. Nature Rev. Cancer 4,
133–142 (2004).
37.Wilson, V. L. & Jones, P. A. DNA methylation decreases in aging but not in immortal cells. Science
220, 1055–1057 (1983).
insight review articles
462 NATURE
|
VOL 429
|
27 MAY 2004
|
www.nature.com/nature
38.Richardson, B. C. Role of DNA methylation in the regulation of cell function: autoimmunity, aging
and cancer. J. Nutr. 132, 2401S–2405S (2002).
39.Issa, J. P. CpG-island methylation in aging and cancer. Curr. Top. Microbiol. Immunol. 249,
101–118 (2000).
40.Beaudet, A. L. & Jiang, Y. H. A rheostat model for a rapid and reversible form of imprinting-
dependent evolution. Am. J. Hum. Genet. 70, 1389–1397 (2002).
41.Laird, P. W. et al. Suppression of intestinal neoplasia by DNA hypomethylation. Cell 81,
197–205 (1995).
42.Sorm, F., Piskala, A., Cihak, A. & Vesely, J. 5-Azacytidine, a new, highly effective cancerostatic.
Experientia 20, 202–203 (1964).
43.Constantinides, P. G., Jones, P. A. & Gevers, W. Functional striated muscle cells from non-myoblast
precursors following 5-azacytidine treatment. Nature 267, 364–366 (1977).
44.Jones, P. A. & Taylor, S. M. Cellular differentiation, cytidine analogs and DNA methylation. Cell 20,
85–93 (1980).
45. Zhou, L. et al. Zebularine: a novel DNA methylation inhibitor that forms a covalent complex with DNA
methyltransferases. J. Mol. Biol. 321, 591–599 (2002).
46.Michalowsky, L. A. & Jones, P. A. Differential nuclear protein binding to 5-azacytosine-containing
DNA as a potential mechanism for 5-aza-2-deoxycytidine resistance. Mol. Cell. Biol. 7,
3076–3083 (1987).
47.Issa, J. P. et al. Phase 1 study of low-dose prolonged exposure schedules of the hypomethylating agent
5-aza-2-deoxycytidine (decitabine) in hematopoietic malignancies. Blood 103, 1635–1640 (2004).
48.Saunthararajah, Y. et al. Effects of 5-aza-2-deoxycytidine on fetal hemoglobin levels, red cell
adhesion, and hematopoietic differentiation in patients with sickle cell disease. Blood 102,
3865–3870 (2003).
49.Cheng, J. C. et al. Inhibition of DNA methylation and reactivation of silenced genes by zebularine.
J. Natl Cancer Inst. 95, 399–409 (2003).
50.Lin, X. et al. Reversal of GSTP1 CpG island hypermethylation and reactivation of pi-class glutathione
S-transferase (GSTP1) expression in human prostate cancer cells by treatment with procainamide.
Cancer Res. 61, 8611–8616 (2001).
51.Fang, M. Z. et al. Tea polyphenol ()-epigallocatechin-3-gallate inhibits DNA methyltransferase and
reactivates methylation-silenced genes in cancer cell lines. Cancer Res. 63, 7563–7570 (2003).
52.Pina, I. C. et al. Psammaplins from the sponge Pseudoceratina purpurea: inhibition of both histone
deacetylase and DNA methyltransferase. J. Org. Chem. 68, 3866–3873 (2003).
53.Yan, L. et al. Specific inhibition of DNMT1 by antisense oligonucleotides induces re-expression of
estrogen receptor- (ER) in ER-negative human breast cancer cell lines. Cancer Biol. Ther. 2, 552–556
(2003).
54.Xiao, H., Hasegawa, T. & Isobe, K. Both Sp1 and Sp3 are responsible for p21waf1 promoter activity
induced by histone deacetylase inhibitor in NIH3T3 cells. J. Cell. Biochem. 73, 291–302 (1999).
55.Marks, P. A., Miller, T. & Richon, V. M. Histone deacetylases. Curr. Opin. Pharmacol. 3,
344–351 (2003).
56.Jahangeer, S., Elliott, R. M. & Henneberry, R. C. -Adrenergic receptor induction in HeLa cells:
synergistic effect of 5-azacytidine and butyrate. Biochem. Biophys. Res. Commun. 108,
1434–1440 (1982).
57.Ginder, G. D., Whitters, M. J. & Pohlman, J. K. Activation of a chicken embryonic globin gene in adult
erythroid cells by 5-azacytidine and sodium butyrate. Proc. Natl Acad. Sci. USA 81,
3954–3958 (1984).
58.Cameron, E. E., Bachman, K. E., Myohanen, S., Herman, J. G. & Baylin, S. B. Synergy of
demethylation and histone deacetylase inhibition in the re-expression of genes silenced in cancer.
Nature Genet. 21, 103–107 (1999).
59.Suzuki, H. et al. A genomic screen for genes upregulated by demethylation and histone deacetylase
inhibition in human colorectal cancer. Nature Genet. 31, 141–149 (2002).
60.Yamashita, K. et al. Pharmacologic unmasking of epigenetically silenced tumor suppressor genes in
esophageal squamous cell carcinoma. Cancer Cell 2, 485–495 (2002).
61.Belinsky, S. A. et al. Inhibition of DNA methylation and histone deacetylation prevents murine lung
cancer. Cancer Res. 63, 7089–7093 (2003).
62.Claus, R. & Lubbert, M. Epigenetic targets in hematopoietic malignancies. Oncogene 22,
6489–6496 (2003).
63.Plumb, J. A., Strathdee, G., Sludden, J., Kaye, S. B. & Brown, R. Reversal of drug resistance in human
tumor xenografts by 2-deoxy-5-azacytidine-induced demethylation of the hMLH1 gene promoter.
Cancer Res. 60, 6039–6044 (2000).
64.Karpf, A. R. & Jones, D. A. Reactivating the expression of methylation silenced genes in human
cancer. Oncogene 21, 5496–5503 (2002).
65.Weber, J. et al. Expression of the MAGE-1 tumor antigen is up-regulated by the demethylating agent
5-aza-2-deoxycytidine. Cancer Res. 54, 1766–1771 (1994).
66.Laird, P. W. The power and the promise of DNA methylation markers. Nature Rev. Cancer 3,
253–266 (2003).
67.Liang, G., Gonzales, F. A., Jones, P. A., Orntoft, T. F. & Thykjaer, T. Analysis of gene induction in
human fibroblasts and bladder cancer cells exposed to the methylation inhibitor 5-aza-2-
deoxycytidine. Cancer Res. 62, 961–966 (2002).
68.Mohandas, T., Sparkes, R. S. & Shapiro, L. J. Reactivation of an inactive X human chromosome:
evidence for X inactivation by DNA methylation. Science 211, 393–396 (1981).
69.Wolf, S. F. & Migeon, B. R. Studies of X chromosome DNA methylation in normal human cells.
Nature 295, 667–671 (1982).
70.Eversole-Cire, P. et al. Activation of an imprinted Igf 2 gene in mouse somatic cell cultures. Mol. Cell.
Biol. 13, 4928–4938 (1993).
71.Jackson-Grusby, L., Laird, P. W., Magge, S. N., Moeller, B. J. & Jaenisch, R. Mutagenicity of 5-aza-2-
deoxycytidine is mediated by the mammalian DNA methyltransferase. Proc. Natl Acad. Sci. USA 94,
4681–4685 (1997).
72.Carr, B. I., Rahbar, S., Asmeron, Y., Riggs, A. & Winberg, C. D. Carcinogenicity and haemoglobin
synthesis induction by cytidine analogues. Br. J. Cancer 57, 395–402 (1988).
73.Sato, N. et al. Frequent hypomethylation of multiple genes overexpressed in pancreatic ductal
adenocarcinoma. Cancer Res. 63, 4158–4166 (2003).
74.Yang, A. S., Estecio, M. R., Garcia-Manero, G., Kantarjian, H. M. & Issa, J. P. Comment on
‘Chromosomal instability and tumors promoted by DNA hypomethylation’ and ‘Induction of
tumors in nice by genomic hypomethylation’. Science 302, 1153 (2003).
JONES new 19/5/04 1:33 pm Page 462
© 2004
Nature
Publishing
Group
insight review articles
NATURE
|
VOL 429
|
27 MAY 2004
|
www.nature.com/nature 463
75.Lubbert, M. et al. Cytogenetic responses in high-risk myelodysplastic syndrome following low-dose
treatment with the DNA methylation inhibitor 5-aza-2-deoxycytidine. Br. J. Haematol. 114,
349–357 (2001).
76.Karpf, A. R., Moore, B. C., Ririe, T. O. & Jones, D. A. Activation of the p53 DNA damage response
pathway after inhibition of DNA methyltransferase by 5-aza-2-deoxycytidine. Mol. Pharmacol. 59,
751–757 (2001).
77.Jackson-Grusby, L. et al. Loss of genomic methylation causes p53-dependent apoptosis and epigenetic
deregulation. Nature Genet. 27, 31–39 (2001).
78.Peterson, E. J., Bogler, O. & Taylor, S. M. p53-mediated repression of DNA methyltransferase 1
expression by specific DNA binding. Cancer Res. 63, 6579–6582 (2003).
79.Juttermann, R., Li, E. & Jaenisch, R. Toxicity of 5-aza-2-deoxycytidine to mammalian cells is
mediated primarily by covalent trapping of DNA methyltransferase rather than DNA demethylation.
Proc. Natl Acad. Sci. USA 91, 11797–11801 (1994).
80.Esteller, M., Corn, P. G., Baylin, S. B. & Herman, J. G. A gene hypermethylation profile of human
cancer. Cancer Res. 61, 3225–3229 (2001).
81.Cui, H. et al. Loss of IGF2 imprinting: a potential marker of colorectal cancer risk. Science 299,
1753–1755 (2003).
82.Gibbons, R. J. & Higgs, D. R. Molecular-clinical spectrum of the ATR-X syndrome. Am. J. Med. Genet.
97, 204–212 (2000).
83.Oostra, B. A. & Willemsen, R. The X chromosome and fragile X mental retardation. Cytogenet.
Genome Res. 99, 257–264 (2002).
84.Ehrlich, M. The ICF syndrome, a DNA methyltransferase 3B deficiency and immunodeficiency
disease. Clin. Immunol. 109, 17–28 (2003).
85.Nicholls, R. D., Saitoh, S. & Horsthemke, B. Imprinting in Prader–Willi and Angelman syndromes.
Tre nds Genet. 14, 194–200 (1998).
86.Goldstone, A. P. Prader–Willi syndrome: advances in genetics, pathophysiology and treatment.
Tre nds Endocrinol. Metab. 15, 12–20 (2004).
87.Maher, E. R. & Reik, W. Beckwith–Wiedemann syndrome: imprinting in clusters revisited. J. Clin.
Invest. 105, 247–252 (2000).
88.Feinberg, A. P. & Tycko, B. The history of cancer epigenetics. Nature Rev. Cancer 4, 143–153 (2004).
89.Soejima, H. et al. Silencing of imprinted CDKN1C gene expression is associated with loss of CpG and
histone H3 lysine 9 methylation at DMR-LIT1 in esophageal cancer. Oncogene published online
8 March 2004 (doi:10.1038/sj.onc.1207576).
90.Ausio, J., Levin, D. B., De Amorim, G. V., Bakker, S. & Macleod, P. M. Syndromes of disordered
chromatin remodeling. Clin. Genet. 64, 83–95 (2003).
91.Yoder, J. A., Walsh, C. P. & Bestor, T. H. Cytosine methylation and the ecology of intragenomic
parasites. Trends Genet. 13, 335–340 (1997).
92.Nguyen, C. T., Gonzales, F. A. & Jones, P. A. Altered chromatin structure associated with methylation-
induced gene silencing in cancer cells: correlation of accessibility, methylation, MeCP2 binding and
acetylation. Nucleic Acids Res. 29, 4598–4606 (2001).
93.Bachman, K. E. et al. Histone modifications and silencing prior to DNA methylation of a tumor
suppressor gene. Cancer Cell 3, 89–95 (2003).
94.El-Osta, A. & Wolffe, A. P. DNA methylation and histone deacetylation in the control of gene
expression: basic biochemistry to human development and disease. Gene Expr. 9, 63–75 (2000).
95.Nguyen, C. T. et al. Histone H3-lysine 9 methylation is associated with aberrant gene silencing in
cancer cells and is rapidly reversed by 5-aza-2-deoxycytidine. Cancer Res. 62, 6456–6461 (2002).
Acknowledgements We thank C. Nguyen for his help with the figures, and A. Yang and
J. Rice for critical reading of the manuscript. This work was supported by the National
Cancer Institute and the Max Kade Foundation.
Competing interests statement The authors declare competing financial interests: details
accompany the paper on www.nature.com/nature.
JONES new 19/5/04 1:33 pm Page 463
© 2004
Nature
Publishing
Group
