© 2008 The Authors
`Journal compilation © 2008 Nordic Pharmacological Society. Basic & Clinical Pharmacology & Toxicology, 103, 389–396
`
`Doi: 10.1111/j.1742-7843.2008.00325.x
`
`Blackwell Publishing Ltd
`
`MiniReview
`
`Epigenetic Changes in Cancer as Potential Targets for Prophylaxis and
`Maintenance Therapy
`
`Kirsten Grønbæk1, Marianne Treppendahl1, Fazila Asmar1 and Per Guldberg2
`1Department of Haematology, Rigshospitalet, Copenhagen, Denmark, and 2Institute of Cancer Biology, Danish Cancer Society,
`Copenhagen, Denmark
`
`(Received June 15, 2008; Accepted July 28, 2008)
`
`Abstract: Epigenetic silencing of gene transcription by methylation of DNA or modification of histones is a key event in
`neoplastic initiation and progression. Alterations of the epigenome have been identified in virtually all types of cancer and
`involve multiple genes and molecular pathways. Recent studies have suggested that epigenetic gene inactivation may repre-
`sent the first step in tumorigenesis, possibly by affecting the normal differentiation of stem cells and by predisposing these
`cells to additional oncogenic insults. The mechanisms that drive epigenetic silencing in pre-malignant cells are still
`unknown, but may reflect simple stochastic events that are beneficial to cancer precursor cells. It is now well established
`that epigenetically silenced genes may be reactivated pharmacologically. Some inhibitors of DNA methyltransferases (5-
`aza-cytidine and 5-aza-2′-deoxycytidine) or histone deacetylases (vorinostat) have been approved for clinical use by the US
`Food and Drug Administration and have reached clinical phase III trials elsewhere. The prospect that epigenetic alterations
`may play an essential role in renewing and maintaining the malignant clone has opened up new perspectives for the use of
`epigenetic therapy in cancer prevention and maintenance.
`
`Disruption of cellular pathways that control the proliferation
`and death of cells is a fundamental event in the initiation
`and progression of cancer. The expression or function of
`proteins in these pathways may be altered by gene mutations
`or by epigenetic changes that alter gene expression without
`disrupting the nucleotide template. Like genetic changes,
`epigenetic alterations are somatically heritable and may lead
`to a clonal expansion of cells. Recent evidence suggests that
`early-stage cancer cells carry epigenetic modifications of
`growth- and differentiation-associated genes, which may
`predispose the cells to accumulation of additional changes
`in oncogenes and tumour suppressor genes and eventually
`lead to a full-blown cancer.
`Changes in the epigenome involve several types of cova-
`lent modifications of DNA and associated proteins. The
`best-studied changes include modifications of the amino-
`terminal ends (‘tails’) of core histones (the proteins around
`which the DNA is wrapped), and methylation of cytosines
`in DNA (fig. 1). It is well established that hypermethylation
`of DNA and deacetylation of histones in promoter regions
`is associated with down-regulation of tumour suppressor
`genes in most cancers. In contrast to genetic alterations,
`gene silencing by epigenetic modifications is potentially
`
`Author for correspondence: Kirsten Grønbæk, Department of
`Haematology, L4042, Rigshospitalet, Blegdamsvej 9, DK-2100
`Copenhagen, Denmark (fax +45 35459542, e-mail kirsten.groenbaek@
`rh.regionh.dk).
`
`reversible. Treatment with drugs that target the epigenome,
`for example, by inhibiting DNA methyltransferases or
`histone deacetylases, can reactivate the transcription of
`silenced genes and restore normal cellular growth and
`differentiation.
`Recently, the effects of DNA methyltransferase inhibitors
`and histone deacetylase inhibitors in advanced cancers
`have been investigated in many clinical trials. However, if
`epigenetic changes are involved in the earliest stages of cancer
`and contribute to maintaining the malignant clone after
`treatment, epigenetic therapy may also become a powerful
`strategy for prevention of cancer initiation and/or relapse.
`
`Epigenetic regulation of gene expression
`
`The transitory up- and down-regulation of gene transcription
`is directed mainly by the ephemeral activities of transcription
`factors. However, more permanent patterns of gene regula-
`tion are established by somatically heritable changes in
`the epigenome. Basically, two main structures may be
`altered by epigenetic modifications; cytosine bases of
`DNA, which can be either methylated or unmethylated, and
`histones, which attain various covalent modifications, such
`as methylation and acetylation, which are critical for the
`regulation of gene transcription. Together, these modifications
`result in a complex series of molecular events that cause
`re-modelling of the chromatin configuration and render
`CELGENE 2087
`genes either active or silent.
`APOTEX v. CELGENE
`IPR2023-00512
`
`

`

`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`390
`
`KIRSTEN GRØNBÆK
`
`ET AL.
`
`MiniReview
`
`Fig. 1. Epigenetic regulation of gene transcription by covalent modifications of DNA and histones. (A) Regulation of gene transcription
`by DNA methylation. Gene silencing by methylation of cytosine bases is catalysed by DNA methyltransferase (DNMT). DNA
`methyltransferases are therapeutic targets of 5-azacytidine, 5-aza-2′-deoxycytidine and zebularine which can reactivate gene transcription.
`(B) Regulation of gene transcription by modification of the histone tails. Histone acetylation is mediated by histone acetyltransferases (HATs)
`and leads to an open chromatin structure at transcriptionally active promotors. Histone deacetylation is mediated by histone deacetylases
`(HDACs) and causes dense condensation of the nucleosomes at transcriptionally inactive promotors. Several histone deacetylase inhibitors
`are listed. Histone lysine methylation is catalysed by histone lysine methyltransferases (KMTs). Methyl groups can be removed by histone
`lysine demethylases (KDMs). Methylation of H3K27 and H3K9 is associated with gene silencing; H3K4 and H3K36 methylation is
`associated transcriptional activation.
`
`Eukaryotic chromatin contains DNA and histones that
`are organized in the nucleosome, which consists of a histone
`octamer (two H2A/H2B dimers and an H3/H4 tetramer),
`around which ~146 bp of DNA is wrapped. Most chromatin
`exists as transcriptionally ineffectual
`heterochromatin
`, in
`which the nucleosomes are densely packaged to form a
`‘closed chromatin structure’. Conversely, an ‘open chromatin
`structure’ is characteristic of transcriptionally competent
`euchromatin
`, which has widely spaced nucleosomes and is
`accessible to the transcriptional machinery. Accordingly,
`euchromatic chromatin is flexible to meet the requests for
`individual proteins under particular cellular circumstances.
`
`DNA methylation.
`5-Methylcytosine is generated when a methyl group from
`the universal methyl donor, S-adenosyl-L-methionine (SAM/
`Ado-Met), is added to the fifth carbon atom in the cytosine
`pyrimidine ring [1]. This process is catalysed by the enzymatic
`activity of DNA methyltransferases, which reacts only with
`cytosine bases of cytidine-phosphate-guanosine (CpG) dinu-
`cleotides in mammalian DNA. During DNA replication, the
`methylation pattern from the parental DNA strand is
`copied onto the newly synthesized DNA strand by the
`maintenance methyltransferase, DNMT 1. In embryonic
`stem (ES) cells and cancer cells, methylation of previously
`
`unmethylated DNA (so-called
`de novo methylation) is catalysed
`by the enzymes DNMT 3a or DNMT 3b [2]. The mecha-
`
`nisms that initiate methylation of mammalian DNA are
`not yet clear, but recent discoveries suggest a link between
`histone modifications in ES cells and the DNA methylation
`patterns of mature cells (see below).
`To inactivate transcription, methylation must occur at
`clusters of CpG sites (‘CpG islands’) in the promoter
`regions of genes. The current definition of a CpG island is a
`>0.5 kb long stretch of DNA with a G + C content greater
`than 55% and an observed CpG/expected CpG ratio greater
`than 0.65 [3]. Approximately, half of all genes harbour CpG
`islands in their promoters. The remainder of the human
`genome is mostly depleted of CpGs because of the spontane-
`ous hydrolytic deamination of 5-methylcytosine to thymine
`(T). During evolution, mutations at CpG sites have caused a
`global reduction in the number of CpGs in the genome, and
`CpG sites in the coding regions of genes remain important
`mutational hot spots in human disease [4].
`During normal embryonic development, cytosine methyl-
`ation is crucial for establishing tissue-specific gene expression,
`for silencing imprinted genes and for X-chromosome inacti-
`vation. Methylation also protects against the transcription
`of parasitic elements that have become integrated in the
`genome [5].
`
`Histone modifications.
`The linkage between methylated cytosines and histone
`modifications has been described in some detail. Methylated
`
`© 2008 The Authors
`Journal compilation
`
` © 2008 Nordic Pharmacological Society.
`
`Basic & Clinical Pharmacology & Toxicology,
`
`103
`
`, 389–396
`
`

`

`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`MiniReview
`
`EPIGENETIC TARGETS FOR CANCER THERAPY
`
`391
`
`CpG islands attract a group of repressive proteins, the
`so-called methyl-CpG-binding-domain proteins, in a complex
`with histone deacetylases, which remove acetyl groups from
`lysine residues in the histone tails [6,7]. Deacetylated lysines
`are positively charged and react strongly with the negatively
`charged DNA. This leads to dense condensation of the
`nucleosomes at transcriptionally inactive promoters. Con-
`versely, acetylation of the lysines neutralizes this charge,
`which generates the open chromatin structure at transcrip-
`tionally active promoters. Histone acetylation is mediated
`by histone acetyl transferases, which form a ‘transcription
`activator complex’ with transcription factors and co-activator
`proteins [8].
`In addition to acetylation, a number of different covalent
`histone modifications have been identified that mark the
`transcriptional status of chromatin. Histone lysines may be
`unmethylated, or mono-, di- or trimethylated (-me1, -me2,
`-me3). The level of methylation at particular residues is
`important for the interaction with cofactors and is therefore
`essential for regulation of transcriptional activity. Repressive
`marks include di- and trimethylation of lysine 9 on histone
`H3 (H3K9me2/me3), trimethylation of lysine 27 on histone
`H3 (H3K27me3) and trimethylation of lysine 20 on histone
`H4 (H4K20me3). Conversely, activation is associated with
`histone H3 di- and trimethylation of lysine 4 (H3K4me2/
`me3) and trimethylation of lysines 36 and 79 (H3K36me3,
`H3K79me3) [9–12] (fig. 1). Methyl groups are provided to
`the histones by SAM/Ado-Met and the methylation reaction
`is catalysed by histone lysine methyltransferases [13]. Methyl
`groups can also be removed from histone lysine residues by
`the action of histone lysine demethylases [12,14].
`
`Polycomb-mediated gene repression in cancer.
`Recent investigations have generated the hypothesis that the
`epigenetic signatures of cancer cells may be predetermined
`already at the stem cell level and suggest the existence of
`early progenitor cells with cancer stem cell features. In ES cells,
`gene transcription is regulated by the so-called polycomb
`group (PcG) complexes, PRC1, PRC2/3 and PRC4 [15].
`These protein complexes can reversibly repress genes that
`encode transcription factors necessary for differentiation.
`The PRC2 polycomb protein EZH2 is a histone lysine
`methyltransferase that catalyses the formation of H3K27me3,
`which is associated with gene silencing. Interestingly, EZH2
`is an oncogene that is up-regulated in several types of cancer
`and associated with aggressive subtypes of melanoma,
`endometrial and breast cancer [16], and is a predictor of
`poor prognosis in metastatic prostate cancer [17] and mantle
`cell lymphoma [18].
`It was recently shown that a significant overlap exists
`between gene promoters that are occupied by PcG complexes
`in ES cells and gene promoters that become methylated in
`cancer [19–21]. In contrast to the reversible gene silencing
`by histone H3K27me3 in ES cells, DNA hypermethylation
`in cancer cells confers deep transcriptional repression that
`can only be reversed by the use of DNA methyltransferase
`inhibitors. It has been suggested that an epigenetic switch
`
`from PcG to DNA methylation-mediated gene silencing
`may be involved in the transformation of normal stem cells
`to cancer stem cells. Although it has been proposed that a
`direct link exists between H3K27me3 and DNA methylation
`[22], this has not been verified in other studies [23]. Further-
`more, new observations indicate that histone H3K27me3
`may be an independent mechanism of gene silencing in
`cancer [24].
`
`Current knowledge and applications of epigenetic therapy
`
`DNA methyltransferase inhibitors.
`Some 30 years ago, it was discovered that analogues of
`′
`cytidine, 5-aza-cytidine (5-aza-CR) and 5-aza-2
`-deoxycytidine
`(5-aza-CdR), could induce differentiation of cultured mouse
`embryo cells to muscle cells [25]. It was demonstrated that
`genes involved in muscular differentiation were inactivated
`by DNA methylation in the mouse embryo cells, and that
`these genes could be reactivated by 5-aza-CR and 5-aza-CdR.
`Within recent years, these drugs have been shown to be
`potent alternatives to conventional chemotherapy, particularly
`in the treatment of myelodysplastic syndrome and acute
`myeloid leukaemia (table 1). 5-Aza-CR and 5-aza-CdR differ
`from their normal counterparts (cytidine and deoxycytidine,
`respectively) by carrying a nitrogen atom in position 5 of the
`pyrimidine ring. Inside the cells, these drugs are converted
`into the corresponding nucleotides. 5-Aza-CdR is incorpo-
`rated into DNA, while 5-aza-CR is preferentially integrated
`into RNA but also into DNA. Both drugs covalently bind
`DNA methyltransferase, which then becomes trapped and
`unable to catalyse the methylation of newly synthesized
`DNA strands [26]. Over subsequent cell divisions, methylation
`is gradually lost leading to cellular differentiation over time.
`Accordingly, the clinical responses are typically detected
`after three to four series of 5-aza-CR treatment [27], and it
`has been shown that DNA methylation gradually disappears
`in the tumour cells from responders [28]. Therefore, in
`contrast to conventional cancer chemotherapy, 5-aza-CR
`and 5-aza-CdR should not be administered at the maximum
`tolerable dose, which will kill the cells, but at smaller doses
`over a longer period of time in order to induce cellular
`differentiation [25,29]. In recent clinical trials, this schedule
`of drug administration has been used in advanced myelod-
`ysplastic syndrome and acute myeloid leukaemia [27,30]. It
`was recently demonstrated that treatment of myelodysplas-
`tic syndrome with 5-aza-CR as compared to conventional
`care regimens doubled the 2-year survival rate, (51% versus
`TM
`)
`26%, respectively; P = 0.0001) [31]. Both 5-aza-CR (Vidaza
`TM
`) have been approved by the US
`and 5-aza-CdR (Dacogen
`Food and Drug Administration for use in the treatment of
`myelodysplastic syndrome [32].
`
`Histone deacetylase inhibitors.
`Histone deacetylase inhibitors are a class of molecules with
`diverse effects on the regulation of cell growth. First, they
`inhibit the deacetylation of histones and thus are directly
`involved in the regulation of gene transcription. Second,
`
`© 2008 The Authors
`Journal compilation
`
` © 2008 Nordic Pharmacological Society.
`
`Basic & Clinical Pharmacology & Toxicology,
`
`103
`
`, 389–396
`
`

`

`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`392
`
`KIRSTEN GRØNBÆK
`
`ET AL.
`
`MiniReview
`
`[60]
`
`[56]
`
`[39]
`
`[57]
`
`[33]
`
`[33]
`
`[36]
`
`[33]
`
`[33]
`
`[33]
`
`[33]
`
`[33]
`
`[53]
`[51]
`
`[48]
`
`[47]
`
`[28,30]
`
`Phase III, maintenance
`
`AML/MDS many cancers in
`
`Yes
`
`therapy
`and combination
`
`combination therapy
`
`[31,42]
`
`Phase III, maintenance
`
`AML/MDS many cancers in
`
`Reference
`
`Clinical trials*
`
`approvedMain clinical* application
`US FDA
`
`Phase I-II
`Pre-clinical
`
`Pre-clinical
`
`Pre-clinical
`therapy
`and combination
`
`combination therapy
`
`–
`
`–
`
`–
`
`Pre-clinical
`
`Pre-clinical
`
`Pre-clinical
`
`Pre-clinical
`
`–
`
`–
`
`–
`
`–
`
`cancers
`
`Phase I-II
`
`Lymphoma and other
`
`Phase I-II
`
`Multiple myeloma and other
`
`cancers
`
`Phase III
`
`Cutaneous T-cell lymphoma
`
`Phase I
`
`EBV-associated lymphoma
`
`Phase I/II
`
`AML/MDS in combination
`
`therapy
`
`Phase II
`
`T-cell lymphoma
`
`therapy
`
`Phase I/II
`
`AML/MDS in combination
`
`in combination therapy
`
`Phase II
`
`AML/MDS and lymphoma
`
`Advanced solid tumours
`
`–
`
`–
`
`–
`
`–
`
`–
`
`Yes
`
`–
`
`Yes
`
`–
`
`–
`
`–
`
`–
`
`–
`
`–
`
`–
`
`*This and further information at
`AML, acute myeloid leukaemia; MDS, myelodysplastic syndrome.
`DNMT, DNA methyltransferase; HDAC, histone deacetylase; HDACI, HDAC inhibitors; SAH, S-adenosylhomocysteine; SAM, S-adenosyl-L-methionine; KDM, histone lysine demethylase;
`
`ClinicalTrials.gov
`
`.
`
`SIRT1 inhibition may activate DNA-
`
`Niacinamide
`
`HDAC inhibition:
`
`methylated genes
`
`(Splitomicin)
`
`class III
`
`↓
`
`EZH2 depletion/H3K27me3
`
`Hydroxamate HDACI
`
`Hydroxamate HDACI
`
`Hydroxamate HDACI
`
`LBH589
`
`(Belinostat)
`
`PXD101
`
`(SAHA)
`Vorinostat
`
`Short chain fatty acid HDACI
`
`Butyric acid
`
`class I, II and IV
`HDAC inhibition:
`
`class I and II
`
`Short chain fatty acid HDACI
`
`Valproic acid
`
`HDAC inhibition:
`
`Cyclic tetrapeptide HDACI
`
`(Romidepsin)
`
`Depsipeptide
`
`Benzamide analogues HDACI
`
`MS275
`
`Benzamide analogues HDACI
`
`MGCD0103
`
`HDAC inhibition:
`
`class I
`
`Antisense to DNMT mRNA
`Binds to the DNMT catalytic site
`
`when incorporated in RNA (and DNA)
`
`For oral use – stable in water. Binds DNMT
`
`incorporated in DNA
`deaminase. Binds DNMT when
`
`MG98
`RG108
`
`Zebularine
`
`Oligonucleotide: no degradation by cytidine
`
`S110: 5-aza-CpG
`
`Dacogen
`
`Binds DNMT when incorporated in DNA
`
`deoxycytidine
`
`5-Aza-2-
`
`Yes
`
`Vidaza
`
`Binds DNMT when incorporated in RNA
`
`5-Azacytidine
`
`DNMT inhibitors
`
`Trade name
`
`Mechanism of action
`
`Drug
`
`Main function
`
`(and DNA)
`
`Table 1.
`
`Epigenetic drugs: functions and current applications.
`
`–
`
`–
`
`–
`
`–
`
`–
`
`–
`
`–
`
`Zolinza
`
`–
`
`others
`and
`
`Depakene
`
`–
`
`–
`
`–
`
`–
`
`–
`
`–
`
`–
`
`,
`
`↑
`
` H3K4me2
`
`⇒
`
`Inhibitor of LSD1
`
`↓
`
`H3K9me2
`
`Inhibits the donation of methyl groups from
`
`
`
`↓
`
` H3K27me3
`
`SAM
`
`⇒
`
`analogues
`Polyamine
`
`A (Dznep)
`
`KDM inhibitor
`
`inhibitor
`
`Deazane-planocin
`
`SAH hydrolase
`
`Journal compilation
`
`© 2008 The Authors
`
` © 2008 Nordic Pharmacological Society.
`
`Basic & Clinical Pharmacology & Toxicology,
`
`103
`
`, 389–396
`
`

`

`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`MiniReview
`
`EPIGENETIC TARGETS FOR CANCER THERAPY
`
`393
`
`they indirectly modulate gene transcription by inhibiting the
`deacetylation of proteins, including transcription factors.
`Third, they may play a significant role in controlling cell
`division and differentiation in cancer, because they are capable
`of inducing G1 cell cycle arrest via induction of p21 in
`tumours that have defective p53 function [33]. Other important
`activities of histone deacetylase inhibitors relate to their
`ability to inhibit proliferation at the G2 cell cycle checkpoint
`and to up-regulate pro-apoptotic molecules.
`At present, 18 different histone deacetylases are known,
`which have been assigned to four classes according to their
`homologues in yeast. Class I and II histone deacetylases
`mainly reside in the nucleus of cells and are involved in the
`deacetylation of histones and other proteins (e.g. p53) [33].
`Class III histone deacetylases (sirtuins) constitute a spe-
`cific class of histone deacetylases that are involved in the
`regulation of gene transcription in response to changes in
`cellular stress, including the levels of reactive oxygen species
`[34]. The specificity of histone deacetylase inhibitors to these
`different histone deacetylase classes varies; some are highly
`specific to one or two classes (e.g. depsipeptide and MS-
`TM
`) and PXD101]
`275), while others [e.g. vorinostat (Zolinza
`cover the whole spectrum of histone deacetylases except
`for class III, which are inhibited by a specific group of
`drugs (e.g. niacinamide). In general, histone deacetylase
`inhibitors can be given with minimal toxicity and some
`drugs are available for oral administration, including
`vorinostat [33,35].
`A large number of histone deacetylase inhibitors of different
`chemical classes are currently under investigation in clinical
`trials, alone or in combination with other anticancer
`therapeutics (table 1). While in most cancers, histone deacety-
`lase inhibitors are most efficient in combination therapy,
`cutaneous T-cell lymphoma was successfully treated by
`vorinostat alone [36,37]. This drug has recently been
`approved by the US Food and Drug Administration for use
`in cutaneous T-cell lymphoma.
`
`Combination epigenetic therapy.
` studies that histone
`in vitro
`It is well established from
`deacetylase class I/II inhibitors given as single drugs do not
`turn on genes that are silenced by promoter hypermethyla-
`in vitro
`tion. However,
` studies of the combination of histone
`deacetylase
`inhibitors and DNA methyltransferase
`inhibitors generally show strong synergy in up-regulating
`hypermethylated genes [26], and clinical studies using
`combination therapy of these drugs induced complete or
`partial responses in acute myeloid leukaemia patients [38].
`In contrast to histone deacetylase class I/II inhibitors,
`histone deacetylase class III inhibitors given as single drugs
`in vitro
`may turn on methylated genes
` [39]; however, it
`remains to be shown if this is also the case
`in vivo.
`in vitro
` In
`studies of acute promyelocytic leukaemia, it was shown that
`α
`the pathognomonic fusion proteins PML-RAR
` and
`α
`PLZF-RAR
` recruit histone deacetylases to initiate their
`tumorigenic effect [40], and the combined use of histone
`deacetylase inhibitors and all-trans retinoic acid induced
`
`differentiation of acute myeloid leukaemia blasts and hae-
`matological response in refractory acute myeloid leukaemia
`patients [41].
`In vitro
` and clinical studies have also shown promising
`effects of combinations with other anticancer agents, such
`as chemotherapeutics. Conventional chemotherapy may be
`
`used
`before DNA methyltransferase inhibitors and histone
`deacetylase inhibitors to debulk solid tumours, which will
`allow subsequent induction of stem cell differentiation by
`the epigenetic therapy. Accordingly, it is investigated in
`clinical trials whether long-term, low-dose DNA methyltrans-
`ferase inhibitor administration after chemotherapy may
`efficiently maintain patients in remission [42]. Alternatively,
`after
`conventional chemotherapy administered
`treatment with
`DNA methyltransferase inhibitors and histone deacetylase
`inhibitors may act via tumour suppressor proteins that are
`reactivated by the epigenetic therapy. Preclinical studies
`have also shown that histone deacetylase inhibitors enhance
`sensitivity to radiotherapy, and clinical trials are ongoing.
`Furthermore, in tumour cell lines, synergism has been
`shown with many of the novel targeted therapies, such as
`TM
`)], kinase
`proteasome inhibitors [e.g. bortezumib (Velcade
`inhibitors and death receptor agonists [43–46].
`
`New drugs for epigenetic therapy.
`Within recent years, a number of attempts have been made
`to identify or design new drugs directed towards epigenetic
`targets. One major focus is to circumvent the inherent limi-
`tations of 5-aza-CR and 5-aza-CdR, including the fact that
`they are unstable in aqueous solution, have transient effects
`and are inactivated by cytidine deaminase to 5-azauridine
`[47]. 5-aza-CR is being developed for oral administration,
`and zebularine, another cytidine analogue that lacks the
`amino group on C-4 of the pyrimidine ring, is stable in
`aqueous solution and can be administered orally [48].
`Unfortunately, this drug is degraded in the liver and, there-
`fore, further development will be required [49]. Short oligo-
`nucleotide derivatives of 5-aza-CdR
`[e.g. 5-aza-CpG
`(S110)], seem to be resistant to degradation by cytidine
`deaminase. However, the stability of 5-aza-CpG in aqueous
`solution is not improved relative to the single nucleotide
`forms [47]. In general, DNA methylation seems to be re-
`established once DNA methyltransferase inhibitors are
`withdrawn; however, this may be overcome by using long-
`term maintenance therapy [50].
`Other approaches to therapeutically down-regulate DNA
`methylation include the DNA methyltransferase 1 antisense
`compound MG98 as well as the small compound RG108,
`which binds to and inhibits the catalytic site of DNA meth-
`yltransferase 1. As RG108 is not incorporated into DNA,
`it is supposed to be less toxic to the cells than the aza-
`nucleotides; however, this compound has not yet entered
`clinical trials [51]. The antisense molecule MG98 showed
`promising results in a phase I clinical trial against renal cell
`carcinoma, but this result could not be confirmed in phase
`II [52]. Further investigations of MG98 are planned in
`combination with interferon [53].
`
`© 2008 The Authors
`Journal compilation
`
` © 2008 Nordic Pharmacological Society.
`
`Basic & Clinical Pharmacology & Toxicology,
`
`103
`
`, 389–396
`
`

`

`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`394
`
`KIRSTEN GRØNBÆK
`
`ET AL.
`
`MiniReview
`
`Toxicity of epigenetic therapy.
`Because epigenetic therapy is not specifically directed
`against cancer cells, a number of concerns have been raised
`regarding the toxicity of these drugs when administered to
`patients. Obviously, in theory, genes and parasitic elements
`that are normally silenced may be inappropriately expressed,
`and oncogenes may be turned on that could aggravate
`the malignant phenotype. These concerns are pertinent;
`however, because the growth rate of cancer cells is typically
`higher than that of normal cells and because DNA methyl-
`transferase inhibitors are only active during cell division,
`they are more likely to reactivate transcription in cancer
`cells. Furthermore, in the majority of genes in normal cells,
`the bulk of DNA methylation is not at gene promoters, but
`within the coding regions of genes where it has no influence
`on gene transcription, and gene expression studies show
`that only a limited number of genes are up-regulated by
`DNA methyltransferase inhibitors in normal fibroblasts
`[54]. However, because global genomic hypomethylation
`may cause genomic instability, leading to chromosomal
`breaks, there is still a reason for concern. Because most
`experience with DNA methyltransferase inhibitors has been
`obtained during treatment of advanced cancers, little is
`known about their long-term side effects, including their
`carcinogenic potential.
`The long-term effects of histone deacetylase inhibitors
`also are not well described. However, valproic acid, a potent
`histone deacetylase inhibitor which is given for years to
`patients suffering from epilepsy, showed no influence on the
`cancer incidence in a recent large population-based case–
`control study [55].
`
`Potential new targets for epigenetic therapy.
`Recently, it has become evident that histone methylation
`is also a reversible epigenetic modification that is essential
`for regulating gene transcription during normal cellular
`differentiation and in disease states. A considerable
`number of histone methyltransferases [13] and demethylases
`[12,14] have been discovered, and many of these enzymes
`are reported to be deregulated in cancer by deletions,
`amplifications and translocations, and have been suggested
`as tumour suppressors and oncogenes, respectively.
`Accordingly, these enzymes may be potential new targets for
`cancer therapy. For example, inhibition of the oncogenic
`histone lysine methyltransferase EZH2 may inhibit the for-
`mation of histone H3K27me3, which is thought to play a role
`in gene silencing in cancer [24]. The S-adenosylhomocysteine
`hydrolase inhibitor 3-deazanaplanocin A (DZNep) was
`recently described as a specific inhibitor of EZH2-mediated
`in vitro
`H3K27me3
`[56]; however, it seems unlikely that
`this drug will not have side effects on other cellular
`methylation reactions. Recent studies have also suggested
`that histone deacetylase inhibitors may deplete EZH2 and
`down-regulate H3K27me3 [24,57]; however, this also most
`likely occurs via an indirect mechanism, and hopefully, new
`and more specific EZH2 inhibitors may be developed for
`clinical use.
`
`Two main types of histone demethylases have been
`discovered: the amino oxidase LSD1 (histone lysine demeth-
`ylase 1) and the Jumonji (JmjC) domain proteins [12,14].
`While LSD1 is capable of demethylating mono- and dimeth-
`ylated lysines, the Jumonji enzymes also target trimethylated
`lysines. LSD1 demethylates both active and repressive
`methylation marks (H3K4me2 and H3K9me2) [58] as well
`as non-histone substrates (e.g. p53) [59]; however, small
`molecule inhibitors of LSD1 have been reported to up-
`regulate silenced tumour suppressors and may potentially
`be effective anticancer agents [60]. Some of the JmjC-domain
`histone enzymes demethylate specific methylation marks
`that may be associated with inappropriate gene transcription
`in cancer, such as H3K4me3 (demethylated by JARID1),
`H3K9me3 (demethylated by JMJD2) and H3K27me3
`(demethylated by UTX/JMJD3) [12]. Accordingly, these
`histone lysine demethylases may also be attractive new targets
`for therapy. However, at least in some cellular contexts, histone
`methyltransferases and demethylases interact in complex in
`transcriptional regulation at the same gene promoters [61],
`and a pharmacological targeting of individual molecules
`might be difficult. Nevertheless, the development of drugs
`that target the Jumonji demethylases is a high priority.
`
`A potential role of epigenetic therapy in cancer prevention and
`maintenance.
`Protection against the development and recurrence of
`cancer is a challenging task and new achievements in this
`field would have great socio-economical impact. If epigenetic
`alterations are involved in the abnormal clonal expansion of
`pre-cancerous cells, epigenetic therapy may potentially both
`reverse the early stages of cancer and maintain patients in
`remission. The most straightforward role for epigenetic
`therapy may be the latter, to ensure maintenance of remission
`after chemotherapy or surgery (e.g. used as so-called adjuvant
`therapy). If remaining cells with epigenetic changes constitute
`a source for cancer relapse, one approach to maintain remis-
`sion would be continuous eradication of these cancer pro-
`genitors. This possibility is already being investigated in
`several clinical trials, including, for example, the ongoing
`Medical Research Council acute myeloid leukaemia16
`(MRC-AML 16) clinical trial. In this study, one arm of the
`≥
`60 years) with
`trial investigates whether older patients (
`acute myeloid leukaemia or advanced myelodysplastic
`syndrome will benefit from maintenance treatment with 5-
`aza-CR after remission has been obtained by conventional
`chemotherapy (www.aml16.bham.ac.uk).
`At the present time, the available drugs that target the
`epigenome may be too toxic for use in cancer prevention.
`However, with the rapid development of more specific drugs
`and with the discovery of new targets, new and milder forms
`of epigenetic therapy may be used in patients with cancer
`predisposition. Obvious candidates for such treatment are
`patients with recognized precancerous lesions such as patients
`in situ
`with carcinoma
` lesions and patients with chronic
`infections that predispose to cancer (e.g. HIV, HHV-6, HPV,
`Hepatitis B and Hepatitis C infection). Other groups of
`
`Journal compilation
`
`© 2008 The Authors
`
` © 2008 Nordic Pharmacological Society.
`
`Basic & Clinical Pharmacology & Toxicology,
`
`
`
`103, 389–396
`
`

`

`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`MiniReview
`
`EPIGENETIC TARGETS FOR CANCER THERAPY
`
`395
`
`individuals that may benefit from a cancer-preventive
`approach are members of cancer families. Although the
`primary inherited component in such families typically is a
`point mutation in a tumour suppressor gene or proto-oncogene
`[62], additional molecular hits are required for a cancer to
`fully develop, and at least some of these additional hits may
`be epigenetic alterations.
`It will be interesting to learn from ongoing and future
`studies whether we can prevent cancer development and/or
`relapse by targeting the epigenome. Hopefully, in the years
`to come, more specific and non-toxic therapies with convenient
`dose schedules, such as oral administration, will provide
`new tools for cancer prevention and maintenance.
`
`References
`
`1 Chiang PK, Gordon RK, Tal J, Zeng GC, Doc