This page has been archived and is no longer updated
CpG island methylator phenotype in cancer
Author: Jean-Pierre Issa
Keywords
Keywords for this Article
Add keywords to your Content
Save
|
Cancel
Share
|
Cancel
Revoke
|
Cancel
Rate & Certify
Rate Me...
Rate Me
!
Comment
Save
|
Cancel
Flag Inappropriate
The Content is
Objectionable
Explicit
Offensive
Inaccurate
Comment
Flag Content
|
Cancel
Delete Content
Reason
Delete
|
Cancel
Close
Full Screen
"depending on which parent that allele was inherited from) and X-chromosome inactiva- tion. Genetic anomalies in DNA methylation (or associated mechanisms) leads to various developmental diseases in humans 5 . Methylation in development and cancer Research into DNA methylation has been progressing at a furious pace, despite uncer- tainty about its origin and physiological function. This much is known: most CpG sites have been lost from mammalian genomes during evolution, but about 1% of human DNA consists of short areas where CpG sites have escaped depletion 1,2 .Most of the remaining CpG sites are normally methy- lated in adult cells. About half of all genes have a CpG island in their promoter region, and this gene configuration is what has recently attracted the most attention. Most promoter CpG islands are normally unmethylated, regardless of the expression state of the associated gene. However, in silenced areas, such as the inactive X-chrom- osome in females and the silenced allele of imprinted genes, promoter-associated CpG islands are generally methylated, and this methylation is essential for maintaining the silenced state. Mechanisms regulating the establishment of methylation remain poorly understood, but the consequences of CpG island methylation are becoming increas- ingly clear. Methylation triggers the binding of methylated DNA-specific binding proteins to CpG sites, attracting histone-modifying enzymes that, in turn, focally establish a silenced chromatin state (FIG. 1). Consistent with a resurgence of interest in the idea that cancer is a disease of faulty development, there has been a revival of interest in the epigenetic processes involved in neoplastic development and progres- sion 5,6 .Epigenetic information, after all, is essential for development, and it is clear that cancer is ultimately a disease of aber- rant gene expression. An epigenetic contri- bution to cancer is no longer in doubt. Early experiments showing that the reprogram- ming of epigenetic information after fertil- ization reverses the malignant phenotype 7 have been confirmed using modern tech- nology such as nuclear transfer 8,9 .The potential reversibility of epigenetic changes through pharmacological manipulation makes this area acutely important in cancer management 5,10 , and a specific DNA methy- lation inhibitor (5-azacytidine) has now been approved for use as an anti-cancer agent in the USA 11 . DNA methylation is central to the aber- rant epigenetics of cancer. As described in extensive, recent reviews, cancer cells often have both a loss of global methylation and a gain of methylation at the promoters of selected CpG islands, resulting in the silencing of hundreds of genes per cancer cell, including tumour-suppressor genes 6,12 . Indeed, DNA methylation is proving to be a useful marker of disease risk, activity and prognosis in various malignancies 13,14 . However, a careful evaluation of the causes of aberrant methylation in cancer has been somewhat overlooked. There is a modest increase in DNA-methyltransferase activity in cancer 15 , and immunohistochemistry studies have confirmed overexpression of DNA (cytosine-5)-methyltransferase 1 (DNMT1) protein in primary cancer cells 16 , albeit with significant heterogeneity. Some of this overexpression might be related to the regulation of DNA-methyl- transferases by the cell cycle 17 , and RNA levels of DNMT1 or DNMT3 have largely not proven useful in explaining aberrant patterns of DNA methylation in cancer 18 . Abstract | DNA hypermethylation in CpG-rich promoters is now recognized as a common feature of human neoplasia. However, the pathophysiology of hyper- methylation (why, when, where) remains obscure. Cancers can be classified according to their degree of methylation, and those cancers with high degrees of methylation (the CpG island methylator phenotype, or CIMP) represent a clinically and aetiologically distinct group that is characterized by ?epigenetic instability?. Furthermore, CIMP-associated cancers seem to have a distinct epidemiology, a distinct histology, distinct precursor lesions and distinct molecular features. DNA methylation is a biochemical modifica- tion that, in human cells, primarily affects cytosines when they are part of the symmetri- cal dinucleotide CpG 1,2 .Cytosine methylation has long been a challenging scientific puzzle. Although it is nearly ubiquitous among multi- cellular organisms, two important biological models ? Saccharomyces cerevisiae and adult Drosophila melanogaster ? have undetectable levels of DNA methylation. In mammals, DNA methylation is essential for normal development 3 ,but its evolutionary raison d?�tre remains controversial. A commonly held hypothesis is that DNA methylation originally evolved to silence repetitive ele- ments 4 , and that this silencing property has also been put to use in other situations where transcriptional silencing is required, such as imprinting (a process whereby one of the two alleles of a gene are permanently inactivated, CpG island methylator phenotype in cancer Jean-Pierre Issa OPINION PERSPECTIVES 988 | DECEMBER 2004 | VOLUME 4 www.nature.com/reviews/cancer cell line 26 .Subsequent studies have added several loci to the list of genes that are pref- erentially hypermethylated in sporadic MSI- positive cases, including HPP1 (hyperplastic polyposis gene 1, also known as TMEFF2) 27 and CDKN2A,which encodes ARF 28 and other proteins. Interestingly, these genes were not as frequently methylated in colon tumours from individuals with inherited MSI caused by germline mutations in mismatch repair genes 29 ,demonstrating that mismatch repair defects per se do not accelerate aberrant methylation. A more extensive study of this phenome- non was made possible by the development of methods to identify CpG islands that are differentially methylated in specific situations 30 .Using 33 such loci to profile the methylation status of normal colonic mucosa compared with colorectal cancer, loci could be divided into two groups 31 .Most loci (26 out of 33 (79%)) are methylated in normal colonic mucosa as a function of age and undergo more extensive methylation in cancer. Therefore, the dominant ?cause? of methylation in cancer is actually methylation in normal tissues that arises as a function of age. A discussion of age- related methylation is beyond the scope of this article, but it is important to note that this find- ing has now been reproduced in several studies and is probably a universal phenomenon in epithelial malignancies 32,33 .However,when age-related methylation was filtered out, it became apparent that methylation of the remaining seven loci, which tended to be lim- ited to neoplastic tissues, was clustered in a spe- cific subset of cases 31 .This phenomenon, descriptively termed CIMP, defined a group of cancers with a 3?5-fold elevated frequency of aberrant gene methylation. Confirmation that this group is distinct came when separate genes were examined. Methylation of each of the genes encoding INK4A, MLH1 and THBS1 was almost exclusively limited to the CIMP- positive subset. In fact, 70?80% of sporadic MSI-positive colon cancers could be attributed to CIMP and associated MLH1 methylation. These data divided sporadic colorectal cancers into four distinct groups: CIMP + MSI + , CIMP + MSI ? , CIMP ? MSI + and CIMP ? MSI ? . The distinct nature of these four groups was confirmed by widely divergent mutation rates in the tumour suppressor gene TP53 and the oncogenes KRAS and BRAF 34,35 .A stark conclusion of this work was that, at the molec- ular level, sporadic colon cancer is in fact a collection of (at least) four different diseases. The findings that led to the definition of CIMP have now been reproduced in several (although not all) studies. Concordant methylation of multiple genes and/or PERSPECTIVES NATURE REVIEWS | CANCER VOLUME 4 | DECEMBER 2004 | 989 Discovery of CIMP Aberrant CpG island methylation events in cancer were first described almost two decades ago 23 ,but most of the early studies in the field focused on the examination of iso- lated genes. A few years ago, as more and more genes affected by the process were iden- tified, puzzling data emerged indicating that in some tumours, groups of genes had con- sistently increased methylation. This was demonstrated statistically by showing that methylation of two separate genes was corre- lated in a given tumour type. This was most apparent in colon cancer, where it was shown that sporadic tumours with microsatellite instability (MSI) had increased frequencies of promoter methylation affect- ing multiple genes such as CDKN2A,which encodes the protein INK4A, and THBS1 (thrombosponsin 1) 24 .Remarkably, parallel studies established that these MSI-positive tumours also had hypermethylation and silencing of the mismatch repair gene MLH1 25 .The suggestion that hypermethyla- tion leads to MSI (through the silencing of MLH1) was supported by the observation that the inhibition of methylation reversed the mismatch repair defect in a colon cancer This remains an area of research interest. Associations between methylation and exposure to carcinogens such as viruses 19 , smoking 20 and radiation 21 have been observed but have not led to mechanistic insights into the process. The central conun- drum in the field remains unresolved: is aberrant DNA methylation a rare, random process that is selected for in neoplastic cells, or does it result from specific defects in the methylation process? This question is similar to one that has divided the field of genetic changes in cancer for years 22 :do mutations arise stochastically and are selected for, or does cancer development require processes that accelerate the rate at which genetic defects are acquired? The dis- covery of multiple mutator phenotypes in cancer, mostly as a result of defects in DNA repair, has resolved the latter question, but the issue of epigenetic changes (and DNA methylation in particular) remains largely controversial. Multiple concordant methyla- tion events ? the CpG island methylator phenotype (CIMP) ? in a subset of colon cancers provides evidence for a process akin to the mutator phenotype that affects CpG island methylation instead (BOX 1). ATCGAATGCTGCGGA TAGCTTACGACGCCT ATCGAATGCTGCGGA TAGCTTACGACGCCT ATCGAATGCTGCGGA TAGCTTACGACGCCT ATCGAATGCTGCGGA TAGCTTACGACGCCT ATCGAATGCTGCGGA TAGCTTACGACGCCT Unmethylated Hemi-methylated Methylated Restored methylation Silenced Epigenic reprogramming (embryogenesis) Silencing complex (histone deacetylation, histone H3K9 methylation, HP1 binding) DNA replication Maintenance methylation (DNMT1) De novo methylation (DNMT3A, DNMT3B) Figure 1 | DNA methylation and gene silencing. In early embryogenesis, DNA is largely devoid of methylation (top left). Post implantation, de novo methylation begins (red circles), mediated primarily by DNA (cytosine-5-)-methyltransferase-3? (DNMT3A) and -3? (DNMT3B) (top). When methylation affects CpG islands, methyl-binding proteins trigger a silencing cascade (activity illustrated by green stars) whereby histone H3K9 is sequentially deaceylated and then methylated, allowing heterochromatin protein 1 (HP1) to bind; eventually resulting in closed chromatin (bottom right). After DNA replication, newly synthesized DNA (in green) is unmethylated. However, DNMT1 rapidly scans DNA and deposits methyl groups on newly synthesized DNA, opposite methyl groups present on the old DNA strand. This results in faithful replication of methylation patterns (bottom left) and the maintenance of silencing. Adult patterns of methylation are erased by epigenetic reprogramming in early embryogenesis (top left). gene inactivation is a cause of cancer develop- ment. This issue cannot be stressed strongly enough. Early developments in the field were met with substantial skepticism, with many investigators arguing that DNA methylation simply results from neoplastic transforma- tion 48 .Indeed, it is likely that much aberrant DNA methylation is a by-product of the neo- plastic process, perhaps an extension of the age-related methylation process. Nevertheless, the presence of CIMP shows that all methyla- tion events are not equal, and that DNA methylation does not necessarily follow malig- nant transformation. Rather, the subset of tumours with intense methylation is proba- bly formed, in part, through changes in cell physiology resulting from epigenetic silencing. CIMP cancers ? unique diseases The molecular identification of CIMP led to a reappraisal of its clinical and pathological fea- tures, and, in many respects, established it as defining a unique subset of colorectal can- cers 31,34,36,38 . CIMP-positive cancers tend to occur in older patients, are overrepresented in proximal tumours of the colon, and occur more often in women. Genetically, CIMP pos- itive cases are also distinct, with a paucity of p53 mutations and a remarkably high rate of mutations in KRAS or BRAF 34,35 ,such that nearly every CIMP-positive tumour has evi- dence of activation of the RAS oncogenic pathway. CIMP-positive cases are also distinct pathologically. Initially, this uniqueness was thought to be related to the strong link between MSI and CIMP. However, CIMP is also associated with distinct features in cases without MSI 36 , and recent studies of stage IV colorectal cancer (in which MSI is rare) identi- fied unique histological features of CIMP- positive cases including poor differentiation and unusual gland architecture (REF. 36; L. Chirieac, personal communication). CIMP- positive cases also show unique clinical attrib- utes. MSI-positive cases tend to have a good prognosis 49 ,whereas CIMP + MSI ? cases have a particularly poor outcome (REF. 50;L.Shen, personal communication). Most remarkably, CIMP is central to recent findings that indicate an alternative precursor lesion to colorectal cancer, a concept that has profound clinical implications 51 .A hitherto overlooked hyper- plastic polyp might well be the precursor to serrated adenomas which, in turn, give rise to MSI-positive cancers. In fact, serrated adeno- mas seem to be a methylation-based disease, showing a high incidence of CIMP, as well as KRAS and BRAF mutations 35,52?54 (BOX 2). Although they are most apparent in colon cancer, the unique features of CIMP extend to other malignancies. CIMP-positive 990 | DECEMBER 2004 | VOLUME 4 www.nature.com/reviews/cancer PERSPECTIVES clustering reminiscent of CIMP has been confirmed in colorectal cancer 36?38 and has also been observed in glioblastomas 39 , gas- tric cancer 40,41 , liver cancer 19 , pancreatic can- cer 42 , oesophageal cancer 14 , ovarian cancer 43 , acute lymphocytic leukaemia 44 and acute myelogenous leukaemia 45 .A few studies, however, have not found evidence of con- cordant methylation in colon cancers or other diseases 18,46 .A common feature of these latter studies is the use of methylation specific PCR (MSP) ? a very sensitive, non-quantitative technique to measure methylation ? and/or the inclusion of all genes in the analyses without attempting to filter out age-related methylation. Indeed, a recent study in colorectal cancer 47 con- cluded that methylation is primarily a char- acteristic of ageing and that, using unse- lected genes, CIMP cannot be identified in this disease, exactly as predicted by the orig- inal observations 31 .In fact, re-examination of data in that paper does reveal clustering of methylation of specific genes (such as CDKN2A,encoding varients INK4A and ARF) in the subset of cases that has MLH1 methylation-based MSI, indicating that CIMP might be discernible with appropri- ate statistical analysis. Every research group that has quantitatively studied the methyla- tion status of the genes originally used to define CIMP has confirmed that there is concordant methylation in colon, gastric, pancreatic and ovarian cancer. The origin of DNA methylation in cancer CIMP has fundamental implications for our understanding of the origin of tumour- suppressor gene methylation in cancer. First and foremost, the fact that different cancers have significantly different rates of tumour- suppressor gene inactivation through DNA methylation-associated processes implies that methylation is not random. In turn, this means that aberrant DNA methylation has a cause that should be searched for and identi- fied. Teleologically, it also implies that this process is selected for during neoplastic trans- formation and, ultimately, it provides strong evidence that epigenetic tumour-suppressor Box 1 | Definition of the CpG island methylator phenotype An important hurdle in the field will be to achieve a consensus definition for the CpG island methylator phenotype (CIMP). This is no trivial issue, given the variety of methods available for studying DNA methylation, each of which might give a slightly different definition 70 . The choice of genes and the minimal number of genes examined is also essential. In all studies of CIMP so far (positive or negative), each group has used different methods and different genes, which can only contribute to the confusion. Moreover, the choice of genes is also tissue-type dependent, and a definition for colon cancer might not be applicable to other cancers. In order to move towards a consensus, two issues are crucial: first, a quantitative method is needed for studying methylation. CIMP seems to affect both the frequency and the extent of methylation 45 , and the latter is missed if one uses a sensitive but non-quantitative method. Second, in order to select genes to define the phenotype, genes with high levels of methylation in normal tissues must be avoided. This is fairly straightforward in the colon, but is problematic in other diseases such as breast cancer, in which the relevant normal cells (breast epithelium) are a small fraction of the tissue that can be obtained at surgical resection or biopsies. It is obvious that large studies are needed (large referring to both the number of genes and the number of cases) before this issue is resolved. An unbiased approach using quantitative analysis of a large number of genes in both normal and neoplastic tissues could definitively help resolve whether CIMP exists and what are the best markers to define it. In colon tumours, appropriate tests to confirm the presence of CIMP should include objective factors unequivocally linked to CIMP, such as histology, as well as mutations in KRAS and BRAF. This endeavour will require an exploration of statistical tools to analyse the data and it could initially follow the methods developed for analysis of microarray data 71 .One possible approach would be to perform a supervised data analysis using CIMP-associated endpoints (for example, the mismatch repair gene MLH1-associated microsatellite instability in sporadic colorectal cancer, or poorly differentiated mucinous sporadic colorectal cancers with mutations in KRAS or BRAF), then identify a minimal set of markers and confirm these in a separate group of tumours. This should be compared with an unsupervised clustering approach. The results of such large studies might then guide appropriate statistical treatment of methylation data sets in other tumours, where CIMP- associated endpoints are not yet clearly defined. Until such experiments are completed, our laboratory has been defining CIMP in colon cancers by quantitatively studying a reduced set of genes , namely MINT1, MINT2, MINT31, CDKN2A and MLH1. One model that accounts for the diverse observations discussed above has been sug- gested by studies on the formation of DNA methylation patterns during embryogenesis. This model 59 , based on initial work by Turker and colleagues (reviewed in REF. 60), proposes the following: methylation initially arises in ?methylation centres?, short sequences that attract DNA methyltrans- ferases based on sequence features. Repetitive elements that can trigger recombination when they are expressed (for example, retro- transposons) form excellent candidates for such methylation centres, but whether they are the only ones involved remains to be seen. This concept solves one of the conundrums in the field ? gene selectivity. Regardless of function, if a gene does not have a methyla- tion centre close to the promoter, it will not be susceptible to DNA methylation in cancer even if it is transcriptionally silenced. Methylation centres are only one half of the equation. DNA methylation, once estab- lished, tends to spread in cis, and this spreading is crucial to eventual gene inacti- vation. CpG islands seem to have developed specific mechanisms to block such methyla- tion spreading, and this barrier is key to maintaining the methylation-free state. Therefore, one view of DNA methylation in predisposed genes is that of a struggle between methylation-promoting events and methylation-protection events. In light of this model, one can propose simple explanations for both age-related methylation and CIMP (FIG. 2).Initially, de novo methylation and spreading is a repli- cation-dependent phenomenon 61 , and is therefore crucially dependent on age. Ageing tissues simply extend the patterns of DNA methylation that were deposited during embryogenesis, and methylation slowly extends towards promoters, piling up at the borders of the islands where protection is presumed to be strong. For some genes, pro- tection against the spreading of methylation is weak, either intrinsically or because protec- tion is mediated by gene expression that can be reduced during ageing, and this will show up as age-related promoter methylation in normal and neoplastic tissues. For other genes, protection is strong, and these genes will only become hypermethylated if there is strong selection for methylation, or if the barrier is somehow lost, possibly by a genetic event. Given that there are probably a limited number of protective mechanisms, the inactivation of a single protecting gene might result in multiple genes being affected simultaneously, exactly as observed in CIMP. Of course, CIMP could also result PERSPECTIVES NATURE REVIEWS | CANCER VOLUME 4 | DECEMBER 2004 | 991 methylation in cancer. MLH1 and MSH2 (mutant S homologue 2) ? two genes that are involved in mismatch repair ? give rise to similar phenotypes (hereditary nonpolyposis colon cancer) when mutated in the familial set- ting, but, in terms of epigenetic inactivation, MLH1 is a frequent target whereas MSH2 is never methylated in cancer, despite having a sizeable CpG island 55 .Second, methylation might follow gene inactivation in some cases 56 . However, hypermethylation cannot be attrib- uted simply to transcriptional inactivation, otherwise, genes showing tissue-specific expression in normal tissues should be hyper- methylated in non-expressing tissues, and this, for the most part, does not occur 2 .Instead,data strongly indicate that gene inactivation might facilitate the establishment of DNA methyla- tion and/or spreading through other mecha- nisms 57 .Third, selection clearly plays a part in the process, during which tumour-suppressor gene methylation is selected for 12 and onco- gene methylation is selected against 58 .Fourth, CIMP implies that a catastrophic event can affect the methylation machinery, thereby rais- ing methylation rates (affecting diverse genes) by several fold in selected cancers. and CIMP-negative gastric cancers tend to have different locations 41 . CIMP-positive acute lymphoblastic leukaemias rarely show the presence of the Philadelphia chromo- some 44 .In general, the simultaneous methy- lation of multiple genes, a hallmark of CIMP, is associated with a poor outcome in multiple malignancies, including head and neck, lung, prostate and oesophageal cancer and acute leukaemias 14 .It is likely that, just as expression profiling is defining pathologi- cally and clinically distinct groups of can- cers, methylation profiling will also achieve a significant reduction in the high degree of heterogeneity that is currently recognized in various malignancies. Causes of CIMP The causes of CIMP remain unknown. More generally, the pathogenesis of aberrant DNA methylation in ageing and cancer and the mechanisms by which genes are selected for methylation have not been elucidated. A series of arguments have to be considered before proposing a unifying theory of the ori- gin of neoplastic methylation patterns. First, not all genes are created equal with regards to Box 2 | The hyperplastic polyp/serrated adenoma pathway Colon cancer has served as a ubiquitous model for the development and progression of neoplasia, from its earliest to most advanced stages 72 .However,recent molecular and pathological findings have led to questions about key parts of that model and are poised to markedly alter our thinking on the prevention of colon cancer 73 .Classical models ascribe all colon tumours to the aberrant crypt focus and adenomatous polyp precursor lesions. Colon cancer prevention by colonoscopy has focused on the adenoma, and adenoma removal is in fact associated with a reduction in colon cancer incidence. Prevention trials have also commonly used the adenoma as a surrogate endpoint. However, for several years, polypoid lesions other than adenomas have been known to occur in the colon. The hyperplastic polyp is an easily recognizable lesion, but it has been given little attention since early studies concluded that it was not a precursor to malignancy. Serrated adenomas (so-called because of their histological appearance) have also been recognized as a rare variant of adenomatous polyps, but have been given little separate attention. The recognition of microsatellite instability (MSI)-positive cancers as a pathologically distinct subset has led investigators to re-evaluate this issue. Histological similarities between lesions, co-occurrence of unusual tumours and compelling molecular data now identify some hyperplastic polyps as precursors to serrated adenomas that seem, in turn, to be precursors to MSI-positive cancers 51 . There are remarkable similarities between proximal hyperplastic polyps, serrated adenomas and CpG island methylator phenotype (CIMP)-positive cancers (a high degree of methylation of multiple genes 53 ,frequent mutations in BRAF 35,54 , and shared histological features such as serration 51 ) and a strong argument can now be made that the serrated pathway is largely a CIMP pathway, and that it could be a precursor of many CIMP- positive cancers. Needless to say, this concept is important clinically. Some hyperplastic polyps (for example, proximal ones and large ones) can no longer be ignored as having no malignant potential. Pathologists are now recognizing serrated lesions at a higher frequency, and a suggestion has emerged that these lesions could be more aggressive in terms of neoplastic transformation. Prevention trials should examine these lesions, given that their molecular characteristics are different than adenomas, and prevention strategies might have to be reviewed. Finally, the identification of individuals with multiple hyperplastic polyposis syndromes 52 might eventually lead to a cause for these lesions that is distinct from classical colon cancer familial genes. that several multiple cancer familial syndromes will eventually be traced to abnormalities in genes that control DNA methylation. Therefore, linkage studies need to be directed by the molecular phenotype of the malignancies of interest, an approach that should increase the chances of correct gene identification by reducing heterogene- ity in the tumours. If, however, CIMP can be traced unequivocally to environmental exposures, this should lead to a careful eval- uation of the epimutagen concept. This issue has a substantial public health impact because the carcinogenic potential of various exposures is currently identified primarily through tests of mutagenicity. Such tests might underestimate the carcinogenic potential of exposures that lead to cancer primarily through epigenetic lesions. The data reviewed above also argue strongly for methylation as an event that is selected for and is therefore physiologically relevant to the neoplastic process. There is still some controversy over the pathophysio- logical role of aberrant methylation in cancer 48 .Nevertheless, the presence of a hypermethylator phenotype provides the same degree of confidence in the role of methylation in cancer as mutator pheno- types do for genetic changes in cancer. In addition, these data indicate that a large number of CpG islands might be methy- lated simultaneously in some (but not all) cancers, which complicates the interpretation of the functional importance of methylation of individual genes, and the use of methyla- tion as a screening tool for neoplasia. Importantly, the frequent association between CIMP and a poor prognosis also indicates a clinical implication. Ultimately, this should lead to a close evaluation of whether epigenetically acting drugs 10 are use- ful in this subset of cases, and whether epige- netic interventions might specifically prevent the emergence of CIMP-positive cancers. Jean-Pierre Issa is at the Department of Leukemia, M. D. Anderson Cancer Center, Unit 425, 1515 Holcombe, Houston, Texas 77030, USA. e-mail: jpissa@mdanderson.org doi: 10.1038/nrc1507 1. Jaenisch, R. & Bird, A. Epigenetic regulation of gene expression: how the genome integrates intrinsic and environmental signals. Nature Genet. 33 (Suppl.), 245?254 (2003). 2. Bird, A. DNA methylation patterns and epigenetic memory. Genes Dev. 16, 6?21 (2002). 3. Li, E., Bestor, T. H. & Jaenisch, R. Targeted mutation of the DNA methyltransferase gene results in embryonic lethality. Cell 69, 915?926 (1992). 4. Yoder, J. A., Walsh, C. P. & Bestor, T. H. Cytosine methylation and the ecology of intragenomic parasites. Trends Genet. 13, 335?340 (1997). 5. Egger, G., Liang, G., Aparicio, A. & Jones, P. A. Epigenetics in human disease and prospects for epigenetic therapy. Nature 429, 457?463 (2004). 992 | DECEMBER 2004 | VOLUME 4 www.nature.com/reviews/cancer PERSPECTIVES from a markedly enhanced de novo methyla- tion pressure, which could result from activating mutations in DNA methyltrans- ferases or histone-modifying proteins. Finally, this model helps to explain why CIMP occurs more frequently in older indi- viduals. Age-related methylation is required to predispose genes to DNA methylation- associated inactivation. Therefore, a genetic event causing CIMP might not affect young tissues, in which methylation is not yet built up at the island borders, ready to spread and inactivate the gene. An alternative explanation for CIMP does not require specific genetic defects, but instead requires repeated exposures to ?epimutagens?. There is mounting evidence for an environ- mental influence on DNA methylation in both normal (ageing) tissues and cancer 32 .Indeed, there is an association between methylation and viral exposures in the liver 19 and in the stomach 62 .Diet might affect DNA methyla- tion patterns 63 , and it would be interesting to determine whether dietary or genetic factors influencing one-carbon metabolism (for example, folate) might affect CIMP. In addition, specific carcinogens seem to have particular preferences for inducing aberrant methylation in rodent models of cancer 21 .It is conceivable, therefore, that CIMP reflects chronic exposure to epimutagens that could then cause or accelerate cancer development through epigenetic pathways. Interestingly, a study of multiple polyps in individuals affected by hyperplastic polyposis showed that there was a high degree of concordance of methylation between the individual lesions 52 , indicating a patient-specific predisposition to CIMP-positive tumours that could be related to either environment and lifestyle, or genetic predisposition. Finally, it is worth mentioning that a universal accelerator of DNA methyla- tion is chronic inflammation, as indicated by studies in preneoplastic colon 64 ,oesophagus 65 , liver 19 and lung 66 .Tumours that arise in the setting of chronic inflammation in the colon (for example, ulcerative colitis) are more likely to be CIMP-positive 67 ,again linking lifestyle and exposures to the phenotype. Implications of CIMP Perhaps the single most important implica- tion of the CIMP concept is that methylation in cancer has a traceable cause. Therefore, a search for mutations in genes that are poten- tially involved in DNA methylation is relevant here, although genes that confer protection against methylation have so far not been identified. The observation of familial clus- tering of CIMP 68 or DNA methylation 69 might lead to the discovery of the gene(s) through classical genetic studies. It is likely Naked DNA DNA?protein complex Border methylation Methylation spreading Ageing CIMP Silenced gene Figure 2 | A model of hypermethylation in cancer. Naked CpG island DNA (top) is unmethylated (yellow) and coated by proteins (green ovals) that protect against DNA methylation establishment and/or spreading. The nature of these proteins is unknown, but probably include transcription factors, co-activators or similar molecules. During repeated rounds of the stem-cell mobilization and replication that accompany ageing, DNA methyltransferases (circles) are recruited to the borders of some CpG islands, depositing methyl groups (red) and creating methylation pressure for these islands. The nature of this initial recruitment is unknown but is probably related to repetitive DNA sequences and/or retrotransposons. The balance of methylation pressure (circles) and methylation protection (ovals) is disrupted in the CpG island methylator phenotype (CIMP), resulting in the spread of methylation into the transcription start area and the triggering of the silencing cascade. As discussed in the text, the disruption of this balance will probably be achieved through the loss of protective proteins (as indicated in the bottom panel), which could occur by mutations that inactivate these proteins or the loss of expression by other mechanisms such as transcription factor loss or histone modifications. Theoretically, this balance could also be disrupted by overactive de novo methylation pressure (circles), for example by activating mutations in DNA methyltransferases. expression in breast cancer. Cell Growth Differ. 8, 773?778 (1997). 57. Song, J. Z., Stirzaker, C., Harrison, J., Melki, J. R. & Clark, S. J. Hypermethylation trigger of the glutathione-S- transferase gene (GSTP1) in prostate cancer cells. Oncogene 21, 1048?1061 (2002). 58. Toyota, M. et al. Aberrant methylation of the cyclooxygenase 2 CpG island in colorectal tumors. Cancer Res. 60, 4044?4048 (2000). 59. Issa, J. P. The epigenetics of colorectal cancer. Ann. NY Acad. Sci. 910, 140?153 (2000). 60. Turker, M. S. Gene silencing in mammalian cells and the spread of DNA methylation. Oncogene 21, 5388?5393 (2002). 61. Velicescu, M. et al. Cell division is required for de novo methylation of CpG islands in bladder cancer cells. Cancer Res. 62, 2378?2384 (2002). 62. Kang, G. H. et al. Epstein-barr virus-positive gastric carcinoma demonstrates frequent aberrant methylation of multiple genes and constitutes CpG island methylator phenotype-positive gastric carcinoma. Am. J. Pathol. 160, 787?794 (2002). 63. Waterland, R. A. & Jirtle, R. L. Transposable elements: targets for early nutritional effects on epigenetic gene regulation. Mol. Cell Biol. 23, 5293?5300 (2003). 64. Issa, J. P., Ahuja, N., Toyota, M., Bronner, M. P. & Brentnall, T. A. Accelerated age-related CpG island methylation in ulcerative colitis. Cancer Res. 61, 3573?3577 (2001). 65. Eads, C. A. et al. Fields of aberrant CpG island hypermethylation in Barrett?s esophagus and associated adenocarcinoma. Cancer Res. 60, 5021?5026 (2000). 66. Zochbauer-Muller, S. et al. Aberrant promoter methylation of multiple genes in non-small cell lung cancers. Cancer Res. 61, 249?255 (2001). 67. Fleisher, A. S. et al. Microsatellite instability in inflammatory bowel disease-associated neoplastic lesions is associated with hypermethylation and diminished expression of the DNA mismatch repair gene, hMLH1. Cancer Res. 60, 4864?4868 (2000). 68. Frazier, M. L. et al. Association of the CpG island methylator phenotype with family history of cancer in patients with colorectal cancer. Cancer Res. 63, 4805?4808 (2003). 69. Ricciardiello, L. et al. Frequent loss of hMLH1 by promoter hypermethylation leads to microsatellite instability in adenomatous polyps of patients with a single first-degree member affected by colon cancer. Cancer Res. 63, 787?792 (2003). 70. Laird, P. W. The power and the promise of DNA methylation markers. Nature Rev. Cancer 3, 253?266 (2003). 71. Leung, Y. F. & Cavalieri, D. Fundamentals of cDNA microarray data analysis. Trends Genet. 19, 649?659 (2003). 72. Kinzler, K. W. & Vogelstein, B. Lessons form hereditary colorectal cancer. Cell 87, 159?170 (1996). 73. Jass, J. R. Limitations of the adenoma-carcinoma sequence in colorectum. Clin. Cancer Res. 10, 5969?5970 (2004). Competing interests statement The author declares no competing financial interests. Online links DATABASES The following terms in this article are linked online to: Entrez Gene: http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=gene DNMT1 | CDKN2A | MLH1 | TMEFF2 | KRAS | BRAF National Cancer Institute: http://www.cancer.gov colon cancer | gastric cancer | liver cancer | pancreatic cancer | oesophageal cancer | ovarian cancer | lymphocytic leukaemia | myelogenous leukaemia FURTHER INFORMATION MD Anderson Cancer Center resource on CpG island methylation: http://www.mdanderson.org/departments/methylation PERSPECTIVES NATURE REVIEWS | CANCER VOLUME 4 | DECEMBER 2004 | 993 30. Toyota, M. et al. Identification of differentially methylated sequences in colorectal cancer by methylated CpG island amplification. Cancer Res. 59, 2307?2312 (1999). 31. Toyota, M. et al. CpG Island methylator phenotype in colorectal cancer. Proc. Natl Acad. Sci. USA 96, 8681?8686 (1999). 32. Issa, J. P. Epigenetic variation and human disease. J. Nutr. 132, 2388S?2392S (2002). 33. Issa, J. P. et al. Methylation of the oestrogen receptor CpG island links ageing and neoplasia in human colon. Nature Genet. 7, 536?540 (1994). 34. Toyota, M., Ohe-Toyota, M., Ahuja, N. & Issa, J. P. Distinct genetic profiles in colorectal tumors with or without the CpG island methylator phenotype. Proc. Natl Acad. Sci. USA 97, 710?715 (2000). 35. Kambara, T. et al. BRAF mutation is associated with DNA methylation in serrated polyps and cancers of the colorectum. Gut 53, 1137?1144 (2004). 36. Whitehall, V. L. et al. Morphological and molecular heterogeneity within nonmicrosatellite instability-high colorectal cancer. Cancer Res. 62, 6011?6014 (2002). 37. Yan, P. S. et al. Use of CpG island microarrays to identify colorectal tumors with a high degree of concurrent methylation. Methods 27, 162?169 (2002). 38. van Rijnsoever, M., Grieu, F., Elsaleh, H., Joseph, D. & Iacopetta, B. Characterisation of colorectal cancers showing hypermethylation at multiple CpG islands. Gut 51, 797?802 (2002). 39. Li, Q. et al. Concordant methylation of the ER and N33 genes in glioblastoma multiforme. Oncogene 16, 3197?3202 (1998). 40. Toyota, M. et al. Aberrant methylation in gastric cancer associated with the CpG island methylator phenotype. Cancer Res. 59, 5438?5442 (1999). 41. Kim, H. et al. Concerted promoter hypermethylation of hMLH1, p16INK4A, and E-cadherin in gastric carcinomas with microsatellite instability. J. Pathol. 200, 23?31 (2003). 42. Ueki, T. et al. Hypermethylation of multiple genes in pancreatic adenocarcinoma. Cancer Res. 60, 1835?1839 (2000). 43. Strathdee, G. et al. Primary ovarian carcinomas display multiple methylator phenotypes involving known tumor suppressor genes. Am. J. Pathol. 158, 1121?1127 (2001). 44. Garcia-Manero, G. et al. DNA methylation of multiple promoter-associated CpG islands in adult acute lymphocytic leukemia. Clin. Cancer Res. 8, 2217?2224 (2002). 45. Toyota, M. et al. Methylation profiling in acute myeloid leukemia. Blood 97, 2823?2829 (2001). 46. Eads, C. A. et al. Epigenetic patterns in the progression of esophageal adenocarcinoma. Cancer Res. 61, 3410?3418 (2001). 47. Yamashita, K., Dai, T., Dai, Y., Yamamoto, F. & Perucho, M. Genetics supersedes epigenetics in colon cancer phenotype. Cancer Cell 4, 121?131 (2003). 48. Bestor, T. H. Unanswered questions about the role of promoter methylation in carcinogenesis. Ann. NY Acad. Sci. 983, 22?27 (2003). 49. Watanabe, T. et al. Molecular predictors of survival after adjuvant chemotherapy for colon cancer. N. Engl. J. Med. 344, 1196?1206 (2001). 50. Ward, R. L. et al. Adverse prognostic effect of methylation in colorectal cancer is reversed by microsatellite instability. J. Clin. Oncol. 21, 3729?3736 (2003). 51. Jass, J. R. Serrated route to colorectal cancer: back street or super highway? J. Pathol. 193, 283?285 (2001). 52. Chan, A. O., Issa, J. P., Morris, J. S., Hamilton, S. R. & Rashid, A. Concordant CpG island methylation in hyperplastic polyposis. Am. J. Pathol. 160, 529?536 (2002). 53. Park, S. J. et al. Frequent CpG island methylation in serrated adenomas of the colorectum. Am. J. Pathol. 162, 815?822 (2003). 54. Chan, T. L., Zhao, W., Leung, S. Y. & Yuen, S. T. BRAF and KRAS mutations in colorectal hyperplastic polyps and serrated adenomas. Cancer Res. 63, 4878?4881 (2003). 55. Cunningham, J. M. et al. Hypermethylation of the hMLH1 promoter in colon cancer with microsatellite instability. Cancer Res. 58, 3455?3460 (1998). 56. Ji, X., Woodard, A. S., Rimm, D. L. & Fearon, E. R. Transcriptional defects underlie loss of e-cadherin 6. Herman, J. G. & Baylin, S. B. Gene silencing in cancer in association with promoter hypermethylation. N. Engl. J. Med. 349, 2042?2054 (2003). 7. Brinster, R. L. Participation of teratocarcinoma cells in mouse embryo development. Cancer Res. 36, 3412?3414 (1976). 8. Li, L., Connelly, M. C., Wetmore, C., Curran, T. & Morgan, J. I. Mouse embryos cloned from brain tumors. Cancer Res. 63, 2733?2736 (2003). 9. Hochedlinger, K. et al. Reprogramming of a melanoma genome by nuclear transplantation. Genes Dev. 18, 1875?1885 (2004). 10. Santini, V., Kantarjian, H. M. & Issa, J. P. Changes in DNA methylation in neoplasia: pathophysiology and therapeutic implications. Ann. Intern. Med. 134, 573?586 (2001). 11. Silverman, L. R. et al. Randomized controlled trial of azacitidine in patients with the myelodysplastic syndrome: a study of the cancer and leukemia group B. J. Clin. Oncol. 20, 2429?2440 (2002). 12. Jones, P. A. & Baylin, S. B. The fundamental role of epigenetic events in cancer. Nature Rev. Genet. 3, 415?428 (2002). 13. Sidransky, D. Emerging molecular markers of cancer. Nature Rev. Cancer 2, 210?219 (2002). 14. Issa, J. P. Methylation and prognosis: of molecular clocks and hypermethylator phenotypes. Clin. Cancer Res. 9, 2879?2881 (2003). 15. Issa, J. P. et al.Increased cytosine DNA- methyltransferase activity during colon cancer progression. J. Natl Cancer Inst. 85, 1235?1240 (1993). 16. De Marzo, A. M. et al. Abnormal regulation of DNA methyltransferase expression during colorectal carcinogenesis. Cancer Res. 59, 3855?3860 (1999). 17. Robertson, K. D., Keyomarsi, K., Gonzales, F. A., Velicescu, M. & Jones, P. A. Differential mRNA expression of the human DNA methyltransferases (DNMTs) 1, 3a and 3b during the G(0)/G(1) to S phase transition in normal and tumor cells. Nucleic Acids Res. 28, 2108?2113 (2000). 18. Eads, C. A. et al.CpG island hypermethylation in human colorectal tumors is not associated with DNA methyltransferase overexpression. Cancer Res. 59, 2302?2306 (1999). 19. Shen, L. et al. DNA methylation and environmental exposures in human hepatocellular carcinoma. J. Natl Cancer Inst. 94, 755?761 (2002). 20. Kim, D. H. et al. p16(INK4a) and histology-specific methylation of CpG islands by exposure to tobacco smoke in non-small cell lung cancer. Cancer Res. 61, 3419?3424 (2001). 21. Issa, J. P., Baylin, S. B. & Belinsky, S. A. Methylation of the estrogen receptor CpG island in lung tumors is related to the specific type of carcinogen exposure. Cancer Res. 56, 3655?3658 (1996). 22. Loeb, L. A. A mutator phenotype in cancer. Cancer Res. 61, 3230?3239 (2001). 23. Baylin, S. B. et al. DNA methylation patterns of the calcitonin gene in human lung cancers and lymphomas. Cancer Res. 46, 2917?2922 (1986). 24. Ahuja, N. et al. Association between CpG island methylation and microsatellite instability in colorectal cancer. Cancer Res. 57, 3370?3374 (1997). 25. 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). 26. Herman, J. G. et al. Incidence and functional consequences of hMLH1 promoter hypermethylation in colorectal carcinoma. Proc. Natl Acad. Sci. USA 95, 6870?6875 (1998). 27. Shibata, D. M. et al. Hypermethylation of HPP1 is associated with hMLH1 hypermethylation in gastric adenocarcinomas. Cancer Res. 62, 5637?5640 (2002). 28. Shen, L., Kondo, Y., Hamilton, S. R., Rashid, A. & Issa, J. P. p14 methylation in human colon cancer is associated with microsatellite instability and wild-type p53. Gastroenterology 124, 626?633 (2003). 29. Yamamoto, H. et al. Differential involvement of the hypermethylator phenotype in hereditary and sporadic colorectal cancers with high-frequency microsatellite instability. Genes Chromosomes Cancer 33, 322?325 (2002). "
Add Content to Group
|
Bookmark
|
Keywords
|
Flag Inappropriate
share
Close
Digg
Facebook
MySpace
Google+
Comments
Close
Please Post Your Comment
*
The Comment you have entered exceeds the maximum length.
Submit
|
Cancel
*
Required
Comments
Please Post Your Comment
No comments yet.
Save Note
Note
View
Public
Private
Friends & Groups
Friends
Groups
Save
|
Cancel
|
Delete
Please provide your notes.
Next
|
Prev
|
Close
|
Edit
|
Delete
Genetics
Gene Inheritance and Transmission
Gene Expression and Regulation
Nucleic Acid Structure and Function
Chromosomes and Cytogenetics
Evolutionary Genetics
Population and Quantitative Genetics
Genomics
Genes and Disease
Genetics and Society
Cell Biology
Cell Origins and Metabolism
Proteins and Gene Expression
Subcellular Compartments
Cell Communication
Cell Cycle and Cell Division
Scientific Communication
Career Planning
Loading ...
Scitable Chat
Register
|
Sign In
Visual Browse
Close
Comments
CloseComments
Please Post Your Comment