Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

Epigenetics as a mechanism driving polygenic clinical drug resistance

Epigenetics as a mechanism driving polygenic clinical drug resistance British Journal of Cancer (2006) 94, 1087 – 1092 & 2006 Cancer Research UK All rights reserved 0007 – 0920/06 $30.00 www.bjcancer.com Minireview Epigenetics as a mechanism driving polygenic clinical drug resistance 1 1 ,1 RM Glasspool , JM Teodoridis and R Brown Centre for Oncology and Applied Pharmacology, Glasgow University, CRUK Beatson Laboratories, Garscube Estate, Glasgow G61 1BD, UK Aberrant methylation of CpG islands located at or near gene promoters is associated with inactivation of gene expression during tumour development. It is increasingly recognised that such epimutations may occur at a much higher frequency than gene mutation and therefore have a greater impact on selection of subpopulations of cells during tumour progression or acquisition of resistance to anticancer drugs. Although laboratory-based models of acquired resistance to anticancer agents tend to focus on specific genes or biochemical pathways, such ‘one gene : one outcome’ models may be an oversimplification of acquired resistance to treatment of cancer patients. Instead, clinical drug resistance may be due to changes in expression of a large number of genes that have a cumulative impact on chemosensitivity. Aberrant CpG island methylation of multiple genes occurring in a nonrandom manner during tumour development and during the acquisition of drug resistance provides a mechanism whereby expression of multiple genes could be affected simultaneously resulting in polygenic clinical drug resistance. If simultaneous epigenetic regulation of multiple genes is indeed a major driving force behind acquired resistance of patients’ tumour to anticancer agents, this has important implications for biomarker studies of clinical outcome following chemotherapy and for clinical approaches designed to circumvent or modulate drug resistance. British Journal of Cancer (2006) 94, 1087–1092. doi:10.1038/sj.bjc.6603024 www.bjcancer.com Published online 21 February 2006 & 2006 Cancer Research UK Keywords: chemotherapy; biomarkers; DNA methylation; drug resistance; epigenetics With the increasing variety of options for the treatment of cancer, in vitro cell lines or transgenic mice. However, there is relatively it is becoming essential that the choice of anticancer therapy, or little evidence that, individually, these mechanisms are able to optimal combination of therapies, is based not only on conven- predict treatment outcome in a manner that is comparable to tional clinical/pathological criteria but also on the molecular known prognostic markers such as stage, performance status and phenotype of the tumour. Many solid tumours are initially histological grade (Agarwal and Kaye, 2003; Hall et al, 2004). The sensitive to chemotherapy, but the vast majority will recur or variability in quality of prognostic and predictive biomarker progress with ultimate failure of conventional cytotoxic che- studies can make reaching a consensus on the value of a given motherapy treatment. In general, novel experimental therapies are marker challenging and recent recommendations have emphasised first examined for efficacy in patients that have failed standard the need for appropriate design and reporting of biomarker studies treatments and whose tumours have acquired resistance to (http://www.cancerdiagnosis.nci.nih.gov/assessment/progress/ cytotoxic drugs. The pattern of gene expression of a tumour that progress/remark.html). Furthermore, response to treatment is only no longer responds to conventional treatment will be very different one factor influencing clinical outcome, numerous other tumour from that of the tumour at presentation due to selection of drug- characteristics, such as capacity for invasion/metastasis or escape resistant subpopulations. However, we know very little about the from the immune response, will also have an impact and may do molecular characteristics of tumours after conventional treatment so irrespective of the therapies used, diluting any association failure or the underlying mechanisms that drive the acquisition of between a marker of drug resistance and clinical outcome. For drug resistance (Agarwal and Kaye, 2003). instance, in the case of ovarian cancer, one of the strongest Laboratory-based studies have identified a wide variety of prognostic markers associate with time to progression of a tumour biochemical pathways and many hundred genes that can after treatment is the number of infiltrating T cells (Zhang et al, potentially influence response to treatment in tumour cells. 2003). However, while these factors may confound the analysis Early work in drug resistance identified genes such as MDR1 of drug-resistance mechanisms, it is also becoming apparent that (P-glycoprotein) (Gottesman, 1993) and p53 (Lowe et al, 1993) as ‘one gene : one outcome’ is an oversimplification for acquired crucial in determining drug resistance in experimental models of resistance to treatment of cancer patients. Thus, it seems increasingly likely that clinical drug resistance is due to polygenic expression changes involving multiple mechanisms rather than to *Correspondence: Professor R Brown; E-mail: [email protected] the alteration of a single pathway or gene. Received 8 November 2005; revised 31 January 2006; accepted 1 An analogy can be made between clinical drug-resistance genes February 2006; published online 21 February 2006 and cancer susceptibility genes. Cancer susceptibility genes such as Epigenetics driving polygenic clinical drug resistance RM Glasspool et al retinoblastoma (RB1) and adenomatois polyposis coli (APC) were It is clear that changes in gene expression do occur following originally identified as rare, mutant alleles that significantly chemotherapy leading to the question, if not gene mutations, what increase the risk of cancer when inherited through the germ line. are the mechanisms leading to changes in gene expression? The More recently, it has been argued that the greater part of cancer answer may lie in the increasing evidence that epigenetic changes predisposition may be due to a combination of weak genetic can be a crucial driving force behind the acquisition of drug variants at many different loci rather than to single high resistance (Teodoridis et al, 2004). Indeed studies of drug-resistant penetrance genes (Balmain et al, 2003). Similarly, the combination cell line models have shown that multiple changes in methylation of weak effects on drug resistance due to expression changes at of CpG islands and epigenetic regulation occur following drug many genes may be more significant than the effect of any single selection (Wei et al, 2003). gene. Since most cytotoxic drugs have a low therapeutic index, Epigenetics changes are heritable changes in gene expression additive effects of multiple low fold changes in drug resistance may that do not involve an alteration in the DNA sequence. Within the be sufficient to cause clinical treatment failure. However, nucleus, DNA is packaged, together with histone proteins, into identification and evaluation of multiple, small additive effects a higher order structure known as chromatin. Interpretation on clinical outcome following chemotherapy will require robust of genetic information coded within the DNA is regulated by and novel statistical and computational approaches that allow mechanisms that involve stable and heritable modifications of nonrandom clustering of effects to be identified. In order to DNA and histones. These modifications include methylation of avoid the pitfalls inherent in analysing high-dimensional data DNA at CpG dinucleotides and methylation, acetylation and sets such as multiple testing and limited sample size, large-scale phosphorylation of histones. Changes in the patterns of these prospective clinical studies are required. In addition, it may modifications are associated with chromatin remodelling and be more informative to study tumours longitudinally, as they can result in changes in gene expression through increasingly acquire resistance during treatment rather than simply sampling understood mechanisms (Lachner et al, 2003). tumours at presentation and to use surrogate end points more DNA methylation involves the transfer of a methyl group to the specific to drug resistance, such as response rather than overall carbon-5 position of cytosine residues, and occurs almost survival. exclusively at cytosines that are followed by a guanine (CpG dinucleotides). CpG dinucleotides are relatively rare in the bulk of the genome and are nearly always methylated, but small stretches of DNA occur that are rich in CpG dinucleotides, so called CpG GENETIC VS EPIGENETIC ALTERATIONS OF islands. These are usually unmethylated in normal cells and are RESISTANCE GENES often associated with the promoter regions of genes (Hendrich and At the time of writing, we have been unable to identify any study of Bird, 2000). Methylation of cytosines within these islands is clinical material that has identified acquisition of a p53 mutation associated with binding of methyl-binding domain (MBD) during treatment of a given patient and similarly gene amplifica- proteins, recruitment of histone deacetylases (HDACs) and histone tion of MDR1, although widely observed in highly resistant cell methyltransferases, histone modification, chromatin condensation lines, is only rarely observed following chemotherapy. Therefore, and transcriptional inactivation of the associated genes. A large although mutations in genes such as MDR1 and p53 confer drug number of genes where aberrant methylation of CpG islands within resistance in vitro and in animal models, and they may have a role their promoters is associated with gene inactivation have now been in inherent resistance, there is little evidence that such genetic identified in tumours (for methods of analysing CpG island changes have a role in acquired clinical resistance following methylation, see Box 1). These include genes involved in all aspects anticancer therapy. of tumour development and also in response to treatment Box 1 Methods for detecting CpG island methylation Methods for the analysis of CpG-island methylation are available both genome-wide and at the single gene level. Restriction landmark genomic scanning (RLGS) is performed by digesting genomic DNA with a methylation-sensitive restriction enzyme, end labelling of the resulting DNA fragments and subsequent digestion with two different restriction enzymes and 2-dimensional gel electrophoresis (Costello et al, 2000). Comparison of signal intensities between tumour and normal DNA after autoradiography allows estimation of the number of aberrantly methylated CpG islands in tumours, and individual aberrantly methylated CpG islands can be identified by sequencing. Differential methylation hybridisation (DMH) is an alternative means of examining genome-wide methylation patterns that uses restriction digestion of genomic DNA and ligation to linkers (Huang et al, 1999), followed by digestion with a methylation-sensitive restriction enzyme such as BstUI, PCR amplification and hybridisation to CpG-rich DNA sequences (representing putative CpG islands). Comparison to hybridisation signals obtained from undigested linker-ligated DNA allowed the identification of aberrantly methylated CpG islands. Methylation sensitive-representational difference analysis (MS-RDA) uses genomic tester and driver DNA samples digested with the methylation-sensitive restriction enzyme HpaII (Ushijima et al, 1997). Sequences that are specific for the tester amplicon are subsequently enriched by repeated cycles of subtractive hybridisations. Several methods for the analysis of the methylation status of individual CpG islands utilise bisulphate treatment of DNA, which has been described in detail (Grunau et al, 2001; Warnecke et al, 2002). Bisulphite treatment of DNA converts unmethylated cytosines into uracil but does not affect methylated cytosines. A difference in methylation is thus converted into a difference in sequence. A widely used method for analysing the methylation status of specific sequences is methylation-specific PCR (MSP) (Herman et al, 1996). Methylation-specific PCR is performed using primers specific for either unmethylated or methylated sequences, thereby allowing the detection of the respective methylation state. Among the advantages of MSP are its easy detection due to its gain-of-signal character and its high sensitivity, allowing the detection of as little as 0.1% methylation in a DNA sample (Herman et al, 1996). The MethyLight technique also involves bisulphite modification. Fluorescence-based PCR is then performed with primers that either overlap CpG methylation sites or that do not overlap any CpG dinucleotides. Sequence discrimination can occur either at the level of the PCR amplification process or at the level of the probe hybridisation process or both (Eads et al, 2000). Combined restriction analysis (COBRA) uses primers that amplify the template irrespective of its methylation state (Xiong and Laird, 1997). The PCR product should therefore be heterogeneous and reflect the various methylation states present in the template. Discrimination of methylation states is achieved by restriction digest using a restriction site whose presence after bisulphite modification depends on the methylation state of the DNA. Combined restriction analysis allows the quantification of the methylation, but its disadvantage is that the methylation of one CpG site is not necessarily representative for the other CpG sites in the analysed sequence. The highest accuracy of methylation density in a region of DNA is achieved by bisulphite sequencing. As in COBRA, the modified DNA is amplified irrespective of its methylation state, but subsequently the amplicon is subcloned and sequenced. This not only allows detection of methylation with a single-nucleotide resolution but also gives information about the distribution of methylated cytosines within individual DNA molecules. The disadvantage is that bisulphite sequencing is relatively labour-intensive. British Journal of Cancer (2006) 94(8), 1087 – 1092 & 2006 Cancer Research UK Epigenetics driving polygenic clinical drug resistance RM Glasspool et al (Teodoridis et al, 2004). Furthermore, for many genes such as selective processes that give rise to specific methylation patterns hMLH1, BRCA1 and E-CADHERIN, aberrant methylation of CpG in tumours remain unclear and are likely to be complex. Changes islands is a far more frequent mechanism of gene inactivation in in cell metabolism (Paz et al, 2002), ‘epigenetic drift’(Egger et al, sporadic tumours than gene mutation or deletion. 2004) and ageing (Richardson, 2002) have all been proposed. For Gene inactivation by DNA methylation can occur at a rate instance, there is a global decrease in global 5 methlycytosine levels several orders of magnitude higher than inactivation of the same in DNA as cells age which is similar to that observed in many gene by mutation (Bhattacharyya et al, 1994). So, if inactivation of tumours (Richardson, 2002). At the same time, localised hyper- a gene is an important mechanism driving the acquisition of drug methylation occurs at some CpG islands. In a restriction landmark resistance, the probability of this occurring by methylation and genome scanning study of CpG island methylation in T being selected for during chemotherapy is much more likely than it lymphocytes from newborn, middle age and elderly subjects, only occurring by mutation. It has also been suggested that some 29 of more than 2000 loci examined were found to alter tumours may acquire a CpG island methylator phenotype, that is, methylation with ageing, with 23 increasing methylation, and six concurrent methylation of genes occurring in a nonrandom decreasing. The same subset also changed methylation status with manner (Toyota et al, 1999). Cellular acquisition of a methylator age in the oesophagus, lung and pancreas, but in variable phenotype could give cells a higher probability of cell transforma- directions (Tra et al, 2002). Thus, age-specific methylation also tion during carcinogenesis, as has been proposed for gene muta- occurs in a nonrandom manner suggesting a tightly controlled tions and the mutator phenotype (Loeb, 1994). Disruption of the process. What ever the process, it seems likely that epigenetic cellular processes involved in methylation could lead to concurrent changes regulating gene expression offer a more rapid means by hypermethylation of multiple genes, including tumour suppressor which tumour cells can adapt to new environment such as genes, and as a result lead to oncogenic transformation. A possible cytotoxic drug therapy than genetic change and because such consequence of this would be that in a tumour with a methylator changes are heritable they can be passed on to daughter cells phenotype there would also be a higher probability of multiple without the need for continuous selection pressure producing drug-resistance/sensitivity genes becoming methylated, with asso- persistent acquired resistance. ciated changes in gene expression. Thus, epigenetic silencing may occur fortuitously during tumour development and only confer an advantage to tumour cells when they are treated with chemo- therapy or radiotherapy. However, the existence of a distinct EVIDENCE FOR THE ROLE OF EPIGENETIC methylator phenotype has been challenged, since a bimodal MECHANISMS IN DRUG RESISTANCE distribution of methylation frequency has not been seen in the same way as observed for gene mutation in tumour cells with the Altered expression of genes involved in apoptosis and DNA repair mutator phenotype (Yamashita et al, 2003; Anacleto et al, 2005). may play an important role in determining response to treatment Nevertheless, the vast majority of tumours, if not all, have and there are many examples of such genes being methylated in aberrant DNA methylation at CpG islands and epigenetic silencing tumours (see Table 1). However, methylation of individual genes of the associated genes. Patterns of CpG island methylation differ may have opposing effects on drug sensitivity. For instance, between and within tumour types in a manner that suggests that methylation of DNA repair genes such as MGMT and FANCF may methylation is not a random process (Costello et al, 2000; Esteller lead to inactivation of DNA repair and confer chemosensitivity, et al, 2001; Wei et al, 2002). Epigenetic inheritance of transcription while methylation and epigenetic silencing of proapoptotic genes patterns has been implicated in the control of cell proliferation such as hMLH1 and APAF1 would confer resistance (Esteller et al, during development, as well as in stem-cell renewal and cancer 2000; Soengas et al, 2001; Taniguchi et al, 2003; Teodoridis et al, (Valk-Lingbeek et al, 2004). However, the mechanisms and 2004). Table 1 Examples of genes associated with drug resistance Gene Function Evidence for role in drug sensitivity Reference Apaf 1 Proapoptotic, binds and promotes Methylation in melanoma cells can be reversed by DNMT inhibitors and this is Soengas et al (2001) caspase 9 activation associated with increased sensitivity to doxorubicin Caspase 8 Proapoptotic Frequently methylated in tumours. Reversal of methylation associated with Fulda et al (2001) increased sensitivity to doxorubicin, etoposide and cisplatin in Ewings sarcoma, neuroblastoma, medulloblastoma and melanoma cell lines hMLH1 DNA mismatch repair protein Methylation and loss of expression associated with resistance to cisplatin in cell Gifford et al (2004) lines, which can be reversed by demethylation with decitabine. Increased frequency of methylation after chemotherapy. Acquisition of hMLH1 methylation during chemotherapy is independently associated with poor overall survival in ovarian patients FancF Activates DNA repair complex Methylation observed in cells with a defective BRCA2 pathway and increased Taniguchi et al (2003) containing BRCA1, and BRCA2 loss sensitivity to cisplatin. Demethylation of FANCF with decitabine reduced cause a decreased ability to repair sensitivity towards cisplatin in these cell line models chemotherapy-induced damage MGMT Removes mutagenic alkyl-groups Methylation and associated loss of expression correlates with response to Paz et al (2004) from the O6-position of guanine temozolamide and BCNU in primary gliomas and overall and progression-free survival in patients with diffuse large B-cell lymphoma treated with cyclophosphamide-containing regimens MCJ Unknown Methylation associated with poor response to therapy and poor overall survival Strathdee et al (2005) in ovarian patients ERb Methylated in 50% of invasive breast cancers. Methylation of ERb less frequent Chang et al (2005) and expression rate was higher in tamoxifen-resistant compared to control tumours & 2006 Cancer Research UK British Journal of Cancer (2006) 94(8), 1087 – 1092 Epigenetics driving polygenic clinical drug resistance RM Glasspool et al The DNA mismatch repair protein, hMLH1, has been shown to that associate with clinical outcome (for instance, see Bair and be necessary for engagement of a variety of downstream cellular Tibshirani, 2004). responses to alkylating agents and cisplatin-induced DNA damage (Papouli et al, 2004). Re-expression of hMLH1 in isogenic model systems has demonstrated that loss of hMLH1 expression confers OVERCOMING EPIGENETIC RESISTANCE resistance to alkylating agents and cisplatin. The frequency of MECHANISMS hMLH1 methylation in ovarian tumours increases after chemo- therapy (Strathdee et al, 1999). Tumours frequently release DNA Epigenetic modifications require active mechanisms of main- which can subsequently be isolated from plasma samples (Johnson tenance and so unlike genetic modifications, they are amenable to and Lo, 2002). Genetic and epigenetic changes that are present in pharmacological manipulation. 5-Azacytidine and its deoxyribose the tumour can be detected in tumour DNA isolated from plasma. analogue, 5-aza-2 -deoxycytidine (decitabine), have been used for Analysis of hMLH1 methylation in tumour DNA isolated from many years to inhibit DNA methyltransferases and reverse DNA plasma of patients with ovarian cancer before chemotherapy and at methylation in tissue culture (Brown and Plumb, 2004). These relapse showed that 25% of patients acquired hMLH1 methylation demethylating agents have been shown to reactivate expression of during chemotherapy and acquisition of hMLH1 methylation was numerous methylation-silenced genes. Decitabine has clinical independently associated with poor overall survival, potentially as activity as a single agent in myelodysplastic syndrome (MDS), a result of poor response to subsequent lines of chemotherapy CML and AML (Issa et al, 2004). Its activity in solid tumours as a (Gifford et al, 2004). single agent has so far been disappointing. However, it may have a In contrast to proapoptotic genes, loss of expression of DNA role in sensitising tumours to other anticancer therapies by repair genes may be associated with increased sensitivity to causing re-expression of genes involved in drug sensitivity (Plumb chemotherapy. The DNA repair enzyme MGMT (O6 methyl et al, 2000). In vitro the differentiating effect of decitabine in guanine methyltransferase) removes mutagenic alkyl-groups from cultured fibroblasts has a narrow dose window with a loss of action the O6-position of guanine, which could otherwise lead to G-A at high doses possibly caused by cytotoxicity as a result of its transitions after DNA replication (Gerson, 2004). As a result, it incorporation into DNA (Taylor and Jones, 1979). It may, inhibits the killing of tumour cells by alkylating agents. therefore, be more appropriate to use demethylating agents at Hypermethylation of the MGMT promoter and associated loss of concentrations below the maximally tolerated dose, but still at a expression correlates with response to temozolamide and BCNU in level where they are known to cause demethylation and induce primary gliomas (Esteller et al, 2000; Paz et al, 2004) and is an gene re-expression. Consistent with this, a low dose schedule independent predictor of overall and progression-free survival in appeared to be superior to schedules using higher doses in a study patients with diffuse large B-cell lymphoma treated with cyclophos- of haematological malignancies (Issa et al, 2004). This has the phamide-containing regimens (Esteller et al, 2002). Importantly, advantage of reducing the bone marrow toxicity of decitabine and the methylation status of MGMT in gliomas at presentation does making it easier to combine it with conventional cytotoxics. not correlate with the clinical response when temozolamide is used Histone deacetylase activity is important in the transcriptional at relapse, demonstrating that the value of biomarkers may depend repression of methylated sequences (Fischle et al, 2003). The on when during tumour progression or treatment they are combination of DNA-demethylating agents and HDAC inhibitors measured. causes synergistic re-expression of epigenetically silenced genes There is thus growing evidence that CpG island methylation of (Cameron et al, 1999). It also produces synergistic antitumour genes with a known direct involvement in drug responses has effects and increased sensitivity to chemotherapeutic agents in cell a potential role in predicting clinical outcome following chemo- line models (Boivin et al, 2002). The potential of this approach is therapy. However, there is a need for studies to investigate the now being assessed in clinical trials (http://www.clinicaltrials.gov/ potential to use methylation patterns of known or unknown genes ct/show/NCT00114257). to identify which patients may benefit from particular chemother- Histone deacetylase inhibitors and demethylating agents, such apeutic regimes or biological therapies. Given the potential of as decitabine, will affect the expression of multiple genes. Given opposing effects depending on which genes are methylated, it is the potential for opposing effects on chemosensitivity when vital to examine whether particular methylation events are different genes are re-expressed, it could be argued that we need dominant in conferring resistance. Methods that allow genome- to develop epigenetic therapies that are more gene specific in their wide analysis of methylation patterns may be particularly mechanism of action. However, if we consider drug resistance to important for these types of study (Box 1). In a study of late- be a polygenic process, then there may be advantages to a stage ovarian tumours, increased methylation of a subset of CpG multitargeted approach. This implies that some patients may islands significantly correlated with worse clinical outcome, as benefit from epigenetic therapies as chemosensitisers, while others defined by the time to clinical disease recurrence after chemo- will not or may even do worse. Therefore, it will be vital to identify therapy (Wei et al, 2002). However in a study of 106 stage III/IV patterns of methylation that reliably predict for response to ovarian cancers, methylation of at least one of a group of genes treatment and whether particular methylation events are dominant involved in DNA repair/drug detoxification (BRCA1, GSTP1, in conferring resistance. In order to do this we need robust MGMT) was associated with improved response to chemotherapy clinically applicable technology to determine methylation patterns (Teodoridis et al, 2005). in tumours both at presentation and at relapse. There is also a need Large-scale analysis of methylation patterns and correlation for pharmacodynamic markers of response to demethylating with response is intrinsically susceptible to the problems of agents. Demethylation can be monitored on a whole-genome level multiple testing. This can be reduced by grouping genes into or by analysis of individual genes (Lyko and Brown, 2005). It has predefined groups according to a biological hypothesis such as been shown that genomic DNA methylation levels are decreased in grouping those with similar biological roles or within the same peripheral blood mononuclear cells from xenograft tumour- pathway, on the assumption that disruption of any one gene within bearing mice treated with 5-aza-2 -deoxycytidine (Plumb et al, a pathway or group will disrupt the functioning of that cellular 2000). This decrease closely coincided with the demethylation of response. This is undoubtedly an oversimplification and the the hMLH1 promoter in the tumours, which indicates that approach will need to be refined as more sophisticated molecular peripheral blood can serve as a surrogate tissue for determining interaction maps and networks are developed (Pommier et al, pharmacodynamic characteristics of DNMT inhibitors. However, 2004). An alternative approach will be to use supervised search although demethylation of individual genes such as p15 has been algorithms that efficiently search array data to identify clusters demonstrated in clinical trials (Daskalakis et al, 2002), the British Journal of Cancer (2006) 94(8), 1087 – 1092 & 2006 Cancer Research UK Epigenetics driving polygenic clinical drug resistance RM Glasspool et al prognostic value of these methylation changes remains to be numbers required for these analyses more feasible. These assays established. need to be conducted in an appropriate quality assured manner and their utility properly evaluated in prospective, randomised trials. Although the epigenetic therapies now undergoing clinical CONCLUSIONS evaluations show promise, there is a need for further agents, which Aberrant epigenetic regulation, such as DNA methylation of CpG are more specific for epigenetic targets. This need not equate to islands, occurs at many genes and in all cancers. CpG island more gene specificity, but rather to less nonspecific toxic effects methylation is a potentially important driving force both for such as the myelosuppression seen with decitabine which may be tumorogenesis and for drug resistance. The use of demethylating the result of direct cytotoxic effects of decitabine rather than agents and HDAC inhibitors offers the potential to favourably alter demethylation. The clinical development of epigenetic therapies the gene expression profile of tumours to cause tumour cell death will require the development of surrogate pharmacodynamic and increased apoptotic response to established cytotoxic agents. markers to assess whether these therapies are having their desired However, we need to identify and evaluate in greater detail the pharmacodynamic effect (e.g. global or gene-specific demethyla- epigenetic characteristics of tumours that predict for lack of tion) and then whether this translates into clinical benefit. response to conventional treatment, so as to identify those patients Epigenetic pharmacodynamic markers can be used as novel end who may particularly benefit from an epigenetic approach. The points in early clinical trials allowing recommended doses to be polygenic nature of these changes will make this challenging. based on maximal biological effect rather than maximum tolerated Towards this objective, genome-wide CpG island methylation of dose. Thus, such pharmacodynamic and predictive epigenetic patients’ tumours can be examined, as well as detailed character- biomarkers, together with targeted drug development, will allow isation of methylation of individual CpG islands. Examination of rational and efficient evaluation of novel epigenetic therapies for tumour DNA released into body fluids may make the large cancer treatment. REFERENCES Agarwal R, Kaye SB (2003) Ovarian cancer: strategies for overcoming Esteller M, Garcia-Foncillas J, Andion E, Goodman SN, Hidalgo OF, resistance to chemotherapy. Nat Rev Cancer 3: 502–516 Vanaclocha V, Baylin SB, Herman JG (2000) Inactivation of the DNA- Anacleto C, Leopoldino AM, Rossi B, Soares FA, Lopes A, Rocha JC, repair gene MGMT and the clinical response of gliomas to alkylating Caballero O, Camargo AA, Simpson AJ, Pena SD (2005) Colorectal cancer agents. N Engl J Med 343: 1350– 1354 ‘methylator phenotype’: fact or artifact? Neoplasia 7: 331–335 Fischle W, Wang Y, Allis CD (2003) Histone and chromatin cross-talk. Curr Bair E, Tibshirani R (2004) Semi-supervised methods to predict patient Opin Cell Biol 15: 172–183 survival from gene expression data. PLoS Biol 2: E108 Fulda S, Kufer MU, Meyer E, van Valen F, Dockhorn-Dworniczak B, Balmain A, Gray J, Ponder B (2003) The genetics and genomics of cancer. Debatin KM (2001) Sensitization for death receptor- or drug-induced Nat Genet 33(Suppl): 238–244 apoptosis by re-expression of caspase-8 through demethylation or gene Bhattacharyya NP, Skandalis A, Ganesh A, Groden J, Meuth M (1994) transfer. Oncogene 20: 5865–5877 Mutator phenotype in human colorectal carcinoma cell lines. Proc Natl Gerson SL (2004) MGMT: its role in cancer aetiology and cancer Acad Sci USA 91: 6319– 6323 therapeutics. Nat Rev Cancer 4: 296 –307 Boivin AJ, Momparler LF, Hurtubise A, Momparler RL (2002) Antineo- Gifford G, Paul J, Vasey PA, Kaye SB, Brown R (2004) The acquisition of plastic action of 5-aza-2 -deoxycytidine and phenylbutyrate on human hMLH1 methylation in plasma DNA after chemotherapy predicts poor lung carcinoma cells. Anticancer Drugs 13: 869–874 survival for ovarian cancer patients. Clin Cancer Res 10: 4420–4426 Brown R, Plumb JA (2004) Demethylation of DNA by decitabine in cancer Gottesman MM (1993) How cancer cells evade chemotherapy. Cancer Res chemotherapy. Expert Rev Anticancer Ther 4: 501–510 53: 747–754 Cameron EE, Bachman KE, Myohanen S, Herman JG, Baylin SB (1999) Grunau C, Clark SJ, Rosenthal A (2001) Bisulfite genomic sequencing: Synergy of demethylation and histone deacetylase inhibition in the re- systematic investigation of critical experimental parameters. Nucleic expression of genes silenced in cancer. Nat Genet 21: 103–107 Acids Res 29: E65-5 Chang HG, Kim SJ, Chung KW, Noh DY, Kwon Y, Lee ES, Kang HS (2005) Hall J, Paul J, Brown R (2004) Critical evaluation of p53 as a prognostic Tamoxifen-resistant breast cancers show less frequent methylation of the marker in ovarian cancer. Expert Rev Mol Med 2004: 1–20 estrogen receptor beta but not the estrogen receptor alpha gene. J Mol Hendrich B, Bird A (2000) Mammalian methyltransferases and methyl- Med 83: 132–139 CpG-binding domains: proteins involved in DNA methylation. Curr Top Costello JF, Fruhwald MC, Smiraglia DJ, Rush LJ, Robertson GP, Gao X, Microbiol Immunol 249: 55–74 Wright FA, Feramisco JD, Peltomaki P, Lang JC, Schuller DE, Yu L, Herman JG, Graff JR, Myohanen S, Nelkin BD, Baylin SB (1996) Bloomfield CD, Caligiuri MA, Yates A, Nishikawa R, Su Huang H, Petrelli Methylation-specific PCR: a novel PCR assay for methylation status of NJ, Zhang X, O’Dorisio MS, Held WA, Cavenee WK, Plass C (2000) CpG islands. Proc Natl Acad Sci USA 93: 9821 –9826 Aberrant CpG-island methylation has non-random and tumour-type- Huang TH, Perry MR, Laux DE (1999) Methylation profiling of CpG islands specific patterns. Nat Genet 24: 132–138 in human breast cancer cells. Hum Mol Genet 8: 459– 470 Daskalakis M, Nguyen TT, Nguyen C, Guldberg P, Kohler G, Wijermans P, Issa JP, Garcia-Manero G, Giles FJ, Mannari R, Thomas D, Faderl S, Bayar Jones PA, Lubbert M (2002) Demethylation of a hypermethylated P15/ E, Lyons J, Rosenfeld CS, Cortes J, Kantarjian HM (2004) Phase 1 study INK4B gene in patients with myelodysplastic syndrome by 5-aza-2 - of low-dose prolonged exposure schedules of the hypomethylating agent deoxycytidine (decitabine) treatment. Blood 100: 2957–2964 5-aza-2 -deoxycytidine (decitabine) in hematopoietic malignancies. Eads CA, Danenberg KD, Kawakami K, Saltz LB, Blake C, Shibata D, Blood 103: 1635 –1640 Danenberg PV, Laird PW (2000) MethyLight: a high-throughput assay to Johnson PJ, Lo YM (2002) Plasma nucleic acids in the diagnosis and measure DNA methylation. Nucleic Acids Res 28: E32 management of malignant disease. Clin Chem 48: 1186 –1193 Egger G, Liang G, Aparicio A, Jones PA (2004) Epigenetics in human Lachner M, O’Sullivan RJ, Jenuwein T (2003) An epigenetic road map for disease and prospects for epigenetic therapy. Nature 429: 457–463 histone lysine methylation. J Cell Sci 116: 2117 –2124 Esteller M, Corn PG, Baylin SB, Herman JG (2001) A gene hypermethylation Loeb LA (1994) Microsatellite instability: marker of a mutator phenotype in profile of human cancer. Cancer Res 61: 3225 –3229 cancer. Cancer Res. 54: 5059– 5063 Esteller M, Gaidano G, Goodman SN, Zagonel V, Capello D, Botto B, Rossi Lowe SW, Ruley HE, Jacks T, Housman DE (1993) p53-dependent apoptosis D, Gloghini A, Vitolo U, Carbone A, Baylin SB, Herman JG (2002) modulates the cytotoxicity of anticancer agents. Cell 74: 957–967 Hypermethylation of the DNA repair gene O(6)-methylguanine DNA Lyko F, Brown R (2005) DNA methyltransferase inhibitors and the methyltransferase and survival of patients with diffuse large B-cell development of epigenetic cancer therapies. J Natl Cancer Inst 97: lymphoma. J Natl Cancer Inst 94: 26–32 1498– 1506 & 2006 Cancer Research UK British Journal of Cancer (2006) 94(8), 1087 – 1092 Epigenetics driving polygenic clinical drug resistance RM Glasspool et al Papouli E, Cejka P, Jiricny J (2004) Dependence of the cytotoxicity of DNA- methylation of DNA damage response genes in advanced ovarian cancer. damaging agents on the mismatch repair status of human cells. Cancer Cancer Res 65: 8961–8967 Res 64: 3391 –3394 Teodoridis JM, Strathdee G, Brown R (2004) Epigenetic silencing mediated Paz MF, Avila S, Fraga MF, Pollan M, Capella G, Peinado MA, Sanchez- by CpG island methylation: potential as a therapeutic target and as a Cespedes M, Herman JG, Esteller M (2002) Germ-line variants in methyl- biomarker. Drug Resist Update 7: 267–278 group metabolism genes and susceptibility to DNA methylation in Tra J, Kondo T, Lu Q, Kuick R, Hanash S, Richardson B (2002) Infrequent normal tissues and human primary tumors. Cancer Res 62: 4519 –4524 occurrence of age-dependent changes in CpG island methylation as Paz MF, Yaya-Tur R, Rojas-Marcos I, Reynes G, Pollan M, Aguirre-Cruz L, detected by restriction landmark genome scanning. Mech Ageing Dev Garcia-Lopez JL, Piquer J, Safont MJ, Balana C, Sanchez-Cespedes M, 123: 1487–1503 Garcia-Villanueva M, Arribas L, Esteller M (2004) CpG island hyper- Toyota M, Ahuja N, Ohe-Toyota M, Herman JG, Baylin SB, Issa JP (1999) methylation of the DNA repair enzyme methyltransferase predicts response CpG island methylator phenotype in colorectal cancer. Proc Natl Acad to temozolomide in primary gliomas. Clin Cancer Res 10: 4933–4938 Sci USA 96: 8681–8686 Plumb JA, Strathdee G, Sludden J, Kaye SB, Brown R (2000) Reversal of Ushijima T, Morimura K, Hosoya Y, Okonogi H, Tatematsu M, Sugimura T, drug resistance in human tumor xenografts by 2 -deoxy-5-azacytidine- Nagao M (1997) Establishment of methylation-sensitive-representa- induced demethylation of the hMLH1 gene promoter. Cancer Res 60: tional difference analysis and isolation of hypo- and hypermethylated 6039– 6044 genomic fragments in mouse liver tumors. Proc Natl Acad Sci USA 94: Pommier Y, Sordet O, Antony S, Hayward RL, Kohn KW (2004) Apoptosis 2284– 2289 defects and chemotherapy resistance: molecular interaction maps and Valk-Lingbeek ME, Bruggeman SW, van Lohuizen M (2004) Stem cells networks. Oncogene 23: 2934–2949 cancer; the polycomb connection. Cell 118: 409–418 Richardson BC (2002) Role of DNA methylation in the regulation of cell Warnecke PM, Stirzaker C, Song J, Grunau C, Melki JR, Clark SJ (2002) function: autoimmunity, aging and cancer. J Nutr 132: 2401S–2405S Identification and resolution of artifacts in bisulfite sequencing. Methods Soengas MS, Capodieci P, Polsky D, Mora J, Esteller M, Opitz-Araya X, 27: 101–107 McCombie R, Herman JG, Gerald WL, Lazebnik YA, Cordon-Cardo C, Wei SH, Brown R, Huang TH (2003) Aberrant DNA methylation in ovarian Lowe SW (2001) Inactivation of the apoptosis effector Apaf-1 in cancer: is there an epigenetic predisposition to drug response? Ann N Y malignant melanoma. Nature 409: 207–211 Acad Sci 983: 243– 250 Strathdee G, MacKean MJ, Illand M, Brown R (1999) A role for methylation Wei SH, Chen CM, Strathdee G, Harnsomburana J, Shyu CR, Rahmatpanah of the hMLH1 promoter in loss of hMLH1 expression and drug resistance F, Shi H, Ng SW, Yan PS, Nephew KP, Brown R, Huang TH (2002) in ovarian cancer. Oncogene 18: 2335–2341 Methylation microarray analysis of late-stage ovarian carcinomas Strathdee G, Vass JK, Oien KA, Siddiqui N, Curto-Garcia J, Brown R (2005) distinguishes progression-free survival in patients and identifies Demethylation of the MCJ gene in stage III/IV epithelial ovarian cancer candidate epigenetic markers. Clin Cancer Res 8: 2246– 2252 and response to chemotherapy. Gynecol Oncol 97: 898–903 Xiong Z, Laird PW (1997) COBRA: a sensitive and quantitative DNA Taniguchi T, Tischkowitz M, Ameziane N, Hodgson SV, Mathew CG, Joenje methylation assay. Nucleic Acids Res 25: 2532–2534 H, Mok SC, D’Andrea AD (2003) Disruption of the Fanconi anemia- Yamashita K, Dai T, Dai Y, Yamamoto F, Perucho M (2003) Genetics BRCA pathway in cisplatin-sensitive ovarian tumors. Nat Med 9: supersedes epigenetics in colon cancer phenotype. Cancer Cell 4: 568–574 121–131 Taylor SM, Jones PA (1979) Multiple new phenotypes induced in 10T1/2 Zhang L, Conejo-Garcia JR, Katsaros D, Gimotty PA, Massobrio M, Regnani and 3T3 cells treated with 5-azacytidine. Cell 17: 771–779 G, Makrigiannakis A, Gray H, Schlienger K, Liebman MN, Rubin SC, Teodoridis JM, Hall J, Marsh S, Kannall HD, Smyth C, Curto J, Siddiqui N, Coukos G (2003) Intratumoral T cells, recurrence, and survival in Gabra H, McLeod HL, Strathdee G, Brown R (2005) CpG island epithelial ovarian cancer. N Engl J Med 348: 203–213 British Journal of Cancer (2006) 94(8), 1087 – 1092 & 2006 Cancer Research UK http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png British Journal of Cancer Springer Journals

Epigenetics as a mechanism driving polygenic clinical drug resistance

Download PDF
6 pages

Loading next page...
 
/lp/springer-journals/epigenetics-as-a-mechanism-driving-polygenic-clinical-drug-resistance-qr75MEIypN

References (111)

Publisher
Springer Journals
Copyright
Copyright © 2006 by The Author(s)
Subject
Biomedicine; Biomedicine, general; Cancer Research; Epidemiology; Molecular Medicine; Oncology; Drug Resistance
ISSN
0007-0920
eISSN
1532-1827
DOI
10.1038/sj.bjc.6603024
Publisher site
See Article on Publisher Site

Abstract

British Journal of Cancer (2006) 94, 1087 – 1092 & 2006 Cancer Research UK All rights reserved 0007 – 0920/06 $30.00 www.bjcancer.com Minireview Epigenetics as a mechanism driving polygenic clinical drug resistance 1 1 ,1 RM Glasspool , JM Teodoridis and R Brown Centre for Oncology and Applied Pharmacology, Glasgow University, CRUK Beatson Laboratories, Garscube Estate, Glasgow G61 1BD, UK Aberrant methylation of CpG islands located at or near gene promoters is associated with inactivation of gene expression during tumour development. It is increasingly recognised that such epimutations may occur at a much higher frequency than gene mutation and therefore have a greater impact on selection of subpopulations of cells during tumour progression or acquisition of resistance to anticancer drugs. Although laboratory-based models of acquired resistance to anticancer agents tend to focus on specific genes or biochemical pathways, such ‘one gene : one outcome’ models may be an oversimplification of acquired resistance to treatment of cancer patients. Instead, clinical drug resistance may be due to changes in expression of a large number of genes that have a cumulative impact on chemosensitivity. Aberrant CpG island methylation of multiple genes occurring in a nonrandom manner during tumour development and during the acquisition of drug resistance provides a mechanism whereby expression of multiple genes could be affected simultaneously resulting in polygenic clinical drug resistance. If simultaneous epigenetic regulation of multiple genes is indeed a major driving force behind acquired resistance of patients’ tumour to anticancer agents, this has important implications for biomarker studies of clinical outcome following chemotherapy and for clinical approaches designed to circumvent or modulate drug resistance. British Journal of Cancer (2006) 94, 1087–1092. doi:10.1038/sj.bjc.6603024 www.bjcancer.com Published online 21 February 2006 & 2006 Cancer Research UK Keywords: chemotherapy; biomarkers; DNA methylation; drug resistance; epigenetics With the increasing variety of options for the treatment of cancer, in vitro cell lines or transgenic mice. However, there is relatively it is becoming essential that the choice of anticancer therapy, or little evidence that, individually, these mechanisms are able to optimal combination of therapies, is based not only on conven- predict treatment outcome in a manner that is comparable to tional clinical/pathological criteria but also on the molecular known prognostic markers such as stage, performance status and phenotype of the tumour. Many solid tumours are initially histological grade (Agarwal and Kaye, 2003; Hall et al, 2004). The sensitive to chemotherapy, but the vast majority will recur or variability in quality of prognostic and predictive biomarker progress with ultimate failure of conventional cytotoxic che- studies can make reaching a consensus on the value of a given motherapy treatment. In general, novel experimental therapies are marker challenging and recent recommendations have emphasised first examined for efficacy in patients that have failed standard the need for appropriate design and reporting of biomarker studies treatments and whose tumours have acquired resistance to (http://www.cancerdiagnosis.nci.nih.gov/assessment/progress/ cytotoxic drugs. The pattern of gene expression of a tumour that progress/remark.html). Furthermore, response to treatment is only no longer responds to conventional treatment will be very different one factor influencing clinical outcome, numerous other tumour from that of the tumour at presentation due to selection of drug- characteristics, such as capacity for invasion/metastasis or escape resistant subpopulations. However, we know very little about the from the immune response, will also have an impact and may do molecular characteristics of tumours after conventional treatment so irrespective of the therapies used, diluting any association failure or the underlying mechanisms that drive the acquisition of between a marker of drug resistance and clinical outcome. For drug resistance (Agarwal and Kaye, 2003). instance, in the case of ovarian cancer, one of the strongest Laboratory-based studies have identified a wide variety of prognostic markers associate with time to progression of a tumour biochemical pathways and many hundred genes that can after treatment is the number of infiltrating T cells (Zhang et al, potentially influence response to treatment in tumour cells. 2003). However, while these factors may confound the analysis Early work in drug resistance identified genes such as MDR1 of drug-resistance mechanisms, it is also becoming apparent that (P-glycoprotein) (Gottesman, 1993) and p53 (Lowe et al, 1993) as ‘one gene : one outcome’ is an oversimplification for acquired crucial in determining drug resistance in experimental models of resistance to treatment of cancer patients. Thus, it seems increasingly likely that clinical drug resistance is due to polygenic expression changes involving multiple mechanisms rather than to *Correspondence: Professor R Brown; E-mail: [email protected] the alteration of a single pathway or gene. Received 8 November 2005; revised 31 January 2006; accepted 1 An analogy can be made between clinical drug-resistance genes February 2006; published online 21 February 2006 and cancer susceptibility genes. Cancer susceptibility genes such as Epigenetics driving polygenic clinical drug resistance RM Glasspool et al retinoblastoma (RB1) and adenomatois polyposis coli (APC) were It is clear that changes in gene expression do occur following originally identified as rare, mutant alleles that significantly chemotherapy leading to the question, if not gene mutations, what increase the risk of cancer when inherited through the germ line. are the mechanisms leading to changes in gene expression? The More recently, it has been argued that the greater part of cancer answer may lie in the increasing evidence that epigenetic changes predisposition may be due to a combination of weak genetic can be a crucial driving force behind the acquisition of drug variants at many different loci rather than to single high resistance (Teodoridis et al, 2004). Indeed studies of drug-resistant penetrance genes (Balmain et al, 2003). Similarly, the combination cell line models have shown that multiple changes in methylation of weak effects on drug resistance due to expression changes at of CpG islands and epigenetic regulation occur following drug many genes may be more significant than the effect of any single selection (Wei et al, 2003). gene. Since most cytotoxic drugs have a low therapeutic index, Epigenetics changes are heritable changes in gene expression additive effects of multiple low fold changes in drug resistance may that do not involve an alteration in the DNA sequence. Within the be sufficient to cause clinical treatment failure. However, nucleus, DNA is packaged, together with histone proteins, into identification and evaluation of multiple, small additive effects a higher order structure known as chromatin. Interpretation on clinical outcome following chemotherapy will require robust of genetic information coded within the DNA is regulated by and novel statistical and computational approaches that allow mechanisms that involve stable and heritable modifications of nonrandom clustering of effects to be identified. In order to DNA and histones. These modifications include methylation of avoid the pitfalls inherent in analysing high-dimensional data DNA at CpG dinucleotides and methylation, acetylation and sets such as multiple testing and limited sample size, large-scale phosphorylation of histones. Changes in the patterns of these prospective clinical studies are required. In addition, it may modifications are associated with chromatin remodelling and be more informative to study tumours longitudinally, as they can result in changes in gene expression through increasingly acquire resistance during treatment rather than simply sampling understood mechanisms (Lachner et al, 2003). tumours at presentation and to use surrogate end points more DNA methylation involves the transfer of a methyl group to the specific to drug resistance, such as response rather than overall carbon-5 position of cytosine residues, and occurs almost survival. exclusively at cytosines that are followed by a guanine (CpG dinucleotides). CpG dinucleotides are relatively rare in the bulk of the genome and are nearly always methylated, but small stretches of DNA occur that are rich in CpG dinucleotides, so called CpG GENETIC VS EPIGENETIC ALTERATIONS OF islands. These are usually unmethylated in normal cells and are RESISTANCE GENES often associated with the promoter regions of genes (Hendrich and At the time of writing, we have been unable to identify any study of Bird, 2000). Methylation of cytosines within these islands is clinical material that has identified acquisition of a p53 mutation associated with binding of methyl-binding domain (MBD) during treatment of a given patient and similarly gene amplifica- proteins, recruitment of histone deacetylases (HDACs) and histone tion of MDR1, although widely observed in highly resistant cell methyltransferases, histone modification, chromatin condensation lines, is only rarely observed following chemotherapy. Therefore, and transcriptional inactivation of the associated genes. A large although mutations in genes such as MDR1 and p53 confer drug number of genes where aberrant methylation of CpG islands within resistance in vitro and in animal models, and they may have a role their promoters is associated with gene inactivation have now been in inherent resistance, there is little evidence that such genetic identified in tumours (for methods of analysing CpG island changes have a role in acquired clinical resistance following methylation, see Box 1). These include genes involved in all aspects anticancer therapy. of tumour development and also in response to treatment Box 1 Methods for detecting CpG island methylation Methods for the analysis of CpG-island methylation are available both genome-wide and at the single gene level. Restriction landmark genomic scanning (RLGS) is performed by digesting genomic DNA with a methylation-sensitive restriction enzyme, end labelling of the resulting DNA fragments and subsequent digestion with two different restriction enzymes and 2-dimensional gel electrophoresis (Costello et al, 2000). Comparison of signal intensities between tumour and normal DNA after autoradiography allows estimation of the number of aberrantly methylated CpG islands in tumours, and individual aberrantly methylated CpG islands can be identified by sequencing. Differential methylation hybridisation (DMH) is an alternative means of examining genome-wide methylation patterns that uses restriction digestion of genomic DNA and ligation to linkers (Huang et al, 1999), followed by digestion with a methylation-sensitive restriction enzyme such as BstUI, PCR amplification and hybridisation to CpG-rich DNA sequences (representing putative CpG islands). Comparison to hybridisation signals obtained from undigested linker-ligated DNA allowed the identification of aberrantly methylated CpG islands. Methylation sensitive-representational difference analysis (MS-RDA) uses genomic tester and driver DNA samples digested with the methylation-sensitive restriction enzyme HpaII (Ushijima et al, 1997). Sequences that are specific for the tester amplicon are subsequently enriched by repeated cycles of subtractive hybridisations. Several methods for the analysis of the methylation status of individual CpG islands utilise bisulphate treatment of DNA, which has been described in detail (Grunau et al, 2001; Warnecke et al, 2002). Bisulphite treatment of DNA converts unmethylated cytosines into uracil but does not affect methylated cytosines. A difference in methylation is thus converted into a difference in sequence. A widely used method for analysing the methylation status of specific sequences is methylation-specific PCR (MSP) (Herman et al, 1996). Methylation-specific PCR is performed using primers specific for either unmethylated or methylated sequences, thereby allowing the detection of the respective methylation state. Among the advantages of MSP are its easy detection due to its gain-of-signal character and its high sensitivity, allowing the detection of as little as 0.1% methylation in a DNA sample (Herman et al, 1996). The MethyLight technique also involves bisulphite modification. Fluorescence-based PCR is then performed with primers that either overlap CpG methylation sites or that do not overlap any CpG dinucleotides. Sequence discrimination can occur either at the level of the PCR amplification process or at the level of the probe hybridisation process or both (Eads et al, 2000). Combined restriction analysis (COBRA) uses primers that amplify the template irrespective of its methylation state (Xiong and Laird, 1997). The PCR product should therefore be heterogeneous and reflect the various methylation states present in the template. Discrimination of methylation states is achieved by restriction digest using a restriction site whose presence after bisulphite modification depends on the methylation state of the DNA. Combined restriction analysis allows the quantification of the methylation, but its disadvantage is that the methylation of one CpG site is not necessarily representative for the other CpG sites in the analysed sequence. The highest accuracy of methylation density in a region of DNA is achieved by bisulphite sequencing. As in COBRA, the modified DNA is amplified irrespective of its methylation state, but subsequently the amplicon is subcloned and sequenced. This not only allows detection of methylation with a single-nucleotide resolution but also gives information about the distribution of methylated cytosines within individual DNA molecules. The disadvantage is that bisulphite sequencing is relatively labour-intensive. British Journal of Cancer (2006) 94(8), 1087 – 1092 & 2006 Cancer Research UK Epigenetics driving polygenic clinical drug resistance RM Glasspool et al (Teodoridis et al, 2004). Furthermore, for many genes such as selective processes that give rise to specific methylation patterns hMLH1, BRCA1 and E-CADHERIN, aberrant methylation of CpG in tumours remain unclear and are likely to be complex. Changes islands is a far more frequent mechanism of gene inactivation in in cell metabolism (Paz et al, 2002), ‘epigenetic drift’(Egger et al, sporadic tumours than gene mutation or deletion. 2004) and ageing (Richardson, 2002) have all been proposed. For Gene inactivation by DNA methylation can occur at a rate instance, there is a global decrease in global 5 methlycytosine levels several orders of magnitude higher than inactivation of the same in DNA as cells age which is similar to that observed in many gene by mutation (Bhattacharyya et al, 1994). So, if inactivation of tumours (Richardson, 2002). At the same time, localised hyper- a gene is an important mechanism driving the acquisition of drug methylation occurs at some CpG islands. In a restriction landmark resistance, the probability of this occurring by methylation and genome scanning study of CpG island methylation in T being selected for during chemotherapy is much more likely than it lymphocytes from newborn, middle age and elderly subjects, only occurring by mutation. It has also been suggested that some 29 of more than 2000 loci examined were found to alter tumours may acquire a CpG island methylator phenotype, that is, methylation with ageing, with 23 increasing methylation, and six concurrent methylation of genes occurring in a nonrandom decreasing. The same subset also changed methylation status with manner (Toyota et al, 1999). Cellular acquisition of a methylator age in the oesophagus, lung and pancreas, but in variable phenotype could give cells a higher probability of cell transforma- directions (Tra et al, 2002). Thus, age-specific methylation also tion during carcinogenesis, as has been proposed for gene muta- occurs in a nonrandom manner suggesting a tightly controlled tions and the mutator phenotype (Loeb, 1994). Disruption of the process. What ever the process, it seems likely that epigenetic cellular processes involved in methylation could lead to concurrent changes regulating gene expression offer a more rapid means by hypermethylation of multiple genes, including tumour suppressor which tumour cells can adapt to new environment such as genes, and as a result lead to oncogenic transformation. A possible cytotoxic drug therapy than genetic change and because such consequence of this would be that in a tumour with a methylator changes are heritable they can be passed on to daughter cells phenotype there would also be a higher probability of multiple without the need for continuous selection pressure producing drug-resistance/sensitivity genes becoming methylated, with asso- persistent acquired resistance. ciated changes in gene expression. Thus, epigenetic silencing may occur fortuitously during tumour development and only confer an advantage to tumour cells when they are treated with chemo- therapy or radiotherapy. However, the existence of a distinct EVIDENCE FOR THE ROLE OF EPIGENETIC methylator phenotype has been challenged, since a bimodal MECHANISMS IN DRUG RESISTANCE distribution of methylation frequency has not been seen in the same way as observed for gene mutation in tumour cells with the Altered expression of genes involved in apoptosis and DNA repair mutator phenotype (Yamashita et al, 2003; Anacleto et al, 2005). may play an important role in determining response to treatment Nevertheless, the vast majority of tumours, if not all, have and there are many examples of such genes being methylated in aberrant DNA methylation at CpG islands and epigenetic silencing tumours (see Table 1). However, methylation of individual genes of the associated genes. Patterns of CpG island methylation differ may have opposing effects on drug sensitivity. For instance, between and within tumour types in a manner that suggests that methylation of DNA repair genes such as MGMT and FANCF may methylation is not a random process (Costello et al, 2000; Esteller lead to inactivation of DNA repair and confer chemosensitivity, et al, 2001; Wei et al, 2002). Epigenetic inheritance of transcription while methylation and epigenetic silencing of proapoptotic genes patterns has been implicated in the control of cell proliferation such as hMLH1 and APAF1 would confer resistance (Esteller et al, during development, as well as in stem-cell renewal and cancer 2000; Soengas et al, 2001; Taniguchi et al, 2003; Teodoridis et al, (Valk-Lingbeek et al, 2004). However, the mechanisms and 2004). Table 1 Examples of genes associated with drug resistance Gene Function Evidence for role in drug sensitivity Reference Apaf 1 Proapoptotic, binds and promotes Methylation in melanoma cells can be reversed by DNMT inhibitors and this is Soengas et al (2001) caspase 9 activation associated with increased sensitivity to doxorubicin Caspase 8 Proapoptotic Frequently methylated in tumours. Reversal of methylation associated with Fulda et al (2001) increased sensitivity to doxorubicin, etoposide and cisplatin in Ewings sarcoma, neuroblastoma, medulloblastoma and melanoma cell lines hMLH1 DNA mismatch repair protein Methylation and loss of expression associated with resistance to cisplatin in cell Gifford et al (2004) lines, which can be reversed by demethylation with decitabine. Increased frequency of methylation after chemotherapy. Acquisition of hMLH1 methylation during chemotherapy is independently associated with poor overall survival in ovarian patients FancF Activates DNA repair complex Methylation observed in cells with a defective BRCA2 pathway and increased Taniguchi et al (2003) containing BRCA1, and BRCA2 loss sensitivity to cisplatin. Demethylation of FANCF with decitabine reduced cause a decreased ability to repair sensitivity towards cisplatin in these cell line models chemotherapy-induced damage MGMT Removes mutagenic alkyl-groups Methylation and associated loss of expression correlates with response to Paz et al (2004) from the O6-position of guanine temozolamide and BCNU in primary gliomas and overall and progression-free survival in patients with diffuse large B-cell lymphoma treated with cyclophosphamide-containing regimens MCJ Unknown Methylation associated with poor response to therapy and poor overall survival Strathdee et al (2005) in ovarian patients ERb Methylated in 50% of invasive breast cancers. Methylation of ERb less frequent Chang et al (2005) and expression rate was higher in tamoxifen-resistant compared to control tumours & 2006 Cancer Research UK British Journal of Cancer (2006) 94(8), 1087 – 1092 Epigenetics driving polygenic clinical drug resistance RM Glasspool et al The DNA mismatch repair protein, hMLH1, has been shown to that associate with clinical outcome (for instance, see Bair and be necessary for engagement of a variety of downstream cellular Tibshirani, 2004). responses to alkylating agents and cisplatin-induced DNA damage (Papouli et al, 2004). Re-expression of hMLH1 in isogenic model systems has demonstrated that loss of hMLH1 expression confers OVERCOMING EPIGENETIC RESISTANCE resistance to alkylating agents and cisplatin. The frequency of MECHANISMS hMLH1 methylation in ovarian tumours increases after chemo- therapy (Strathdee et al, 1999). Tumours frequently release DNA Epigenetic modifications require active mechanisms of main- which can subsequently be isolated from plasma samples (Johnson tenance and so unlike genetic modifications, they are amenable to and Lo, 2002). Genetic and epigenetic changes that are present in pharmacological manipulation. 5-Azacytidine and its deoxyribose the tumour can be detected in tumour DNA isolated from plasma. analogue, 5-aza-2 -deoxycytidine (decitabine), have been used for Analysis of hMLH1 methylation in tumour DNA isolated from many years to inhibit DNA methyltransferases and reverse DNA plasma of patients with ovarian cancer before chemotherapy and at methylation in tissue culture (Brown and Plumb, 2004). These relapse showed that 25% of patients acquired hMLH1 methylation demethylating agents have been shown to reactivate expression of during chemotherapy and acquisition of hMLH1 methylation was numerous methylation-silenced genes. Decitabine has clinical independently associated with poor overall survival, potentially as activity as a single agent in myelodysplastic syndrome (MDS), a result of poor response to subsequent lines of chemotherapy CML and AML (Issa et al, 2004). Its activity in solid tumours as a (Gifford et al, 2004). single agent has so far been disappointing. However, it may have a In contrast to proapoptotic genes, loss of expression of DNA role in sensitising tumours to other anticancer therapies by repair genes may be associated with increased sensitivity to causing re-expression of genes involved in drug sensitivity (Plumb chemotherapy. The DNA repair enzyme MGMT (O6 methyl et al, 2000). In vitro the differentiating effect of decitabine in guanine methyltransferase) removes mutagenic alkyl-groups from cultured fibroblasts has a narrow dose window with a loss of action the O6-position of guanine, which could otherwise lead to G-A at high doses possibly caused by cytotoxicity as a result of its transitions after DNA replication (Gerson, 2004). As a result, it incorporation into DNA (Taylor and Jones, 1979). It may, inhibits the killing of tumour cells by alkylating agents. therefore, be more appropriate to use demethylating agents at Hypermethylation of the MGMT promoter and associated loss of concentrations below the maximally tolerated dose, but still at a expression correlates with response to temozolamide and BCNU in level where they are known to cause demethylation and induce primary gliomas (Esteller et al, 2000; Paz et al, 2004) and is an gene re-expression. Consistent with this, a low dose schedule independent predictor of overall and progression-free survival in appeared to be superior to schedules using higher doses in a study patients with diffuse large B-cell lymphoma treated with cyclophos- of haematological malignancies (Issa et al, 2004). This has the phamide-containing regimens (Esteller et al, 2002). Importantly, advantage of reducing the bone marrow toxicity of decitabine and the methylation status of MGMT in gliomas at presentation does making it easier to combine it with conventional cytotoxics. not correlate with the clinical response when temozolamide is used Histone deacetylase activity is important in the transcriptional at relapse, demonstrating that the value of biomarkers may depend repression of methylated sequences (Fischle et al, 2003). The on when during tumour progression or treatment they are combination of DNA-demethylating agents and HDAC inhibitors measured. causes synergistic re-expression of epigenetically silenced genes There is thus growing evidence that CpG island methylation of (Cameron et al, 1999). It also produces synergistic antitumour genes with a known direct involvement in drug responses has effects and increased sensitivity to chemotherapeutic agents in cell a potential role in predicting clinical outcome following chemo- line models (Boivin et al, 2002). The potential of this approach is therapy. However, there is a need for studies to investigate the now being assessed in clinical trials (http://www.clinicaltrials.gov/ potential to use methylation patterns of known or unknown genes ct/show/NCT00114257). to identify which patients may benefit from particular chemother- Histone deacetylase inhibitors and demethylating agents, such apeutic regimes or biological therapies. Given the potential of as decitabine, will affect the expression of multiple genes. Given opposing effects depending on which genes are methylated, it is the potential for opposing effects on chemosensitivity when vital to examine whether particular methylation events are different genes are re-expressed, it could be argued that we need dominant in conferring resistance. Methods that allow genome- to develop epigenetic therapies that are more gene specific in their wide analysis of methylation patterns may be particularly mechanism of action. However, if we consider drug resistance to important for these types of study (Box 1). In a study of late- be a polygenic process, then there may be advantages to a stage ovarian tumours, increased methylation of a subset of CpG multitargeted approach. This implies that some patients may islands significantly correlated with worse clinical outcome, as benefit from epigenetic therapies as chemosensitisers, while others defined by the time to clinical disease recurrence after chemo- will not or may even do worse. Therefore, it will be vital to identify therapy (Wei et al, 2002). However in a study of 106 stage III/IV patterns of methylation that reliably predict for response to ovarian cancers, methylation of at least one of a group of genes treatment and whether particular methylation events are dominant involved in DNA repair/drug detoxification (BRCA1, GSTP1, in conferring resistance. In order to do this we need robust MGMT) was associated with improved response to chemotherapy clinically applicable technology to determine methylation patterns (Teodoridis et al, 2005). in tumours both at presentation and at relapse. There is also a need Large-scale analysis of methylation patterns and correlation for pharmacodynamic markers of response to demethylating with response is intrinsically susceptible to the problems of agents. Demethylation can be monitored on a whole-genome level multiple testing. This can be reduced by grouping genes into or by analysis of individual genes (Lyko and Brown, 2005). It has predefined groups according to a biological hypothesis such as been shown that genomic DNA methylation levels are decreased in grouping those with similar biological roles or within the same peripheral blood mononuclear cells from xenograft tumour- pathway, on the assumption that disruption of any one gene within bearing mice treated with 5-aza-2 -deoxycytidine (Plumb et al, a pathway or group will disrupt the functioning of that cellular 2000). This decrease closely coincided with the demethylation of response. This is undoubtedly an oversimplification and the the hMLH1 promoter in the tumours, which indicates that approach will need to be refined as more sophisticated molecular peripheral blood can serve as a surrogate tissue for determining interaction maps and networks are developed (Pommier et al, pharmacodynamic characteristics of DNMT inhibitors. However, 2004). An alternative approach will be to use supervised search although demethylation of individual genes such as p15 has been algorithms that efficiently search array data to identify clusters demonstrated in clinical trials (Daskalakis et al, 2002), the British Journal of Cancer (2006) 94(8), 1087 – 1092 & 2006 Cancer Research UK Epigenetics driving polygenic clinical drug resistance RM Glasspool et al prognostic value of these methylation changes remains to be numbers required for these analyses more feasible. These assays established. need to be conducted in an appropriate quality assured manner and their utility properly evaluated in prospective, randomised trials. Although the epigenetic therapies now undergoing clinical CONCLUSIONS evaluations show promise, there is a need for further agents, which Aberrant epigenetic regulation, such as DNA methylation of CpG are more specific for epigenetic targets. This need not equate to islands, occurs at many genes and in all cancers. CpG island more gene specificity, but rather to less nonspecific toxic effects methylation is a potentially important driving force both for such as the myelosuppression seen with decitabine which may be tumorogenesis and for drug resistance. The use of demethylating the result of direct cytotoxic effects of decitabine rather than agents and HDAC inhibitors offers the potential to favourably alter demethylation. The clinical development of epigenetic therapies the gene expression profile of tumours to cause tumour cell death will require the development of surrogate pharmacodynamic and increased apoptotic response to established cytotoxic agents. markers to assess whether these therapies are having their desired However, we need to identify and evaluate in greater detail the pharmacodynamic effect (e.g. global or gene-specific demethyla- epigenetic characteristics of tumours that predict for lack of tion) and then whether this translates into clinical benefit. response to conventional treatment, so as to identify those patients Epigenetic pharmacodynamic markers can be used as novel end who may particularly benefit from an epigenetic approach. The points in early clinical trials allowing recommended doses to be polygenic nature of these changes will make this challenging. based on maximal biological effect rather than maximum tolerated Towards this objective, genome-wide CpG island methylation of dose. Thus, such pharmacodynamic and predictive epigenetic patients’ tumours can be examined, as well as detailed character- biomarkers, together with targeted drug development, will allow isation of methylation of individual CpG islands. Examination of rational and efficient evaluation of novel epigenetic therapies for tumour DNA released into body fluids may make the large cancer treatment. REFERENCES Agarwal R, Kaye SB (2003) Ovarian cancer: strategies for overcoming Esteller M, Garcia-Foncillas J, Andion E, Goodman SN, Hidalgo OF, resistance to chemotherapy. Nat Rev Cancer 3: 502–516 Vanaclocha V, Baylin SB, Herman JG (2000) Inactivation of the DNA- Anacleto C, Leopoldino AM, Rossi B, Soares FA, Lopes A, Rocha JC, repair gene MGMT and the clinical response of gliomas to alkylating Caballero O, Camargo AA, Simpson AJ, Pena SD (2005) Colorectal cancer agents. N Engl J Med 343: 1350– 1354 ‘methylator phenotype’: fact or artifact? Neoplasia 7: 331–335 Fischle W, Wang Y, Allis CD (2003) Histone and chromatin cross-talk. Curr Bair E, Tibshirani R (2004) Semi-supervised methods to predict patient Opin Cell Biol 15: 172–183 survival from gene expression data. PLoS Biol 2: E108 Fulda S, Kufer MU, Meyer E, van Valen F, Dockhorn-Dworniczak B, Balmain A, Gray J, Ponder B (2003) The genetics and genomics of cancer. Debatin KM (2001) Sensitization for death receptor- or drug-induced Nat Genet 33(Suppl): 238–244 apoptosis by re-expression of caspase-8 through demethylation or gene Bhattacharyya NP, Skandalis A, Ganesh A, Groden J, Meuth M (1994) transfer. Oncogene 20: 5865–5877 Mutator phenotype in human colorectal carcinoma cell lines. Proc Natl Gerson SL (2004) MGMT: its role in cancer aetiology and cancer Acad Sci USA 91: 6319– 6323 therapeutics. Nat Rev Cancer 4: 296 –307 Boivin AJ, Momparler LF, Hurtubise A, Momparler RL (2002) Antineo- Gifford G, Paul J, Vasey PA, Kaye SB, Brown R (2004) The acquisition of plastic action of 5-aza-2 -deoxycytidine and phenylbutyrate on human hMLH1 methylation in plasma DNA after chemotherapy predicts poor lung carcinoma cells. Anticancer Drugs 13: 869–874 survival for ovarian cancer patients. Clin Cancer Res 10: 4420–4426 Brown R, Plumb JA (2004) Demethylation of DNA by decitabine in cancer Gottesman MM (1993) How cancer cells evade chemotherapy. Cancer Res chemotherapy. Expert Rev Anticancer Ther 4: 501–510 53: 747–754 Cameron EE, Bachman KE, Myohanen S, Herman JG, Baylin SB (1999) Grunau C, Clark SJ, Rosenthal A (2001) Bisulfite genomic sequencing: Synergy of demethylation and histone deacetylase inhibition in the re- systematic investigation of critical experimental parameters. Nucleic expression of genes silenced in cancer. Nat Genet 21: 103–107 Acids Res 29: E65-5 Chang HG, Kim SJ, Chung KW, Noh DY, Kwon Y, Lee ES, Kang HS (2005) Hall J, Paul J, Brown R (2004) Critical evaluation of p53 as a prognostic Tamoxifen-resistant breast cancers show less frequent methylation of the marker in ovarian cancer. Expert Rev Mol Med 2004: 1–20 estrogen receptor beta but not the estrogen receptor alpha gene. J Mol Hendrich B, Bird A (2000) Mammalian methyltransferases and methyl- Med 83: 132–139 CpG-binding domains: proteins involved in DNA methylation. Curr Top Costello JF, Fruhwald MC, Smiraglia DJ, Rush LJ, Robertson GP, Gao X, Microbiol Immunol 249: 55–74 Wright FA, Feramisco JD, Peltomaki P, Lang JC, Schuller DE, Yu L, Herman JG, Graff JR, Myohanen S, Nelkin BD, Baylin SB (1996) Bloomfield CD, Caligiuri MA, Yates A, Nishikawa R, Su Huang H, Petrelli Methylation-specific PCR: a novel PCR assay for methylation status of NJ, Zhang X, O’Dorisio MS, Held WA, Cavenee WK, Plass C (2000) CpG islands. Proc Natl Acad Sci USA 93: 9821 –9826 Aberrant CpG-island methylation has non-random and tumour-type- Huang TH, Perry MR, Laux DE (1999) Methylation profiling of CpG islands specific patterns. Nat Genet 24: 132–138 in human breast cancer cells. Hum Mol Genet 8: 459– 470 Daskalakis M, Nguyen TT, Nguyen C, Guldberg P, Kohler G, Wijermans P, Issa JP, Garcia-Manero G, Giles FJ, Mannari R, Thomas D, Faderl S, Bayar Jones PA, Lubbert M (2002) Demethylation of a hypermethylated P15/ E, Lyons J, Rosenfeld CS, Cortes J, Kantarjian HM (2004) Phase 1 study INK4B gene in patients with myelodysplastic syndrome by 5-aza-2 - of low-dose prolonged exposure schedules of the hypomethylating agent deoxycytidine (decitabine) treatment. Blood 100: 2957–2964 5-aza-2 -deoxycytidine (decitabine) in hematopoietic malignancies. Eads CA, Danenberg KD, Kawakami K, Saltz LB, Blake C, Shibata D, Blood 103: 1635 –1640 Danenberg PV, Laird PW (2000) MethyLight: a high-throughput assay to Johnson PJ, Lo YM (2002) Plasma nucleic acids in the diagnosis and measure DNA methylation. Nucleic Acids Res 28: E32 management of malignant disease. Clin Chem 48: 1186 –1193 Egger G, Liang G, Aparicio A, Jones PA (2004) Epigenetics in human Lachner M, O’Sullivan RJ, Jenuwein T (2003) An epigenetic road map for disease and prospects for epigenetic therapy. Nature 429: 457–463 histone lysine methylation. J Cell Sci 116: 2117 –2124 Esteller M, Corn PG, Baylin SB, Herman JG (2001) A gene hypermethylation Loeb LA (1994) Microsatellite instability: marker of a mutator phenotype in profile of human cancer. Cancer Res 61: 3225 –3229 cancer. Cancer Res. 54: 5059– 5063 Esteller M, Gaidano G, Goodman SN, Zagonel V, Capello D, Botto B, Rossi Lowe SW, Ruley HE, Jacks T, Housman DE (1993) p53-dependent apoptosis D, Gloghini A, Vitolo U, Carbone A, Baylin SB, Herman JG (2002) modulates the cytotoxicity of anticancer agents. Cell 74: 957–967 Hypermethylation of the DNA repair gene O(6)-methylguanine DNA Lyko F, Brown R (2005) DNA methyltransferase inhibitors and the methyltransferase and survival of patients with diffuse large B-cell development of epigenetic cancer therapies. J Natl Cancer Inst 97: lymphoma. J Natl Cancer Inst 94: 26–32 1498– 1506 & 2006 Cancer Research UK British Journal of Cancer (2006) 94(8), 1087 – 1092 Epigenetics driving polygenic clinical drug resistance RM Glasspool et al Papouli E, Cejka P, Jiricny J (2004) Dependence of the cytotoxicity of DNA- methylation of DNA damage response genes in advanced ovarian cancer. damaging agents on the mismatch repair status of human cells. Cancer Cancer Res 65: 8961–8967 Res 64: 3391 –3394 Teodoridis JM, Strathdee G, Brown R (2004) Epigenetic silencing mediated Paz MF, Avila S, Fraga MF, Pollan M, Capella G, Peinado MA, Sanchez- by CpG island methylation: potential as a therapeutic target and as a Cespedes M, Herman JG, Esteller M (2002) Germ-line variants in methyl- biomarker. Drug Resist Update 7: 267–278 group metabolism genes and susceptibility to DNA methylation in Tra J, Kondo T, Lu Q, Kuick R, Hanash S, Richardson B (2002) Infrequent normal tissues and human primary tumors. Cancer Res 62: 4519 –4524 occurrence of age-dependent changes in CpG island methylation as Paz MF, Yaya-Tur R, Rojas-Marcos I, Reynes G, Pollan M, Aguirre-Cruz L, detected by restriction landmark genome scanning. Mech Ageing Dev Garcia-Lopez JL, Piquer J, Safont MJ, Balana C, Sanchez-Cespedes M, 123: 1487–1503 Garcia-Villanueva M, Arribas L, Esteller M (2004) CpG island hyper- Toyota M, Ahuja N, Ohe-Toyota M, Herman JG, Baylin SB, Issa JP (1999) methylation of the DNA repair enzyme methyltransferase predicts response CpG island methylator phenotype in colorectal cancer. Proc Natl Acad to temozolomide in primary gliomas. Clin Cancer Res 10: 4933–4938 Sci USA 96: 8681–8686 Plumb JA, Strathdee G, Sludden J, Kaye SB, Brown R (2000) Reversal of Ushijima T, Morimura K, Hosoya Y, Okonogi H, Tatematsu M, Sugimura T, drug resistance in human tumor xenografts by 2 -deoxy-5-azacytidine- Nagao M (1997) Establishment of methylation-sensitive-representa- induced demethylation of the hMLH1 gene promoter. Cancer Res 60: tional difference analysis and isolation of hypo- and hypermethylated 6039– 6044 genomic fragments in mouse liver tumors. Proc Natl Acad Sci USA 94: Pommier Y, Sordet O, Antony S, Hayward RL, Kohn KW (2004) Apoptosis 2284– 2289 defects and chemotherapy resistance: molecular interaction maps and Valk-Lingbeek ME, Bruggeman SW, van Lohuizen M (2004) Stem cells networks. Oncogene 23: 2934–2949 cancer; the polycomb connection. Cell 118: 409–418 Richardson BC (2002) Role of DNA methylation in the regulation of cell Warnecke PM, Stirzaker C, Song J, Grunau C, Melki JR, Clark SJ (2002) function: autoimmunity, aging and cancer. J Nutr 132: 2401S–2405S Identification and resolution of artifacts in bisulfite sequencing. Methods Soengas MS, Capodieci P, Polsky D, Mora J, Esteller M, Opitz-Araya X, 27: 101–107 McCombie R, Herman JG, Gerald WL, Lazebnik YA, Cordon-Cardo C, Wei SH, Brown R, Huang TH (2003) Aberrant DNA methylation in ovarian Lowe SW (2001) Inactivation of the apoptosis effector Apaf-1 in cancer: is there an epigenetic predisposition to drug response? Ann N Y malignant melanoma. Nature 409: 207–211 Acad Sci 983: 243– 250 Strathdee G, MacKean MJ, Illand M, Brown R (1999) A role for methylation Wei SH, Chen CM, Strathdee G, Harnsomburana J, Shyu CR, Rahmatpanah of the hMLH1 promoter in loss of hMLH1 expression and drug resistance F, Shi H, Ng SW, Yan PS, Nephew KP, Brown R, Huang TH (2002) in ovarian cancer. Oncogene 18: 2335–2341 Methylation microarray analysis of late-stage ovarian carcinomas Strathdee G, Vass JK, Oien KA, Siddiqui N, Curto-Garcia J, Brown R (2005) distinguishes progression-free survival in patients and identifies Demethylation of the MCJ gene in stage III/IV epithelial ovarian cancer candidate epigenetic markers. Clin Cancer Res 8: 2246– 2252 and response to chemotherapy. Gynecol Oncol 97: 898–903 Xiong Z, Laird PW (1997) COBRA: a sensitive and quantitative DNA Taniguchi T, Tischkowitz M, Ameziane N, Hodgson SV, Mathew CG, Joenje methylation assay. Nucleic Acids Res 25: 2532–2534 H, Mok SC, D’Andrea AD (2003) Disruption of the Fanconi anemia- Yamashita K, Dai T, Dai Y, Yamamoto F, Perucho M (2003) Genetics BRCA pathway in cisplatin-sensitive ovarian tumors. Nat Med 9: supersedes epigenetics in colon cancer phenotype. Cancer Cell 4: 568–574 121–131 Taylor SM, Jones PA (1979) Multiple new phenotypes induced in 10T1/2 Zhang L, Conejo-Garcia JR, Katsaros D, Gimotty PA, Massobrio M, Regnani and 3T3 cells treated with 5-azacytidine. Cell 17: 771–779 G, Makrigiannakis A, Gray H, Schlienger K, Liebman MN, Rubin SC, Teodoridis JM, Hall J, Marsh S, Kannall HD, Smyth C, Curto J, Siddiqui N, Coukos G (2003) Intratumoral T cells, recurrence, and survival in Gabra H, McLeod HL, Strathdee G, Brown R (2005) CpG island epithelial ovarian cancer. N Engl J Med 348: 203–213 British Journal of Cancer (2006) 94(8), 1087 – 1092 & 2006 Cancer Research UK

Journal

British Journal of Cancer – Springer Journals

Published: Feb 21, 2006

There are no references for this article.