Introductory Review Cellscience Reviews Vol.2 No.3 ISSN 1742-8130

p53 and deregulation of DNA methylation in cancer

Shirley M. Taylor

Introduction

Initiation and progression of human tumors is generally believed to be a result of oncogene activation and tumor suppressor gene inactivation by both genetic and epigenetic means. In most tumors studied to date, aberrant DNA methylation patterns include localized increases in CpG island methylation (regional hypermethylation), with concomitant loss of transcriptional activity, accompanied by a global decrease in total methylation (global hypomethylation). When CpG island methylation and transcriptional silencing occurs at tumor suppressor gene loci, a selective growth advantage is conferred, and cells progress to a more malignant phenotype. Different tumor types have significantly different rates and patterns of epigenetic tumor suppressor gene inactivation (Costello et al., 2000), implying that this process is not random, but is perhaps selected for during neoplastic transformation. The underlying cause of aberrant DNA methylation appears intimately connected with the process of neoplastic transformation, rather than a consequence of transformation, although this is still a subject of intense debate. The events that alter the methylation machinery, and thereby raise the rates of CpG island methylation in most human cancers, remain poorly understood. This review explores the possibility that disruption of the p53 pathway is causally related to the loss of regulation of DNA methylation, with its downstream effects on tumor suppressor gene inactivation, genome instability and tumor progression.

Epigenetic alterations in cancer

Epigenetics, a term that describes the stable alteration of gene expression potential, encompasses both DNA methylation and histone modification. Epigenetic processes cooperate to modify either the DNA itself or the proteins that intimately associate with the DNA to alter chromatin configuration, thus establishing the characteristic gene expression profile of individual somatic tissues. DNA methylation is a post-replication addition of methyl groups to cytosine residues in DNA, predominantly in the sequence CpG. In collaboration with the proteins that modify nucleosomes (Wolffe and Matzke, 1999), DNA methylation establishes a silent chromatin state, which is presumably interpreted by proteins that recognize the modified state and bring about the silencing of transcription. Patterns of DNA methylation change dramatically during embryonic development, with demethylation occurring during early cleavage, followed by widespread de novo methylation shortly after implantation (Jaenisch, 1997). This methylation pattern is then fine-tuned during subsequent development to yield a pattern characteristic of a particular tissue. Somatic cells in general are methylated at approximately 80% of their CpG sites. Interspersed in this global methylation are CpG islands, which are short regions relatively rich in CpG dinucleotides that generally remain free of methylation, regardless of gene expression (Bird, 1997). In a few specific cases, CpG islands become methylated during development, leading to permanent silencing of the associated gene (Jones, 1982, Wolf et al, 1984), but how CpG islands escape methylation under circumstances where most of the genome is methylated, remains an open question (Brandeis et al., 1994, MacLeod et al, 1994).

In contrast to somatic cells, cancer cells show a dramatic alteration in the distribution of methyl groups. Globally, most human tumors have a significantly reduced frequency of CpG methylation (global hypomethylation, Feinberg and Vogelstein, 1983, Gama-Sosa et al, 1983), while CpG islands become hypermethylated (regional hypermethylation, Jones, 2002). These epigenetic changes contribute to genome instability with increased loss of heterozygosity (LOH, Chen et al., 1998, Chan et al, 2001), activation of oncogenes (Laird et al., 1995), silencing of tumor suppressor genes (Jones, 2002) and inactivation of DNA repair systems (Esteller et al., 2000, Kane et al, 1997). The phenomenon of regional hypermethylation has received a greater degree of scrutiny than has global hypomethylation, and the list of genes shown to be transcriptionally inactivated in various human tumors continues to grow. Aberrant methylation in cancer has been the subject of many excellent reviews (Ehrlich, 2002, Jones and Baylin, 2002). However, the mechanisms underlying the establishment of abnormal methylation processes, or the deregulation of methylation enzymes which might allow these abnormalities to accumulate, are poorly understood and have been largely absent from this extensive literature. Several studies published in the past few years have, however, begun to shed light on these mechanisms, and have implicated p53 in the control of DNA methylation. This review explores the possibility that disruption of the p53 pathway is causally related to the loss of regulation of DNA methylation, with its downstream effects on tumor suppressor gene inactivation and genome instability.

DNA methyltransferases contribute to the process of oncogenic transformation

To date three enzymatically active DNA methyltransferases have been identified in mammals. The first of these to be cloned and studied in depth was murine Dnmt1 (Bestor), which has become known as the “maintenance methyltransferase” for its predominant role in maintaining the pattern of DNA methylation in somatic cells through successive rounds of DNA replication (Wigler et al., 1981, Taylor & Jones 1982). DNMT3a and 3b are two related enzymes that carry out de novo DNA methylation and play a major role in establishing the somatic pattern of DNA methylation during embryogenesis (Okano, 1999). In addition to these enzymes, DNMT2 (Okano, 1998, Yoder and Bestor, 1998) is perhaps the most highly conserved of all DNMT family members, with DNMT2-like proteins in yeast, plants and Drosphila (Lyko et al., 2000). This protein contains all of the conserved motifs found in the enzymatically active DNMTs, including the critical proline-cysteine dipeptide at the active site (Okano, 1998, Yoder and Bestor, 1998). It was initially thought to have no DNA methyl transfer activity, but more recently has been shown to be active, albeit with a very low turnover number (Hermann et al., 2003, Tang et al 2003, Kunert et al , 2003). DNMT3L (DNMT3-like) was isolated using genome database searches (Aapola, 2000) and lacks most of the motifs required for catalytic activity, including the catalytic PC motif. Despite having no catalytic activity, DNMT3L interacts with DNMT3a to stimulate its activity (Gowher et al.), and is critical in the establishment of imprints during gametogenesis (Hata et al., 2002, Bourc’his and Bestor, 2004). The properties, functions and roles of the various DNA methyltransferases in the etiology of disease states has recently been the subject of a number of excellent reviews (Herman, 2003, Robertson, 2005). Evidence for the involvement of DNA methyltransferases 1 and 3b in the oncogenic process is substantial, as discussed below. Whether the other family members play a role in this disease remains an open question.

Figure 1. DNA Methylation A. Methylation by DNA methyltransferases at CpG islands. B. DNA demethylation relaxes chromatin structure allowing histone acetylation and the binding of transcriptional complexes. C. Tumor cells are characterised by hypermethylation of CpG islands and general DNA hypomethylation.

DNMT1, 3a and 3b are all moderately overexpressed at both the mRNA and protein level in many different tumor types (Robertson et al., 1999, De Marzo et al, 1999), although overexpression of Dnmt3b seems to be more significant (Robertson et al., 1999). Ectopic overexpression of DNMT1 caused neoplastic transformation in murine 3T3 cells (Wu, 1993), and de novo methylation of CpG islands in human fibroblasts (Vertino, 1996). Jaenisch and colleagues (Biniszkiewicz et al., 2002) generated a series of ES cells with a wide range of Dnmt1 expression levels and demonstrated that elevated expression resulted in genomic hypermethylation, loss of imprinting in the Igf2 and H19 loci, although not in other imprinted loci, and embryonic lethality similar to that seen in mice lacking Dnmt1 (Leach et al., 1993). On the other hand, reduced expression of Dnmt1 using antisense technology in murine adrenocortical tumor cells resulted in reduced tumorigenesis in syngeneic mice and in vitro (MacLeod, 1995). Heterozygous Dnmt1 knockout mice, with expression levels approximately half that of wild type mice, significantly reduced Apc^min-induced intestinal neoplasia (Laird et al., 1995). Pharmacological reduction in DNA methylation using 5-aza-deoxycytidine further reduced the incidence of Apc^min-induced tumors in these mice. In mice carrying a hypomorphic allele of Dnmt1 (Dnmt1^Chip/-), which reduces Dnmt1 expression to 10% of wild type levels, the resulting DNA hypomethylation was accompanied by the induction of T cell lymphomas (Gaudet et al., 2003). Thus, removing methyl groups from tumor cells by limiting expression of Dnmt1 minimizes the tumor phenotype, while significant alteration, in either direction, of the level of methylation in normal somatic cells appears to promote tumorigenesis.

Studies in cultured tumor cell lines have yielded somewhat conflicting results. Targeted disruption of DNMT1 or 3b in the human colon carcinoma cell line, HCT116 had little effect on DNA methylation or cell growth (Rhee, 2000, 2002). Simultaneous disruption of both genes was required to obtain a 95% reduction of total methylation, demethylation of repeated sequences, reversal of silencing of the p16 tumor suppressor gene and growth suppression. Thus, it appeared that cooperation between DNMT1 and DNMT3b was required to silence tumor suppressor genes by methylation. Other studies, however, have shown that reduction of DNMT1 alone, by either antisense RNA or RNA interference in various tumor cell lines, including HCT116, resulted in substantial DNA methylation loss and reduction of cell growth (Robert, 2003). Also, targeted deletion of dnmt3b, but not dnmt3a, in murine embryonic fibroblasts (MEFs) resulted in partial loss of DNA methylation, chromosome instability and immortalization (Dodge et al.). These somewhat disparate results likely reflect differences in the cellular context in which the experiments were conducted. Nonetheless, it is probable that deregulated expression of DNMT1 and/or 3b contributes to tumor initiation and progression.

p53 tumor suppressor as a regulator of DNA methylation

The p53 tumor suppressor plays a critical role in the cellular response to genotoxic stress. When activated, p53 is stabilized by posttranslational modification and accumulates in the nucleus, where it acts as both transcriptional activator and repressor. The initial response to p53 activation is the transcriptional activation of p21, engaging cell cycle checkpoints, and the activation of DNA repair pathways (Jin & Levine, 2001). In response to excessive damage, p53 initiates apoptosis through activation of pro-apoptotic genes and repression of anti-apoptotic genes (el-Deiry, 1998).

p53 is the most frequently altered gene in human cancer, with roughly 50% of all types of cancers carrying a mutation in p53, usually accompanied by LOH (Lain & Lane, 2003). Unlike the majority of tumor suppressor genes that are mutated in human tumors, the wild type allele of p53 is seldom transcriptionally silenced through methylation (Agirre et al., 2003), but is lost from the genome through deletions of varying degree on chromosome 17 (Bieche & Lidereau, 1995). Mutations in p53 occur predominantly in the DNA binding domain and many appear to result in gain of function. Familial studies indicate that many of the mutations occurred via C to T transitions at CpG dinucleotides that were methylated (Rideout et al.). Deamination of 5-methyl-cytosine to thymidine can occur either spontaneously or through the action of methyltransferases (Shen et al., 1992) and is thought to be the cause of loss of CpG dinucleotides throughout evolution of the mammalian genome, a phenomenon called CpG suppression (Jones et al., 1992).

Although most studies on the role of p53 as a transcription factor have focused on activated (stabilized) p53 following DNA damage, the tumor suppressor appears to also play a role in transcriptional regulation without activation. In liver, p53 represses αfetoprotein (AFP) expression during development by direct binding to a p53 consensus sequence, apparently in the absence of activation (Lee et al., 1999). In this model, activation of p53 further represses transcription of AFP. The DNMT1 gene is also a target of repression by unactivated or latent p53 (Peterson et al., 2003). In both mouse and human, the promoter region of DNMT1 contains 3 consensus binding sites for p53; the concensus binding site located in exon 1, upstream of the translation start site, binds p53 in vitro and in vivo. Loss of p53 by genetic deletion in either human colon carcinoma cells (HCT116p53-/-) or in primary astrocyte cultures from p53 knockout mice (Donehower, 1992) resulted in release of repression and a dramatic increase in expression of DNMT1 at both the RNA and protein level (Peterson et al., 2003). Interestingly, activation of p53 by genotoxic stress in HCT116 cells resulted in release of p53 from the DNMT1 promoter concomitant with release of repression. This study was the first indication that defects in the p53 pathway might be responsible for elevated expression of DNMT1. In brain tissue from neonatal p53 knockout mice, 2-fold elevated expression of dnmt1 was apparent in the absence of one allele of p53 (p53+/- heterozygotes), while homozygous knockout mice showed a 10-fold overexpression (Peterson, Bogler and Taylor, in preparation). Downstream effects of this overexpression in the neonatal mice included low but detectable levels of methylation in the CpG islands of two tumor suppressor genes, MGMT and TIMP3 (assayed using the PCR-based methylLight technique (Eads et al., 2000) which is capable of detecting changes in as few as 0.1% of the alleles present), along with reduced expression of these genes. No evidence of increased CpG island methylation or reduced expression was found for the p16INK4a tumor suppressor gene. Accompanying these changes was a decrease in PCNA expression, which is believed to interact with Dnmt1 to recruit it to the replication complex (Maga and Hubscher, 2003). It is tempting to speculate that changes in the balance between the enzymes responsible for catalysis of DNA methylation and the proteins that control their accessibility to the appropriate regions of DNA might be extremely important in maintaining the somatic pattern of DNA methylation.

In this study, no evidence was found for global hypomethylation in neonatal tissues. Establishment of primary cultures of p53 heterozygous knockout astrocytes results in spontaneous transformation with a finite frequency and lag time (Yahanda, 1995, Bogler et al, 1999). This transformation is accompanied by spreading of methylation in the CpG islands of the MGMT and TIMP3 tumor suppressor genes, loss of the wild type p53 allele, transcriptional silencing by methylation of the p16INK4a locus and global hypomethylation (Peterson, Bogler & Taylor, in preparation). Thus, haploinsufficiency at the p53 locus appears to be sufficient to initiate CpG island methylation associated with tumor suppressor gene silencing in human cancer. Loss of the second p53 allele leads to more dramatic deregulation of DNA methylation, with widespread CpG island methylation. The two epigenetic changes associated with human tumors (global hypomethylation and regional hypermethylation) are clearly distinct processes that are mechanistically and temporally separable. This model system therefore provides an opportunity to explore the mechanisms governing loss of genomic methylation as well as the accumulation of CpG island methylation.

In a separate study, Park and colleagues studied the profile of DNA methyltransferases and DNA methylation in p53 knockout mice (p53+/+, p53+/- and p53-/-) at 7 weeks of age, which is prior to tumor development (Park et al., 2005b). Moderate increases in Dnmt1 and 3b were observed in liver and thymus of these animals (maximally 2-3-fold in p53-/- tissues), while similar increases in Dnmt3a were seen in liver only. Expression of Dnmt3L in thymus was abolished by loss of one or both p53 alleles. Accompanying these changes in methyltransferase level was an increase in total 5meC content in heterozygous and homozygous tissue, most notably in liver. Methylation-specific PCR (Herman et al., 1996) did not detect gene specific changes in methylation in the CpG islands of tumor suppressor genes p16INK4a, p15INK4b, E-cadherin and Rassf1A. This is perhaps not surprising given the limits of detection of this assay. Hypomethylation of the imprint control region of the oppositely imprinted H19 and Igf2 loci was detected using COBRA analysis (Xiong & Laird, 1997), although the parental methylation pattern of other differentially methylated regions remained intact. It would appear likely, therefore, that normal p53 function is required for the maintenance of DNA methylation patterns in vivo.

Mutant p53 and DNA methylation

As mentioned above, human tumors that carry defects in p53 usually present with a mutant allele accompanied by genomic loss of the wild type allele (LOH). In addition to these tumors, p53 point mutations occur in patients suffering from Li-Fraumeni syndrome, a familial syndrome resulting from germ-line transmission of these mutations and characterized by development of a broad spectrum of tumors (Kleihues et al., 1997). The p53 knockout mouse (Donehower, 1992, Jacks et al, 1994) attempts to model the Li-Fraumeni syndrome in mice. However, the tumor spectrum in p53 knockout mice is largely restricted to sarcomas and lymphomas (Purdie et al., 1994). Other mouse models include the overexpression of mutant p53 transgenes, which generally results in enhanced spontaneous and carcinogen induced tumor formation, thus demonstrating the gain-of-function properties of some p53 mutants (Harvey et al., 1995, Wang et al, 1998). in vitro models created by transfection with expression vectors carrying mutant p53 cDNAs are similarly confounded by overexpression of the mutant transgene. True models of Li-Fraumeni were recently generated in the laboratories of Jacks at MIT (Olive et al., 2004) and Lozano at MD Anderson (Lang et al., 2004). These models result from the targeted insertion of DNA binding domain mutations (R172H and R270H) into an endogenous allele. The resulting backcrossed mice develop a tumor spectrum that is much more representative of the human syndrome. Data presented in these reports provide the strongest evidence to date for gain-of-function properties of mutant p53. Significantly, most tumors that developed in these animals had lost the wild type allele. In addition, over-expression of the mutant allele was seen in tumors but not in the normal stroma surrounding the tumor (Olive et al., 2004), suggesting a loss of proteasomal function in tumor cells that contain high levels of mutant p53.

If one considers the sequence of events that needs to occur in order to generate a tumor carrying a p53 mutant allele with deletion of the wild type allele, it is reasonable to propose that the earliest event is likely to be acquisition of an allele carrying the point mutation. Gain-of-function effects of this mutant p53 (Cadwell & Zambetti, 2001), such as transactivation of novel target genes, novel protein-protein interactions, interference with the normal function of p53 family members, p63 and p73 (Di Como, 1999, Marin et al, 2000) or alterations in epigenetic regulation could then set the stage for loss of the wild type allele through deletion, as well as for epigenetic silencing of other tumor suppressor genes. In proposing a central role for p53 in the maintenance of genomic patterns of methylation, it becomes imperative that we understand the effects of normal expression levels of mutant p53, in the presence of wild type p53, on the proteins involved in epigenetic modification. The Li-Fraumeni mouse models will be critical in pursuing this understanding.

p53 cooperates with DNA methylation to silence gene expression

In addition to a role for p53 in the regulation of expression of DNMT1, several studies have demonstrated physical interaction between p53 and DNA methyltransferases, and between p53 and histone deacetylases and co-repressors on the promoters of genes classically considered to be transcriptionally repressed by p53. Murphy et al (Murphy et al.) reported that p53 physically interacts with histone deacteylases (HDACs), mediated by the co-repressor mSin3a, and that this interaction is essential for repression of Map4 and stathmin by p53. Similarly, activation of p53 leads to down-regulation of the anti-apoptotic gene, surviving via direct interaction with a p53 consensus binding site, and this repression requires the presence of Dnmt1 (Esteve et al., 2005). MASPIN and desmocolin 3 are both transcriptionally activated by wild type p53 (Oshiro et al., 2003), and both promoters undergo CpG methylation and silencing in breast cancer cell lines harboring mutant p53. Restoration of wild type p53 function by overexpression of the wild type cDNA partially reactivated both genes without affecting the methylation of their promoters. Pharmacological inhibition of Dnmt1, in combination with restoration of wild type p53 function resulted in synergistic reactivation to near- normal levels, suggesting cooperation between mutant p53 and cytosine methylation. Finally a recent report suggests that Dnmt3a is capable of interacting with p53 to repress p53-mediated activation of p21, which apparently does not require the methyltransferase activity of Dnmt3a (Wang et al., 2005). Thus, it appears that signaling through p53 is intimately connected with various proteins required for epigenetic modification of gene expression.

Mechanisms for disruption of the p53 pathway

As noted above, mutations in the p53 gene occur in over 50% of all human tumors (Lain and Lane, 2003). Genetic or functional aberrations of genes that act as upstream effectors or negative regulators of p53 function ultimately lead to disabled p53 function and are common features of the remainder of human tumors (Vousden and Lu, 2002). One such candidate is MDM2, a negative regulator of p53 that directly binds to and inhibits p53, regulating its location, stability and activity as a transcription activator and repressor (Michael and Oren, 2003). Knockout of MDM2 in the mouse is lethal prior to implantation, and this lethal phenotype is rescued by knockout of p53 function (Jones et al., 1995, Montes de Oca Luna et al, 1995). Several human tumors have been reported to overexpress MDM2 as a consequence of gene amplification, increased transcript levels or enhanced translation (Momand et al., 2000). In a subset of these tumors. MDM2 overexpression was mutually exclusive to p53 mutation, suggesting that overexpression of MDM2 could substitute for inactivating p53 mutations (Oliner et al., 1992, Leach et al 1993). Levine and colleagues discovered a single nucleotide polymorphism in the promoter of the MDM2 gene that increases the binding affinity of Sp1 for the promoter and dramatically increases MDM2 expression (Bond et al.). The SNP is associated with accelerated tumor formation in sporadic cancers as well as in Li-Fraumeni individuals carrying germline mutation of the p53 gene.

The tumor suppressor p18 interacts directly with ATM/ATR for the activation of p53 in response to DNA damage (Park et al., 2005a), and is required for activation of ATM. Homozygous knockout of p18 in mice is early embryonic lethal, but heterozygous animals developed a broad spectrum of tumors as they aged. Perhaps the most significant finding in this study is the demonstration that p18 is haploinsufficient and expressed at low levels in many tumor cell lines with apparently normal p53, while normal p18 expression levels are seen in cells lacking functional p53. Low expression of p18 abrogates the p53 response to DNA damage due to a defect in the activation of ATM/ATR. This reverse correlation between low expression levels of p18 and functionality of p53 implies that defects in the p53 pathway could be much more frequent than the 50% reported for p53 mutations.

Correlation between defective p53 signaling and altered expression of DNA methyltransferases

The examples cited above illustrate the breadth of molecular defects that impact the p53 pathway, and that are common in human tumors. In order to make a case for an essential role for p53 in maintaining the somatic pattern of DNA methylation, or for the disruption of these patterns by defects in signaling through p53, it is important to understand whether a correlation exists between the p53 status of tumors and the expression patterns of DNA methyltransferases. It is equally important to realize such a relationship is not likely to be a simple correlation between mRNA expression levels and loss of wild type p53 function. Small changes in expression level could well have a dramatic effect on the stability of DNA methylation patterns when examined in the context of proteins that modify DNA methyltransferase activity or the accessibility of substrate DNA within the nucleus. Mutant forms of p53, as well as defects in pathway components upstream of p53 are as likely to impact DNA methylation indirectly as they are to have a direct effect on mRNA or protein expression levels. Nonetheless, the ubiquitous nature of defects in p53 pathway components and in epigenetic modification of the genome of cancer cells leads to the conclusion that both are early events in the process of neoplastic transformation, and are likely to be mechanistically connected regulatory mechanisms.

References

Aapola, U., Kawasaki, K., Scott, H.S., Ollila, J., Vihinen, M., Heino, M., Shintani, A., Minoshima, S., Krohn, K., Antonarakis, S.E., Shimizu, N., Kudoh, J. and Peterson, P. (2000). Isolation and initial characterization of a novel zinc finger gene, Dnmt3L, on 21q22.3, related to the cytosine-5-methyltransferase 3 gene family. Genomics 65, 293-298.

Agirre, X., Vizmanos, J. L., Calasanz, M. J., Garcia-Delgado, M., Larrayoz, M. J., and Novo, F. J. (2003). Methylation of CpG dinucleotides and/or CCWGG motifs at the promoter of TP53 correlates with decreased gene expression in a subset of acute lymphoblastic leukemia patients. Oncogene 22, 1070-1072.

Bestor, T. H., Laudano, A., Mattaliano, R. and Ingram, V. (1988). Cloning and sequencing of a cDNA encoding DNA methyltransferase of mouse cells. Tha carboxy-terminal domain is related to bacterial restriction methyltransferases. J MolBiol 203, 971-983.

Bieche, I., and Lidereau, R. (1995). Genetic alterations in breast cancer. Genes Chromosomes Cancer 14, 227-251.

Biniszkiewicz, D., Gribnau, J., Ramsahoye, B., Gaudet, F., Eggan, K., Humpherys, D., Mastrangelo, M. A., Jun, Z., Walter, J., and Jaenisch, R. (2002). Dnmt1 overexpression causes genomic hypermethylation, loss of imprinting, and embryonic lethality. Mol Cell Biol 22, 2124-2135.

Bird, A. (1997). Does DNA methylation control transposition of selfish elements in the germline? Trends Genet 13, 469-472.

Bond, G. L., Hu, W., Bond, E. E., Robins, H., Lutzker, S. G., Arva, N. C., Bargonetti, J., Bartel, F., Taubert, H., Wuerl, P., et al. (2004). A single nucleotide polymorphism in the MDM2 promoter attenuates the p53 tumor suppressor pathway and accelerates tumor formation in humans. Cell 119, 591-602.

Brandeis, M., Frank, D., Keshet, I., Siegfried, Z., Mendelsohn, M., Nemes, A., Temper, V., Razin, A., and Cedar, H. (1994). Sp1 elements protect a CpG island from de novo methylation. Nature 371, 435-438.

Cadwell, C., and Zambetti, G. P. (2001). The effects of wild-type p53 tumor suppressor activity and mutant p53 gain-of-function on cell growth. Gene 277, 15-30.

Chen, R. Z., Pettersson, U., Beard, C., Jackson-Grusby, L., and Jaenisch, R. (1998). DNA hypomethylation leads to elevated mutation rates. Nature 395, 89-93.

Costello, J. F., Fruhwald, M. C., Smiraglia, D. J., Rush, L. J., Robertson, G. P., Gao, X., Wright, F. A., Feramisco, J. D., Peltomaki, P., Lang, J. C., et al. (2000). Aberrant CpG-island methylation has non-random and tumour-type-specific patterns. Nat Genet 24, 132-138.

Di Como, C. J., Gaiddon, C., Prives, C. (1999). p73 function in inhibited by tumor-derived p53 mutants in mammalian cells. Mol Cell Biol 19, 1438-1449.

Dodge, J. E., Okano, M., Dick, F., Tsujimoto, N., Chen, T., Wang, S., Ueda, Y., Dyson, N., and Li, E. (2005). Inactivation of Dnmt3b in mouse embryonic fibroblasts results in DNA hypomethylation, chromosomal instability, and spontaneous immortalization. J Biol Chem 280, 17986-17991.

Donehower, L. A., Harvey, M., Slagle, B.L., McArthur, M.J., Mongomery, C.A., Butel, J.S. and Bradley, A. (1992). Mice deficient for p53 are developmentally normal but susceptible to spontanteous tumours. Nature 356, 215-221.

Eads, C. A., Danenberg, K. D., Kawakami, K., Saltz, L. B., Blake, C., Shibata, D., Danenberg, P. V., and Laird, P. W. (2000). MethyLight: a high-throughput assay to measure DNA methylation. Nucleic Acids Res 28, E32.

Ehrlich, M. (2002). DNA methylation in cancer: too much, but also too little. Oncogene 21, 5400-5413.

el-Deiry, W. (1998). Regulation of p53 downstream genes. Semin Cancer Biol 8, 345-347.

Esteller, M., Garcia-Foncillas, J., Andion, E., Goodman, S. N., Hidalgo, O. F., Vanaclocha, V., Baylin, S. B., and Herman, J. G. (2000). Inactivation of the DNA-repair gene MGMT and the clinical response of gliomas to alkylating agents. N Engl J Med 343, 1350-1354.

Esteve, P. O., Chin, H. G., and Pradhan, S. (2005). Human maintenance DNA (cytosine-5)-methyltransferase and p53 modulate expression of p53-repressed promoters. Proc Natl Acad Sci U S A 102, 1000-1005.

Feinberg, A. P., and Vogelstein, B. (1983). Hypomethylation distinguishes genes of some human cancers from their normal counterparts. Nature 301, 89-92.

Gaudet, F., Hodgson, J. G., Eden, A., Jackson-Grusby, L., Dausman, J., Gray, J. W., Leonhardt, H., and Jaenisch, R. (2003). Induction of tumors in mice by genomic hypomethylation. Science 300, 489-492.

Gowher, H., Liebert, K., Hermann, A., Xu, G., and Jeltsch, A. (2005). Mechanism of stimulation of catalytic activity of Dnmt3A and Dnmt3B DNA-(cytosine-C5)-methyltransferases by Dnmt3L. J Biol Chem 280, 13341-13348.

Harvey, M., Vogel, H., Morris, D., Bradley, A., Bernstein, A., and Donehower, L. A. (1995). A mutant p53 transgene accelerates tumour development in heterozygous but not nullizygous p53-deficient mice. Nat Genet 9, 305-311.

Hata, K., Okano, M., Lei, H., and Li, E. (2002). Dnmt3L cooperates with the Dnmt3 family of de novo DNA methyltransferases to establish maternal imprints in mice. Development 129, 1983-1993.

Herman, J. G., Graff, J. R., Myohanen, S., Nelkin, B. D., and Baylin, S. B. (1996). Methylation-specific PCR: a novel PCR assay for methylation status of CpG islands. Proc Natl Acad Sci U S A 93, 9821-9826.

Herman, J. G. a. B., S.B. (2003). Gene silencing in cancer in association with promoter hypermthylation. N Engl J Med 349, 2042-2054.

Hermann, A., Schmitt, S., and Jeltsch, A. (2003). The human Dnmt2 has residual DNA-(cytosine-C5) methyltransferase activity. J Biol Chem 278, 31717-31721.

Jaenisch, R. (1997). DNA methylation and imprinting: why bother? Trends Genet 13, 323-329.

Jin, S., and Levine, A. J. (2001). The p53 functional circuit. J Cell Sci 114, 4139-4140.

Jones, P. A., Rideout, W. M., 3rd, Shen, J. C., Spruck, C. H., and Tsai, Y. C. (1992). Methylation, mutation and cancer. Bioessays 14, 33-36.

Jones, P. A., Taylor S.M., Mohandas T, Shapiro L.J. (1982). Cell cycle-specific reactivation of an inactive X-chromosome locus by 5-azadeoxycytidine. Proc Natl Acad Sci USA 79, 1215-1219.

Jones, P. A. a. B., S.B. (2002). The fundamental role of epigenetic events in cancer. Nat Rev Genet 3, 514-528.

Jones, S. N., Roe, A. E., Donehower, L. A., and Bradley, A. (1995). Rescue of embryonic lethality in Mdm2-deficient mice by absence of p53. Nature 378, 206-208.

Kleihues, P., Schauble, B., zur Hausen, A., Esteve, J., and Ohgaki, H. (1997). Tumors associated with p53 germline mutations: a synopsis of 91 families. Am J Pathol 150, 1-13.

Lain, S., and Lane, D. (2003). Improving cancer therapy by non-genotoxic activation of p53. Eur J Cancer 39, 1053-1060.

Laird, P. W., Jackson-Grusby, L., Fazeli, A., Dickinson, S. L., Jung, W. E., Li, E., Weinberg, R. A., and Jaenisch, R. (1995). Suppression of intestinal neoplasia by DNA hypomethylation. Cell 81, 197-205.

Lang, G. A., Iwakuma, T., Suh, Y. A., Liu, G., Rao, V. A., Parant, J. M., Valentin-Vega, Y. A., Terzian, T., Caldwell, L. C., Strong, L. C., et al. (2004). Gain of function of a p53 hot spot mutation in a mouse model of Li-Fraumeni syndrome. Cell 119, 861-872.

Leach, F. S., Tokino, T., Meltzer, P., Burrell, M., Oliner, J. D., Smith, S., Hill, D. E., Sidransky, D., Kinzler, K. W., and Vogelstein, B. (1993). p53 Mutation and MDM2 amplification in human soft tissue sarcomas. Cancer Res 53, 2231-2234.

Lee, K. C., Crowe, A. J., and Barton, M. C. (1999). p53-mediated repression of alpha-fetoprotein gene expression by specific DNA binding. Mol Cell Biol 19, 1279-1288.

Lyko, F., Ramsahoye, B. H., and Jaenisch, R. (2000). DNA methylation in Drosophila melanogaster. Nature 408, 538-540.

MacLeod, A., Szyf M. (1995). Expression of antisense to DNA methyltransferase mRNA induces DNA demethylation and inhibits tumorigenesis. J Biol Chem 270, 8037-8043.

Maga, G., and Hubscher, U. (2003). Proliferating cell nuclear antigen (PCNA): a dancer with many partners. J Cell Sci 116, 3051-3060.

Michael, D., and Oren, M. (2003). The p53-Mdm2 module and the ubiquitin system. Semin Cancer Biol 13, 49-58.

Momand, J., Wu, H. H., and Dasgupta, G. (2000). MDM2--master regulator of the p53 tumor suppressor protein. Gene 242, 15-29.

Murphy, M., Ahn, J., Walker, K. K., Hoffman, W. H., Evans, R. M., Levine, A. J., and George, D. L. (1999). Transcriptional repression by wild-type p53 utilizes histone deacetylases, mediated by interaction with mSin3a. Genes Dev 13, 2490-2501.

Okano, M., Bell, D.W., Haber, D.A. and Li, E. (1999). DNA methyltransferases Dnmt3a and Dnmt3b are essential for de novo methylation and mammalian development. Cell 99, 247-257.

Okano, M., Xie, S. and Li, E. (1998). Cloning and characterization of a family of novel mammalian DNA (cytosine-5) methyltransferases. Nature Genetics 19, 219-220.

Oliner, J. D., Kinzler, K. W., Meltzer, P. S., George, D. L., and Vogelstein, B. (1992). Amplification of a gene encoding a p53-associated protein in human sarcomas. Nature 358, 80-83.

Olive, K. P., Tuveson, D. A., Ruhe, Z. C., Yin, B., Willis, N. A., Bronson, R. T., Crowley, D., and Jacks, T. (2004). Mutant p53 gain of function in two mouse models of Li-Fraumeni syndrome. Cell 119, 847-860.

Oshiro, M. M., Watts, G. S., Wozniak, R. J., Junk, D. J., Munoz-Rodriguez, J. L., Domann, F. E., and Futscher, B. W. (2003). Mutant p53 and aberrant cytosine methylation cooperate to silence gene expression. Oncogene 22, 3624-3634.

Park, B. J., Kang, J. W., Lee, S. W., Choi, S. J., Shin, Y. K., Ahn, Y. H., Choi, Y. H., Choi, D., Lee, K. S., and Kim, S. (2005a). The haploinsufficient tumor suppressor p18 upregulates p53 via interactions with ATM/ATR. Cell 120, 209-221.

Park, I. Y., Sohn, B. H., Choo, J. H., Joe, C. O., Seong, J. K., Lee, Y. I., and Chung, J. H. (2005b). Deregulation of DNA methyltransferases and loss of parental methylation at the insulin-like growth factor II (Igf2)/H19 loci in p53 knockout mice prior to tumor development. J Cell Biochem 94, 585-596.

Peterson, E. J., Bogler, O., and Taylor, S. M. (2003). p53-mediated repression of DNA methyltransferase 1 expression by specific DNA binding. Cancer Res 63, 6579-6582.

Purdie, C. A., Harrison, D. J., Peter, A., Dobbie, L., White, S., Howie, S. E., Salter, D. M., Bird, C. C., Wyllie, A. H., Hooper, M. L., and et al. (1994). Tumour incidence, spectrum and ploidy in mice with a large deletion in the p53 gene. Oncogene 9, 603-609.

Rhee, I., Jair, K-W., Yen, R-W. C., Lenglauer, C., Herman, J.G., Kinzler, K.W., Vogelstein, B., Baylin, S.B. and Schuebel, K.E. (2000). CpG methylation is maintained in human cancer cells lacking Dnmt1. Nature 404, 1003-1007.

Rideout, W. M., 3rd, Coetzee, G. A., Olumi, A. F., and Jones, P. A. (1990). 5-Methylcytosine as an endogenous mutagen in the human LDL receptor and p53 genes. Science 249, 1288-1290.

Robert, M.-F., Morin, S., Beaulieu, N., Gauthier, F., Chute, I.C., Barsalou, A. and MacLeod, R. (2003). DNMT1 is required to maintain CpG methylation and aberrant gene silencing in human cancer cells. Nature Genetics 33, 61-65.

Robertson, K. D., Uzvolgyi, E., Liang, G., Talmadge, C., Sumegi, J., Gonzales, F. A., and Jones, P. A. (1999). The human DNA methyltransferases (DNMTs) 1, 3a and 3b: coordinate mRNA expression in normal tissues and overexpression in tumors. Nucleic Acids Res 27, 2291-2298.

Shen, J. C., Creighton, S., Jones, P. A., and Goodman, M. F. (1992). A comparison of the fidelity of copying 5-methylcytosine and cytosine at a defined DNA template site. Nucleic Acids Res 20, 5119-5125.

Vertino, P. M., Yen, R.C., Gao, J. and Baylin, S.B. (1996). de novo methylation of CpG island sequences in human fibroblasts verexpressing DNA (cytosine-5-) methyltransferase. Mol Cell Biol 16, 4555-4565.

Vousden, K. H., and Lu, X. (2002). Live or let die: the cell's response to p53. Nat Rev Cancer 2, 594-604.

Wang, Y. A., Kamarova, Y., Shen, K. C., Jiang, Z., Hahn, M. J., Wang, Y., and Brooks, S. C. (2005). DNA Methyltransferase-3a Interacts with p53 and Represses p53-Mediated Gene Expression. Cancer Biol Ther 4, 1138-1143.

Wigler, M., Levy, D., and Perucho, M. (1981). The somatic replication of DNA methylation. Cell 24, 33-40.

Wolffe, A. P., and Matzke, M. A. (1999). Epigenetics: regulation through repression. Science 286, 481-486.

Wu, J., Issa, J.. Herman, j.,Bassett DE Jr, Nelkin BD and SB Baylin SB (1993). Expression of an Exogenous Eukaryotic DNA Methyltransferase Gene Induces Transformation of NIH 3T3 Cells. Proc Natl Acad Sci USA 90, 8891-8895.

Xiong, Z., and Laird, P. W. (1997). COBRA: a sensitive and quantitative DNA methylation assay. Nucleic Acids Res 25, 2532-2534.

Yahanda, A. M., Bruner, J. M., Donehower, L. A., and Morrison, R. S. (1995). Astrocytes derived from p53-deficient mice provide a multistep in vitro model for development of malignant gliomas. Mol Cell Biol 15, 4249 - 4259.