DNA Methylation and Cancer1 - cancerres.aacrjournals.orgcancerres.aacrjournals.org/content/canres/46/2/461.full.pdf · DNA METHYLATION AND CANCER dehydrogenase activity and hypermethylation

(CANCER RESEARCH 46, 461-466, February 1986]

Perspectives in Cancer Research

DNA Methylation and Cancer1

Peter A. Jones

University of Southern California Comprehensive Cancer Center, Los Angeles, California 90033

The progressive nature of malignant neoplasms and theirinherent heterogeneity ultimately limit our ability to understandand treat them. The fundamental concepts of tumor progressionelaborated by Foulds (1) and Nowell (2) have as their centralthemes the idea that increasing genetic alterations generated byrandom somatic mutational events are responsible for this heterogeneity and progression of tumor cells to increasingly malignant and less responsive states. However, recent information onthe unexpected flexibility of the eukaryotic genome with regardto gene amplification, rearrangement, deletion, and methylationmakes the concept of somatic mutation-induced tumor progres

sion perhaps too simplistic. For example, not all neoplastic cellsshow a higher mutational frequency than normal cells (3) and therate of metastatic variation can be as much as 2 orders ofmagnitude higher than the rate of mutation in the same cells (4).

Thus, processes such as amplification, rearrangement, deletion, and methylation, which are components of normal cellularontogeny, become increasingly attractive candidates for mediators of tumor cell progression. Importantly, these changes canbe ephemeral in nature and can be used to explain the rapidchanges in metastatic abilities which occur in several experimental systems (5, 6). In this article, I will expand upon the suggestions of several authors (7-10) that changes in DNA methylation,

which are known to be implicated in eukaryotic gene control,may play a central role in the generation of heterogeneity andphenotypic instability in cancer.

DNA Methylation Is Implicated in Eukaryotic Gene Expression

The evidence that the methylation of specific cytosine residuesin DNA controls the expression of some, but not all, eukaryoticgenes has strengthened over the last few years. There areseveral features of eukaryotic DNA methylation which make it aparticularly attractive component of a multilevel control mechanism (8). Methylation patterns are symmetrically distributed inCpG doublets (11) and are tissue specific (12, 13). The methylation status of specific CpG sites is maintained after DNA synthesis by methyltransferase enzymes, a fact that ensures thatmodification patterns, once established, are somatically heritable(14, 15). The existence of tissue specific DNA methylation patterns and a mechanism for their accurate copying are importantin suggesting how this information coding system may be implicated in differentiation. Thus, a pattern once established can befaithfully copied so that differentiated cells breed true.

The existence of methyl groups on CpG doublets is known tohave profound effects on the expression of eukaryotic genes. A

Received 8/6/85; revised 10/11/85; accepted 10/14/85.1This work was supported by grants CA 40422 and CA 39913 from the National

Cancer Institute.

large number of studies have shown that active genes arehypomethylated in expressing cells whereas they are methylatedin other cell types in which they are not being transcribed (16).These experiments do not, in themselves, show a cause andeffect relationship between methylation and gene expression,but several elegant experiments have indicated that the in vitromethylation of genes before their introduction into eukaryoticcells results in transcriptional inactivity (17, 18). These studieshave also given strong support to the idea that the methylationof the 5' region of the gene may be more important in silencing

genes than methylation of other gene regions (19) and that geneswhich are not active within cells become de novo methylated(20).

Not all eukaryotic genes seem to be regulated by DNA methylation (21). Well known exceptions to the general rule thathypomethylation is associated with gene expression are rRNAand vitellogenin genes which are expressed from heavily methylated configurations (22, 23). Also, changes in methylationstatus do not always occur during alterations in gene expression(24). These experiments do, however, have the acknowledgedweakness that only a small percentage of total CpG sites wereexamined due to the sequence specificities of restriction enzymes used for the analysis. There is therefore a danger indrawing global conclusions from such a limited analysis. However, the prevailing feeling in the field is that DNA methylationrepresents one part of a multilevel control mechanism in eukaryotic cells with undermethylation being a necessary but notsufficient condition for the activation of some but not all genes.

Methylation Domains

Initial experiments in the area of DNA methylation examinedthe methylation status of genes on a rather global basis. However, considerable excitement has recently been generated bythe realization that potential methylation sites are not randomlydistributed throughout the genome but often appear in highlyCpG enriched domains or clusters which are associated withgenes. Bird ef al. (25) have estimated that about 30,000 suchCpG rich clusters are present per haploid genome in the mouse.Clusters rich in CpG are found in the major histocompatibilitycomplex (26), Â«2(1) collagen (24), glucose-6-phosphate dehy-drogenase (27), adenosine phosphoribosyltransferase (28), di-

hydrofolate reducÃase (28, 29), hypoxanthine phosphoribosyltransferase (30), and metallothionein HA (31) genes. The CpGclusters appear to have been protected from methylation andare often found in the 5' region of genes.

The exception to the general rule that the housekeeping geneclusters are unmethylated occurs when they are located oninactive X-chromosomes. Wolf ef al. (27) have recently found acomplete concordance between lack of glucose-6-phosphate

CANCER RESEARCH VOL. 46 FEBRUARY 1986

461

on May 18, 2018. © 1986 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

DNA METHYLATION AND CANCER

dehydrogenase activity and hypermethylation of CpG clusterson inactive X-chromosomes. Clusters on hypoxanthine phos-

phoribosyltransferase are undermethylated and hypersensitiveto nuclease digestion on the active X-chromosome whereas thesame clusters are hypermethylated on inactive X-chromosomes(32). Interestingly, CpG rich clusters are not invariably located 5'

to the coding regions of genes; those associated with g6pdoccur in the 3' region of the gene. Assuming that this CpG

cluster is involved in regulating expression of g6pd, its positionsuggests the possibility that CpG clusters function as enhancer-

like elements. Since methylation of CpG rich clusters prior tointroduction into eukaryotic cells can inactivate housekeepinggenes (19) we are left with the exciting concept that methylationof CpG sites in enhancers might block enhancing activity.

The existence of CpG rich clusters in association with housekeeping genes and their undermethylation when associated withactive rather than with inactive X-chromosomes are powerful

evidence that these sequences have an important role in geneexpression. Also, the fact that housekeeping genes can beinactivated if the clusters are methylated before introduction intoeukaryotic cells (19) is further evidence for their potential importance. However, they are not only confined to housekeepinggenes since a cluster occurs on the 2(1) collagen gene (24) whichis not generally considered to fulfill a housekeeping function.

5-Azacytidine

Another series of studies which strongly supports a causativerole for methylation in suppressing gene activity comes fromexperiments using the nucleoside analogue 5-aza-Cyd.2 5-aza-

Cyd was originally developed as a cancer chemotherapeuticagent (33) but also has marked effects on the stability of thedifferentiated state of cultured cells (34, 35). The analogue isthought to act by incorporation into DNA where it functions as apowerful inhibitor of the methylation of newly incorporated cy-

tosine residues (36). It is quite clear that the fraudulent nucleosidemust be incorporated into DNA in order to inhibit methylation(37, 38). The result of this incorporation is a loss of active DNAmethyltransferase activity (38-40) which is due to the formation

of a tight noncovalent complex (40) or a covalent linkage (41)between the enzyme and 5-azacytosine residues in DNA.

Several lines of evidence support the idea that inhibition ofDNA methylation is indeed the mechanism of drug action: (a)biologically active doses of 5-aza-Cyd are strong inducers of

gene expression and potent inhibitors of DNA methylation; (b)the effects on cell differentiation are specific for position 5 ofcytosine and can be mimicked by other analogues such aspseudoisocytidine and 5-fluorodeoxycytidine which also inhibit

methylation and change the differentiated state of cells (36); (c)the changes in gene expression are often heritable for manygenerations in the absence of further drug treatment; (d) theactivation frequencies observed are sometimes 5-6 orders ofmagnitude greater than those expected for the activity of muta-genie agents (42); (e) the effects of 5-aza-Cyd on methylation

have been localized to genes which become transcriptionallyactivated (43); (/) genes which have been inactivated by methylation before introduction into eukaryotic cells can be reactivated by 5-aza-Cyd treatment (44); (g) in an experiment which

2The abbreviation used is: 5-aza-Cyd, 5-azacytidine.

is formally analogous to the classic study of Avery ef al. (45)showing that DNA was the carrier of genetic information, severalinvestigators have shown that the heritable changes induced by5-aza-Cyd are localized to the level of DNA. Thus, DNA extractedfrom cells reactivated by 5-aza-Cyd can successfully transfect

recipient cells with newly acquired characteristics (46,47). Sincethese experiments were conducted with purified naked DNAmolecules they demonstrate that the expression of a gene canbe regulated by modification of DNA. While the transfectionexperiments do not exclude the possibility that other covalentchanges such as rearrangements occurred, they are highlysuggestive of an important role for methylation in suppressinggene activity.

How Well Are Methylation Patterns Copied?

The presence of maintenance DNA methyltransferases ensures that established methylation patterns can be faithfullycopied. The existence of tissue specific patterns of methylationimplies that this is the case during normal development. However, the mechanisms governing the acquisition of new methylgroups or the loss of preexisting methyl groups are not wellunderstood. What has become clear in recent years is thatsubstantial changes in genomic methylation including loss ofmethyl groups and de novo methylation may occur in culturedcells. If DNA methylation plays an important role in controllinggene expression then it is important that we understand howmethylation patterns are changed and altered during normalcellular development and possibly during tumor formation andprogression.

Early experiments using DNA molecules transfected into recipient cells suggested that methylation patterns could be inheritedwith a high (48, 49) but not absolute fidelity (15). Thus, methylation patterns in rapidly dividing cells may not be copied withcomplete accuracy. Shmookler-Reis and Goldstein (50,51 ) foundconsiderable variability in DNA methylation patterns during theserial passage of human diploid fibroblasts and considerableinterclonal variations in methylation patterns for expressed andnonexpressed genes in 8 clones isolated from a mass culture ofhuman diploid fibroblasts. Different clone specific patterns ofDNA methylation including increased methylation were foundindicating a striking degree of interclonal heterogeneity, particularly for those genes not normally expressed in diploid fibroblasts(e.g., y-globin and B-globin). Considerable heterogeneity withregard to X-chromosome DNA methylation has been found in

normal euploid human cells during replication in culture (52).These changes in methylation patterns seen in cultured cellsmay have important implications in the generation of new cellulardiversity.

We have found considerable decreases in DNA methylation inaging but not in immortal cells in culture (53). When normaldiploid fibroblasts from mice, hamsters, and humans were grownin culture, DNA methylation decreased markedly with the greatest rate of loss of 5-methylcytosine observed in mouse cells

which survived the least number of divisions in culture. On theother hand, the immortal cell lines had more stable rates ofmethylation.

These data therefore show that methylation patterns are notinherited with the high degree of fidelity which was previouslythought. The mechanisms for the selective loss of methyl groups


462




during differentiation or for the generation of these abnormalnew methylation patterns are not known. Demethylation couldoccur by an active or passive process, but the prevailing feelingin the field is that it is the failure to methylate during critical celldivisions which is responsible for hypomethylation. Earlier reportson the evidence of demethylating enzymes (54) have not beenconfirmed and the removal of the methyl group from the 5'

position of the cytosine ring may be highly unfavorable from athermodynamic standpoint.

Hemimethylated sites are very inaccessible to the methylatingenzyme when complexed with histones or in the form of nucleo-

somes (55, 56). Thus, if modification is not completed beforeDMA is complexed with chromatin proteins, hypomethylation mayresult. We have recently found that the level of extractablemethyltransferase in cultured mouse cells drops with passage(57) which might explain the lowered methylation level seen asmouse cells age in culture (53).

In addition to the decreases in DMA methylation which occurduring development or propagation of cells in culture, de novomethylation occurs in many physiological situations and also incultured cells. The clearest example of the occurrence of denovo methylation as a physiological function is the de novomethylation of DMA during spermatogenesis (58). This de novomethylation proceeds so that regions within and around genesbecome methylated but excludes the methylation of sequenceswithin nuclease hypersensitive sites and may therefore play animportant role in the activation of the paternal genome duringembryogenesis. There is also considerable evidence that de novomethylation of viral genomes occurs in embryonic cells (reviewedin Ref. 20) and can extend to flanking host sequences (59). Thede novo methylation is seen not only with exogenously addedDNA but also with cellular sequences during embryonic development, an example being mouse repetitive sequences (60).

The de novo methylation which occurs with viral sequences inembryonic cells is rapid but a slower methylation process canalso take place during the passage of cultured cells. For example,Gasson et al. (61) demonstrated spontaneous de novo methylation in a T-lymphoid cell line previously treated with 5-aza-Cyd

to generate glucocorticoid sensitivity. This de novo methylationwas accompanied by the acquisition of the glucocorticoid resistant phenotype. We have also observed considerable de novomethylation in mouse cell lines with artificially depressed 5-

methylcytosine values (62). Some of the phenotypic changesinduced by 5-aza-Cyd appear to be ephemeral in nature (e.g.,

Refs. 61 and 63) and we have observed a metastable acquisitionof the muscle phenotype in myogenic cell lines derived after 5-aza-Cyd treatment.3

Taken together, these data show that considerable under-

methylation and remethylation of genes can occur in normaldevelopment and in culture environments. These changes mayreflect the kinds of alterations needed to occur in genomic DNAduring differentiation. However, abnormal control of methylationin cancer cells may result in the generation of random modification patterns which may serve ultimately to unleash new genesfor transcription. The fluctuations in methylation patterns observed in tumor cells and normal cells in culture may thereforebe a molecular manifestation of phenotypic heterogeneity.

3 L. Liu. M. S. Baker, and P. A. Jones. Characterization of myogenic cell linesderived by 5-azacytidine treatment, submitted for publication.

Methylation Changes in Cancer Cells

A large body of evidence shows that methylation levels andpatterns are deranged in tumor cells (reviewed in Ref. 8). Under-methylation of DNA has been seen in a large number of animaland human tumor cells (64-66) and might contribute to aberrant

gene expression. However, hypomethylation in the DNA of tumors freshly excised from children with a variety of neoplasmsis not always observed (67) although some tumors, particularlyneuroblastomas, were significantly hypomethylated compared tohuman fibroblasts. These results suggest that significantchanges in the overall level of DNA methylation can occur intumors although it is not possible to generalize these findings.The relationship between decreased levels of overall DNA methylation and gene expression is not known, although it is clear,but not widely recognized, that tissue specific differences inoverall genomic methylation levels exist (68).

Substantial hypomethylation of specific sites within growthhormone and globin genes in primary human colonie tumors wasobserved in 4 of 5 tumors by Feinberg and Vogelstein (69).These studies, which were conducted using colon carcinomacells, were particularly significant because of the availability ofnormal mucosal tissue in the vicinity of the tumor which couldeasily be separated from the underlying stroma. Since the neoplasms themselves were derived from the epithelial cell layer, itwas possible to conduct carefully controlled studies when comparing the normal cells to the tumor cells.

Feinberg and Vogelstein (70) extended their studies to the rasgene family in primary colon and lung carcinomas and observedundermethylation in 6 of 8 tumors relative to normal coloniemucosa. Interestingly, the degree of hypomethylation appearedto be greater in metastatic nodules than in the correspondingprimaries suggesting that progressive undermethylation mightoccur during the metastatic process. This was extended in themore recent study by Goelz ef al. (71) who observed altered 5-methylcytosine patterns in a series of genes in 23 neoplasms.Hypomethylation of a specific site in the c-myc gene in human

tumor cell lines has also been observed (72) and changes inmethylation within specific genes have also been observed inanimal tumor cells (73, 74). However, several of the genesinvestigated in these studies were not expressed even thoughthey were hypomethylated (71). Future work will therefore haveto concentrate on genes known to be involved in transformationand progression before the exact significance of methylationchanges can be understood.

Nevertheless, these studies demonstrate that considerableheterogeneity of methylation patterns occurs in tumors derivedfrom humans and animals and in cultured cell lines. This suggeststhat methylation patterns are not as rigorously controlled in thesecells as they are in their normal counterparts but says nothingas to the potential relevance of these findings to the phenomenonof tumor progression. More direct evidence that the 2 phenomena are linked comes from studies in which methylation patternswithin tumor cells have been artificially changed and these experiments suggest a strong causal relationship between methylation changes and increased phenotypic diversity in tumor cells.

Changes in Malignant Behavior Induced by 5-Azacytidine

5-aza-Cyd is a useful tool for assessing the role of epigenetic

phenomena in cancer because several studies have shown that


463




the agent is not demonstrably mutagenic in eukaryotic cells (75-

77). The drug is capable of the direct transformation of culturedcells (78, 79) and is tumorigenic in whole animals (80). However,recent excitement has been generated by the finding that 5-aza-

Cyd can have substantial effects on malignant behavior.Frost et al. (81) found that 5-aza-Cyd influenced the expression

of tumor antigens on tumorigenic mouse cell lines which resultedin some of them becoming nontumorigenic in syngeneic animals.These experiments were expanded in a later study by Kerbel efal. (77) who found that phenotypically unstable clones of meta-

static tumor cells could be obtained from the TA3 mammarycarcinoma cell line which in itself is nonmetastatic. A key pointin this study was the fact that the cell lines derived were notcompletely stable with regard to their phenotype, which raisedthe possibility that epigenetic rather than genetic mechanismswere responsible for the fluctuation in biological behavior. Ols-son's group (82, 83) have also found that 5-aza-Cyd can induce

metastatic potential in nonmetastatic sublines or extinguish met-

astatic potential in cells already metastatic.These results therefore show that 5-aza-Cyd can have variable

effects on tumor progression and may activate genes necessaryfor progression in some cases but serve to activate genes whichrepress progression in others. A recent paper suggests that thismight be the case in rat embryo cells transformed by adenovirustype 5 (84). Progression was not correlated with major changesin the pattern of integration of viral DNA sequences but wasassociated with an increased methylation of integrated viralsequences other than those corresponding to the transforminggenes of the virus. A single exposure of progressed cells to 5-aza-Cyd resulted in the stable reversion to the unprogressed

state of the original parental clone. The observations in the viralsystem thus suggest that progression is a reversible process,which might be associated with changes in the state of methylation of one or more specific genes.

Before leaving the subject of the effects of 5-aza-Cyd on

tumorigenicity and metastatic potential it should also be remembered that the drug can induce the end-stage differentiation of

tumor cells. Examples of this are the formation of contractilemuscle cells from tumorigenic transformed derivatives of 10T1/2cells (85) and the end-stage differentiation of Friend cells (40).

The differentiation may result in the formation of cells with nofurther division potential so that changes in DNA methylationpatterns can result in the apparently normal differentiation ofmalignant cells.

Summary

The main thrusts of the arguments that aberrant DNA methylation is involved in the generation of tumor heterogeneity andprogression can be summarized as follows. The methylation ofspecific cytosine residues in DNA is certainly an important component in multilevel gene control in eukaryotes. The discovery ofCpG clusters in the flanking regions of genes and their under-

methylation on housekeeping genes, except those located oninactive X-chromosomes, strongly suggests a controlling func

tion for modification in these regions. Since methylation plays animportant role in controlling normal cellular development, it follows that aberrations within this mechanism may be implicatedin the abnormal gene control which characterizes cancer.

Methylation patterns are not copied rigorously in rapidly divid

ing cells. This may be because there is normally a close coordination between DNA synthesis, DNA methylation, and DNApackaging, and changes in the timing of these processes couldconceivably result in hypomethylation at some sites and de novomethylation at others. Since the greatest variability of methylationpatterns is seen in nonexpressed genes, it is possible that thereis a tendency for cells to activate genes when dividing in aninappropriate growth environment.

The constant evolution and shuffling of methylation patternswhich occur during division might play a role in the developmentof new phenotypes within cell populations. One might predictthat selective pressures within the host would select for thosecells with specific new methylation patterns allowing for theexpression of genes necessary for survival in a particular environment. Many experiments have in fact shown that methylationlevels and patterns and indeed methyltransferase levels (57) arealtered in cancer cells. Thus, there is considerable heterogeneitywithin tumor populations with regard to this fundamental biological control mechanism. The fact that direct intervention by theuse of 5-aza-Cyd can result in dramatic alterations in malignant

potential allows this hypothesis to be tested more critically.Hopefully, the use of 5-aza-Cyd in defined systems will allow

us to isolate genes which might become activated by drugtreatment and which might contribute to metastatic potential. Anunderstanding of the fundamental aspects of the enzymologyand control of DNA methylation might therefore allow us to makesignificant inroads into understanding how heterogeneity is generated and what we might do about it.

References

1. Foulds, L. Neoplastic Development. New York: Academic Press, 1975.2. Nowell, P. C. The clonal evolution of tumor cell populations. Science (Wash.

DC), 794: 23-28, 1976.3. Elmore, K., Kakunaga, T., and Barrett, J. C. Comparison of spontaneous

mutation rates of normal and chemically transformed human skin fibroblasts.Cancer Res., 43: 1650-1655,1983.

4. Rubin, H. Cancer as a dynamic developmental disorder. Cancer Res., 45:2935-2942, 1985.

5. Kerbel, R. S., Frost, P., and Ã¼teplo,R. G. Genetic and epigenetic regulationsof the metastatic phenotype: a basis for resolving the controversy regardingits selective or random nature and variable phenotypic stability. In: L. Liotta,C. Rabson, and B. Chambers (eds.), Biochemistry and Molecular Genetics.The Hague: Martinus Nijhoff, 1985.

6. Chambers, A. F., Harris, J. F., Ling, V., and Hill, R. P. Rapid phenotypevariation in cells derived from lung mÃ©tastasesof KHT fibrosarcoma. InvasionMetastasis, 4: 255-237,1984.

7. Holliday. R. A new theory of carcinogenesis. Br. J. Cancer, 40: 513-522,1979.8. Riggs, A. D., and Jones, P. A. 5-Methylcytosine, gene regulation and cancer.

Adv. Cancer Res., 40:1-30, 1983.9. Nyce, J. W., Weinhouse, S., and Magee, P. N. 5-Methylcytosine depletion

during tumor development: an extension of the miscoding concept. Br. J.Cancer, 48: 463-475,1983.

10. Frost, P., and Kerbel, R. S. On a possible epigenetic mechanism(s) of tumorcell heterogeneity: the role of DNA methylation. Cancer Metastasis Rev., 2:375-378, 1983.

11. Bird, A. P. Use of restriction enzymes to study eukaryotic DNA methylation. II.The symmetry of methylated sites supports semi-conservative copying of themethylation pattern. J. Mol. Biol., 778: 49-60, 1978.

12. Waalwijk, C., and Flavell, R. A. DNA methylation at a CCGG sequence in thelarge intron of the rabbit B-globin gene: tissue specific variations. NucÃ.AcidsRes., 5: 4631-4641, 1981.

13. McGhee, J. D., and Ginder, G. D. Specific DNA methylation sites in the vicinityof the chicken B-globin genes. Nature (Lond.), 280: 419-420, 1979.

14. Pollack, Y., Stein, R., Razin, A., and Cedar, H. Methylation of foreign DNAsequences in eukaryotic cells. Proc. Nati. Acad. Sci. USA, 77: 6463-6467.1980.

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

16. Doerfler, W. DNA methylation and gene activity. Annu. Rev. Biochem., 52: 93-124, 1983.


464




17. Busslinger, M., Hurst, J., and Flavell, R. A. DNA methylation and the regulationof globin gene expression. Cell, 34: 197-206, 1983.

18. Langer, K-D., Vardimon, L., Renz, D., and Doerfler, W. DMA methylation ofthree 5'-CCGG-3' sites in the promoter and 5' region inactivate the E2a gene

of adenovirus type 2. Proc. Nati. Acad. Sci. USA, 81: 2950-2954, 1984.19 Keshet, I., Yisraeli, J., and Cedar, H. Effect of regional DMA methylation on

gene expression. Proc. Nati. Acad. Sci. USA, 82: 2560-2564, 1985.20 Jaenisch, R., and Jahner, D. Methylation, expression and chromosomal posi

tion of genes in mammals. Biochim. Biophys. Acta, 782: 1-9, 1984.21. Bird, A. P. DNA methylationâ€”how important in gene control? Nature (Lond.),

308. 503-504, 198422. Macleod, D., and Bird, A. P. Transcription in oocytes of highly methylated

rDNA from Xenopus laevis sperm. Nature (Lond.), 306: 200-203, 1983.23. Gerber-Huber, S., May, F. E. B., Westley, B. R., Felber, B. K., Hosbach, H. A.,

Andres, A-C., and Ryffel, G. U. In contrast to other Xenopus genes theestrogen-inducible vitellogenin genes are expressed when totally methylated.Cell, 33. 43-51, 1983.

24. McKeon, C., Ohkubo, H., Pastan, I., and deCrombruggle, B. Unusual methylation pattern of the Â«2(1) collagen gene. Cell, 29: 203-210, 1982.

25. Bird, A., Taggart, M., Frommer, M., Miller. 0. J., and Macleod, D. A fraction ofthe mouse genome that is derived from islands of nonmethylated, CpG-richDNA. Cell, 40:91-99, 1985.

26. Tykocinski, M. L., and Max, E. E. CG dinucleotide clusters in MHC genes and5' demethylated genes. NucÃ.Acids Res., 12: 4385-4396, 1984.

27 Wolf, S. F., Dintzis. S., Tortolo, D.. Persico, G., Lunnen, K. D., Axelman, J.,and Migeon, B. R. Complete concordance between glucose-6-phosphatedehydrogenase activity and hypomethylation of CpG clusters: implications forX-chromosome dosage compensation. NucÃ.Acids Res., 12: 9333-9348,

1984.28. Stein, R., Sciaky-Gallili. N., Razin, A., and Cedar, H. Pattern of methylation of

two genes coding for housekeeping functions. Proc. Nati. Acad. Sci. USA, 80:2422-2426,1983.

29. Shimada, T., and Nienhuis, A. W. Only the promoter region of the constitutivelyexpressed normal and amplified human dihydrofolate reducÃasegene is DNAseI hypersensitive and undermethylated. J. Biol. Chem., 260. 2468-2474,1985.

30 Melton, D. W., Konechki, D. S., Brennard, J.. and Caskey, C. T. Structure,expression and mutation of the hypoxanthine phosphoribosyl transferase gene.Proc. Nati. Acad. Sci. USA, 81: 2147-2151, 1984.

31. Karin. M., and Richards. R. I. Human metallothionein genesâ€”primary structureof the metallothionein ll-gene and a related processed gene. Nature (Lond.),299:797-802, 1982.

32. Wolf. S. F., and Migeon, B. R. Clusters of CpG dinucleotides implicated bynuclease hypersensitivity as control elements of housekeeping genes. Nature(Lond.), 314: 467-469, 1985.

33. Vesely, J., and Cihak, A. 5-Azacytidine: mechanism of action and biologicaleffects in mammalian cells. Pharmacol. Ther., 2: 813-840, 1978.

34 Constantinides, P. G.. Jones, P. A., and Gevers, W. Functional striated musclecells from non-myoblast precursors following 5-azacytidine treatment. Nature(Lond.). 267. 364-366, 1977.

35. Jones, P. A. Altering gene expression with 5-azacytidine. Cell, 40: 485-486,

1985.36. Jones, P. A., and Taylor, S. M. Cellular differentiation, cytidine analogs and

DNA methylation. Cell, 20: 85-93, 1980.37. Jones, P. A., and Taylor, S. M. Hemimethylated duplex DNAs prepared from

5-azacytidine treated cells. NucÃ.Acids Res., 9. 2933-2947.1981.

38. Taylor, S. M., and Jones, P. A. Mechanism of action of eukaryotic DNAmethyltransferase: use of 5-azacytosine containing DNA. J. Mol. Biol., )62:679-692, 1982.

39. Creusot, F., Acs, G., and Christman, J. K. Inhibition of DNA methyltransferaseand induction of Friend erythroleukemia cell differentiation by 5-azacytidineand 5-aza-2'-deoxycytidine. J. Biol. Chem., 257: 2041-2048, 1982.

40. Christman, J. K., Schneiderman, N., and Acs, G. Formation of highly stablecomplexes between 5-azacytosine-substituted DNA and specific non-histonenuclear proteins. J. Biol. Chem., 260: 4059-4068,1985.

41. Santi, D. V., Norment, A., and Garrett, C. E. Covalent bond formation betweena DNA-cytosine methyltransferase and DNA containing 5-azacytosine. Proc.Nati. Acad. Sci. USA, 81: 6993-6997,1984.

42. Harris, M. Induction of thymidine kinase in enzyme-deficient Chinese hamstercells. Cell, 29: 483-492, 1982.

43. Groudine, M., Eisenman, R., and Weintraub, H. Chromatin structure of endogenous retroviral genes and activation by an inhibitor of DNA methylation.Nature (Lond.). 292:311-317, 1981.

44. Simon, D., Stuhlman, H., Jahner, D., Wagner. H., Werner, E., and Jaenisch,R. Retrovirus genomes methylated by mammalian but not bacterial methylaseare noninfectious. Nature (Lond.). 304: 275-277, 1983.

45. Avery, O. T., MacLeod, C. M., and McCarty, M. Studies on the chemical natureof the substance inducing transformation of pneumococcal types. Induction oftransformation by a desoxyribonucleic acid fraction isolated from pneumococ-cus type III. J. Exp. Med.. 79: 137-158, 1944.

46. Venolia, L.. Gartler, S. M., Wassman, E. R., Yen, P.. Mohandas, T., andShapiro. L. J. Transformation with DNA from 5-azacytidine reactivated X-chromosomes. Proc. Nati. Acad. Sci. USA, 79: 2352-2354, 1982.

47. Lester, S. C., Korn, N.J., and DeMars, R. Derepression of genes on the humaninactive X-chromosome: evidence for differences in locus-specific rates ofderepression and rates of transfer of active and inactive gene after DNA-mediated transformation. Somatic Cell Genet., 8: 265-284,1982.

48. Stein, R., Gruenbaum. Y., Pollack. Y., Razin, A., and Cedar, H. Clonal inheritance of the pattern of DNA methylation in mouse cells. Proc. Nati. Acad. Sci.USA, 79. 61-65. 1982.

49. Harland, R. M Inheritance of DNA methylation in microinjected eggs ofXenopus laevis. Proc. Nati. Acad. Sci. USA, 79: 2323-2327,1982.

50. Shmookler-Reis, R. J.. and Goldstein, S. Variability of DNA methylation patterns during serial passage of human diploid fibroWasts. Proc. Nati. Acad. Sci.USA, 79. 3949-3953,1982.

51. Shmookler-Reis, R. J.. and Goldstein. S. Interclonal variation in methylationpatterns for expressed and non-expressed genes. NucÃ.Acids Res., 70: 4293-4304, 1982.

52. Wolf, S.F., and Migeon, B. R. Studies of X-chromosome DNA methylation innormal human cells. Nature (Lond.), 295: 667-671, 1982.

53. Wilson, V. L., and Jones, P. A. DNA methylation decreases in aging but not inimmortal cells. Science (Wash. DC), 220: 1055-1057,1983.

54. Gjerset, R. A., and Martin, D. W. Presence of a DNA demethylating activity inthe nucleus of murine erythroleukemia cells. J. Biol. Chem., 257: 8581-8583,1982.

55. Creusot, F., and Christman, J. K. Localization of DNA methyltransferase in theChromatin of Friend erythroleukemia cells. NucÃ.Acids Res., 9: 5359-5381.1981.

56. Kautiainen, T. L.. and Jones, P. A. Effects of DNA binding proteins on DNAmethylation in vitro. Biochemistry. 24: 1193-1196. 1985.

57. Kautiainen, T. L.. and Jones, P. A. DNA methyltransferase levels in tumorigenicand nontumorigenic cells in culture. J. Biol. Chem., in press, 1985.

58. Groudine, M., and Conklin, K. F. Chromatin structure and de novo methylationof sperm DNA: implication for activation on the paternal genome. Science(Wash. DC), 228: 1061-1068,1985.

59. Jahner, D., and Jaenisch, R. Retrovirus-induced de novo methylation of flankinghost sequences correlates with gene inactivity. Nature (Lond.), 375: 594-597,1985.

60. Chapman, V., Forrester. C., Sanford, J., Hastie, N.. and Rossant, J. Celllineage-specific undermethylation of mouse repetitive DNA. Nature (Lond.),307:284-286, 1984.

61. Gasson, J. C.. Ryden, T., and Bourgeois, S. Role of de novo DNA methylationin the glucocorticoid resistance of T-lymphoid cell line. Nature (Lond.). 302:621-623, 1983.

62. Flatau, E., Gonzales, F. A., Michalowsky, L. A., and Jones, P. A. DNAmethylation in 5-aza-2'-deoxycytidine resistant variants of C3H10TV2 CI 8

cells. Mol. Cell. Biol., 4: 2098-2102, 1984.

63. Compere, S. J., and Palmiter, R. D. DNA methylation controls the inducibilityof the mouse metallothionein-l gene in lymphoid cells. Cell, 25: 233-240,1981.

64. Lapeyre. J-N., Walker, M.S., and Becker, F. F. DNA methylation and methylaselevels in normal and malignant mouse hepatic tissues. Carcinogenesis (Lond.),2:873-878, 1981.

65. Diala, E. S.. Cheah, M. S. C., Rowitch, D., and Hoffman. R. M. Extent of DNAmethylation in human tumor cells. J. Nati. Cancer Inst., 71: 755-764. 1983.

66. Gama-Sosa. M. A., Slagel, V. A., Trewyn, R. W.. Oxenhandler, R., Kuo, K.,Gehrke, W., and Ehrlich, M. The 5-methylcytosine content of DNA from humantumors. NucÃ.Acids Res., 11: 6883-6894, 1983.

67. Flatau, E.. Bogenmann, E.. and Jones. P. A. Variable 5-methylcytosine levelsin human tumor cell lines and fresh pediatrie tumor expiants. Cancer Res.. 43:4901-4905. 1983.

68. Ehrlich, M., Gama-Sosa, M A., Huang, L-H., Midgett, R. M., Kuo, K. C..McCune, R. A., and Gehrke, C. Amount and distribution of 5-methylcytosinein human DNA from different types of tissues or cells. NucÃ.Acids Res., 70:2709-2721,1983.

69. Feinberg. A. P., and Vogelstein, B. Hypomethylation distinguishes genes ofsome human cancers from their normal counterparts. Nature (Lond.), 307: 89-91, 1983.

70. Feinberg, A. P., and Vogelstein, B. Hypomethylation of ras oncogenes inprimary human cancers. Biochem. Biophys. Res. Commun., 777:47-54,1983.

71. Goelz, S. E., Vogelstein, B., Hamilton, S. R., and Feinberg, A. P. DNA frombenign and malignant human colon neoplasms is hypomethylated. Science(Wash. DC), 228: 187-190, 1985.

72. Cheah, M. S. C., Wallace, C. D., and Hoffman, R M. Hypomethylation of DNAin human cancer cells: a site-specific change in the c-myc oncogene. J. Nati.Cancer Inst., 73: 1057-1065.1984.

73. Mays-Hoopes, L., Brown, A., and Huang, R. C. C. Methylation and rearrange

ment of mouse intracisternal A particle genes in development, aging andmyeloma. Mol. Cell. Biol., 3: 1371-1380,1983.

74. Kuo, M. T., Iyer, B., Wu, J. R., Lapeyre. J-N., and Becker, F. F. Methylation ofthe Â«-fetoprotein gene in productive and nonproductive rat hepatocellularcarcinomas. Cancer Res., 44: 1642-1647. 1984.

75. Landolph, J. R., and Jones. P. A. Mutagenicity of 5-azacytidine and relatednucleosides in C3H/10TV2 C18 and V79 cells. Cancer Res., 42. 817-823.1982.

76. Delers, A., Szpirer, J., Szpirer. C., and Saggioro, D. Spontaneous and 5-


465




azacytidine-induced re-expression of ornithine carbamoyl transferase Â¡nhep- 5: 1583-1590,1984.atoma cells. Mol. Cell. Biol., 4: 809-812, 1984. 81. Frost, P., Liteplo, R. G., Donaghue, R. P., and Kerbel, R. S. The selection of

77. Kerbel, R. S., Frost, P., Liteplo, R., Cariow, D., and Elliott, B. E. Possible strongly immunogenic "Turn"" variants from tumors at high frequency using 5-

epigenetic mechanisms of tumor progression: induction of high frequency azacytidine. J. Exp. Med., 759: 1491-1501,1984.heritable but phenotypically unstable changes in the tumorigenic and meta- 82. Olsson, L, and Forchhammer, J. Induction of the metastatic phenotype in astatic properties of tumor cell populations by 5-azacytidine treatment. J. Cell. mouse tumor model by 5-azacytidine, and characterization of an antigenPhysiol. Suppl., 3: 87-97, 1984. associated with metastatic activity. Proc. Nati. Acad. Sci. USA, 87: 3389-

78. Benedict, W. F., Banerjee, A., Gardner, A., and Jones, P. A. Induction of 3393, 1984.morphological transformation in mouse C3H/10T'/2 clone 8 cells and chrome- 83. Olsson, L., Due, C., and Diamont, M. Treatment of human cell lines with 5-somal damage in hamster A(T,)CI-3 cells by cancer chemotherapeutic agents. azacytidine may result in profound alterations in clonogenicity and growth rate.Cancer Res., 37.; 2202-2208, 1977. J. Cell Biol., 700: 508-513, 1985.

79. Harrison, J. J., Anisowicz, A., Gadi, I. K., Raffeld, M., and Sager, R. Azacytidine- 84. Babiss, L. E., Zimmer, S. G., and Fisher, P. B. Reversibility of progression ofinduced tumorigenesis of CHEF/18 cells: correlated DMA methylation and the transformed phenotype in AdS-transformed rat embryo cells. Sciencechromosome changes. Proc. Nati. Acad. Sci. USA, 80: 6606-6610,1983. (Wash. DC), 228: 1099-1101,1985.

80. Carr, B. I., Reilly, J. G., Smith, S. S., Winberg, C., and Riggs, A. The 85. Taylor, S. M., and Jones, P. A. Multiple new phenotypes induced in 10T1/2andtumorigenicity of 5-azacytidine in the male Fischer rat. Carcinogenesis (Lond.), 3T3 cells treated with 5-azacytidine. Cell, 77: 771 -779,1979.


466



1986;46:461-466. Cancer Res Peter A. Jones DNA Methylation and Cancer

Updated version

http://cancerres.aacrjournals.org/content/46/2/461.citation

Access the most recent version of this article at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected] at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/46/2/461.citationTo request permission to re-use all or part of this article, use this link



http://cancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]



Documents

DNA Methylation and Cancer1 - cancerres.aacrjournals.orgcancerres.aacrjournals.org/content/canres/46/2/461.full.pdf · DNA METHYLATION AND CANCER dehydrogenase activity and hypermethylation