Cancer Epigenetics

Epigenetics Controls Reading of Genetic Code

Epigentics refers to high level information residing above the genetic code. While each cell in the body is equipped with the same genetic manual, epigenetic instructions tell cells how to make a difference. These instructions determine the access to pages with genetic information by directing the way the DNA is packaged into chromatin. DNA organized in loose chromatin is readily available for gene expression. Conversely, DNA tightly packed into dense chromatin has the letters of genetic code effectively buried and unavailable for reading and transcription. Distinct epigenetic marks decide which sets of genes may be expressed and which genes are kept silent. Once established, epigenetic patterns are stable and are transferred into subsequent generations of somatic cells. Permanent inactivation of one X chromosome in females and parental imprinting of specific genes on autosomes are the best examples of stable epigenetic silencing exerted on one allele. Transcription factors present in the cell nucleus regulate expression of only the other nonsilenced allele.

Note: If you’d like to print out a hard copy of these reviews, you can find it at Zymo Research’s website in their Publications.

Histones and methylation of DNA are key molecular players with well documented roles in epigenetic molding of chromatin. Histones are small proteins that form spools for wrapping of DNA into nucleosomes. Nucleosomes are the smallest structural unit of chromatin and consist of 8 histone core molecules (doublets of histone 2A, 2B, 3, and 4) with two loops of DNA (147 bp) coiled around them. N-terminal parts of histone proteins protrude from nucleosome cores. Amino acids in histone tails can be modified by numerous enzymes bringing acetylation, methylation, phosphorylation, ubiquitination and other substitutions, creating a complex ‘bar’ code1,2. Based on this histone code, activating or repressive complexes are attracted to DNA and shape chromatin into relaxed or tightly packed structures. Acetylation and phosphorylation of histones are usually activation marks associated with open chromatin structure. Methylation could also be an activation mark, such as trimethylation of lysine 4 in histone3. Remarkably, several amino acids further on the same histone molecule methylation of lysine 9 is a mark associated with silencing of euchromatin or densely packed pericentromeric heterochromatin3. Methylation of lysine 27 mediated by the Polycomb repressive complex4,5 is an important silencing mechanism of key differentiation factors in embryonic stem cells6.

In contrast to a dazzling variety of histone modifications7,8 DNA methylation is just a simple addition of a single methyl (CH3-) group to cytosine at position 5. In mammals, it can almost exclusively happen only on cytosines preceding guanine (CpG) in the DNA sequence. The majority of CpG sites in human DNA are methylated9. Methyl groups flag approximately 5-6% of cytosines in healthy cells10. CpG methylation is a silencing epigenetic mark and may have developed as a defense against expression of parasitic DNA elements11. Moreover, 5-methylcytosine is prone to spontaneous deamination and point mutation to thymine12. Consequently, CpG dinucleotides are depleted almost fivefold in the human genome13. Although 5-methylcytosine represents only 1% bases in the human genome, its potential mutagenic hazard is well illustrated by the fact that CpG dinucleotide is involved in one third of point mutations causing human genetic disorders14 and a similar proportion of single nucleotide polymorphisms detected in gene exons15.

Despite the mutational threat it poses, CpG methylation is essential for life. DNA methylation is established and maintained by specific DNA methyltransferases (DNMT1, 3a and 3b). In mammals, DNA methylation is required for proper embryogenesis and development16, for sustaining chromosomal stability17, telomere length18, and predetermined gene expression states19 (see review, page 5). The methylation pattern is faithfully passed into dividing somatic cells in a similar fashion as the genetic code20. It is ‘photocopied’ on newly made DNA strands by DNA methyltransferase 1 (DNMT1) that works together with the protein UHRF1 as an all-in-one scanner and copier21,22. UHRF1 detects CpG sites methylated only on a single DNA strand and flips the existing 5-methylcytosine out of the DNA helix. DNMT1 then adds a new methyl group on CpG on the complementary DNA strand23-25.

There is an important exception to a general rule that CpG sites are both sparse and methylated. About half of human genes have their transcription start sites flanked with islands containing tens or hundreds of CpG sites crammed in hundreds to thousands bases-long stretches of DNA. Cytosines in these CpG islands are typically unmethylated which makes them less susceptible to mutational pressure and may explain their survival throughout evolution. By keeping CpG islands methylation-free, normal cells maintain passageways for gene expression open, clean and clear of obstructions9,19.


Epigenetics and Genetics Are Accomplices in Cancer Development

Cancer is caused by failure of checks and balances that control cell numbers in response to the needs of the whole organism. Inappropriate function of genes that promote or inhibit cell growth or survival can be caused by errors introduced into the genetic code itself or by faulty epigenetic mechanisms deciding which genes can and cannot be expressed. Epigenetic lesions and genetic mutations are acquired during the life of an individual and accumulate with aging. Both types of events, either individually or in cooperation, can result in the loss of control over cell growth and development of cancer26.

DNA methylation patterns undergo complex changes in cancer. The total amount of methylated cytosine is usually decreased resulting in global hypomethylation. Decreased cytosine methylation typically affects satellite DNA, repetitive sequences, and CpG sites located in gene bodies (introns and inner exons). Mice with crippled function of DNMT1 have widespread genomic hypomethylation, activated endogenous retroviral elements, and develop aggressive T lymphomas27,28. The cause of reduced amount of methylcytosine observed in human tumors has not been determined. Despite global hypomethylation, high activity of DNA methyltransferases has been detected in multiple human tumor types. This increase may be related to higher proliferation rate of malignant cells29.

Besides global hypomethylation, most cancers also show focal hypermethylation in distinct subsets of promoter-associated CpG islands. Affected genes are permanently silenced, since methylation marks are propagated through mitosis and are maintained in the malignant clone. Aberrant de novo hypermethylation occurring in transformed cells serves as an alternative mechanism for inactivation of tumor suppressor genes. Hundreds to thousands of genes can be epigenetically silenced by CpG island hypermethylation in human cancer26,30,31, suggesting a general disturbance of epigenetic memory. Methylation affects individual cancer patients with varying extent. While some patients have minimal changes, others show concordant hypermethylation of multiple genes. This phenomenon was first described as CpG island methylator phenotype (CIMP) in colorectal cancer and confirmed in many other types of cancer and leukemia29. Epigenetic DNA methylation changes in cancer appear to be considerably more frequent events than genetic mutations. Mass sequencing of more than 20,000 transcripts in breast and colorectal cancers revealed about 80 harmless and less than 15 potentially oncogenic mutations per tumor32.

Most mutations and epigenetic alterations occurring in cancer cells are probably harmless. Many genes methylated in tumors are not expressed in relevant normal tissues and their silencing by methylation is inconsequential. However, disabling of genes that are critically important to controlling cell proliferation contributes to the development of a malignant phenotype in the same manner as inactivating mutations of tumor suppressor genes33. Epigenetic silencing by DNA methylation of cyclin-dependent kinase inhibitors34-37, DNA repair genes38, apoptosis mediators39,40, nuclear receptors41-43, transcription factors44, cell adhesion molecules45, and many other genes has been reported in multiple cancer types46-49. Perhaps the most convincing evidence for CpG island methylation causing the same harmful effect as inactivating mutations of tumor suppressor genes in cancer is documented in studies describing RB150, p16 (CDKN2A)51, VHL52, and MLH153 genes. Each of these genes can be inactivated by either DNA methylation or a mutation with equal consequences for cancer development.

A complex code created by covalent modifications of amino acids in histone tails is also heavily involved in chromatin disturbances in cancer54-56. Some genes mutated in cancer recruit histone modifying enzymes and thus alter gene expression. For example, the PML-RAR gene translocation in acute promyelocytic leukemia recruits histone deacetylases that change the chromatin structure from active to silenced and contribute to leukemic transformation57. Mutations also directly target histone-modifying enzymes. For example, histone acetylase CBP is mutated, and histone 3 lysine 4 activating methyltransferase MLL gets rearranged in leukemia58,59. EZH2, a histone 3 lysine 27 trimethyltransferase is frequently overexpressed in cancer and likely plays an important role in tumorigenesis60.

Stem Cell Epigenome and Cancer

Temporary silencing of developmental genes in stem cells is an important mechanism for keeping their chromatin in a plastic state. Silencing is brought by EZH2, a member of the Polycomb repressive complex 2 causing methylation of lysine 27 on histone 3. Embryonic stem cells display the silencing mark together with an activating methylation mark on lysine 4. Genes with this bivalent chromatin structure are poised for expression driving differentiation into a specific lineage just after removal of the silencing histone mark61. Interestingly, developmental genes silenced by EZH2 in embryonic stem cells appear to be more prone to DNA methylation in cancer61-65. The exact relationship between histone 3 lysine 27 methylation and DNA methylation is not clear. It has been proposed that EZH2 and other proteins of the Polycomb repressive complex recruit DNA methyltransferases to specific promoters66. Recent studies in prostate cancer suggest that EZH2-mediated histone methylation and DNA methylation are not concurrent mechanisms of gene silencing. They may occur side by side67 or in sequence, where relatively plastic Polycomb-mediated repression is replaced by DNA hypermethylation that seals reprogramming of the cancer epigenome68. Alterations of epigenomes in cancer stem cells may promote their self-renewal and make them less responsive to differentiation signals than their stem cell counterparts in normal tissues. Tight heritable silencing of crucial gene sets would hold cells in a stem cell state; make them vulnerable to mutations and further epigenetic changes in cancer progression26.

Epigenome Mapping

With the impact of epigenetics in health and disease firmly established, national and international initiatives have been established to coordinate the scientific community’s efforts. The National Institutes of Health recently announced a Roadmap Epigenomics Program to define and map epigenetic modifications genome wide. An international Alliance for the Human Epigenome and Disease (AHEAD) has been formed by the American Association for Cancer Research Epigenome Task Force and the European Union Network of Excellence Scientific Advisory Board. Inspired by the great success of the human genome sequencing, the goals of the AHEAD project are to generate highresolution reference epigenome maps. First, a defined subset of robust epigenetic markers will be analyzed in a limited number of human tissues at different stages. To support epigenetic data, a bioinformatics infrastructure will be established69.

Rapidly developing technologies greatly transformed our capacity to map epigenetic marks in a large scale. Most recently, ultra-highthroughput technologies using massive parallel sequencing brought a new exciting tool for epigenomic analyses. These techniques can simultaneously read millions of short DNA fragments. Genome wide maps of specific histone marks captured by chromatin immunoprecipitation were generated using this deep sequencing technology70,71.

DNA methylated regions in the genome can be detected on high density microarrays after isolation by restriction digests with methylation-sensitive enzymes72,73, methylcytosine-specific antibodies74 or methyl binding proteins75. Replacing microarrays with deep sequencing of isolated methylated DNA further broadens the spectrum of analyzed sequences and also provides more accurate quantitative information than microarray detection.

The gold standard of DNA methylation analysis is sequencing of DNA chemically modified by bisulfite treatment that converts unmethylated cytosines to uracils while leaving methylated cytosines intact76. Massive sequencing of the whole bisulfite converted genome will provide DNA methylation maps with a single nucleotide resolution. This task was successfully accomplished in Arabidopsis with a relatively small 120 megabase genome77 and in a fraction of mammalian genome enriched for CpG islands78.

Epigenetic Therapy of Cancer

Mapping and characterizing epigenomic changes will transform our understanding of pathology and enhance our ability to diagnose and treat cancer. Epigenetic alterations are easier to reverse than mutations affecting the genetic code. Two inhibitors of DNA methyltransferases, azacytidine and deoxyazacytidine, have already been approved by the Food and Drug Administration as effective drugs for treatment of patients with myelodysplastic syndromes. An inhibitor of histone deacetylases, vorinostat (suberoylanilide hydroxamic acid), is approved for the treatment of cutaneous T-cell lymphoma. Other epigenetic drugs targeting histone modifying enzymes or DNA methylation are in clinical trials or development79. Detailed understanding of chromatin dysregulation will undoubtedly be translated into new and more effective ways of cancer treatment.

Epigenetics References

  1. Jenuwein, T & Allis, CD. Science 293:1074-1080 (2001).
  2. Strahl, BD & Allis, CD. Nature 403:41-45 (2000).
  3. Lachner, M et al. Curr Opin Cell Biol 14:286-298 (2002).
  4. Czermin, B et al. Cell 111:185-196 (2002).
  5. Muller, J et al. Cell 111:197-208 (2002).
  6. Lee, TI et al. Cell 125:301-313 (2006).
  7. Bernstein, BE et al. Cell 128:669-681 (2007).
  8. Kouzarides, T. Cell 128:693-705 (2007).
  9. Bird, A. Genes Dev 16:6-21 (2002).
  10. Romerio, AS et al. Anal Biochem 336:158-163 (2005).
  11. Yoder, JA et al. Trends Genet 13:335-340 (1997).
  12. Gonzalgo, ML & Jones PA. Mutat Res 386:107-118 (1997).
  13. Simmen, MW. Genomics 92:33-40 (2008).
  14. Cooper, DN & Youssoufian, H. Hum Genet 78:151-155 (1998).
  15. Jiang, C & Zhao, Z. Genomics 88:527-534 (2006).
  16. Li, E et al. Cell 69:915-926 (1992).
  17. Eden, A et al. Science 300:455 (2003).
  18. Vera, E et al. Oncogene (in press) doi: 10.1038/onc.2008.289 (2008).
  19. Jaenisch, R & Bird, A. Nat Genet 33 Suppl: 245-254 (2003).
  20. Ooi, SK & Bestor, TH. Curr Biol 18:R174-176 (2008).
  21. Sharif, J et al. Nature 450:908-912 (2007).
  22. Bostick, M et al. Science 317:1760-1764 (2007).
  23. Avvakumov, GV et al. Nature (in press) doi: 10.1038/nature07273 (2008).
  24. Arita, K et al. Nature (in press) doi: 10.1038/nature07249 (2008).
  25. Hashimoto, H et al. Nature (in press) doi: 10.1038/nature07280 (2008).
  26. Jones, PA & Baylin, SB. Cell 128:683-692 (2007).
  27. Gaudet, F et al. Science 300:489-492 (2003).
  28. Howard, G et al. Oncogene 27:404-408 (2008).
  29. Issa, JP. Nat Rev Cancer 4:988-993 (2004).
  30. Kroeger, H et al. Blood 112:1366-1373 (2008).
  31. Kuang, SQ et al. Leukemia 22:1529-1538 (2008).
  32. Wood, LD et al. Science 318:1108-1113 (2007).
  33. Toyota, M & Issa, JP. Semin Oncol 32:521-530 (2005).
  34. Herman JG, et al. Cancer Res 56:722-727 (1996).
  35. Corn, PG et al. Cancer Res 59:3352-3356 (1999).
  36. Kikuchi, T et al. Oncogene 21:2741-2749 (2002).