Epigenetics in Cancer
Introduction
Epigenetic Variation in Humans
Genetic and Epigenetic Heterogeneity in Cancer
Techniques and Applications
Mediators and dynamics of DNA methylation
DNA Methylation Alterations in Multiple Myeloma as a Model for Epigenetic Changes in Cancer
Ink4-Arf locus in cancer and aging
Protocols
Simultaneous Single-Molecule Mapping of Protein-DNA Interactions and DNA Methylation by MAPit
Comprehensive High-Throughput Arrays for Relative Methylation (CHARM)
Methylation-specific PCR
Further Reading
eLS
WIREs
Current Protocols
John Wiley & Sons, Ltd.
ELS subject area: Evolution and Diversity of Life
How to cite: Wilkins, Jon F (July 2008) Epigenetic Variation in Humans. In: Encyclopedia of Life Sciences (ELS). John Wiley & Sons, Ltd: Chichester.
DOI: 10.1002/9780470015902.a0020811
Advanced article
Epigenetic modifications alter the expression behaviour of genes without involving changes to the DNA (deoxyribonucleic acid) sequence itself. These modifications take a variety of forms, most notably DNA methylation and histone modification. Epigenetic differences allow alleles, cells or even individuals that are genetically identical to exhibit radically different phenotypes. Differentiation among cells underlies tissue differentiation and development. Differentiation between alleles within a cell is used to enhance clonal diversity, as in the immune system, and has been highjacked in the service of evolutionary genetic conflicts, as seen in the phenomenon of genomic imprinting. Epigenetic differences among individuals may account for some of the differences between monozygotic (identical) twins. Some recent studies have even suggested that these epigenetic mechanisms may allow organisms to adapt to environmental changes on very short timescales. Environmental adaptations could be epigenetically encoded and passed on to offspring, providing a potential mechanism for a neo-Lamarckian mode of evolution.
There is more to a gene than its deoxyribonucleic acid (DNA) sequence. CH Waddington coined the term ‘epigenetic’ to describe biological differences among tissues resulting from the process of development (Waddington, 1939, 1942). Waddington needed a new term to describe this type of variation, which is not the product of genotypic differences among the cells. Neither are these tissue-level differences well described as phenotypic variation, since the variation occurs within an individual organism. We now understand that heritable modifications to the DNA (such as cytosine methylation) and aspects of chromatin structure (including histone modifications) are the mechanisms that underlie Waddington’s ‘epigenotype’. The DNA and its associated proteins are modified in particular cells during development, and those modifications are propagated across multiple cell divisions. Those modifications are responsible for the variation in patterns of gene expression across cell types. In contemporary usage, the term epigenetic refers to any heritable change in gene expression that is not coded in the DNA sequence itself (Egger et al., 2004).
In humans, there are two principal mechanisms of epigenetic modification. The first is cytosine methylation. This occurs overwhelmingly at CpG dinucleotides. These dinucleotides are found in clusters throughout the genome, known as CpG islands. The CpG dinucleotide is a simple example of a ‘pallindromic sequence’. That is, since C pairs with G, a CpG in one strand of the DNA is paired with another CpG in the complementary strand.
The propagation of methylation states across cell divisions depends on this pallindromic structure and the action of the protein Dnmt1, a maintenance methyltransferase, or hemimethylase. The fully methylated form has methyl groups attached to the cytosine on each strand. When the DNA is replicated, these two methylated strands are separated, and new (unmethylated) DNA is synthesized on each. The two daughter cells are then set to inherit hemimethylated DNA (DNA with methyl groups on one strand, but not the other). Dnmt1 specifically recognizes this hemimethylated DNA and attaches a methyl group to the newly synthesized strand. Thus, once the pattern of methylation is set in a particular cell, Dnmt1 acts to propagate this pattern to all of that cell’s descendants.
The second mechanism of epigenetic modification involves various modifications to histones, which are proteins that are physically associated with the DNA. Humans express four distinct histone proteins H2A, H2B, H3 and H4. These histones are subject to a large number of modifications, many of which correlate with the transcriptional state of nearby genes. For instance, H3–K9 methylation (methylation of the lysine at position nine of the H3 histone) is characteristics of transcriptionally inactive heterochromatin. H3–H4 acetylation and H3–K4 methylation are characteristic of transcriptionally active euchromatin.
The stable propagation of epigenetic states across multiple cell divisions relies, in part, on a set of positive feedback mechanisms that link cytosine methylation and histone modification (Wilkins, 2005). The activity of Dnmt1 at a particular hemimethylated site is enhanced both by the presence of nearby DNA methylation and by particular histone modifications. De novo methyltransferases (which act on unmethylated DNA) are also recruited to particular loci by the proximity of these modifications. Likewise, the enzymes responsible for establishing these histone modifications are recruited by the presence of methylated cytosine.
This system of positive feedbacks between DNA methylation and histone modification leads to a kind of ‘epigenetic canalization’. The result is a bistable system in which two states – locally fully methylated and locally fully unmethylated – are each stable with respect to small perturbations. Once methylation is established across a local cluster of CpG dinucleotides, these mechanisms will maintain that methylated state across many cell divisions. For instance, if a single methylation site were to fail to be copied following DNA replication, the existence of neighbouring methylated sites and the presence of modified histones would cause the site to become remethylated at a later time.
The genomic DNA sequence encodes the instructions for building the many components that make up a cell. The DNA sequence also includes regulatory motifs, which partially determine when, where and at what level a particular gene product will be produced. It may be best, however, to think of the DNA as encoding the potential for a very broad range of possible expression patterns. Epigenetic modifications provide a kind of filter, which can stably select a particular pattern of gene expression from this possible set. This permits the same genome to stably follow different expression strategies depending on environmental conditions, or to simultaneously exhibit different behaviours in different (genetically identical) cell types in a multicellular organism.
This behavioural differentiation among cell lineages is just one of the roles played by epigenetic modifications. These modifications can also produce differences in the expression behaviour of the two alleles within a single cell. Typically, this involves silencing one of the two copies. This allele-specific silencing is involved in X-chromosome inactivation and parent-of-origin effects, as well as generating variation in the olfactory and immune systems.
DNA methylation and its associated histone modifications also increase the stability of various repeat sequences, including minor satellites, subtelomeric satellites and microsatellites, as well as a variety of repeat-containing transposable elements. Demethylation of these elements results in increased recombination and structural instability of the chromosome (Gonzalo et al., 2006; Guo et al., 2004; Hansen et al., 1999; Kim et al., 2004; Okano et al., 1999; Wang and Shen, 2004; Xu et al., 1999).
The third context in which epigenetic modifications play a crucial role is maintaining these transposable elements in an inactive state. When methylation is lost from transposable elements, those elements become active, and insert additional copies of themselves at random locations throughout the genome, with strongly deleterious effects. In fact, it has been suggested that the ability to protect the genome from these parasitic elements was the primary selective force behind the evolution of our sophisticated epigenetic machinery (Bestor, 2003).
Development is the most prominent, and best understood, context in which we find epigenetic variation. By the term ‘developmental variation’, I mean to refer to the variation among cells that is generated during development. This variation is not ‘heritable’ in the sense that it is recreated each generation, and does not necessarily imply heterogeneity among individuals. However, it is heritable within a given cell lineage. This type of epigenetic variation is not the primary focus of this article, but will be discussed briefly here.
One of the surprising features of the diploid genome is that many genes are expressed from only one of the two alleles. Epigenetic differences between the alleles can generate differences in expression behaviour, even in the absence of any differences at the DNA sequence level.
The most familiar example is likely to be the random inactivation of one of the two X-chromosomes in females. This is commonly interpreted as a mechanism of dosage compensation: males have a single X-chromosome, whereas females have two. Inactivation of one of the two X-chromosomes in females results in similar gene expression levels for X-linked loci in the two sexes.
In some cell types, epigenetic modifications are used to establish clonally stable expression patterns. For instance, several of the interleukins, including IL2, IL3, IL4, IL5 and IL13, are randomly expressed monoallelically in particular subpopulations of T lymphocytes (Bix and Locksley, 1998; Reinhardt et al., 2006). This random inactivation generates a combinatorial assortment of cell lines, each with a distinct, but stable, expression profile. A similar process occurs in the olfactory system, where individual neurons express only one of the order of 1000 distinct odourant receptor genes.
The types of epigenetic variation discussed in section Developmental Epigenetic Variation are heritable in two senses. First, these epigenetic states are propagated through multiple cell divisions, and are therefore inherited by the progeny of a particular cell. Second, the epigenetic machinery and the developmental programme that it executes are encoded by the genome. However, neither of these senses implies the concept of ‘heritability’ as it is typically conceived in population genetics. In this section, we will focus on those patterns of epigenetic variation that are passed on from parent to offspring.
In humans and other mammals, approximately 1 per cent of genes are subject to genomic imprinting (Wilkins and Haig, 2003). Imprinting represents a type of allelic silencing, but, in contrast to the forms of allelic silencing discussed in section Developmental Epigenetic Variation, imprinting is the product of epigenetic modifications that are established in the germlines of the parents and inherited by the offspring.
The discovery of the first imprinted genes was precipitated by the discovery that the maternally and paternally derived genomes are not equivalent in mammals. Nuclear transplant experiments showed that gynogenetic embryos (in which both copies of each gene are derived from a female) do not develop into viable offspring. Androgenetic embryos (with two paternally derived genomes) are similarly inviable. In mammals, successful development requires both maternally and paternally derived alleles.
The epigenetic differences associated with imprinted genes are reproduced every generation, and in that way are similar to other forms of developmental epigenetic variation. However, those epigenetic marks are established in the parents and inherited by the offspring.
Most imprinted genes are expressed monoallelically in one or more tissues. This makes these genes particularly susceptible to dysregulation either through mutations in the DNA sequence, or through the loss or reprogramming of epigenetic information. Loss of function or of transcription of the single active allele can result from a single mutation or epimutation. In certain cases, it is possible that such an epimutation may be passed from one generation to another.
One disorder known to be associated with epigenetic defects is familial pseudohypoparathyroidism type Ib (PHP-Ib) (Bastepe and Juppner, 2005; Bastepe et al., 2001a, 2001b; Ding et al., 1996; Levine et al., 1983; Silve et al., 1986). PHP-Ib results from the loss of the G-protein stimulatory α subunit (Gsα) in particular tissues, including the renal proximal tubules. Patients with PHP-Ib do not appear to have any coding mutations in Gsα. Rather, these patients show epigenetic defects of the stimulatory G-protein α subunit (GNAS) locus, which produces multiple transcripts whose expression is coregulated. The most common epigenetic defect associated with PHP-Ib is a loss of imprinting of the exon A/B differentially methylated region. This epimutation appears to silence maternal expression of the Gsα transcript. In those tissues where Gsα is predominantly expressed from the maternally derived allele, this silencing results in a dramatic reduction in Gsα protein levels.
It should be clear from the preceding sections that the establishment, propagation and interpretation of epigenetic modifications involve the interaction of multiple cis- and trans-acting factors. In principle, genetic variation in these factors among individuals might generate epigenetic variation. Most obviously, if an individual is lacking a particular cytosine, that site will not be subject to methylation.
Less trivially, mutations in trans-acting factors involved in targeting the epigenetic machinery to particular loci might lead to reproducible differences in epigenetic state. We would expect these differences to be heritable in a classical Mendelian fashion, but the epigenetic states at the affected loci will follow the inheritance pattern of the trans-acting factor in question, rather than that of the epigenetically modified loci themselves.
The effects of this type of variation are not yet well understood. Trans-acting factors whose involvement in epigenetic regulation is well established are largely restricted to those enzymes that are directly involved in the methylation reaction. Because these proteins act on epigenetic regulation throughout the genome, loss-of-function mutations have devastating effects and are not found segregating in the population. Mutations with less pronounced effects might well exist at higher frequencies in the population, and might be associated with variations in tissue and/or locus specificity. However, the existence and function of these variants have not yet been studied systematically.
There is some evidence of the consequences of mutations in the epigenetic machinery that comes from naturally occurring mutations in humans. In particular, mutations in the Dnmt3b gene, which encodes one of the de novo methyltransferases, are associated with immunodeficiency, centromeric instability and facial anomalies (ICF) syndrome (Hansen et al., 1999; Okano et al., 1999; Xu et al., 1999). Patients with ICF syndrome have hypomethylation of classical satellites and some interspersed repeats. This hypomethylation results in chromosomal instability. In particular, chromosomes containing sat2 or sat3 sequences in the pericentromeric regions show expansion of these repeats and increased recombination between repeat regions. This inappropriate recombination can result in the duplication and/or deletion of entire chromosome arms.
The Dnmt3L protein, which is orthologous to Dnmt3b, also appears to play an important role in genomic stability and the epigenetic silencing of transposable elements (Bourc’his and Bestor, 2004; Webster et al., 2005). In Dnmt3L mutant males, transposable elements of the intracisternal A particle (IAP) and L1 families show demethylation and increased transcription. This activity leads to meiotic instability, most likely due to increased recombination activity between nonhomologous loci, which leads to chromosomal deletions and translocations. The mechanism through which mutations in Dnmt3L have this effect, however, is not clear, since Dnmt3L lacks the active site and catalytic activity associated with other members of the methyltransferase family.
As we have already discussed, most epigenetic marks appear to be erased each generation in the germline. However, there are three classes of repeat sequence found in the genome that appear to undergo only partial demethylation (Hajkova et al., 2002; Lane et al., 2003; Lees-Murdock et al., 2003). The first of these is the class of IAPs, which are endogenous retroviral elements featuring long terminal repeats (LTRs). IAPs contain all of the elements required for the element to make a copy of itself and integrate that copy into a new site in the genome. L1 elements form a second class that contains a promoter and two open reading frames, but lacks the LTRs. The third class contains nontranscribed minor satellite sequences consisting of 20–200 bp repeats and is found close to the centromere.
In mice it has been shown that all three classes are partially demethylated when they first enter the germline. In the male germline, this hypomethylated state persists until the prospermatogonia enter mitotic arrest, at which point de novo methylation restores these sequences to their fully methylated state. In the female germline, the hypomethylated state persists longer, at least until the primary oocytes pass through the meiotic prophase.
At these repeat sequences, therefore, the epigenetic state that is passed on to the offspring represents some combination of modifications that have been passively propagated throughout the development of the parents. It is not yet known whether epimutations could be passed from generation to generation in this way, or if the de novo methylation process that occurs later in gametogenesis would be expected to erase any such variation.
Recent work by Michael Meaney’s group has identified an alternative mechanism by which the epigenetic state of offspring can be influenced by parents. In mice, the epigenetic status of the glucocorticoid receptor and oestrogen receptor genes is influenced by tactile stimulation – specifically, maternal licking (Champagne et al., 2006; Szyf et al., 2005; Weaver et al., 2004, 2005). Whether or not analogous mechanisms exist in humans remains to be seen.
In the preceding sections, we have introduced the mechanisms of epigenetic regulation of gene expression. Although it is clear that these epigenetic modifications play a central role in generating a broad range of phenotypic diversity among genetically identical cells within a single organism, it is not yet clear to what extent these epigenetic patterns vary among different individuals.
In most of the preceding examples, the complex patterns of epigenetic diversity are recreated every generation as a part of the normal developmental process. The patterns of epigenetic silencing found in genomic imprinting are technically inherited from one generation to another, but can also be viewed as a reproducible part of the extended developmental process. That is, genomic imprinting represents an increase in the diversity of allelic behaviour, but that does not necessarily imply any greater degree of epigenetic variation among individuals.
In fact, there is some reason to believe that heritable epigenetic variation may be relatively rare. In both the male and female germlines, the genome is subject to a massive epigenetic reprogramming. Most of the epigenetic modifications that are inherited by the cells in the germline are actively removed during meiosis. Even those modifications that are subsequently reestablished are first erased. For example, at a maternally inherited imprinted locus in the female germline, the maternal-specific methylation pattern is first erased, and then the identical methylation pattern is reestablished.
Although it appears that the vast majority of the genome’s cytosine methylation is erased every generation, other aspects of the epigenetic state may be more persistent, and may even permit the possibility that particular epigenetic states could be propagated across multiple generations. Let us return to the example of epigenetic reprogramming of an imprinted locus in the female germline. The germ cells contain two differently marked alleles (maternally inherited and paternally inherited). Methylation is removed from both of these alleles, and then the maternal-specific methylation pattern is established on both.
For the maternally derived allele, this is the same methylation pattern that was erased. For the paternally derived allele, this is a new pattern. Although the methylation process begins at the same time, it proceeds at different rates on the two alleles. The reestablishment of the maternal pattern on the maternally derived allele happens much more quickly than the establishment of the same pattern on the paternally derived allele (Davis et al., 1999, 2000; Hiura et al., 2006; Lucifero et al., 2002, 2004; Obata and Kono, 2002; Ueda et al., 2000).
This difference in rates of methylation suggests that other epigenetic differences (such as patterns of histone modification) between the two alleles persist beyond the removal of the DNA methylation. The positive feedback between DNA methylation and histone modification means that epigenetic state of the maternally derived allele facilitates its own propagation, even if the methylation is transiently removed. It is possible that this same process could facilitate the inheritance of epigenetic states across multiple generations.
The ability to reliably propagate epigenetic states across multiple generations could potentially open up new modes of plasticity and adaptation that operate on a timescale shorter than that typically associated with heritable mutations in the DNA, but longer than that associated with the development of the individual organism.
The idea of transgenerational inheritance of epigenetic states to adapt to fluctuating environmental conditions is an appealing one. Parents would pass on information about the environment to their offspring, allowing them to tailor their development to current conditions. However, the plausibility of this evolutionary argument demands that several conditions be met. First, environmental conditions would need to fluctuate widely enough to necessitate different developmental strategies. These fluctuations would need to be on a sufficiently slow timescale that the environments experienced by parents and offspring would be highly correlated. However, fluctuations would need to occur frequently enough, and over a sufficiently long period, that there would be a selective advantage to simultaneously adapting to multiple environmental conditions.
Most meiotically heritable epimutations will be those that occur in the germline and are passed on to offspring. In some cases, we expect the epimutation to persist only for a single generation. In other cases, a germline epimutation might be more stable, persisting for several generations, or, potentially, even being assimilated into the genotype (Hanson and Gluckman, 2005; Jablonka and Lamb, 1998; Waterland and Jirtle, 2004).
The adaptive significance of meiotically heritable epimutations has been a source of widespread speculation. In principle, these epimutations could form the basis of a type of transgenerational phenotypic plasticity, allowing offspring to respond adaptively to environmental conditions experienced by the parents (Gluckman et al., 2005; Gorelick, 2004; Hanson and Gluckman, 2005; Jablonka and Lamb, 1998; Waterland and Jirtle, 2004). Whether any of the documented epimutations truly embody this type of preadaptation is not yet clear. It is also possible that the phenotypic consequences of these epimutations will be best understood as epiphenomena.
One type of epimutation that has received some attention is changes in methylation status resulting from dietary conditions. In particular, certain dietary perturbations at critical ontogenic stages can result in a shortage or excess of methyl donors (Gallou-Kabani and Junien, 2005; Hanson and Gluckman, 2005; Waterland and Jirtle, 2004). An example of this type of environmentally driven epimutation with transgenerational consequences is ‘metabolic syndrome’ (Gallou-Kabani and Junien, 2005). Mothers who are subjected to nutritional constraints during pregnancy often have descendants who suffer from a variety of metabolic dysregulation as adults, including glucose and insulin metabolism disorders, weight problems, hypertension, diabetes and cardiovascular disease (Barker, 2002; Gallou-Kabani and Junien, 2005; Hales and Barker, 2001). What is most compelling about this example is the fact that these disorders are not limited to the immediate descendants of the mother. Some of these effects appear to persist for two or even three generations.
It is conventional to partition phenotypic variation into its genetic and environmental components. However, our increasing understanding of epigenetic variation suggests that this may represent an important third aspect of phenotypic variation. Accounting for this epigenetic variation will be crucial to research programmes involving techniques such as association mapping, which have traditionally focused specifically on segregating variation at the DNA sequence level.
Analysis of twin studies, both in humans and in mice, has suggested that there are significant patterns of phenotypic variation between genetically identical individuals that are not accounted for by environmental differences (Wong et al., 2005). In the previous sections, we have encountered a number of processes that might result in the accumulation of epigenetic differences over the course of development. More speculatively, some of the mechanisms described earlier might mean that siblings could share inherited epigenetic marks that would make them phenotypically more similar than would be predicted by their shared genetic material alone.
It has become clear that epigenetic modifications play an important role in determining individual phenotypes. Some of these modifications are established in the germlines of the parents and inherited by offspring, such as those associated with imprinted gene expression, and those involved in transposon silencing and chromosome stabilization. There is some indirect evidence that has fuelled speculation that some epigenetic modifications might be passed on for multiple generations. Although we know that this type of epigenetic inheritance is common in plants, it is not yet clear whether or not it plays an important role in humans. Such transgenerational epigenetic inheritance mechanisms would demand a fundamental shift in our understanding of heritability, plasticity and the relationship between phenotype, genetics and environment.
Barker DJP (2002) Fetal programming of coronary heart disease. Trends in Endocrinology and Metabolism 13: 364–368.
Bastepe M and Juppner H (2005) GNAS locus and pseudohypoparathyroidism. Hormone Research 63: 65–74.
Bastepe M, Lane AH and Yuppner H (2001a) Paternal uniparental isodisomy of chromosome 20q – and the resulting changes in GNAS1 methylation – as a plausible cause of pseudohypoparathyroidism. American Journal of Human Genetics 68: 1283–1289.
Bastepe M, Pincus JE, Sugimoto T et al. (2001b) Positional dissociation between the genetic mutation responsible for pseudohypoparathyroidism type Ib and the associated methylation defect at exon A/B: evidence for a long-range regulatory element within the imprinted GNAS1 locus. Human Molecular Genetics 10: 1231–1241.
Bestor TH (2003) Cytosine methylation mediates sexual conflict. Trends in Genetics 19: 185–190.
Bix M and Locksley RM (1998) Independent and epigenetic regulation of the interleukin-4 alleles in CD4(+) T cells. Science 281: 1352–1354.
Bourc’his D and Bestor TH (2004) Meiotic catastrophe and retrotransposon reactivation in male germ cells lacking Dnmt3L. Nature 431: 96–99.
Champagne FA, Weaver ICG, Diorio J et al. (2006) Maternal care associated with methylation of the estrogen receptor-alpha 1b promoter and estrogen receptor-alpha expression in the medial preoptic area of female offspring. Endocrinology 147: 2909–2915.
Davis TL, Trasler JM, Moss SB, Yang GJ and Bartolomei MS (1999) Acquisition of the H19 methylation imprint occurs differentially on the parental alleles during spermatogenesis. Genomics 58: 18–28.
Davis TL, Yang GJ, McCarrey JR and Bartolomei MS (2000) The H19 methylation imprint is erased and re-established differentially on the parental alleles during male germ cell development. Human Molecular Genetics 9: 2885–2894.
Ding CL, Usdin TB, Labuda M and Levine MA (1996) Molecular genetic analysis of pseudohypoparathyroidism type 1B: exclusion of the genes encoding the type 1 and type 2 PTH receptors. Journal of Bone and Mineral Research 11: M483–M483.
Egger G, Liang GN, Aparicio A and Jones PA (2004) Epigenetics in human disease and prospects for epigenetic therapy. Nature 429: 457–463.
Gallou-Kabani C and Junien C (2005) Nutritional epigenomics of metabolic syndrome – new perspective against the epidemic. Diabetes 54: 1899–1906.
Gluckman PD, Hanson MA and Spencer HG (2005) Predictive adaptive responses and human evolution. Trends in Ecology & Evolution 20: 527–533.
Gonzalo S, Jaco I, Fraga MF et al. (2006) DNA methyltransferases control telomere length and telomere recombination in mammalian cells. Nature Cell Biology 8: U416–U466.
Gorelick R (2004) Neo-Lamarckian medicine. Medical Hypotheses 62: 299–303.
Guo G, Wang W and Bradley A (2004) Mismatch repair genes identified using genetic screens in Blm-deficient embryonic stem cells. Nature 429: 891–895.
Hajkova P, Erhardt S, Lane N et al. (2002) Epigenetic reprogramming in mouse primordial germ cells. Mechanisms of Development 117: 15–23.
Hales NC and Barker DJP (2001) The thrifty phenotype hypothesis. British Medical Bulletin 60: 5–20.
Hansen RS, Wijmenga C, Luo P et al. (1999) The DNMT3B DNA methyltransferase gene is mutated in the ICF immunodeficiency syndrome. Proceedings of the National Academy of Sciences of the USA 96: 14412–14417.
Hanson MA and Gluckman PD (2005) Developmental processes and the induction of cardiovascular function: conceptual aspects. Journal of Physiology – London 565: 27–34.
Hiura H, Obata Y, Komiyama J, Shirai M and Kono T (2006) Oocyte growth-dependent progression of maternal imprinting in mice. Genes to Cells 11: 353–361.
Jablonka E and Lamb MJ (1998) Epigenetic inheritance in evolution. Journal of Evolutionary Biology 11: 159–183.
Kim M, Trinh BN, Long TI, Oghamian S and Laird PW (2004) Dnmt1 deficiency leads to enhanced microsatellite instability in mouse embryonic stem cells. Nucleic Acids Research 32: 5742–5749.
Lane N, Dean W, Erhardt S et al. (2003) Resistance of IAPs to methylation reprogramming may provide a mechanism for epigenetic inheritance in the mouse. Genesis 35: 88–93.
Lees-Murdock DJ, De Felici M and Walsh CP (2003) Methylation dynamics of repetitive DNA elements in the mouse germ cell lineage. Genomics 82: 230–237.
Levine MA, Downs RW, Moses AM et al. (1983) Resistance to multiple hormones in patients with pseudohypoparathyroidism – association with deficient activity of guanine-nucleotide regulatory protein. American Journal of Medicine 74: 545–556.
Lucifero D, Mann MRW, Bartolomei MS and Trasler JM (2004) Gene-specific timing and epigenetic memory in oocyte imprinting. Human Molecular Genetics 13: 839–849.
Lucifero D, Mertineit C, Clarke HJ, Bestor TH and Trasler JM (2002) Methylation dynamics of imprinted genes in mouse germ cells. Genomics 79: 530–538.
Obata Y and Kono T (2002) Maternal primary imprinting is established at a specific time for each gene throughout oocyte growth. Journal of Biological Chemistry 277: 5285–5289.
Okano M, Bell DW, Haber DA and Li E (1999) DNA methyltransferases Dnmt3a and Dnmt3b are essential for de novo methylation and mammalian development. Cell 99: 247–257.
Reinhardt RL, Kang SJ, Liang HE and Locksley RM (2006) T helper cell effector fates – who, how and where? Current Opinion in Immunology 18: 271–277.
Silve C, Santora A, Breslau N, Moses A and Spiegel A (1986) Selective resistance to parathyroid hormone in cultured skin fibroblasts from patients with pseudohypoparathyroidism type-Ib. Journal of Clinical Endocrinology and Metabolism 62: 640–644.
Szyf M, Weaver ICG, Champagne FA, Diorio J and Meaney MJ (2005) Maternal programming of steroid receptor expression and phenotype through DNA methylation in the rat. Frontiers in Neuroendocrinology 26: 139–162.
Ueda T, Abe K, Miura A et al. (2000) The paternal methylation imprint of the mouse H19 locus is acquired in the gonocyte stage during foetal testis development. Genes to Cells 5: 649–659.
Waddington CH (1939) An Introduction to Modern Genetics. London: Allen & Unwin.
Waddington CH (1942) The epigenotype. Endeavour 1: 18–20.
Wang KY and Shen CKJ (2004) DNA methyltransferase Dnmt1 and mismatch repair. Oncogene 23: 7898–7902.
Waterland RA and Jirtle RL (2004) Early nutrition, epigenetic changes at transposons and imprinted genes, and enhanced susceptibility to adult chronic diseases. Nutrition 20: 63–68.
Weaver ICG, Cervoni N, Champagne FA et al. (2004) Epigenetic programming by maternal behavior. Nature Neuroscience 7: 847–854.
Weaver ICG, Champagne FA, Brown SE et al. (2005) Reversal of maternal programming of stress responses in adult offspring through methyl supplementation: altering epigenetic marking later in life. Journal of Neuroscience 25: 11045–11054.
Webster KE, O’Bryan MK, Fletcher S et al. (2005) Meiotic and epigenetic defects in Dnmt3L-knockout mouse spermatogenesis. Proceedings of the National Academy of Sciences of the USA 102: 4068–4073.
Wilkins JF (2005) Genomic imprinting and methylation: epigenetic canalization and conflict. Trends in Genetics 21: 356–365.
Wilkins JF and Haig D (2003) What good is genomic imprinting: the function of parent-specific gene expression. Nature Reviews. Genetics 4: 359–368.
Wong AHC, Gottesman II and Petronis A (2005) Phenotypic differences in genetically identical organisms: the epigenetic perspective. Human Molecular Genetics 14: R11–R18.
Xu GL, Bestor TH, Bourc’his D et al. (1999) Chromosome instability and immunodeficiency syndrome caused by mutations in a DNA methyltransferase gene. Nature 402: 187–191.
Burt A and Trivers R (2006) Genes in Conflict: The Biology of Selfish Genetic Elements. Cambridge, MA: Harvard University Press.
Haig D (2004) Genomic Imprinting and Kinship. Piscataway, NJ: Rutgers University Press.
Jablonka E and Lamb MJ (2005) Evolution in Four Dimensions: Genetic, Epigenetic, Behavioral, and Symbolic. Cambridge, MA: MIT Press.
Rakyan VK and Beck S (2006) Epigenetic variation and inheritance in mammals. Current Opinion in Genetics and Development 16(6): 573–577.
Richards EJ (2006) Inherited epigenetic variation – revisiting soft inheritance. Nature Reviews. Genetics 7(5): 395–401.
John Wiley & Sons, Ltd.
eLS subject area: Genetics & Disease
How to cite: Stevens, Joshua B; Abdallah, Batoul Y; Horne, Steven D; Liu, Guo; Bremer, Steven W; and Heng, Henry H (October 2011) Genetic and Epigenetic Heterogeneity in Cancer. In: eLS. John Wiley & Sons, Ltd: Chichester.
DOI: 10.1002/9780470015902.a0023592
Advanced article
Cancer is commonly viewed as a disease of the stepwise accumulation of gene mutations. However, genetic and epigenetic heterogeneity (GEH) is pervasive in cancer, playing a key role in promoting cancer progression. GEH occurs at three primary levels, the genome, gene and epigenetic levels and increases during the aging process and during stress. GEH at the genome level plays the largest role of the three in cancer evolution, as genome level change creates new cellular systems whereas genetic and epigenetic level change mainly modify the existing system. This system replacement achieved by genome level change is essential for cancer evolution. GEH challenges traditional molecular-based cancer research that focuses on common gene mutations in linear models of progression. Clinically, GEH must be monitored in order to determine the evolutionary potential of a tumour. Increased knowledge about GEH will benefit basic research and cancer treatment.
Cancer progression is traditionally thought to be a disease of a stepwise accumulation of common gene mutations (Nowell, 1976; Hanahan and Weinberg, 2000; Vogelstein and Kinzler, 2004). Genetic change that is not shared is thought to be unimportant and is treated as noise from which a distinct genetic pattern can be derived (Heng, 2007a, b; Heng et al., 2011a, b). Recent work has shown that cancer is a system disease defined by a high degree of genetic and epigenetic heterogeneity (GEH) particularly at the genome level (Heng et al., 2009, 2010a, b). By definition, heterogeneity is the lack of uniformity in a substance. Although the presence of heterogeneity within biological systems is generally acknowledged, it has often been ignored to simplify the analysis. For example, cells, tissues and individuals are often treated as homogeneous entities. Every level of the bio-system exhibits heterogeneity, from gene expression, protein folding, up through organisation of the genome, tissue and individual. GEH is critically important as it is heritable (Table 1), and heritability is a key factor of somatic evolution. Broadly, heterogeneity impacts cancer at two fundamental levels, the intrapopulation and the intraorganism levels. At the intrapopulation level, population heterogeneity is well accepted; forming the basis for population-based genetic studies. Cancer research has taken a traditional genetic approach, focusing on identification of disease contributing alleles. This has met with success in a limited number of cancer types with a strong genetic component such as breast cancer patients with BRCA gene defects. However in the case of sporadic cancer, specific gene defects are not shared between patients (Heng, 2010).
Table 1 Contributors to genetic and epigenetic heterogeneity
| Gene level heterogeneity Gene amplification Insertions Loss of heterozygosity Minor alleles Moveable elements Mutations Single-nucleotide polymorphisms Deletions Splice site variants Tri-nucleotide repeats |
| Epigenetic level heterogeneity 3-dimensional chromatin domain Chromatin folding DNA methylation Histone modification Noncoding RNA Nuclear matrix attachment/loop regulation Transcription factor binding |
| Genome level heterogeneity Chromosome level Aneuploidy Polyploidy Chromosome fragmentation Defective mitotic figures Micronuclei Multiple nuclei 3D-telomere behaviour Double minute chromosomes Inversions Large region amplification Deletions Translocations |
| Subchromosome level Copy number variation Cryptic translocations |