Epigenetics in evolution and disease

Abstract

We are not our genes. Genes are just part of the story. We cannot fully blame our genome for our behaviour and susceptibility to disease. In Lehninger's classic textbook for students of medicine and biology we can find a more accurate definition: we are our proteins (and our carbohydrates, fat, and so on). This more precise definition relates to the central dogma of molecular biology, that our proteins are generated from our DNA via an RNA intermediate state. These RNA molecules are translated into proteins. However, some RNA molecules do not encode proteins, and they have developed very specific and important cellular functions.The base-pair nucleotide sequence of DNA, the typical subject of study in classic genetics, cannot completely explain the functionality of our cells, their disruption in complex diseases, or the definition of our species. We need something else. Part of the explanation is provided by epigenetics. Conrad Waddington defined epigenetics in 1939 as “the causal interactions between genes and their products, which bring the phenotype into being”. From our current knowledge we can define epigenetics as “the inheritance of DNA activity that does not depend on the naked DNA sequence”.Epigenetics refers to the dynamic chemical modifications that occur to our DNA and its subsequent association with regulatory proteins. The best recognised epigenetic modifications, or marks, are the addition of a methyl group to DNA and the post-translational modifications to histones. The histone proteins are responsible for organising DNA into nucleosomes and higher-order chromatin fibres. Epigeneticists are still discovering additional epigenetic systems, with recent descriptions of mechanisms involving the formation of high-order structures by DNA-histone complexes, the positioning of nucleosomes, and the activity of many non-protein-coding RNAs (such as microRNAs and antisense RNAs), but these would be the subject of another article.DNA methylation has crucial roles in the control of gene activity and nuclear architecture. In human beings, DNA methylation occurs at cytosine within CpG dinucleotides. These sites are not randomly distributed in the human genome; CpG-rich regions, known as CpG islands, are commonly associated with the 5L' regulatory region of many genes and are usually unmethylated in normal cells. This unmethylated status enables genes that contain CpG islands to be transcribed in the presence of the necessary transcriptional activators. However, there is a subset of promoter CpG islands that are heavily methylated and these are often associated with tissue-specific and germline-specific genes, imprinted genes (ie, genes that are expressed only from one copy, the maternal or paternal), and genes that undergo X-chromosome inactivation in females. In addition, repetitive genomic sequences are also heavily methylated. The maintenance of the methylation state may have a role in protecting chromosome integrity by preventing chromosomal instability. DNA methylation is not an isolated epigenetic mark. It is commonly associated with chemical modifications to the N-terminal tails of histone proteins. Once considered mere DNA-packaging proteins, histones now take centre stage as stores of epigenetic information through a complex set of post-translational modifications such as lysine acetylation, arginine and lysine methylation, and serine phosphorylation. It has been proposed that distinct patterns of modifications presented on histone tails form an inherited histone code for gene activity.If we move from the cellular standpoint to the whole individual, even more factors have roles in epigenetic modifications and their inheritance, such as inter-individual and environmental differences. Genetic variants in the enzymes involved in the metabolism of the methyl and acetyl chemical groups, for both DNA and histones, and dietary factors may be important. For example, diets deficient in methyl-group donors such as choline and methionine, or in coenzymes of methyl-group metabolism such as folate and vitamin B12, have long been known to affect concentrations of the universal methyl-donor S-adenosyl-L-methionine, causing DNA hypomethylation. Other examples include the genetic variants of the methylenetetrahydrofolate reductase gene that have been linked with the risk of neural-tube defects and vascular disease; and methionine synthase deficiency leading to megaloblastic anaemia. Thus, a flexible, but loyal, pattern of DNA methylation and histone modification is essential for cell, tissue, and organism functionality.There are many instances of real-life observations confronting traditional genetics. The most obvious cases involve monozygotic twins. These individuals are natural clones, identical at the level of DNA sequence. However, the penetrance of various diseases in these individuals can differ substantially, and discordant twins for a particular disorder have long puzzled biomedical researchers and physicians. One example of discordance occurs in monozygotic twin sisters carrying the same penetrant germline mutation for the breast-cancer hereditary gene BRCA1: one twin develops breast cancer at age 35 years, but the disease might not develop in the other for another 30 years. The discordance also occurs in carriers of high-penetrant genetic alterations in other cancer-prone syndromes such as hereditary non-polyposis colorectal cancer associated with hMLH1 mutations. These observations therefore suggest a role for the environment in the susceptibility and presentation of disease. But how is the environment involved? It seems easier to change the epigenetic setting of a cell than to change its genetic material. In this regard, epigenetics has also been proposed as a translator between the environment and genetics. We have recently solved a small part of the mystery by showing that monozygotic twins present an epigenetic drift in their epigenetic modifications, which is accentuated by their increasing age, decreasing amount of time shared together, and lifestyle differences, including smoking. Other researchers have confirmed these findings, and we now have to address future challenges by identifying the epigenetic changes in particular loci that could explain discordance for a phenotype in monozygotic twins, with attention focusing on cancer, diabetes, and autoimmune disorders.Monozygotic twins are not the only puzzle: there are many more cases in which epigenetics might explain how the same genotype can produce different phenotypes. For example, cloned animals artificially generated from the same donor DNA have the same DNA sequence; they should be identical. But they are not. Individual cloned mice, cats, and sheep are not identical to their unique donor “parent”: they develop diseases with different penetrance and they show disrupted epigenetic patterns. A controversial area is assisted reproduction: epidemiological data suggest that children born by use of these technologies have a higher likelihood of developing imprinting disorders, such as Beckwith-Wiedemann syndrome, with the possibility that a distinct DNA methylation profile is involved. Assisted reproductive techniques have helped many families, but some researchers have suggested a call for future research, particularly thinking about the children of those born after use of these methods.One of the most surprising results from the completion of species-specific genome sequencing is how similar we all are. The mice genome does not differ much from the human genome. How then, can we explain the evident differences? The chimpanzee genome is almost a copy of the human genome. What then are the mechanisms involved in making us human? Epigenetics can come to the rescue for some of these questions. Although the discipline of comparative epigenetics is in its infancy, there are DNA methylation patterns that are conserved between species, whereas others are distinct, without any evident explanation in the underlying DNA sequence. Conservation of gene expression among different organs in different species is associated with conservation of DNA methylation profiles: for example, a mouse and a human being might both have DNA methylation in a gene that is not expressed in the brain, but in other organs of both species where the gene is expressed, that gene locus would remain unmethylated. Even more interestingly, we might imagine that subfamilies of odour receptors are unmethylated in chimpanzee DNA, but methylated in human DNA, since people, unlike chimpanzees, do not need to distinguish numerous subtypes of odours to survive. The converse is probably also true; particular subgroups of neurotransmitters in regions of the brain could be unmethylated in people but methylated in chimpanzee DNA.Another unresolved issue is the adaptation of species to a changing environment. How can the rigid structure of our genome, with its very efficient and loyal DNA repair system, bring about all the necessary evolutionary changes observed in a short period? By contrast, epigenetics is dynamic and can adapt easily. Of course, Lamarck's theory that an organism can pass on characteristics acquired during its lifetime to its offspring, taken to its extreme in experiments in which the tails of mice were cut off, only demonstrates that generational change is not a simple thing. However, we may feel more comfortable with the idea that DNA methylation and histone modification changes caused by environment can affect our germinal cells, and this new subtle difference in epigenetic setting could be transmitted to our offspring. In human beings, epidemiological studies suggest that mothers who were growth restricted in utero may themselves give birth to low-birthweight offspring. The first altered epi-alleles (epigenetic variants of alleles) for human disease are being reported, with the description of caudal duplication anomalies as an example.Epigenetic disruption is a characteristic of human cancer. The reduction of the total amounts of DNA methylation in human tumours compared with their normal counterparts was one of the first epigenetic alterations described in tumours. This reduction arises mainly through DNA hypomethylation of repetitive DNA sequences and demethylation of the gene bodies (coding regions and introns). Global DNA hypomethylation contributes to the origin of cancer cells by generation of chromosomal instability, reactivation of transposable elements, and loss of imprinting. Most importantly, the DNA methylation paradox, there are local areas of DNA that gain CpG methylation—the promoter CpG islands of many tumour-suppressor genes, such as hMLH1, BRCA1, and p16INK4a—leading to the inactivation of these cancer-protecting proteins. MicroRNAs with tumour-suppressor functions are also silenced in cancer cells by DNA hypermethylation. Human tumours also show a distorted histone code. In leukaemia, we know that the pathognomonic translocations involve histone acetyltransferase and methyltransferase genes.Epigenetics seems to have a central role in cancer at a cellular evolution level. Human tumours undergo massive and adaptive changes in their natural history: the cancer can metastasise to distant sites, it can create new blood and lymph vessels to feed on and eliminate its metabolites, it can also change in response to treatment with drugs, hormones, or radiation. The cancer cell has limited ability to undergo fast genetic changes to adapt to the hostile cellular microenvironment. However, selection of cancer cells is permitted by the production of “fitter” cells through rapid, randomly occurring epigenetic changes; within 48 h of an external stimulus, the DNA methylation and histone modification patterns of transformed cells can be completely altered. We can take the example of a breast cancer to illustrate this process. The cell adherence E-cadherin gene can become methylated and silenced, inducing the formation of metastases in the rib. If the cancer cells now located in that bone are to survive, they have to establish an interaction with their new surroundings; subsequent loss of the DNA methylation at this locus will promote their survival. Another interesting example is a glioma with DNA methylation-associated inactivation of the DNA repair enzyme O6-methylguanine DNA methyltransferase, which is predictive of a good response to chemotherapy. However, once the treatment has started the tumour can evolve, selecting those cells that are unmethylated at the gene for this enzyme, producing chemoresistance with a purely epigenetic basis.Cactus flowers.View Large Image Copyright © 2008 PhotolibraryOne of the essential differences between human cancer genetics and epigenetics is that DNA methylation and histone modification changes are reversible under the right circumstances. Thus, epigenetic alterations are one of the weakest points in the defences of the cancer cell, because those hypermethylated tumour-suppressor genes in their long “sleep” can be awoken and reactivated with the right drug regimens and exert their normal growth-inhibitory functions. Two families of epigenetic drugs, DNA-demethylating agents and inhibitors of histone deacetylase, have emerged as the most promising compounds in this area, and four pharmaceutical compounds have received approval for the treatment of specific leukaemia and lymphoma subtypes. The successful story in these malignant disorders needs now to be translated to epithelial solid tumours.The recognition of the importance of epigenetics in human disease initiation and progression started in oncology, but other specialties are now involved, such as neurodevelopment and neurodegenerative disorders, psychiatric entities, cardiovascular diseases, and autoimmune disorders. There are many classic monogenic syndromes in which the mutated gene is an epigenetic regulator. For example, one of the most common causes of mental retardation in women is Rett's syndrome, which is associated with the disruption of MeCP2, a protein that binds to methylated DNA. Other genetic disorders affecting epigenetic genes are the Rubenstein-Taybi syndrome (CBP/p300 mutations), X-linked α-thalassaemia/mental retardation syndrome (ATRX mutations), and immunodeficiency, centromere instability, facial anomalies syndrome (mutations in DNA methyltransferase 3B). Autoimmune diseases such as systemic lupus erythematosus and rheumatoid arthritis are characterised by important genomic DNA hypomethylation events. DNA methylation is also relevant for the histo-blood group ABO genes, the class I human leucocyte antigens, and the melanoma antigen genes. Imprinting diseases are also examples of a precise allelic defect in DNA methylation causing the development of reciprocal phenotypes such as Beckwith-Wiedemann and Silver-Russell syndromes, and Prader-Willi and Angelman syndromes. Finally, the contribution of epigenetic aberrations in the formation and progression of mental health disorders (schizophrenia and bipolar entities) will become clear in the next few years, and in Alzheimer's disease for which the first aberrant DNA methylation maps are appearingSome readers might think that the contribution of genetics is not sufficiently acknowledged here. This is not the case. In fact, in cancer research, no human tumour has been studied that can be completely explained by epigenetics. However, nor have we found any neoplasm that is completely caused by genetic alterations. Even in the most high-risk familial cancer cases, the tumours are the result of a mixture of genetic and epigenetic alterations. Genetics has had several jump-starts since the late recognition of Mendel's work, the elucidation of the double-strand DNA structure by Watson, Crick, Wilkins, and Franklin, and the completion of the human genome projects led by Francis Collins and Craig Venter. In epigenetics, we are still trying to catch up. What lies ahead could be even more exciting. We are not our genes. Genes are just part of the story. We cannot fully blame our genome for our behaviour and susceptibility to disease. In Lehninger's classic textbook for students of medicine and biology we can find a more accurate definition: we are our proteins (and our carbohydrates, fat, and so on). This more precise definition relates to the central dogma of molecular biology, that our proteins are generated from our DNA via an RNA intermediate state. These RNA molecules are translated into proteins. However, some RNA molecules do not encode proteins, and they have developed very specific and important cellular functions. The base-pair nucleotide sequence of DNA, the typical subject of study in classic genetics, cannot completely explain the functionality of our cells, their disruption in complex diseases, or the definition of our species. We need something else. Part of the explanation is provided by epigenetics. Conrad Waddington defined epigenetics in 1939 as “the causal interactions between genes and their products, which bring the phenotype into being”. From our current knowledge we can define epigenetics as “the inheritance of DNA activity that does not depend on the naked DNA sequence”. Epigenetics refers to the dynamic chemical modifications that occur to our DNA and its subsequent association with regulatory proteins. The best recognised epigenetic modifications, or marks, are the addition of a methyl group to DNA and the post-translational modifications to histones. The histone proteins are responsible for organising DNA into nucleosomes and higher-order chromatin fibres. Epigeneticists are still discovering additional epigenetic systems, with recent descriptions of mechanisms involving the formation of high-order structures by DNA-histone complexes, the positioning of nucleosomes, and the activity of many non-protein-coding RNAs (such as microRNAs and antisense RNAs), but these would be the subject of another article. DNA methylation has crucial roles in the control of gene activity and nuclear architecture. In human beings, DNA methylation occurs at cytosine within CpG dinucleotides. These sites are not randomly distributed in the human genome; CpG-rich regions, known as CpG islands, are commonly associated with the 5L' regulatory region of many genes and are usually unmethylated in normal cells. This unmethylated status enables genes that contain CpG islands to be transcribed in the presence of the necessary transcriptional activators. However, there is a subset of promoter CpG islands that are heavily methylated and these are often associated with tissue-specific and germline-specific genes, imprinted genes (ie, genes that are expressed only from one copy, the maternal or paternal), and genes that undergo X-chromosome inactivation in females. In addition, repetitive genomic sequences are also heavily methylated. The maintenance of the methylation state may have a role in protecting chromosome integrity by preventing chromosomal instability. DNA methylation is not an isolated epigenetic mark. It is commonly associated with chemical modifications to the N-terminal tails of histone proteins. Once considered mere DNA-packaging proteins, histones now take centre stage as stores of epigenetic information through a complex set of post-translational modifications such as lysine acetylation, arginine and lysine methylation, and serine phosphorylation. It has been proposed that distinct patterns of modifications presented on histone tails form an inherited histone code for gene activity. If we move from the cellular standpoint to the whole individual, even more factors have roles in epigenetic modifications and their inheritance, such as inter-individual and environmental differences. Genetic variants in the enzymes involved in the metabolism of the methyl and acetyl chemical groups, for both DNA and histones, and dietary factors may be important. For example, diets deficient in methyl-group donors such as choline and methionine, or in coenzymes of methyl-group metabolism such as folate and vitamin B12, have long been known to affect concentrations of the universal methyl-donor S-adenosyl-L-methionine, causing DNA hypomethylation. Other examples include the genetic variants of the methylenetetrahydrofolate reductase gene that have been linked with the risk of neural-tube defects and vascular disease; and methionine synthase deficiency leading to megaloblastic anaemia. Thus, a flexible, but loyal, pattern of DNA methylation and histone modification is essential for cell, tissue, and organism functionality. There are many instances of real-life observations confronting traditional genetics. The most obvious cases involve monozygotic twins. These individuals are natural clones, identical at the level of DNA sequence. However, the penetrance of various diseases in these individuals can differ substantially, and discordant twins for a particular disorder have long puzzled biomedical researchers and physicians. One example of discordance occurs in monozygotic twin sisters carrying the same penetrant germline mutation for the breast-cancer hereditary gene BRCA1: one twin develops breast cancer at age 35 years, but the disease might not develop in the other for another 30 years. The discordance also occurs in carriers of high-penetrant genetic alterations in other cancer-prone syndromes such as hereditary non-polyposis colorectal cancer associated with hMLH1 mutations. These observations therefore suggest a role for the environment in the susceptibility and presentation of disease. But how is the environment involved? It seems easier to change the epigenetic setting of a cell than to change its genetic material. In this regard, epigenetics has also been proposed as a translator between the environment and genetics. We have recently solved a small part of the mystery by showing that monozygotic twins present an epigenetic drift in their epigenetic modifications, which is accentuated by their increasing age, decreasing amount of time shared together, and lifestyle differences, including smoking. Other researchers have confirmed these findings, and we now have to address future challenges by identifying the epigenetic changes in particular loci that could explain discordance for a phenotype in monozygotic twins, with attention focusing on cancer, diabetes, and autoimmune disorders. Monozygotic twins are not the only puzzle: there are many more cases in which epigenetics might explain how the same genotype can produce different phenotypes. For example, cloned animals artificially generated from the same donor DNA have the same DNA sequence; they should be identical. But they are not. Individual cloned mice, cats, and sheep are not identical to their unique donor “parent”: they develop diseases with different penetrance and they show disrupted epigenetic patterns. A controversial area is assisted reproduction: epidemiological data suggest that children born by use of these technologies have a higher likelihood of developing imprinting disorders, such as Beckwith-Wiedemann syndrome, with the possibility that a distinct DNA methylation profile is involved. Assisted reproductive techniques have helped many families, but some researchers have suggested a call for future research, particularly thinking about the children of those born after use of these methods. One of the most surprising results from the completion of species-specific genome sequencing is how similar we all are. The mice genome does not differ much from the human genome. How then, can we explain the evident differences? The chimpanzee genome is almost a copy of the human genome. What then are the mechanisms involved in making us human? Epigenetics can come to the rescue for some of these questions. Although the discipline of comparative epigenetics is in its infancy, there are DNA methylation patterns that are conserved between species, whereas others are distinct, without any evident explanation in the underlying DNA sequence. Conservation of gene expression among different organs in different species is associated with conservation of DNA methylation profiles: for example, a mouse and a human being might both have DNA methylation in a gene that is not expressed in the brain, but in other organs of both species where the gene is expressed, that gene locus would remain unmethylated. Even more interestingly, we might imagine that subfamilies of odour receptors are unmethylated in chimpanzee DNA, but methylated in human DNA, since people, unlike chimpanzees, do not need to distinguish numerous subtypes of odours to survive. The converse is probably also true; particular subgroups of neurotransmitters in regions of the brain could be unmethylated in people but methylated in chimpanzee DNA. Another unresolved issue is the adaptation of species to a changing environment. How can the rigid structure of our genome, with its very efficient and loyal DNA repair system, bring about all the necessary evolutionary changes observed in a short period? By contrast, epigenetics is dynamic and can adapt easily. Of course, Lamarck's theory that an organism can pass on characteristics acquired during its lifetime to its offspring, taken to its extreme in experiments in which the tails of mice were cut off, only demonstrates that generational change is not a simple thing. However, we may feel more comfortable with the idea that DNA methylation and histone modification changes caused by environment can affect our germinal cells, and this new subtle difference in epigenetic setting could be transmitted to our offspring. In human beings, epidemiological studies suggest that mothers who were growth restricted in utero may themselves give birth to low-birthweight offspring. The first altered epi-alleles (epigenetic variants of alleles) for human disease are being reported, with the description of caudal duplication anomalies as an example. Epigenetic disruption is a characteristic of human cancer. The reduction of the total amounts of DNA methylation in human tumours compared with their normal counterparts was one of the first epigenetic alterations described in tumours. This reduction arises mainly through DNA hypomethylation of repetitive DNA sequences and demethylation of the gene bodies (coding regions and introns). Global DNA hypomethylation contributes to the origin of cancer cells by generation of chromosomal instability, reactivation of transposable elements, and loss of imprinting. Most importantly, the DNA methylation paradox, there are local areas of DNA that gain CpG methylation—the promoter CpG islands of many tumour-suppressor genes, such as hMLH1, BRCA1, and p16INK4a—leading to the inactivation of these cancer-protecting proteins. MicroRNAs with tumour-suppressor functions are also silenced in cancer cells by DNA hypermethylation. Human tumours also show a distorted histone code. In leukaemia, we know that the pathognomonic translocations involve histone acetyltransferase and methyltransferase genes. Epigenetics seems to have a central role in cancer at a cellular evolution level. Human tumours undergo massive and adaptive changes in their natural history: the cancer can metastasise to distant sites, it can create new blood and lymph vessels to feed on and eliminate its metabolites, it can also change in response to treatment with drugs, hormones, or radiation. The cancer cell has limited ability to undergo fast genetic changes to adapt to the hostile cellular microenvironment. However, selection of cancer cells is permitted by the production of “fitter” cells through rapid, randomly occurring epigenetic changes; within 48 h of an external stimulus, the DNA methylation and histone modification patterns of transformed cells can be completely altered. We can take the example of a breast cancer to illustrate this process. The cell adherence E-cadherin gene can become methylated and silenced, inducing the formation of metastases in the rib. If the cancer cells now located in that bone are to survive, they have to establish an interaction with their new surroundings; subsequent loss of the DNA methylation at this locus will promote their survival. Another interesting example is a glioma with DNA methylation-associated inactivation of the DNA repair enzyme O6-methylguanine DNA methyltransferase, which is predictive of a good response to chemotherapy. However, once the treatment has started the tumour can evolve, selecting those cells that are unmethylated at the gene for this enzyme, producing chemoresistance with a purely epigenetic basis. One of the essential differences between human cancer genetics and epigenetics is that DNA methylation and histone modification changes are reversible under the right circumstances. Thus, epigenetic alterations are one of the weakest points in the defences of the cancer cell, because those hypermethylated tumour-suppressor genes in their long “sleep” can be awoken and reactivated with the right drug regimens and exert their normal growth-inhibitory functions. Two families of epigenetic drugs, DNA-demethylating agents and inhibitors of histone deacetylase, have emerged as the most promising compounds in this area, and four pharmaceutical compounds have received approval for the treatment of specific leukaemia and lymphoma subtypes. The successful story in these malignant disorders needs now to be translated to epithelial solid tumours. The recognition of the importance of epigenetics in human disease initiation and progression started in oncology, but other specialties are now involved, such as neurodevelopment and neurodegenerative disorders, psychiatric entities, cardiovascular diseases, and autoimmune disorders. There are many classic monogenic syndromes in which the mutated gene is an epigenetic regulator. For example, one of the most common causes of mental retardation in women is Rett's syndrome, which is associated with the disruption of MeCP2, a protein that binds to methylated DNA. Other genetic disorders affecting epigenetic genes are the Rubenstein-Taybi syndrome (CBP/p300 mutations), X-linked α-thalassaemia/mental retardation syndrome (ATRX mutations), and immunodeficiency, centromere instability, facial anomalies syndrome (mutations in DNA methyltransferase 3B). Autoimmune diseases such as systemic lupus erythematosus and rheumatoid arthritis are characterised by important genomic DNA hypomethylation events. DNA methylation is also relevant for the histo-blood group ABO genes, the class I human leucocyte antigens, and the melanoma antigen genes. Imprinting diseases are also examples of a precise allelic defect in DNA methylation causing the development of reciprocal phenotypes such as Beckwith-Wiedemann and Silver-Russell syndromes, and Prader-Willi and Angelman syndromes. Finally, the contribution of epigenetic aberrations in the formation and progression of mental health disorders (schizophrenia and bipolar entities) will become clear in the next few years, and in Alzheimer's disease for which the first aberrant DNA methylation maps are appearing Some readers might think that the contribution of genetics is not sufficiently acknowledged here. This is not the case. In fact, in cancer research, no human tumour has been studied that can be completely explained by epigenetics. However, nor have we found any neoplasm that is completely caused by genetic alterations. Even in the most high-risk familial cancer cases, the tumours are the result of a mixture of genetic and epigenetic alterations. Genetics has had several jump-starts since the late recognition of Mendel's work, the elucidation of the double-strand DNA structure by Watson, Crick, Wilkins, and Franklin, and the completion of the human genome projects led by Francis Collins and Craig Venter. In epigenetics, we are still trying to catch up. What lies ahead could be even more exciting.