Epigenetics of discordant monozygotic twins: implications for disease
Genome Medicinevolume 6, Article number: 60 (2014)
Monozygotic (MZ) twins share nearly all of their genetic variants and many similar environments before and after birth. However, they can also show phenotypic discordance for a wide range of traits. Differences at the epigenetic level may account for such discordances. It is well established that epigenetic states can contribute to phenotypic variation, including disease. Epigenetic states are dynamic and potentially reversible marks involved in gene regulation, which can be influenced by genetics, environment, and stochastic events. Here, we review advances in epigenetic studies of discordant MZ twins, focusing on disease. The study of epigenetics and disease using discordant MZ twins offers the opportunity to control for many potential confounders encountered in general population studies, such as differences in genetic background, early-life environmental exposure, age, gender, and cohort effects. Recently, analysis of disease-discordant MZ twins has been successfully used to study epigenetic mechanisms in aging, cancer, autoimmune disease, psychiatric, neurological, and multiple other traits. Epigenetic aberrations have been found in a range of phenotypes, and challenges have been identified, including sampling time, tissue specificity, validation, and replication. The results have relevance for personalized medicine approaches, including the identification of prognostic, diagnostic, and therapeutic targets. The findings also help to identify epigenetic markers of environmental risk and molecular mechanisms involved in disease and disease progression, which have implications both for understanding disease and for future medical research.
Epigenetics emerged during the first half of the 20th century as the study of biological mechanisms involved in embryonic development and cell differentiation []. More recently, it has been defined as the study of nuclear inheritance through cell division that is not based on differences in DNA sequence [] and includes any mechanism that alters gene expression without altering DNA sequence. DNA methylation of cytosines at CpG dinucleotides was proposed as a mechanism of mammalian gene regulation in 1975 [,], and as it is the best studied epigenetic mechanism in human populations we will concentrate on it in this review. Typically, CpG methylation at the 5' regions of genes reduces gene expression. This downregulation is a result of either the inability of specific transcription factors to bind methylated CpGs or the recruitment of methyl-CpG-binding proteins (such as MeCP) with transcription repression activity [-]. Conversely, in gene body coding regions, patterns of high levels of methylation have been found in transcriptionally active genes []. Histone modifications, which are the next most studied epigenetic mark, are very diverse and may include acetylation, methylation, phosphorylation, ubiquitination, ADP-ribosylation, and others. It has been proposed that combinatorial modifications at selected residues trigger specific gene expression activity []. Less studied epigenetic regulators include histone variants [,], ATP-dependent chromatin remodeling complexes [], and non-coding RNAs [].
Apart from its key role in developmental biology, epigenetics has recently become relevant to epidemiology because it offers the promise of unraveling the biological mechanisms underlying disease and has potential as a biomarker of disease or of disease progression. In 1983, Feinberg and Vogelstein [] reported epigenetic alterations of the human growth hormone and γ-globin genes in colon cancer patients. Since then, epigenetic alterations have been reported in many other types of cancer [], autoimmune diseases [], diabetes [], Alzheimer's disease [], Parkinson's disease [], asthma [], and multiple other human complex traits [-]. The majority of reported associations between epigenetic changes and phenotypic variation were observed in population samples of unrelated individuals. However, a number of studies have also explored epigenetic profiles in twins during normal development, aging, and in the context of disease, using disease-discordant MZ twins.
Epigenetic disease studies can particularly benefit from the unique study design of disease-discordant MZ twins. The use of MZ twins allows us to study the role of epigenetics in disease by controlling for many potential confounders, such as genetic factors, age, gender, maternal effects, cohort effects, and most in utero and environmental influences (Figure 1). Both population-based and twin-based epigenetic studies are susceptible to bias from potential unobserved confounders, and require replication to minimize false positive findings. Here, we discuss the benefits, challenges (Box 1) and limitations (Box 2) of epigenetic studies using disease-discordant monozygotic (MZ) twins (also called identical twins), and we review recent findings and their implications for medical research.
Monozygotic twins and epigenetics
MZ twins arise from the same single cell and therefore share almost all of their genetic variants (Figure 1). Nevertheless, these isogenic individuals are not completely identical, but show phenotypic discordance for many traits from birth weight to a range of complex diseases []. Proband-wise concordance rate estimates show that MZ twins are relatively often discordant for most common complex diseases, such as type 1 diabetes (T1D; 61%) [], type 2 diabetes (41%) [], autism (58 to 60%) [], schizophrenia (58%) [], and different types of cancer (0 to 16%) []. This observation supports the finding that for many complex traits, genotype alone may not fully determine phenotypic variation, and the interplay between genes and environment needs to be considered. Epigenetics has been proposed to be one of the main mediators of this interaction [].
Fraga et al. [] first demonstrated that MZ twin pairs exhibit epigenetic differences by comparing total content of methylcytosine and histone acetylation levels of peripheral lymphocytes. Twin pairs who were older, spent less time together, or had more dissimilar health or medical histories showed greater differences in both types of epigenetic marks. In 2009, Kaminsky et al. [] confirmed the presence of DNA methylation differences between co-twins across three different tissues (white blood cells (WBCs), buccal epithelial cells, and gut biopsies) in an age- and sex-matched twin sample. They also observed greater similarity between DNA methylation profiles within MZ twins than within dizygotic (DZ) twins. Furthermore, they proposed a functional stratification of the epigenome based on the finding of lower variability between co-twins in CpG islands and promoters than in all non-CpG island loci. Similar region-specific variation in DNA methylation was found from profiling of the major histocompatibility complex (MHC) in CD4+ lymphocytes of 49 Norwegian MZ twins []. CpG islands, 5' regions, and conserved non-coding regions were less variable than CpG-poor regions within twins. The reported epigenetic differences between MZ co-twins appear to be established early on in life, as they were also observed in the neonatal epigenome in a longitudinal genome-wide study of DNA methylation that profiled 10 MZ and 5 DZ twin pairs at birth and at 18 months []. Interestingly, in the early years epigenetic profiles tended to diverge, converge, or remain discordant in a twin-pair-specific manner, showing high variability compared with adults.
The process of establishment and maintenance of epigenetic marks may be one explanation for the epigenetic discordance between MZ co-twins. DNA methylation is laid down by DNA methyltransferases and inherited through cell division, and epimutations (differences in epigenetic marks) can arise during this process. The DNA methyltransferases (DNMTs) Dnmt3a and Dnmt3b are involved in de novo methylation, whereas Dnmt1 maintains the marks across cell division. Methylcytosines are not added during DNA replication; as a consequence, the newly synthesized strands will lack methylation marks and the DNA molecules would be methylated only on one strand. Dnmt1 uses this hemi-methylated molecule as a substrate and methylates the corresponding nucleotides on the complementary strand. Epimutations, which can be inherited through cell division, may occur as probabilistic errors in the correct function of these enzymes. For example, Dnmt1 skips 4 to 5% of methylation sites and shows de novo methylation activity near densely methylated regions [].
Failure in the correct establishment of DNA methylation patterns has a key role in human genetic diseases known as imprinting disorders. Parent-of-origin-specific gene expression, known as genomic imprinting, is an epigenetic mechanism in which the imprinted gene is differentially methylated at the paternal and maternal alleles []. The region that controls the expression of an imprinted gene is known as an imprinting center. For example, methylation of the Prader-Willi imprinting center in chromosome 15q on the paternal allele causes Prader-Willi syndrome [], while failure to methylate the same locus on the maternal allele causes Angelman syndrome [], and loss of DNA methylation at the KCNQ1OT1 gene on the maternal allele causes Beckwith-Wiedemann syndrome []. In two separate studies using twins discordant for Beckwith-Wiedemann syndrome, Weksberg et al. [] and Bliek et al. [] found that the affected twins (5 of 5 MZ twin pairs and 10 of 10 MZ twin pairs, respectively) had the epimutation. Both studies also emphasize the fact that there is an excess of discordant female MZ twins in Beckwith-Wiedemann syndrome patients (8% [] and 2.5% []) compared with the general population (0.3 to 0.4%). This observation has suggested a link between epigenetic maintenance failures and the MZ twinning event, which has also previously been proposed by Bestor []. MZ twin discordance for other imprinting disorders has also been described, for example, in Silver-Russell syndrome [,], but in relatively small numbers.
Epigenetic differences in MZ twins may also arise as a result of environmental differences. MZ twins are exposed to different nutritional conditions related to the formation and vascularization of the placenta []. Because the intrauterine environment of MZ twins can differ, it is possible that intrauterine differences contribute to differences in epigenetic states (Box 2). DNA methylation profiles are less similar within pairs of monochorionic (shared placenta) MZ twins than within pairs of dichorionic (non-shared placenta) MZ twins [,]. This observation suggests that sharing a placenta may cause imbalanced in utero conditions and, consequently, more discordant epigenetic profiles. Another factor that may result in early-life epigenetic differences in twins is an unequal allocation of cells during the twinning event; however, not much detail is known about the precise mechanism of twinning. Environmental factors may also contribute to epigenetic differences in adult twins over time. Cigarette smoking is a known environmental modifier of DNA methylation. Smoking was first associated with differential DNA methylation in an adult population (50 to 60 years of age) at the F2RL3 gene, which encodes coagulation factor II receptor-like 3 []. Since then, multiple studies have identified and replicated smoking-associated changes in DNA methylation at several genes, including the well replicated aryl hydrocarbon receptor repressor (AHRR) gene [-] and several other regions. Interestingly, at multiple smoking-associated sites DNA methylation levels of former smokers are generally similar to levels observed in non-smokers, which suggests that smoking cessation may allow methylation levels to revert back. In addition to smoking, many other environmental factors have been linked to changes in DNA methylation, although not in humans, including diet, exposure to environmental toxins, temperature changes, and others []. As adult lifestyles of twins diverge, it is likely that their epigenomes may too. Traditionally, MZ twin discordance has been attributed only to environmental differences, but stochastic events, which are difficult to measure, may also play a role.
Lastly, although MZ twins are often referred to as being genetically identical, post-zygotic mutation events can occur (Box 2). Somatic point mutations during early development occur with a frequency of 1.2 × 10−7 per base pair per twin pair []. Similarly, somatic copy number variations have been found in normal concordant and in disease-discordant MZ twin pairs []. The influence of genetic variation on epigenetic marks has been previously demonstrated by the association of specific genetic variants at single nucleotide polymorphisms with specific DNA methylation sites []; it is therefore possible that somatic mutations or copy number variations in MZ twins may lead to epimutations in a cell- and tissue-specific manner.
Twin discordance for disease and environmental factors
Most complex phenotypes arise as a result of the interplay between genetics and environment. In epidemiology, it is of interest to determine what proportion of the phenotypic variance each of these factors can explain. Classical twin studies make use of MZ and DZ twins to decipher these influences. Because MZ twins are assumed to share almost 100% of their genetic variants, and DZ twins share, on average, only 50% of their variants, the difference in phenotype concordance levels between these two groups can indicate a genetic influence on the phenotype. Greater phenotype concordance in MZ twins would point to a higher contribution of genetics to the disease.
Heritability in the broad sense (H2) has been used to define the fraction of the total phenotypic variance () in a population that can be attributed to genetic variance () . Genetic variance can further be partitioned into the variance attributed to additive (), dominant (), and epistatic () effects. This partition gives origin to the definition of heritability (h2) in the narrow sense, which only considers the fraction of the total phenotypic variance attributed to additive genetic effects []. The rest of the variation that is not attributed to genetic effects is attributed to environmental influences, which can be divided into shared and unique environmental influences and random error. The classical twin estimate of heritability, known as h2, is defined as twice the difference between MZ and DZ intra-pair correlation coefficients (ICCs) of the trait of interest [h2 = 2(ICC MZ − ICC DZ )] []. The twin model also assumes that MZ and DZ twins equally share the environment and that a fraction of the correlation is due to the shared environmental influences. The fraction corresponding to shared environmental effects is estimated as the difference between the total correlation and the heritability estimate. Finally, the missing variation is attributed to unique environmental effects plus error. Figure 2 shows these proportions estimated by different studies for several diseases, including diabetes, psoriasis, autism, myocardial infarction, and different types of cancer.
Heritabilities can be calculated for particular epigenetic variants at specific epigenetic loci by treating them like phenotypes. Kaminsky et al. [] estimated the locus-specific DNA methylation ICC differences (ICC MZ − ICC DZ ) of 40 age- and sex-matched MZ and DZ twin pairs in buccal epithelial cells and WBCs at about 6,000 loci. The differences were significant in both tissues, but greater in buccal epithelial cells (mean ICC MZ − ICC DZ = 0.15 ± 0.0039, P = 1.2 × 10− 294) than in WBCs (mean ICC MZ − ICC DZ = 0.0073 ± 0.0034, P = 0.044), showing that at a proportion of sites in the genome DNA methylation levels show evidence for heritability. Gervin et al. [] estimated heritabilities of DNA methylation across CpG sites in the human MHC in CD4+ lymphocytes using bisulfite sequencing. Their relatively low heritability estimates (2 to 16%) suggested that DNA methylation variation in the MHC is mostly due to environmental factors. A subsequent larger study using whole blood samples from 172 female twins examined 26,690 CpG sites on the Illumina Infinium HumanMethylation27 BeadChip (Illumina 27 K) that includes CpG sites only at promoter regions, and revealed slightly higher DNA methylation heritability per CpG site, with an average of 18.2% across the genome []. DNA methylation heritability was supported by the significant association of genetic variants with 6.3% of the probes. More recently, Gordon et al. [], using the Illumina 27 K array, estimated heritabilities on three different tissues at birth: human umbilical vascular endothelial cells (HUVECs), cord blood mononuclear cells, and placenta. The results suggested a tissue-specific effect on heritability, in which only three of the top 5% most heritable sites were shared across all three tissues. However, given the low coverage of the Illumina 27 K array, it may be difficult to generalize these results. Decomposition of the epigenetic variance into genetic, shared environmental, and non-shared environmental effects in different twin studies so far show that, overall, non-shared environmental and stochastic factors may have a greater effect on methylation sites than genetic and shared environmental effects [,]. However, many of these studies rely on array technologies that may be subject to technical noise that may affect the accurate measurement of DNA methylation, particularly subtle changes []. Heritability estimates may also be confounded by the presence of cellular heterogeneity in the sample, which may introduce DNA methylation differences, for example, as shown in saliva samples [], and to a lesser extent in buccal cells [].
A different approach to study genetic effects on epigenetics focused on chromatin states, which are key to epigenetic mechanisms []. The authors explored discordant chromatin accessibility states associated with genetic variants in lymphoblastoid cells of 36 MZ twin pairs. Within-pair differences in chromatin accessibility were increased in the presence of somatic mutations and had a significant effect on gene expression discordance. Furthermore, quantitative trait locus mapping revealed chromatin accessibility differences at 1,325 loci associated with genotypic variants at a false discovery rate of 0.01. This integrative analysis using MZ twins shows how genetic factors affect chromatin states, potentially through epigenetic mechanisms, and ultimately gene expression.
Discordant lifestyles are obvious potential triggers for disease discordance. Epidemiological studies using disease-discordant MZ twins have identified environmental or lifestyle factors that increase or decrease the risk of developing a disease. Increased risk for basal cell carcinoma of the skin was associated with smoking status in females and decreased risk with outdoor work [], and because smoking is strongly linked to DNA methylation changes at multiple genes [], DNA methylation is a potential mediator of the molecular impacts of smoking on the disease in this case. Additional findings include moderate alcohol consumption, which associates positively with bone mineral density []. Furthermore, a study of MZ twins from the Older Australian Twins Study revealed that cognitive and social activity was associated with performance on some cognitive tasks []. Finally, habitual physical activity during adulthood enlarges the distal aorta and iliac and femoral artery diameters according to a study that included six middle-aged MZ twins with 32 years of discordance for physical activity []. Although the link between most lifestyle changes and specific epigenetic marks that may lead to disease is still unknown, MZ twins provide a good system to identify environmentally associated DNA methylation modifications and study their role in disease.
Low birth weight is an environmentally associated phenotype that has become of interest in epigenetics and disease because epidemiological studies have associated it with increased rates of coronary heart disease, stroke, hypertension, and type 2 diabetes []. Epigenetic studies have taken advantage of the birth weight differences that occur between MZ twin siblings, due to the competition for resources during their intrauterine development, to study methylation differences associated with birth weight. Gordon et al. [] identified differential methylation of a gene with vascular function, APOLD1, as being associated with birth weight after adjustment for multiple testing using umbilical vascular endothelial cells of 14 MZ twin pairs. A later study was conducted by Souren et al. [] using DNA extracted from saliva of 17 monochorionic MZ twin pairs with relative birth weight differences greater than 20%. At nominal significance and with absolute methylation differences greater than 0.05, 45 CpG sites had an uncorrected P-value <0.01 and absolute methylation differences greater than 0.05, but validation of those loci by bisulfite sequencing showed that the methylation differences may be due to technical variation.
Epigenetic studies of twins and disease
Multiple studies have explored the epigenetic changes in twins discordant for a range of diseases (Table 1). Here, we focus predominantly on recent genome-wide efforts across four major types of disease.
Many epigenetic aberrations have been identified in cancer. Two recent studies of cancer-discordant MZ twins [,] have provided interesting findings. High-resolution profiling of DNA methylation in whole blood samples from breast-cancer-discordant MZ twins led to the identification of DOK7 hypermethylation as a blood-based epigenetic biomarker that can be traced years before tumor diagnosis []. DOK7 encodes a docking protein that acts as a substrate and activator of receptor tyrosine kinase. Heyn et al. [] used the Infinium HumanMethylation450 BeadChip (Illumina 450 K array) to profile 15 MZ twin pairs at about 485,000 CpG sites genome-wide, focusing on genes, including promoters and gene bodies, as well as intergenic regions. The authors identified 403 positions that were differentially methylated in breast cancer in whole blood from 15 MZ twin pairs discordant for the disease. Among these positions they found a hypermethylated site located in an alternative promoter of the DOK7 gene with a methylation difference of only 2%. Further analysis revealed that the methylation signal spreads upstream of the promoter. Interestingly, the methylation of DOK7 was also found in samples taken, on average, 4.7 years before tumor diagnosis, implying that the epigenetic change occurs early in the development of the disease. Although the consequences of an altered DOK7 methylation status are unknown, these results indicate a promising role for DOK7 as a blood-based biomarker and serve as a model for other larger studies. In addition to this finding, the authors [] identified a breast cancer epigenetic signature that comprises 46 of the differentially methylated positions that can cluster samples according to the presence of cancer. Further study of the associated genes could give valuable clues to the pathogenesis of the disease.
Galetzka et al. [] studied an interesting case that suggests that epigenetic mosaicism is the source of variation between one pair of MZ twins discordant for childhood leukemia and secondary thyroid carcinoma. Using bisulfite pyrosequencing of target genes known to be associated with cancer (ATM, BRCA1, BRCA2, MLH1, RAD51C, and TP53), they found differential methylation of BRCA1 in skin fibroblasts and saliva. The difference in methylation levels in fibroblasts was 9% and in saliva was 7%. The methylation patterns of individual DNA molecules revealed epigenetic mosaicism in the affected twin, with 13% of her alleles exhibiting the epimutation. Although sequence variants have been linked to aberrant methylation patterns, no sequence alteration was found in the 5' regulatory region of the gene. Protein expression analysis demonstrated lower levels of BRCA1 in the affected twin, consistent with the idea that hypermethylation of tumor suppressor gene promoters is associated with gene silencing. Although the possibility of late somatic events as a consequence of treatment cannot be ruled out, the epigenetic mosaicism may have resulted from an epimutation during early embryonic stages that affected a subpopulation of cells. This study [] highlights the importance of early embryonic stages during the development of some epigenetic marks and not just adult-life environmental factors.
MZ twins are also useful for identifying genetic factors linked to epigenetic mechanisms that may lead to cancer. The study of a pair of MZ twins discordant for MLL gene-rearranged leukemia helped to identify mutations in a histone H3 lysine36 methyltransferase, SETD2, that causes changes in the epigenetic state of the cell []. A different approach used two pairs of MZ twins with acute lymphoblastic leukemia to study the timing of genomic lesions that may lead to this disease []. Thus, the interplay between genetics and epigenetics can also be studied using MZ twins.
One of the first genome-wide DNA methylation studies of autoimmune disease was performed by Javierre et al. [], who explored three autoimmune diseases with overlapping clinical signs and symptoms: systemic lupus erythematosus (SLE), rheumatoid arthritis, and dermatomyositis. DNA methylation of WBCs was assayed with a platform designed for cancer analysis (GoldenGate Methylation Cancer Panel I) that included 1,505 CpG sites from 807 genes. The cohort included five discordant MZ twin pairs for each disease and 30 unrelated normal controls. The authors [] found no differences in DNA methylation of the five MZ twin pairs for rheumatoid arthritis and dermatomyositis, but they found a set of 49 genes differentially methylated in SLE discordant twins with absolute mean methylation differences >0.1. Ontology analysis revealed that the differentially methylated regions (DMRs) were linked to genes enriched in functional processes related to autoimmune inflammatory diseases. Successful confirmation of the methylation differences at eight genes was carried out by pyrosequencing of additional related samples.
A more extensive recent epigenetic study of T1D used the Illumina 27 K array for genome-wide DNA methylation analysis of a specific subset of immune cells, monocytes, from 15 MZ twins discordant for T1D. The results revealed the presence of T1D-specific methylation variable positions in the T1D-affected co-twins, with effect sizes of small magnitude, from 0.13% to 6.6% []. The T1D-specific methylation variable positions were replicated in a set of four independent MZ pairs discordant for the disease. Furthermore, using samples from seven singletons before and after diagnosis, they demonstrated that methylation differences at these sites precede clinical diagnosis. DNA methylation changes associated with T1D were also observed by Stefan et al. [] using the same array and immortalized B cell lines from only three T1D-discordant and six T1D-concordant MZ twin pairs. The authors identified 88 differentially methylated sites present in all three discordant pairs and, interestingly, six of the identified genes are known to be associated with the disease from genome-wide association studies. These findings in part overlap with the results of Rakyan et al. []: both reported differentially methylated positions in the MHC region. Replication of these findings is pending, but may contribute to understanding the role of epigenetic mechanisms in the etiology of T1D.
Using an integrative approach, Gervin et al. [] combined genome-wide DNA methylation and gene expression profiles to study psoriasis in CD4+ cells from 17 discordant MZ twins and CD8+ cells from 13 discordant MZ twins pairs. The DNA methylation profiles were once again obtained using the Illumina 27 K array, and gene expression was explored using the Illumina HT-12 array. The independent methylation and expression analyses did not detect any significant difference in DNA methylation or expression, but the combined analysis revealed correlation between the co-twin DNA methylation difference and the log fold expression ratio of genes involved with immune response in CD4+ cells. These findings highlight important pathways for the study of the disease, but require further study with more extensive DNA methylation coverage, as well as replication in a larger cohort. Another integrative approach included the analysis of differentially methylated CpG sites characterized with the Illumina 27 K array, DMRs obtained by methylated DNA immunoprecipitation and array hybridization (MeDIP-chip), and differential gene expression interrogated with Affymetrix UG 133 Plus 2.0 arrays to study ulcerative colitis with intestinal biopsies of 10 discordant MZ twin pairs []. In total, 61 genes with differential expression and at least one differentially methylated position or region were defined as disease-associated. These integrative approaches allow the association of methylation with gene expression, but cannot determine causality. Different approaches are needed to establish whether DNA methylation is controlling gene expression.
One early study of epigenetics in schizophrenia used lymphocytes of one discordant and one concordant pair of MZ twins to characterize the 5’ regulatory region of the dopamine D2 receptor gene (DRD2) through bisulfite sequencing []. The results showed that the epigenetic profile of the affected twin was more similar to the profile of the affected concordant twins than to the profile of their unaffected co-twin. The finding was of relevance for studying epigenetics and schizophrenia, but involved only a single gene.
Recent genome-wide epigenetic studies have identified interesting findings in major psychosis and depression. Dempster et al. [] studied epigenetic changes in schizophrenia and bipolar disorder (BD) independently and as a major psychosis group. DNA methylation was assayed in blood samples from 22 MZ twin pairs discordant for schizophrenia or BD using the Illumina 27 K array. The strongest associated DMR in the schizophrenia group was found upstream of the tRNA pseudouridine synthase 3 gene (PUS3), in the BD group was upstream of the melanin-concentrating hormone receptor 1 gene (GPR24), and in the major psychosis group it was found in the promoter region of ST6GALNAC1. ST6GALNAC1 encodes a protein involved in glycosylation and cell-cell interactions, and the identified epigenetic change overlapped a previously reported rare genomic duplication observed in schizophrenia. The reported effect size at this epigenetic variant was relatively modest: differences in methylation between affected and unaffected twins were, on average, less than 10%. Some of the top-ranked differentially methylated sites in the two groups were shared, but disorder-specific alterations were also observed with DNA methylation changes in opposite directions. This study also supported the idea of epimutations across tissues with the observation of hypomethylation of the ST6GALNAC1 DMR in 4 out of 30 post-mortem brain tissue samples of affected individuals.
A different approach to study BD used methylation-sensitive representational difference analysis to characterize DNA methylation in lymphoblastoid cell lines of only one pair of discordant MZ twins []. One of the isolated differentially methylated fragments corresponded to the regulatory region of the pseudogene PPIEL. Its function is unknown, but a strong inverse correlation between the expression and DNA methylation levels of PPIEL was found.
Depression is another psychiatric disorder with interesting epigenetic findings. Using methylated DNA immunoprecipitation (MeDIP) with sequencing, Davies et al. [] identified hypermethylation of a region within the ZBTBT20 gene associated with major depressive disorder (MDD). ZBTB20 has an important role in the developing hippocampus, a region previously implicated in the development of MDD. The discovery cohort included 50 MDD-discordant MZ twin pairs and the finding was replicated in a cohort of 356 unrelated case–control individuals. Although a previous study using the 450 K array was unable to observe significant DNA methylation differences using 12 MDD-discordant MZ twin pairs [], differences in sample size and much greater genome coverage of the MeDIP-seq method in comparison with the 450 K array (which only covers 8.9% of dynamic CpGs []) may account for the differences. However, both studies reported a greater genome-wide methylation variance in the MDD-affected twins than in the unaffected siblings. This study is among the largest epigenetic twin studies published so far (50 MZ discordant twin pairs), with an independent replication sample, but lacks validation of the findings using an alternative DNA methylation detection technique.
Adolescent depression was also studied in 18 pairs of MZ twins with discordant scores on self-rated depression []. Buccal cells were characterized in this study because of their low heterogeneity compared with blood, and the DNA methylation assay was performed with the Illumina 450 K array. Two of the top 10 differentially methylated sites with methylation differences less than 10% were also found in post-mortem brain samples of patients with MDD. The methylation of these two sites shared across tissues has a potential use for development as biomarkers.
Neurological diseases and pain
One of the earliest epigenetic studies using disease-discordant MZ twins explored caudal duplication anomaly [], a condition characterized by duplication of organs in the caudal region. The authors used a candidate gene approach to examine DNA methylation at the promoter region of a gene implicated in these anomalies, AXIN1. Hypermethylation of the region was detected in the affected twin in peripheral blood mononuclear cells (PBMCs), but not in buccal cells. Because blood is mesodermal tissue whereas buccal cells are ectodermal, this result suggests that the epimutation occurred after the differentiation of the three germ layers and highlights the tissue-specific role of epigenetic marks.
The epigenetics of Alzheimer’s disease has also been studied, though with a limited sample size. An immunochemistry assay was performed on the temporal neocortex of a pair of discordant MZ twins []. With this technique it was only possible to establish that the affected twin possessed reduced levels of DNA methylation in temporal neocortex, but not in other parts of the brain. This study highlights the importance of tissue selection in epigenetic studies and the difficulty of finding markers shared across tissues.
Subsequent genome-wide efforts included a small epigenetic twin study of multiple sclerosis, which is a demyelinating disease that causes neurodegeneration. Baranzini et al. [] compared the genome-wide methylation of three MZ twin pairs discordant for multiple sclerosis using reduced representation bisulfite sequencing. CpG methylation differences were few, and no differential methylation was common in at least two pairs. This was the first attempt to study this disease in discordant MZ twins at this level of resolution, but the results were based on only three twin pairs. A larger sample size is needed to have reasonable statistical power to detect epigenetic changes. Furthermore, methylation was interrogated by a sequencing technology that only captures 11.5% of the dynamic CpG sites [].
Autism is a neurodevelopmental disorder with considerable twin discordance. One early approach with only three pairs of discordant MZ twins allowed the identification of 73 differentially methylated CpG islands in lymphoblastoid cell lines []. Methylation was interrogated with the 8.1 K CpG island microarray (UHN Microarray Center). The protein levels of two of the candidate genes, RORA and BCL-2, were decreased in autistic brains. In this case there is a link between the methylation levels in peripheral cells and the brain. Following this study [], whole blood DNA methylation profiles of MZ twins discordant for autism spectrum disorder (ASD) allowed the identification of DMRs associated with ASD-related traits: social autistic traits, autistic restricted repetitive behaviors and interests (RRBIs), and communication autistic traits []. The genome-wide study characterized 34 MZ twin pairs discordant for ASD or a related trait, 5 concordant for ASD, and 11 concordant for no autistic phenotype using the Illumina 27 K array. No global DNA methylation differences within discordant twin pairs were found, but site-specific differences were abundant. The top-ranked DMRs included genes already mentioned in the ASD literature (GABRB3, AFF2, NLGN2, JMJD1C, SNRPN, SNURF, UBE3A, and KCNJ10). Although this study explored a relatively large number of twin pairs, the findings have not yet been replicated and were based on a relatively low-coverage array (Illumina 27 K) that examines only 0.7% of variable CpG sites [].
Pain sensitivity is a complex phenotype, and genetic effects have been unable to fully explain the variation in pain sensitivity. Bell et al. [] examined DMRs associated with pain using 50 MZ twins discordant for heat pain tolerance and 50 unrelated individuals in two separate epigenome-wide association studies. Pain sensitivity was determined using the heat pain suprathreshold and DNA methylation profiles of whole blood were characterized by genome-wide MeDIP-seq. The highest association of the combined results was in the promoter of TRPA1, a known pain-sensitivity gene that encodes an ion channel expressed in sensory neurons, and the finding was validated using bisulfite sequencing. The observed effect at the CpG site validated by bisulfite sequencing was a 10% change in DNA methylation, which corresponded to a 2°C change in heat pain sensitivity. Pain sensitivity discordance was defined as a heat pain tolerance difference >2°C between twins. Additionally, most of the 100 top-ranked DMRs were found to be associated with genetic variants and to be stable over time. This study used a moderate sample size of 25 MZ twin pairs, combined with an unrelated sample of individuals, to identify and validate DMRs and to gain insight into underlying mechanisms of epigenetic association.
Multiple previous studies have also explored predominantly candidate genes in other complex human traits with high MZ twin discordance rates. A recent genome-wide study focused on a case of congenital renal agenesis []. In this study, no genetic alterations were confirmed between a pair of discordant MZ twins, but 514 DMRs were detected using reduced representation bisulfite sequencing. However, because DNA methylation differences were present in normal concordant MZ twins, further studies and larger cohorts would be needed to reach conclusive results.
The study of Down syndrome has also benefited from the use of MZ twins. Using one pair of MZ twins discordant for trisomy 21 (which causes Down syndrome) it was possible to determine that the differential expression between the twins was organized in domains along all chromosomes and that the H3K4me3 profile was altered []. Further study of the chromosomal arrangements caused by alterations in epigenetic states might help to provide further insights into this condition.
Implications for disease and medicine
Twins are a unique resource for elucidating complex molecular mechanisms of disease. Epigenetic findings from disease-discordant MZ twin studies so far have identified DNA methylation changes in multiple genes across a wide range of phenotypes. In some cases, such as DOK7 in breast cancer, the identified changes are likely biomarkers of disease but may also be involved in disease susceptibility as they pre-date the cancer diagnosis. In other examples, such as TRPA1 in pain susceptibility, ST6GALNAC1 in BD, ZBTB20 in MDD, and several genes identified in ASD, the peak findings are in genes that have already been implicated in the trait. However, it remains unclear whether these changes are causal or secondary to disease, and longitudinal studies are required to address this question. However, even if the identified epigenetic changes are secondary to disease, they can still improve our understanding of disease progression, especially for relapsing disorders such as BD.
Recent findings that smoking can be reliably detected in epigenetic patterns across tissues in humans also pave the way for further use of epigenetics as an epidemiology tool - using epigenetic changes as surrogates of the environment or risk factors. The use of epigenetic markers of environmental risk would greatly improve our understanding of the molecular basis of disease, as many complex traits have an environmental risk component that is often difficult to define and assess. Therefore, using epigenetic markers of environmental disease risk would help to identify environmentally driven disease mechanisms, including gene-environment interactions.
The study of epigenetics in disease using discordant MZ twins has present and future medical implications and translational potential. Finding novel epigenetic associations with disease has implications for the identification of potential targets of drugs and biomarkers for diagnosis and prognosis. Epigenetic drugs influencing DNA methylation and chromatin modifiers are capable of reversing aberrant epigenetic states and therefore the gene regulation of abnormal cells. Examples of such drugs include the DNMT inhibitors decitabine and azacitidine for the treatment of myelodysplastic syndrome, and the combined use of a DNMT inhibitor (hydralazine) and a histone deacetylase (HDAC) inhibitor (valproate) []. At present, at least another 40 epigenetic drugs are reported to be in development for cancer [].
New technologies also have the potential to allow the development of drugs for targeted epigenetic editing in the near future. Zinc-finger nucleases, transcription activator-like effector nucleases (TALENs), and the clustered regularly interspaced short palindromic repeats (CRISPRs) have allowed biological researchers to edit the genomes of various organisms []. The CRISPR/Cas complex, a RNA guided nuclease, is probably the most promising tool because of the relative ease with which experiments can be designed and loci targeted. The system consists of foreign DNA sequences integrated within CRISPR loci that are later transcribed. The transcript is recognized by Cas proteins and together causes the cleavage of specific sequences. The fusion of this technology with epigenetic modifiers can potentially allow targeted epigenome editing []. Zinc fingers coupled to the catalytic domain of Dnmt3a have been used successfully for the epigenetic reprogramming of a tumor suppressor gene (MASPIN) and oncogene (SOX2) in cancer cells []. Epigenetic editing targets the gene itself to silence expression, rather than targeting multiple copies of the transcript as other technologies such as RNA interference do, which makes silencing by epigenetic editing potentially more effective. Many of the findings derived from epigenetic studies could provide therapeutic targets for these types of new epigenetic drugs.
Epigenetic comparisons of MZ twins and heritability studies of DNA methylation highlight that most epigenetic variability is unique to the individual. This property makes epigenetics potentially valuable for personalized medicine. Future implications of these findings include the role of epigenetics in P4 medicine (personalized, predictive, preventive, and participative medicine), as a stable yet potentially reversible molecular mechanism.
At present, key challenges in epigenetic studies of disease-discordant MZ twins predominantly relate to sample size and power, choice and availability of appropriate tissue for the trait of interest, maximizing DNA methylation array coverage and sensitivity, and integrating epigenetic profiles with genetic, transcription, and environmental datasets (Box 1). The awareness of the value of discordant MZ studies is spreading and has led to the consolidation of international efforts, such as the EUroDiscotwin Consortium aimed at the combined collection of disease-discordant MZ twins from European and Australian Twin Registries. Furthermore, several large-scale epigenetic studies in existing twin cohorts are already under way. Whole blood is currently the sample of choice, and adjusting for cellular heterogeneity in blood samples should be a standard analytical tool in epigenetic analysis. In future, access to reference datasets of epigenetic profiles across multiple tissues will enable the establishment of tissue-specificity rates at particular epigenetic marks. As costs of sequencing technologies are constantly decreasing, reasonable genome-wide methylation coverage and sensitivity will become feasible in the near future. Lastly, methods for integrating multiple layers of functional genomic data, including epigenetic data, should be explored in the context of epigenome-wide studies of MZ discordant twins. With the use of systems biology it is increasingly possible to integrate all molecular levels of information from a patient to provide a personalized diagnosis and treatment options.
The main priorities for future research would be to explore in depth epigenetic profiles in disease-discordant MZ twins using high-resolution epigenetic assays, and replicate the findings in independent cohorts. Because twinning is a rare event, to accelerate progress large-scale efforts across international twin cohorts will be needed to reach sample sizes required to detect moderate epigenetic changes. Ultimately, longitudinal studies will be necessary to establish the timing of the epigenetic change with respect to disease onset, and to investigate its role in disease susceptibility or progression.
Autism spectrum disorder
Clustered regularly interspaced short palindromic repeats
Differentially methylated region
Intra-pair correlation coefficient
Major depressive disorder
Methylated DNA immunoprecipitation
Major histocompatibility complex
Peripheral blood mononuclear cells
Type 1 diabetes
White blood cell
Waddington CH: The epigenotype. Endeavour. 1942, 1: 18-20.
Holliday R: Epigenetics: an overview. Dev Genet. 1994, 15: 453-457.
Holliday R, Pugh JE: DNA modification mechanisms and gene activity during development. Science. 1975, 187: 226-232.
Riggs AD: X inactivation, differentiation, and DNA methylation. Cytogenet Cell Genet. 1975, 14: 9-25.
Iguchi-Ariga SM, Schaffner W: CpG methylation of the cAMP-responsive enhancer/promoter sequence TGACGTCA abolishes specific factor binding as well as transcriptional activation. Genes Dev. 1989, 3: 612-619.
Boyes J, Bird A: DNA methylation inhibits transcription indirectly via a methyl-CpG binding protein. Cell. 1991, 64: 1123-1134.
Nan X, Campoy FJ, Bird A: MeCP2 is a transcriptional repressor with abundant binding sites in genomic chromatin. Cell. 1997, 88: 471-481.
Ball MP, Li JB, Gao Y, Lee J-H, LeProust EM, Park I-H, Xie B, Daley GQ, Church GM: Targeted and genome-scale strategies reveal gene-body methylation signatures in human cells. Nat Biotechnol. 2009, 27: 361-368.
Strahl BD, Allis CD: The language of covalent histone modifications. Nature. 2000, 403: 41-45.
Sullivan KF, Hechenberger M, Masri K: Human CENP-A contains a histone H3 related histone fold domain that is required for targeting to the centromere. J Cell Biol. 1994, 127: 581-592.
Barrero MJ, Sese B, Martí M, Izpisua Belmonte JC: Macro histone variants are critical for the differentiation of human pluripotent cells. J Biol Chem. 2013, 288: 16110-16116.
Hirschhorn JN, Brown SA, Clark CD, Winston F: Evidence that SNF2/SWI2 and SNF5 activate transcription in yeast by altering chromatin structure. Genes Dev. 1992, 6: 2288-2298.
Aravin AA, Sachidanandam R, Bourc’his D, Schaefer C, Pezic D, Toth KF, Bestor T, Hannon GJ: A piRNA pathway primed by individual transposons is linked to de novo DNA methylation in mice. Mol Cell. 2008, 31: 785-799.
Feinberg AP, Vogelstein B: Hypomethylation distinguishes genes of some human cancers from their normal counterparts. Nature. 1983, 301: 89-92.
Feinberg AP, Tycko B: The history of cancer epigenetics. Nat Rev Cancer. 2004, 4: 143-153.
Ballestar E: Epigenetics lessons from twins: prospects for autoimmune disease. Clin Rev Allergy Immunol. 2010, 39: 30-41.
Rakyan VK, Beyan H, Down TA, Hawa MI, Maslau S, Aden D, Daunay A, Busato F, Mein CA, Manfras B, Dias K-RM, Bell CG, Tost J, Boehm BO, Beck S, Leslie RD: Identification of type 1 diabetes-associated DNA methylation variable positions that precede disease diagnosis. PLoS Genet. 2011, 7: e1002300-
Sanchez-Mut JV, Aso E, Panayotis N, Lott I, Dierssen M, Rabano A, Urdinguio RG, Fernandez AF, Astudillo A, Martin-Subero JI, Balint B, Fraga MF, Gomez A, Gurnot C, Roux J-C, Avila J, Hensch TK, Ferrer I, Esteller M: DNA methylation map of mouse and human brain identifies target genes in Alzheimer’s disease. Brain. 2013, 136: 3018-3027.
Masliah E, Dumaop W, Galasko D, Desplats P: Distinctive patterns of DNA methylation associated with Parkinson disease: identification of concordant epigenetic changes in brain and peripheral blood leukocytes. Epigenetics. 2013, 8: 1030-1038.
Salam MT, Zhang Y, Begum K: Epigenetics and childhood asthma: current evidence and future research directions. Epigenomics. 2012, 4: 415-429.
Kaminsky Z, Petronis A, Wang S-C, Levine B, Ghaffar O, Floden D, Feinstein A: Epigenetics of personality traits: an illustrative study of identical twins discordant for risk-taking behavior. Twin Res Hum Genet. 2008, 11: 1-11.
Heyn H, Ferreira HJ, Bassas L, Bonache S, Sayols S, Sandoval J, Esteller M, Larriba S: Epigenetic disruption of the PIWI pathway in human spermatogenic disorders. PLoS One. 2012, 7: e47892-
Zhang R, Miao Q, Wang C, Zhao R, Li W, Haile CN, Hao W, Zhang XY: Genome-wide DNA methylation analysis in alcohol dependence. Addict Biol. 2013, 18: 392-403.
Boomsma D, Busjahn A, Peltonen L: Classical twin studies and beyond. Nat Rev Genet. 2002, 3: 872-882.
Condon J, Shaw JE, Luciano M, Kyvik KO, Martin NG, Duffy DL: A study of diabetes mellitus within a large sample of Australian twins. Twin Res Hum Genet. 2008, 11: 28-40.
Lehtovirta M, Pietiläinen KH, Levälahti E, Heikkilä K, Groop L, Silventoinen K, Koskenvuo M, Kaprio J: Evidence that BMI and type 2 diabetes share only a minor fraction of genetic variance: a follow-up study of 23,585 monozygotic and dizygotic twins from the Finnish Twin Cohort Study. Diabetologia. 2010, 53: 1314-1321.
Hallmayer J, Cleveland S, Torres A, Phillips J, Cohen B, Torigoe T, Miller J, Fedele A, Collins J, Smith K, Lotspeich L, Croen LA, Ozonoff S, Lajonchere C, Grether JK, Risch N: Genetic heritability and shared environmental factors among twin pairs with autism. Arch Gen Psychiatry. 2011, 68: 1095-1102.
Beckmann H, Franzek E: The genetic heterogeneity of “schizophrenia”. World J Biol Psychiatry. 2000, 1: 35-41.
Lichtenstein P, Holm NV, Verkasalo PK, Iliadou A, Kaprio J, Koskenvuo M, Pukkala E, Skytthe A, Hemminki K: Environmental and heritable factors in the causation of cancer–analyses of cohorts of twins from Sweden, Denmark, and Finland. N Engl J Med. 2000, 343: 78-85.
Jaenisch R, Bird A: Epigenetic regulation of gene expression: how the genome integrates intrinsic and environmental signals. Nat Genet. 2003, 33: 245-254.
Fraga MF, Ballestar E, Paz MF, Ropero S, Setien F, Ballestar ML, Heine-Suñer D, Cigudosa JC, Urioste M, Benitez J, Boix-Chornet M, Sanchez-Aguilera A, Ling C, Carlsson E, Poulsen P, Vaag A, Stephan Z, Spector TD, Wu Y-Z, Plass C, Esteller M: Epigenetic differences arise during the lifetime of monozygotic twins. Proc Natl Acad Sci U S A. 2005, 102: 10604-10609.
Kaminsky ZA, Tang T, Wang S-C, Ptak C, Oh GHT, Wong AHC, Feldcamp LA, Virtanen C, Halfvarson J, Tysk C, McRae AF, Visscher PM, Montgomery GW, Gottesman II, Martin NG, Petronis A: DNA methylation profiles in monozygotic and dizygotic twins. Nat Genet. 2009, 41: 240-245.
Gervin K, Hammerø M, Akselsen HE, Moe R, Nygård H, Brandt I, Gjessing HK, Harris JR, Undlien DE, Lyle R: Extensive variation and low heritability of DNA methylation identified in a twin study. Genome Res. 2011, 21: 1813-1821.
Martino D, Loke YJ, Gordon L, Ollikainen M, Cruickshank MN, Saffery R, Craig JM: Longitudinal, genome-scale analysis of DNA methylation in twins from birth to 18 months of age reveals rapid epigenetic change in early life and pair-specific effects of discordance. Genome Biol. 2013, 14: R42-
Vilkaitis G, Suetake I, Klimasauskas S, Tajima S: Processive methylation of hemimethylated CpG sites by mouse Dnmt1 DNA methyltransferase. J Biol Chem. 2005, 280: 64-72.
Swain JL, Stewart TA, Leder P: Parental legacy determines methylation and expression of an autosomal transgene: A molecular mechanism for parental imprinting. Cell. 1987, 50: 719-727.
Buiting K, Dittrich B, Gross S, Lich C, Färber C, Buchholz T, Smith E, Reis A, Bürger J, Nöthen MM, Barth-Witte U, Janssen B, Abeliovich D, Lerer I, van den Ouweland AM, Halley DJ, Schrander-Stumpel C, Smeets H, Meinecke P, Malcolm S, Gardner A, Lalande M, Nicholls RD, Friend K, Schulze A, Matthijs G, Kokkonen H, Hilbert P, Van Maldergem L, Glover G, et al: Sporadic imprinting defects in Prader-Willi syndrome and Angelman syndrome: implications for imprint-switch models, genetic counseling, and prenatal diagnosis. Am J Hum Genet. 1998, 63: 170-180.
Lee MP, DeBaun MR, Mitsuya K, Galonek HL, Brandenburg S, Oshimura M, Feinberg AP: Loss of imprinting of a paternally expressed transcript, with antisense orientation to KVLQT1, occurs frequently in Beckwith-Wiedemann syndrome and is independent of insulin-like growth factor II imprinting. Proc Natl Acad Sci U S A. 1999, 96: 5203-5208.
Weksberg R: Discordant KCNQ1OT1 imprinting in sets of monozygotic twins discordant for Beckwith-Wiedemann syndrome. Hum Mol Genet. 2002, 11: 1317-1325.
Bliek J, Alders M, Maas SM, Oostra R-J, Mackay DM, van der Lip K, Callaway JL, Brooks A, van ’t Padje S, Westerveld A, Leschot NJ, Mannens MMAM: Lessons from BWS twins: complex maternal and paternal hypomethylation and a common source of haematopoietic stem cells. Eur J Hum Genet. 2009, 17: 1625-1634.
Bestor TH: Imprinting errors and developmental asymmetry. Philos Trans R Soc Lond B Biol Sci. 2003, 358: 1411-1415.
Yamazawa K, Kagami M, Fukami M, Matsubara K, Ogata T: Monozygotic female twins discordant for Silver-Russell syndrome and hypomethylation of the H19-DMR. J Hum Genet. 2008, 53: 950-955.
Begemann M, Spengler S, Kanber D, Haake A, Baudis M, Leisten I, Binder G, Markus S, Rupprecht T, Segerer H, Fricke-Otto S, Mühlenberg R, Siebert R, Buiting K, Eggermann T: Silver-Russell patients showing a broad range of ICR1 and ICR2 hypomethylation in different tissues. Clin Genet. 2011, 80: 83-88.
Machin GA: Some causes of genotypic and phenotypic discordance in monozygotic twin pairs. Am J Med Genet. 1996, 61: 216-228.
Gordon L, Joo JE, Powell JE, Ollikainen M, Novakovic B, Li X, Andronikos R, Cruickshank MN, Conneely KN, Smith AK, Alisch RS, Morley R, Visscher PM, Craig JM, Saffery R: Neonatal DNA methylation profile in human twins is specified by a complex interplay between intrauterine environmental and genetic factors, subject to tissue-specific influence. Genome Res. 2012, 22: 1395-1406.
Breitling LP, Yang R, Korn B, Burwinkel B, Brenner H: Tobacco-smoking-related differential DNA methylation: 27 K discovery and replication. Am J Hum Genet. 2011, 88: 450-457.
Monick MM, Beach SRH, Plume J, Sears R, Gerrard M, Brody GH, Philibert RA: Coordinated changes in AHRR methylation in lymphoblasts and pulmonary macrophages from smokers. Am J Med Genet B Neuropsychiatr Genet. 2012, 159B: 141-151.
Shenker NS, Polidoro S, van Veldhoven K, Sacerdote C, Ricceri F, Birrell MA, Belvisi MG, Brown R, Vineis P, Flanagan JM: Epigenome-wide association study in the European Prospective Investigation into Cancer and Nutrition (EPIC-Turin) identifies novel genetic loci associated with smoking. Hum Mol Genet. 2013, 22: 843-851.
Zeilinger S, Kühnel B, Klopp N, Baurecht H, Kleinschmidt A, Gieger C, Weidinger S, Lattka E, Adamski J, Peters A, Strauch K, Waldenberger M, Illig T: Tobacco smoking leads to extensive genome-wide changes in DNA methylation. PLoS One. 2013, 8: e63812-
Feil R, Fraga MF: Epigenetics and the environment: emerging patterns and implications. Nat Rev Genet. 2011, 13: 97-109.
Li R, Montpetit A, Rousseau M, Wu SY, Greenwood CM, Spector TD, Pollak M, Polychronakos C, Richards JB: Somatic point mutations occurring early in development: a monozygotic twin study. J Med Genet. 2014, 51: 28-34.
Bruder CEG, Piotrowski A, Gijsbers AACJ, Andersson R, Erickson S, Diaz de Ståhl T, Menzel U, Sandgren J, von Tell D, Poplawski A, Crowley M, Crasto C, Partridge EC, Tiwari H, Allison DB, Komorowski J, van Ommen G-JB, Boomsma DI, Pedersen NL, den Dunnen JT, Wirdefeldt K, Dumanski JP: Phenotypically concordant and discordant monozygotic twins display different DNA copy-number-variation profiles. Am J Hum Genet. 2008, 82: 763-771.
Bell JT, Pai AA, Pickrell JK, Gaffney DJ, Pique-Regi R, Degner JF, Gilad Y, Pritchard JK: DNA methylation patterns associate with genetic and gene expression variation in HapMap cell lines. Genome Biol. 2011, 12: R10-
Visscher PM, Hill WG, Wray NR: Heritability in the genomics era–concepts and misconceptions. Nat Rev Genet. 2008, 9: 255-266.
Falconer DS: Introduction to Quantitative Genetics. 1960, Robert MacLehose and Company, Glasgow
Grjibovski AM, Olsen AO, Magnus P, Harris JR: Psoriasis in Norwegian twins: contribution of genetic and environmental effects. J Eur Acad Dermatol Venereol. 2007, 21: 1337-1343.
Zdravkovic S, Wienke A, Pedersen NL, de Faire U: Genetic susceptibility of myocardial infarction. Twin Res Hum Genet. 2007, 10: 848-852.
Bell JT, Tsai P-C, Yang T-P, Pidsley R, Nisbet J, Glass D, Mangino M, Zhai G, Zhang F, Valdes A, Shin S-Y, Dempster EL, Murray RM, Grundberg E, Hedman AK, Nica A, Small KS, Dermitzakis ET, McCarthy MI, Mill J, Spector TD, Deloukas P: Epigenome-wide scans identify differentially methylated regions for age and age-related phenotypes in a healthy ageing population. PLoS Genet. 2012, 8: e1002629-
Naeem H, Wong NC, Chatterton Z, Hong MKH, Pedersen JS, Corcoran NM, Hovens CM, Macintyre G: Reducing the risk of false discovery enabling identification of biologically significant genome-wide methylation status using the HumanMethylation450 array. BMC Genomics. 2014, 15: 51-
Souren NY, Lutsik P, Gasparoni G, Tierling S, Gries J, Riemenschneider M, Fryns J-P, Derom C, Zeegers MP, Walter J: Adult monozygotic twins discordant for intra-uterine growth have indistinguishable genome-wide DNA methylation profiles. Genome Biol. 2013, 14: R44-
Kim K, Ban H-J, Seo J, Lee K, Yavartanoo M, Kim SC, Park K, Cho SB, Choi JK: Genetic factors underlying discordance in chromatin accessibility between monozygotic twins. Genome Biol. 2014, 15: R72-
Milán T, Verkasalo PK, Kaprio J, Koskenvuo M: Lifestyle differences in twin pairs discordant for basal cell carcinoma of the skin. Br J Dermatol. 2003, 149: 115-123.
Williams FMK, Cherkas LF, Spector TD, MacGregor AJ: The effect of moderate alcohol consumption on bone mineral density: a study of female twins. Ann Rheum Dis. 2005, 64: 309-310.
Lee T, Lipnicki DM, Crawford JD, Henry JD, Trollor JN, Ames D, Wright MJ, Sachdev PS: Leisure activity, health, and medical correlates of neurocognitive performance among monozygotic twins: the older Australian twins study. J Gerontol B Psychol Sci Soc Sci. 2014, 69: 514-522.
Leskinen T, Usenius J-P, Alen M, Kainulainen H, Kaprio J, Kujala UM: Leisure-time physical activity and artery lumen diameters: a monozygotic co-twin control study. Scand J Med Sci Sports. 2011, 21: e208-e214.
Barker DJP: Adult consequences of fetal growth restriction. Clin Obstet Gynecol. 2006, 49: 270-283.
Dempster EL, Wong CCY, Lester KJ, Burrage J, Gregory AM, Mill J, Eley TC: Genome-wide methylomic analysis of monozygotic twins discordant for adolescent depression. Biol Psychiatry. 2014, ᅟ: ᅟ-doi:10.1016/j.biopsych.2014.04.013
Mastroeni D, McKee A, Grover A, Rogers J, Coleman PD: Epigenetic differences in cortical neurons from a pair of monozygotic twins discordant for Alzheimer’s disease. PLoS One. 2009, 4: e6617-
Nguyen A, Rauch TA, Pfeifer GP, Hu VW: Global methylation profiling of lymphoblastoid cell lines reveals epigenetic contributions to autism spectrum disorders and a novel autism candidate gene, RORA, whose protein product is reduced in autistic brain. FASEB J. 2010, 24: 3036-3051.
Wong CCY, Meaburn EL, Ronald A, Price TS, Jeffries AR, Schalkwyk LC, Plomin R, Mill J: Methylomic analysis of monozygotic twins discordant for autism spectrum disorder and related behavioural traits. Mol Psychiatry. 2014, 19: 495-503.
Dempster EL, Pidsley R, Schalkwyk LC, Owens S, Georgiades A, Kane F, Kalidindi S, Picchioni M, Kravariti E, Toulopoulou T, Murray RM, Mill J: Disease-associated epigenetic changes in monozygotic twins discordant for schizophrenia and bipolar disorder. Hum Mol Genet. 2011, 20: 4786-4796.
Kuratomi G, Iwamoto K, Bundo M, Kusumi I, Kato N, Iwata N, Ozaki N, Kato T: Aberrant DNA methylation associated with bipolar disorder identified from discordant monozygotic twins. Mol Psychiatry. 2008, 13: 429-441.
Heyn H, Carmona FJ, Gomez A, Ferreira HJ, Bell JT, Sayols S, Ward K, Stefansson OA, Moran S, Sandoval J, Eyfjord JE, Spector TD, Esteller M: DNA methylation profiling in breast cancer discordant identical twins identifies DOK7 as novel epigenetic biomarker. Carcinogenesis. 2013, 34: 102-108.
Oates NA, van Vliet J, Duffy DL, Kroes HY, Martin NG, Boomsma DI, Campbell M, Coulthard MG, Whitelaw E, Chong S: Increased DNA methylation at the AXIN1 gene in a monozygotic twin from a pair discordant for a caudal duplication anomaly. Am J Hum Genet. 2006, 79: 155-162.
Galetzka D, Hansmann T, El Hajj N, Weis E, Irmscher B, Ludwig M, Schneider-Rätzke B, Kohlschmidt N, Beyer V, Bartsch O, Zechner U, Spix C, Haaf T: Monozygotic twins discordant for constitutive BRCA1 promoter methylation, childhood cancer and secondary cancer. Epigenetics. 2012, 7: 47-54.
Jin M, Zhu S, Hu P, Liu D, Li Q, Li Z, Zhang X, Xie Y, Chen X: Genomic and epigenomic analyses of monozygotic twins discordant for congenital renal agenesis. Am J Kidney Dis. 2014, 64: 119-122.
Davies MN, Krause L, Bell JT, Gao F, Ward KJ, Wu H, Lu H, Liu Y, Tsai P-C, Collier DA, Murphy T, Dempster E, Mill J, Battle A, Mostafavi S, Zhu X, Henders A, Byrne E, Wray NR, Martin NG, Spector TD, Wang J: Hypermethylation in the ZBTB20 gene is associated with major depressive disorder. Genome Biol. 2014, 15: R56-
Baranzini SE, Mudge J, van Velkinburgh JC, Khankhanian P, Khrebtukova I, Miller NA, Zhang L, Farmer AD, Bell CJ, Kim RW, May GD, Woodward JE, Caillier SJ, McElroy JP, Gomez R, Pando MJ, Clendenen LE, Ganusova EE, Schilkey FD, Ramaraj T, Khan OA, Huntley JJ, Luo S, Kwok P-Y, Wu TD, Schroth GP, Oksenberg JR, Hauser SL, Kingsmore SF: Genome, epigenome and RNA sequences of monozygotic twins discordant for multiple sclerosis. Nature. 2010, 464: 1351-1356.
Bell JT, Loomis AK, Butcher LM, Gao F, Zhang B, Hyde CL, Sun J, Wu H, Ward K, Harris J, Scollen S, Davies MN, Schalkwyk LC, Mill J, Williams FMK, Li N, Deloukas P, Beck S, McMahon SB, Wang J, John SL, Spector TD: Differential methylation of the TRPA1 promoter in pain sensitivity. Nat Commun. 2014, 5: 2978-
Gervin K, Vigeland MD, Mattingsdal M, Hammerø M, Nygård H, Olsen AO, Brandt I, Harris JR, Undlien DE, Lyle R: DNA methylation and gene expression changes in monozygotic twins discordant for psoriasis: identification of epigenetically dysregulated genes. PLoS Genet. 2012, 8: e1002454-
Petronis A, Gottesman II, Kan P, Kennedy JL, Basile VS, Paterson AD, Popendikyte V: Monozygotic twins exhibit numerous epigenetic differences: clues to twin discordance?. Schizophr Bull. 2003, 29: 169-178.
Javierre BM, Fernandez AF, Richter J, Al-Shahrour F, Martin-Subero JI, Rodriguez-Ubreva J, Berdasco M, Fraga MF, O’Hanlon TP, Rider LG, Jacinto FV, Lopez-Longo FJ, Dopazo J, Forn M, Peinado MA, Carreño L, Sawalha AH, Harley JB, Siebert R, Esteller M, Miller FW, Ballestar E: Changes in the pattern of DNA methylation associate with twin discordance in systemic lupus erythematosus. Genome Res. 2010, 20: 170-179.
Stefan M, Zhang W, Concepcion E, Yi Z, Tomer Y: DNA methylation profiles in type 1 diabetes twins point to strong epigenetic effects on etiology. J Autoimmun. 2014, 50: 33-37.
Häsler R, Feng Z, Bäckdahl L, Spehlmann ME, Franke A, Teschendorff A, Rakyan VK, Down TA, Wilson GA, Feber A, Beck S, Schreiber S, Rosenstiel P: A functional methylome map of ulcerative colitis. Genome Res. 2012, 22: 2130-2137.
Zhu X, He F, Zeng H, Ling S, Chen A, Wang Y, Yan X, Wei W, Pang Y, Cheng H, Hua C, Zhang Y, Yang X, Lu X, Cao L, Hao L, Dong L, Zou W, Wu J, Li X, Zheng S, Yan J, Zhou J, Zhang L, Mi S, Wang X, Zhang L, Zou Y, Chen Y, Geng Z, et al: Identification of functional cooperative mutations of SETD2 in human acute leukemia. Nat Genet. 2014, 46: 287-293.
Ma Y, Dobbins SE, Sherborne AL, Chubb D, Galbiati M, Cazzaniga G, Micalizzi C, Tearle R, Lloyd AL, Hain R, Greaves M, Houlston RS: Developmental timing of mutations revealed by whole-genome sequencing of twins with acute lymphoblastic leukemia. Proc Natl Acad Sci U S A. 2013, 110: 7429-7433.
Byrne EM, Carrillo-Roa T, Henders AK, Bowdler L, McRae AF, Heath AC, Martin NG, Montgomery GW, Krause L, Wray NR: Monozygotic twins affected with major depressive disorder have greater variance in methylation than their unaffected co-twin. Transl Psychiatry. 2013, 3: e269-
Ziller MJ, Gu H, Müller F, Donaghey J, Tsai LT-Y, Kohlbacher O, De Jager PL, Rosen ED, Bennett DA, Bernstein BE, Gnirke A, Meissner A: Charting a dynamic DNA methylation landscape of the human genome. Nature. 2013, 500: 477-481.
Letourneau A, Santoni FA, Bonilla X, Sailani MR, Gonzalez D, Kind J, Chevalier C, Thurman R, Sandstrom RS, Hibaoui Y, Garieri M, Popadin K, Falconnet E, Gagnebin M, Gehrig C, Vannier A, Guipponi M, Farinelli L, Robyr D, Migliavacca E, Borel C, Deutsch S, Feki A, Stamatoyannopoulos JA, Herault Y, van Steensel B, Guigo R, Antonarakis SE: Domains of genome-wide gene expression dysregulation in Down’s syndrome. Nature. 2014, 508: 345-350.
Candelaria M, Herrera A, Labardini J, González-Fierro A, Trejo-Becerril C, Taja-Chayeb L, Pérez-Cárdenas E, de la Cruz-Hernández E, Arias-Bofill D, Vidal S, Cervera E, Dueñas-Gonzalez A: Hydralazine and magnesium valproate as epigenetic treatment for myelodysplastic syndrome. Preliminary results of a phase-II trial. Ann Hematol. 2011, 90: 379-387.
Ellis L, Atadja PW, Johnstone RW: Epigenetics in cancer: targeting chromatin modifications. Mol Cancer Ther. 2009, 8: 1409-1420.
Gaj T, Gersbach CA, Barbas CF: ZFN, TALEN, and CRISPR/Cas-based methods for genome engineering. Trends Biotechnol. 2013, 31: 397-405.
Rusk N: CRISPRs and epigenome editing. Nat Methods. 2014, 11: 28-
Rivenbark AG, Stolzenburg S, Beltran AS, Yuan X, Rots MG, Strahl BD, Blancafort P: Epigenetic reprogramming of cancer cells via targeted DNA methylation. Epigenetics. 2012, 7: 350-360.
Houseman EA, Accomando WP, Koestler DC, Christensen BC, Marsit CJ, Nelson HH, Wiencke JK, Kelsey KT: DNA methylation arrays as surrogate measures of cell mixture distribution. BMC Bioinformatics. 2012, 13: 86-
Zou J, Lippert C, Heckerman D, Aryee M, Listgarten J: Epigenome-wide association studies without the need for cell-type composition. Nat Methods. 2014, 11: 309-311.
Houseman EA, Molitor J, Marsit CJ: Reference-free cell mixture adjustments in analysis of DNA methylation data. Bioinformatics. 2014, 30: 1431-1439.
Schreiber J, Wescoe ZL, Abu-Shumays R, Vivian JT, Baatar B, Karplus K, Akeson M: Error rates for nanopore discrimination among cytosine, methylcytosine, and hydroxymethylcytosine along individual DNA strands. Proc Natl Acad Sci U S A. 2013, 110: 18910-18915.
Evans DM, Frazer IH, Martin NG: Genetic and environmental causes of variation in basal levels of blood cells. Twin Res. 1999, 2: 250-257.
This work was funded by EU-FP7 project EpitTrain (316758). TDS is an NIHR Senior Investigator and a European Research Council Senior Research Investigator (ERC 250157). The authors’ research also received support from the TwinsUK resource, funded by the Wellcome Trust, with support from the NIHR-funded BioResource, Clinical Research Facility and Biomedical Research Centre based at Guy's and St Thomas' NHS Foundation Trust in partnership with King's College London.
The authors declare that they have no competing interests.