Skip to main content

Thank you for visiting You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Molecular studies of major depressive disorder: the epigenetic perspective


Major depressive disorder (MDD) is a common and highly heterogeneous psychiatric disorder encompassing a spectrum of symptoms involving deficits to a range of cognitive, psychomotor and emotional processes. As is the norm for aetiological studies into the majority of psychiatric phenotypes, particular focus has fallen on the interplay between genetic and environmental factors. There are, however, several epidemiological, clinical and molecular peculiarities associated with MDD that are hard to explain using traditional gene- and environment-based approaches. Our goal in this study is to demonstrate the benefits of looking beyond conventional ‘DNA+environment’ and ‘DNA × environment’ aetiological paradigms. Epigenetic factors – inherited and acquired modifications of DNA and histones that regulate various genomic functions occurring without a change in nuclear DNA sequence – offer new insights about many of the non-Mendelian features of major depression, and provide a direct mechanistic route via which the environment can interact with the genome. The study of epigenetics, especially in complex diseases, is a relatively new field of research, and optimal laboratory techniques and analysis methods are still being developed. Incorporating epigenetic research into aetiological studies of MDD thus presents a number of methodological and interpretive challenges that need to be addressed. Despite these difficulties, the study of DNA methylation and histone modifications has the potential to transform our understanding about the molecular aetiology of complex diseases.


Major depressive disorder (MDD) is defined by episodes of depressed mood lasting for greater than 2 weeks accompanied by additional symptoms including disturbed sleep and appetite, reduced concentration and energy, excessive guilt, slowed movements and suicidal thoughts.1 Depression is an extremely common disorder, ranking second in the global burden of disease,2 with the overall lifetime risk of MDD estimated to be 16.2% in the general US population.3 The social and economic consequences of depression are huge, eclipsing those of many other mental and somatic illnesses, and as a result funding agencies across the world have invested huge sums of money into aetiological research.

As is the norm for aetiological studies into the majority of psychiatric phenotypes, particular focus has fallen on the interplay between genetic and environmental factors. Using these traditional research paradigms, however, progress in understanding the neurobiology of MDD has been slow. We are some distance from our ultimate goal of revealing clear risk factors that can help in the diagnosis, prevention and treatment of major depression. The first studies into the genetics of depression began over 30 years ago.4, 5, 6, 7, 8 Since then it has been repeatedly demonstrated that affective disorders run in families, with certain gene polymorphisms and environmental stressors postulated to increase susceptibility. Despite significant progress in our understanding of the probable causes of MDD, however, we are still some distance from identifying proven aetiological risks for depression and understanding the mechanisms behind their action.

Contemporary theories about the causes of depression usually separate inherited biological factors from the effect of hazardous environmental exposure. A common assumption is that ‘genes’, that is, DNA sequence variants, and ‘environments’ are the only sets of factors influencing susceptibility, and that anything not caused by the former must be owing to the latter. Only very recently have researchers started to move away from these traditional aetiological models, and look beyond the role of simple additive genetic and environmental effects. Cohort-based studies have started to address the problem of why only certain individuals – carriers of some specific genetic variants – exposed to a putative risk environment actually develop depression, finding evidence for interactions between specific environmental factors and genotype.9

There are, in fact, numerous clinical and epidemiological peculiarities associated with major depression that are hard to explain – not only in terms of traditional genetic and environmental approaches, but also by gene–environment interactions. If changes to the nuclear DNA sequence, and exposure to certain ‘risky environments’ is all that matters, why are so many identical twins raised in the same way discordant for the symptoms of depression? Why is the prevalence of MDD in women approximately double that seen in men after puberty? Why does depression follow such a striking developmental trajectory in women, with a sharp rise in prevalence following puberty? Why do some genes appear to increase the risk of developing major depression only if they are inherited from one parent but not the other?

Our goal in this study is to demonstrate the advantages of looking beyond the conventional ‘DNA+environment’ and ‘DNA × environment’ aetiological paradigms that dominate research into MDD. We briefly review the current state of traditional molecular genetic and environmental research into MDD and show that (1) it may be impossible to accurately estimate the relative contributions of either genes or the environment in MDD; (2) the boundary between genetic and environmental factors is less clear-cut than is widely believed and the environment may actually be a proxy for a more complex set of phenomenon containing a significant inherited component; (3) the inherited, biological component of MDD susceptibility may comprise of much more than simple DNA sequence variation; and (4) epigenetic factors, that is, inherited and acquired mechanisms regulating gene function occurring without a change in nuclear DNA sequence, can offer new insights about many of the non-Mendelian features of MDD including the discordance of identical twins, sex- and parental origin- effects, as well as some controversial findings in traditional genetic studies (see Table 1). It is our view that by focussing only on the sequence of the ‘genome’ and the impact of the ‘environome’, the importance of a third set of aetiological influences, namely those that act upon the ‘epigenome’, has been largely neglected.

Table 1 The epigenetic perspective on the aetiological complexities of MDD

Genes and environments: limitations to the traditional research approach

Genetics and MDD: progress, but few convincing risk loci identified

It is clear that MDD aggregates in families, and a meta-analysis of quantitative genetic studies reveals a relative risk of 2.84 for the first-degree relatives of affected sibs.10 Conventional twin analyses have concluded that much of this familial clustering is due to inherited factors. A recent meta-analysis of five twin studies, for example, concluded that the heritability of depression is 37%,10 with severe, recurrent and early-onset forms of the disorder demonstrating an elevated genetic contribution. Another way to tease out the contribution of genetic and environmental factors in the aetiology of MDD is to examine the prevalence of depression in the biological and adoptive relatives of adoptees with MDD compared to matched unaffected adoptees. Only a couple of adoption studies have been performed for depression, and while these suffer certain methodological problems,10 they provide additional evidence to indicate a strong genetic aetiology in MDD (e.g. Wender PH et al.11). Although behavioural genetic studies generally report similar heritabilities for males and females, several authors report significantly higher heritability in females (see e.g. Bierut et al.12 and Kendler et al.13), a potentially interesting observation given the significantly higher prevalence of MDD in women. Also of note is the considerable familial comorbidity observed between MDD and other affective disorders, including anxiety and bipolar disorder, suggesting that these conditions may be aetiologically related and perhaps share common inherited risk factors.14

The apparently clear contribution of inherited factors to MDD led to early optimism among the psychiatric genetics research community that loci involved in aetiology would be identified with ease. As was the case with research into other forms of psychopathology, perhaps most notably schizophrenia, considerable effort has been expended on genetic linkage and association studies with researchers around the world being swept along by the tide of enthusiasm spawned by the rapid progress of the Human Genome Project in the late 1990s. It soon became apparent that no major MDD gene was going to be identified using the classical ‘candidate gene’ or ‘whole genome linkage’ approaches that were commonly employed in these studies. As with other psychiatric conditions, the widely accepted aetiological doctrine now sees the role of genes as much more complex, involving numerous loci of small effect interacting epistatically with each other and with a range of environmental pathogens. Although some progress in identifying these risk loci has undoubtedly been made, we are still a long way from being able to definitively prove the role of specific molecular factors in MDD aetiology.

Despite the publication of numerous genetic linkage scans and a plethora of association studies, progress in identifying risk loci for MDD remains slow. Results from a number of genome-wide linkage studies for MDD have highlighted a swathe of potential susceptibility regions, although there are numerous inconsistencies between studies (for a comprehensive review see Camp and Cannon-Albright15). Perhaps the most thorough study for MDD to date investigated linkage in 110 large extended pedigrees, comprising of 1890 individuals, with a strong family history of major depression.16 Significant linkage to major depression in males was identified at marker D12S1300 (multipoint heterogeneity LOD score 4.6; P=0.00003) suggesting the existence of a sex-specific predisposition gene to major depression at 12q22–q23.2. Other studies, however, comprising of smaller sample sets, have highlighted a number of different chromosomal regions, with the overall picture looking highly complicated. In short, all that can be taken from the research effort expended on linkage studies in MDD to date are a number of large chromosomal regions that exhibit relatively low LOD scores, but show some degree of overlap between studies (e.g. 1p, 2q, 3centr, 8p, 12q, 15q and 18q). Although these regions may be candidates for future dissection by positional cloning, the interpretation of the available linkage data is difficult because of the considerable heterogeneity between studies resulting from differences in study design, sample recruitment and analytical methods employed.

Association studies of a priori candidate genes in MDD research have primarily focussed on genes involved in the serotoninergic system. The rationale behind this approach is that considerable evidence implicates dysregulation of this system as being pivotal in the development of depression. Of particular note is the effectiveness of selective serotonin reuptake inhibitors (SSRIs), which block the reuptake of serotonin at the synapse, in the treatment of depression. Studies have also demonstrated considerable impairment in serotoninergic function in a number of brain regions in individuals with major depression.17 SLC6A4, located on chromosome 17q, encodes the serotonin transporter, the primary target of SSRIs and has been implicated in the aetiology of MDD by several studies. Replicated evidence exists for an association between polymorphic variants in SLC6A4 and MDD, in particular with the short allele of a repeat polymorphism in the promoter region of the gene. A meta-analysis including data from several thousand MDD patients demonstrates that the overall effect of this polymorphism is relatively small, with an overall odds ratio in individuals homozygous for the short allele being 1.16.18 Paradoxically, this association with the short SLC6A4 allele in depression counteracts what would be predicted from functional studies of the polymorphism. The short allele is generally associated with reduced transporter activity and lower SLC6A4 expression19 – precisely the functional effect of the SSRIs used to treat depression – although a recent study investigating allelic expression in the serotonin transporter gene found no correlation between expression levels and the promoter polymorphism.20 Other genes that have been putatively associated with MDD include those encoding tryptophan hydroxylase,21 brain-derived neurotrophic factor,22, 23 catechol-O-methyl transferase,24 phospholipase A2,25 the glucocorticoid receptor26 and the serotonin receptor 1A,27 although these findings still await convincing replication in other samples (see Levinson28 a recent review).

Environmental influences: modest evidence for direct causal effects

Given that the heritability estimates for MDD are well below 100%, most quantitative geneticists have argued that, in addition to genetic factors, environmental influences are likely to be important in the aetiology of the disorder.10 Indeed, there is circumstantial evidence to link exposure to a range of specific psychosocial environmental pathogens with the development of depression. These include exposure to stressful life events, the death of a spouse or close relative, prolonged medical illness and injuries, disability and functional decline and social isolation.29 One caveat of studies into the environmental causes of MDD relates to the actual overall contribution that the environment makes. A major tenet of quantitative genetic theory argues that phenotypic variation not attributable to genetic factors must be environmental in origin. In this way, it is argued, all the phenotypic differences observed between monozygotic (MZ) twins, who have their entire genome in common, must be owing to non-shared environmental factors. Given that the proband-wise MZ concordance for MDD is only 31% for men and 48% for female MZ twins,30 this theory proposes a very large non-shared environmental contribution to the aetiology of MDD. However, a recent review of numerous behavioural studies measuring the environments of twins and non-twin siblings, and relating them to differences in their developmental outcomes, has shown that whereas over 50% of phenotypic variance is accounted for by factors attributed to be ‘non-shared environment’, actual measured environmental variables account for only a very small proportion of this variability.31 Furthermore, several other studies allude to the notion that behavioural differences observed between MZ twins may not be entirely environmental in origin. For example, normal MZ twins reared together show similar correlations for various behavioural characteristics as do MZ twins reared apart.32, 33 Adoption studies also indicate that the rearing environment may be secondary to inherited factors in mediating susceptibility to MDD.11 This is certainly the case for other psychiatric disorders, often related aetiologically to depression, in which better controlled studies have been performed – for example, it has been shown that the risk of schizophrenia does not decrease if a child born to an affected parent is raised in a healthy family.34

Another problem is that while certain environmental factors such as serious illness, divorce, violent crime and sexual assault are strongly correlated with MDD, it is not clear whether they are directly causal in their action. As a result, teasing their effect apart from that of genes may be more difficult than is often realized. It has been demonstrated that individual exposure to environmental stressors is itself strongly influenced by genetic factors, that is, predisposed individuals may choose ‘risky’ environments.35, 36 The inherited risk factors for some of the environmental factors linked to MDD, in particular stressful life events, are actually strongly correlated with those influencing susceptibility to depression itself37 indicating that the same genes are likely to be involved in both. So while exposure to stressful life events and the development of depression are almost definitely related, we can conclude relatively little about the causality behind this relationship. Another example is urbanicity, an ‘environmental’ factor commonly associated with MDD, schizophrenia and other forms of psychopathology.38 The prevalence rate of depression is clearly higher in individuals residing in urban areas with numerous studies reporting a supposedly clear cause–effect relationship between city living and depression, an observation often attributed to social isolation, increased stress, poverty and a reduced role of the family unit.39 It now appears, however, that where people live can't always be considered a true ‘environmental’ factor. Recent research suggests that amongst adults in Australia there is actually a clear inherited contribution to where people choose to live.40 It is interesting to note that this finding is not universal, and may be not applicable to some densely populated countries such as the Netherlands41 highlighting the context-dependent complexities of aetiological research.

A major dilemma facing epidemiologists investigating the impact of environmental pathogens on the development of complex disorders is that it is exceedingly difficult to initiate a well-designed, fully controlled study. The huge number of confounding variables means that the investigation of specific environmental factors, particularly in human subjects, is hard. The most powerful method with which to assess an environments' contribution to MDD aetiology would be via longitudinal randomized trial experiments in which subjects are randomly exposed to risk factors and their outcome followed up compared to controls. For obvious ethical reasons, such approaches are not viable in the investigation of variables such as severe stressful life events, and thus firm empirical evidence directly linking exposure to outcome is not available. Although animal research has given us valuable insight into the neurochemical response to environmental stressors, the degree to which these findings can be extrapolated from the laboratory rodents to real human subjects is not known.

Gene–environment interactions: causal mechanisms needed

Most behavioural geneticists would concur that the traditional idea of ‘nature’ and ‘nurture’ as distinct entities is outdated, with neither genes nor the environment likely to act in isolation to increase susceptibility. The study of gene–environment interaction effects are still in their infancy, with the first direct evidence for such interplay in the aetiology of depression being reported by Caspi et al.9 who found that a polymorphism in the serotonin transporter gene regulates the effect of stressful life events on susceptibility to depression. Citing evidence demonstrating that the serotonin transporter moderates the biological response to stressful experiences, they found that the effect of stressful life events on depression was significantly stronger in individuals carrying at least one ‘short’ allele of the 5-HTTLPR polymorphism in a large, epidemiologically ascertained birth cohort sample. Similar findings have been subsequently reported by other groups (e.g. Kendler et al.,42 Wilhelm et al.43), although both studies differ in a number of ways from the original report by Caspi et al.9

Although reports of gene–environment interactions have breathed new life into aetiological research, a number of important caveats should be considered when interpreting the results from these studies. First, given the current high level of interest in gene–environment interaction analyses and the potential for false-positive findings, credence should only be given to carefully planned hypothesis-driven research. The replication of gene–environment interaction findings is needed before firm conclusions can be drawn, but will prove extremely difficult given the unique nature of the samples used to accurately detect these effects. In this regard, it is important that studies reporting gene–environment interactions are scrutinized for proper scientific rigour, and follow a clearly defined set of strategies such as those set out by Moffitt et al.44 Second, as mentioned above, it has been shown that genes may influence the risk of developing depression by mediating an individuals' sensitivity to stressful life-events. Given the likely genetic contribution to environmental exposure, it is plausible that at least some observed gene–environment interactions are actually the result of gene–gene interactions, with the environment in question being a proxy for inherited factors. In other words, the role of the environment in gene–environment interactions may be overplayed and conclusions regarding environmental intervention strategies may be premature. Third, it is important to acknowledge that statistical evidence of an interaction between a measured gene and a measured environment does not provide any clues about the molecular mechanisms behind the interaction. Elucidating the precise mechanism(s) through which gene–environment interactions operate is vital if any diagnostic, therapeutic or preventative benefits are to result. Although progress in identifying the causes of depression has no doubt been enhanced by investigating interactions between genes and environments, it is our belief that the relatively unexplored area of epigenetics can provide new insights and new experimental strategies to uncover the molecular mechanisms behind MDD (see Figure 1).

Figure 1

MDD results from a combination of interacting genetic, environmental and epigenetic factors.

Epigenetics: from basic mechanisms to aetiology of human disease

Epigenetics refers to the heritable, but reversible, regulation of various genomic functions mediated principally through changes in DNA methylation and chromatin structure.45 Epigenetic processes are essential for normal cellular development and differentiation, and allow the regulation of gene function through non-mutagenic mechanisms. Of particular interest is the phenomenon of cytosine methylation, occurring at position 5 of the cytosine pyrimidine ring in CpG dinucleotides. This process is intrinsically linked to the regulation of gene expression, with many genes demonstrating an inverse correlation between the degree of methylation and the level of expression.46 The methylation of these CpG sites, overrepresented in CpG islands in the promoter regulatory regions of many genes, disrupts the binding of transcription factors and attracts methyl-binding proteins that are associated with gene silencing and chromatin compaction. Histone modification, the other major type of epigenetic mechanism mediating gene expression, affects chromatin structure via the processes of histone acetylation, histone methylation and histone phosphorylation.47 Interestingly, these two broad types of epigenetic mechanism are not mutually exclusive and interact in a number of ways. The methyl-binding protein MeCP2, for example, binds specifically to methylated cytosines, attracting histone deacetylases, which hypoacetylate histones.48 Transcriptionally competent chromatin is generally enriched with acetylated histones, but transcriptionally silent chromatin is normally deacetylated.49

Like the DNA sequence, the epigenetic profile of somatic cells is inherited from maternal to daughter chromatids during mitosis. Unlike the DNA sequence, which is stable and strongly conserved, epigenetic processes are highly dynamic even within an individual: they can be tissue-specific, developmentally regulated, and induced by exposure to a range of environmental factors. In addition, stochastic factors are important in determining the epigenetic milieu of the genome. Experiments tracking the inheritance of epigenetic marks through generations of genetically identical cells in tissue culture have indicated that there is considerable infidelity in the maintenance of methylation patterns in mammalian cells, and that de novo methylation events are fairly common during mitosis.50 The fidelity of the methylation pattern varies across the genome, with unmethylated regions showing a higher error rate than predominantly methylated regions, particularly in areas of the genome outside promoter regions.51 In conclusion, whereas DNA sequences exhibit nearly complete interclonal fidelity, the epigenetic profile of a cell lineage can display substantial intergenerational differences. Because processes like DNA methylation are integral in determining when and where certain genes are expressed – the precise coordination of gene expression is crucial to the correct development of any organism – this epigenetic metastability could have profound effects. While the DNA sequence of an organism dictates the physical structure of proteins, epigenetic mechanisms control the quantity, location and timing of gene expression.

It has been traditionally believed that epigenetic profiles are reset and erased during gametogenesis, thus preventing the meiotic transmission of epigenetic information between generations. Evidence is mounting, however, that the epigenetic marks of at least some mammalian genes are not fully erased during meiosis and can thus be transmitted from generation to generation.52 It appears that such meiotic transmission of epigenetic alleles, or ‘soft inheritance’, may be a fairly common phenomenon in a number of eukaryotic organisms.53 While the process of epigenetic inheritance is clearly less stable than DNA sequence inheritance, it provides another, often ignored, mechanism via which information can be passed on transgenerationally. This has obvious ramifications for traditional approaches to disease gene mapping in which all inherited traits are assumed to result from the transmission of DNA sequence changes, and may partly explain why researchers are finding it hard to pinpoint specific causal gene polymorphisms in apparently highly heritable disorders like MDD.

Epigenetic research indicates that the regulation of gene activity is critically important for normal functioning of the genome. Genes, even the ones that carry no mutations or disease predisposing polymorphisms, may be useless or even harmful if not expressed in the appropriate amount, at the right time of the cell cycle or in the right compartment of the nucleus. Cells operate normally only if both DNA sequence and epigenetic components of the genome function properly.46 In this regard, epigenetic abnormalities have been well characterized in several rare diseases such as Prader–Willi, Angelman and Beckwith–Wiedemann syndromes.54 Other syndromes postulated to have an epigenetic aetiology include: immunodeficiency, centromeric region instability, facial anomalies syndrome (ICF) which is caused by mutations in the DNA–methyltransferase 3B (DNMT3B) gene;55 Rett syndrome, a neurological disorder that occurs in females and is caused by mutations in the methylcytosine-binding protein (MeCP2) gene;56 X-linked α-thalassaemia/mental retardation (ATRX) Syndrome, caused by mutations in the X-chromosome gene ATR-X which result in the hypomethylation of repeat sequences;57 and Fragile-X syndrome, which is caused by a combination of both genetic and epigenetic mechanisms.58 It is becoming increasing clear that epigenetic processes are important in the development of cancer. In various cancers there is some degree of epigenetic misregulation, including both global genome-wide hypomethylation and the CpG island promoter hypermethylation of tumour suppressor genes.59 Complex non-malignant diseases such as MDD, however, have hardly been investigated from an epigenetic perspective. As will be discussed below, the partial stability of epigenetic signals during meiosis, and their partial instability during mitosis resulting from a host of developmental, environmental and stochastic events, makes epigenetics an attractive aetiological candidate for such diseases (see Figure 2).

Figure 2

The epigenetic model of MDD. (a) Traditional aetiological models of MDD have considered only genetic polymorphisms and environmental factors. (b) The epigenetic model of MDD postulates that epigenetic changes can result from environmental, hormonal and random stochastic factors to increase susceptibility to MDD.

Relevance of epigenetics to aetiological studies of MDD

Discordance of MZ twins and environmental impact

Twin studies have strongly implicated genetic factors in the aetiology of MDD, but as is the case for all complex psychiatric conditions, MZ twin concordances are observed to be considerably less than 100%. In the classical twin-study approach, in which MZ twins are assumed to be genetically identical, any discordance between MZ twins is attributed solely to ‘non-shared’ environmental factors. An alternative explanation, however, is that some of the observed phenotypic differences between MZ twins may be the result of epigenetic factors.60, 61 A recent study by Fraga et al.62 has demonstrated that fairly profound epigenetic differences across the genome do arise during the lifetime of MZ twins, highlighting the dynamic nature of epigenetic processes. Interestingly, MZ twin methylation differences have been reported for CpG sites in a number of specific genes that have been associated with psychiatric illness including the dopamine D2 receptor (DRD2) gene63 and the catechol-O-methyltransferase (COMT) gene.64 Similarly, genetically identical inbred animals have been shown to demonstrate considerable epigenetic differences that may be linked to gene expression differences resulting in marked phenotypic variation.65 It is becoming increasingly apparent that many of the observed epigenetic differences between MZ twins and inbred animals may be the result of random stochastic events.60 It can be envisaged that such stochastic epigenetic ‘mutations’ may accumulate over the millions of mitotic divisions occurring during the lifetime of two MZ twins, and could lead to profound gene expression alterations if present in regulatory regions of the genome. It has been proposed that such stochastic epigenetic variation may be more important in complex psychiatric disorders than is currently recognized, perhaps accounting for some of the risk currently attributed to environmental factors.66

Of course, the fact that epigenetic differences exist between MZ twins does not rule out the role of environmental factors in disease susceptibility. In fact, there is growing evidence that the environment can itself influence the epigenetic status of the genome – either globally or at specific loci.67 It is becoming increasingly apparent, for example, that a range of environmental toxins, both chemical and psychosocial, can lead to long-lasting alterations to the epigenetic profile of the genome or specific genes through processes such as the modification of DNA and histones. The fact that epigenetic marks can directly link environmental factors to gene function makes them attractive targets for a mechanistic role in the gene–environment interaction effects that are increasingly being uncovered in aetiological studies of complex diseases. Many environmentally induced epigenetic changes appear to be related to dietary intake (for a detailed review see Waterland and Jirtle68). It appears that DNA methylation can be affected by the dietary levels of methyl-donor components, for example folic acid, and that maternal dietary methyl supplements can increase DNA methylation and alter methylation-dependent phenotypes in mammalian offspring. One classic example is that of the epigenetic state of the agouti viable yellow locus in mice, which contains a gene determining coat colour and can be manipulated by altering the diet of the pregnant female.69 Certain drugs may also modify epigenetic regulation – the use of metamphetamine, for example, linked to psychiatric illness in humans, alters the DNA methylation profile of genes expressed in the brain.70 Intriguingly, and perhaps more pertinent to the aetiology of depression, it appears that the psychosocial environment may also mediate gene expression epigenetically. Research by Meaney and colleagues has shown that postnatal maternal care in rats, as measured by increased pup licking, grooming, and arch-backed nursing, leads to epigenetic modification of a NGF1-A transcription factor binding site in the promoter region of the glucocorticoid receptor gene (NR3C1).71 This finding is of particular relevance to the study of depression given the likely role of both the early rearing environment and the HPA axis in the aetiology of affective disorders.

Proband sex effects

As discussed earlier, the prevalence of MDD is approximately doubled in females compared to males, and several studies suggest that the heritability of the disorder is significantly higher in women.12, 72 Furthermore, it appears that the genetic aetiology of MDD may differ between the sexes with the overlap of putative genetic factors involved in the disorder being incomplete.72 Following the investigation of X-linked Mendelian diseases that exhibit sexual dimorphism, one potential explanation is that the overrepresentation of depression in women results from a MDD risk gene on the X chromosome. Several segregation analyses have suggested involvement of X-chromosome genes in MDD (see e.g. Vaillant et al.73) and certain X-linked genes, for example, GPR50, have been associated with the disorder,74 although such conclusions are not ubiquitous (see e.g. Faraone et al.75) and no firm conclusions can yet been drawn about the involvement of X-linked sequence variants in depression.

Skewed X-chromosome inactivation is one X-linked epigenetic process that could potentially cause the excess rates of MDD in women, and also explain female MZ twin discordance.76 X-inactivation silences gene expression and is initially instigated by expression of the XIST gene and then maintained via DNA methylation and histone modifications. The role of X inactivation is to compensate for the greater dosage of X-linked genes in women (who have two X chromosomes) compared to men (who have one X chromosome). It is generally assumed that in most women X inactivation is stochastic for each cell lineage – either the maternally or paternally inherited X can be silenced – but is maintained throughout subsequent cell divisions. There is, however, increasing evidence that in some cases the selection process is not random. It has been demonstrated, for example, that approximately half the female carriers of X-linked mental retardation exhibit skewed X-inactivation where the activation ratio between the two X chromosomes is 80:20% or higher.77 Non-random X-chromosome inactivation has been observed in a number of disorders, including several X-linked immunodeficiencies, Lesch–Nyhan disease, incontinentia pigmenti, focal dermal hypoplasia and adrenoleukodystrophy.78 In the normal population of women without a family history of X-linked disorders, 5–20% of women have constitutional skewing of X inactivation.79 According to other authors, 30–40% of females exhibit ratios of 60:40% or more, and 10% of normal females demonstrate even more extreme ratios.78 In studies of MDD, therefore, we propose that the ‘skewedness’ of the X chromosome should be thoroughly investigated, and if present taken into account when performing genetic studies because the results of linkage and association studies may be mediated by the proportion of females with skewed X-inactivation. Skewed X-inactivation also provides a mechanism for phenotypic discordance between female MZ twins. Male MZ twins will only express X-linked genes inherited from their mothers, but females can express either maternally or paternally inherited genes. Loat and colleagues.80 have argued that given the random nature of X inactivation, female MZ twin pairs do not necessarily inactivate the same X chromosome and will thus show higher levels of discordance than male MZ twin pairs. To our knowledge, no study has yet tested this theory on individuals with MDD, but analyses on MZ twins assessed at ages 2, 3 and 4 suggests that this pattern is observed for a number of behavioural traits closely related to depression.80 Another epigenetic phenomenon that may confound X-linked gene studies in MDD is the existence of genes escaping X inactivation. Not all genes on the X chromosome are inactivated, with as many as 15% showing evidence of expression from both X chromosomes.76, 81 Genes escaping X inactivation are obvious candidates for explaining sexual dimorphism in disease prevalence, especially if no Y-chromosome homologue for those genes exists.

Sex effects have also been recently detected in genetic linkage and association studies of autosomal genes. For example, genome-wide linkage scans of MDD families provide evidence for male-only linkage on chromosome 12q22–q23.216 and evidence for female-only linkage on chromosome 2q33–35.82 There are numerous other examples across the spectrum of psychiatric diseases where autosomal loci exhibit apparently clear-cut sex differences. Sex hormones are the usual ‘culprit’ used to explain such gender effects in complex disease, based in part on the myriad of data associating hormonal differences with disease states, and their critical involvement in numerous regulatory processes. Little is known, however, about the underlying mechanisms of how sex hormones predispose to or protect from a disease. The gender-specific effects observed in genetic linkage and association studies suggest that chromosomal regions and individual genes may be the target of sex hormone action, but the molecular mechanisms behind these interactions are yet to be elucidated. Although sex hormones cannot change DNA sequence, it is known they can be potent modifiers of epigenetic status and gene expression. There are several reports of the female sex hormone oestrogen altering the chromatin configuration of certain genes.83, 84 Furthermore, it has been proposed that sex hormones may act to alter the DNA methylation profile of specific loci in the genome,85, 86 controlling gene expression in a sex-specific manner. It is thus plausible that susceptibly to depression is mediated by the hormone-specific epigenetic modification of genes contained within nominated linkage regions. This may explain some of the sex effects observed widely in genetic linkage and association studies: a specific allele or haplotype may be a risk factor only after epigenetic modification by some aspect of the endocrinological millieu.

Parental origin effects

Another non-Mendelian feature that is commonly observed in studies of psychopathology is a parent-of-origin effect, where disease susceptibility is mediated by parental factors in a sex-specific manner. Parent-of-origin effects are apparent at both a phenotypic level, where the risk to offspring depends on the sex of the affected parent, and also at a molecular level when risk alleles only confer increased susceptibility if they are transmitted from the mother or the father. These effects have been observed in a number of psychiatric phenotypes, including several closely related to major depression. For example, it has been shown that the risk of developing bipolar disorder is higher in offspring with affected mothers than in those with affected fathers.87 Genome-wide linkage-scans and candidate–gene association studies have also provided some preliminary evidence for parent-of-origin effects in depression. Zill et al.88 for example, find evidence for a parent-of-origin effect with the GOLF gene, and Schiffer and Heinemann89 report preferential maternal transmission of a GluR7 gene risk allele to MDD patients. Many other examples, however, have probably gone unnoticed because parent-of-origin effects are not routinely tested in traditional genetic studies. In fact, it is likely that parental origin effects are a much wider phenomenon in psychiatric diseases than it is generally thought given that linkage analyses, as a rule, are performed in a sex-averaged way.

The most likely mechanism behind such parent-of-origin effects is genomic imprinting – the differential expression of the genetic material at either a chromosomal or allelic level depending on whether the genetic material has been transmitted from the paternal or the maternal side. Genomic imprinting involves the epigenetic marking of chromatin, via the modifications of DNA and histones, which results in expression of only one allele at that locus. Genomic imprinting is fundamental to normal mammalian development and growth, and plays an important role in brain function and behaviour.90 Genomic imprinting is traditionally viewed as an all-or-nothing phenomenon that occurs in only small, clustered regions of the genome. In these imprinted regions only maternally or paternally inherited genes are turned on, but not both, resulting in monoallelic expression, for example, at both H19 and IGF2. It is estimated that there are between 100 and 200 imprinted loci in the human genome (see Recent evidence, however, suggests that imprinting may be more widespread and less clear-cut than originally believed, and this could have important ramifications for our understanding of the aetiology of complex disorders such as depression. Preis et al.91 used nine transgenic mouse lines to uncover a much more subtle form of imprinting in the murine genome. They found that a high proportion of transgenic lines displayed lower expression of a reporter transgene when it was passed through the maternal line although paternal alleles were expressed as well. It is proposed that these observations may be representative of imprinting patterns across all mammalian genomes, with perhaps the majority of genes demonstrating some degree of skewed allelic expression in a parent-of-origin dependent manner. Conversely, it has also been shown that genes previously believed to exhibit total monoallelic expression as a result of imprinting may actually show considerable heterogeneity. A study examining the allelic expression of three imprinted genes (SNRPN, IMPT1, IGF2) in the normal population found evidence for significant biallelic expression indicating that functional variation at imprinted loci may be a common feature of the mammalian genome.92

Association of DNA sequence variation with epigenetic variation – epi-alleles, epi-genotypes and epi-haplotypes

As discussed above, few convincing associations between specific polymorphisms and susceptibility to MDD have been revealed by candidate gene studies. One major problem is the heterogeneity apparent in the data generated from these analyses. The pattern of association observed is rarely consistent between studies, and it is not uncommon for different alleles and haplotypes to be associated with the disorder in different study populations. One possible explanation for this heterogeneity comes from the observation that genetic–epigenetic interactions may be commonplace across the genome. ‘Epi-alleles’ and ‘epi-haplotypes’ that combine both DNA sequence and epigenetic information may thus be better predictors of the risk for disorders like depression than any of the two components analysed separately. There is increasing evidence that some DNA alleles and haplotypes tend to be associated with a specific epigenetic profile. Of particular interest in this regard is increasing evidence that DNA methylation profiles may be associated with specific DNA alleles in genes putatively linked to mood disorders. For example, the T allele of the C677T polymorphism in the methylenetetrahydrofolate reductase gene (MTHFR), recently implicated in depression,93 is associated with an increased risk of imprinting defects in the Prader–Willi syndrome/Angelman syndrome region of 15q.94 Furthermore, the C102T polymorphism in the serotonin 5-HT2A receptor gene (5HT2AR), which has been associated with several psychiatric disorders including depression, was found to be methylated specifically on the C allele.95 Dempster et al.96 found evidence that methylation in the promoter region of the COMT gene, another locus implicated in numerous psychiatric phenotypes, was associated in a dose-dependent manner with genotype at the val158met polymorphism, with val158 homozygotes exhibiting lower levels of methylation than met158 homozygotes. Finally, a recent epigenetic study of a C/G single nucleotide polymorphism (SNP) in the CDH13 gene in male germline cells revealed that C alleles are predominantly unmethylated, while G alleles are predominantly methylated.97 The notion that epigenetic changes may be associated with specific DNA sequence changes sheds new light on the inconsistent genetic association studies observed in complex diseases, and suggests that a comprehensive epigenetic analysis of candidate SNPs and haplotypes is warranted.

Epigenetic studies in MDD: technological and methodological complexities

While epigenetics provides a new perspective on the aetiology of major depression, it would be naïve to expect that applying epigenetic theory to molecular-based studies of MDD psychopathology is a straightforward task that can be achieved without encountering a number of technological and methodological complexities. The study of epigenetics, especially in complex diseases, is a relatively new field of research, and optimal laboratory techniques and analysis methods are still being developed. It is vital that laboratory approaches are refined and verified before sweeping conclusions are drawn from experimental data. This is particularly important because epigenetic experiments need to overcome a number of inherent complexities that are not encountered when investigating simple DNA sequence variation. Potential problems surrounding the epigenetic studies can be subdivided into several groups, which will be discussed briefly in the sections below: (1) the technological limitations in identifying epigenetic variation; (2) sample- and target tissue-related issues; (3) the confounding effects of DNA sequence variation; and (4) the interpretation of a cause-effect relationship between epigenetic changes and disease aetiology.

Technological limitations

Whilst it is easy to theorize about the role of epigenetic processes in mediating susceptibility to psychiatric disorders, actually investigating these modifications at a molecular level is not so straightforward. To date, only a few empirical studies have been performed, and it is clear that a number of methodological difficulties need to be overcome before large-scale epigenetic studies can be viably executed.

A number of approaches to investigate DNA methylation have been developed (see Figure 3). Currently, the ‘gold standard technique’ for fine mapping of methylated cytosines is based on the treatment of genomic DNA with sodium bisulfite, which converts unmethylated cytosines into uracils (and subsequently, via polymerase chain reaction (PCR), to thymidines), whereas methylated cytosines are resistant to bisulfite and remain unchanged.98 After sodium bisulfite treatment, one crude method to examine DNA methylation is via methylation-specific PCR (MSP). In this approach two PCR reactions are performed on bisulfite-treated DNA with primer sets specific to (i) methylated DNA and (ii) unmethylated DNA. Such an approach can only investigate methylation at a very small number of CpG sites and gives no quantitative information about the degree of methylation. Furthermore, data from experiments utilizing MSP should be treated with caution given the numerous confounding factors that can affect PCR efficiency on bisulfite-treated DNA. A more rigorous method to investigate DNA methylation is bisulfite genomic sequencing. DNA regions of interest are amplified and sequenced to identify C>T transitions or stable C positions, respectively corresponding to unmethylated and methylated cytosines in the native DNA. Typically, PCR amplicons are either sequenced directly to provide a strand-specific average sequence for the population of DNA molecules, or cloned and sequenced to provide methylation maps of single DNA molecules.98, 99 An alternative to the laborious and expensive sequencing of cloned bisulfite-PCR products is the use of pyrosequencing to accurately quantify the degree of CpG methylation. Pyrosequencing is a sequencing-by-synthesis method that relies on the luminometric detection of pyrophosphate release upon nucleotide incorporation via an enzymatic cascade. Because there is a direct correlation between sequence data and the amount of nucleotides incorporated during the reaction, it provides an accurate method for detecting the precise methylation level at any CpG site giving truly quantitative data that are not achievable using standard bisulfite-sequencing protocols.100 A major problem with the sodium bisulfite approach is that it results in considerable DNA degradation, and large quantities of genomic DNA material are needed if numerous genomic regions are to be profiled – obviously an issue if valuable, relatively inaccessible tissues such as the brain are to be profiled. One potential method to overcome problems associated with tissue availability is the use of whole-genome amplification of bisulfite-treated DNA.101 In addition, no consensus has yet been reached on how to best statistically analyse bisulfite-sequence data. Such analyses are unlikely to be straightforward, especially if the differences observed across individuals are subtle as would be expected for disorders such as MDD. One possibility is to use an approach based on epigenetic distance, which reduces bisulfite-sequencing data from cloned PCR products to a binary code for each DNA molecule, and allows patterns of ‘epigenetic drift’ to be easily ascertained.102

Figure 3

Experimental approaches to investigate DNA methylation. (a) Microarray-based methylation profiling using microarrays can investigate DNA methylation changes at a genome-wide level. In this example the unmethylated fraction of genomic DNA is enriched using methylation-sensitive restriction enzymes and adaptor-ligation PCR and subsequently hybridized on microarrays (see Schumacher et al.103); (b) locus-specific methylation analysis using sodium bisulfite treatment and PCR-based sequence analysis.

These traditional, locus-specific techniques limit the number of the CpG sites that can be feasibly assessed, and thus exclude the possibility of performing a genome-wide ‘epigenome-scan’ to identify novel regions of importance. In recent years, a number of novel, high-throughput microarray-based methods have been developed, which should enable future epigenetic studies to overcome many of these issues.103, 104 At present these methods have not been widely used and are relatively untested. Two general methods exist for treating and enriching DNA samples before array hybridization: the digestion of genomic DNA with restriction-sensitive enzymes, and the use of antibodies to specifically pull-down methylated cytosines. The enzyme approach cannot be used to interrogate every potential methylated cytosine in the genome, whereas the antibody approach can only examine the methylated fraction of the genome, which is far less informative than the unmethylated fraction.103 As we have learnt from the use of gene expression microarrays in psychiatry, differences in sample preparation, microarray type, sample labeling and data analysis procedures pose a tremendous challenge to the field,105 and some degree of standardization is needed across epigenetic studies. A major hurdle to the widespread use of methylation microarrays is cost – for example it is currently prohibitively expensive to scan the entire genome using DNA from a large sample of MDD patients and controls, especially if tissue-specific effects are important and numerous brain regions need to be assessed. Finally, until recently, the resolution of the available low-density arrays has been relatively crude, and the available methodology has not been particularly sensitive to subtle methylation differences. Current genome-wide methylation studies have been biased by the fact that current microarray designs focus primarily on cDNA, promoters, and 5′ regulatory sequences of known genes only. As array technology advances, however, with the development of commercially produced whole-genome ‘tiling’ arrays containing millions of features allowing an unbiased assessment of the entire genome, microarray studies have the potential to rapidly advance our understanding of the epigenetic mileau of the genome, and provide novel insights into the aetiology of MDD.

This article has been primarily dedicated to the role of DNA methylation in MDD aetiology. This focus partly reflects the fact that methods to examine other epigenetic processes such as histone modifications in psychiatric disorders are a step behind those currently available for DNA methylation analysis. The method used to examine the epigenetic state of histones is chromatin immunoprecipitation (ChIP). One limitation to this approach is the large number of cells needed for each experiment, although recent methodological advances may permit ChIP analysis on as few as 100 cells.106 Another major problem with ChIP analysis is that, compared to DNA methylation marks, histone modifications are far less stable in post-mortem tissues. In this regard, specific immunoprecipitation protocols applicable to post-mortem brain tissue need to be developed, and the impact of potential confounds such as autolysis time or tissue pH need to be thoroughly examined. Some promising data come from recent investigations that suggest the immunoprecipitation of micrococcal nuclease-digested tissue extracts is a feasible approach to profile histone methylation at defined genomic loci in post-mortem brain.107 Once histone-stability problems have been addressed, the development of powerful microarray-based ‘ChIP-on-chip’ methods, in which DNA isolated from ChIPs is hybridized to a microarry chip, should allow genome-wide analysis of histone modifications and provide additional epigenetic information to complement data from genome-wide DNA methylation studies.

Sample and target tissue

To identify epigenetic changes associated with depression, the ideal experiment would investigate prospectively the dynamics of genome-wide epigenetic changes in the brains of individuals who eventually become affected with MDD compared to unaffected control individuals. At the present time, however, limitations with our current range of epigenetic methodologies render such an approach impossible. As it is not yet possible to perform in vivo epigenetic studies, particularly for tissues like the brain, only retrospective study designs using post-mortem brain samples are viable.

Given the role of epigenetics in mediating when and where genes are expressed, an obvious confounding factor in epigenetic studies is tissue heterogeneity. Because different cell types exhibit quite different epigenetic profiles across different genomic regions, specific tissues – that is, the primary sites of disease manifestation – are deemed preferable for aetiological studies. It is likely that for studies of psychiatric disorders such as MDD, the primary focus of aetiological studies will be the numerous brain regions implicated in the disorder. Even within specific-tissue types, however, there is considerable cellular heterogeneity. Tissues like the cortex, for example, consist of numerous different types of neuronal and glial cells, and the detection of cell-type specific epimutations is likely to be difficult unless the epigenetic profile of specific cells can be investigated. This point is highlighted by a recent study identifying epigenetic changes occurring specifically in cortical interneuron cells, suggesting that different cell populations within a single tissue-type can have quite distinct epigenetic profiles.108 An increasingly popular solution in the epigenetics research field is to apply ‘Laser Capture Microdissection’ technology, which enables the isolation of specific cell types or single cells from whole tissue avoiding the confounding effects of cell type variation.109

Despite the problems of tissue specificity, because the availability of fresh brain tissue from well-phenotyped individuals with disorders like MDD is non-existent, one compromise is to use other, more accessible tissue sources like leukocytes for epigenetic profiling in the hope that the patterns observed will reflect those of the real tissue of interest. There is increasing evidence that many epimutations are not limited to the affected tissue or cell type, but can also be detected in other tissues. Good examples are the epimutations at IGF2110 in lymphocytes and MLH1 in sperm cells,111 observed in colon cancer patients and epimutations at KCNQ1OT1 in the lymphocytes and skin fibroblasts in Beckwith–Wiedemann syndrome.112 In fact, peripheral blood-based studies may be useful in revealing epigenetic changes resulting from early embryogenesis, or even highlighting inherited epigenetic variation. The study of peripheral cells may even have some advantages over the use of brain tissue in that they are likely to accumulate fewer epigenetic changes induced by disease-related external factors, such as medications, recreational drugs and stress, which could all be confounding factors to epigenetic studies. The use of peripheral cells would also allow the longitudinal study of epigenetic changes throughout the lifecourse of individuals diagnosed with MDD, providing information about the epigenetic factors associated with the development, remission and relapse of the disorder. Finally, because there have been no thorough epigenetic studies in MDD to date, it is not known how large a sample needs to be investigated. At this stage the degree of any potential epigenetic differences between cases and controls is unknown, although the effects observed are likely to be small.

Defining a ‘normal’ epigenetic profile and proving a cause–effect relationship between epigenetic differences and MDD

Despite the current interest in the possible role of epigenetic processes in disorders like MDD, very little empirical research has been performed to characterize what actually comprises a ‘normal’ epigenome. Before we can systematically investigate the epigenetic changes associated with disease, we must first catalogue the genome-wide DNA methylation patterns in all major tissues. This is one of the targets of the Human Epigenome Project (, which aims to map all methylation variable positions in the genome.113 Producing a reference map of the entire human epigenome is going to be a gargantuan task, especially given the huge epigenetic differences that are likely to occur both between tissue/cell types, and within a specific tissue/cell type over the course of development. To date, this private/public consortium has released DNA methylation profiles of the major histocompatibility complex114 and chromosomes 6, 20 and 22.115

As is the case for all non-DNA sequence based (neurochemical, pathophysiological and environmental) studies, proving a direct causal link between epigenetic factors and disease is not straightforward. For example, the observation of a different epigenetic profile between depressed and non-depressed individuals does not necessarily indicate a causal relationship between these epigenetic changes and depression. It is possible that any epigenetic changes in depressed individuals are not a cause of, but actually the result of MDD-induced changes in the brain. Given that several exogenous chemicals have been linked with processes like DNA methylation, it is also possible, for example, that the medications used to treat depression may cause epigenetic changes. In this regard, tissues that are not the disease site or directly affected by antidepressant medications, for example peripheral blood cells, could be very useful in elucidating inherited epigenetic changes associated with MDD. The ideal study design to identify causal relationships between epigenetic factors and MDD would be a prospective longitudinal study in which individuals at risk for depression are investigated before they develop the disease.

Confounding effects of DNA variation

A major issue in epigenetic studies is that putative epigenetic differences can be masked by the thousands of DNA sequence variants that exist between individuals. Owing to the nature of the technology currently employed to examine epigenetic changes, DNA sequence polymorphisms are not always easy to differentiate from epigenetic differences. This problem applies to all DNA methylation technologies currently utilized in aetiological studies. Cut sites for methylation-sensitive restriction enzymes, for example, can be created or removed by sequence changes. Sodium-bisulfite conversion reduces methylation status to standard sequence information, and can thus be confounded by SNPs. Two possible study designs can be used to address these problems. First, methylation differences in MZ twins discordant for MDD can be examined. MZ twins have identical genomes, so phenotypic differences cannot be accounted for by DNA sequence changes. However, a potential limitation of using MZ twins is that unaffected co-twins may also be carriers of epigenetic changes that may not be too far from a hypothetical ‘threshold’ of epigenetic misregulation, above which the twin would present with a clinical diagnosis of MDD. Typical microarray experiments are based on relative differences between two individuals, rather than the absolute size of a putative epigenetic change. It is possible that this difference, between an affected twin (a little bit above the ‘threshold’) and an unaffected co-twin (a little bit below the ‘threshold’) may be relatively small and difficult to differentiate from experimental errors. A second approach is to instigate a longitudinal study of epigenetic changes during the course of disease in MDD patients, using DNA taken from easily accessible tissue (e.g. peripheral blood cells). In such an experiment, the epigenetic profile of DNA samples taken at various time-points could be compared (e.g. at first diagnosis, remission and relapse). A similar approach has been suggested as a substitute to twin studies in pharmacogenetic investigations.116 Such a design is powerful as it can inform us about the extent that epigenetic status fluctuates in peripheral blood cells, and may provide an insight into relapse-associated epigenetic changes.

Epigenetics and MDD: vogue, hassle or necessity?

So what is the role of epigenetics in aetiological studies of MDD: vogue, hassle or necessity? The answer is probably a bit of all three. Largely neglected until very recently, the study of epigenetics has become increasingly ‘mainstream’. There is a certain buzz amongst the molecular psychiatry research community about the potential role of DNA methylation and histone modifications in a range of psychiatric phenotypes. It is now generally accepted that these epigenetic factors may sit alongside DNA sequence polymorphisms and environmental factors in mediating susceptibility to disorders like MDD. In this article we have discussed the benefits of including epigenetic theory into models of disease aetiology, highlighting how epigenetic mechanisms can account for the many features of MDD that are unexplainable using traditional genetic and environmental approaches (see Table 1).

Theoretical notions about a role for epigenetic processes in the development of MDD do not prove such a link, however. There is a need to move from theory into empirical laboratory-based research, and this is where the current ‘vogue’ for epigenetics becomes more of a ‘hassle’. The study of epigenetics, especially in complex diseases, is a relatively new field of research, and optimal laboratory techniques and analysis methods are still being developed. We should be under no illusion that such experiments are going to be easy. If we are to really fulfill the novel heuristic opportunities that epigenetics provides, the development of novel high-resolution genome-wide methods to scan the epigenome for DNA methylation and histone modifications are vital.

Despite these difficulties, it is also clear that studies investigating epigenetic factors in MDD (as well as various other complex diseases) are a ‘necessity’. Years of research effort have been spent attempting to identify the specific genes and environmental factors that were traditionally believed to determine individual susceptibility to MDD. While some progress has been achieved, our focus on only genes, environments and the interactions between them, has left many questions unanswered. Although it would be premature to conclude that epigenetics will lead to revolutionary discoveries in non-Mendelian biology, the study of DNA methylation and histone modifications has the potential to transform our understanding about the molecular aetiology of complex diseases.


  1. 1

    American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders, 4th edn. American Psychiatric Association: Washington, DC, 1994.

  2. 2

    Murray CJ, Lopez AD . Global mortality, disability, and the contribution of risk factors: global burden of disease study. Lancet 1997; 349: 1436–1442.

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  3. 3

    Kessler RC, Berglund P, Demler O, Jin R, Koretz D, Merikangas KR et al. The epidemiology of major depressive disorder: results from the national comorbidity survey replication (NCS-R). JAMA 2003; 289: 3095–3105.

    PubMed  PubMed Central  Article  Google Scholar 

  4. 4

    Winokur G, Pitts Jr FN . Affective disorder: VI. A family history study of prevalences, sex differences and possible genetic factors. J Psychiatr Res 1965; 3: 113–123.

    CAS  PubMed  Article  PubMed Central  Google Scholar 

  5. 5

    Stenstedt A . Genetics of neurotic depression. Acta Psychiatr Scand 1966; 42: 392–409.

    CAS  PubMed  Article  PubMed Central  Google Scholar 

  6. 6

    Reich T, Clayton PJ, Winokur G . Family history studies: V. The genetics of mania. Am J Psychiatry 1969; 125: 1358–1369.

    CAS  PubMed  Article  PubMed Central  Google Scholar 

  7. 7

    Gershon ES, Dunner DL, Goodwin FK . Toward a biology of affective disorders. Genetic Contributions. Arch Gen Psychiatry 1971; 25: 1–15.

    CAS  PubMed  Article  PubMed Central  Google Scholar 

  8. 8

    Murphy DL, Wyatt RJ . Neurotransmitter-related enzymes in the major psychiatric disorders: I. Catechol-O-Methyl transferase, monoamine oxidase in the affective disorders, and factors affecting some behaviorally correlated enzyme activities. Res Publ Assoc Res Nerv Ment Dis 1975; 54: 277–288.

    CAS  PubMed  PubMed Central  Google Scholar 

  9. 9

    Caspi A, Sugden K, Moffitt TE, Taylor A, Craig IW, Harrington H et al. Influence of life stress on depression: moderation by a polymorphism in the 5-HTT gene. Science 2003; 301: 386–389.

    CAS  Article  Google Scholar 

  10. 10

    Sullivan PF, Neale MC, Kendler KS . Genetic epidemiology of major depression: review and meta-analysis. Am J Psychiatry 2000; 157: 1552–1562.

    CAS  Article  Google Scholar 

  11. 11

    Wender PH, Kety SS, Rosenthal D, Schulsinger F, Ortmann J, Lunde I . Psychiatric disorders in the biological and adoptive families of adopted individuals with affective disorders. Arch Gen Psychiatry 1986; 43: 923–929.

    CAS  PubMed  Article  Google Scholar 

  12. 12

    Bierut LJ, Heath AC, Bucholz KK, Dinwiddie SH, Madden PA, Statham DJ et al. Major depressive disorder in a community-based twin sample: are there different genetic and environmental contributions for men and women? Arch Gen Psychiatry 1999; 56: 557–563.

    CAS  PubMed  Article  Google Scholar 

  13. 13

    Kendler KS, Gardner CO, Neale MC, Prescott CA . Genetic risk factors for major depression in men and women: similar or different heritabilities and same or partly distinct genes? Psychol Med 2001; 31: 605–616.

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  14. 14

    Rush AJ, Zimmerman M, Wisniewski SR, Fava M, Hollon SD, Warden D et al. Comorbid psychiatric disorders in depressed outpatients: demographic and clinical features. J Affect Disord 2005; 87: 43–55.

    PubMed  PubMed Central  Article  Google Scholar 

  15. 15

    Camp NJ, Cannon-Albright LA . Dissecting the genetic aetiology of major depressive disorder using linkage analysis. Trends Mol Med 2005; 11: 138–144.

    CAS  PubMed  Article  Google Scholar 

  16. 16

    Abkevich V, Camp NJ, Hensel CH, Neff CD, Russell DL, Hughes DC et al. Predisposition locus for major depression at chromosome 12q22–12q23.2. Am J Hum Genet 2003; 73: 1271–1281.

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  17. 17

    Arango V, Underwood MD, Mann JJ . Serotonin brain circuits involved in major depression and suicide. Prog Brain Res 2002; 136: 443–453.

    CAS  Article  PubMed  Google Scholar 

  18. 18

    Lotrich FE, Pollock BG . Meta-analysis of serotonin transporter polymorphisms and affective disorders. Psychiatr Genet 2004; 14: 121–129.

    PubMed  PubMed Central  Article  Google Scholar 

  19. 19

    Lesch KP, Bengel D, Heils A, Sabol SZ, Greenberg BD, Petri S et al. Association of anxiety-related traits with a polymorphism in the serotonin transporter gene regulatory region. Science 1996; 274: 1527–1531.

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  20. 20

    Lim JE, Papp A, Pinsonneault J, Sadee W, Saffen D . Allelic expression of serotonin transporter (SERT) MRNA in human pons: lack of correlation with the polymorphism SERTLPR. Mol Psychiatry 2006; 11: 649–662.

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  21. 21

    Gizatullin R, Zaboli G, Jonsson EG, Asberg M, Leopardi R . Haplotype analysis reveals tryptophan hydroxylase (TPH) 1 gene variants associated with major depression. Biol Psychiatry 2006; 59: 295–300.

    CAS  PubMed  Article  Google Scholar 

  22. 22

    Schumacher J, Jamra RA, Becker T, Ohlraun S, Klopp N, Binder EB et al. Evidence for a relationship between genetic variants at the brain-derived neurotrophic factor (bdnf) locus and major depression. Biol Psychiatry 2005; 58: 307–314.

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  23. 23

    Strauss J, Barr CL, George CJ, Devlin B, Vetro A, Kiss E et al. Brain-derived neurotrophic factor variants are associated with childhood-onset mood disorder: confirmation in a hungarian sample. Mol Psychiatry 2005; 10: 861–867.

    CAS  Article  PubMed  Google Scholar 

  24. 24

    Massat I, Souery D, Del Favero J, Nothen M, Blackwood D, Muir W et al. Association between COMT (Val158Met) functional polymorphism and early onset in patients with major depressive disorder in a European multicenter genetic association study. Mol Psychiatry 2005; 10: 598–605.

    CAS  Article  PubMed  Google Scholar 

  25. 25

    Papadimitriou GN, Dikeos DG, Souery D, Del Favero J, Massat I, Avramopoulos D et al. Genetic association between the phospholipase A2 gene and unipolar affective disorder: a multicentre case–control study. Psychiatr Genet 2003; 13: 211–220.

    PubMed  Article  Google Scholar 

  26. 26

    van Rossum EF, Binder EB, Majer M, Koper JW, Ising M, Modell S et al. Polymorphisms of the glucocorticoid receptor gene and major depression. Biol Psychiatry 2006; 59: 681–688.

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  27. 27

    Arias B, Catalan R, Gasto C, Gutierrez B, Fananas L . Evidence for a combined genetic effect of the 5-HT(1A) receptor and serotonin transporter genes in the clinical outcome of major depressive patients treated with citalopram. J Psychopharmacol 2005; 19: 166–172.

    CAS  Article  Google Scholar 

  28. 28

    Levinson DF . The genetics of depression: a review. Biol Psychiatry 2006; 60: 84–92.

    CAS  Article  Google Scholar 

  29. 29

    Bruce ML . Psychosocial risk factors for depressive disorders in late life. Biol Psychiatry 2002; 52: 175–184.

    PubMed  Article  Google Scholar 

  30. 30

    Kendler KS, Prescott CA . A population-based twin study of lifetime major depression in men and women. Arch Gen Psychiatry 1999; 56: 39–44.

    CAS  PubMed  Article  Google Scholar 

  31. 31

    Turkheimer E, Waldron M . Nonshared environment: a theoretical, methodological, and quantitative review. Psychol Bull 2000; 126: 78–108.

    CAS  PubMed  Article  Google Scholar 

  32. 32

    Bouchard Jr TJ, Lykken DT, McGue M, Segal NL, Tellegen A . Sources of human psychological differences: the minnesota study of twins reared apart. Science 1990; 250: 223–228.

    Article  Google Scholar 

  33. 33

    Bouchard Jr TJ, McGue M . Genetic and environmental influences on human psychological differences. J Neurobiol 2003; 54: 4–45.

    PubMed  PubMed Central  Article  Google Scholar 

  34. 34

    Gottesman II . Schizophrenia Genesis. Freeman: New York, 1991.

    Google Scholar 

  35. 35

    Plomin R, Lichtenstein P, Pedersen NL, McClearn GE, Nesselroade JR . Genetic influence on life events during the last half of the life span. Psychol Aging 1990; 5: 25–30.

    CAS  PubMed  Article  Google Scholar 

  36. 36

    Kendler KS, Neale M, Kessler R, Heath A, Eaves L . A twin study of recent life events and difficulties. Arch Gen Psychiatry 1993; 50: 789–796.

    CAS  PubMed  Article  Google Scholar 

  37. 37

    Kendler KS, Karkowski-Shuman L . Stressful life events and genetic liability to major depression: genetic control of exposure to the environment? Psychol Med 1997; 27: 539–547.

    CAS  PubMed  Article  Google Scholar 

  38. 38

    Paykel E, Abbott R, Jenkins R, Brugha T, Meltzer H . Urban–rural mental health differences in great britain: findings from the national morbidity survey. Int Rev Psychiatry 2003; 15: 97–107.

    CAS  PubMed  Article  Google Scholar 

  39. 39

    Wang JL . Rural-urban differences in the prevalence of major depression and associated impairment. Soc Psychiatry Psychiatr Epidemiol 2004; 39: 19–25.

    PubMed  Article  Google Scholar 

  40. 40

    Whitfield JB, Zhu G, Heath AC, Martin NG . Choice of residential location: chance, family influences, or genes? Twin Res Hum Genet 2005; 8: 22–26.

    PubMed  Article  Google Scholar 

  41. 41

    Willemsen G, Posthuma D, Boomsma DI . Environmental factors determine where the Dutch live: results from the Netherlands twin register. Twin Res Hum Genet 2005; 8: 312–317.

    PubMed  Article  Google Scholar 

  42. 42

    Kendler KS, Kuhn JW, Vittum J, Prescott CA, Riley B . The interaction of stressful life events and a serotonin transporter polymorphism in the prediction of episodes of major depression: a replication. Arch Gen Psychiatry 2005; 62: 529–535.

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  43. 43

    Wilhelm K, Mitchell PB, Niven H, Finch A, Wedgwood L, Scimone A et al. Life events, first depression onset and the serotonin transporter gene. Br J Psychiatry 2006; 188: 210–215.

    PubMed  PubMed Central  Article  Google Scholar 

  44. 44

    Moffitt TE, Caspi A, Rutter M . Strategy for investigating interactions between measured genes and measured environments. Arch Gen Psychiatry 2005; 62: 473–481.

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  45. 45

    Henikoff S, Matzke MA . Exploring and explaining epigenetic effects. Trends Genet 1997; 13: 293–295.

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  46. 46

    Jaenisch R, Bird A . Epigenetic regulation of gene expression: how the genome integrates intrinsic and environmental signals. Nat Genet 2003; 33: 245–254.

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  47. 47

    Jenuwein T, Allis CD . Translating the histone code. Science 2001; 293: 1074–1080.

    CAS  PubMed  Article  Google Scholar 

  48. 48

    Jones PL, Veenstra GJ, Wade PA, Vermaak D, Kass SU, Landsberger N et al. Methylated DNA and MeCP2 recruit histone deacetylase to repress transcription. Nat Genet 1998; 19: 187–191.

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  49. 49

    Robertson KD, Wolffe AP . DNA methylation in health and disease. Nat Rev Genet 2000; 1: 11–19.

    CAS  Article  Google Scholar 

  50. 50

    Riggs AD, Xiong Z, Wang L, LeBon JM . Methylation dynamics, epigenetic fidelity and X chromosome structure. Novartis Found Symp. 1998; 214: 214–225.

    CAS  PubMed  Google Scholar 

  51. 51

    Ushijima T, Watanabe N, Okochi E, Kaneda A, Sugimura T, Miyamoto K . Fidelity of the methylation pattern and its variation in the genome. Genome Res 2003; 13: 868–874.

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  52. 52

    Rakyan VK, Preis J, Morgan HD, Whitelaw E . The marks, mechanisms and memory of epigenetic states in mammals. Biochem J 2001; 356: 1–10.

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  53. 53

    Richards EJ . Inherited epigenetic variation – revisiting soft inheritance. Nat Rev Genet 2006; 7: 395–401.

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  54. 54

    Nicholls RD, Knepper JL . Genome organization, function, and imprinting in Prader–Willi and Angelman syndromes. Annu Rev Genomics Hum Genet 2001; 2: 153–175.

    CAS  PubMed  Article  Google Scholar 

  55. 55

    Xu GL, Bestor TH, Bourc'his D, Hsieh CL, Tommerup N, Bugge M et al. chromosome instability and immunodeficiency syndrome caused by mutations in a DNA methyltransferase gene. Nature 1999; 402: 187–191.

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  56. 56

    Amir RE, Zoghbi HY . Rett Syndrome: Methyl-CpG-binding protein 2 mutations and phenotype–genotype correlations. Am J Med Genet 2000; 97: 147–152.

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  57. 57

    Gibbons RJ, McDowell TL, Raman S, O'Rourke DM, Garrick D, Ayyub H et al. Mutations in ATRX, encoding a SWI/SNF-like protein, cause diverse changes in the pattern of DNA methylation. Nat Genet 2000; 24: 368–371.

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  58. 58

    Tapscott SJ, Klesert TR, Widrow RJ, Stoger R, Laird CD . Fragile-X syndrome and myotonic dystrophy: parallels and paradoxes. Curr Opin Genet Dev 1998; 8: 245–253.

    CAS  PubMed  Article  Google Scholar 

  59. 59

    Jones PA, Baylin SB . The fundamental role of epigenetic events in cancer. Nat Rev Genet 2002; 3: 415–428.

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  60. 60

    Wong AH, Gottesman II, Petronis A . Phenotypic differences in genetically identical organisms: the epigenetic perspective. Hum Mol Genet 2005; 14: R11–R18.

    CAS  PubMed  Article  Google Scholar 

  61. 61

    Kato T, Iwamoto K, Kakiuchi C, Kuratomi G, Okazaki Y . Genetic or epigenetic difference causing discordance between monozygotic twins as a clue to molecular basis of mental disorders. Mol Psychiatry 2005; 10: 622–630.

    CAS  PubMed  Article  PubMed Central  Google Scholar 

  62. 62

    Fraga MF, Ballestar E, Paz MF, Ropero S, Setien F, Ballestar ML et al. Epigenetic differences arise during the lifetime of monozygotic twins. Proc Natl Acad Sci USA 2005; 102: 10604–10609.

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  63. 63

    Petronis A, Gottesman II, Kan P, Kennedy JL, Basile VS, Paterson AD et al. Monozygotic twins exhibit numerous epigenetic differences: clues to twin discordance? Schizophr Bull 2003; 29: 169–178.

    PubMed  PubMed Central  Article  Google Scholar 

  64. 64

    Mill J, Dempster E, Caspi A, Williams B, Moffitt T, Craig I . Evidence for monozygotic twin (MZ) discordance in methylation level at two CpG Sites in the promoter region of the Catechol-O-Methyltransferase (COMT) gene. Am J Med Genet B Neuropsychiatr Genet 2006; 141: 421–425.

    Article  CAS  Google Scholar 

  65. 65

    Rakyan VK, Blewitt ME, Druker R, Preis JI, Whitelaw E . Metastable epialleles in mammals. Trends Genet 2002; 18: 348–351.

    CAS  PubMed  Article  Google Scholar 

  66. 66

    Petronis A . The origin of schizophrenia: genetic thesis, epigenetic antithesis, and resolving synthesis. Biol Psychiatry 2004; 55: 965–970.

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  67. 67

    Sutherland JE, Costa M . Epigenetics and the environment. Ann N Y Acad Sci 2003; 983: 151–160.

    CAS  PubMed  Article  PubMed Central  Google Scholar 

  68. 68

    Waterland RA, Jirtle RL . Early nutrition, epigenetic changes at transposons and imprinted genes, and enhanced susceptibility to adult chronic diseases. Nutrition 2004; 20: 63–68.

    CAS  PubMed  Article  Google Scholar 

  69. 69

    Cooney CA, Dave AA, Wolff GL . Maternal methyl supplements in mice affect epigenetic variation and DNA methylation of offspring. J Nutr 2002; 132: 2393S–2400S.

    CAS  PubMed  Article  Google Scholar 

  70. 70

    Numachi Y, Yoshida S, Yamashita M, Fujiyama K, Naka M, Matsuoka H et al. Psychostimulant alters expression of DNA methyltransferase MRNA in the rat brain. Ann N Y Acad Sci 2004; 1025: 102–109.

    CAS  PubMed  Article  PubMed Central  Google Scholar 

  71. 71

    Weaver IC, Cervoni N, Champagne FA, D'Alessio AC, Sharma S, Seckl JR et al. Epigenetic programming by maternal behavior. Nat Neurosci 2004; 7: 847–854.

    CAS  Article  PubMed  Google Scholar 

  72. 72

    Kendler KS, Gatz M, Gardner CO, Pedersen NL . A Swedish national twin study of lifetime major depression. Am J Psychiatry 2006; 163: 109–114.

    PubMed  PubMed Central  Article  Google Scholar 

  73. 73

    Vaillant GE, Batalden M, Orav J, Roston D, Barrett JE . Evidence for a possibly X-linked trait related to affective illness. Aust N Z J Psychiatry 2005; 39: 730–735.

    PubMed  Article  PubMed Central  Google Scholar 

  74. 74

    Thomson PA, Wray NR, Thomson AM, Dunbar DR, Grassie MA, Condie A et al. Sex-specific association between bipolar affective disorder in women and GPR50, an X-linked orphan G protein-coupled receptor. Mol Psychiatry 2005; 10: 470–478.

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  75. 75

    Faraone SV, Lyons MJ, Tsuang MT . Sex differences in affective disorder: genetic transmission. Genet Epidemiol 1987; 4: 331–343.

    CAS  PubMed  Article  PubMed Central  Google Scholar 

  76. 76

    Craig IW, Harper E, Loat CS . The genetic basis for sex differences in human behaviour: role of the sex chromosomes. Ann Hum Genet 2004; 68: 269–284.

    CAS  PubMed  Article  PubMed Central  Google Scholar 

  77. 77

    Plenge RM, Stevenson RA, Lubs HA, Schwartz CE, Willard HF . Skewed X-chromosome inactivation is a common feature of X-linked mental retardation disorders. Am J Hum Genet 2002; 71: 168–173.

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  78. 78

    Willard HF . The sex chromosomes and X chromosome inactivation. In: Scriver CR, Beaudet AL, Sly WS, Valle D, Childs B, Vogelstein B (eds). The Metabolic and Molecular Bases of Inherited Disease. McGraw-Hill: New York, 2000, pp 1191–1221.

    Google Scholar 

  79. 79

    Belmont JW . Genetic control of X inactivation and processes leading to X-inactivation skewing. Am J Hum Genet 1996; 58: 1101–1108.

    CAS  PubMed  PubMed Central  Google Scholar 

  80. 80

    Loat CS, Asbury K, Galsworthy MJ, Plomin R, Craig IW . X inactivation as a source of behavioural differences in monozygotic female twins. Twin Res 2004; 7: 54–61.

    PubMed  Article  Google Scholar 

  81. 81

    Carrel L, Willard HF . X-inactivation profile reveals extensive variability in X-linked gene expression in females. Nature 2005; 434: 400–404.

    CAS  Article  PubMed  Google Scholar 

  82. 82

    Zubenko GS, Maher B, Hughes III HB, Zubenko WN, Stiffler JS, Kaplan BB et al. L. genome-wide linkage survey for genetic loci that influence the development of depressive disorders in families with recurrent, early onset, major depression. Am J Med Genet B Neuropsychiatr Genet 2003; 123: 1–18.

    Article  Google Scholar 

  83. 83

    Csordas A, Puschendorf B, Grunicke H . Increased acetylation of histones at an early stage of oestradiol-mediated gene activation in the liver of immature chicks. J Steroid Biochem 1986; 24: 437–442.

    CAS  PubMed  Article  Google Scholar 

  84. 84

    Pasqualini JR, Mercat P, Giambiagi N . Histone acetylation decreased by estradiol in the MCF-7 human mammary cancer cell line. Breast Cancer Res Treat 1989; 14: 101–105.

    CAS  PubMed  Article  PubMed Central  Google Scholar 

  85. 85

    Yokomori N, Moore R, Negishi M . Sexually dimorphic DNA demethylation in the promoter of the Slp (sex-limited protein) gene in mouse liver. Proc Natl Acad Sci USA 1995; 92: 1302–1306.

    CAS  PubMed  Article  PubMed Central  Google Scholar 

  86. 86

    Saluz HP, Jiricny J, Jost JP . Genomic Sequencing reveals a positive correlation between the kinetics of strand-specific DNA Demethylation of the overlapping estradiol/glucocorticoid-receptor binding sites and the rate of avian vitellogenin MRNA synthesis. Proc Natl Acad Sci USA 1986; 83: 7167–7171.

    CAS  PubMed  Article  PubMed Central  Google Scholar 

  87. 87

    McMahon FJ, Stine OC, Meyers DA, Simpson SG, DePaulo JR . Patterns of maternal transmission in bipolar affective disorder. Am J Hum Genet 1995; 56: 1277–1286.

    CAS  PubMed  PubMed Central  Google Scholar 

  88. 88

    Zill P, Engel R, Baghai TC, Zwanzger P, Schule C, Minov C et al. Analysis of polymorphisms in the olfactory G-Protein Golf in major depression. Psychiatr Genet 2002; 12: 17–22.

    PubMed  Article  PubMed Central  Google Scholar 

  89. 89

    Schiffer HH, Heinemann SF . Association of the human kainate receptor GluR7 Gene (GRIK3) with recurrent major depressive disorder. Am J Med Genet B Neuropsychiatr Genet 2006 [E-pub ahead of print].

  90. 90

    Davies W, Isles AR, Wilkinson LS . Imprinted gene expression in the brain. Neurosci Biobehav Rev 2005; 29: 421–430.

    CAS  PubMed  Article  Google Scholar 

  91. 91

    Preis JI, Downes M, Oates NA, Rasko JE, Whitelaw E . Sensitive flow cytometric analysis reveals a novel type of parent-of-origin effect in the mouse genome. Curr Biol 2003; 13: 955–959.

    CAS  PubMed  Article  Google Scholar 

  92. 92

    Sakatani T, Wei M, Katoh M, Okita C, Wada D, Mitsuya K et al. Epigenetic heterogeneity at imprinted loci in normal populations. Biochem Biophys Res Commun 2001; 283: 1124–1130.

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  93. 93

    Lewis SJ, Lawlor DA, Davey Smith G, Araya R, Timpson N, Day IN et al. The thermolabile variant of MTHFR is associated with depression in the British women's heart and health study and a meta-analysis. Mol Psychiatry 2006; 11: 352–360.

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  94. 94

    Zogel C, Bohringer S, Gross S, Varon R, Buiting K, Horsthemke B . Identification of Cis- and trans-acting factors possibly modifying the risk of epimutations on chromosome 15. Eur J Hum Genet 2006; 14: 752–758.

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  95. 95

    Polesskaya OO, Aston C, Sokolov BP . Allele C-specific methylation of the 5-HT2A receptor gene: evidence for correlation with its expression and expression of DNA methylase DNMT1. J Neurosci Res 2006; 83: 362–373.

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  96. 96

    Dempster EL, Mill J, Craig IW, Collier DA . The quantification of COMT MRNA in post mortem cerebellum tissue: diagnosis, genotype, methylation and expression. BMC Med Genet 2006; 7: 10.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  97. 97

    Flanagan JM, Popendikyte V, Pozdniakovaite N, Sobolev M, Assadzadeh A, Schumacher A et al. Intra- and interindividual epigenetic variation in human germ cells. Am J Hum Genet 2006; 79: 67–84.

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  98. 98

    Frommer M, McDonald LE, Millar DS, Collis CM, Watt F, Grigg GW et al. A genomic sequencing protocol that yields a positive display of 5-methylcytosine residues in individual DNA strands. Proc Natl Acad Sci USA 1992; 89: 1827–1831.

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  99. 99

    Paul CL, Clark SJ . Cytosine methylation: quantitation by automated genomic sequencing and GENESCAN analysis. Biotechniques 1996; 21: 126–133.

    CAS  PubMed  Article  Google Scholar 

  100. 100

    Tost J, Dunker J, Gut IG . Analysis and quantification of multiple methylation variable positions in CpG islands by pyrosequencing. Biotechniques 2003; 35: 152–156.

    CAS  PubMed  Article  Google Scholar 

  101. 101

    Mill J, Yazdanpanah S, Gukel E, Ziegler S, Kaminsky Z, Petronis A . Whole genome amplification of sodium bisulfite-treated DNA allows the accurate estimate of methylated cytosine density in limited DNA resources. Biotechniques 2006; 41: 603–607.

    CAS  PubMed  Article  Google Scholar 

  102. 102

    Yatabe Y, Tavare S, Shibata D . Investigating stem cells in human colon by using methylation patterns. Proc Natl Acad Sci USA 2001; 98: 10839–10844.

    CAS  PubMed  Article  Google Scholar 

  103. 103

    Schumacher A, Kapranov P, Kaminsky Z, Flanagan J, Assadzadeh A, Yau P et al. Microarray-based DNA methylation profiling: technology and applications. Nucleic Acids Res 2006; 34: 528–542.

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  104. 104

    Zhou D, Qiao W, Wan Y, Lu Z . Microarray-based methylation analysis using dual-color fluorescence hybridization. J Biochem Biophys Methods 2006; 66: 33–43.

    CAS  PubMed  Article  Google Scholar 

  105. 105

    Mirnics K . Microarrays in brain research: the good, the bad and the ugly. Nat Rev Neurosci 2001; 2: 444–447.

    CAS  PubMed  Article  Google Scholar 

  106. 106

    O'Neill LP, VerMilyea MD, Turner BM . Epigenetic characterization of the early embryo with a chromatin immunoprecipitation protocol applicable to small cell populations. Nat Genet 2006; 38: 835–841.

    CAS  PubMed  Article  Google Scholar 

  107. 107

    Huang HS, Matevossian A, Jiang Y, Akbarian S . Chromatin immunoprecipitation in postmortem brain. J Neurosci Methods 2006; 156: 284–292.

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  108. 108

    Veldic M, Guidotti A, Maloku E, Davis JM, Costa E . In psychosis, cortical interneurons overexpress DNA-methyltransferase 1. Proc Natl Acad Sci USA 2005; 102: 2152–2157.

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  109. 109

    Bahn S, Augood SJ, Ryan M, Standaert DG, Starkey M, Emson PC . Gene expression profiling in the post-mortem human brain – no cause for dismay. J Chem Neuroanat 2001; 22: 79–94.

    CAS  Article  Google Scholar 

  110. 110

    Cui H, Cruz-Correa M, Giardiello FM, Hutcheon DF, Kafonek DR, Brandenburg S et al. Loss of IGF2 imprinting: a potential marker of colorectal cancer risk. Science 2003; 299: 1753–1755.

    CAS  PubMed  Article  PubMed Central  Google Scholar 

  111. 111

    Suter CM, Martin DI, Ward RL . Germline epimutation of MLH1 in individuals with multiple cancers. Nat Genet 2004; 36: 497–501.

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  112. 112

    Weksberg R, Shuman C, Caluseriu O, Smith AC, Fei YL, Nishikawa J et al. Discordant KCNQ1OT1 imprinting in sets of monozygotic twins discordant for beckwith-wiedemann syndrome. Hum Mol Genet 2002; 11: 1317–1325.

    CAS  PubMed  Article  PubMed Central  Google Scholar 

  113. 113

    Eckhardt F, Beck S, Gut IG, Berlin K . Future potential of the human epigenome project. Expert Rev Mol Diagn 2004; 4: 609–618.

    CAS  PubMed  Article  PubMed Central  Google Scholar 

  114. 114

    Rakyan VK, Hildmann T, Novik KL, Lewin J, Tost J, Cox AV et al. DNA methylation profiling of the human major histocompatibility complex: a pilot study for the human epigenome project. PLoS.Biol 2004; 2: e405.

    PubMed  PubMed Central  Article  CAS  Google Scholar 

  115. 115

    Eckhardt F, Lewin J, Cortese R, Rakyan VK, Attwood J, Burger M et al. DNA methylation profiling of human chromosomes 6, 20 and 22. Nat Genet 2006; 38: 1378–1385.

    CAS  PubMed  PubMed Central  Article  Google Scholar 

  116. 116

    Kalow W, Tang BK, Endrenyi L . Hypothesis: comparisons of inter- and intra-individual variations can substitute for twin studies in drug research. Pharmacogenetics 1998; 8: 283–289.

    CAS  PubMed  Article  Google Scholar 

Download references


Research in the Krembil Family Epigenetics laboratory is supported by the National Institute of Mental Health (R01MH074127-01), the Canadian Institutes of Health Research (CIHR), and NARSAD. JM is supported by a CIHR postdoctoral fellowship.

Author information



Corresponding author

Correspondence to A Petronis.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Mill, J., Petronis, A. Molecular studies of major depressive disorder: the epigenetic perspective. Mol Psychiatry 12, 799–814 (2007).

Download citation


  • depression
  • epigenetics
  • methylation
  • genetics
  • environment
  • sex-effects

Further reading


Quick links